Inspection method of multi-layer multi-material photomask
By constructing a three-dimensional data architecture and a two-dimensional design layout database, performing light intensity calibration and die-to-database detection, the imaging quality and signal-to-noise ratio problems of multi-layer, multi-material photomasks were solved, and reliable defect identification was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NEW RAY MASK TECHNOLOGY CORP
- Filing Date
- 2026-04-03
- Publication Date
- 2026-05-12
Smart Images
Figure CN122018229A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of semiconductor photolithography technology, and in particular to a method for inspecting multilayer, multimaterial photomasks. Background Technology
[0002] Photomasks are key components in photolithography, acting as templates for patterning semiconductor wafers. With the continued pursuit of Moore's Law and the emergence of new device architectures such as 3D integrated circuits, microelectromechanical systems (MEMS), and advanced packaging, photomask technology has become increasingly complex. These advanced photomasks are no longer simple planar patterns made of a single material. Instead, they incorporate 3D topologies with varying heights on a single substrate and employ multiple materials (e.g., chromium layers of different thicknesses and types) to achieve specific optical and performance goals.
[0003] In the field of photomask inspection, there is no experience or precedent to refer to for this type of product. Due to the parameter limitations of the inspection equipment itself, focusing and optical calibration are required for each material and height to confirm the type of material and the corresponding optical signal value. Using a single focal plane on the inspection equipment cannot produce clear imaging across the entire area, and the differences in the optical properties of the materials mean that a single inspection parameter cannot simultaneously obtain the optimal signal-to-noise ratio for all areas, thus making it impossible to identify whether defects exist on the product.
[0004] The technical challenges existing in the current technology include: 1. Limited Depth of Focal Length and Defocus Issues: Traditional optical inspection systems have a limited depth of focal length. When inspecting masks with significant height differences, it is impossible to keep all structures in sharp focus simultaneously.
[0005] 2. Material Response Inhomogeneity: Different materials (e.g., chromium films of varying thicknesses) possess unique optical properties. Therefore, they will produce different responses when interacting with a given test wavelength and illumination conditions.
[0006] 3. Signal crosstalk and false defects: When attempting to inspect a specific layer or height, signals from adjacent or lower-level structures can cause interference. This interference manifests as optical crosstalk, complicating image analysis. Summary of the Invention
[0007] Therefore, the purpose of this invention is to provide an inspection method for multi-layer, multi-material photomasks, which can effectively adapt to the height differences and material optical property differences on the photomask surface, improve imaging quality and signal-to-noise ratio, reduce false defect signals, and thus reliably identify real defects.
[0008] To address the aforementioned technical problems, this invention provides a method for inspecting multilayer, multimaterial photomasks, comprising the following steps: S1. Verify database construction: S1-1. Through data processing, based on the design pattern, material distribution, and different process steps of the photomask, establish a three-dimensional data architecture that characterizes the material and height properties of the corresponding position of the photomask. S1-2. Obtain multi-layer exposure data for constructing the photomask. Based on the logic of the three-dimensional data architecture, perform Boolean logic operations on the multi-layer exposure data to map the features of different three-dimensional heights onto a two-dimensional plane, and finally generate a two-dimensional design layout database containing only the features of the light-transmitting area and the light-blocking area. S2. Optical calibration and inspection: S2-1, Light Intensity Calibration: Based on the height and material information in the three-dimensional data architecture, select the area corresponding to the maximum light-shielding layer height and the light-transmitting substrate area without light-shielding material on the actual photomask as calibration points. In the transmitted light mode, the minimum luminous flux value of the maximum light-shielding layer area is calibrated as the lower limit of gray level, and the maximum luminous flux value of the light-transmitting substrate area is calibrated as the upper limit of gray level, thereby establishing a full-coverage detection gray level range. S2-2, Die-to-Database Defect Detection: Image acquisition of the actual photomask is performed using transmitted light. The grayscale range established in step S2-1 is used to ensure that the image signals of all other light-shielding material areas below the maximum light-shielding layer height are included in the grayscale threshold of the light-shielding area. The actual acquired image is compared and detected with the two-dimensional design layout database generated in step S1-2 using Die-to-Database. S3. Analysis and Judgment: After acquiring the image of the actual photomask under optimal optical parameters and focusing conditions, the system compares the acquired actual photomask image data with the corresponding expected data in the two-dimensional design layout database to distinguish the defect locations of the actual photomask.
[0009] In one embodiment of the present invention, in steps S1-2 above, the two-dimensional design layout database is a black and white area database, which is a two-dimensional layout located at the same height.
[0010] In one embodiment of the present invention, in step S2-1 above, the region with the maximum chromium layer height on the photomask is selected as the lower limit of the grayscale value, and the chromium-free quartz substrate region is selected as the upper limit of the grayscale value, establishing a dynamic detection range of 0-255, so that the grayscale value of the chromium layer height below the maximum height naturally falls into the threshold judgment range of the shading area under transmitted light imaging.
[0011] In one embodiment of the present invention, in step S2-2 above, the system automatically focuses on different height regions of the photomask to ensure that layers of different heights are at the optimal focal plane when acquiring images.
