Defect inspection apparatus and defect inspection method

Abstract

The present application provides a kind of defect inspection device and defect inspection method, which can check a variety of defect types in composite image.The defect inspection device synthesizes first detection signal from first detector and second detection signal from second detector with first synthesis ratio to generate first composite image, and synthesizes the first detection signal and the second detection signal with second synthesis ratio different from the first synthesis ratio to generate second composite image.Then, first inspection image is generated based on the first composite image, and second inspection image is generated based on the second composite image.And, logical operation of the first inspection image and the second inspection image is performed to generate composite inspection image.Defect determination is performed on the composite inspection image.

Description

Technical Field

[0001] This invention relates to a defect inspection apparatus and a defect inspection method using the same apparatus. Background Technology

[0002] As a method for detecting defects in the circuit pattern of a semiconductor, there is a known method that compares the shape of a normal circuit pattern without defects with the shape of the circuit pattern of the object to be inspected. Typically, defect detection using a scanning electron microscope is performed using a secondary electron image (SE image). However, as described in Patent Document 1, in cases where more accurate detection of scum is desired, a method for defining the inspection area using a composite image obtained by combining a secondary electron image (SE image) and a reflected electron image (BSE) is also known.

[0003] Defects in semiconductor circuit patterns are diverse, and depending on the type, some defects can only be visually identified in SE (Self-Exposed) images, while others can only be visually identified in BSE (Browser-Exposed) images. For example, defects such as potential contrast caused by the electrical characteristics of semiconductor devices can only be visually identified in SE images, while defects such as unevenness or deformation of the circuit pattern can only be visually identified in BSE images. In such cases, methods are known for inspection using an image obtained by combining SE and BSE images (composite image).

[0004] However, while defect inspection using such composite images can cover a wide variety of defects, there is a trade-off: the defect detection sensitivity decreases compared to the original image. Specifically, defects that can be visually identified solely by SE images may become undetectable in composite images that combine SE and BSE images due to reduced sensitivity. Similarly, defects that can be visually identified solely by BSE images may be undetectable in the composite image. In recent years, due to the miniaturization of semiconductor patterns, the size of defects to be detected has gradually decreased. Consequently, in the latest generation of semiconductor devices, there are increasing instances where composite images cannot be used to inspect all types of defects.

[0005] Patent Document 1: Japanese Patent Application Publication No. 2014-77798

[0006] Patent Document 2: Japanese Patent No. 6281019 Summary of the Invention

[0007] This invention provides a defect inspection device and method capable of inspecting a wide variety of defect types in synthetic images.

[0008] In view of the aforementioned problems, the defect inspection apparatus of the present invention comprises: an image synthesis unit that synthesizes a first detection signal from a first detector and a second detection signal from a second detector at a first synthesis ratio to generate a first composite image, and synthesizes the first detection signal and the second detection signal at a second synthesis ratio different from the first synthesis ratio to generate a second composite image; a defect inspection processing unit that generates a first inspection image based on the first composite image, and generates a second inspection image based on the second composite image; an inspection image calculation unit that performs logical operations on the first inspection image and the second inspection image to generate a composite inspection image; and a defect determination processing unit that performs defect determination on the composite inspection image.

[0009] According to the defect inspection apparatus and method of the present invention, a defect inspection apparatus and method capable of inspecting a wide variety of defect types in a composite image can be provided. Attached Figure Description

[0010] Figure 1 An example of the structure of an image generation tool, namely a scanning electron microscope, is shown in the defect inspection apparatus of the first embodiment, which acquires an image of the object to be inspected.

[0011] Figure 2 It indicates that the passage Figure 1 A block diagram of an example of a defect inspection system (defect inspection device) that performs image synthesis and defect detection processing on image data obtained from a scanning electron microscope (image generation tool) as illustrated, and outputs the result of defect determination.

[0012] Figure 3 This is an explanatory diagram illustrating the defect determination operation in the computer system 201 of the first embodiment.

[0013] Figure 4 This is an explanatory diagram illustrating the defect determination operation in the computer system 201 of the first embodiment.

