Defect classification method and defect classification apparatus

Abstract

The disclosure provides a defect classification method and a defect classification device. The defect classification method comprises: obtaining an original image of a defect area and obtaining initial position information of a defect on the original image; obtaining an enlarged collection area of the defect area according to the initial position information; adjusting the enlarged collection area so that the defect is located in a preset area of the enlarged collection area and obtaining accurate position information of the defect; obtaining a fine image of the defect area with reference to the accurate position information; obtaining attribute information of the defect with reference to the fine image; and classifying the defect according to the attribute information. The defect classification method provided by the embodiment of the disclosure can correct the position of the defect based on the enlarged collection area, the deviation of the defect in the fine image obtained on this basis is reduced, or even there is no deviation, the accuracy of the attribute information of the extracted defect is high, and the classification accuracy is improved.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of integrated circuits, and in particular, to a defect classification method and a defect classification device. BACKGROUND

[0002] Defect classification is one of the important steps of online defect detection in the semiconductor manufacturing process. The traditional defect classification method is usually performed by engineers through visual judgment. This kind of classification method is time-consuming and laborious, and has low accuracy and poor stability. Defect review machines have automatic defect classification functions, which can achieve good automatic defect classification through the method of collecting a large amount of data and deep learning. However, due to the complex types and similar characteristics of defects, the actual defect classification still has the problems of low accuracy and poor stability, which cannot meet the needs. SUMMARY

[0003] Embodiments of the present disclosure provide a defect classification method and a defect classification device, which can improve the accuracy and stability of defect classification.

[0004] In one aspect, the present disclosure provides a defect classification method, which includes: obtaining an original image of a defect region, and obtaining initial position information of a defect on the original image; obtaining an enlarged collection area of the defect region according to the initial position information; adjusting the enlarged collection area so that the defect is located in a preset area of the enlarged collection area, and obtaining accurate position information of the defect; obtaining a fine image of the defect region with reference to the accurate position information; obtaining attribute information of the defect with reference to the fine image; and classifying the defect according to the attribute information.

[0005] In one embodiment, the step of obtaining the original image of the defect region includes: obtaining basic information of the defect, the basic information including coordinate information of the defect; and obtaining the original image of the defect region with reference to the basic information.

[0006] In one embodiment, the basic information further includes size information of the defect.

[0007] In the step of obtaining the original image of the defect region, the collection range of the original image is greater than the size range of the defect.

[0008] In one embodiment, the step of adjusting the enlarged collection area so that the defect is located in the preset area of the enlarged collection area includes: adjusting the enlarged collection area so that the center of the defect coincides with the center of the enlarged collection area.

[0009] In one embodiment, the step of adjusting the enlarged collection area includes: adjusting the position of the enlarged collection area.

[0010] In an embodiment, the adjusting the enlarged collection area comprises adjusting a window range of the enlarged collection area.

[0011] In an embodiment, the obtaining the original image of the defect area comprises obtaining a reference image of a non-defect area, and obtaining the initial position information of the defect according to a comparison between the reference image and the original image.

[0012] In an embodiment, the original image of the defect area and the reference image of the non-defect area are located in corresponding areas in different memory cells of a semiconductor structure, respectively.

[0013] In an embodiment, the original image of the defect area and the reference image of the non-defect area are located in corresponding areas in adjacent memory cells of a semiconductor structure, respectively.

[0014] In an embodiment, the method further comprises obtaining a fine image at a corresponding position of the non-defect area according to the accurate position information of the defect, and comparing the fine image of the defect area with the fine image at the corresponding position of the non-defect area to obtain attribute information of the defect.

[0015] In an embodiment, the method further comprises obtaining and recording a deviation between the accurate position information and the initial position information.

[0016] In an embodiment, after the step of adjusting the enlarged collection area so that the defect is located in a preset area of the enlarged collection area, the method further comprises obtaining an image of the enlarged collection area and storing the image as classification data of the defect.