[0012] In one embodiment of the present invention, in step S2-1 above, the light intensity is calibrated as follows: Based on the height and material information in the three-dimensional data architecture, the region corresponding to the highest chromium layer height and the chromium-free quartz substrate region are selected as calibration points on the actual photomask; in the transmitted light and / or reflected light mode, the minimum luminous flux value of the highest chromium layer region is calibrated as the lower limit of grayscale, and the maximum luminous flux value of the quartz substrate region is calibrated as the upper limit of grayscale, thereby establishing a full-coverage detection grayscale range.
[0013] The above-mentioned inspection method for multi-layer, multi-material photomasks of the present invention includes steps such as inspection database construction, optical calibration and inspection, analysis and judgment. By constructing a three-dimensional data architecture, generating a two-dimensional design layout database, performing light intensity calibration and die-to-database detection, it solves the problems of defocusing, low signal-to-noise ratio and false defects caused by height differences and material differences in multi-layer, multi-material photomasks that traditional methods cannot handle. It can effectively adapt to height differences on the photomask surface and differences in material optical properties, improve imaging quality and signal-to-noise ratio, reduce false defect signals, and thus reliably identify real defects. Attached Figure Description
[0014] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0015] Figure 1 These are actual product images of a single layer or the same height of the present invention; Figure 2 This is the invention Figure 1 The database format corresponding to the physical products in the product; Figure 3 These are actual product images of the present invention, showing multiple layers or different heights. Figure 4 This is the invention Figure 3 The database format corresponding to the physical product in the product; Figure 5 It is the luminous flux intensity in the transmitted light mode of this invention; Figure 6 This is a specific embodiment of the inspection method for multilayer, multimaterial photomasks of the present invention. Detailed Implementation
[0016] The embodiments of this application will now be described in detail with reference to the accompanying drawings.
[0017] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0018] It should be noted that the following description covers various aspects of embodiments within the scope of the appended claims. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.
[0019] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.
[0020] Additionally, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that practice can be carried out without these specific details.
[0021] Traditional photomask inspection methods suffer from limitations when dealing with multi-layered, multi-material photomasks. These limitations stem from the complex three-dimensional topology and the varying optical properties of different materials, resulting in a limited depth of focus that prevents clear imaging across the entire surface. Furthermore, the uneven response of different materials to the detection light signal can lead to signal crosstalk, causing false defects and severely impacting the accuracy of defect identification.
[0022] Therefore, this invention proposes a method for inspecting multilayer, multimaterial photomasks, comprising the following steps: S1. Verify database construction: S1-1. Through data processing, based on the design pattern, material distribution, and different process steps of the photomask, establish a three-dimensional data architecture that characterizes the material and height properties of the corresponding position of the photomask. S1-2. Obtain multi-layer exposure data for constructing the photomask. Based on the logic of the three-dimensional data architecture, perform Boolean logic operations on the multi-layer exposure data to map the features of different three-dimensional heights onto a two-dimensional plane, and finally generate a two-dimensional design layout database containing only the features of the light-transmitting area and the light-blocking area. S2. Optical calibration and inspection: S2-1, Light Intensity Calibration: Based on the height and material information in the three-dimensional data architecture, select the area corresponding to the maximum light-shielding layer height and the light-transmitting substrate area without light-shielding material on the actual photomask as calibration points. In the transmitted light mode, the minimum luminous flux value of the maximum light-shielding layer area is calibrated as the lower limit of gray level, and the maximum luminous flux value of the light-transmitting substrate area is calibrated as the upper limit of gray level, thereby establishing a full-coverage detection gray level range. S2-2, Die-to-Database Defect Detection: Image acquisition of the actual photomask is performed using transmitted light. The grayscale range established in step S2-1 is used to ensure that the image signals of all other light-shielding material areas below the maximum light-shielding layer height are included in the grayscale threshold of the light-shielding area. The actual acquired image is compared and detected with the two-dimensional design layout database generated in step S1-2 using Die-to-Database. S3. Analysis and Judgment: After acquiring the image of the actual photomask under optimal optical parameters and focusing conditions, the system compares the acquired actual photomask image data with the corresponding expected data in the two-dimensional design layout database to distinguish the defect locations of the actual photomask.
[0023] The data structure comprises several key components: a three-dimensional data architecture representing the material type and height information of the photomask at different locations; a data processing architecture built based on design graphics, material distribution, and process steps; and a multi-layer exposure data set representing the graphic data of each layer used to construct the photomask, typically containing geometric shape and position information at different levels, reflecting the photomask manufacturing process. Boolean logic operations, a Boolean algebra-based arithmetic method, are used to combine, subtract, or intersect graphic data to generate or simplify complex graphics. A two-dimensional design layout database, after processing and mapping, is a set of planar graphic data containing only the features of the transparent and shaded areas, serving as a benchmark for comparing actual photomask images. Luminous flux represents the amount of light energy passing through a cross-section, reflecting the optical transmittance or reflectance of the transparent or shaded area in photomask detection. A grayscale lower limit, set in image processing as the lowest light signal value for a shaded area, is used to classify areas below this value as shaded. Grayscale upper limit refers to the threshold set in image processing as the highest light signal value in a transparent area; areas exceeding this value are considered transparent. Die-to-Database comparison inspection is a defect detection method that identifies deviations or defects on the actual photomask by comparing the actual acquired photomask image with a pre-generated two-dimensional design layout database at the pixel level or feature level.