[0014] Figure 5 This is an example of a GUI screen displayed on the display of the operation unit 206 in the defect inspection device of the first embodiment.

[0015] Figure 6 This is a flowchart explaining the defect inspection sequence using the defect inspection device of the first embodiment.

[0016] Figure 7 This is a block diagram illustrating an example of a defect inspection system (defect inspection device) according to the second embodiment.

[0017] Figure 8This is an explanatory diagram illustrating the defect determination operation in the computer system 901 of the second embodiment.

[0018] Figure 9 This is an explanatory diagram illustrating the defect determination operation in the computer system 901 of the second embodiment.

[0019] Figure 10 This is an example of a GUI screen displayed on the display of the operation unit 206 in the defect inspection device of the second embodiment.

[0020] Figure 11 This is a flowchart explaining the defect inspection sequence using the defect inspection device of the second embodiment.

[0021] Figure 12 This is a flowchart explaining the defect inspection sequence using the defect inspection device of the second embodiment. Detailed Implementation

[0022] Hereinafter, this embodiment will be described with reference to the accompanying drawings. In the drawings, functionally identical elements are sometimes shown with the same numbers. Furthermore, the drawings illustrate embodiments and installation examples that follow the principles of this disclosure, but these are for understanding this disclosure and are in no way intended to limit its interpretation. The descriptions in this specification are merely typical examples and do not limit the scope of patent protection or application of this disclosure in any sense.

[0023] In this embodiment, a thorough and detailed description has been provided by those skilled in the art for the purpose of implementing this disclosure. However, it should be understood that other installations and methods are also possible, and structural and construction modifications and substitutions of various elements can be made without departing from the scope and spirit of the technical concept of this disclosure. Therefore, the following description should not be interpreted as limited to this.

[0024] [First Implementation Method]

[0025] Reference Figures 1-6 The defect inspection apparatus and method of the first embodiment will be described. Figure 1 An example of the structure of an image generation tool that obtains an image of the object to be inspected in the defect inspection apparatus of the first embodiment is a scanning electron microscope. Furthermore, the image generation tool applicable to this invention is not limited to a scanning electron microscope; for example, a focused ion beam apparatus that generates images based on scanning of an ion beam can also be used as an image generation tool.

[0026] Figure 1 The illustrated scanning electron microscope includes: an imaging unit 101, a computer system 102, a signal processing unit 103, an input / output unit 104, and a storage unit 105.

[0027] The imaging unit 101 includes: an electron gun 106 that irradiates an electron beam 107, a focusing lens 108 that focuses the electron beam 107, and a focusing lens 109 that further focuses the electron beam 107 after it has passed through the focusing lens 108. The imaging unit 101 also includes a deflector 110 that deflects the electron beam 107 and an objective lens 111 that controls the focusing height of the electron beam 107.

[0028] An electron beam 107 passing through the optical system of the imaging unit 101 irradiates the sample 112 placed on the sample stage 113. Secondary electrons (SE) 114 emitted from the sample 112 due to irradiation by the electron beam 107 are mainly detected by a secondary electron detector (upper detector) 115. In addition, backscattered electrons (BSE) 116 generated from the sample 112 are mainly detected by a backscattered electron detector (lower detector) 117.

[0029] Computer system 102 controls imaging unit 101. Signal processing unit 103 generates SEM images (SE image, BSE image) based on the outputs of upper detector 115 and lower detector 117. Synchronously with the scanning of a scanning deflector (not shown), signal processing unit 103 stores detection signals in a frame memory or similar device to generate image data. When storing detection signals in the frame memory, a signal profile (one-dimensional information) and an SEM image (two-dimensional information) are generated by storing the detection signals at positions corresponding to the scanning positions in the frame memory. Storage unit 105 also functions as a non-temporary recording medium for storing computer programs that manage the operations of this system. Input / output unit 104 inputs various instructions from the operator and outputs inspection results, etc., to a display (not shown).

