JP2010165876A - Defect correlation device, substrate inspection system, and method of correlating defects - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To display a relation of inspection results as to a top surface, a reverse surface and bevels of a substrate by low processing load. <P>SOLUTION: Two or more of inspection images are obtained by imaging at least two out of the top surface, the reverse surface and the bevels of the substrate in one or more of manufacturing process of a semiconductor device. Defects imaged in the respective selected inspection images are correlated, and processing is carried out to compose one two-dimensional composition image 100 from the respective selected inspection images, based on defect position information input as information indicating positions of the defects imaged in the respective selected inspection images, when receiving an instruction to select at least two out of the two or more of inspection images, and the composition image 100 is displayed. For example, the bevel defect is correlated with the other surface defects 103a-103d and the reverse surface defect 104, by projecting a position of the bevel defect in projection points 105a-105f, in the composition image 100. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to inspection in a manufacturing process of a semiconductor device.

  Conventionally, various inspections are performed in a manufacturing process of a semiconductor device such as an integrated circuit (sometimes referred to as “chip”) manufactured from a semiconductor wafer (hereinafter simply referred to as “wafer”). One of the purposes of the inspection is to improve the yield by detecting the defect, pursuing the cause of the detected defect, and performing improvement according to the cause.

  In recent years, in the manufacturing process before dicing, not only the inspection of the front surface of the wafer but also the inspection of the back surface and bevel of the wafer has been performed. The bevel is a portion corresponding to a side surface when the wafer is regarded as a very thin substantially cylindrical shape, in other words, a portion corresponding to the edge of a disk-shaped wafer.

  One reason for inspecting the backside of the wafer is that as the wafer becomes thinner, the possibility of defects occurring on the surface of the wafer has increased due to the effects of defects on the backside of the wafer. Further, one of the reasons for inspecting the wafer bevel is that it has been found that there is a possibility that a defect such as a damage to the front and back surfaces of the wafer may occur due to the influence of a defect such as a crack generated on the bevel. is there.

  In particular, in recent years, miniaturization of semiconductor devices has progressed, and since any part of the wafer is valuable, the arrangement pattern of the die is often determined so as not to waste the wafer. When a circuit pattern is formed very close to the edge of the wafer in an attempt to manufacture as many semiconductor devices as possible from a single wafer, the influence of the bevel on the surface of the wafer cannot be ignored.

  For example, if the bevel is slightly chipped when the vicinity of the bevel is held by a clamp or the like for transporting the wafer, particles scattered at that time may adhere to the surface. Alternatively, metal contaminants originating from clamps may adhere to the wafer surface. Thus, the particles or metal contaminants that wrap around and adhere to the surface of the wafer may cause uneven coating of the resist, which may cause the circuit pattern not to be formed correctly.

  There are various types of defects such as scratches, chips, contaminants such as metals, adhesion of foreign matters such as particles floating in the air and residual resist, and exposure failure due to defocus during exposure.

  In general, since a semiconductor device is manufactured through a plurality of manufacturing processes, inspection may be performed in a plurality of manufacturing processes. Further, in one manufacturing process, one to three types of inspections are appropriately performed among the three types of inspections, that is, the front surface, the back surface, and the bevel. One defective device may inspect the surface according to each of a plurality of inspection conditions, and the same applies to the back surface and the bevel.

  The defect inspection apparatus detects defects by, for example, imaging the front surface, back surface, or bevel of a wafer with an imaging device and comparing the obtained image with a reference image. The reference image is, for example, an image obtained by imaging a wafer that has been confirmed in advance to have no defects. For example, the defect inspection apparatus detects a defect based on a difference in luminance between an image obtained by imaging a wafer to be inspected and a reference image. The defect inspection apparatus may display the inspection result on a display or the like.

  Furthermore, the result of statistically processing the inspection results for a plurality of wafers is displayed, and images obtained by imaging the plurality of wafers are displayed side by side. The display method may be devised to improve the efficiency of analysis of the cause of the defect.

  For example, by superimposing a macro inspection image and a micro inspection result so that the state of association between the macro defect portion and the micro defect portion can be checked, the inspection efficiency for the glass substrate can be improved (for example, , See Patent Document 1).

  Alternatively, on the defect thumbnail display screen, an image that most significantly represents the feature of the defect may be determined and displayed for each defect from inspection information, defect type, and the like. In addition, on the detailed display screen of the defect, there are cases where the image to be displayed and the display order of the image are determined so as to express the feature of the defect remarkably based on the inspection information, the defect type, etc. (for example, (See Patent Document 2).