[0024] The above method begins with the construction of a verification database. This step aims to provide an accurate design baseline for subsequent defect detection. For example, the geometry and number of material layers of the photomask can be manually entered, or preliminary data can be obtained by reading standard design documents. In some cases, design data may exist in multiple formats, requiring format conversion for unified processing.
[0025] Specifically, in step S1-1, through data processing, a three-dimensional data architecture characterizing the material and height attributes of corresponding locations on the photomask is established based on the photomask's design, material distribution, and different process steps. This architecture can be constructed in various ways. For example, it can be based on 3D model data exported from CAD software, converted into an internally recognizable data format containing the material type and Z-axis height information of each voxel. Alternatively, according to design specifications, layers of graphics can be superimposed programmatically, with material attributes and relative heights assigned to each layer, thus forming a logical 3D structural description. The establishment of this 3D data architecture provides accurate raw information for subsequently mapping complex 3D features to a 2D plane.
[0026] Further, in steps S1-2, multi-layer exposure data for constructing the photomask is acquired. Based on the logic of the aforementioned three-dimensional data architecture, Boolean logic operations are performed on the multi-layer exposure data to map features at different three-dimensional heights onto a two-dimensional plane, ultimately generating a two-dimensional design layout database containing only the features of the light-transmitting and light-shielding areas. Specifically, the multi-layer exposure data can be viewed as a series of two-dimensional graphic layers, each representing a pattern under a specific process step. Boolean logic operations can include superposition, subtraction, or intersection operations of graphics; for example, performing a logical "OR" operation on all light-shielding material layers to obtain a comprehensive light-shielding area. When mapping to the two-dimensional plane, a simple orthogonal projection method can be used to project the light-shielding features at all heights onto a single plane, forming a two-dimensional planar image that clearly distinguishes between the light-transmitting and light-shielding areas. This two-dimensional design layout database serves as a reference for subsequent die-to-database comparison, simplifying the complexity of the three-dimensional structure and facilitating rapid image comparison.
[0027] Following this, optical calibration and inspection are performed. This step aims to ensure accurate and consistent image data during the inspection process and to effectively identify defects. For example, the light source intensity and focus position can be manually adjusted to accommodate the inspection requirements of different photomasks. During image acquisition, a standard CCD camera can be used to capture images, and the image data is then transmitted to the processing unit.
[0028] Specifically, in step S2-1, light intensity calibration is performed. Based on the height and material information in the aforementioned three-dimensional data architecture, regions corresponding to the maximum shading layer height and transparent substrate regions without shading material are selected as calibration points on the actual photomask. In transmitted light mode, the minimum luminous flux value of the maximum shading layer region is calibrated as the lower grayscale limit, and the maximum luminous flux value of the transparent substrate region is calibrated as the upper grayscale limit, thereby establishing a full-coverage detection grayscale range. For example, the region with the thickest shading material on the photomask can be manually selected as the calibration point for the lower grayscale limit, and a completely transparent substrate region can be selected as the calibration point for the upper grayscale limit. By measuring the actual luminous flux at these points and mapping it to preset grayscale values, such as 0 and 255, a linear or non-linear grayscale mapping relationship is established. This calibration process ensures that image signals from regions with different materials and heights can be uniformly quantized, avoiding misjudgments caused by differences in optical response.
[0029] Further, in step S2-2, die-to-database defect detection is performed. Transmitted light is used to acquire images of the actual photomask. Utilizing the grayscale range established in step S2-1, the image signals of all other light-shielding material areas below the maximum light-shielding layer height are included in the grayscale threshold of the light-shielding area. The acquired images are then compared with the two-dimensional design layout database generated in step S1-2 using a die-to-database comparison. For example, a fixed-focal-length microscopic imaging system can be used to scan the photomask, acquiring images region by region. After grayscale calibration, the pixel values of the acquired image data are normalized to a preset grayscale range. Subsequently, these image data are subjected to a pixel-level logical XOR operation with the corresponding areas in the two-dimensional design layout database; any mismatched areas are marked as potential defects. This comparison detection method can quickly identify differences between the actual photomask and the design layout.
[0030] Finally, analysis and judgment are performed. After acquiring images of the actual photomask under optimal optical parameters and focusing conditions, the system compares the acquired image data with the corresponding expected data in the two-dimensional design layout database to identify the locations of defects in the actual photomask. For example, the comparison results can generate a defect map, marking all areas that do not conform to the design. These areas can be further classified using image processing algorithms, such as distinguishing different types of defects like dimensional deviations, edge roughness, foreign objects, or missing patterns. Ultimately, the system can output a detailed defect report for operators to review and make decisions. This analysis and judgment process ensures the accuracy and efficiency of defect identification.
