Wafer defect detection apparatus and method

Abstract

The application discloses a wafer defect detection device and method. The device comprises an illumination system, a motion platform, a detection system and a signal processing system. The illumination system provides at least three different incident direction illumination beams to the wafer surface and forms overlapping linear illumination spots; the motion platform drives the wafer to perform linear scanning; the detection system comprises multiple detection channels with coplanar optical axes, collects scattered light and converts it into electrical signals, the multiple detection channels comprise large-angle detection channels with a large angle between the optical axis and the normal line of the wafer, small-angle detection channels with a small angle between the optical axis and the normal line of the wafer, at least one detection channel and at least one illumination beam are located on both sides of the normal line of the wafer; the signal processing system processes the electrical signals to determine the defect information. The application realizes high-sensitivity detection of all types of spherical, spherical-like, flat contaminants and damage defects by configuring large-angle and small-angle detection channels with specific angles and illumination beams with specific directions, and avoids missing detection.

Description

Technical Field

[0001] This application relates to the field of semiconductor manufacturing technology, and in particular to a wafer defect detection device and method. Background Technology

[0002] Wafer defect inspection is a critical quality control step in semiconductor manufacturing, used to detect and locate physical or chemical defects such as spherical or near-spherical additional defects and scratches on the wafer surface. As semiconductor chip manufacturing processes continue to shrink, the sensitivity requirements for detecting harmful defects are becoming increasingly stringent. Dark-field inspection technology, which illuminates the wafer surface and collects scattered light generated by defects, offers advantages in terms of high precision and speed. Among these, line scanning combined with multi-channel imaging detection significantly improves detection speed compared to single-point scanning. However, existing multi-channel inspection schemes still face key technical challenges in detecting all types of defects, making it difficult to achieve high signal-to-noise ratio and high sensitivity for all types of defects, thus leading to missed detections of certain defects. Summary of the Invention

[0003] In view of this, this application provides a wafer defect detection device and method to solve the problem of insufficient sensitivity of multi-channel detection for all types of defects in the prior art.

[0004] To solve the above-mentioned technical problems, one technical solution adopted in this application is: providing a wafer defect detection device, comprising: an illumination system configured to provide at least three illumination beams with different incident directions to the wafer surface, forming overlapping linear illumination spots on the wafer surface; a motion platform for carrying and driving the wafer to move relative to the linear illumination spots; a detection system including multiple detection channels, each detection channel configured to collect scattered light generated by illumination in the linear illumination spot area and convert it into an electrical signal; and a signal processing system configured to process the electrical signal output by the detection system to determine defect information; wherein, The optical axes of multiple detection channels are all located on the same detection plane, which is perpendicular to the wafer surface. The long axis of the linear illumination spot has a first angle with the detection plane, which is in the range of 85 degrees to 95 degrees. The multiple detection channels include at least one large-angle detection channel and at least one small-angle detection channel. The angle between the optical axis of the large-angle detection channel and the wafer normal is greater than a second preset angle, and the angle between the optical axis of the small-angle detection channel and the wafer normal is less than the second preset angle. The projection of one of the at least three illumination beams onto the detection plane and at least one detection channel are located on opposite sides of the wafer normal.

[0005] As a further improvement of this application, at least three illumination beams include a first illumination beam, a second illumination beam, and a third illumination beam. The second illumination beam is located on the detection plane. The first illumination beam is perpendicular to the plane formed by the wafer normal, the detection plane, and the plane where the wafer is located. The second illumination beam and at least one detection channel are located on opposite sides of the wafer normal.

[0006] As a further improvement of this application, the angle between the first illumination beam and the wafer normal is between 60° and 80°, the angle between the second illumination beam and the wafer normal is between 30° and 60°, and the angle between the third illumination beam and the wafer normal is between 0° and 5°.

[0007] As a further improvement of this application, the first illumination beam, the second illumination beam, and the third illumination beam have the same wavelength and are sequentially illuminated during wafer scanning to complete the scan; or, the first illumination beam, the second illumination beam, and the third illumination beam all have different wavelengths and are simultaneously illuminated during wafer scanning to complete the scan, and the detection system is equipped with an element that distinguishes different wavelengths; or, two of the first illumination beam, the second illumination beam, and the third illumination beam have the same wavelength, and the third illumination beam has a different wavelength, and during wafer scanning, the illumination beam with the different wavelength is simultaneously illuminated by one of the two illumination beams with the same wavelength to complete the scan, and the detection system is equipped with an element that distinguishes different wavelengths.

[0008] As a further improvement of this application, the multiple detection channels include a large-angle detection channel and a small-angle detection channel, with the large-angle detection channel and / or the small-angle detection channel located on opposite sides of the wafer normal, respectively, along with the second illumination beam.

[0009] As a further improvement of this application, the multiple detection channels include two large-angle detection channels and one small-angle detection channel. The two large-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two large-angle detection channels and the wafer normal is within a preset angle difference range; or, the multiple detection channels include one large-angle detection channel and two small-angle detection channels. The two small-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two small-angle detection channels and the wafer normal is within a preset angle difference range.

[0010] As a further improvement of this application, the multiple detection channels include two large-angle detection channels and two small-angle detection channels. The two large-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two large-angle detection channels and the wafer normal is within a preset angle difference range. The two small-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two small-angle detection channels and the wafer normal is within a preset angle difference range.

[0011] As a further improvement of this application, the angle between the optical axis of the large-angle detection channel and the wafer normal is between 45° and 65°, and the angle between the optical axis of the small-angle detection channel and the wafer normal is between 15° and 25°.

[0012] As a further improvement of this application, the long axis of the linear illumination spot is perpendicular to the detection plane, and the width of the linear illumination spot is preset according to the depth of field of the detection channel.

[0013] To address the aforementioned technical problems, another technical solution adopted in this application is: providing a wafer defect detection method, which is applied to one of the wafer defect detection devices mentioned above; the method includes: an illumination system irradiating the wafer surface with illumination beams from at least three different incident directions to form linear illumination spots; a motion platform driving the wafer to move, causing the linear illumination spots to scan the wafer surface; a detection system using multiple detection channels to collect scattered light generated in the linear illumination spot area during the scanning process and converting it into multi-channel electrical signals; and a signal processing system processing the multi-channel electrical signals and fusing information from different illumination beams and different detection channels to identify defects and defect size information.

[0014] As a further improvement of this application, the signal processing system processes multi-channel electrical signals and fuses information from different illumination beams and different detection channels to identify defects and their size information. This includes: acquiring the location of the defect signal intensity and the corresponding defect signal intensity under any illumination beam and any detection channel, and determining whether the defect signal intensity meets a predefined saturation state; mapping the defect signal intensity to intermediate defect sizes based on the state of the defect signal intensity; using predefined fusion logic to fuse multiple intermediate defect sizes from different illumination beams and different detection channels in pairs until the final defect size is obtained; and outputting the location of the defect and the final defect size.

[0015] As a further improvement of this application, mapping the defect signal intensity to the intermediate defect size based on the saturation state includes: when the defect signal intensity is less than or equal to 0, the intermediate defect size is 0; when the defect signal intensity is greater than 0 and less than or equal to a first preset threshold, a pre-calibrated calibration curve is queried to obtain the intermediate defect size, where the calibration curve refers to the mapping curve between the defect signal intensity and the defect size; when the defect signal intensity is greater than the first preset threshold, the pixels in the location area corresponding to the defect signal are clustered and merged according to the intensity distribution to obtain the total pixel area occupied by the defect signal, and then the intermediate defect size is calculated by combining the total pixel area and the magnification of the corresponding detection channel.

[0016] As a further improvement of this application, according to a predefined fusion logic, multiple intermediate defect sizes from different illumination beams and different detection channels are fused in pairs, including: determining whether the two intermediate defect sizes to be fused are being fused for the first time; if so, then taking a value from the two intermediate defect sizes based on the collection efficiency of the detection channel, or taking the average value of the two intermediate defect sizes as the fusion value; if not, then fusion is performed based on the source type of the two intermediate defect sizes.

[0017] As a further improvement of this application, the collection efficiency of the detection channel is used to select values ​​from two intermediate defect sizes, including: taking the larger value of the two intermediate defect sizes and comparing it with a second preset threshold; when the larger value of the two intermediate defect sizes does not exceed the second preset threshold, taking the intermediate defect size corresponding to the detection channel with higher collection efficiency as the fusion value; when the larger value of the two intermediate defect sizes exceeds the second preset threshold, taking the intermediate defect size corresponding to the detection channel with lower collection efficiency as the fusion value.

[0018] As a further improvement of this application, fusion is performed based on the source type of the two intermediate defect sizes, including: when both intermediate defect sizes are obtained from defect signal intensity in a non-saturated state, the larger of the two values ​​is taken as the fusion value; when the two intermediate defect sizes are obtained from defect signal intensity in a non-saturated state and defect signal intensity in a saturated state, respectively, the intermediate defect size obtained from the defect signal intensity in a saturated state is taken as the fusion value; when both intermediate defect sizes are obtained from defect signal intensity in a saturated state, the average of the two values ​​is taken as the fusion value; when at least one of the two intermediate defect sizes is obtained by calculating the total pixel area of ​​the defect signal, the larger of the two values ​​is taken as the fusion value.

