Semiconductor testing method, electronic device, storage medium, and program product

Abstract

The application provides a semiconductor testing method, an electronic device, a storage medium and a program product, comprising: obtaining flatness detection information of a product; obtaining surface defect detection information of the product; and obtaining flatness classification results and / or surface defect classification results according to the flatness detection information and the surface defect detection information. The application combines flatness detection and surface defect detection, and can obtain flatness classification results and / or surface defect classification results. The comprehensive testing method can fully consider the flatness and surface defect of the semiconductor product, provide comprehensive quality evaluation, and provide support for product improvement.

Description

Technical Field

[0001] This application relates to the fields of semiconductor testing and artificial intelligence, and in particular to semiconductor testing methods, electronic devices, computer-readable storage media, and computer program products. Background Technology

[0002] In the semiconductor industry, product testing is a crucial step in ensuring product quality and performance. Flatness testing and surface defect detection are commonly used testing methods.

[0003] TTV (Transient Voltage Transmission) is a wafer flatness inspection method that generates a three-dimensional image of a semiconductor sample. By utilizing oblique incidence laser interference, TTV can measure the flatness of the wafer surface and generate a three-dimensional surface topography image. However, existing TTV methods can only characterize the flatness of the sample surface and cannot directly characterize the location and size of dust or particles on the sample surface. These particles or dust may exist on the sample surface, affecting the measurement results.

[0004] The Candela surface defect detection method is a surface defect detection method that uses the Candela surface defect detector to detect impurities and defects on the surface of semiconductor samples. The two-dimensional images generated by Candela can show the location of impurities and defects on the sample surface. However, these images do not visually characterize the morphology of the sample surface.

[0005] Therefore, both the existing TTV detection method and the Candela surface defect detection method have certain limitations.

[0006] Based on this, this application provides semiconductor testing methods, electronic devices, computer-readable storage media, and computer program products to improve the prior art. Summary of the Invention

[0007] The purpose of this application is to provide semiconductor testing methods, electronic devices, computer-readable storage media, and computer program products that combine two testing methods: flatness detection and surface defect detection. This approach fully considers the flatness and surface defects of semiconductor products and provides a comprehensive quality assessment.

[0008] The objective of this application is achieved through the following technical solution:

[0009] This application provides a semiconductor testing method, including:

[0010] Obtain flatness test information for the product;

[0011] Obtain surface defect detection information of the product;

[0012] Based on the flatness detection information and the surface defect detection information, obtain the flatness classification result and / or the surface defect classification result.

[0013] The beneficial effects of this technical solution are as follows: This method combines two testing methods—flatness detection and surface defect detection—to comprehensively evaluate the quality and performance of semiconductor products. By acquiring and analyzing the flatness and surface defect detection information of the product, flatness classification results and / or surface defect classification results can be obtained. This provides a comprehensive understanding of the surface condition of the semiconductor product. Flatness detection provides information about the smoothness of the sample surface, while surface defect detection can detect and locate potential impurities and defects. By combining these two types of information, the quality of semiconductor products can be evaluated more accurately, and potential problems can be intuitively identified. Furthermore, this method can provide valuable data support for subsequent process improvements and product optimization. By analyzing the flatness and surface defect classification results, problems in the product manufacturing process can be identified, and corresponding improvement measures can be taken. This helps improve the manufacturing quality and performance of semiconductor products. In summary, this comprehensive testing method can fully consider the flatness and surface defect conditions of semiconductor products, provide a comprehensive quality assessment, and offer strong support for product improvement, which is of great significance for the development of the semiconductor industry and the improvement of product quality.

[0014] In some possible implementations, the flatness detection information includes a three-dimensional flatness image and its corresponding flatness data, wherein the flatness data is used to indicate the flatness of each unit area;

[0015] The surface defect detection information includes two-dimensional particle detection images and their corresponding defect type data and defect location data;

[0016] The flatness classification result includes the flatness category of each unit area in the flatness detection information;

[0017] The surface defect classification result includes the defect category of each target defect in the surface defect detection information.

[0018] The beneficial effects of this technical solution are as follows: By comprehensively utilizing three-dimensional flatness images and corresponding flatness data, as well as two-dimensional particle detection images and corresponding defect type and location data, this method can provide more comprehensive and detailed flatness and surface defect information. Three-dimensional flatness images and data can intuitively display the flatness of the sample surface, and the flatness level of each unit area can be quantified through the flatness data. This information can more accurately assess the surface quality of the sample and provide important basis for product process improvement. Meanwhile, two-dimensional particle detection images and their corresponding defect type and location data can intuitively display the particles and defects on the sample surface. Through image analysis and data processing, defects can be classified and their specific locations on the sample surface can be determined. This helps to promptly identify and locate potential problems, providing accurate guidance for quality control and repair work. By obtaining the flatness classification results and surface defect classification results, operators can directly understand the flatness category of the sample and the defect category of each defect. Such classification results facilitate comparative analysis, statistical summarization, and problem localization, providing strong support for process improvement and quality control.

[0019] In some possible implementations, the method further includes:

[0020] Based on the flatness category of each unit area in the flatness detection information, image encoding data for each unit area is obtained, and a first display image of the product is generated based on the image encoding data of each unit area and the three-dimensional flatness image. In the first display image, the set of display parameters corresponding to unit areas of different flatness categories is different; and / or,

[0021] Based on the defect type data and defect location data in the surface defect detection information, the area of ​​each defect category in the three-dimensional flatness image is obtained; based on the area of ​​each defect category in the three-dimensional flatness image, the image encoding data of the area of ​​each defect category is obtained; based on the image encoding data of the area of ​​each defect category and the three-dimensional flatness image, a second display image of the product is generated, in which the set of display parameters corresponding to the area of ​​different defect categories is different.

[0022] The beneficial effects of this technical solution are as follows: On the one hand, based on the flatness category of each unit area, image-coded data for each unit area is obtained, and a first display image of the product is generated using the three-dimensional flatness image. In the first display image, the unit area of ​​different flatness categories will correspond to different sets of display parameters. On the other hand, based on the defect type data and defect location data in the surface defect detection information, the area of ​​each defect category in the three-dimensional flatness image is obtained. Then, based on the area of ​​each defect category in the three-dimensional flatness image, image-coded data of the area of ​​each defect category is obtained. Finally, a second display image of the product is generated using these image-coded data and the three-dimensional flatness image. In the second display image, the area of ​​different defect categories will correspond to different sets of display parameters. The advantages of this approach are twofold: First, generating a first display image visually shows the unit area of ​​different flatness categories and uses different sets of display parameters to distinguish them. This display method helps operators quickly understand the flatness distribution of the sample and more accurately judge the surface quality of the product. Second, generating a second display image clearly shows the area of ​​different defect categories and uses different sets of display parameters to distinguish them. This display method makes defect location and analysis more intuitive and convenient, facilitating timely quality control and problem handling by operators. In summary, generating different display images and distinguishing flatness and defect categories based on different sets of display parameters provides more intuitive and detailed surface quality information. This contributes to the manufacturing and quality control of semiconductor products, improving product reliability and performance.

[0023] In some possible implementations, the product has positioning edges;

[0024] The step of obtaining flatness classification results and / or surface defect classification results based on the flatness detection information and the surface defect detection information includes:

[0025] Based on the flatness data, find one or more unit areas whose flatness is less than the preset flatness.

[0026] For each identified unit area, based on the positioning edge, the corresponding defect type data is obtained from the surface defect detection information to detect whether there is a target defect within the unit area;

[0027] Set the flatness category of the unit area with the target defect to the first flatness category, and set the defect category of the target defect within the unit area of ​​the first flatness category to the first defect category.

[0028] Set the flatness category per unit area for defects without a target to the second flatness category.

[0029] The beneficial effects of this technical solution are as follows: First, based on flatness data, one or more unit areas in the product with flatness less than a preset flatness standard are identified, indicating potential flatness deficiencies. Then, for each identified unit area, based on the positioning edges, corresponding defect type data is obtained from surface defect detection information to detect the presence of a target defect within that unit area. Analysis of the defect type data determines the existence of a target defect. Next, the flatness category of the unit area containing the target defect is set as the first flatness category, and the defect category of the target defect within the unit area of ​​the first flatness category is set as the first defect category. Simultaneously, the flatness category of the unit area without a target defect is set as the second flatness category. The advantages of this approach are that by combining flatness detection and surface defect detection, this method can identify unit areas with substandard flatness based on a preset flatness standard and determine their location using positioning edges; by analyzing surface defect detection information, the presence of a target defect within these unit areas can be detected and determined; and by setting flatness and defect classifications, the quality status of the product can be more accurately assessed.

[0030] In some possible implementations, after setting the defect category of the target defect within a unit area of ​​the first flatness category as the first defect category, the method further includes:

[0031] For each unit area of ​​the first flatness category, perform the following processing:

[0032] Based on the positioning edge, the size information of the target defect within the unit area is obtained from the surface defect detection information, and it is determined whether the size of the target defect within the unit area is greater than a preset size.

[0033] When the size of the target defect within the unit area is greater than the preset size: the flatness category of the unit area is adjusted to the third flatness category, and / or the defect category of the target defect within the unit area is adjusted to the second defect category;

[0034] When the size of the target defect within the unit area is not greater than the preset size: the flatness category of the unit area is adjusted to the fourth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the third defect category.

[0035] The beneficial effects of this technical solution are as follows: by further judging and processing the size of the target defect within a unit area of ​​the first flatness category, the unit area can be more accurately classified and evaluated. If the size of the target defect exceeds a preset size, the flatness category will be adjusted to the third flatness category, and the defect category will be adjusted accordingly to the second defect category. This processing method helps to more accurately identify serious defects so that they can be handled and repaired in a timely manner. For cases where the target defect size does not exceed the preset size, the flatness category will be adjusted to the fourth flatness category, and the defect category will be adjusted to the third defect category. This adjustment reflects that although a defect exists, its size is within an acceptable range and has a minimal impact on product performance. In summary, by adjusting the flatness category of a unit area and the corresponding defect category of the target defect based on the size of the target defect, this technical solution can more accurately classify and evaluate products and take corresponding measures for different situations, further improving the quality control level of semiconductor products.

[0036] In some possible implementations, the aspect ratio gradually decreases from the first preset aspect ratio to the Nth preset aspect ratio, where N is a positive integer;

[0037] The method further includes:

[0038] For each unit area of ​​either the first or third flatness category, perform the following processing:

[0039] Obtain the aspect ratio corresponding to the unit area;

[0040] When the aspect ratio is greater than the first preset aspect ratio: the flatness category of the unit area is adjusted to the fifth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the fourth defect category;

[0041] When the aspect ratio is greater than the kth preset aspect ratio and not greater than the (k-1)th preset aspect ratio: the flatness category of the unit area is adjusted to the (k+4)th flatness category, and / or the defect category of the target defect within the unit area is adjusted to the (k+3)th defect category, where k is an integer greater than 1 and not greater than N.