[0031] This method effectively solves the signal crosstalk and pseudo-defect problems caused by the complex three-dimensional structure of multi-layer, multi-material photomasks by constructing a three-dimensional data architecture and mapping it to a two-dimensional design layout database. Simultaneously, by using light intensity calibration based on height and material information, it achieves the normalization of the optical responses of different materials, overcoming the challenge of signal inhomogeneity. Therefore, it ensures that clear and accurate detection images can still be obtained under limited depth of focus conditions, improving the reliability of defect identification.
[0032] Furthermore, in steps S1-2 above, the two-dimensional design layout database is a black and white area database, which is a two-dimensional layout located at the same height.
[0033] Specifically, the two-dimensional design layout database is limited to a black-and-white area database. This means that the graphic data stored in the database contains only two distinct states: completely transparent (usually represented as white) and completely opaque (usually represented as black), without any intermediate grayscale values. This binarization is crucial for ensuring the accuracy of defect detection. In the detection of multi-layer, multi-material photomasks, subtle differences in transmittance due to different materials and thicknesses can occur. If the database contains grayscale information, these normal optical differences may be misjudged as defects when compared with the actual acquired images, or the true defects may be masked by signal crosstalk. By strictly limiting the database to a black-and-white area, this ambiguity can be eliminated. For example, when generating the database, the original design data can be thresholded, setting one or more thresholds to mark all areas with transmittance above a specific threshold as transparent areas and areas below that threshold as opaque areas. Alternatively, based on design rules and material properties, specific material areas with light-blocking functions (such as chromium layers) in the design graphic can be directly defined as light-blocking areas, and areas without material (such as quartz substrates) can be defined as light-transmitting areas, thereby directly generating binary graphic data.
[0034] Meanwhile, the 2D design layout database is limited to 2D layouts located at the same height. This means that all features in the 3D structure of the photomask, regardless of their original design height, are logically projected or mapped onto a single 2D plane, forming a planar layout without height information. This approach aims to solve the problem of limited depth of focus in traditional optical inspection systems. By unifying all features to the same height plane, image blurring or inaccurate detection due to defocusing of features at different heights during the inspection process can be avoided. This allows the inspection system to perform comparisons on a unified focal plane, thereby simplifying the focusing process and improving inspection efficiency and accuracy. For example, the maximum projection method can be used to project all stacked light-blocking features in the 3D data vertically onto a 2D plane. As long as light-blocking material exists at a certain vertical position, the corresponding position on that 2D plane is marked as light-blocking. Alternatively, based on the Boolean logic operation results of the 3D solid data architecture, a logical "OR" operation is performed on the light-blocking layers at different heights in the vertical direction, ultimately generating a composite light-blocking region on a 2D plane.
[0035] By defining the two-dimensional design layout database as a black-and-white area database and a two-dimensional layout at the same height, the problems of signal crosstalk, false defects, and defocusing caused by height differences due to unclear database structure are effectively solved. The black-and-white area database eliminates the risk of misjudgment caused by uneven optical responses of different materials, thereby improving the accuracy of defect identification. At the same time, mapping all features to a two-dimensional layout at the same height allows the detection system to perform image acquisition and comparison on a unified focal plane, avoiding image blurring and detection errors caused by limited depth of focus, and significantly improving the reliability and efficiency of detection.
[0036] Furthermore, referring to Figure 5 As shown, in step S2-1 above, the region with the maximum chromium layer height on the photomask is selected as the lower limit of the grayscale value, and the chromium-free quartz substrate region is selected as the upper limit of the grayscale value, establishing a dynamic detection range of 0-255. This ensures that the grayscale value of chromium layer heights below the maximum height naturally falls within the threshold judgment range of the shading area under transmitted light imaging. In transmitted light mode, the grayscale value of chromium layer regions with light flux intensity below the maximum height naturally falls within the 0-255 range of the shading area threshold.
[0037] Specifically, the region with the highest chromium layer height on the photomask is selected as the lower limit of grayscale value, aiming to set a reliable grayscale lower limit benchmark for the shading area for the detection system. In multi-layer, multi-material photomasks, the chromium layer is the main shading material, and chromium layers of different heights exhibit different luminous flux values under transmitted light. The region with the highest chromium layer height usually corresponds to the strongest shading effect, i.e., the lowest transmitted luminous flux. Using it as the lower limit of grayscale ensures that the grayscale values of all other chromium layers with lower heights (whose transmittance may be slightly higher) are included within the determination range of the shading area. For example, by analyzing the design data of the photomask in advance, a specific pattern or region with the highest height in all chromium layer regions can be identified. During actual detection, the detection beam is focused on this region to measure the luminous flux, and its corresponding image signal value is set as the lower limit of grayscale. Alternatively, the entire photomask can be scanned in transmitted light mode to find the region with the lowest luminous flux, and combined with the design data to confirm that this region is the region with the highest chromium layer height, and then its luminous flux value is calibrated as the lower limit of grayscale.