[0019] The beneficial effects of this application are: The wafer defect detection device of this application defines multiple detection channels with coplanar optical axes and the detection plane perpendicular to the wafer surface. The detection system includes at least one large-angle detection channel and one small-angle detection channel, and ensures that at least one illumination beam and one detection channel are located on opposite sides of the wafer normal. The large-angle detection channel collects scattering images of spherical or near-spherical additional defects, and the small-angle detection channel collects scattering images of damaged defects. The illumination beam and detection channel distributed on both sides of the normal are combined to realize scattering images of disk-shaped, flat cylindrical, hemispherical additional contamination and other defects, thereby ensuring high signal-to-noise ratio detection for various defect types and avoiding missed detection of certain defects.

[0020] The wafer defect detection device achieves clear imaging across the entire long axis by making the long axis of the linear illumination spot perpendicular or nearly perpendicular to the detection plane, i.e., the first angle is in the range of 85° to 95°. The object distance from all points of the linear spot to the detection lens in the long axis direction is the same or nearly equal (1% error), thereby improving the detection accuracy. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the structure of one embodiment of the wafer defect detection device of the present invention; Figure 2 This is a schematic diagram of the optical path principle of one embodiment of the wafer defect detection device of the present invention; Figure 3 These are top views (a) and side views (b) of an embodiment of the wafer defect detection device of the present invention when the optical axis is perpendicular to the wafer surface. Figure 4 This is a diagram showing the substrate noise and scattered light distribution of different types of defects in one embodiment of the wafer defect detection device of the present invention; Figure 5 This is a schematic diagram of the optical path principle of the dual-detection channel of one embodiment of the wafer defect detection device of the present invention; Figure 6 This is a schematic diagram of the optical path principle of a three-detection channel in one embodiment of the wafer defect detection device of the present invention; Figure 7 This is a schematic diagram of the optical path principle of a four-detection channel in one embodiment of the wafer defect detection device of the present invention; Figure 8 This is a schematic flowchart of an embodiment of the wafer defect detection method of the present invention; Figure 9 This is a schematic diagram of the calibration curve of one embodiment of the wafer defect detection method of the present invention; Figure 10 This is a schematic diagram of the four-channel defect size fusion process under three illumination beams in one embodiment of the wafer defect detection method of the present invention; Figure 11 This is a schematic diagram of the four-channel defect size fusion process under three illumination beams in another embodiment of the wafer defect detection method of the present invention; Figure 12 This is a schematic diagram of the two-channel defect size fusion process under three illumination beams in one embodiment of the wafer defect detection method of the present invention; Figure 13 This is a schematic diagram of the two-channel defect size fusion process under three illumination beams in another embodiment of the wafer defect detection method of the present invention; Figure 14 This is a schematic diagram of the three-channel defect size fusion process under three illumination beams in one embodiment of the wafer defect detection method of the present invention; Figure 15 This is a schematic diagram of the four-channel defect size fusion process under two illumination beams in one embodiment of the wafer defect detection method of the present invention; Figure 16 This is a schematic diagram of the two-channel defect size fusion process under two illumination beams in one embodiment of the wafer defect detection method of the present invention; Figure 17 This is a schematic diagram of the three-channel defect size fusion process under two illumination beams in one embodiment of the wafer defect detection method of the present invention; Figure 18 This is a schematic diagram of the four-channel defect size fusion process under two illumination beams in another embodiment of the wafer defect detection method of the present invention; Figure 19 This is a schematic diagram of the two-channel defect size fusion process under two illumination beams in another embodiment of the wafer defect detection method of the present invention; Figure 20 This is a schematic diagram of the three-channel defect size fusion process under two illumination beams in another embodiment of the wafer defect detection method of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0023] The terms "first," "second," and "third" in this application are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this application are only used to explain the relative spatial positions and movements of components in a specific orientation (as shown in the figures). If the specific orientation changes, the directional indications also change accordingly. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0024] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0025] We found that the scattered light distribution of defects is directional and closely related to the defect type. The scattering characteristics of different types of defects, such as spherical, near-spherical, flat contaminants, and damage defects, are significantly different.

[0026] Figure 1 This is a schematic diagram of the structure of a wafer defect detection device according to an embodiment of the present invention. Figure 1 As shown, the wafer defect detection device 100 includes: an illumination system 1, a motion platform 2, a detection system 3, and a signal processing system 4.

[0027] Please refer to the following: Figure 2 The illumination system 1 is configured to provide at least three illumination beams with different incident directions to the wafer surface, forming an overlapping linear illumination spot on the wafer surface. Specifically, the illumination source in the illumination system 1 preferably uses a laser of one or more wavelengths in the 266 to 532 nm band. The illumination system 1 provides at least three illumination beams with different incident directions, and the illumination beams are precisely optically adjusted and calibrated to overlap and form a linear illumination spot on the wafer surface.

[0028] The motion platform 2 is used to carry and drive the wafer to move relative to the linear illumination spot. Specifically, the motion platform 2 carries the wafer and makes it move in a straight line with high precision and uniform speed relative to the fixed illumination spot, so that the linear illumination spot scans the entire surface area of ​​the wafer.

[0029] The detection system 3 includes multiple detection channels, each configured to collect scattered light generated by the linear illumination spot area and convert it into an electrical signal. Specifically, the detection system 3 includes multiple detection channels, each of which is a complete optical imaging system, including an objective lens, a lens group, and a detector. The detector is preferably a line detector, parallel to the linear illumination spot. The optical axes of all detection channels lie within the same detection plane to form structural constraints, greatly simplifying the system calibration process.

[0030] The signal processing system 4 is configured to process the electrical signals output by the detection system 3 to determine defect information. Specifically, the signal processing system 4 is used to receive and process the electrical signals from the detection system 3.

[0031] In this configuration, the optical axes of multiple detection channels are all located on the same detection plane, which is perpendicular to the wafer surface. The long axis of the linear illumination spot forms a first angle with the detection plane, ranging from 85 to 95 degrees. The multiple detection channels include at least one large-angle detection channel and at least one small-angle detection channel. The angle between the optical axis of the large-angle detection channel and the wafer normal is greater than a second preset angle, while the angle between the optical axis of the small-angle detection channel and the wafer normal is less than the second preset angle. The projection of one of the at least three illumination beams onto the detection plane and at least one detection channel are located on opposite sides of the wafer normal. It should be noted that the second preset angle is preset. In some embodiments, the first angle is preferably set to 90°, and the second preset angle is preferably set to 45°.

[0032] In this embodiment, the scattered light distribution of defects on the wafer is affected by the defect morphology. The main defect morphologies include: spherical, near-spherical, disk-shaped, flat cylindrical, and hemispherical additional defects, as well as damage defects. Additional defects mainly originate from contaminants falling into the cavity, while damage defects mainly originate from wafer manufacturing process errors such as polishing and epitaxial processes. In this embodiment, the scattered light distribution characteristics of different types of defects are analyzed, and damage defects are classified into three categories: The first type is low to medium aspect ratio additional defects. These defects are characterized by their similar lateral and longitudinal dimensions (lateral refers to the direction parallel to the wafer surface, and longitudinal refers to the direction perpendicular to the wafer surface), and are mainly spherical or near-spherical additional contaminants; for example... Figure 4 Figure (b) shows the intensity distribution of a typical scattered field from a spherical defect with a diameter on the order of tens of nanometers, where the arrows indicate the incident direction. It can be observed that spherical defects with sizes on the order of tens of nanometers exhibit greater scattered light intensity at large exit angles, especially in the region with the incident angle reversed. However, the unavoidable roughness of the wafer surface also causes scattering, resulting in system stray light, which is mainly distributed in the region with the incident angle reversed.

[0033] The second category is additional defects with high aspect ratios, mainly including additional contamination in the form of discs, flat cylinders, and hemispherical shapes. Figure 4 Figure (c) shows the typical scattered field intensity distribution of disk-like defects with lateral dimensions on the order of tens of nanometers, where the arrows indicate the incident direction. Unlike the wide range of scattered light angle distribution of defects with medium to low aspect ratios, defects with high aspect ratios are mainly distributed in the reverse region of the incident angle.

[0034] The third type is damage defects, which manifest as partial missing parts of the wafer substrate. Figure 4 Figure (e) describes a typical scattered light distribution of a damage defect with a lateral size on the order of tens of nanometers, where the incident light is incident vertically downwards. Figure 4(d) shows the substrate noise distribution under approximately perpendicular incidence. Since the reflected light corresponds to the center position and has extremely high intensity, the central reflected light portion is not shown in the figure. Both damage-related defects and the background noise at this point are mainly distributed in the small-angle annular region, with the background noise being more concentrated.

[0035] To address the three types of defects mentioned above, this embodiment employs an illumination system 1 to provide at least three illumination beams with different incident directions, illuminating the wafer surface to form linear illumination spots. Multiple detection channels are used: a large-angle detection channel with an angle greater than a second preset angle is used to cover the detection of low to medium aspect ratio defects; a small-angle detection channel with an angle less than a second preset angle is used to cover the detection of damage defects; and an illumination beam located on either side of the wafer normal, along with at least one detection channel, is used to cover the detection of high aspect ratio defects, thus completing the detection of all types of wafer defects.