[0042] The beneficial effects of this technical solution are as follows: For each unit area of ​​the first or third flatness category, the following processing is performed: First, the aspect ratio corresponding to the unit area is obtained. Then, adjustments are made based on a comparison between the aspect ratio and a preset aspect ratio. When the aspect ratio is greater than the first preset aspect ratio, the flatness category of the unit area is adjusted to the fifth flatness category, and the defect category of the target defect within the unit area of ​​the fifth flatness category is adjusted to the fourth defect category. This adjustment method can reflect that an abnormal deviation of the aspect ratio from the expected range may adversely affect the product performance. In the subsequent processing, adjustments are made based on a comparison between the aspect ratio and each preset aspect ratio. When the aspect ratio is greater than the kth preset aspect ratio but not greater than the (k-1)th preset aspect ratio, the flatness category of the unit area is adjusted to the (k+4)th flatness category, and the defect category of the target defect within the unit area of ​​the (k+4)th flatness category is adjusted to the (k+3)th defect category. This processing method classifies and adjusts according to the specific situation of the aspect ratio to more comprehensively reflect the characteristics of the product surface. The advantage of doing this is that by considering the relationship between the aspect ratio and the preset aspect ratio, the unit area can be further classified and adjusted. This approach can more accurately reflect the situation of aspect ratio abnormalities, help operators better judge the characteristics and potential defects of the product surface, and by adjusting the flatness category and defect category, the changes in surface features can be described and located more precisely, providing more targeted guidance for product manufacturing and quality control.

[0043] In some possible implementations, obtaining the aspect ratio corresponding to the unit area includes:

[0044] Based on the flatness detection information, the lateral distance and longitudinal distance of the undulating region corresponding to the unit area are obtained, wherein the lateral distance between the nearest valley points at both ends of the undulating region is taken as the lateral distance of the undulating region, and the longitudinal distance between the peak point of the undulating region and its nearest valley point is taken as the longitudinal distance of the undulating region.

[0045] The ratio of the longitudinal distance and the lateral distance of the undulating region is calculated and used as the aspect ratio corresponding to the unit area.

[0046] The beneficial effects of this technical solution are as follows: First, based on flatness detection information, the lateral and longitudinal distances of the undulating regions per unit area are obtained. The lateral distance of the undulating region refers to the lateral distance between the nearest valleys at both ends of the undulating region, while the longitudinal distance refers to the longitudinal distance between the peak of the undulating region and its nearest valley. Then, the quotient between the longitudinal and lateral distances of the undulating regions is calculated. This quotient is the aspect ratio per unit area. Through these steps, the aspect ratio per unit area can be obtained to reflect the characteristics of the surface undulating regions. The aspect ratio is calculated by comparing the longitudinal and lateral distances of the undulating regions to determine the relative degree of surface undulation characteristics. The advantage of this approach is that by calculating the aspect ratio of the undulating regions, the surface undulation characteristics can be objectively evaluated. The aspect ratio value can quantitatively describe the degree and direction of undulation within a unit area. This evaluation method is of great significance for the quality control and manufacturing process optimization of semiconductor products. Through the calculation of the aspect ratio, operators can better understand and judge the surface undulation of the product and take targeted measures to improve product quality and performance.

[0047] In some possible implementations, the method further includes:

[0048] Based on the flatness detection information, the flatness category of each unit area, and the defect category of each target defect, a third display image of the product is generated. In the third display image, the set of display parameters corresponding to the unit area of ​​different flatness categories is different, and the set of display parameters corresponding to the target defects of different defect categories is different.

[0049] The beneficial effects of this technical solution are as follows: A third display image of the product is generated based on flatness detection information, the flatness category of each unit area, and the defect category of each target defect. In this third display image, the set of display parameters corresponding to the unit area for different flatness categories is different, and the set of display parameters corresponding to the target defects for different defect categories is also different. By generating the third display image, the flatness and defect information of the product surface can be presented intuitively. The unit area for different flatness categories is assigned a corresponding set of display parameters based on its flatness, thus presenting different representations in the image. Similarly, target defects for different defect categories are assigned a corresponding set of display parameters based on their defect types, clearly displaying these defects in the third display image. Through this display method, operators can intuitively understand the flatness of the product surface and the types of defects present, thereby better enabling quality control and decision-making. The third display image provides visualized information, helping operators to conduct comprehensive evaluation and analysis of the product, and can identify and locate different types of defects based on the set of display parameters in the image. Therefore, by generating a third-party display image, the flatness and defect information of the product can be effectively displayed, providing an intuitive visual reference, which has obvious advantages for product quality control and problem diagnosis.

[0050] Secondly, this application provides an electronic device including a memory and at least one processor, the memory storing a computer program, and the at least one processor being configured to execute the computer program to perform the following steps:

[0051] Obtain flatness test information for the product;

[0052] Obtain surface defect detection information of the product;

[0053] Based on the flatness detection information and the surface defect detection information, obtain the flatness classification result and / or the surface defect classification result.

[0054] Thirdly, this application provides a computer-readable storage medium storing a computer program that, when executed by at least one processor, implements the steps of any of the above-described methods or the functions of any of the above-described electronic devices.

[0055] Fourthly, this application provides a computer program product comprising a computer program that, when executed by at least one processor, implements the steps of any of the above methods or the functions of any of the above electronic devices. Attached Figure Description

[0056] This application will be further described below with reference to the accompanying drawings and specific embodiments.

[0057] Figure 1 This is a three-dimensional flatness image obtained by flatness detection of a wafer, provided in an embodiment of this application.

[0058] Figure 2 yes Figure 1 The corresponding two-dimensional local flatness image.

[0059] Figure 3 yes Figure 1 The corresponding two-dimensional contour map.

[0060] Figure 4 This application provides a method for obtaining sample surface information by generating a scattering signal on the wafer surface through oblique incident laser light.

[0061] Figure 5 This application provides a two-dimensional particle detection image obtained by using a Candela surface defect detector to detect surface defects on a wafer.

[0062] Figure 6 This is a schematic flowchart of a semiconductor testing method provided in an embodiment of this application.

[0063] Figure 7 This is a schematic diagram of the horizontal and vertical distances corresponding to an undulating region provided in an embodiment of this application.

[0064] Figure 8 This is a schematic diagram of the peaks and valleys corresponding to an undulating region provided in an embodiment of this application.

[0065] Figure 9 This is a schematic diagram of a process for obtaining flatness classification results and / or surface defect classification results provided in an embodiment of this application.

[0066] Figure 10 This is a structural block diagram of an electronic device provided in an embodiment of this application.

[0067] Figure 11 This is a schematic diagram of the structure of a computer program product provided in an embodiment of this application. Detailed Implementation

[0068] The technical solutions in this application will be described below with reference to the accompanying drawings and specific embodiments. It should be noted that, without conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments.

[0069] In this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any implementation or design described as "exemplary" or "for example" in this application should not be construed as being better or more advantageous than other implementations or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0070] The descriptions of "first," "second," etc., appearing in the embodiments of this application are for illustrative purposes and to distinguish the objects being described. They have no order and do not indicate any special limitation on the quantity in the embodiments of this application, nor do they constitute any limitation on the embodiments of this application.

[0071] The technical field and related terms of the embodiments of this application are briefly described below.

[0072] Artificial Intelligence (AI) is the theory, methods, technology, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. AI studies the design principles and implementation methods of various intelligent machines, enabling them to possess perception, reasoning, and decision-making capabilities. AI technology is a comprehensive discipline involving a wide range of fields, encompassing both hardware and software technologies. Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly include computer vision, speech processing, natural language processing, as well as machine learning / deep learning, autonomous driving, and intelligent transportation.

[0073] Machine Learning (ML) is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. A computer program can learn experience E given a certain type of task T and a performance metric P. If its performance on task T can be measured by P, it improves with experience E. Machine learning specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence.

[0074] Deep learning is a special type of machine learning that learns to represent the world using nested hierarchical concepts, achieving tremendous functionality and flexibility. Each concept is defined as being associated with a simpler one, while more abstract representations are computed in a less abstract manner. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and learning by demonstration.

[0075] Semiconductor products, including wafers, epitaxial wafers, integrated circuits, optoelectronic devices, discrete devices, and sensors, are widely used in the electronics and communications fields. For example, integrated circuits include operational amplifiers, voltage regulator integrated circuits, and audio integrated circuits. Optoelectronic devices include photodiodes, phototransistors, photoresistors, and optocouplers. Discrete devices include transistors, diodes, and field-effect transistors. Sensors include temperature sensors, humidity sensors, and pressure sensors.

[0076] A wafer is a chip used in the fabrication of semiconductor integrated circuits. It is called a wafer because of its circular shape. Various circuit element structures can be fabricated on the wafer to create IC products with specific electrical functions.

[0077] Semiconductor defects are common problems in the manufacturing process, potentially leading to performance degradation or complete chip failure. Therefore, semiconductor manufacturers typically need to use defect detection technologies to identify and correct these problems early on. This technology can detect potential issues at an early stage of the manufacturing process and guide subsequent process control. Semiconductor defect detection technologies include various methods such as optical inspection, electron microscopy, X-ray inspection, and scanning probe microscopy. These technologies help manufacturers locate and identify different types of defects, such as crystal defects and inhomogeneous doping, thereby guiding process control. Current semiconductor sample (or product) testing methods use different detection equipment to detect various characterization parameters, such as crystallographic, electrical, optical, and thermodynamic parameters.

[0078] In the semiconductor industry, product testing is a crucial step in ensuring product quality and performance. Flatness inspection and surface defect inspection are commonly used testing methods. The principle behind wafer flatness inspection is based on the interference of obliquely incident laser light; the interference of light produces fringes that can be used to measure distance and surface topography. Flatness is synonymous with smoothness. Current wafer flatness inspection technologies cannot detect particles.

[0079] TTV (Through-Layer Television) inspection is a wafer flatness inspection method that generates a three-dimensional image of a semiconductor sample. By utilizing oblique incidence laser interference, TTV inspection can measure the flatness of the wafer surface and generate a three-dimensional surface topography image. However, existing TTV inspection methods can only characterize the flatness of the sample surface and cannot directly characterize the location and size of dust or particles on the sample surface. These particles or dust may exist on the sample surface, affecting the measurement results. The Candela surface defect inspection method is a surface defect detection method that uses the Candela surface defect inspector to detect impurities and defects on the semiconductor sample surface. The two-dimensional images generated by Candela can show the location of impurities and defects on the sample surface. However, these images do not intuitively characterize the morphology of the sample surface.

[0080] Therefore, both the existing TTV detection method and the Candela surface defect detection method have certain limitations.

[0081] Based on this, this application provides semiconductor testing methods, electronic devices, computer-readable storage media, and computer program products to improve the prior art.

[0082] The solutions provided in this application relate to technologies such as semiconductor product manufacturing and testing, and artificial intelligence, and are specifically illustrated through the following embodiments. It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments.

[0083] (Semiconductor testing methods)

[0084] See Figures 1 to 6 , Figure 1 This application provides a three-dimensional flatness image obtained by performing flatness detection on a wafer. Figure 2 yes Figure 1 The corresponding two-dimensional local flatness image, Figure 3 yes Figure 1 The corresponding two-dimensional contour map, Figure 4 This application provides an embodiment of sample surface information obtained by generating a scattering signal on the wafer surface using an obliquely incident laser. Figure 5 This application provides a two-dimensional particle detection image obtained by detecting surface defects on a wafer using a Candela surface defect detector, as provided in an embodiment of the present application. Figure 6 This is a schematic flowchart of a semiconductor testing method provided in an embodiment of this application.