  In addition, the defect inspection apparatus performs end face (that is, bevel) inspection, front surface inspection, and back surface inspection, and converts the end surface image, the front surface image, and the back surface image into a three-dimensional coordinate system image based on the notch position and the center position of the substrate. In some cases, a three-dimensional shape is displayed on the display unit (see, for example, Patent Document 3).

Japanese Patent Laid-Open No. 2002-277412 JP 2007-225351 A JP 2008-196975 A

  In order to assist in improving the efficiency of analysis of the cause of defects in a substrate such as a wafer, it is considered effective to display the relationship between the inspection results of the front surface, back surface, and bevel of the substrate. Further, by reducing the processing load for displaying the relationship, the relationship can be displayed quickly and efficiently, and it is effective for improving the efficiency of the cause analysis. However, in the conventional substrate inspection in the manufacturing process of the semiconductor device, it is difficult to easily grasp the relationship between the inspection results of the front surface, back surface, and bevel of the substrate.

  Therefore, an object of the present invention is to provide a technique for displaying the relationship between the inspection results of the front surface, back surface, and bevel of a substrate with a light processing load.

According to one aspect of the present invention, there is provided a defect association apparatus including a selection instruction unit, a synthesis unit, and a display unit.
The selection instruction unit selects at least two of a plurality of inspection images obtained by imaging at least two of the front surface, the back surface, and the bevel of the substrate in one or more manufacturing steps of the semiconductor device. Accept instructions.

  The synthesizing unit is selected based on defect position information input as information indicating the position of a defect in each of two or more inspection images selected by the instruction received by the selection instruction unit. Further, the defects appearing in two or more inspection images are associated with each other, and a composite image that is one two-dimensional image is generated from the selected two or more inspection images.

The display means displays the synthesized image generated by the synthesizing means.
According to another aspect of the present invention, a substrate inspection system including the defect association apparatus is provided.

  The substrate inspection system further obtains the inspection image by imaging the front surface, the back surface, and the bevel of the substrate, and detects the position of the defect based on the inspection image to obtain the defect position information. A defect inspection device is provided. The substrate inspection system also includes a storage device that stores the plurality of inspection images obtained by the defect inspection apparatus.

  In the substrate inspection system, the defect position information is input from the defect inspection apparatus to the defect association apparatus, and the combining unit acquires the two or more selected inspection images from the storage device and Generate a composite image.

  According to still another aspect of the present invention, there is provided a defect association method that is executed when an information processing apparatus functions as the defect association apparatus.

  According to the present invention, two or more sheets obtained by imaging the front surface, the back surface, and the bevel of each of the substrates in one or more manufacturing steps of the semiconductor device with a lighter processing load than generating a three-dimensional image. It is possible to display the relevance between defects in the inspection image.

It is a figure which shows the example of the synthesized image in 1st Embodiment. It is a block diagram of a board | substrate inspection system. It is a flowchart which shows the operation | movement which one board | substrate inspection apparatus performs regarding one wafer in one manufacturing process. It is a figure explaining a wafer and a coordinate system. It is a figure explaining imaging of a bevel image. It is a figure which shows the example of the surface image used as the origin of the synthesized image, the back surface image, and the bevel image. It is a figure which shows the example of the screen which displays a test result. It is a figure which shows the example of the table | surface displayed on a wafer information display area. It is a figure which shows the example of the table | surface respectively displayed on a bevel defect summary display area, a surface defect summary display area, and a back surface defect summary display area. It is a figure which shows the example of the table | surface displayed on the same chip defect information display area. It is a figure which shows the example of the table | surface displayed on a surface inspection information display area, a back surface inspection information display area, and a bevel inspection information display area. It is a figure which shows the example of the bevel image of the whole 1 lot. It is an example of the graph which shows distribution of the bevel defect of the whole 1 lot. It is a figure which shows the example of the synthesized image in 2nd Embodiment. It is a flowchart of the process which highlights a bevel defect in association with a front surface defect or a back surface defect. It is a figure which shows the example of the synthesized image in 3rd Embodiment.