[0036] This embodiment defines multiple detection channels with coplanar optical axes and a detection plane perpendicular to the wafer surface. Simultaneously, it ensures a first angle between the long axis of the linear illumination spot and the detection plane, ranging from 85 to 95 degrees. The object distance from all points of the linear spot along its long axis to the detection lens is the same or nearly equal (1% error), thus achieving relatively clear imaging along the long axis. Furthermore, the detection system 3 includes at least one large-angle detection channel and one small-angle detection channel, ensuring that at least one illumination beam and one detection channel are located on opposite sides of the wafer normal. The large-angle detection channel collects scattering images of spherical and near-spherical additional defects, while the small-angle detection channel collects scattering images of damaged defects. Combined with the illumination beam and detection channel distributed on both sides of the normal, scattering images of disk-shaped, flat cylindrical, and hemispherical additional contamination defects are obtained, ensuring high signal-to-noise ratio detection for various defect types and avoiding missed detection of certain defects.

[0037] Furthermore, existing multi-channel detection technologies suffer from defocusing issues. When the optical axis of the detection channel is not perpendicular to the wafer surface, its object-side focal plane forms an angle with the wafer surface, causing different locations within the linear detection area to be unable to be clearly imaged simultaneously. Existing technologies typically employ Scham's law to design the imaging system, but this system is inferior to classical imaging methods in terms of key optical parameters such as aberrations and field of view. Therefore, in some embodiments, the long axis of the linear illumination spot is perpendicular to the detection plane, and the width of the linear illumination spot is preset according to the depth of field of the detection channel.

[0038] It should be noted that depth of field refers to the range of object space depth within which an optical system can obtain a clear image. For channels where the optical axis is not perpendicular to the wafer surface, the focal plane is tilted. By designing the width of the linear illumination spot (i.e., the direction of the minor axis of the linear illumination spot) to be sufficiently narrow, the change in object distance (i.e., the defocusing amount) within the entire width range is less than the depth of field of the lens, thereby ensuring that the entire line area is clear in the image. Therefore, in this embodiment, the width of the linear illumination spot is preset according to the depth of field of the detection channel. Figure 3 As shown, Figure 3 The diagram describes a top view (a) and a side view (b) when the optical axis is perpendicular to the wafer surface. The illumination area of ​​the illumination system 1 in this embodiment is designed as a linear illumination spot with a large aspect ratio. The long axis of the linear illumination spot is set perpendicular or nearly perpendicular (within ±5°) to the optical axis of the detection channel, and parallel or nearly parallel (within ±5°) to the detection plane of the line detector. At each position along the long axis of the linear illumination spot, it is in a focused or slightly defocused state. At this time, the object-side focal plane along the short axis of the linear illumination spot cannot be parallel to the linear illumination spot on the crystal surface. The center of the object-side focal plane of the detection channel along the short axis coincides with the center of the short axis of the linear illumination spot, that is, the center of the linear illumination spot CD coincides with the center of the object-side focal plane PQ of the detection channel. Figure 3 Point O in (b) is defocused because the defocus direction is located along the minor axis of the linear illumination spot. Under this setting, the maximum defocus CQ value of the linear illumination spot is less than or equal to the depth of field of the detection channel, thus ensuring that all positions within the linear illumination spot are simultaneously in focus. Furthermore, the imaging quality and stability of the detection channel can be improved by compressing the width of the linear illumination spot.

[0039] This embodiment defines multiple detection channels with coplanar optical axes and the detection plane is perpendicular to the wafer surface. At the same time, the long axis of the linear illumination spot is perpendicular or nearly perpendicular to the detection plane. This ensures that the object distance from all points of the linear spot to the detection lens is exactly the same or nearly the same along its long axis, thereby achieving clear imaging across the entire long axis. Furthermore, by pre-setting the width of the linear spot according to the depth of field of the detection channel, the maximum defocusing amount along the short axis of the spot is controlled within the depth of field tolerance of the detection channel, thereby achieving simultaneous focusing on the linear spot area and avoiding defocusing.

[0040] Furthermore, to achieve even better imaging results, in some embodiments, such as Figure 2 As shown, at least three illumination beams include a first illumination beam, a second illumination beam, and a third illumination beam. The second illumination beam is located on the detection plane. The first illumination beam is perpendicular to the plane formed by the wafer normal, the detection plane, and the plane where the wafer is located. The second illumination beam and at least one detection channel are located on opposite sides of the wafer normal.

[0041] Specifically, for defects with medium to low aspect ratios, the detection sensitivity directly depends on the signal-to-noise ratio. Considering the distribution of both defect-scattered light and background stray light, the collection channel should be positioned in a large-angle region of the vertical plane (e.g., Figure 4 (b) shown in white circle has a higher signal-to-noise ratio. For high aspect ratio defects, considering that the background noise is concentrated in a large angle range in the reverse region, the illumination beam and the detection channel are set to be coplanar, and the center of the detection channel is located in a small angle range in the reverse region of the incident plane (e.g., Figure 4 (c) shown in white circle has a higher signal-to-noise ratio.

[0042] Furthermore, in some embodiments, the angle between the first illumination beam and the wafer normal is between 60° and 80°, the angle between the second illumination beam and the wafer normal is between 30° and 60°, and the angle between the third illumination beam and the wafer normal is between 0° and 5°.

[0043] It should be noted that in this embodiment, the first illumination beam lies in a plane perpendicular to both the detection plane and the wafer plane. Using the projection of the plane formed by the first illumination beam and the wafer normal onto the wafer plane as a reference, the azimuth angle of the first illumination beam can be either 0° or 180°, as long as the angle between the first illumination beam and the wafer normal is maintained between 60° and 80°. The angle between the third illumination beam and the wafer normal is between 0° and 5°. Using the projection of the third illumination beam onto the wafer plane as a reference, the azimuth angle of the third illumination beam is between 0° and 360°.

[0044] Specifically, the angle between the first illumination beam and the wafer normal is between 60° and 80° primarily to target low-to-medium aspect ratio defects, such as spherical and near-spherical defects, which are subject to large-angle incidence. For spherical and near-spherical defects located in the Mie scattering region, when incident at a large angle, their scattered light energy will significantly concentrate in the large-angle lateral region. Setting the first illumination beam within this range can most effectively excite the scattering intensity of spherical and near-spherical defects in the direction of the large-angle detection channel. The angle between the second illumination beam and the wafer normal is between 20° and 45° primarily to target high aspect ratio defects, such as disks and flat cylinders. These defects behave similarly to micromirrors. According to the law of reflection, their strongest scattered light (specular reflection light) is located in the opposite direction of the incident light and at the same angle to the normal. The moderate incident angle of the second illumination beam ensures that its strongest reflected light energy falls precisely within the collection range of the small-angle detection channel located on the opposite side. The angle between the third illumination beam and the wafer normal is between 0° and 5° to target damage defects. These defects are essentially topological changes on the wafer substrate and can be regarded as phase objects. Vertical or near-vertical illumination can most effectively convert this phase difference into intensity modulation, highlighting the defect signal. Under vertical illumination, the scattered light from damage defects is mainly distributed in a small-angle annular region close to the normal. The small-angle detection channel is designed to efficiently collect signals from this region.

[0045] Preferably, in some embodiments, for the three types of defects mentioned above—low aspect ratio defects, high aspect ratio defects, and damage defects—a customized lighting system 1 and a detection system 3 can be selected according to requirements. Specifically: When only low-to-medium aspect ratio additional defects and high aspect ratio additional defects need to be detected, the illumination system 1 of this embodiment preferably uses a combination of a first illumination beam and a second illumination beam, and the detection system 3 preferably uses a large-angle detection channel with an angle greater than a second preset angle to cover the detection of low-to-medium aspect ratio additional defects, and uses an illumination beam located on both sides of the wafer normal and at least one detection channel to cover the detection of high aspect ratio additional defects.

[0046] When only high aspect ratio additional defects and damage defects need to be detected, the illumination system 1 of this embodiment preferably uses a combination of a second illumination beam and a third illumination beam, and the detection system 3 preferably uses a small-angle detection channel with an angle less than a second preset angle to cover the detection of damage defects, and uses an illumination beam located on both sides of the wafer normal and at least one detection channel to cover the detection of high aspect ratio additional defects.

[0047] This embodiment adjusts the number and incident angle of the illumination beam, as well as the number and detection angle of the detection channels, according to requirements to ensure that specific types of defects can be detected as needed, thus making it suitable for more application scenarios.

[0048] Furthermore, in some embodiments, the first illumination beam, the second illumination beam, and the third illumination beam have the same wavelength and are illuminated sequentially during the wafer scanning process to complete the scan.