[0030] Figure 2 It is through Figure 1 This is an example of a defect inspection system (defect inspection device) that performs image synthesis and defect detection processing on image data obtained from a scanning electron microscope (image generation tool) and outputs the result of defect determination. The defect inspection system consists of a computer system 201 and an operation unit 206.

[0031] Computer system 201 comprises one or more computer subsystems, each containing one or more CPUs. Each of the one or more computer subsystems can implement the processing described later using software via one or more processors, or can implement part or all of the processing via hardware such as electronic circuits.

[0032] As an example, the computer system 201 includes an image compositing unit 202, a defect inspection processing unit 203, an inspection image processing unit 204, and a defect determination processing unit 205. The image compositing unit 202, the defect inspection processing unit 203, the inspection image processing unit 204, and the defect determination processing unit 205 can be implemented virtually in software or in hardware such as electronic circuits.

[0033] The computer system 201 is configured to receive defect inspection parameters and defect judgment thresholds from the operation unit 206, and the operation unit 206 displays the inspection results processed by the computer system 201 on a display screen or the like.

[0034] In computer system 201, image compositing unit 202 composites SEM images 207 (SE images, BSE images) detected by multiple detectors (115, 117) at a pre-specified compositing rate. As described later, image compositing unit 202 is configured to composite SE images and BSE images at multiple compositing rates (two in the following example) to generate various composite images.

[0035] The defect inspection processing unit 203 performs defect inspection processing on the inspection image synthesized by the image compositing unit 202 and outputs the inspection image to the inspection image storage unit 208. At this time, the image compositing unit 202 outputs multiple inspection images at various compositing ratios. The multiple inspection images are processed by the inspection image processing unit 204, and a composite inspection image is output as the result of this processing. The composite inspection image undergoes defect determination by the defect determination processing unit 205, and the inspection result is displayed on the operation unit 206 and output to the inspection result file 209.

[0036] Reference Figure 3 and Figure 4 The operation of defect determination in the computer system 201 of the first embodiment will be described. As mentioned above, there are many types of defects. Some defects can be visually recognized in SE images but cannot be visually recognized or are difficult to visually recognize in BSE images. On the other hand, there are defects that can be visually recognized in BSE images but cannot be visually recognized or are difficult to visually recognize in SE images. In the defect inspection method using a composite image of SE and BSE images, any defect can be inspected from a composite image. However, due to the image synthesis, a trade-off problem may arise where defects that can be visually recognized in SE images alone are either not visually recognized or are difficult to visually recognize in the composite image. In the first embodiment, by performing... Figure 3 as well as Figure 4 The method described in the text eliminates this trade-off problem. Figure 3 This illustrates a case where defect A, which is visually identifiable only in the SE image, occurs in an area of ​​the inspected object. Figure 4This illustrates a case where defect B, which is visually identifiable only in BSE images, occurs in an area of ​​the inspected object.

[0037] like Figure 3 and Figure 4 As shown, in this first embodiment, the image compositing unit 202 combines SE and BSE images captured in the area targeted for defect inspection at two compositing ratios a and b to generate two composite images MP1 and MP2. Referring to images RP1 and RP2, SE and BSE images captured in a normal area (circuit pattern, etc.) where no defects have occurred are also generated by combining them at the same compositing ratios a and b. As an example, images RP1 and RP2 can be generated (composite) based on SE and BSE images in an area close to the area where SE and BSE images were captured for defect inspection, but which are not defective.

[0038] Then, the defect inspection processing unit 203 calculates the difference between the synthesized image MP1 and the reference image RP1 using inspection parameter 1 optimized for the synthesis ratio a, and generates an inspection image DMP1. Similarly, the defect inspection processing unit 203 calculates the difference between the synthesized image MP2 and the reference image RP2 using inspection parameter 2 optimized for the synthesis ratio b, and generates an inspection image DMP2. That is, inspection image DMP1 is a difference image generated based on the synthesized image MP1 at synthesis ratio a and the reference image RP1, and inspection image DMP2 is a difference image generated based on the synthesized image MP2 at synthesis ratio b and the reference image RP2. Inspection parameters 1 and 2 define the degree of sensitivity at which defects are detected. If inspection parameters 1 and 2 are set to increase sensitivity, smaller defects can be detected. However, in this case, random noise components in the image are also detected simultaneously, increasing the risk of misdetecting normal circuit patterns as defect candidate areas. Therefore, it is preferable to optimize inspection parameters 1 and 2 to detect defects to the extent necessary to ensure the performance of the circuit pattern, and to minimize noise components as much as possible.