Hereinafter, a plurality of embodiments will be described in detail in the following order with reference to the drawings.
First, the composite image generated in the first embodiment will be described with reference to FIG. 1, and how the composite image is generated will be described with reference to FIGS. 2 to 6. Furthermore, an example of a screen that displays a combined image and other information in combination will be described with reference to FIGS. In addition, with reference to FIG. 12 and FIG. 13, information supplementarily displayed in the first embodiment is illustrated.

Thereafter, the second embodiment will be described with reference to FIGS. 14 and 15, and the third embodiment will be described with reference to FIG.
FIG. 1 is a diagram illustrating an example of a composite image in the first embodiment. In the first embodiment, the composite image 100 of FIG. 1 is generated and displayed in order to assist the investigation of the cause of the defect on the surface of the wafer in the manufacturing process of the semiconductor device. The synthesized image 100 is an image synthesized so that an operator (hereinafter referred to as “user”) of a device in a semiconductor device manufacturing factory can easily find the relationship between defects generated on the front surface, back surface, and bevel of the wafer. is there.

  In FIG. 1, a composite image 100 represents a substantially circular shape of a wafer 101 arranged in a predetermined direction. In the example of FIG. 1, the predetermined direction is a direction in which the notch 102 formed in the wafer 101 is directed downward.

  Hereinafter, defects detected on the surface of the wafer 101 (that is, the surface on which the circuit pattern is created) are referred to as “surface defects”. A defect detected on the back surface of the wafer 101 is referred to as a “back surface defect”, and a defect detected on the side surface of the wafer 101, that is, on the bevel, is referred to as a “bevel defect”.

The composite image 100 further represents the shape of the four surface defects 103a to 103d, and also represents the shape of the back surface defect 104 projected onto the surface.
The composite image 100 also represents six projection points 105 a to 105 f obtained by projecting the positions of the six bevel defects on the same plane as the surface of the wafer 101. Furthermore, the composite image 100 includes six line segments 106 a to 106 f connecting the six projection points 105 a to 105 f with the center point defined on the surface of the wafer 101.

  Further, the composite image 100 includes a plurality of small rectangles showing each die 107 in the wafer 101. The composite image 100 includes a plurality of rectangles arranged in a two-dimensional array. Each of these rectangles is a shot area 108 that is an area exposed by one exposure (that is, one shot). Indicates. In the example of FIG. 1, 4 × 6 dies 107 are included in one shot area 108.

  The composite image 100 as described above is a two-dimensional image expressed in a plane, but the inspection results regarding the front surface, the back surface, and the bevel of the wafer 100 are combined into one sheet. That is, the front surface defect, the back surface defect, and the bevel defect that are actually three-dimensionally distributed are associated with each other by a two-dimensional expression and expressed in the composite image 100.

  As is well known, the processing load for generating and displaying a two-dimensional image is generally lighter than the processing load for generating and displaying a three-dimensional image. This is because the two-dimensional image generation and display processing requires fewer parameters, the algorithm is simpler, and the required calculation amount is smaller than the three-dimensional image generation and display processing.

  Therefore, by generating and displaying the composite image 100 as described above, it is possible to comprehensively display the relationship between defects with a light processing load without sacrificing the function of displaying the relationship between defects. Become. Therefore, the user can quickly confirm the relevance of the defects and comprehensively guess the cause of the defect, and the composite image 100 is effective in improving the efficiency of the analysis of the cause of the defect.

  Specifically, the front surface defects 103a to 103d and the back surface defect 104 are displayed in the composite image 100 according to positions on a two-dimensional coordinate system (that is, positions represented by an x coordinate and a y coordinate described later). In other words, the relevance between the front surface defects 103a to 103d and the back surface defect 104 in the composite image 100 is represented by the distance and arrangement of positions on the two-dimensional coordinate system.

  Therefore, for example, the user can intuitively grasp from the composite image 100 that the distance between the front surface defect 103 a and the back surface defect 104 is shorter than the distance between the front surface defect 103 b and the back surface defect 104. . It is assumed that the shorter the distance, the higher the relevance between defects.

  Further, the relationship between the front surface defects 103a to 103d and the bevel defect and the relationship between the back surface defect 104 and the bevel defect are represented by the projection points 105a to 105f and also by the line segments 106a to 106f. Yes.

  For example, as understood from the fact that the line segment 106e crosses the range of the surface defect 103d, the direction of the line connecting the position of the surface defect 103d and the center point is very close to the direction of the line segment 106e. The direction of the line connecting the position of the surface defect 103d and the center point is also close to the direction of the line segment 106f.