[0049] Specifically, the first, second, and third illumination beams use lasers of the exact same wavelength. During the scanning process, they are controlled to illuminate sequentially. For example, within one line scan integration cycle, illumination 1 is illuminated for the first 1 / 3 of the time, illumination 2 for the middle 1 / 3, and illumination 3 for the last 1 / 3. The detection system 3 continuously acquires data throughout the cycle, but the signal processing system 4 separates the signals according to the timing and assigns them to the corresponding illumination conditions. Only one beam is active at any given time, avoiding signal crosstalk between different illumination beams. Therefore, it is suitable for applications where detection speed requirements are not extremely stringent, but high accuracy, low noise, and system stability are sought.

[0050] Furthermore, in some other embodiments, the first illumination beam, the second illumination beam, and the third illumination beam all have different wavelengths and are illuminated simultaneously during the wafer scanning process to complete the scan. The detection system 3 is equipped with elements that distinguish between different wavelengths.

[0051] Specifically, the three illumination beams employ three distinct wavelengths (e.g., first illumination beam: 266nm deep ultraviolet, second illumination beam: 355nm ultraviolet, third illumination beam: 532nm green light). During scanning, the three beams simultaneously illuminate the wafer surface. The detection system 3 requires elements to distinguish between the different wavelengths, such as a combination of dichroic mirrors and bandpass filters before each detection channel, to separate the mixed scattered light by wavelength and guide it to different detectors or different regions of the detectors. This method offers extremely high detection speed, acquiring defect information under three illumination conditions simultaneously in a single scan. The detection throughput is theoretically nearly three times higher than that of the time-division multiplexing mode, making it suitable for large-scale production lines with extremely high detection speed requirements.

[0052] Furthermore, in some other embodiments, two of the first illumination beam, the second illumination beam, and the third illumination beam have the same wavelength, and the other has a different wavelength. During the wafer scanning process, the illumination beam with the different wavelength is simultaneously illuminated by one of the two illumination beams with the same wavelength to complete the scan. The detection system 3 is equipped with an element that distinguishes between different wavelengths.

[0053] Specifically, of the three illumination beams, two have the same wavelength (e.g., the first and third illumination beams use wavelength λ1), and the third has a different wavelength (the second illumination beam uses wavelength λ2). During scanning, a partially simultaneous strategy is adopted: that is, the second illumination beam with a different wavelength can work simultaneously with either the first or third illumination beam with the same wavelength. The detection system 3 also needs to be configured with a beam splitter, but only needs to distinguish between the λ1 and λ2 bands. This method strikes a balance between speed and complexity. It is faster than a purely time-division multiplexing mode (because partial illumination is simultaneous), and simpler and less expensive than a three-wavelength simultaneous mode system. Moreover, it offers flexible scheduling, allowing simultaneous illumination combinations to be allocated according to the priority of defect detection. For example, if high aspect ratio defects are the main problem on the production line, the second illumination beam (specifically targeting such defects) and the third illumination beam (specifically targeting damage defects) can work simultaneously for focused monitoring. Therefore, it is suitable for most semiconductor manufacturing and packaging processes that have certain requirements for speed and cost.

[0054] Furthermore, based on the above embodiments, in other embodiments, the angle between the optical axis of the large-angle detection channel and the wafer normal is between 45° and 65°, and the angle between the optical axis of the small-angle detection channel and the wafer normal is between 15° and 25°.

[0055] Furthermore, based on the above embodiments, in other embodiments, such as Figure 5 As shown, the multiple detection channels include a large-angle detection channel and a small-angle detection channel, with the large-angle detection channel and / or the small-angle detection channel located on opposite sides of the wafer normal, respectively, along with the second illumination beam.

[0056] Specifically, in this embodiment, the detection system 3 adopts a simplified dual-channel configuration, including a large-angle detection channel and a small-angle detection channel. The angle between the optical axis of the large-angle detection channel and the wafer normal is set in the range of 45° to 65°, preferably 55°. The numerical aperture (NA) of this channel is 0.1 to 0.3, preferably 0.2, and the magnification is set to 1 to 3X. The angle between the optical axis of the small-angle detection channel and the wafer normal is set in the range of 15° to 25°, preferably 20°. The numerical aperture (NA) of this channel is 0.1 to 0.2, preferably 0.15, and the magnification is also 1 to 3X. The large-angle detection channel and the small-angle detection channel can be located on opposite sides or on the same side of the wafer normal, wherein at least one detection channel needs to be located on opposite sides of the wafer normal from the second illumination beam.

[0057] Furthermore, based on the above embodiments, in other embodiments, such as Figure 6As shown, the multiple detection channels include two large-angle detection channels and one small-angle detection channel. The two large-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two large-angle detection channels and the wafer normal is within a preset angle difference range. Alternatively, multiple detection channels may include a large-angle detection channel and two small-angle detection channels. The two small-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two small-angle detection channels and the wafer normal is within a preset angle difference range.

[0058] Specifically, this embodiment provides two extended three-channel configuration schemes, which improve system performance by adding a channel of the same type to the dual-channel configuration. It should be noted that the preset angle difference range is pre-set, preferably 0°~5°.

[0059] The first scheme includes two large-angle detection channels and one small-angle detection channel. The optical axes of the two large-angle detection channels are located on opposite sides of the wafer normal, and the angle difference between them and the normal is less than or equal to 5°, i.e., they are approximately symmetrically arranged. For example, the first large-angle detection channel has an angle of θ=50°, and the second large-angle detection channel has an angle of θ=55°. The small-angle detection channel is on the same side as either the first or second large-angle detection channel.

[0060] It should be noted that in this embodiment, the two large-angle detection channels can be set to different collection efficiencies (such as different NA values ​​or gains). When the defect size is small, a high-sensitivity detection channel is used; when the defect size is large and causes the high-sensitivity channel to saturate, it automatically switches to a low-sensitivity detection channel, thereby expanding the effective detection range of the system. Furthermore, a separate detection channel can be set for the second illumination beam among the three detection channels, enabling simultaneous detection of the second illumination beam and the first or third illumination beam, thus improving the detection speed.

[0061] The second scheme includes a large-angle detection channel and two small-angle detection channels. The optical axes of the two small-angle detection channels are located on opposite sides of the wafer normal, and the angle difference between them and the normal is less than or equal to 5°, i.e., they are approximately symmetrically arranged. For example, the first small-angle detection channel has an angle of θ=18°, and the second small-angle detection channel has an angle of θ=22°. This configuration allows for a dedicated detection channel for the second illumination beam, enabling simultaneous operation when the second illumination beam uses different wavelengths from the first and third illumination beams. For example, one detection channel collects only the signal from the second illumination beam through a filter, while the other two detection channels collect signals from either the first or third illumination beam, thereby improving the detection speed.

[0062] Furthermore, based on the above embodiments, in other embodiments, such as Figure 7As shown, the multiple detection channels include two large-angle detection channels and two small-angle detection channels. The two large-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two large-angle detection channels and the wafer normal is within a preset angle difference range. The two small-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two small-angle detection channels and the wafer normal is within a preset angle difference range.

[0063] Specifically, this embodiment provides the most comprehensive four-channel configuration, including two large-angle detection channels and two small-angle detection channels. The optical axes of the two large-angle detection channels are located on opposite sides of the wafer normal, with an angle difference ≤ 5°. Similarly, the optical axes of the two small-angle detection channels are also located on opposite sides of the wafer normal, with an angle difference ≤ 5°. This embodiment can provide the most comprehensive detection coverage for three illumination beams. For example, the second and third illuminations can be configured to operate simultaneously, each with one or even two dedicated detection channels. This achieves the fastest detection speed while ensuring high sensitivity. For any illumination direction, the system can acquire signals from two symmetrical channels for fusion processing, maximizing the signal-to-noise ratio, reliability, and detection accuracy.

[0064] In this disclosure, each detection channel includes a corresponding detection module, and each detection module includes a detection lens and a detector. The detection lens is used to collect the scattered light beam from the object to be detected and to image the object to be detected onto the detection camera.

[0065] In this disclosure, the detector may include, but is not limited to, a CCD detector, a TDI detector, a photomultiplier tube (PMT), an avalanche photodiode (APD), etc.

[0066] The light source is used to send a beam of light of a preset wavelength to the object to be detected at a third preset angle. The light source can be a red laser, a blue laser, an ultraviolet light collector, an LED light source, etc. The embodiments of this application do not specifically limit the type of light source.

[0067] Optionally, the lighting system 1 may be configured with a light source modulation component, which includes a beam expander module and a shaping module.

[0068] A beam expander module can be used to enlarge the beam diameter to meet the requirements of subsequent shaping modules. The beam expander module can be a one-dimensional beam expander element, designed based on cylindrical lens pairs and deformable prism pairs. It is mainly used to adjust the beam size along the narrow axis, ensuring the output beam reaches a predetermined size to meet the needs of subsequent shaping optical elements. The larger the beam size, the narrower the compressible linewidth. The narrow axis is the linewidth direction corresponding to the final line spot, also known as the short side direction of the line spot.