[0017] In an embodiment, the attribute information of the defect comprises shape, size, gray scale and brightness of the defect.

[0018] Another aspect of the present disclosure also provides a defect classification device, comprising: an image coarse collection apparatus for obtaining an original image of a defect area and an enlarged collection area; a position collection apparatus for obtaining initial position information and accurate position information of a defect; an adjusting apparatus for adjusting the enlarged collection area so that the defect is located in a preset area of the enlarged collection area; an image fine collection apparatus for obtaining a fine image of the defect; and a classification apparatus for obtaining attribute information of the defect by taking the fine image as a reference, and for classifying the defect according to the attribute information.

[0019] In an embodiment, the image coarse collection apparatus is also used for obtaining a reference image of a non-defect area, and the image fine collection apparatus is also used for obtaining a fine image of the non-defect area.

[0020] In an embodiment, the method further comprises comparing the reference image with the original image, and comparing the fine image of the defect area with the fine image of the non-defect area.

[0021] The defect classification method provided by the embodiments of the present disclosure can correct the position of the defect based on the enlarged collection area, and the deviation of the defect in the fine image obtained based on this is reduced or even eliminated, so that the accuracy of the attribute information of the extracted defect is high, and the classification accuracy is improved. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required to be used in the embodiments of the present disclosure will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and other drawings can also be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0023] Figure 1A is a defect image provided by the first embodiment of the present disclosure;

[0024] Figure 1B is another defect image provided by the first embodiment of the present disclosure;

[0025] Figure 1C is still another defect image provided by the first embodiment of the present disclosure;

[0026] Figure 2 is a schematic diagram of the steps of the defect classification method provided by the second embodiment of the present disclosure;

[0027] Figure 3A is an original image of a defect area obtained by the present embodiment;

[0028] Figure 3B is a reference image of a non-defect area obtained by the present embodiment;

[0029] Figure 4 is an image of an enlarged collection area obtained by the present embodiment;

[0030] Figure 5A is a fine image of a defect area obtained by the present embodiment;

[0031] Figure 5B is a fine image of a non-defect area obtained by the present embodiment;

[0032] Figure 5C is a schematic diagram of superimposed comparison of the fine image of the defect area and the fine image at the corresponding position of the non-defect area obtained by the present embodiment;

[0033] Figure 6 is a schematic diagram of the defect classification device provided by the embodiments of the present disclosure. DETAILED DESCRIPTION

[0034] In order to make the objectives, technical solutions and effects of the present disclosure clearer, further explanations will be given below with reference to the drawings. It should be understood that the embodiments described herein are only a part of the embodiments of the present disclosure, rather than all the embodiments of the present disclosure, and are not used to limit the present disclosure. Based on the embodiments in the present disclosure, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present disclosure.

[0035] The first embodiment of the present disclosure provides a defect classification method. The defect classification method can be applied to classification of defects of a semiconductor structure. The method comprises the following steps: a defect scanning device obtains relevant information of defects of a semiconductor structure; a defect photographing device photographs an image of a defect according to the relevant information of the defect; and a defect automatic classification device classifies the defect according to the attribute of the image.

[0036] The inventors found that the defect classification method provided in the first embodiment has the problems of low accuracy and poor stability. Further analysis by the inventors found that the cause of the problem is that the image photographed by the defect photographing device has a problem, and the attribute information of the defect obtained has a deviation, so that accurate classification cannot be performed.