[0038] Simultaneously, a chromium-free quartz substrate area is selected as the upper limit of grayscale values to set the benchmark for the upper limit of grayscale values in the transparent area of the detection system. The chromium-free quartz substrate area represents the completely transparent area on the photomask, with the highest transmitted light flux. Using it as the upper limit ensures that the grayscale values of all transparent areas (including areas with possible slight absorption or scattering) are included within the judgment range of the transparent area and effectively distinguished from the shaded areas. For example, by analyzing the design data of the photomask, an area entirely composed of a quartz substrate can be identified, and during detection, the detection beam can be focused on this area to measure the light flux, and the corresponding image signal value can be set as the upper limit of grayscale. Alternatively, a chromium-free quartz substrate area can be reserved at the edge of the photomask or in a specific test structure as a calibration point, and its light flux can be measured before detection, with the value calibrated as the upper limit of grayscale.
[0039] Building upon this, establishing a dynamic detection range of 0-255 means mapping the aforementioned lower and upper grayscale limits to a standard digital image grayscale range, such as 0-255 for an 8-bit image. This dynamic range allows the detection system to fully utilize the dynamic range of the image sensor, accurately distinguishing different luminous flux levels, thereby improving detection sensitivity and accuracy. For example, a linear mapping can be used to map the minimum measured luminous flux value (corresponding to the lower grayscale limit) to 0, the maximum luminous flux value (corresponding to the upper grayscale limit) to 255, and intermediate luminous flux values proportionally to grayscale values between 0 and 255. Alternatively, non-linear mapping, such as gamma correction or logarithmic mapping, can be used to better adapt to human visual characteristics or specific detection needs, but the core principle remains the same: mapping the physical luminous flux range to the digital grayscale range.
[0040] Through the above technical solution, this application proposes a more accurate and robust method for establishing grayscale ranges during the light intensity calibration process, addressing the complexity of multi-layer, multi-material photomasks. By selecting the region with the highest chromium layer height on the photomask as the lower limit of grayscale values, the system can capture the strongest shading signal. Using this as a benchmark, it ensures that image signals from all other lower-height chromium layers (whose transmittance may be slightly higher than the highest chromium layer) can be correctly identified as shading areas, avoiding signal deviations caused by uneven material response. Simultaneously, selecting a chromium-free quartz substrate region as the upper limit of grayscale values provides a clear boundary for the transparent regions. Based on this, a dynamic detection range of 0-255 is established, enabling the entire detection process to fully utilize the dynamic range of the image sensor, achieving comprehensive coverage and accurate differentiation of all transparent and shading areas on the photomask. This calibration strategy ensures that the grayscale value of chromium layers below the maximum height naturally falls within the threshold judgment range of the shaded area under transmitted light imaging, eliminating the need for complex layer-by-layer calibration or adjustment for chromium layers of different heights. This significantly reduces signal crosstalk and the occurrence of false defects, improving the accuracy and efficiency of defect detection.
[0041] Furthermore, in step S2-2 above, the system automatically focuses on different height regions of the photomask to ensure that layers at different heights are at the optimal focal plane during image acquisition. Automatic focusing refers to the detection system's ability to dynamically adjust the focal plane of the optical system based on the actual height information of the area to be detected, ensuring that it remains consistent with the surface of the current detection area. Its function is to overcome the limitations of the limited depth of focus in traditional optical systems, ensuring that features at different heights can be clearly imaged on complex structures with significant height differences, such as multi-layer, multi-material photomasks.
[0042] Specifically, the methods for achieving autofocus can include, but are not limited to, the following: One implementation involves the system pre-storing the height information of the photomask's three-dimensional data architecture. During image acquisition, the system automatically calculates and adjusts the Z-axis position of the objective lens or stage based on the height value of the currently scanned area within the three-dimensional data architecture, ensuring that the focal plane precisely falls on that height region.
[0043] Another approach is for the system to determine whether the focal plane is in an optimal state by analyzing the acquired image signals in real time. For example, the system can use image processing algorithms such as contrast detection, edge sharpness analysis, or Fourier transform to evaluate image sharpness. When the system detects that the image sharpness has reached its maximum value, it considers the focal plane to be in the optimal position. The system will continuously perform this real-time feedback adjustment to adapt to the focusing needs of different height areas.
[0044] Another approach is to integrate a laser ranging module into the system to measure the precise height of each area on the photomask surface in real time before or during image acquisition. Based on the measured height data, the system drives a Z-axis actuator (such as a piezoelectric ceramic driver or a voice coil motor) to quickly adjust the focal plane, thereby achieving precise focusing on areas at different heights.
[0045] Ensuring that layers at different heights are at the optimal focal plane during image acquisition refers to the plane where the optical system images the sharpest. Through an autofocus mechanism, the surface of any area of the photomask can be clearly imaged by the optical system, avoiding image blurring and loss of detail caused by defocusing. This is crucial for accurately identifying minute defects.