[0069] The shaping module can be a collimated flat-top shaping element or a freeform surface shaping element. The freeform surface shaping element can provide single-degree-of-freedom linewidth compression along the narrow side direction. The shaping module can include any of the following: diffractive optical elements (DOE), microlens arrays, prisms, etc. The shaping module can also include a one-dimensional beam expander, a collimated flat-top shaping element, a freeform surface shaper, or any other suitable shaping module known in the art. Figure 8 A schematic flowchart of one embodiment of the wafer defect detection method of the present invention is shown. This wafer defect detection method is applied to the wafer defect detection apparatus of one of the above embodiments. Figure 8 As shown, the wafer defect detection method includes: Step S1: The illumination system illuminates the wafer surface with illumination beams from at least three different incident directions to form linear illumination spots.

[0070] Specifically, in this embodiment, at least three illumination beams with different incident directions can be illumination beams of the same wavelength or illumination beams of different wavelengths, or some illumination beams can have the same wavelength. The on and off of the illumination beams are related to the wavelength setting. For example, taking three illumination beams as an example, it can be specifically described as follows: 1. Time-sharing illumination with the same wavelength: Controls three illumination beams of the same wavelength to be lit in turn in a strict time sequence.

[0071] 2. Simultaneous illumination of multiple wavelengths: Control three different wavelengths of illumination beams to light up simultaneously, using wavelength differences to distinguish the signal source.

[0072] 3. Mixed lighting: Controlling two lights with the same wavelength to partially illuminate another light with a different wavelength simultaneously.

[0073] Step S2: The motion platform moves the wafer, causing the linear illumination spot to scan the wafer surface.

[0074] Specifically, a high-precision air-bearing platform is employed, with its motion control strictly synchronized with the row trigger signal of the detection system. The scanning path typically consists of closely adjacent parallel lines, ensuring complete coverage of the entire wafer surface. The scanning speed is optimized based on the illumination mode, detector integration time, and required detection accuracy.

[0075] Step S3: The detection system uses multiple detection channels to collect the scattered light generated by the linear illumination spot area during the scanning process and converts it into a multi-channel electrical signal.

[0076] Specifically, the linear array detectors in all detection channels receive a unified trigger signal, ensuring that they capture images of scattered light from identical linear regions on the wafer surface at the same time. Each channel independently performs photoelectric conversion, outputting a one-dimensional sequence of electrical signals containing spatial and intensity information.

[0077] Step S4: The signal processing system processes the multi-channel electrical signals and fuses information from different illumination beams and different detection channels to identify defects and their size information.

[0078] Specifically, the signal processing system receives and processes multiple electrical signals from all detection channels. By fusing information from different lighting conditions and different detection channels, it ultimately identifies defects and determines their size information.

[0079] Furthermore, step S4 specifically includes: 1. Obtain the location of the defect signal intensity under any illumination beam and any detection channel, the corresponding defect signal intensity, and determine whether the defect signal intensity meets the predefined saturation state.

[0080] Specifically, the signal processing system receives the one-dimensional line scan electrical signal (which can be considered as a one-dimensional image) from each illumination beam and each detection channel. After filtering and denoising each line of signal, an adaptive thresholding algorithm (the threshold is dynamically calculated based on the local background) is used to identify continuous pixel regions with significantly higher intensity than the background. These regions are identified as potential defects. The location (coordinates in the scanning direction) of each potential defect and the calculated defect signal intensity are recorded. The calculation process for the defect signal intensity is as follows: ; in, It is the defect signal strength. It is the signal strength of the pixel area detected by the detection channel. It is the noise intensity generated by the wafer substrate in the pixel area. It can be the intensity of the defect-free location on the wafer or the average intensity, or the average intensity of the pixels around the defect signal.

[0081] It should be noted that photodetectors all have a certain saturation threshold (i.e., when the light intensity density incident on a pixel exceeds a certain value, the output of that pixel no longer increases with increasing light intensity density, and directly outputs saturated intensity), which affects signal strength. The validity of the numerical values ​​requires determining whether the signal is in a saturated state after pixel fusion. The determination method is related to the defect area delineation method. It can be determined by the number of saturated pixels (e.g., greater than or equal to 10 pixels) or the proportion of saturated pixels in the total defect signal pixel area (e.g., greater than 30%).

[0082] 2. Map the defect signal intensity to the intermediate defect size based on the state of the defect signal intensity.

[0083] Specifically, after confirming whether the defect signal intensity is in a non-saturated or saturated state, an appropriate mapping rule is selected based on the state of the defect signal intensity to map the defect signal intensity to an intermediate defect size.

[0084] Furthermore, the step of mapping the defect signal intensity to the intermediate defect size based on the saturation state specifically includes: 2.1 When the defect signal strength is less than or equal to 0, the intermediate defect size is 0.

[0085] Specifically, when the calculated defect signal intensity is less than or equal to 0, it indicates that no valid defect signal was detected in that area. The defect signal intensity is the result of subtracting local background noise from the total intensity of the defect pixel area. A defect signal intensity ≤ 0 means that the signal intensity in that area is not higher than the surrounding background noise, and its fluctuations are entirely within the noise range. Therefore, it can be determined that there is no defect at that location or that the defect size is below the system's detection limit.

[0086] 2.2 When the defect signal strength is greater than 0 and less than or equal to the first preset threshold, query the pre-calibrated calibration curve to obtain the intermediate defect size. The calibration curve refers to the mapping curve between the defect signal strength and the defect size.

[0087] Specifically, the first preset threshold is preset. In this embodiment, the first preset threshold may be greater than the critical value of the signal strength in the saturation state. That is, when the defect signal strength is greater than 0 and less than or equal to the first preset threshold, the defect signal strength may be in a non-saturated state or in a saturated state. At this time, the intermediate defect size corresponding to the defect signal strength is confirmed by querying the pre-calibrated calibration curve.

[0088] It should be noted that, Figure 9 A schematic diagram of a calibration curve is shown, such as... Figure 9 As shown, the horizontal axis of the calibration curve represents the defect signal intensity, and the vertical axis represents the size of the intermediate defect. The specific distribution of the calibration curve is directly related to the hardware settings. It can be obtained by experimental calibration. The specific steps are as follows: prepare a series of standard films containing defects of a specified size, use the same equipment and different lighting modes to detect and obtain the signal intensity values ​​of each collection channel, and then substitute the known defect size information to obtain the signal calibration curve.

[0089] 2.3 When the intensity of the defect signal is greater than the first preset threshold, the pixels in the location area corresponding to the defect signal are clustered and merged according to the intensity distribution to obtain the total area of ​​pixels occupied by the defect signal. Then, the intermediate defect size is calculated by combining the total area of ​​pixels and the magnification of the corresponding detection channel.

[0090] Specifically, when the defect signal intensity is greater than the first preset threshold, it indicates that the calibration curve is distorted and has no reference value. The intermediate defect size is no longer obtained based on the signal intensity, but is determined based on the pixel area occupied by the defect signal. The specific method is as follows: the pixels in the location area corresponding to the defect signal are clustered and merged according to the intensity distribution to obtain the total pixel area occupied by the signal. Then, the intermediate defect size is obtained by dividing the total pixel area occupied by the signal by the magnification factor of the detection channel.

[0091] 3. Using predefined fusion logic, multiple intermediate defect sizes from different illumination beams and different detection channels are fused in pairs until the final defect size is obtained.

[0092] Specifically, after obtaining the intermediate defect size, multiple intermediate sizes from different sources, each with a confidence level label, are merged into a final, most reliable defect size.

[0093] For further details, please refer to Figure 10 , Figure 10A flowchart illustrating a four-channel fusion process for obtaining defect size information is provided. Illumination 1, Illumination 2, and Illumination 3 represent the first, second, and third illumination beams, respectively. Channels l1_N1, l1_N2, l1_W1, and l1_W2 represent the defect signal intensities detected by the two small-angle detection channels l1_N1 and l1_N2 and the two large-angle detection channels l1_W1 and l1_W2 under the first illumination beam, respectively. DN1 represents the intermediate defect size after fusing the defect signal intensities from the two small-angle detection channels l1_N1 and l1_N2. DW1... The term "DW2" represents the intermediate defect size after fusing the defect signal intensities of the two large-angle detection channels l1_W1 and l1_W2. Channels l2_N1, l2_N2, l2_W1, and l2_W2 represent the defect signal intensities detected by the two small-angle detection channels l2_N1 and l2_N2 and the two large-angle detection channels l2_W1 and l2_W2 under the second illumination beam, respectively. The defect signal intensities of channels l2_W2 and l2_N2 are not used. DW2 represents the defect signal intensity of channel l2_W1. The intermediate defect size is obtained from the intensity query calibration curve of channel L2_N1. DN2 represents the intermediate defect size obtained from the defect signal intensity query calibration curve of channel L2_N1. DW12 represents the intermediate defect size obtained by fusing DW1 and DW2. Do represents the intermediate defect size obtained by fusing DN1 and DW12. Channels l3_N1, l3_N2, l3_W1, and l3_W2 represent the two small-angle detection channels l3_N1 and l3_N2 and the two large-angle detection channels l3_W1 and l3_W2 under the third illumination beam, respectively. The defect signal intensity detected by each channel is DN3, which represents the intermediate defect size after fusing the defect signal intensity of the two small-angle detection channels l3_N1 and l3_N2. DW3 represents the intermediate defect size after fusing the defect signal intensity of the two large-angle detection channels l3_W1 and l3_W2. DN23 represents the intermediate defect size after fusing DN2 and DN3. Dn represents the intermediate defect size after fusing DN23 and DW3. D represents the final defect size obtained after fusing Do and Dn. When fusing Do and Dn, the larger value of the two is taken.