[0037] For example, in the image obtained by the defect photographing device, only part of the defect is located in the image, and the other part is located outside the image. As shown in FIG. 1, part of the defect A is located at the lower edge of the image, and the other part is not located in the image, so that accurate attribute information of the defect cannot be obtained. Further research by the inventors found that the cause of the phenomenon is that there is a position deviation between the defect scanning device and the defect photographing device, for example, a relative deviation of the coordinates of the two devices, so that the position of the defect in the image photographed by the defect photographing device according to the relevant information (for example, coordinate information) of the defect provided by the defect scanning device changes, and a complete image of the defect cannot be obtained. Figure 1A

[0037] For example, in the image obtained by the defect photographing device, only part of the defect is located in the image, and the other part is located outside the image. As shown in FIG. 1, part of the defect A is located at the lower edge of the image, and the other part is not located in the image, so that accurate attribute information of the defect cannot be obtained. Further research by the inventors found that the cause of the phenomenon is that there is a position deviation between the defect scanning device and the defect photographing device, for example, a relative deviation of the coordinates of the two devices, so that the position of the defect in the image photographed by the defect photographing device according to the relevant information (for example, coordinate information) of the defect provided by the defect scanning device changes, and a complete image of the defect cannot be obtained. Figure 1A

[0037] For example, in the image obtained by the defect photographing device, only part of the defect is located in the image, and the other part is located outside the image. As shown in FIG. 1, part of the defect A is located at the lower edge of the image, and the other part is not located in the image, so that accurate attribute information of the defect cannot be obtained. Further research by the inventors found that the cause of the phenomenon is that there is a position deviation between the defect scanning device and the defect photographing device, for example, a relative deviation of the coordinates of the two devices, so that the position of the defect in the image photographed by the defect photographing device according to the relevant information (for example, coordinate information) of the defect provided by the defect scanning device changes, and a complete image of the defect cannot be obtained.

[0037] For example, in the image obtained by the defect photographing device, only part of the defect is located in the image, and the other part is located outside the image. As shown in FIG. 1, part of the defect A is located at the lower edge of the image, and the other part is not located in the image, so that accurate attribute information of the defect cannot be obtained. Further research by the inventors found that the cause of the phenomenon is that there is a position deviation between the defect scanning device and the defect photographing device, for example, a relative deviation of the coordinates of the two devices, so that the position of the defect in the image photographed by the defect photographing device according to the relevant information (for example, coordinate information) of the defect provided by the defect scanning device changes, and a complete image of the defect cannot be obtained.

[0038]

[0037] For example, in the image obtained by the defect photographing device, only part of the defect is located in the image, and the other part is located outside the image. As shown in FIG. 1, part of the defect A is located at the lower edge of the image, and the other part is not located in the image, so that accurate attribute information of the defect cannot be obtained. Further research by the inventors found that the cause of the phenomenon is that there is a position deviation between the defect scanning device and the defect photographing device, for example, a relative deviation of the coordinates of the two devices, so that the position of the defect in the image photographed by the defect photographing device according to the relevant information (for example, coordinate information) of the defect provided by the defect scanning device changes, and a complete image of the defect cannot be obtained. Figure 1B

[0037] For example, in the image obtained by the defect photographing device, only part of the defect is located in the image, and the other part is located outside the image. As shown in FIG. 1, part of the defect A is located at the lower edge of the image, and the other part is not located in the image, so that accurate attribute information of the defect cannot be obtained. Further research by the inventors found that the cause of the phenomenon is that there is a position deviation between the defect scanning device and the defect photographing device, for example, a relative deviation of the coordinates of the two devices, so that the position of the defect in the image photographed by the defect photographing device according to the relevant information (for example, coordinate information) of the defect provided by the defect scanning device changes, and a complete image of the defect cannot be obtained. Figure 1BThe reason for the phenomenon shown is that the information provided by the defect scanning device is incorrect, causing the defect photographing device to take an abnormal photograph, for example, the defect scanning device considers that the defect is a large defect, but in fact the defect is a small defect, and the defect photographing device can only take a large range photograph according to the information provided by the defect scanning device, and cannot adjust the photographing range according to the actual situation of the defect, so that the obtained defect cannot be clearly displayed in the image.