[0046] By employing the aforementioned technical solution, this application addresses the defocusing problem in multi-layer, multi-material photomasks at different heights. It introduces an autofocus mechanism that enables the detection system to dynamically adjust the focal plane based on the actual height of different areas of the photomask. Specifically, during image acquisition, the system no longer uses a single fixed focal plane but intelligently identifies the height of the area to be inspected and adjusts the focus position of the optical system accordingly. This ensures that image acquisition is performed on the optimal focal plane regardless of the height layer of the photomask. This autofocus capability directly solves the defocusing problem caused by limited depth of focus in traditional detection methods, avoiding image blurring and loss of detail. By ensuring clear, high-resolution images across all height areas, the accuracy and reliability of defect detection are greatly improved. Even minute defects located at different heights can be clearly captured and identified, effectively preventing missed detections or misjudgments and enhancing the overall performance of the photomask inspection method.
[0047] In step S2-1 above, light intensity calibration is performed as follows: Based on the height and material information in the three-dimensional data architecture, the region corresponding to the highest chromium layer height and the chromium-free quartz substrate region are selected as calibration points on the actual photomask; in transmitted light and / or reflected light modes, the minimum luminous flux value of the highest chromium layer region is calibrated as the lower limit of grayscale, and the maximum luminous flux value of the quartz substrate region is calibrated as the upper limit of grayscale, thereby establishing a full-coverage detection grayscale range.
[0048] Specifically, "based on the height and material information in the aforementioned three-dimensional data architecture" refers to fully utilizing the three-dimensional structure and material distribution data of the photomask determined during the design phase when performing light intensity calibration. The three-dimensional data architecture records in detail the material type (e.g., chromium, quartz, etc.) and its height information in the vertical direction at each location on the photomask. This utilization of precise design data ensures that the calibration process accurately reflects the actual physical characteristics of the photomask. In implementation, the system can directly read and parse the photomask's design files (such as GDSII files or CAD data), which typically contain geometry, layer information, and material properties. By processing and parsing these raw design data, the height and material information of each region can be extracted. Alternatively, a dedicated database can be pre-built, generated based on the processed and optimized raw design files. In this database, each spatial coordinate point is pre-associated with its corresponding material type and height value, allowing for direct querying and retrieval during calibration. "Selecting the region corresponding to the highest chromium layer height and the chromium-free quartz substrate region on the actual photomask as calibration points" clarifies the selection strategy for light intensity calibration points. Selecting the "region with the highest chromium layer height" aims to capture the most opaque or reflective state on the photomask, as the chromium layer is typically used for light blocking, and the higher the height, the more significant the light-blocking effect. Selecting the "chromium-free quartz substrate region" aims to capture the most transparent or least reflective state (for transmission mode) on the photomask, since the quartz substrate is transparent. By selecting these two extreme regions as calibration points, it can be ensured that the established grayscale range covers all possible optical responses on the photomask. In practice, the system can automatically identify and locate specific patterned regions with the highest chromium layer height and blank areas consisting entirely of a quartz substrate on the actual photomask using image recognition algorithms combined with information from the three-dimensional data architecture. Alternatively, the operator can manually select and mark the highest chromium layer region and the chromium-free quartz substrate region as calibration points on the graphical user interface of the detection system, based on the visualization of the three-dimensional data architecture. The "in transmitted light and / or reflected light mode" option introduces flexibility in detection modes, allowing light intensity calibration in transmitted light mode, reflected light mode, or a combination of both. Transmitted light mode is suitable for detecting the attenuation or phase change of light after it passes through a material, while reflected light mode is suitable for detecting the reflective properties of a material surface. This flexibility in mode selection is particularly important for multi-layer, multi-material photomasks, as different materials and structures may exhibit optimal signal responses under different illumination modes. In practice, the detection system can be equipped with a switchable illumination module, for example, selectively activating the transmitted or reflected light source via mechanical switching or electronic control. When transmitted light mode is selected, the light source is located below the photomask; when reflected light mode is selected, the light source is located above the photomask and works in conjunction with the detector.Alternatively, the detection system can employ an integrated illumination design, such as using a beam splitter and multiple light sources, allowing transmitted and reflected light to simultaneously or alternately illuminate the photomask. The system can automatically select or combine these two modes for luminous flux measurement based on preset material properties or detection requirements. The description of "calibrating the minimum luminous flux value of the highest chromium layer region as the lower grayscale limit and the maximum luminous flux value of the quartz substrate region as the upper grayscale limit" details the quantitative calibration process for the grayscale range. For the highest chromium layer region, its "minimum luminous flux value" represents the darkest signal achievable in that region under the selected detection mode, and is set as the lowest point of the grayscale (e.g., 0 in 8-bit grayscale). For the chromium-free quartz substrate region, its "maximum luminous flux value" represents the brightest signal achievable in that region, and is set as the highest point of the grayscale (e.g., 255 in 8-bit grayscale). This calibration method ensures that the physical meaning of the grayscale range closely corresponds to the actual optical response of the photomask. In practice, the detection system uses a high-precision photodetector (such as a CCD or CMOS sensor) to perform multiple luminous flux measurements at selected calibration points, taking the statistically minimum (for the highest chromium layer region) and maximum (for the quartz substrate region) values. These values are then directly mapped to the upper and lower limits of a preset grayscale range. Alternatively, after acquiring image data from the calibration points, the system performs pixel value analysis on these images. For the highest chromium layer region, the lowest point in its pixel value distribution is identified as the lower grayscale limit; for the quartz substrate region, the highest point in its pixel value distribution is identified as the upper grayscale limit. "Establishing a fully covered detection grayscale range" indicates the ultimate purpose and effect of the above calibration process. By setting the luminous flux values of the darkest and brightest areas on the photomask as the lower and upper grayscale limits, respectively, the established grayscale range can completely cover all intermediate optical signals that may be generated by materials and height structures on the photomask. This means that in subsequent image acquisition and defect detection processes, the image signals of all regions will fall within the defined grayscale range, avoiding signal overflow or truncation, thus ensuring the comprehensiveness and accuracy of the detection. In implementation, the system transforms the original pixel values of all subsequently acquired actual photomask images to this established full-coverage grayscale range according to a linear or non-linear mapping relationship, ensuring the consistency and comparability of image data. Alternatively, this grayscale range can be used as a benchmark for defect detection algorithms. For example, when performing threshold segmentation, a reasonable threshold can be set based on this range to distinguish between transparent areas, shading areas, and potential defect areas.