[0085] This application provides a semiconductor testing method, including:

[0086] Step S101: Obtain the flatness detection information of the product;

[0087] Step S102: Obtain surface defect detection information of the product;

[0088] Step S103: Based on the flatness detection information and surface defect detection information, obtain the flatness classification result and / or surface defect classification result.

[0089] In the embodiments of this application, the product is a semiconductor product, such as a wafer or epitaxial wafer, etc., and this application does not set any limitation on it.

[0090] Flatness inspection information refers to the relevant data and information for detecting and measuring the flatness of a product surface. Flatness refers to the degree of flatness or smoothness of a surface. Step S101 specifically includes using oblique incidence interference of a laser; the interference of light produces fringes from which distance and surface morphology can be measured. For example, by using existing flatness inspection equipment (e.g., by measuring with oblique incidence interferometry of a laser using a TTV instrument), elevation data of the wafer surface can be obtained, thereby obtaining flatness inspection information.

[0091] Surface defect detection information refers to the relevant data and information for detecting and identifying surface defects in products. Surface defects include undesirable features such as impurities, scratches, and cracks. Step S102 specifically includes detecting defects by generating a scattered signal on the sample surface through oblique incident laser light. For larger particles, the light scattering signal is stronger; smaller particles can be measured through the reflection of the oblique incident laser. For example, using an existing Candela surface defect detector, an image of the product surface can be acquired, and surface defect information can be identified through image analysis and algorithms.

[0092] The above steps S101 and S102 are not limited to being obtained by testing separately with a flatness tester and a Candela surface defect detector, but can also integrate the principle of "measuring distance and surface morphology by generating fringes through light interference" with the principle of "detecting defects by generating scattering signals on the sample surface through oblique incident laser action" to obtain flatness detection information and surface defect detection information.

[0093] However, defect detection based on the scattering signal of obliquely incident laser cannot characterize the longitudinal dimension of the defect, but can only obtain the transverse dimension. Therefore, when comparing flatness detection information and surface defect detection information, it is impossible to accurately determine whether the fluctuation in flatness is caused by surface defects.

[0094] Flatness classification results are the results of product classification based on flatness inspection information. According to different degrees of flatness, the product surface can be divided into different flatness categories. For example, the flatness of a product can be classified into excellent, good, and average. Alternatively, the flatness of a product can be classified into different categories such as caused by defects, caused by inherent unevenness, and caused by a combination of defects and inherent unevenness. Defect types include, for example, particles, impurities, scratches, and cracks.

[0095] Surface defect classification results refer to the classification of products based on surface defect detection information. According to different types of defects (particles, impurities, scratches, cracks, etc.) and their characteristics, product surface defects can be divided into different defect categories. For example, product surface defects can be classified into different categories such as causing unevenness, causing unevenness collectively, and not causing unevenness.

[0096] By acquiring flatness detection information and surface defect detection information, and performing corresponding classification analysis, the flatness and surface defects of semiconductor products can be evaluated and judged, which helps to improve product quality control and optimize the manufacturing process.

[0097] Therefore, this method combines flatness testing and surface defect detection to comprehensively evaluate the quality and performance of semiconductor products. By acquiring and analyzing flatness and surface defect detection information, flatness classification results and / or surface defect classification results can be obtained. The advantage of this approach is that it provides a comprehensive understanding of the surface condition of semiconductor products. Flatness testing provides information on the smoothness of the sample surface, while surface defect detection can detect and locate potential impurities and defects. By combining these two types of information, the quality of semiconductor products can be more accurately assessed, and potential problems can be intuitively identified. Furthermore, this method can provide valuable data support for subsequent process improvements and product optimization. By analyzing the flatness and surface defect classification results, problems in the product manufacturing process can be identified, and corresponding improvement measures can be taken. This contributes to improving the manufacturing quality and performance of semiconductor products. In conclusion, this comprehensive testing method fully considers the flatness and surface defect conditions of semiconductor products, provides a comprehensive quality assessment, and offers strong support for product improvement, which is of great significance for the development of the semiconductor industry and the improvement of product quality.

[0098] In some embodiments, the flatness detection information includes a three-dimensional flatness image and its corresponding flatness data, wherein the flatness data is used to indicate the flatness of each unit area.

[0099] Surface defect detection information includes two-dimensional particle detection images and their corresponding defect type data and defect location data;

[0100] The flatness classification results include the flatness category for each unit area in the flatness detection information;

[0101] The surface defect classification results include the defect category for each target defect in the surface defect detection information.

[0102] The unit area refers to one unit of area used for processing and analysis in this method. This unit area can be determined according to specific needs and is typically used for defect detection, flatness assessment, or other characterization analyses on semiconductor products.

[0103] Flatness inspection information includes three-dimensional flatness images and their corresponding flatness data. Flatness data indicates the flatness level per unit area, quantifying the smoothness of the surface. Flatness inspection information may also include two-dimensional contour images. A two-dimensional contour image is an image that displays the height variations of a product surface in two dimensions. It converts the height data of the product surface into contour lines, using the density and shape of these lines to represent the flatness of different areas. For example, in wafer flatness inspection, contour images can be used to display the undulations of the wafer surface, with the numerical values ​​of different contour lines indicating the degree of height difference.

[0104] Surface defect detection information includes two-dimensional particle detection images and their corresponding defect type and location data. The particle detection image is an image of the product surface acquired through optical microscopy or image processing techniques, containing information about particles or other defects. The defect type data describes the type of particles or defects detected, while the defect location data indicates their position on the product surface. Surface defect detection information may also include the size information of each defect. Size information refers to data describing the physical dimensional characteristics of the target defect. Size information may include parameters such as the defect's length, width, and diameter. By obtaining the size information of the target defect, the severity and extent of its impact can be understood.

[0105] Each unit area in the flatness inspection information is categorized into different flatness categories based on its flatness level. For example, the flatness of a unit area can be classified into different results such as caused by defects, caused by its own unevenness, or caused by a combination of defects and its own unevenness, or into different flatness classification results such as the probability level of being caused by defects, so as to quantitatively assess and compare the flatness of the product surface.

[0106] Each target defect in the surface defect detection information is categorized into different defect categories. For example, detected defects can be classified into different categories such as causing unevenness, jointly causing unevenness, and not causing unevenness, or into different defect classifications such as the probability level of the defect, in order to describe the undesirable features of the product surface.

[0107] Therefore, by comprehensively utilizing three-dimensional flatness images and corresponding flatness data, as well as two-dimensional particle detection images and corresponding defect type and location data, this method can provide more comprehensive and detailed flatness and surface defect information. Three-dimensional flatness images and data can intuitively display the flatness of the sample surface, and the flatness level per unit area can be quantified through the flatness data. This information can more accurately assess the surface quality of the sample and provide important basis for product process improvement. Two-dimensional particle detection images and their corresponding defect type and location data can intuitively display the particles and defects on the sample surface. Through image analysis and data processing, defects can be classified and their specific locations on the sample surface can be determined. This helps to promptly identify and locate potential problems, providing accurate guidance for quality control and repair work. By obtaining the flatness classification results and surface defect classification results, operators can directly understand the flatness category of the sample and the defect category of each defect. Such classification results facilitate comparative analysis, statistical summarization, and problem localization, providing strong support for process improvement and quality control.

[0108] In some embodiments, the method further includes:

[0109] Based on the flatness category of each unit area in the flatness detection information, image encoding data for each unit area is obtained. A first display image of the product is generated based on the image encoding data of each unit area and the 3D flatness image. In the first display image, the set of display parameters corresponding to unit areas of different flatness categories is different; and / or,

[0110] Based on the defect type data and defect location data in the surface defect detection information, the area of ​​each defect category in the 3D flatness image is obtained; based on the area of ​​each defect category in the 3D flatness image, the image encoding data of the area of ​​each defect category is obtained; based on the image encoding data of the area of ​​each defect category and the 3D flatness image, a second display image of the product is generated. In the second display image, the set of display parameters corresponding to the area of ​​different defect categories is different.

[0111] Image encoding data is encoded data used to represent image information for each unit area. Based on the flatness category of each unit area in the flatness detection information, the corresponding image encoding data can be obtained. This application does not limit the image encoding data; it may include, for example, at least one of the data values ​​of a two-dimensional image such as RGB values, coordinate data, grayscale values, bitmaps, histogram encoding, wavelet transform encoding, JPEG encoding, and PNG encoding.

[0112] A 3D flatness image is an image with three-dimensional topographic information generated by measuring and analyzing the surface of a wafer. Based on flatness detection information and image encoding data, a first display image of a product can be generated using the 3D flatness image. In the first display image, the unit area of ​​different flatness categories will correspond to different sets of display parameters. For example, the unit area of ​​the first flatness category may be presented in the display image as a first grayscale, while the unit area of ​​the second flatness category may be displayed as a second grayscale.

[0113] Defect type data describes the detected defect categories (or defect types), while defect location data indicates the regional locations of these defects in the 3D flatness image. Based on the data in the surface defect detection information, the area of ​​each defect category in the 3D flatness image can be obtained. It is important to note that "defect type" and "defect category" are two different concepts in this paper. "Defect type" can be categorized, for example, as particles, impurities, scratches, cracks, etc., while "defect category" can be categorized, for example, as causing unevenness, contributing to unevenness, not causing unevenness, etc.

[0114] The second display image refers to the product display image generated based on image-encoded data of the area of ​​the defect category and a three-dimensional flatness image. In the second display image, the area of ​​different defect categories will correspond to different sets of display parameters. For example, defects causing unevenness may be represented in the display image with a first-texture color, while defects that collectively cause unevenness may be represented with a second-texture color.

[0115] Through the above steps, first and second display images with different sets of display parameters can be generated based on flatness detection information and surface defect detection information. Thus, on one hand, image-coded data for each unit area is obtained according to the flatness category of each unit area, and the first display image of the product is generated using the three-dimensional flatness image. In the first display image, the unit area of ​​different flatness categories will correspond to different sets of display parameters. On the other hand, the area of ​​each defect category in the three-dimensional flatness image is obtained based on the defect type data and defect location data in the surface defect detection information. Then, image-coded data of the area of ​​each defect category in the three-dimensional flatness image is obtained. Finally, the second display image of the product is generated using this image-coded data and the three-dimensional flatness image. In the second display image, the area of ​​different defect categories will correspond to different sets of display parameters.

[0116] By generating a first display image, the unit area of ​​different flatness categories can be intuitively shown, and different sets of display parameters can be used to distinguish them. This display method helps operators quickly understand the flatness distribution of the sample and more accurately judge the surface quality of the product. By generating a second display image, the area of ​​different defect categories can be clearly displayed, and different sets of display parameters can be used to distinguish them. This display method makes defect location and analysis more intuitive and convenient, helping operators to perform timely quality control and problem handling. In summary, by generating different display images and distinguishing flatness and defect categories according to different sets of display parameters, more intuitive and detailed surface quality information can be provided. This is helpful for the manufacturing and quality control of semiconductor products, improving product reliability and performance.