[0039] For another example, in the image obtained by the defect photographing device, only part of the image of the defect can be observed, such as Figure 1C As shown in the image, only part of the image of the defect A can be observed, and accurate attribute information of the defect cannot be obtained. The inventors have found through in-depth research that the cause of Figure 1C The reason for the phenomenon shown is that the information provided by the defect scanning device is incorrect, causing the defect photographing device to take an abnormal photograph, for example, the defect scanning device considers that the defect is a small defect, but in fact the defect is a large defect, and the defect photographing device can only take a small range photograph according to the information provided by the defect scanning device, and cannot adjust the photographing range according to the actual situation of the defect, so that the obtained defect cannot be completely displayed in the image.

[0040] To solve the above problems, the second embodiment of the present disclosure provides a defect classification method, which can correct the position of the defect based on the enlarged collection area, so that the position deviation of the defect in the fine image obtained based thereon is reduced or even eliminated, thereby improving the accuracy of the extracted attribute information of the defect, and further improving the classification accuracy.

[0041] Figure 2 is a schematic diagram of the steps of the defect classification method provided by the second embodiment of the present disclosure. Please refer to Figure 2 , the defect classification method comprises:

[0042] In step S20, an original image of a defect area is obtained, and initial position information of the defect on the original image is obtained. Figure 3A For the original image of the defect area obtained by the present embodiment, please refer to Figure 3A , the defect is located in the area circled by the dashed line B.

[0043] The defect area refers to an area containing the defect. It can be understood that in the present embodiment, the defect area needs to be large enough so that the defect is entirely located in the defect area, that is, the original image can display the complete defect.

[0044] In this step, an image coarse collection device can be used to obtain the original image of the defect area. In some embodiments, the image coarse collection device can be integrated in a defect photographing device, and the original image of the defect area can be obtained by a scanning electron microscope (SEM) or the like.

[0045] As an example, in the present embodiment, the image rough collection device obtains the original image of the defect according to the basic information of the defect. The specific description is as follows:

[0046] The image rough collection device obtains the basic information of the defect. The basic information can be provided by the defect scanning device. For example, the defect scanning device performs scanning analysis on a semiconductor structure to obtain the basic information of the defect on the surface of the semiconductor structure, the image rough collection device obtains the basic information of the defect from the defect scanning device, or the defect scanning device sends the basic information of the defect to the image rough collection device.

[0047] The original image of the defect region is obtained according to the basic information. In some embodiments, the basic information includes coordinate information of the defect, and in some embodiments, the position collection device determines the position of the defect on the surface of the semiconductor structure according to the coordinate information, and then the image rough collection device photographs a region formed by expanding a preset distance from the center of the coordinate information as a defect region to obtain the original image. For example, the coordinate information of the defect is (X, Y), the image rough collection device determines the position of the defect on the surface of the semiconductor structure according to the coordinate (X, Y), and then the image rough collection device photographs a region with a horizontal coordinate range of (X-X1, X+X1) and a vertical coordinate range of (Y-Y1, Y+Y1) as a defect region, as shown in Figure 3A The coordinates of the four corners of the defect region are (X-X1, Y-Y1), (X-X1, Y+Y1), (X+X1, Y+Y1), and (X+X1, Y-Y1), respectively.

[0048] In some embodiments, the basic information further includes size information of the defect, and the size information can include a size range or an area of the defect. When the original image of the defect region is obtained, in order to ensure that the original image can display the complete defect, the collection range of the original image is greater than the size range of the defect, that is, the defect region covers the complete defect.

[0049] As shown in Figure 3A In the original image obtained according to the coordinate information of the defect provided by the defect scanning device, the defect is not located at the center position of the original image, but is offset relative to the center position, that is, the actual coordinate of the defect is not (X, Y). The reason for this phenomenon can be that there is a deviation in the coordinates between the defect scanning device and the image rough collection device. In other embodiments of the present disclosure, the defect can also be located at the center position of the original image, that is, the actual coordinate of the defect is (X, Y), and the present disclosure is not limited thereto.