[0039] The inspection image processing unit 204 applies a prescribed logical operation (in this case, logical OR) to the two inspection images DMP1 and DMP2 to generate a composite inspection image SDMP. Then, the defect determination processing unit 205 analyzes the composite inspection image SDMP according to a predetermined threshold to determine defects. In this example, the logical operation performed by the inspection image processing unit 204 is logical OR, but the type of logical operation is not limited to logical OR and can be changed according to the selected synthesis ratio a, b, inspection parameters 1, 2, and other factors.

[0040] exist Figure 3In the process, the two composite images MP1 and MP2 generated by the image compositing unit 202 are respectively calculated and differ from the reference images RP1 and RP2 in the defect inspection processing unit 203. As the result of this difference operation, the defect inspection processing unit 203 outputs inspection images DMP1 and DMP2. Figure 3 As shown, in the inspection images DMP1 and DMP2, if a predetermined defect exists, the part corresponding to the defect can be displayed, for example, in a form where it is painted white on a black background.

[0041] In the case where a defect A exists in the region corresponding to the SE image and the BSE image, which can only be visually recognized in the SE image, for example, we obtain... Figure 3 The inspection images DMP1 and DMP2 are shown as shown.

[0042] For example, in Figure 3 In the case where defect A exists, defect A cannot be detected only in the inspection image DMP2, which is based on the synthetic image synthesized at a synthesis ratio b. Furthermore, in Figure 4 In the case where defect B exists, defect B cannot be detected only in the inspection image DMP1, which is based on the composite image synthesized at a synthesis ratio of a. However, in this embodiment, defect detection is performed on the composite inspection image SDMP, which is obtained by performing a logical operation (e.g., an OR operation) on the inspection image DMP1, which is the composite image MP1 synthesized at a synthesis ratio of a, and the inspection image DMP2, which is the composite image MP2 synthesized at a synthesis ratio of b. Therefore, either defect A or defect B can be detected. Figure 3 , Figure 4 The examples illustrating defects A and B are provided, but the same discussion applies even when the types of defects increase. That is, according to this embodiment, the aforementioned trade-offs can be eliminated.

[0043] The synthesis ratio 'a' of the composite image MP1 is preferably set to a value where the SE image is larger than the BSE image, and conversely, the synthesis ratio of the composite image MP2 is preferably set to a value where the BSE image is larger than the SE image. Therefore, compared to using a single composite image, the number of detectable defects can be increased.

[0044] Even when using a single composite image to detect defects, it is considered that by setting the composite ratio to around 50%, both defects A and B can be detected. However, the types of defects that can occur on semiconductor devices typically range from hundreds, making it extremely difficult to set an appropriate composite ratio for all of these defects. With this embodiment, even in such cases where a defect candidate region appears only in either inspection image DMP1 or DMP2, the defect candidate region can always be ensured in the composite inspection image SDMP, enabling the simultaneous detection of both defects A and B.

[0045] Figure 5 This is an example of a GUI screen displayed on the monitor of the operation unit 206 in the defect inspection device of the first embodiment. On this GUI screen, settings can be input (configured). Figure 3 and Figure 4 The defect inspection method shown includes the synthesis ratios a and b, inspection parameters 1 and 2, and the types of logical operations. Additionally, it can display the synthesized image, reference image, inspection images DMP1 and DMP2, and the synthesized inspection image SDMP, given the input information.