[0049] By utilizing the height and material information in the aforementioned three-dimensional data architecture, this application can accurately identify and select representative calibration points on the photomask, namely the region with the highest chromium layer height and the chromium-free quartz substrate region. This calibration point selection method based on actual physical properties ensures high accuracy and specificity in establishing the grayscale range, effectively avoiding the problem of incomplete grayscale range caused by improper selection of calibration points. Furthermore, by introducing transmitted light and / or reflected light modes for light intensity calibration, this application overcomes the limitations that a single transmitted light mode may encounter when processing multi-layer, multi-material photomasks. Different materials and structures have different optical responses in transmitted and reflected modes. Flexible selection or combination of detection modes allows the system to better adapt to the differences in optical properties of various materials on the photomask, thereby obtaining superior signal quality and signal-to-noise ratio. Based on this, the minimum luminous flux value of the highest chromium layer region is precisely calibrated as the lower limit of the grayscale, and the maximum luminous flux value of the quartz substrate region is precisely calibrated as the upper limit of the grayscale. This ensures that the established detection grayscale range can completely cover all possible optical signals on the photomask, including the entire dynamic range from the least opaque to the most transparent. This allows the image signals of all other light-shielding material regions below the maximum light-shielding layer height to be accurately included within the grayscale threshold of the light-shielding area in the subsequent Die-to-Database defect detection step S2-2, significantly improving the accuracy and reliability of defect identification, especially for photomasks with complex three-dimensional structures and multiple materials.
[0050] The following provides a more specific example to illustrate the above technical solution in greater detail. Figure 6 As shown: Suppose we need to inspect a multilayer, multimaterial photomask used in advanced semiconductor manufacturing. This photomask comprises chromium layers of varying thicknesses (e.g., a thicker chromium layer and a thinner chromium layer) and a quartz substrate, forming a three-dimensional topological structure with varying heights on the photomask. Traditional inspection methods struggle to accurately identify defects in such complex structures due to limited depth of focus, inhomogeneous optical responses of different materials, and signal crosstalk.
[0051] To address these issues, this method first constructs a verification database. In step S1-1, the system establishes a three-dimensional data architecture characterizing the material type and height attributes at corresponding locations on the photomask, based on the photomask's design pattern, material distribution (e.g., thick chromium layer, thin chromium layer, quartz), and different manufacturing processes. This architecture precisely records the material type and height information on the Z-axis for each point on the photomask. For example, it records that one region is a thick chromium layer with a height of H1; another region is a thin chromium layer with a height of H2; and the height of the quartz substrate region is H0.
[0052] Next, in steps S1-2, the system acquires the multi-layer exposure data used to construct the photomask. Based on the previously established three-dimensional data architecture, the system performs Boolean logic operations on this multi-layer exposure data. Through this operation, features at different heights in three-dimensional space (such as chromium layers of different thicknesses) are mapped onto a two-dimensional plane. Ultimately, a two-dimensional design layout database containing only the features of the light-transmitting and light-shielding areas is generated. This database is a black-and-white area database that uniformly represents all light-shielding materials (regardless of their original height and thickness) as light-shielding areas and all light-transmitting areas as light-transmitting areas, forming a two-dimensional layout at the same height. Unlike traditional methods that directly use single-layer design data, this step effectively avoids the complexity caused by multi-layer structures and different material thicknesses through three-dimensional to two-dimensional mapping, providing a unified reference benchmark for subsequent defect detection.
[0053] Subsequently, the optical calibration and verification stage begins. In the optical intensity calibration step S2-1, the system selects two key calibration points on the actual photomask based on the height and material information in the three-dimensional data architecture. Specifically, the region corresponding to the highest chromium layer height (e.g., the thickest chromium layer region) is selected as the lower grayscale limit calibration point, and the chromium-free quartz substrate region is selected as the upper grayscale limit calibration point. In transmitted light mode, the system calibrates the minimum luminous flux value of the highest chromium layer region as the lower grayscale limit (e.g., set to grayscale value 0), and calibrates the maximum luminous flux value of the quartz substrate region as the upper grayscale limit (e.g., set to grayscale value 255). In this way, a fully covered dynamic detection grayscale range of 0-255 is established. This calibration method ensures that even other chromium layers (such as thin chromium layers) with a height lower than the highest chromium layer height can naturally fall within the threshold judgment range of the shaded area under transmitted light imaging, thereby solving the problem of uneven optical response of different materials and avoiding missed detections or misjudgments caused by differences in grayscale values.