[0094] For further details, please refer to Figure 11 , Figure 11A flowchart illustrating another embodiment of obtaining defect size information through four-channel fusion is shown. Here, illumination 1, illumination 2, and illumination 3 represent the first, second, and third illumination beams, respectively; channels l1_N1, l1_N2, l1_W1, and l1_W2 represent the defect signal intensities detected by the two small-angle detection channels l1_N1 and l1_N2 and the two large-angle detection channels l1_W1 and l1_W2 under the first illumination beam, respectively; DN1 represents the intermediate defect size after fusing the defect signal intensities from the two small-angle detection channels l1_N1 and l1_N2; and DW1 represents the defect size. The term "DW2" represents the intermediate defect size after fusing the defect signal intensities of the two large-angle detection channels l1_W1 and l1_W2. Channels l2_N1, l2_N2, l2_W1, and l2_W2 represent the defect signal intensities detected by the two small-angle detection channels l2_N1 and l2_N2 and the two large-angle detection channels l2_W1 and l2_W2 under the second illumination beam, respectively. The defect signal intensities of channels l2_W2 and l2_N2 are not used. DW2 represents the defect signal intensity of channel l2_W1. The intermediate defect size obtained from the intensity query calibration curve is DN2, which represents the intermediate defect size obtained from the defect signal intensity query calibration curve of channel L2_N1. DN12 represents the intermediate defect size obtained by fusing DN1 and DN2. Do represents the intermediate defect size obtained by fusing DN12 and DW1. Channels l3_N1, l3_N2, l3_W1, and l3_W2 represent the two small-angle detection channels l3_N1 and l3_N2 and the two large-angle detection channels l3_W1 and l3_W2 under the third illumination beam, respectively. The defect signal intensity detected by each channel is DN3, which represents the intermediate defect size after fusing the defect signal intensity of the two small-angle detection channels l3_N1 and l3_N2. DW3 represents the intermediate defect size after fusing the defect signal intensity of the two large-angle detection channels l3_W1 and l3_W2. DW23 represents the intermediate defect size after fusing DW2 and DW3. Dn represents the intermediate defect size after fusing DW23 and DN3. D represents the final defect size obtained after fusing Do and Dn. When fusing Do and Dn, the larger value of the two is taken.

[0095] For further details, please refer to Figure 12 , Figure 12A flowchart illustrating another embodiment of obtaining defect size information through two-channel fusion is provided. In this flowchart, illumination 1, illumination 2, and illumination 3 represent the first, second, and third illumination beams, respectively; the l1_W1 channel and l1_N1 channel represent the defect signal intensities detected by the large-angle detection channel l1_W1 and the small-angle detection channel l1_N1 under the first illumination beam, respectively; DN1 represents the intermediate defect size obtained by querying the calibration curve based on the defect signal intensity of the l1_N1 channel; DW1 represents the intermediate defect size obtained by querying the calibration curve based on the defect signal intensity of the l1_W1 channel; the L2_W1 channel and l2_N1 channel represent the defect signal intensities detected by the large-angle detection channel l2_W1 and the small-angle detection channel l2_N1 under the second illumination beam, respectively; DN2 represents the intermediate defect size obtained by querying the calibration curve based on the defect signal intensity of the l2_N1 channel; DW2 represents the intermediate defect size obtained by querying the calibration curve based on the defect signal intensity of the l2_W1 channel. Defect size: DW12 represents the intermediate defect size after fusing DW1 and DW2; Do represents the intermediate defect size after fusing DN1 and DW12; L3_W1 channel and l3_N1 channel represent the defect signal intensities detected by the large-angle detection channel l3_W1 and the small-angle detection channel l3_N1 under the third illumination beam, respectively; DN3 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of l3_N1 channel; DW3 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of l3_W1 channel; DN23 represents the intermediate defect size after fusing DN2 and DN3; Dn represents the intermediate defect size after fusing DN23 and DW3; D represents the final defect size obtained by fusing Do and Dn. When fusing Do and Dn, the larger of the two values ​​is taken.

[0096] For further details, please refer to Figure 13 , Figure 13A flowchart illustrating another embodiment of obtaining defect size information through two-channel fusion is presented. In this flowchart, Illumination 1, Illumination 2, and Illumination 3 represent the first, second, and third illumination beams, respectively. Channels l1_W1 and l1_N1 represent the defect signal intensities detected by the large-angle detection channel l1_W1 and the small-angle detection channel l1_N1 under the first illumination beam, respectively. DN1 represents the intermediate defect size obtained by querying the calibration curve using the defect signal intensity of channel l1_N1, DW1 represents the intermediate defect size obtained by querying the calibration curve using the defect signal intensity of channel l1_W1, and channel L2_W1 represents the defect signal intensity detected by the large-angle detection channel l2_W1 under the second illumination beam. DW2 represents the intermediate defect size obtained from the calibration curve of the defect signal intensity query of the l2_W1 channel; DW12 represents the intermediate defect size after the fusion of DW1 and DW2; Do represents the intermediate defect size after the fusion of DN1 and DW12; L3_W1 channel and l3_N1 channel represent the defect signal intensities detected by the large-angle detection channel l3_W1 and the small-angle detection channel l3_N1 under the third illumination beam, respectively; DN3 represents the intermediate defect size obtained from the calibration curve of the defect signal intensity query of the l3_N1 channel; DW3 represents the intermediate defect size obtained from the calibration curve of the defect signal intensity query of the l3_W1 channel; Dn represents the intermediate defect size after the fusion of DN3 and DW3; D represents the final defect size obtained after the fusion of Do and Dn, where the larger value of Do and Dn is taken during fusion.

[0097] For further details, please refer to Figure 14 , Figure 14A flowchart illustrating another embodiment of obtaining defect size information through three-channel fusion is presented. Here, illumination 1, illumination 2, and illumination 3 represent the first, second, and third illumination beams, respectively; channels l1_N1, l1_W1, and l1_W2 represent the defect signal intensities detected by the small-angle detection channel l1_N1 and the two large-angle detection channels l1_W1 and l1_W2 under the first illumination beam, respectively; DN1 represents the intermediate defect size obtained by querying the calibration curve based on the defect signal intensity of channel l1_N1; DW1 represents the intermediate defect size after fusing the defect signal intensities of the two large-angle detection channels l1_W1 and l1_W2; channels l2_N1, l2_W1, and l2_W2 represent the defect signal intensities detected by the small-angle detection channel l2_N1 and the two large-angle detection channels l2_W1 and l2_W2 under the second illumination beam, respectively; DN2 represents the intermediate defect size obtained by querying the calibration curve based on the defect signal intensity of channel l2_N1; and DW2 represents the intermediate defect size obtained by fusing the defect signal intensities of the two large-angle detection channels l1_W1 and l1_W2. The intermediate defect size is obtained by fusing the defect signal intensities of channels l2_W1 and l2_W2. DW12 represents the intermediate defect size obtained by fusing DW1 and DW2. Do represents the intermediate defect size obtained by fusing DN1 and DW12. Channels l3_N1, l3_W1, and l3_W2 represent the defect signal intensities detected by the small-angle detection channel l3_N1 and the two large-angle detection channels l3_W1 and l3_W2 under the third illumination beam, respectively. DN3 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of channel l3_N1. DW3 represents the intermediate defect size obtained by fusing the defect signal intensities of the two large-angle detection channels l3_W1 and l3_W2. DN23 represents the intermediate defect size obtained by fusing DN2 and DN3. Dn represents the intermediate defect size obtained by fusing DN23 and DW3. D represents the final defect size obtained by fusing Do and Dn. When fusing Do and Dn, the larger value of the two is taken.

[0098] Furthermore, in some embodiments, for the three types of defects mentioned above—low aspect ratio defects, high aspect ratio defects, and damage defects—customized lighting and detection systems can be selected according to requirements, detecting only two types of defects. Therefore, the design of the lighting beam and detection channel also differs; please refer to [link / reference] for details. Figures 15-20 As shown, Figure 15 , Figure 16 , Figure 17 The fusion process is demonstrated when using a first illumination beam and a second illumination beam. Figure 18 , Figure 19 , Figure 20 The fusion process using a second and a third illumination light was demonstrated.