[0117] In some embodiments, the product has a positioning edge;

[0118] Based on flatness detection information and surface defect detection information, obtain flatness classification results and / or surface defect classification results, including:

[0119] Based on the flatness data, find one or more unit areas where the flatness is less than the preset flatness.

[0120] For each identified unit area, based on the positioning edge, the corresponding defect type data is obtained from the surface defect detection information to detect whether there is a target defect within the unit area;

[0121] Set the flatness category of the unit area with the target defect to the first flatness category, and / or set the defect category of the target defect within the unit area with the target defect to the first defect category.

[0122] Set the flatness category per unit area for defects without a target to the second flatness category.

[0123] In this embodiment, the greater the flatness, the flatter the surface; the smaller the flatness, the less flat the surface.

[0124] Products with locating edges refer to specific edge structures or markings on semiconductor products used for positioning and alignment during the manufacturing process. These locating edges are used to ensure the correct positioning and alignment of the product during processing, testing, or assembly. For example, in wafer manufacturing, the edges of wafers typically have locating edges with markings or special shapes to accurately position the wafer in subsequent process steps.

[0125] Based on the flatness data in the flatness detection information, identify unit areas whose flatness is below a preset threshold. For example, set a preset flatness threshold; if the flatness of a unit area is less than this threshold, it is considered a unit area that does not meet the flatness standard.

[0126] A target defect refers to a defect that requires attention and treatment as defined in surface defect detection information; that is, a target defect is one of many types of defects. Target defects may negatively impact the performance, reliability, or appearance of a product. For example, on a semiconductor chip, a target defect might be a particle, scratch, or lattice defect.

[0127] By using the location information of the positioning edge, the region corresponding to each unit area is found in the surface defect detection information, and the defect type data in the region is extracted. This allows us to determine whether there is a target defect in the unit area.

[0128] For a unit area containing a target defect, its flatness category is marked as the first flatness category, and the defect category of the target defect within that unit area is set as the first defect category. This associates the flatness category and the defect category, facilitating further analysis and processing.

[0129] Furthermore, the flatness category of the unit area without target defects is set as the second flatness category. The cause of unevenness corresponding to the unit area of ​​the second flatness category is excluded as being caused by target defects (particles).

[0130] In this embodiment, the target defect is a particle. The above method can be used to obtain the flatness category or the particle category, or both the flatness category and the particle category can be obtained simultaneously.

[0131] Therefore, firstly, based on flatness data, one or more unit areas in the product with flatness less than a preset flatness standard are identified, indicating potential flatness deficiencies. Then, for each identified unit area, based on the positioning edges, the corresponding defect type data is obtained from surface defect detection information to detect the presence of a target defect within that unit area. Analysis of the defect type data determines the existence of a target defect. Next, the flatness category of the unit area containing the target defect is set as the first flatness category, and the defect category of the target defect within the unit area of ​​the first flatness category is set as the first defect category. Simultaneously, the flatness category of the unit area without a target defect is set as the second flatness category. The advantages of this approach are that by combining flatness detection and surface defect detection, this method can identify unit areas with substandard flatness based on a preset flatness standard and determine their location using positioning edges; by analyzing surface defect detection information, the presence of a target defect within these unit areas can be detected and determined; and by setting flatness and defect classifications, the product's quality status can be assessed more accurately.

[0132] In some embodiments, after setting the defect category of the target defect within a unit area of ​​the first flatness category as the first defect category, the method further includes:

[0133] For each unit area of ​​the first flatness category, perform the following processing:

[0134] Based on the positioning edge, the size information of the target defect within a unit area is obtained from the surface defect detection information, and it is determined whether the size of the target defect within a unit area is greater than the preset size.

[0135] When the size of the target defect within the unit area is greater than the preset size: the flatness category of the unit area is adjusted to the third flatness category, and / or the defect category of the target defect within the unit area is adjusted to the second defect category;

[0136] When the size of the target defect within the unit area is not greater than the preset size: the flatness category of the unit area is adjusted to the fourth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the third defect category.

[0137] The preset size is a pre-defined dimensional value used to determine whether a target defect meets a certain dimensional limit. By comparing it with the preset size, it can be determined whether the size of the target defect exceeds the set limit.

[0138] The third and fourth flatness categories are flatness categories obtained by reclassifying the unit area of ​​the first flatness category. That is, the flatness category of the unit area of ​​some units of the first flatness category is adjusted to the third flatness category, and the flatness category of the unit area of ​​some units of the first flatness category is adjusted to the fourth flatness category.

[0139] In this method, if the size of the target defect per unit area exceeds a preset size, the flatness category of that unit area will be adjusted to the third flatness category, and the defect category of the corresponding target defect will also be adjusted to the second defect category. If the size of the target defect does not exceed the preset size, the flatness category of the unit area will be adjusted to the fourth flatness category, and the defect category of the target defect will also be adjusted to the third defect category.

[0140] Therefore, by further judging and processing the size of the target defect per unit area in the first flatness category, the unit area can be more accurately classified and evaluated. If the size of the target defect exceeds the preset size, the flatness category will be adjusted to the third flatness category, and the defect category will be adjusted accordingly to the second defect category. This processing method helps to more accurately identify serious defects so that they can be dealt with and repaired in a timely manner. For cases where the size of the target defect does not exceed the preset size, the flatness category will be adjusted to the fourth flatness category, and the defect category will be adjusted to the third defect category. This adjustment reflects that although a defect exists, its size is within an acceptable range and has a small impact on product performance. In summary, by adjusting the flatness category per unit area and the corresponding defect category of the target defect according to the size of the target defect, this technical solution can more accurately classify and evaluate products and take corresponding measures for different situations, thereby further improving the quality control level of semiconductor products.

[0141] In some embodiments, the first preset aspect ratio to the Nth preset aspect ratio gradually decrease, where N is a positive integer;

[0142] The method also includes:

[0143] For each unit area of ​​either the first or third flatness category, perform the following processing:

[0144] Obtain the aspect ratio per unit area;

[0145] When the aspect ratio is greater than the first preset aspect ratio: the flatness category of the unit area is adjusted to the fifth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the fourth defect category;

[0146] When the aspect ratio is greater than the kth preset aspect ratio and not greater than the (k-1)th preset aspect ratio: the flatness category of the unit area is adjusted to the (k+4)th flatness category, and / or the defect category of the target defect within the unit area is adjusted to the (k+3)th defect category, where k is an integer greater than 1 and not greater than N.

[0147] The aspect ratio gradually decreases from the first preset aspect ratio to the Nth preset aspect ratio. Specifically, this method defines a series of preset aspect ratio values, where the first preset aspect ratio is the largest and the Nth preset aspect ratio is the smallest. This gradual decrease in preset aspect ratios can be used to describe the trend of aspect ratio per unit area changing with flatness. For example, when N=4, there are four preset aspect ratio values: 2, 1, 0.5, and 0.33, where 2 is the first preset aspect ratio and 0.33 is the fourth preset aspect ratio.

[0148] Aspect ratio represents the length-to-width ratio of a unit area of ​​undulating region; that is, the ratio of the longitudinal dimension to the transverse dimension of the undulating region per unit area. Aspect ratio can be used to describe the shape characteristics of a unit area. For example, if the aspect ratio of a unit area is 2:1, it means that the longitudinal dimension of its corresponding undulating region is twice its transverse dimension. If there are particles within this unit area, the particles may be fibrous or strip-shaped.

[0149] In this method, for a unit area of ​​either the first or third flatness category, if the aspect ratio of that unit area is greater than a first preset aspect ratio, the flatness category of that unit area will be adjusted to the fifth flatness category. Correspondingly, the defect category of the target defect within a unit area of ​​the fifth flatness category will be adjusted to the fourth defect category.

[0150] The (k+4)th flatness category and the (k+4)th defect category refer to the different flatness levels and defect categories defined when classifying a unit area based on flatness detection information and defect detection information. According to the aspect ratio threshold, when the aspect ratio of a unit area is greater than the k-th preset aspect ratio but not greater than the (k-1)-th preset aspect ratio, the flatness category of that unit area will be adjusted to the (k+4)th flatness category, and the defect category of the target defect under that flatness category will be adjusted to the (k+3)th defect category.

[0151] Therefore, for each unit area of ​​the first or third flatness category, the following processing is performed: First, the aspect ratio corresponding to the unit area is obtained. Then, adjustments are made based on a comparison of the aspect ratio with a preset aspect ratio. When the aspect ratio is greater than the first preset aspect ratio, the flatness category of the unit area is adjusted to the fifth flatness category, and the defect category of the target defect within the unit area of ​​the fifth flatness category is adjusted to the fourth defect category. This adjustment method can reflect that an abnormal deviation of the aspect ratio from the expected range may have an adverse effect on the product performance. In the following processing, adjustments are made based on a comparison of the aspect ratio with each preset aspect ratio. When the aspect ratio is greater than the kth preset aspect ratio but not greater than the (k-1)th preset aspect ratio, the flatness category of the unit area is adjusted to the (k+4)th flatness category, and the defect category of the target defect within the unit area of ​​the (k+4)th flatness category is adjusted to the (k+3)th defect category. This processing method classifies and adjusts according to the specific aspect ratio to more comprehensively reflect the characteristics of the product surface. The advantage of doing this is that by considering the relationship between the aspect ratio and the preset aspect ratio, the unit area can be further classified and adjusted. This approach can more accurately reflect the situation of aspect ratio abnormalities, help operators better judge the characteristics and potential defects of the product surface, and by adjusting the flatness category and defect category, the changes in surface features can be described and located more precisely, providing more targeted guidance for product manufacturing and quality control.

[0152] See Figure 7 and Figure 8 , Figure 7 This is a schematic diagram illustrating the lateral and longitudinal distances corresponding to an undulating region, provided in an embodiment of this application. Figure 8 This is a schematic diagram of the peaks and valleys corresponding to an undulating region provided in an embodiment of this application.

[0153] In some embodiments, obtaining the aspect ratio per unit area includes:

[0154] Based on the flatness detection information, the horizontal and vertical distances of the undulating regions corresponding to a unit area are obtained. The horizontal distance between the nearest valleys at both ends of the undulating region is taken as the horizontal distance of the undulating region, and the vertical distance between the peak of the undulating region and its nearest valley is taken as the vertical distance of the undulating region.

[0155] Calculate the quotient of the longitudinal distance and the transverse distance of the undulating region, and use it as the aspect ratio per unit area.

[0156] An undulating region refers to a region per unit area that exhibits variations in elevation, typically consisting of peaks and valleys. In flatness measurement information, the peaks and valleys of an undulating region can be defined and measured. The two ends of an undulating region refer to the two boundaries or endpoints of its horizontal central axis. The horizontal central axis is a horizontal line passing through the center point of the undulating region. Peaks represent the highest points of the undulating region, and valleys represent its lowest points.

[0157] Lateral distance refers to the width or distance of an undulating region in the horizontal direction. For example, the horizontal distance between the nearest valleys at both ends of an undulating region can be used as the lateral distance of the region.

[0158] Longitudinal distance refers to the height or distance of an undulating region in the vertical direction. For example, the longitudinal distance of an undulating region is the vertical distance between the peak of the undulating region and its nearest valley.