[0046] The GUI screen includes a parameter setting unit 503, an image display unit 504, a logic operation / threshold setting unit 505, and an inspection result image display unit 506. The parameter setting unit 503 is configured to accept, for example, the aforementioned synthesis ratios a and b, and inspection parameters 1 and 2 into a numerical input box. When these values ​​and parameters are input to the parameter setting unit 503, the image display unit 504 displays the synthesized images MP1 and MP2, reference images RP1 and RP2, and inspection images DMP1 and DMP2 according to those values. Furthermore, the logic operation / threshold setting unit 505 can set the logic operation used when synthesizing the inspection images DMP1 and DMP2, and can also set the threshold for defect determination. The inspection result image display unit 506 displays an inspection result image showing the result of defect inspection performed on the synthesized inspection image SDMP obtained in this way.

[0047] Next, refer to Figure 6 The flowchart below describes the defect inspection process using the defect inspection apparatus of the first embodiment. First, in step S801, the electron beam 107 is moved to the inspection target position (step S801), and an SE image and a BSE image are obtained at the inspection target position (step S802). Then, the obtained SE image and BSE image are combined at a synthesis ratio a to generate a composite image MP1 (step S803), and the obtained SE image and BSE image are combined at a synthesis ratio b to generate a composite image MP2 (step S804).

[0048] Next, the electron beam 107 is moved to a reference position close to the location of the object being inspected (step S805), and an SE image and a BSE image are acquired at this reference position (step S806). Then, the obtained SE image and BSE image are combined at a synthesis ratio a to generate a reference image RP1 (step S807), and the obtained SE image and BSE image are combined at a synthesis ratio b to generate a reference image RP2 (step S808).

[0049] Next, in step S809, a subtraction operation is performed between the composite image MP1 and the reference image RP1, thereby generating an inspection image DMP1 as the difference between the two. Additionally, in step S810, a subtraction operation is performed between the composite image MP2 and the reference image RP2, thereby generating an inspection image DMP2 as the difference between the two. Then, in step S811, a logical operation (logical OR operation) is performed between the inspection images DMP1 and DMP2 to generate a composite inspection image SDMP. Based on this composite inspection image SDMP, a defect determination at the inspection target location is performed (step S812), and the result of the defect determination is output to the operation unit 206, etc. (step S813). The above operations continue until all inspection target locations are completed.

[0050] As explained above, according to the defect inspection apparatus and method of the first embodiment, multiple inspection images are generated at various synthesis ratios, and defect inspection is performed based on a composite inspection image formed by synthesizing the multiple inspection images. Therefore, according to this first embodiment, a wide variety of defects can be inspected using a single composite inspection image.

[0051] [Second Implementation]

[0052] Next, refer to Figures 7-12 The defect inspection apparatus and defect inspection method of the second embodiment will be described. Figure 7 The block diagram illustrates an example of a defect inspection system (defect inspection device) according to the second embodiment. Additionally, Figure 8 as well as Figure 9 This diagram illustrates the defect determination process in the computer system 901 that constitutes the defect inspection system of the second embodiment. Figure 3 , Figure 4 Corresponding. Furthermore, in Figure 7 In China, regarding and Figure 2 Same constituent elements, label and Figure 2 The same reference symbols are used, therefore repeated descriptions are omitted below.

[0053] In addition to the components of the first embodiment, the defect inspection system of this second embodiment also includes an AI model generation unit 902 and an inference image generation unit 903. Instead of generating the aforementioned reference images, the defect inspection system of the second embodiment generates a learned AI (artificial intelligence) model in the AI ​​model generation unit 902 for generating inference images, and generates inference images according to this learned AI model. Then, based on the inference images and the synthesized images MP1 and MP2, first inspection images DMP1 and DMP2 are generated. The learned AI model includes previously captured SE images, BSE images, design data, defect determination result data, etc., as a dataset. The inference image generation unit 903 infers the shape of the inspection object (e.g., a normal circuit pattern) without defects according to the learned AI model and outputs it as an inference image.