[0054] In the Die-to-Database defect detection step S2-2, the system uses transmitted light to acquire images of the actual photomask. Unlike traditional methods that struggle to simultaneously and clearly image all height layers with limited depth of focus, this method automatically focuses on different height regions of the photomask during image acquisition, ensuring that layers of different heights are at the optimal focal plane, thus solving the problems of limited depth of focus and defocus. Using the grayscale range established in step S2-1, the system processes the acquired image signals, ensuring that image signals from all other light-shielding material regions below the height of the highest chromium layer are included within the grayscale threshold of the light-shielding area. This means that regardless of whether the chromium layer is thick or thin, as long as it is a light-shielding material, its image signal will be correctly identified as a light-shielding area. Subsequently, the system performs Die-to-Database comparison and detection between the actual acquired images and the two-dimensional design layout database generated in step S1-2. This comparison is based on a unified black-and-white two-dimensional layout, effectively avoiding signal crosstalk and false defects caused by differences in three-dimensional structure and materials in traditional methods.
[0055] Finally, in the analysis and judgment of step S3, after acquiring an image of the actual photomask under optimal optical parameters and focusing conditions, the system accurately compares the acquired actual photomask image data with the corresponding expected data in the two-dimensional design layout database. Through this comparison, the system can accurately distinguish the location of defects on the actual photomask; for example, light transmission occurs in areas that should be light-blocking areas, or light blocking occurs in light-transmitting areas, thereby achieving accurate inspection of multi-layer, multi-material photomasks.
[0056] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.
Claims
1. A method for inspecting multilayer, multi-material photomasks, characterized in that, Includes the following steps: S1. Verify database construction: S1-1. Through data processing, based on the design pattern, material distribution, and different process steps of the photomask, establish a three-dimensional data architecture that characterizes the material and height properties of the corresponding position of the photomask. S1-2. Obtain multi-layer exposure data for constructing the photomask. Based on the logic of the three-dimensional data architecture, perform Boolean logic operations on the multi-layer exposure data to map the features of different three-dimensional heights onto a two-dimensional plane, and finally generate a two-dimensional design layout database containing only the features of the light-transmitting area and the light-blocking area. S2. Optical calibration and inspection: S2-1, Light Intensity Calibration: Based on the height and material information in the three-dimensional data architecture, select the area corresponding to the maximum light-shielding layer height and the light-transmitting substrate area without light-shielding material on the actual photomask as calibration points. In the transmitted light mode, the minimum luminous flux value of the maximum light-shielding layer area is calibrated as the lower limit of gray level, and the maximum luminous flux value of the light-transmitting substrate area is calibrated as the upper limit of gray level, thereby establishing a full-coverage detection gray level range. S2-2, Die-to-Database Defect Detection: Image acquisition of the actual photomask is performed using transmitted light. The grayscale range established in step S2-1 is used to ensure that the image signals of all other light-shielding material areas below the maximum light-shielding layer height are included in the grayscale threshold of the light-shielding area. The actual acquired image is compared and detected with the two-dimensional design layout database generated in step S1-2 using Die-to-Database. S3. Analysis and Judgment: After acquiring the image of the actual photomask under optimal optical parameters and focusing conditions, the system compares the acquired actual photomask image data with the corresponding expected data in the two-dimensional design layout database to distinguish the defect locations of the actual photomask.
2. The inspection method for multilayer, multi-material photomasks according to claim 1, characterized in that: In steps S1-2 above, the two-dimensional design layout database is a black and white area database, which is a two-dimensional layout located at the same height.
3. The inspection method for multilayer, multi-material photomasks according to claim 1, characterized in that: In step S2-1 above, the region with the maximum chromium layer height on the photomask is selected as the lower limit of the grayscale value, and the chromium-free quartz substrate region is selected as the upper limit of the grayscale value. A dynamic detection range of 0-255 is established so that the grayscale value of the chromium layer height below the maximum height naturally falls into the threshold judgment range of the shading area under transmitted light imaging.
4. The inspection method for multilayer, multi-material photomasks according to claim 1, characterized in that: In step S2-2 above, the system will automatically focus on different height areas of the photomask to ensure that layers of different heights are at the optimal focal plane when acquiring images.
5. The inspection method for multilayer, multi-material photomasks according to claim 1, characterized in that: In step S2-1 above, light intensity calibration is performed as follows: Based on the height and material information in the three-dimensional data architecture, the region corresponding to the highest chromium layer height and the chromium-free quartz substrate region are selected as calibration points on the actual photomask; in transmitted light and / or reflected light modes, the minimum luminous flux value of the highest chromium layer region is calibrated as the lower limit of grayscale, and the maximum luminous flux value of the quartz substrate region is calibrated as the upper limit of grayscale, thereby establishing a full-coverage detection grayscale range.