[0159] Aspect ratio is used to describe the ratio between the longitudinal and lateral distances of an undulating area. By calculating the quotient of the longitudinal and lateral distances of the undulating area, the aspect ratio per unit area can be obtained, which can be used to describe the shape characteristics of a unit area, i.e., the ratio of the longitudinal dimension to the lateral dimension.

[0160] like Figure 7 and Figure 8 As shown, for an undulating region on a wafer, flatness detection information is analyzed to determine the horizontal coordinates of the valley points c and d at both ends of the undulating region in the wafer coordinate system as x1 and x2, respectively. The vertical coordinate of the peak point a in the undulating region is y1 in the wafer coordinate system. The vertical coordinate of the valley point c (between c and d) closest to the peak point a is y2 in the wafer coordinate system. The vertical distance h and the horizontal distance w are calculated, where h = y1 - y2 and w = |x2 - x1|. Therefore, the aspect ratio of the undulating region is calculated as h / w.

[0161] For example, for a unit area of ​​a wafer surface, based on flatness detection information, the lateral and longitudinal distances of the undulating regions corresponding to that unit area can be found. Assuming the lateral distance of the undulating regions is 10 micrometers and the longitudinal distance is 5 micrometers, then the aspect ratio corresponding to that unit area is 5 / 10 = 0.5.

[0162] Aspect ratio is an indicator describing the shape characteristics of undulations. It can be expressed as the ratio of longitudinal distance to lateral distance, reflecting the radius of curvature of the undulation. The advantage of this is that by calculating the aspect ratio, the shape characteristics of the undulation can be quantitatively characterized, and the degree of curvature can be objectively assessed. Furthermore, the aspect ratio can provide a reference for further explaining the causes of undulations (or unevenness). By comparing the aspect ratio per unit area with a preset aspect ratio, it is possible to infer whether the unevenness is caused by particles or by the unevenness of the particles themselves, thus contributing to a deeper understanding of the causes of unevenness.

[0163] Therefore, firstly, based on the flatness detection information, the lateral and longitudinal distances of the undulating regions per unit area are obtained. The lateral distance of the undulating region refers to the lateral distance between the nearest valleys at both ends of the undulating region, while the longitudinal distance refers to the longitudinal distance between the peak of the undulating region and its nearest valley. Then, the quotient between the longitudinal and lateral distances of the undulating regions is calculated. This quotient is the aspect ratio per unit area. Through these steps, the aspect ratio per unit area can be obtained to reflect the characteristics of the surface undulating regions. The aspect ratio is calculated by comparing the longitudinal and lateral distances of the undulating regions to determine the relative degree of surface undulation characteristics. The advantage of this approach is that by calculating the aspect ratio of the undulating regions, the surface undulation characteristics can be objectively assessed. The aspect ratio value can quantitatively describe the degree and direction of undulation within a unit area. This assessment method is of great significance for the quality control and manufacturing process optimization of semiconductor products. Through the calculation of the aspect ratio, operators can better understand and judge the surface undulation of the product and take corresponding measures to improve product quality and performance.

[0164] In some embodiments, the method further includes:

[0165] Based on the flatness detection information, the flatness category of each unit area, and the defect category of each target defect, a third display image of the product is generated. In the third display image, the set of display parameters corresponding to the unit area of ​​different flatness categories is different, and the set of display parameters corresponding to the target defects of different defect categories is different.

[0166] Therefore, based on flatness detection information, the flatness category of each unit area, and the defect category of each target defect, a third display image of the product is generated. In this third display image, the set of display parameters corresponding to the unit area for different flatness categories is different, and the set of display parameters corresponding to the target defects for different defect categories is also different. By generating the third display image, the flatness and defect information of the product surface can be presented intuitively. The unit area of ​​different flatness categories is assigned a corresponding set of display parameters according to its flatness, thus presenting different representations in the image. Similarly, target defects of different defect categories are assigned a corresponding set of display parameters according to their defect types, clearly displaying these defects in the third display image. Through this display method, operators can intuitively understand the flatness of the product surface and the types of defects present, thereby better enabling quality control and decision-making. The third display image provides visualized information, helping operators to comprehensively evaluate and analyze the product, and can identify and locate different types of defects based on the set of display parameters in the image. Therefore, by generating a third display image, the flatness and defect information of the product can be effectively displayed, providing an intuitive visual reference, which has significant advantages for product quality control and problem diagnosis.

[0167] In some embodiments, the set of display parameters includes one or more of the following display parameters: color, shape, texture, size, transparency, brightness, contrast, saturation, edge sharpness, and gradient category.

[0168] Therefore, by using different display parameters, different categories of particles can be effectively distinguished and represented. For example, using different colors to represent different categories of particles allows for intuitive differentiation through color differences. Furthermore, using different shapes, textures, or sizes can further emphasize the characteristics of different particle categories. Adjusting parameters such as transparency, brightness, contrast, and saturation can also enhance the visual effect of different particle categories. Controlling edge sharpness and gradient categories can affect the edge features and gradient transition effects of particles. By selecting an appropriate set of display parameters, the characterization image can more accurately reflect the characteristics of different particle categories, helping researchers and engineers better understand and analyze the particle situation on the product surface. This contributes to further improving semiconductor testing methods, optimizing manufacturing processes, and improving product quality and performance. In summary, using a set of display parameters, including color, shape, texture, size, transparency, brightness, contrast, saturation, edge sharpness, and gradient categories, as a visual expression method for characterization images can effectively convey information about different particle categories, improving the accuracy of analysis and the effectiveness of visualization.

[0169] As an example, if the target defect is a particle, the set of display parameters per unit area for the fifth flatness category is:

[0170] Color: Blue;

[0171] Transparency: Semi-transparent;

[0172] Edge sharpness: Sharp.

[0173] The set of display parameters per unit area for the sixth flatness category is as follows:

[0174] Color: Yellow;

[0175] Transparency: Semi-transparent;

[0176] Edge sharpness: blurry.

[0177] The set of display parameters per unit area for the seventh flatness category is as follows:

[0178] Color: Green;

[0179] Transparency: Opaque;

[0180] Edge sharpness: Sharp.

[0181] The set of display parameters corresponding to the particles in the fourth defect category is as follows:

[0182] Color: Red;

[0183] Shape: Rectangle;

[0184] Size: Large;

[0185] Contrast ratio: High contrast.

[0186] The set of display parameters corresponding to the particles in the fifth defect category is as follows:

[0187] Color: Purple;

[0188] Shape: Triangle;

[0189] Size: Medium;

[0190] Contrast ratio: Medium.

[0191] The set of display parameters corresponding to particles in the sixth defect category is as follows:

[0192] Color: Orange;

[0193] Shape: Circular;

[0194] Size: Small;

[0195] Contrast ratio: Low contrast.

[0196] These numerical examples are for illustrative purposes only and are selected based on common characteristics of the display parameters. In practical applications, the values ​​of these parameters can be adjusted according to specific needs and application scenarios. The selection and adjustment of these display parameters can help accurately display and highlight the characteristics corresponding to different flatness and defect categories, providing more detailed information about semiconductor products.

[0197] See Figure 6 and Figure 9 , Figure 9 This is a schematic diagram of a process for obtaining flatness classification results and / or surface defect classification results provided in an embodiment of this application.

[0198] In a specific application scenario, this application embodiment also provides a semiconductor testing method, including:

[0199] Obtain flatness detection information of the product; the product has positioning edges; the flatness detection information includes a three-dimensional flatness image and its corresponding flatness data, the flatness data is used to indicate the flatness of each unit area;

[0200] Acquire surface defect detection information of the product; surface defect detection information includes two-dimensional particle detection images and their corresponding defect type data and defect location data;

[0201] Based on the flatness detection information and surface defect detection information, obtain the flatness classification results and / or surface defect classification results; the flatness classification results include the flatness category of each unit area in the flatness detection information; the surface defect classification results include the defect category of each target defect in the surface defect detection information;

[0202] Based on the flatness category of each unit area in the flatness detection information, image encoding data of each unit area is obtained, and the first display image of the product is generated based on the image encoding data of each unit area and the three-dimensional flatness image. In the first display image, the set of display parameters corresponding to the unit area of ​​different flatness categories is different.

[0203] Based on the defect type data and defect location data in the surface defect detection information, obtain the area of ​​each defect category in the 3D flatness image; based on the area of ​​each defect category in the 3D flatness image, obtain the image encoding data of the area of ​​each defect category; generate a second display image of the product based on the image encoding data of the area of ​​each defect category and the 3D flatness image, in which the set of display parameters corresponding to the area of ​​different defect categories is different;

[0204] Based on the flatness detection information, the flatness category of each unit area, and the defect category of each target defect, a third display image of the product is generated. In the third display image, the set of display parameters corresponding to the unit area of ​​different flatness categories is different, and the set of display parameters corresponding to the target defects of different defect categories is different.

[0205] The process of obtaining flatness classification results and / or surface defect classification results includes:

[0206] Based on the flatness data, find one or more unit areas where the flatness is less than the preset flatness.

[0207] For each identified unit area, based on the positioning edge, the corresponding defect type data is obtained from the surface defect detection information to detect whether there is a target defect within the unit area;

[0208] Set the flatness category of the unit area with the target defect to the first flatness category, and set the defect category of the target defect within the unit area of ​​the first flatness category to the first defect category.

[0209] Set the flatness category per unit area of ​​the defect without a target to the second flatness category;

[0210] For each unit area of ​​the first flatness category, perform the following processing:

[0211] Based on the positioning edge, the size information of the target defect within a unit area is obtained from the surface defect detection information, and it is determined whether the size of the target defect within a unit area is greater than the preset size.

[0212] When the size of the target defect within the unit area is greater than the preset size: the flatness category of the unit area is adjusted to the third flatness category, and / or the defect category of the target defect within the unit area is adjusted to the second defect category;

[0213] When the size of the target defect within the unit area is not greater than the preset size: the flatness category of the unit area is adjusted to the fourth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the third defect category;

[0214] For each unit area of ​​the third flatness category, perform the following processing:

[0215] Obtain the aspect ratio per unit area;

[0216] When the aspect ratio is greater than the first preset aspect ratio: the flatness category of the unit area is adjusted to the fifth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the fourth defect category;

[0217] When the aspect ratio is greater than the kth preset aspect ratio and not greater than the (k-1)th preset aspect ratio: the flatness category of the unit area is adjusted to the (k+4)th flatness category, and / or the defect category of the target defect within the unit area is adjusted to the (k+3)th defect category, where k is an integer greater than 1 and not greater than N.

[0218] The process of obtaining the aspect ratio corresponding to a unit area includes:

[0219] Based on the flatness detection information, the horizontal and vertical distances of the undulating regions corresponding to a unit area are obtained. The horizontal distance between the nearest valleys at both ends of the undulating region is taken as the horizontal distance of the undulating region, and the vertical distance between the peak of the undulating region and its nearest valley is taken as the vertical distance of the undulating region.

[0220] Calculate the quotient of the longitudinal distance and the transverse distance of the undulating region, and use it as the aspect ratio per unit area.