[0050] The initial position information includes actual coordinates of the defect on the original image. As an example, the embodiments of the present disclosure also provide a method for obtaining initial position information of the defect on the original image. The method includes:

[0051] A reference image of a non-defect region is obtained. The non-defect region refers to a region on the semiconductor surface that does not contain the defect. In order to make the reference image have reference significance, the reference image includes all structural information in the original image except the defect. As shown in Figure 3B , which is a reference image of a non-defect region obtained by the embodiments, it can be seen that the reference image includes all structural information in the original image except the defect. Figure 3A and Figure 3B It can be seen that the reference image includes all structural information in the original image except the defect. In some embodiments of the present disclosure, the reference image can be stored in advance to improve the efficiency of defect classification and reduce the time cost. In the embodiments, when the original image of the defect region is obtained by using an image coarse collection device, the reference image of the non-defect region is obtained at the same time, which can obtain the initial position information of the defect more quickly compared with storing the reference image in advance.

[0052] In some embodiments, the original image and the reference image are located in corresponding regions in different storage units in the semiconductor structure, so that the reference image includes all structural information in the original image except the defect. In other embodiments, the original image of the defect region and the reference image of the non-defect region are located in corresponding regions in adjacent storage units in the semiconductor structure, so as to facilitate image acquisition.

[0053] According to the comparison between the reference image and the original image, the initial position information of the defect is obtained. Specifically, the reference image and the original image are compared in terms of shape, size, gray level, brightness and other characteristics, so as to determine the actual coordinates of the defect, i.e. the initial position information of the defect. For example, if the brightness at the coordinate (X3, Y3) of the original image is greater than the brightness at the coordinate (X3, Y3) of the reference image, it indicates that the defect is located at the coordinate of the original image, and the initial position information of the defect is the coordinate (X3, Y3).

[0054] In some embodiments, the initial position information of the defect is a coordinate point, which can be the center coordinate of the defect, and in other embodiments, the initial position information of the defect can also be a coordinate range, which includes the complete defect, and the present disclosure does not limit this.

[0055] Step S21: Obtain the magnified acquisition area of ​​the defect region based on the initial position information.

[0056] Since the initial location information is the location information of the defect in the original image, and the defect area is very small relative to the original image, the initial location information may still have deviations. Therefore, in this step, the defect area is magnified within a set range with the initial location information as the center to obtain the magnified acquisition area of ​​the defect area.

[0057] For example, in one embodiment, the initial position information is coordinates (X3, Y3). This step uses the area centered at coordinates (X3, Y3), with an abscissa range of (X3-X2, X3+X2) and a ordinate range of (Y3–Y2, Y3+Y2) as a set area to magnify the defect area, thus obtaining the magnified acquisition area. Here, X2 is less than X1, and Y2 is less than Y1, so that the magnified acquisition area remains within the defect area.

[0058] Step S22: Adjust the magnified acquisition area so that the defect is placed in a preset area of ​​the magnified acquisition area, and obtain the accurate location information of the defect.

[0059] In this step, the position of the defect in the magnified acquisition area is changed by adjusting the magnified acquisition area to ensure that the defect can be completely placed within the magnified acquisition area.

[0060] In some embodiments, adjusting the magnified acquisition area to place the defect within a preset area of ​​the magnified acquisition area includes: adjusting the magnified acquisition area so that the center of the defect coincides with the center of the magnified acquisition area, i.e., adjusting the magnified acquisition area to center the defect. In some embodiments, adjusting the magnified acquisition area includes adjusting the position of the magnified acquisition area or adjusting the window range of the magnified acquisition area. Adjusting the position of the magnified acquisition area means keeping the window range of the magnified acquisition area unchanged and only changing its edge coordinates to achieve position adjustment. Adjusting the window range of the magnified acquisition area means keeping the center coordinates of the magnified acquisition area unchanged and expanding or shrinking the edge coordinates by a set value to achieve window range adjustment.

[0061] like Figure 4 As shown, it is an image of the magnified acquisition area obtained in this embodiment. Figure 4 It can be seen that the defects ( Figure 3AThe center of the defect coincides with the center of the enlarged collection area. In some embodiments, an image of the enlarged collection area is obtained and stored as classification data of the defect, avoiding obtaining an enlarged image of the defect after attribute classification, and simplifying the operation steps. The image of the enlarged collection area can be obtained by SEM, and the magnification of the image of the enlarged collection area is greater than the magnification of the original image.