[0054] Reference Figure 8 and Figure 9 This section explains the defect detection process in computer system 901. Figure 8 , Figure 9 They respectively showed the same as Figure 3 , Figure 4 Similarly, a defect A was found in the area of ​​the inspected object that could only be visually identified in the SE image. Figure 4 This illustrates a case where defect B, which is visually identifiable only in BSE images, occurs in an area of ​​the inspected object. Figure 8 and Figure 9 and Figure 3 and Figure 4 The difference lies in that an AI-based comparison (AI inspection) is performed between the synthesized image MP1 and the inference image generated by the learned AI model LP1, and an inspection image DMP1 is generated based on the result. Similarly, an AI-based comparison (AI inspection) is performed between the synthesized image MP2 and the inference image generated by the learned AI model LP2, and an inspection image DMP2 is generated based on the result. Furthermore, as inspection parameter 1, parameters optimized for the image at synthesis ratio a and the learned AI model LP1 are used. Additionally, as inspection parameter 2, parameters optimized for the image at synthesis ratio b and the learned AI model LP2 are used. According to this second embodiment, similar to the first embodiment, it is possible to simultaneously detect defects that are visually identifiable only in SE images and defects that are visually identifiable only in BSE images within a single synthesized inspection image. Furthermore, since it is not necessary to obtain reference images, a higher speed inspection is possible compared to the first embodiment.

[0055] Figure 10This is an example of a GUI screen displayed on the operation unit 206 in the defect inspection apparatus of the second embodiment. This GUI screen differs from the first embodiment in that it includes a parameter setting unit 503, an image display unit 504, a logic operation setting unit 505, and an inspection result image display unit 504. Figure 5 The same. However, Figure 10 In the GUI screen, one model can be selected from multiple learned AI models in the AI ​​model selection units 1001 and 1002 of the parameter setting unit 503. In addition, the inference images 1003 and 1004 corresponding to the selected learned AI model can be displayed in the image display unit 504.

[0056] Figure 11 This is a flowchart for generating a learned AI model as a prior preparation for checking the learned AI model. First, through having Figure 1 The scanning electron microscope with the optical system shown is moved to the normal circuit pattern position set as the learning object (learning object position) (step S1101), and SE image and BSE image are captured based on the detection signals of the upper detector 115 and the lower detector 117 (steps S1101, S1102).

[0057] Next, the SE and BSE images of the obtained learning object locations are combined at a synthesis ratio of a to generate learning image 1, and similarly, the SE and BSE images of the learning object locations are combined at a synthesis ratio of b to generate learning image 2 (steps S1103 and S1104). If there is another shooting location designated as the learning object, steps S1101 to S1104 are repeated at other shooting locations. After shooting is completed at all learning object locations, Figure 9 In the AI ​​model generation unit 902 of the illustrated computer system 101, a learned AI model LP1 is generated based on a learning image set with a synthesis ratio of a. Similarly, a learned AI model LP2 is generated based on a learning image set with a synthesis ratio of b, thus completing the prior preparation.

[0058] Figure 12 It shows the use according to Figure 11 This is an example of a flowchart illustrating the defect checking process of the learned AI models LP1 and LP2. (The last sentence appears to be incomplete and possibly refers to a separate process.) Figure 1 The scanning electron microscope with the illustrated optical system moves to the position of the object to be examined and acquires BSE and SE images based on the detection signals from the upper detector 115 and the lower detector 117 (steps S1201, 1202). Then, the SE and BSE images are synthesized at a synthesis ratio a to generate a composite image MP1 (step S1203).

[0059] Next, in the inference image generation unit 903, an AI inspection is performed on the synthesized image MP1 using the learned AI model LP1 (its inference image), generating an inspection image DMP1 (step S1204). Using the same process, an AI inspection is also performed on the synthesized image MP2 with a synthesis ratio b using the learned AI model LP2 (its inference image), generating an inspection image DMP2 (steps S1205, S1206). Next, an OR operation is performed on the inspection images DMP1 and DMP2 to generate a synthesized inspection image SDMP (step S1207). Finally, a defect determination is performed on the synthesized inspection image SDMP (step S1208), and the defect determination result is output to a specific inspection result file (step S1209). If a next inspection target point exists, the inspection continues, and the same process is repeated from step S1201. If no next inspection target point exists, the process ends. The computer system 901, for example, obtains the data from... Figure 10 The GUI screen shown in the example specifies the parameters and AI model used to handle these processes.