[0221] Taking a wafer as an example, firstly, the wafer is flattened using a TTV (Total Television) inspection device, collecting three-dimensional flatness images, two-dimensional contour images, and corresponding flatness data. Then, a Candela surface defect detector is used to detect defects on the wafer surface, collecting two-dimensional particle detection images and corresponding defect type and location data. For example, the Candela surface defect detector can detect different types of defects on the wafer surface, such as particles, scratches, and stains. In this embodiment, the wafer is, for example, a silicon wafer or a gallium nitride wafer; this application is not limited to either.

[0222] After collecting the above data, the wafers are classified based on flatness detection information and surface defect detection information. For example, if the preset flatness is 0.02 mm, areas with flatness less than this threshold are detected.

[0223] The unit area with flatness lower than the preset flatness and containing target defects (such as particles) is marked as the first flatness category, and the particles within the unit area are marked as the first defect category; for areas with flatness lower than the threshold but no particles, they are marked as the second flatness category.

[0224] For each unit area of ​​the first flatness category, further processing is performed based on the size information of the defects. For example, if the particle size exceeds a preset size, such as 0.5 μm, its flatness category is adjusted to the third flatness category, and the defect category of the particles within that unit area is adjusted to the second defect category; conversely, if the particle size does not exceed the preset size, its flatness category is adjusted to the fourth flatness category, and the defect category of the particles within that unit area is adjusted to the third defect category.

[0225] For a unit area of ​​the third flatness category, its aspect ratio information is further obtained. For example, a set of preset aspect ratio thresholds are set. If the aspect ratio of a unit area is greater than the first preset aspect ratio, its flatness category is adjusted to the fifth flatness category, and the defect category of the particles within that unit area is adjusted to the fourth defect category. If the aspect ratio of the unit area is between two preset aspect ratio thresholds, the flatness category of the unit area and the defect category of the particles within that unit area are adjusted to the corresponding categories.

[0226] As an example, the unevenness corresponding to the fifth flatness category and the fourth defect category is caused by particles; the unevenness corresponding to the sixth flatness category and the fifth defect category is caused by a combination of particles and their own unevenness, with weights of 0.8 and 0.2 respectively; the unevenness corresponding to the seventh flatness category and the sixth defect category is caused by a combination of particles and their own unevenness, with weights of 0.6 and 0.4 respectively; the unevenness corresponding to the eighth flatness category and the seventh defect category is caused by a combination of particles and their own unevenness, with weights of 0.5 and 0.5 respectively; the unevenness corresponding to the ninth flatness category and the eighth defect category is caused by a combination of particles and their own unevenness, with weights of 0.4 and 0.6 respectively; and so on.

[0227] For wafers in semiconductor manufacturing processes, controlling their flatness and surface defect states is crucial. In the example above, different flatness and defect categories correspond to different causes of unevenness, as well as the relative weights among these causes, thus providing a more refined and scientific way to identify and address wafer unevenness issues.

[0228] First, a detailed classification of wafer surface defects and flatness provides a more accurate reflection of the wafer's true condition. For example, the fifth flatness category and the fourth defect category are caused solely by particles, while categories six through nine are caused by a combination of particles and wafer-level unevenness, with each contributing differently to the unevenness problem. This classification not only provides a clear understanding of the various issues on the wafer surface but also helps identify the root causes, enabling targeted improvements.

[0229] Secondly, clarifying the different causes of unevenness and their relative weights helps to better understand the complexity and diversity of wafer unevenness problems. For example, for the sixth to ninth flatness categories, although the causes of unevenness are all the same—the unevenness of the particles and the wafer itself—their relative weights differ. This difference may mean that different strategies and methods are needed when dealing with these problems.

[0230] Finally, this meticulous classification and weighting provides rich information for subsequent data analysis and model optimization. For each flatness and defect category, a large amount of experimental data can be collected and analyzed to explore the impact of different factors on wafer flatness and the relationships between them. This not only improves our understanding of wafer flatness issues but also helps in developing more effective process optimization strategies.

[0231] In this embodiment, "particles" refers to tiny particles or impurities on the surface of a semiconductor product. These particles may originate from impurities in the manufacturing environment or process, such as dust or metal particles, and may affect the flatness and performance of the product. "Intrinsic unevenness" refers to surface irregularities caused by internal factors during manufacturing or the inherent properties of the material itself, rather than the influence of external particles. For example, structural changes in the wafer or internal material stress may lead to intrinsic unevenness.

[0232] Therefore, it is possible to comprehensively analyze the flatness inspection information and surface defect inspection information of semiconductor products, and determine the possible causes of flatness non-compliance for each unit area. By acquiring flatness inspection information and surface defect inspection information, this method can identify and distinguish the possible causes of undulations, including undulations caused by particles and unevenness of the product itself. The advantage of this approach is that it considers not only the product's flatness inspection information but also combines it with surface defect inspection information, enabling the evaluation and analysis of the product's comprehensive surface performance. By comprehensively analyzing information from different data sources, a more complete understanding of the causes of undulations can be achieved. This method can clearly identify each unit area where flatness fails to meet standards, making problem localization more accurate. This helps to carry out targeted improvements and treatments, thereby improving product quality and performance. This method can determine the possible causes of each undulation, including undulations caused by particles and / or unevenness of the product itself. This provides targeted references for further process optimization, equipment improvement, and production control, reducing the generation of defective products. By acquiring flatness detection information and surface defect detection information and conducting comprehensive analysis, this method can improve the efficiency of testing and analysis. Compared with performing flatness detection and surface defect detection separately, comprehensive analysis can more accurately determine the problem area and cause, with high accuracy and saving time and resource costs.

[0233] Taking wafers as an example, firstly, flatness testing is performed, and the flatness testing information of the wafers is obtained, such as... Figures 1 to 3 As shown. For example, the undulations of the wafer surface can be measured, and the height value of each point can be recorded. This data can be used to represent the flatness of the wafer surface.

[0234] Next, surface defect detection is performed, and surface defect detection information of the wafer is obtained, such as... Figure 4 and Figure 5 As shown. For example, a Candela surface defect detector can be used to inspect a wafer and generate a two-dimensional particle detection image. Figure 4 Figures 1-6 show the location of the particles on the wafer surface.

[0235] Pay attention to areas with poor flatness, as these may be caused by unevenness of the particles or the wafer itself. For example, if a particle with a large diameter is found in a unit area, it can be inferred that the presence of the particle affects the flatness of that area. Conversely, if an undulating region does not have obvious particles but still fails to achieve the required flatness, it can be inferred that the unevenness in that region may be caused by the wafer itself.

[0236] Defect location data refers to the coordinate information of defects on the product surface, used to accurately locate and describe the specific location of each defect. For example, for a particle located on a wafer, its X and Y coordinate values ​​relative to the wafer coordinate system can be recorded to determine its position on the wafer surface.

[0237] Diameter is a parameter describing particle size, typically referring to the maximum line segment length of the particle. It can be calculated by measuring the pixel length in a particle image and converting it to its actual length based on the corresponding scale relationships. For example, for a single particle, the diameter can be calculated by measuring the number of pixels it occupies in a microscope image and applying known scale relationships.

[0238] By matching and aligning the scales of flatness detection information and surface defect detection information, the scale transformation relationship between the two can be obtained. This relationship can be acquired by calculating transformation coefficients or establishing coordinate mappings. By determining the correspondence between the positioning markers in the two types of data, it can be ensured that the flatness data and surface defect data are compared and analyzed under the same scale and coordinate system.

[0239] In the analysis of flatness and surface defect detection information, overlay refers to combining the two types of information to obtain a more comprehensive analysis result. For example, a three-dimensional flatness image (or a two-dimensional contour image) and a two-dimensional particle detection image can be overlaid on the same image to show the undulations of the wafer surface and the location distribution of particles. Through overlay analysis, the presence of particles in each unit area can be observed intuitively, and the impact of particles on flatness can be further understood.

[0240] In the analysis of flatness detection information and surface defect detection information, comparison refers to comparing the two to determine their relationship and mutual influence. For example, a three-dimensional flatness image (or a two-dimensional contour image) and a two-dimensional particle detection image can be compared and analyzed. Through comparative analysis, the spatial distribution characteristics of undulating regions and particles can be observed, and the causes of the undulating regions can be further inferred, such as whether they are caused by particles or by the surface's own unevenness.

[0241] In some embodiments, the scaling relationship is represented by a scaling transformation matrix.

[0242] For example, the scaling transformation matrix is ​​a two-dimensional matrix that contains the parameters required for scaling. Specifically, the scaling transformation matrix can be represented as a 2x2 matrix, as shown below:

[0243] [b11 b12]

[0244] [b21 b22]

[0245] In this matrix, b11 and b22 represent the scale factors in the horizontal and vertical directions, respectively, while b12 and b21 represent the rotation or shearing transformation of the scale.

[0246] In some embodiments, the scale transformation relationship is represented by a scale transformation model, which is either a linear model or a deep learning-based model.

[0247] A scale transformation model is a mathematical model used to describe scale transformation relationships. It can be a linear model or a deep learning-based model.

[0248] Linear models can use linear transformation matrices (such as the two-dimensional matrix mentioned above) to describe the scaling relationship, where the elements of the matrix represent the scaling factor and parameters of rotation or shearing transformations. Linear models are suitable for some simple scaling tasks, and their advantages are high computational efficiency and strong interpretability.

[0249] Deep learning-based scale transformation models leverage the powerful representation learning capabilities of neural networks to automatically learn scale transformation relationships in input data. Deep learning models can be trained with large amounts of sample data to learn complex nonlinear scale transformation rules. While deep learning models perform well in some complex scale transformation tasks, they have a large number of parameters and require substantial training data and computational resources. Deep learning-based scale transformation models can employ various deep neural network architectures, such as convolutional neural networks (CNNs) or generative adversarial networks (GANs).

[0250] Based on flatness and surface defect detection information, it can be determined whether particles are present within each unit area where flatness fails to meet standards. When particles are present within a unit area, the degree of influence of the particles on that unit area can be inferred by comparing the particle diameter with the aspect ratio of that unit area. Larger particles and larger aspect ratios are more likely to be caused by particles, while smaller aspect ratios may indicate fluctuations caused by other factors. The advantage of this approach is that by detecting the presence of particles within a unit area, it is possible to determine whether the cause of the fluctuation is related to particles; by analyzing the diameter and aspect ratio of the particles within that unit area, the degree of influence of the particles on the formation of that unit area can be inferred, thereby identifying the possible causes of fluctuations; and by understanding the possible causes of fluctuations, targeted subsequent process optimization and quality control can be implemented to improve the flatness performance of the product.

[0251] In this embodiment, if the particle size is large and the curvature of the undulation is high, the particle may be the cause of the undulation. If the aspect ratio of the undulation region is not greater than a first preset aspect ratio, it means that the particle size is large, but the curvature of the undulation region is small, and there may be other factors besides the particle causing the undulation. By judging the particle diameter, the degree of influence of particle size on the undulation region can be inferred, thereby better understanding the cause of the undulation; by comprehensively considering the particle diameter and the aspect ratio of the undulation region, the possible causes of the undulation region can be analyzed more comprehensively, including factors such as particle and its own unevenness. This is because the aspect ratio corresponding to a unit area may reflect the shape and distribution of the particles. For example, a high aspect ratio may indicate that the particles have a significant distribution along a specific direction, which may affect the microstructure of the wafer.