[0062] After adjusting the enlarged collection area, accurate position information of the defect is obtained, and the accurate position information includes coordinate information of the defect. For example, after adjusting the enlarged collection area, the defect is located at the center of the enlarged collection area, and the coordinate (X4, Y4) of the center of the enlarged collection area is taken as the accurate position information. In some embodiments, the accurate position information of the defect is a coordinate point, which can be the coordinate of the center of the defect, and in other embodiments, the accurate position information of the defect can also be a coordinate range, which includes the complete defect, and the disclosure does not limit this.

[0063] In step S23, a fine image of the defect region is obtained with reference to the accurate position information. Figure 5A is a fine image of the defect region obtained by the embodiment.

[0064] In this step, the accurate position information includes coordinate information of the defect, and the fine image of the defect region is obtained with reference to the coordinate information of the defect. The fine image can be obtained by SEM. The magnification and resolution of the fine image are greater than the magnification and resolution of the original image.

[0065] After the adjustment of the enlarged collection area, the defect is accurately positioned, and the accurate position information is closer to the actual position of the defect compared with the initial position information, so that the obtained fine image can include the complete defect.

[0066] In step S24, attribute information of the defect is obtained with reference to the fine image.

[0067] In some embodiments, the attribute information of the defect includes shape, size, gray scale, brightness, etc. of the defect.

[0068] The embodiment provides a method for obtaining attribute information of the defect with reference to the fine image, and the method comprises the following steps: Figure 5B comparing the fine image of the defect region with the fine image at the corresponding position of the non-defect region to obtain the attribute information of the defect. Please refer to Figure 5C, which is a schematic diagram of superimposed comparison of the fine image of the defect area and the fine image at the corresponding position of the non-defect area, and the defect is clearly presented (as indicated by the arrow in the figure), so that the attribute information of the defect can be determined.

[0069] In step S25, the defect is classified according to the attribute information.

[0070] In this step, the defect can be classified according to the shape, size, gray scale, brightness, etc. of the defect. In some embodiments, the classification can be performed according to one parameter, and in other embodiments, the classification can be performed according to two or more parameters. For example, in one embodiment, the defect is classified only according to the shape, and in another embodiment, the defect is classified according to the shape and brightness.

[0071] In this embodiment, the method further comprises obtaining and recording the deviation of the accurate position information and the initial position information as a reference for subsequent products of the same type, thereby improving the classification processing speed. Specifically, the deviation of the accurate position information and the initial position information is recorded, so that when subsequent products of the same type are classified, the accurate position information can be obtained directly according to the deviation, and then the fine image is obtained, thereby omitting the step of obtaining the accurate position information, and improving the classification processing speed.

[0072] The defect classification method provided by the embodiments of the present disclosure can correct the position of the defect based on the enlarged collection area, and the deviation of the defect in the fine image obtained based on this is reduced or even eliminated, so that the attribute information of the extracted defect is accurate, and the classification accuracy is improved. The defect classification method provided by the embodiments of the present disclosure can reduce the difference between the basic information related to the defect provided by the defect scanning device and the defect photographing device, and can reduce the problem of collecting incorrect defects or incomplete defects caused by the defect photographing device only being able to take pictures according to the information provided by the defect scanning device.

[0073] The embodiments of the present disclosure also provide a defect classification device for implementing the above defect classification method. Please refer to Figure 6 , which is a schematic diagram of the defect classification device provided by the embodiments of the present disclosure, and the defect classification device comprises an image coarse collection device 60, a position collection device 61, an adjustment device 62, an image fine collection device 63, a classification device 64, and a comparison device 65.