[0099] like Figure 15 As shown, illumination 1 and illumination 2 represent the first illumination beam and the second illumination beam, respectively. Channels l1_N1, l1_N2, l1_W1, and l1_W2 represent the defect signal intensities detected by the two small-angle detection channels l1_N1 and l1_N2 and the two large-angle detection channels l1_W1 and l1_W2 under the first illumination beam, respectively. DN1 represents the intermediate defect size after fusing the defect signal intensities of the two small-angle detection channels l1_N1 and l1_N2. DW1 represents the intermediate defect size after fusing the defect signal intensities of the two large-angle detection channels l1_W1 and l1_W2. Channels l2_N1, l2_N2, l2_W1, and l2_W2... The channels represent the defect signal intensities detected by the two small-angle detection channels l2_N1 and l2_N2 and the two large-angle detection channels l2_W1 and l2_W2 under the second illumination beam, respectively. The defect signal intensities of the l2_W2 and l2_N2 channels are not used. DW2 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of the l2_W1 channel. DN2 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of the l2_N1 channel. DW represents the intermediate defect size obtained by fusing DW1 and DW2. DN represents the intermediate defect size obtained by fusing DN1 and DN2. D represents the final defect size obtained by fusing DW and DN. When fusing DW and DN, the larger value of the two is taken.

[0100] like Figure 16 As shown, Illumination 1 and Illumination 2 represent the first and second illumination beams, respectively. Channels l1_N1 and l1_W2 represent the defect signal intensities detected by the small-angle detection channel l1_N1 and the large-angle detection channel l1_W2 under the first illumination beam, respectively. DN1 represents the intermediate defect size obtained from the calibration curve using the defect signal intensity of channel l1_N1, and DW1 represents the intermediate defect size obtained from the calibration curve using the defect signal intensity of channel l1_W2. Channels l2_N1 and l2_W1 represent the defect signal intensities detected by the second illumination beam under the first illumination beam, respectively. The defect signal intensities detected by the small-angle detection channel l2_N1 and the large-angle detection channel l2_W1 are respectively. DW2 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of the l2_W1 channel. DN2 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of the l2_N1 channel. DW represents the intermediate defect size obtained by fusing DW1 and DW2. DN represents the intermediate defect size obtained by fusing DN1 and DN2. D represents the final defect size obtained by fusing DW and DN. When fusing DW and DN, the larger value of the two is taken.

[0101] like Figure 17 As shown, illumination 1 and illumination 2 represent the first illumination beam and the second illumination beam, respectively. The l1_N1 channel, l1_N2 channel, and l1_W1 channel represent the defect signal intensities detected by the two small-angle detection channels l1_N1 and l1_N2 and the large-angle detection channel l1_W1 under the first illumination beam, respectively. DN represents the intermediate defect size after fusing the defect signal intensities of the two small-angle detection channels l1_N1 and l1_N2. The l2_W1 channel represents the defect signal intensity detected by the large-angle detection channel l2_W1 under the second illumination beam. DW represents the intermediate defect size after fusing the defect signal intensities of the two large-angle detection channels l1_W1 and l2_W1. D represents the final defect size obtained by fusing DW and DN. When fusing DW and DN, the larger value of the two is taken.

[0102] like Figure 18 As shown, illumination 2 and illumination 3 represent the second and third illumination beams, respectively. Channels l2_N1, l2_N2, l2_W1, and l2_W2 represent the defect signal intensities detected by the two small-angle detection channels l2_N1 and l2_N2 and the two large-angle detection channels l2_W1 and l2_W2, respectively, under the second illumination beam. The defect signal intensities of channels l2_W2 and l2_N2 are not used. DW2 represents the intermediate defect size obtained from the calibration curve of the defect signal intensity of channel l2_W1, and DN2 represents the intermediate defect size obtained from the calibration curve of the defect signal intensity of channel l2_N1. Channels l3_N1, l3_N2, and l3_W1... The l3_W2 channel represents the defect signal intensity detected by the two small-angle detection channels l3_N1 and l3_N2 and the two large-angle detection channels l3_W1 and l3_W2 under the third illumination beam, respectively. DN3 represents the intermediate defect size after fusing the defect signal intensity of the two small-angle detection channels l3_N1 and l3_N2. DW3 represents the intermediate defect size after fusing the defect signal intensity of the two large-angle detection channels l3_W1 and l3_W2. DN represents the intermediate defect size after fusing DN2 and DN3. DW represents the intermediate defect size after fusing DW2 and DW3. D represents the final defect size obtained after fusing DN and DW. When fusing DN and DW, the larger value of the two is taken.

[0103] like Figure 19As shown, Illumination 2 and Illumination 3 represent the second and third illumination beams, respectively. The l2_N and l2_W channels represent the defect signal intensities detected by the small-angle detection channel l2_N and the large-angle detection channel l2_W, respectively, following the second illumination beam. DW2 represents the intermediate defect size obtained from the calibration curve using the defect signal intensity of the l2_W channel. DN2 represents the intermediate defect size obtained from the calibration curve using the defect signal intensity of the l2_N channel. D2 represents the intermediate defect size after fusing DW2 and DN2. The l3_N1 channel... The l3_W1 channel represents the defect signal intensity detected by the small-angle detection channel l3_N1 and the large-angle detection channel l3_W1 under the third illumination beam, respectively. DN3 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of the l3_N1 channel. DW3 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of the l3_W1 channel. D3 represents the intermediate defect size after fusing DW3 and DN3. D represents the final defect size obtained after fusing D2 and D3. When fusing D2 and D3, the larger value of the two is taken.

[0104] like Figure 20 As shown, illumination 2 and illumination 3 represent the second and third illumination beams, respectively. Channels l2_N1, l2_W1, and l2_W2 represent the defect signal intensities detected by the small-angle detection channel l2_N1 and the two large-angle detection channels l2_W1 and l2_W2, respectively, following the second illumination beam. The defect signal intensity of channel l2_W2 is not used. DW2 represents the intermediate defect size obtained from the calibration curve using the defect signal intensity of channel l2_W1, and DN2 represents the intermediate defect size obtained from the calibration curve using the defect signal intensity of channel l2_N1. Channels l3_N1, l3_W1, and l3_W2... The channels represent the defect signal intensities detected by the small-angle detection channel l3_N1 and the two large-angle detection channels l3_W1 and l3_W2 under the third illumination beam, respectively. DN3 represents the intermediate defect size obtained by querying the calibration curve of the defect signal intensity of the l3_N1 channel. DW3 represents the intermediate defect size after fusing the defect signal intensities of the two large-angle detection channels l3_W1 and l3_W2. DN represents the intermediate defect size after fusing DN2 and DN3. DW represents the intermediate defect size after fusing DW2 and DW3. D represents the final defect size obtained after fusing DN and DW. When fusing DN and DW, the larger value of the two is taken.

[0105] Furthermore, according to the predefined fusion logic, the specific steps of fusing multiple intermediate defect sizes from different illumination beams and different detection channels in pairs include: 3.1 Determine whether the dimensions of the two intermediate defects to be fused are being fused for the first time.

[0106] It should be noted that the first fusion specifically refers to fusing the defect signal intensities of two similar detection channels (e.g., two large-angle detection channels under the first illumination beam) under the same illumination condition and mapping them to a single intermediate defect size. This is the initial stage of the fusion process. Subsequent fusions refer to the further fusion of already generated intermediate defect sizes. The two sizes involved in the fusion may originate from: the initial fusion results of different types of channels under the same illumination (e.g., a large-angle detection channel and a small-angle detection channel under the first illumination beam), or intermediate defect sizes obtained under different illumination conditions. These represent the intermediate and final stages of the fusion process. By distinguishing these stages, the system can apply the most suitable fusion algorithm. The initial fusion focuses on optimizing the original signal, while subsequent fusions focus on confidence-based decisions.

[0107] 3.2 If so, the value is taken from the two intermediate defect sizes based on the collection efficiency of the detection channel, or the average value of the two intermediate defect sizes is taken as the fusion value.

[0108] Furthermore, the steps for determining the collection efficiency from the two intermediate defect sizes based on the detection channel specifically include: 3.21. Compare the larger of the two intermediate defect sizes with the second preset threshold.

[0109] Specifically, the intermediate defect size calculated from the two detection channels is taken, and the larger value is compared with a second preset threshold. This second preset threshold is pre-set to mark the critical point at which the efficient detection channel may begin to saturate.

[0110] 3.22. When the larger of the two intermediate defect sizes does not exceed the second preset threshold, the intermediate defect size corresponding to the detection channel with higher collection efficiency is taken as the fusion value.

[0111] Specifically, if the larger value does not exceed the second preset threshold, it indicates that the defect size is small and both detection channels are in the linear operating region. In this case, the detection channel with higher collection efficiency has better data because it is more sensitive to weak signals and has a higher signal-to-noise ratio. Therefore, its corresponding intermediate defect size is adopted as the fusion value.

[0112] 3.23 When the larger of the two intermediate defect sizes exceeds the second preset threshold, the intermediate defect size corresponding to the detection channel with lower collection efficiency is taken as the fusion value.

[0113] Specifically, if a large value exceeds the second preset threshold, it is a strong signal indicating that the defect size is large, the efficient detection channel may be saturated or near saturation, and its data is beginning to distort. Meanwhile, the detection channel with lower collection efficiency may still be in the linear region, providing more reliable measurements. Therefore, the system intelligently switches to adopting the intermediate defect size corresponding to the detection channel with lower collection efficiency as the fused value.

[0114] In other embodiments, the average value of the two defect signal intensities can be directly taken, and then the average value can be used to query the calibration curve to obtain the fused intermediate defect size.

[0115] 3.3 If not, then the fusion is based on the source type of the two intermediate defect sizes.