[0252] When the target defect is a particle, classifying the particles into multiple categories allows for a more accurate description of their size, shape, and characteristics. Different categories of particles may have varying degrees of impact on unevenness, so further subdivision provides more detailed information. Determining the particle category based on its diameter and / or aspect ratio per unit area allows for a quantitative description of the particles, enabling a more objective analysis of the characteristics of particles per unit area and providing a basis for subsequent analysis and judgment. Identifying the particle category allows for further research and analysis of the characteristics, origins, and impacts on undulating areas of different particle categories. This helps to deepen the understanding of the causes and characteristics of unevenness and provides a reference for developing corresponding improvement measures.

[0253] As an example, the product is a gallium nitride wafer sample. See also... Figures 1 to 5 , Figure 1 This is a three-dimensional image (i.e., a three-dimensional flatness image) obtained when using wafer flatness inspection. Figure 2 This is a local flatness map from TTV testing. Figure 3 TTV testing in progress Figure 1 The corresponding plan view (i.e., a two-dimensional contour map). Figure 1 and Figure 3 All measurements were obtained under vacuum adsorption. The vacuum condition forces the wafer to be confined to the surface of the chuck, and the chuck surface matches the back side of the wafer. Using the back side of the wafer as a reference surface, the total thickness change (TTV) of the wafer surface can be measured. Figure 3 It is a two-dimensional contour image of the sample surface, with different colored areas representing the undulations of the wafer surface. Figure 1 It generates a three-dimensional surface topography map, which can more intuitively display the flatness of the sample surface. Figure 2 It is the measured local thickness variation (LTV value) of the wafer, which is used to evaluate the wafer in a local area.

[0254] If there are particles on the upper and lower surfaces of the sample, they will appear as a raised area in the TTV image (e.g. Figure 1 In regions A and B of the sample, the dust or particles will protrude due to the vacuum adsorption, affecting the measurement results. However, the wafer flatness instrument cannot display the size and precise location of the particles.

[0255] Figure 4 The Candela test obtains information about the sample surface by generating a scattered signal from the surface of the wafer (i.e., a crystal wafer) using obliquely incident laser light, and then detects particles (positions 1-6 circled in the figure correspond to relatively large particles). For larger particles, the light scattering signal is stronger, so the Candela test also includes an obliquely incident laser reflection mode for measuring smaller particles.

[0256] Candela's testing equipment integrates particle detection functionality, such as... Figure 5 As shown. However, this function can only measure the lateral diameter of the particle, and cannot determine the vertical diameter of the particle.

[0257] Since TTV (Transient Volume Television) detection can obtain the ratio of the longitudinal distance to the lateral distance of an undulating region, i.e., the aspect ratio, it can indirectly reflect the radius of curvature of that undulating region. Specifically, the process of measuring the aspect ratio is as follows: the longitudinal distance (h) can be defined as the distance from the peak point (h) to the lateral distance (h). Figure 8 The nearest valley point to the peak point (i.e., point a) Figure 8 The longitudinal distance of valley point c) in the middle is based on the TTV measurement. Figure 3 The lateral distance (w) is determined by the numerical values ​​of the two-dimensional contour image. Figure 8 The distance from midpoint c to point d can be calculated using the coordinates of each point in the wafer coordinate system. The aspect ratio of the undulating region is defined as h / (w).

[0258] When performing TTV and Candela detection, the same sample has positioning edges. In this embodiment, the two images can be superimposed or compared based on the positioning edges (the order of comparison is not limited to the order of TTV image and Candela image). Here, TTV image refers to the three-dimensional flatness image and two-dimensional contour image obtained by TTV detection, and Candela image refers to the two-dimensional particle detection image obtained by Candela detection.

[0259] The particles on the sample surface measured by Candela are all introduced particles from external sources. If particles are present on the upper surface of the sample, Candela can display their location and size. Based on the particle location, corresponding undulating areas (such as...) can be found in the TTV. Figure 1 The bulge at position B is consistent with the measurement by Candela. Figure 4 (Corresponding to position 5). For Figure 1 The protrusion at position A Figure 4 The fact that no corresponding particles were found directly suggests that these particles may exist on the lower surface of the sample or on the wafer chuck.

[0260] Therefore, by overlaying and comparing the information from TTV and Candela tests, it is possible to determine whether the undulations in the 3D flatness image obtained from TTV testing are caused by externally introduced particles or by inherent unevenness in the semiconductor itself. Furthermore, the positions of particles can be marked on the 3D TTV image, addressing the lack of intuitiveness of Candela testing. Additionally, the precise location of particles can be accurately pinpointed based on the position of the TTV test image.

[0261] Specifically, TTV testing can obtain the flatness and local flatness of a sample; if the flatness of a unit area obtained from TTV testing is greater than the threshold, it can be understood that the flatness of that unit area does not meet the standard.

[0262] By analyzing the coordinates of the particles, we can determine whether particles exist within a unit area where flatness is substandard. This allows us to determine whether particles are present within that unit area, categorizing substandard flatness units into: 1. A first flatness category containing particles; 2. A second flatness category without particles. However, it's impossible to determine whether the unevenness corresponding to the unit area in the first flatness category is caused by particles (because the vertical dimensions of the particles are unknown).

[0263] Determining whether the particle diameter exceeds a certain threshold initially suggests that the undulations in the area are likely caused by particles. Therefore, the unit area of ​​the first flatness category is divided into: 1. A third flatness category with a diameter greater than a preset diameter; 2. A fourth flatness category with a diameter not greater than a preset diameter. A large lateral diameter of particles is highly likely to be caused by particle-induced undulations, with an accuracy rate of up to 80%. However, there is a very small possibility that the undulations are caused by a small longitudinal diameter of the particles, resulting from a combination of particles and other factors.

[0264] When the particle diameter is greater than the diameter threshold, the aspect ratio of the undulating region per unit area is obtained from the flatness detection information. The undulating region per unit area is, for example, the undulating region within that unit area.

[0265] Judging from the aspect ratio, a large aspect ratio indicates a steeper undulation (which can be understood as a small radius of curvature). The undulation is caused by external factors. In addition, the particle size is also large, which further indicates that the undulation is caused by external particles.

[0266] When the aspect ratio is small, it indicates that the fluctuations are relatively gentle (which can be understood as a large radius of curvature). Although the particle size at this point is also larger than the threshold, the fluctuations may be caused by particles and internal bulges.

[0267] To further understand the cause of the second situation mentioned above, we can classify the units based on the aspect ratio, thereby dividing the unit area of ​​the third flatness category into the fifth flatness category to the N+4 flatness category, and correspondingly, dividing the particles within that unit area into the fourth defect category to the N+3 defect category.

[0268] Specifically, for a unit area of ​​the third flatness category, when the aspect ratio of the unit area is greater than the first preset aspect ratio, the flatness category of the unit area is adjusted to the fifth flatness category. In the three-dimensional flatness image, the particles in the unit area can be marked as the fourth defect category and matched with the color of the fourth defect category (such as red), indicating that the unevenness of the unit area may be caused by particles.

[0269] When the aspect ratio is between the first preset aspect ratio and the second preset aspect ratio, it means that although the particle has a large lateral size, the undulation is not very steep, but it is not very gentle either. This means that the particle still has a certain impact, but the impact of the particle is not as great as in the first case (possibly because the particle's longitudinal size is not high). Adjust the flatness category of the unit area to the sixth flatness category, mark the particle category within the unit area as the fifth defect category, match the color of the fifth defect category (such as yellow), and so on.

[0270] For external particles, the coordinates of each particle within a unit area can be precisely located and then integrated into the original TTV 3D image. Furthermore, particles can be categorized and integrated into the original 3D flatness image using different colors. The advantage of this approach is that it not only provides a comprehensive and intuitive TTV image but also accurately displays external particles within the TTV.

[0271] This application embodiment combines flatness testing and particle size testing to speculate on the possible causes of changes in product surface flatness, classify particle types, and distinguish them with different display effects. By combining the wafer flatness test results and particle test results, the wafer flatness test results are more intuitive and can help speculate on the causes of flatness changes.

[0272] (Electronic devices)

[0273] This application also provides an electronic device, the specific embodiments of which are consistent with the embodiments and technical effects achieved in the above method embodiments, and some contents will not be repeated.

[0274] The electronic device includes a memory and at least one processor. The memory stores a computer program, and the at least one processor is configured to perform the following steps when executing the computer program:

[0275] Obtain flatness test information for the product;

[0276] Obtain surface defect detection information for products;

[0277] Based on the flatness detection information and surface defect detection information, obtain the flatness classification results and / or surface defect classification results.

[0278] In some embodiments, the flatness detection information includes a three-dimensional flatness image and its corresponding flatness data, wherein the flatness data is used to indicate the flatness of each unit area.

[0279] Surface defect detection information includes two-dimensional particle detection images and their corresponding defect type data and defect location data;

[0280] The flatness classification results include the flatness category for each unit area in the flatness detection information;

[0281] The surface defect classification results include the defect category for each target defect in the surface defect detection information.

[0282] In some embodiments, when at least one processor is configured to execute a computer program, the following steps are also performed:

[0283] Based on the flatness category of each unit area in the flatness detection information, image encoding data for each unit area is obtained. A first display image of the product is generated based on the image encoding data of each unit area and the 3D flatness image. In the first display image, the set of display parameters corresponding to unit areas of different flatness categories is different; and / or,

[0284] Based on the defect type data and defect location data in the surface defect detection information, the area of ​​each defect category in the 3D flatness image is obtained; based on the area of ​​each defect category in the 3D flatness image, the image encoding data of the area of ​​each defect category is obtained; based on the image encoding data of the area of ​​each defect category and the 3D flatness image, a second display image of the product is generated. In the second display image, the set of display parameters corresponding to the area of ​​different defect categories is different.

[0285] In some embodiments, the product has a positioning edge;

[0286] When at least one processor is configured to execute a computer program, it acquires flatness classification results and / or surface defect classification results based on flatness detection information and surface defect detection information in the following manner:

[0287] Based on the flatness data, find one or more unit areas where the flatness is less than the preset flatness.

[0288] For each identified unit area, based on the positioning edge, the corresponding defect type data is obtained from the surface defect detection information to detect whether there is a target defect within the unit area;

[0289] Set the flatness category of the unit area with the target defect to the first flatness category, and set the defect category of the target defect within the unit area of ​​the first flatness category to the first defect category.

[0290] Set the flatness category per unit area for defects without a target to the second flatness category.

[0291] In some embodiments, after setting the defect category of a target defect within a unit area of ​​a first flatness category to the first defect category, at least one processor is configured to execute a computer program to further implement the following steps:

[0292] For each unit area of ​​the first flatness category, perform the following processing:

[0293] Based on the positioning edge, the size information of the target defect within a unit area is obtained from the surface defect detection information, and it is determined whether the size of the target defect within a unit area is greater than the preset size.

[0294] When the size of the target defect within the unit area is greater than the preset size: the flatness category of the unit area is adjusted to the third flatness category, and / or the defect category of the target defect within the unit area is adjusted to the second defect category;

[0295] When the size of the target defect within the unit area is not greater than the preset size: the flatness category of the unit area is adjusted to the fourth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the third defect category.