[0074] The image rough collection device 60 is used to obtain the original image of the defect area and the enlarged collection area. In some embodiments, the image rough collection device 60 obtains the original image of the defect according to the basic information of the defect. The image rough collection device 60 changes the magnification, so that the enlarged collection area can be obtained in the original image. The image rough collection device can be a scanning electron microscope (SEM), and the image rough collection device 60 can be integrated in a defect photographing device.

[0075] In some embodiments, the image rough collection device 60 is also used to obtain the image of the enlarged collection area, which is used as the classification data of the defect, so as to avoid obtaining the enlarged image of the defect after attribute classification, and to simplify the operation steps.

[0076] In this embodiment, the image rough collection device is also used to obtain the reference image of the non-defect area. The reference image includes all structural information in the original image except the defect. In some embodiments, when the image rough collection device obtains the original image of the defect area, the reference image of the non-defect area is also obtained.

[0077] The position collection device 61 is used to obtain the initial position information and the accurate position information of the defect. The initial position information and the accurate position information can include the coordinate information of the defect. In some embodiments, the position collection device 61 determines the position of the defect on the surface of the semiconductor structure according to the basic information of the defect provided by the defect scanning device. In some embodiments, the basic information includes the coordinate information of the defect.

[0078] The adjustment device 62 is used to adjust the enlarged collection area, so that the defect is placed in the preset area of the enlarged collection area. The adjustment device 62 adjusts the enlarged collection area, so that the center of the defect coincides with the center of the enlarged collection area, that is, the defect is placed in the center. In some embodiments, the adjustment device 62 adjusts the enlarged collection area, which includes adjusting the position of the enlarged collection area or adjusting the window range of the enlarged collection area. The adjustment device 62 adjusts the position of the enlarged collection area, which means that the window range of the enlarged collection area is kept unchanged, and only the edge coordinates are adjusted to realize the position adjustment. The adjustment device 62 adjusts the window range of the enlarged collection area, which means that the center coordinates of the enlarged collection area are kept unchanged, and only the edge coordinates are expanded or reduced by a certain value to realize the adjustment of the window range.

[0079] The image fine collection device 63 is configured to obtain a fine image of the defect. In some embodiments, the image fine collection device 63 can be a scanning electron microscope (SEM), and the image fine collection device 63 can be integrated in a defect photographing device. The image fine collection device 63 can be different from the image coarse collection device 60 in resolution, and the resolution of the image fine collection device 63 is greater than that of the image coarse collection device 60, so as to obtain images with different resolutions.

[0080] In the embodiment, the image fine collection device 63 is also configured to obtain a fine image of the non-defect region. The fine image of the non-defect region includes all structural information in the fine image of the defect region except the defect. In some embodiments, the image fine collection device 63 obtains the fine image of the defect region, and simultaneously obtains the fine image of the non-defect region.

[0081] The classification device 64 is configured to obtain attribute information of the defect based on the fine image, and to classify the defect based on the attribute information. The classification device 64 can classify the defect based on shape, size, gray level, brightness, etc. of the defect. In some embodiments, the classification device 64 can classify the defect based on one parameter, and in other embodiments, the classification device 64 can classify the defect based on two or more parameters. For example, in one embodiment, the defect is classified based on shape only, and in another embodiment, the defect is classified based on shape and brightness.

[0082] The comparison device 65 is configured to compare the reference image with the original image, and to compare the fine image of the defect region with the fine image of the non-defect region.

[0083] The defect classification device provided in the embodiments of the present disclosure can correct the position of the defect based on the magnified collection region, and the deviation of the defect in the fine image obtained based on the same is reduced or even eliminated, so that the accuracy of the attribute information of the extracted defect is high, and the classification accuracy is improved.

[0084] The above only describes the preferred embodiments of the present disclosure, and it should be noted that those skilled in the art can make several improvements and refinements without departing from the principles of the present disclosure, and these improvements and refinements should also be considered as the protection scope of the present disclosure.