[0116] Furthermore, the steps for fusing based on the source types of the two intermediate defect sizes specifically include: 3.31 When the dimensions of two intermediate defects are obtained from the defect signal intensity in the unsaturated state, the larger of the two values ​​shall be taken as the fusion value.

[0117] 3.32 When two intermediate defect sizes are obtained from the defect signal intensity in the unsaturated state and the defect signal intensity in the saturated state, respectively, the intermediate defect size obtained from the defect signal intensity in the saturated state is taken as the fusion value.

[0118] 3.33 When the dimensions of two intermediate defects are obtained from the defect signal intensity in a saturated state, the average of the two values ​​is taken as the fusion value.

[0119] 3.34. When at least one of the two intermediate defect sizes is calculated from the total pixel area of ​​the defect signal, the larger of the two values ​​is taken as the fusion value.

[0120] 4. Output the location of the defect and the final defect size.

[0121] The wafer defect detection method of this invention achieves high sensitivity coverage and accurate identification of all types of defects by using multi-angle composite illumination, combined with large and small angle detection channels, and key contralateral arrangement. After acquiring image data collected by multiple illumination methods and multiple detection channels, the method intelligently fuses defect signals under different illumination methods and different detection channels based on the state of defect signal intensity, thereby obtaining high-precision and high-robust detection results, which greatly improves the final accuracy of wafer defect detection.

[0122] The above are merely embodiments of this application and do not limit the scope of this patent application. Any equivalent structural or procedural changes made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of this application.

Claims

1. A wafer defect detection device, characterized in that, It includes: An illumination system is configured to provide at least three illumination beams with different incident directions to the wafer surface and form overlapping linear illumination spots on the wafer surface; A motion platform for carrying and driving the wafer to move relative to the linear illumination spot; The detection system includes multiple detection channels, each configured to collect scattered light generated by illumination in the linear illumination spot area and convert it into an electrical signal; A signal processing system configured to process the electrical signals output by the detection system to determine defect information; The optical axes of the multiple detection channels are all located on the same detection plane, which is perpendicular to the wafer surface. The long axis of the linear illumination spot has a first angle with the detection plane, which is in the range of 85 to 95 degrees. The multiple detection channels include at least one large-angle detection channel and at least one small-angle detection channel. The angle between the optical axis of the large-angle detection channel and the wafer normal is greater than a second preset angle, and the angle between the optical axis of the small-angle detection channel and the wafer normal is less than the second preset angle. The projection of one of the at least three illumination beams onto the detection plane and at least one detection channel are located on opposite sides of the wafer normal.

2. The wafer defect detection device according to claim 1, characterized in that, The at least three illumination beams include a first illumination beam, a second illumination beam, and a third illumination beam. The second illumination beam is located on the detection plane. The first illumination beam is perpendicular to the plane formed by the wafer normal, the detection plane, and the plane where the wafer is located. The second illumination beam and at least one detection channel are located on opposite sides of the wafer normal.

3. The wafer defect detection device according to claim 2, characterized in that, The angle between the first illumination beam and the wafer normal is between 60° and 80°, the angle between the second illumination beam and the wafer normal is between 30° and 60°, and the angle between the third illumination beam and the wafer normal is between 0° and 5°.

4. The wafer defect detection device according to claim 2, characterized in that, The first illumination beam, the second illumination beam, and the third illumination beam have the same wavelength and are illuminated sequentially during the wafer scanning process to complete the scan; Alternatively, the first illumination beam, the second illumination beam, and the third illumination beam all have different wavelengths and are illuminated simultaneously during the wafer scanning process to complete the scan, and the detection system is equipped with elements that distinguish between different wavelengths; Alternatively, two of the first, second, and third illumination beams have the same wavelength, while the third has a different wavelength. During the wafer scanning process, the illumination beam with the different wavelength is simultaneously illuminated by one of the two illumination beams with the same wavelength to complete the scan. The detection system is equipped with elements that distinguish between different wavelengths.

5. The wafer defect detection device according to claim 2, characterized in that, The plurality of detection channels includes a large-angle detection channel and a small-angle detection channel, wherein the large-angle detection channel and / or the small-angle detection channel are located on opposite sides of the wafer normal, respectively, from the second illumination beam.

6. The wafer defect detection device according to claim 1, characterized in that, The plurality of detection channels include two large-angle detection channels and one small-angle detection channel. The two large-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two large-angle detection channels and the wafer normal is within a preset angle difference range. Alternatively, the plurality of detection channels may include a large-angle detection channel and two small-angle detection channels, wherein the two small-angle detection channels are located on opposite sides of the wafer normal, and the difference between the angles between the two small-angle detection channels and the wafer normal is within a preset angle difference range.

7. The wafer defect detection device according to claim 1, characterized in that, The plurality of detection channels include two large-angle detection channels and two small-angle detection channels. The two large-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two large-angle detection channels and the wafer normal is within a preset angle difference range. The two small-angle detection channels are located on both sides of the wafer normal, and the difference between the angles between the two small-angle detection channels and the wafer normal is within a preset angle difference range.

8. The wafer defect detection device according to claim 1, characterized in that, The angle between the optical axis of the large-angle detection channel and the wafer normal is between 45° and 65°, and the angle between the optical axis of the small-angle detection channel and the wafer normal is between 15° and 25°.

9. The wafer defect detection device according to claim 1, characterized in that, The long axis of the linear illumination spot is perpendicular to the detection plane, and the width of the linear illumination spot is preset according to the depth of field of the detection channel.

10. A method for detecting wafer defects, characterized in that, It is applied to the wafer defect detection apparatus according to any one of claims 1-9; the method includes: The illumination system illuminates the wafer surface with illumination beams from at least three different incident directions, forming linear illumination spots; The motion platform moves the wafer, causing the linear illumination spot to scan the wafer surface; The detection system uses multiple detection channels to collect the scattered light generated in the linear illumination spot area during the scanning process and converts it into multi-channel electrical signals; The signal processing system processes the multi-channel electrical signals and fuses information from different illumination beams and different detection channels to identify defects and their size information.

11. The wafer defect detection method according to claim 10, characterized in that, The signal processing system processes the multi-channel electrical signals, fusing information from different illumination beams and different detection channels to identify defects and their size information, including: The location of the defect signal intensity under any illumination beam and any detection channel, the corresponding defect signal intensity, and whether the defect signal intensity meets a predefined saturation state are obtained. Based on the state of the defect signal strength, the defect signal strength is mapped to the intermediate defect size; Using predefined fusion logic, multiple intermediate defect sizes from different illumination beams and different detection channels are fused in pairs until the final defect size is obtained; Output the location of the defect and the final defect size.

12. The wafer defect detection method according to claim 11, characterized in that, Mapping the defect signal intensity to the intermediate defect size based on the saturation state includes: When the defect signal strength is less than or equal to 0, the intermediate defect size is 0; When the defect signal strength is greater than 0 and less than or equal to a first preset threshold, the intermediate defect size is obtained by querying the pre-calibrated calibration curve. The calibration curve refers to the mapping curve between the defect signal strength and the defect size. When the intensity of the defect signal is greater than the first preset threshold, the pixels in the location area corresponding to the defect signal are clustered and merged according to the intensity distribution to obtain the total area of ​​pixels occupied by the defect signal. Then, the size of the intermediate defect is calculated by combining the total area of ​​pixels and the magnification of the corresponding detection channel.

13. The wafer defect detection method according to claim 12, characterized in that, The method utilizes predefined fusion logic to fuse multiple intermediate defect sizes from different illumination beams and different detection channels pairwise, including: Determine whether the dimensions of the two intermediate defects to be fused are being fused for the first time; If so, the value is taken from the two intermediate defect sizes based on the collection efficiency of the detection channel, or the average of the two intermediate defect sizes is taken as the fusion value; If not, then fusion is performed based on the source type of the two intermediate defect sizes.

14. The wafer defect detection method according to claim 13, characterized in that, The collection efficiency based on the detection channel is taken from two intermediate defect sizes, including: The larger of the two intermediate defect sizes is compared with a second preset threshold. When the larger of the two intermediate defect sizes does not exceed the second preset threshold, the intermediate defect size corresponding to the detection channel with higher collection efficiency is taken as the fusion value. When the larger of the two intermediate defect sizes exceeds the second preset threshold, the intermediate defect size corresponding to the detection channel with lower collection efficiency is taken as the fusion value.

15. The wafer defect detection method according to claim 13, characterized in that, The fusion based on the source types of two intermediate defect sizes includes: When both intermediate defect sizes are obtained through the defect signal intensity in the unsaturated state, the larger of the two values ​​is taken as the fusion value. When two intermediate defect sizes are obtained from the defect signal intensity in the unsaturated state and the defect signal intensity in the saturated state, respectively, the intermediate defect size obtained from the defect signal intensity in the saturated state is taken as the fusion value. When the dimensions of two intermediate defects are obtained from the defect signal intensity in the saturated state, the average of the two is taken as the fusion value. When at least one of the two intermediate defect sizes is calculated from the total pixel area of ​​the defect signal, the larger of the two values ​​is taken as the fusion value.

Images