[0060] Furthermore, the present invention is not limited to the embodiments described above, and includes various modifications. For example, the embodiments described above are those that have been explained in detail for the purpose of easily understanding the present invention, and are not limited to having all the structures described. In addition, a part of the structure of a certain embodiment can be replaced with the structure of another embodiment, and it is also possible to add the structure of another embodiment to the structure of a certain embodiment. Furthermore, for a part of the structure of each embodiment, other structures can be added, deleted, or replaced.

Claims

1. A defect inspection device, characterized in that, have: The image synthesis unit, for a region designated as a defect inspection target, synthesizes a first detection signal detected by a first detector and a second detection signal detected by a second detector at a first synthesis ratio to generate a first synthesized image, and synthesizes the first detection signal and the second detection signal at a second synthesis ratio different from the first synthesis ratio to generate a second synthesized image. The defect inspection processing unit generates a first inspection image based on the first composite image and inspection parameters optimized for the first composite ratio, and generates a second inspection image based on the second composite image and inspection parameters optimized for the second composite ratio. The image processing unit performs logical operations on the first and second inspected images to generate a composite inspected image. as well as The defect determination processing unit performs defect determination on the synthesized inspection image. The first detector is a secondary electron detector, and the second detector is a backscattering electron detector.

2. The defect inspection device according to claim 1, characterized in that, The defect inspection processing unit generates the first inspection image and the second inspection image based on the first composite image or the second composite image and the reference image, wherein the reference image is an image generated based on the first detection signal and the second detection signal obtained at the reference position.

3. The defect inspection device according to claim 1, characterized in that, The defect inspection processing unit generates the first inspection image and the second inspection image based on the first synthetic image or the second synthetic image and the inference image, wherein the inference image is an image obtained by artificial intelligence learning the first detection signal and the second detection signal.

4. The defect inspection device according to claim 1, characterized in that, The defect inspection device also includes: The input unit receives the first synthesis ratio and the second synthesis ratio; and The display unit shows the first composite image and the second composite image generated by the first composite ratio and the second composite ratio input by the input unit.

5. The defect inspection device according to claim 4, characterized in that, Together with the first composite image and the second composite image, the display unit displays a reference image or an inferred image, wherein the reference image is an image generated based on the first detection signal and the second detection signal obtained at the reference position, and the inferred image is an image obtained by artificial intelligence learning the first detection signal and the second detection signal.

6. The defect inspection device according to claim 4, characterized in that, The input section is configured to input the types of logical operations.

7. A defect inspection method, characterized in that, The following steps are required: The steps are as follows: for an area designated as a defect inspection target, a first detection signal detected by a first detector and a second detection signal detected by a second detector are combined at a first synthesis ratio to generate a first composite image, and the first detection signal and the second detection signal are combined at a second synthesis ratio different from the first synthesis ratio to generate a second composite image. The steps are as follows: generating a first inspection image based on the first synthesized image and inspection parameters optimized for the first synthesis ratio, and generating a second inspection image based on the second synthesized image and inspection parameters optimized for the second synthesis ratio; The step of performing logical operations on the first inspection image and the second inspection image to generate a composite inspection image; as well as Perform a defect determination step on the synthesized inspection image. The first detector is a secondary electron detector, and the second detector is a backscattering electron detector.

8. The defect inspection method according to claim 7, characterized in that, The first inspection image and the second inspection image are generated based on the first composite image or the second composite image and the reference image, wherein the reference image is an image generated based on the first detection signal and the second detection signal obtained at the reference position.

9. The defect inspection method according to claim 7, characterized in that, The first inspection image and the second inspection image are generated based on the first synthetic image or the second synthetic image and the inferred image, wherein the inferred image is an image obtained by artificial intelligence learning the first detection signal and the second detection signal.

10. The defect inspection method according to claim 7, characterized in that, The defect inspection method also includes the following steps: The steps of inputting the first synthesis ratio and the second synthesis ratio; and The steps for displaying the first composite image and the second composite image generated using the input first composite ratio and second composite ratio are shown.

Images