Claims

1. A defect classification method characterized by, The method comprises the following steps: obtaining an original image of a defect region and obtaining initial position information of the defect on the original image; obtaining an enlarged collection area of the defect region according to the initial position information; adjusting the enlarged collection area so that the defect is located in a preset area of the enlarged collection area and obtaining accurate position information of the defect; obtaining a fine image of the defect region with reference to the accurate position information; obtaining attribute information of the defect with reference to the fine image; classifying the defect according to the attribute information; wherein the step of obtaining the original image of the defect region comprises: scanning and analyzing a semiconductor structure to obtain basic information of a defect on a surface of the semiconductor structure, the basic information comprising coordinate information of the defect; determining a position of the defect on the surface of the semiconductor structure according to the coordinate information, and taking a region formed by enlarging a preset distance from the coordinate information as a defect region for shooting to obtain the original image.

2. The defect classification method according to claim 1, characterized by, The basic information further comprises size information of the defect; in the step of obtaining the original image of the defect region, a collection range of the original image is larger than a size range of the defect.

3. The defect classification method of claim 1, wherein The step of adjusting the enlarged collection area so that the defect is located in the preset area of the enlarged collection area comprises: adjusting the enlarged collection area so that a center of the defect coincides with a center of the enlarged collection area.

4. The defect classification method of claim 1, wherein The step of adjusting the enlarged collection area comprises adjusting a position of the enlarged collection area.

5. The defect classification method of claim 1, wherein The step of adjusting the enlarged collection area comprises adjusting a window range of the enlarged collection area.

6. The defect classification method of claim 1, wherein The step of obtaining the original image of the defect region comprises: obtaining a reference image of a non-defect region, and obtaining the initial position information of the defect according to a comparison between the reference image and the original image.

7. The defect classification method of claim 6, wherein The original image and the reference image are located in corresponding regions in different storage units of a semiconductor structure.

8. The defect classification method of claim 6, wherein, The original image of the defect region and the reference image of the non-defect region are located in corresponding regions in adjacent storage units of a semiconductor structure.

9. The defect classification method of claim 6, wherein, The method further comprises: obtaining a fine image at a corresponding position of the non-defect region according to the accurate position information of the defect; comparing the fine image of the defect region with the fine image at the corresponding position of the non-defect region to obtain attribute information of the defect.

10. The defect classification method of claim 1, wherein The method further comprises obtaining and recording a deviation between the accurate position information and the initial position information.

11. The defect classification method of claim 1, wherein After the step of adjusting the enlarged collection area so that the defect is located in the preset area of the enlarged collection area, the method further comprises: obtaining an image of the enlarged collection area and storing the image as classification data of the defect.

12. The defect classification method of claim 1, wherein, The attribute information of the defect comprises shape, size, gray scale and brightness of the defect.

13. A defect classification apparatus characterized by comprising: The method comprises the following steps: a defect scanning device for scanning and analyzing a semiconductor structure to obtain basic information of a defect on a surface of the semiconductor structure; an image coarse collection device for obtaining an original image of a defect region and an enlarged collection area; a position collection device for obtaining initial position information and accurate position information of the defect; an adjusting device for adjusting the enlarged collection area so as to place the defect in a preset area of the enlarged collection area; an image fine collection device for obtaining a fine image of the defect; a classifying device for obtaining attribute information of the defect with reference to the fine image, and for classifying the defect according to the attribute information; wherein the basic information comprises coordinate information of the defect, the position collection device is used for determining the position of the defect on the surface of the semiconductor structure according to the coordinate information, and the image coarse collection device is used for taking a picture of an area formed by enlarging a preset distance from the coordinate information as the defect area to obtain the original image.

14. The defect classification apparatus according to claim 13, characterized by The image coarse collection device is also used for obtaining a reference image of a non-defect area, and the image fine collection device is also used for obtaining a fine image of the non-defect area.

15. The defect classification apparatus according to claim 14, characterized by Further comprising: a comparing device for comparing the reference image with the original image, and for comparing the fine image of the defect area with the fine image of the non-defect area.

Images