[0296] In some embodiments, the first preset aspect ratio to the Nth preset aspect ratio gradually decrease, where N is a positive integer;

[0297] When at least one processor is configured to execute a computer program, the following steps are also performed:

[0298] For each unit area of ​​either the first or third flatness category, perform the following processing:

[0299] Obtain the aspect ratio per unit area;

[0300] When the aspect ratio is greater than the first preset aspect ratio: the flatness category of the unit area is adjusted to the fifth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the fourth defect category;

[0301] When the aspect ratio is greater than the kth preset aspect ratio and not greater than the (k-1)th preset aspect ratio: the flatness category of the unit area is adjusted to the (k+4)th flatness category, and / or the defect category of the target defect within the unit area is adjusted to the (k+3)th defect category, where k is an integer greater than 1 and not greater than N.

[0302] In some embodiments, at least one processor is configured to obtain the aspect ratio per unit area when executing a computer program in the following manner:

[0303] Based on the flatness detection information, the horizontal and vertical distances of the undulating regions corresponding to a unit area are obtained. The horizontal distance between the nearest valleys at both ends of the undulating region is taken as the horizontal distance of the undulating region, and the vertical distance between the peak of the undulating region and its nearest valley is taken as the vertical distance of the undulating region.

[0304] Calculate the quotient of the longitudinal distance and the transverse distance of the undulating region, and use it as the aspect ratio per unit area.

[0305] In some embodiments, when at least one processor is configured to execute a computer program, the following steps are also performed:

[0306] Based on the flatness detection information, the flatness category of each unit area, and the defect category of each target defect, a third display image of the product is generated. In the third display image, the set of display parameters corresponding to the unit area of ​​different flatness categories is different, and the set of display parameters corresponding to the target defects of different defect categories is different.

[0307] See Figure 10 , Figure 10 This is a structural block diagram of an electronic device 10 provided in an embodiment of this application.

[0308] Electronic device 10 may include, for example, at least one memory 11, at least one processor 12, and a bus 13 connecting different platform systems.

[0309] Memory 11 may include a computer-readable medium in the form of volatile memory, such as random access memory (RAM) 111 and / or cache memory 112, and may further include read-only memory (ROM) 113. Memory 11 also stores a computer program that can be executed by processor 12, causing processor 12 to perform the steps of any of the methods described above. Memory 11 may also include a utility 114 having at least one program module 115, such program module 115 including, but not limited to, an operating system, one or more application programs, other program modules, and program data; each or some combination of these examples may include an implementation of a network environment.

[0310] Accordingly, processor 12 can execute the aforementioned computer program and can execute utility 114. Processor 12 may employ one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0311] Bus 13 can represent one or more types of bus structures, including a memory bus or memory controller, peripheral bus, graphics acceleration port, processor, or a local bus using any bus structure with multiple bus structures.

[0312] Electronic device 10 can also communicate with one or more external devices, such as keyboards, pointing devices, Bluetooth devices, etc., and with one or more devices capable of interacting with it, and / or with any device that enables it to communicate with one or more other computing devices (e.g., routers, modems, etc.). This communication can be performed through input / output interface 14. Furthermore, electronic device 10 can communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via network adapter 15. Network adapter 15 can communicate with other modules of electronic device 10 via bus 13. It should be understood that, although not shown in the figures, in practical applications, other hardware and / or software modules can be used in conjunction with electronic device 10, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0313] (Computer-readable storage medium)

[0314] This application also provides a computer-readable storage medium, the specific embodiments of which are consistent with the embodiments and technical effects achieved in the above method embodiments, and some contents will not be repeated.

[0315] A computer-readable storage medium stores a computer program that, when executed by at least one processor, implements the steps of any of the above methods or the functions of any of the above electronic devices.

[0316] A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. In embodiments of this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0317] Computer-readable storage media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable storage medium may also be any computer-readable medium capable of sending, propagating, or transmitting a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, or any suitable combination thereof. Program code for performing operations of the present invention may be written in any combination of one or more programming languages, including Java, C++, Python, C#, JavaScript, PHP, Ruby, Swift, Go, Kotlin, etc. The program code may be executed entirely on a user's computing device, partially on a user's device, as a standalone software package, partially on a user's device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing devices can be connected to user equipment via any type of network, including local area networks (LANs) or wide area networks (WANs), or they can be connected to external computing devices (e.g., via the Internet using an Internet service provider).

[0318] (Computer program products)

[0319] This application also provides a computer program product, the specific embodiments of which are consistent with the embodiments and technical effects achieved in the above method embodiments, and some contents will not be repeated.

[0320] This application provides a computer program product comprising a computer program that, when executed by at least one processor, implements the steps of any of the above-described methods or the functions of any of the above-described electronic devices.

[0321] See Figure 11 , Figure 11 This is a schematic diagram of the structure of a computer program product provided in an embodiment of this application.

[0322] The computer program product is used to implement the steps of any of the above methods or to implement the functions of any of the above electronic devices. The computer program product may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the computer program product of the present invention is not limited thereto, and the computer program product may employ any combination of one or more computer-readable media.

[0323] This application describes the invention from the perspectives of purpose, performance, progress, and novelty, and it meets the functional enhancement and use requirements emphasized by the Patent Law. The above description and drawings are merely preferred embodiments of this application and are not intended to limit this application. Therefore, all structures, devices, features, etc., that are similar to or identical to those of this application, i.e., all equivalent substitutions or modifications made in accordance with the scope of this patent application, shall fall within the scope of protection of this patent application.

Claims

1. A semiconductor testing method, characterized in that, include: Obtain flatness test information for the product; The product has positioning edges, and the flatness detection information includes flatness data, which is used to indicate the flatness of each unit area. Obtain surface defect detection information of the product; The surface defect detection information includes defect type data; Based on the flatness detection information and the surface defect detection information, obtain the flatness classification result and / or the surface defect classification result; The flatness classification result includes the flatness category of each unit area in the flatness detection information, and the surface defect classification result includes the defect category of each target defect in the surface defect detection information; The step of obtaining flatness classification results and / or surface defect classification results based on the flatness detection information and the surface defect detection information includes: Based on the flatness data, find one or more unit areas whose flatness is less than the preset flatness. For each identified unit area, based on the positioning edge, the corresponding defect type data is obtained from the surface defect detection information to detect whether there is a target defect within the unit area; Set the flatness category of the unit area with the target defect to the first flatness category, and / or set the defect category of the target defect within the unit area with the target defect to the first defect category. Set the flatness category per unit area for defects without a target to the second flatness category.

2. The semiconductor testing method according to claim 1, characterized in that, The flatness detection information also includes a three-dimensional flatness image corresponding to the flatness data; The surface defect detection information also includes a two-dimensional particle detection image corresponding to the defect type data and its corresponding defect location data.

3. The semiconductor testing method according to claim 2, characterized in that, The method further includes: Based on the flatness category of each unit area in the flatness detection information, image encoding data for each unit area is obtained, and a first display image of the product is generated based on the image encoding data of each unit area and the three-dimensional flatness image. In the first display image, the set of display parameters corresponding to unit areas of different flatness categories is different; and / or, Based on the defect type data and defect location data in the surface defect detection information, the area of ​​each defect category in the three-dimensional flatness image is obtained; based on the area of ​​each defect category in the three-dimensional flatness image, the image encoding data of the area of ​​each defect category is obtained; based on the image encoding data of the area of ​​each defect category and the three-dimensional flatness image, a second display image of the product is generated, in which the set of display parameters corresponding to the area of ​​different defect categories is different.

4. The semiconductor testing method according to claim 1, characterized in that, After setting the defect category of the target defect within a unit area of ​​the first flatness category to the first defect category, the method further includes: For each unit area of ​​the first flatness category, perform the following processing: Based on the positioning edge, the size information of the target defect within the unit area is obtained from the surface defect detection information, and it is determined whether the size of the target defect within the unit area is greater than a preset size. When the size of the target defect within the unit area is greater than the preset size: the flatness category of the unit area is adjusted to the third flatness category, and / or the defect category of the target defect within the unit area is adjusted to the second defect category; When the size of the target defect within the unit area is not greater than the preset size: the flatness category of the unit area is adjusted to the fourth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the third defect category.

5. The semiconductor testing method according to claim 1 or 4, characterized in that, The aspect ratio gradually decreases from the first preset aspect ratio to the Nth preset aspect ratio, where N is a positive integer. The method further includes: For each unit area of ​​either the first or third flatness category, perform the following processing: Obtain the aspect ratio corresponding to the unit area; When the aspect ratio is greater than the first preset aspect ratio: the flatness category of the unit area is adjusted to the fifth flatness category, and / or the defect category of the target defect within the unit area is adjusted to the fourth defect category; When the aspect ratio is greater than the kth preset aspect ratio and not greater than the (k-1)th preset aspect ratio: the flatness category of the unit area is adjusted to the (k+4)th flatness category, and / or the defect category of the target defect within the unit area is adjusted to the (k+3)th defect category, where k is an integer greater than 1 and not greater than N.

6. The semiconductor testing method according to claim 5, characterized in that, The step of obtaining the aspect ratio corresponding to the unit area includes: Based on the flatness detection information, the lateral distance and longitudinal distance of the undulating region corresponding to the unit area are obtained, wherein the lateral distance between the nearest valley points at both ends of the undulating region is taken as the lateral distance of the undulating region, and the longitudinal distance between the peak point of the undulating region and its nearest valley point is taken as the longitudinal distance of the undulating region. The ratio of the longitudinal distance and the lateral distance of the undulating region is calculated and used as the aspect ratio corresponding to the unit area.

7. The semiconductor testing method according to claim 5, characterized in that, The method further includes: Based on the flatness detection information, the flatness category of each unit area, and the defect category of each target defect, a third display image of the product is generated. In the third display image, the set of display parameters corresponding to the unit area of ​​different flatness categories is different, and the set of display parameters corresponding to the target defects of different defect categories is different.

8. An electronic device, characterized in that, The electronic device includes a memory and at least one processor, the memory storing a computer program, and the at least one processor being configured to execute the computer program to perform the following steps: Obtain flatness detection information of the product; the product has positioning edges, and the flatness detection information includes flatness data, which is used to indicate the flatness of each unit area; Obtain surface defect detection information of the product; The surface defect detection information includes defect type data; Based on the flatness detection information and the surface defect detection information, obtain the flatness classification result and / or the surface defect classification result; The flatness classification result includes the flatness category of each unit area in the flatness detection information, and the surface defect classification result includes the defect category of each target defect in the surface defect detection information; The step of obtaining flatness classification results and / or surface defect classification results based on the flatness detection information and the surface defect detection information includes: Based on the flatness data, find one or more unit areas whose flatness is less than the preset flatness. For each identified unit area, based on the positioning edge, the corresponding defect type data is obtained from the surface defect detection information to detect whether there is a target defect within the unit area; Set the flatness category of the unit area with the target defect to the first flatness category, and / or set the defect category of the target defect within the unit area with the target defect to the first defect category. Set the flatness category per unit area for defects without a target to the second flatness category.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by at least one processor, implements the steps of the method of any one of claims 1-7 or the function of the electronic device of claim 8.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by at least one processor, implements the steps of the method of any one of claims 1-7 or the function of the electronic device of claim 8.

Images