  Such “closeness” between the directions is determined based on, for example, whether or not the image server 231 described later includes both of the two directions within the specified angle range specified with the center point as a reference. Alternatively, the determination may be made based on whether the angle difference between the two directions is equal to or less than a threshold value.

Next, how the composite image 100 in FIG. 1 is generated will be described.
FIG. 2 is a configuration diagram of the substrate inspection system. The board inspection system 200 of FIG. 2 includes a board inspection apparatus group 210, a DB (database) group 220, a server group 230, an image PC (Personal Computer) group 240, and a client PC 250. Since a wafer is a kind of substrate, hereinafter, the wafer may be simply referred to as “substrate”.

  The substrate inspection device group 210 includes n (n ≧ 1) substrate inspection devices 211-1 to 211-n. The board inspection apparatus 211-1 includes a plane inspection unit 212-1 for inspecting the front and back surfaces of the substrate and a bevel inspection unit 213-1 for inspecting the bevel of the substrate.

  The board inspection apparatus 211-1 includes a control unit (not shown) that controls the entire board inspection apparatus 211-1. The control unit can also be realized by a dedicated hardware circuit. In addition, a CPU (Central Processing Unit), a nonvolatile memory such as a ROM (Read Only Memory) for storing a program, a memory such as a RAM (Random Access Memory) for a working area, a storage device such as a hard disk device, The control unit can also be realized by a general-purpose computer having a connection interface with an external device.

  Further, the substrate inspection apparatus 211-1 may include an aligner (not shown) that aligns the wafer based on the position of the notch or orientation flat of the wafer. Moreover, the board | substrate inspection apparatus 211-1 may be provided with the display.

  The planar inspection unit 212-1 includes, for example, a first holding unit that holds a wafer at a bevel portion, a first imaging unit that images the front or back surface of the held wafer, and the front or back surface of an imaging target. The 1st illumination part to illuminate is provided.

The first holding unit may suck the wafer by vacuum suction or electrostatic chuck, or may hold the wafer by mechanical parts such as a clamp.
As the first imaging unit, a charge coupled device (CCD) image sensor or a complementary metal-oxide semiconductor (CMOS) image sensor can be used. The first imaging unit may capture a monochrome image called a luminance image, or may capture a color image. The first imaging unit may be a line sensor or an area sensor.

  Hereinafter, an image obtained by the first imaging unit imaging the front surface of the wafer is referred to as a “front image”, and an image obtained by the first imaging unit imaging the back surface of the wafer is referred to as a “back image”. That's it. The front image and the back image may be images obtained directly from imaging by the first imaging unit, or may be images indirectly obtained through some processing. For example, examples of processing include a process of combining a linear image captured by a line sensor into a planar image, luminance correction, shape distortion correction, noise removal, binarization, and the like.

  The bevel inspection unit 213-1 includes, for example, a second holding unit that holds the wafer by sucking the back surface of the substrate by vacuum suction or electrostatic chuck, and a second imaging unit that captures an image of the held bevel. And a second illumination unit that illuminates the bevel.

  A CCD image sensor or a CMOS image sensor can be used as the second imaging unit. The second imaging unit may capture a luminance image or a color image, and may be a line sensor or an area sensor. Hereinafter, an image obtained directly or indirectly by the second imaging unit imaging a wafer bevel is referred to as a “bevel image”. Further, as will be described later with reference to FIG. 5, the second imaging unit may include a plurality of image sensors.

  For example, the second imaging unit is realized by a fixed line sensor, the second holding unit holds the wafer on the rotation stage, and the rotation stage 360 is rotated at a rotation speed in accordance with the imaging timing by the second imaging unit. You may rotate it. In this case, the second imaging unit repeats imaging while the rotary stage rotates. A belt-like bevel obtained by observing the bevel along the substantially circular outer periphery of the wafer by synthesizing and outputting a planar image from the linear image obtained by the imaging by the second imaging unit which is a line sensor. An image is obtained.

The other board inspection apparatuses 211-2 to 211-n are configured in the same manner as the board inspection apparatus 211-1.
The DB group 220 includes an inspection DB 221 and a recipe DB 222. In FIG. 2, the inspection DB 221 and the recipe DB 222 are one each, but a configuration in which a plurality of DBs are distributed is also possible.

  The inspection DB 221 receives, accumulates, and manages inspection result data from the substrate inspection apparatuses 211-1 to 211-n via the network. The inspection result data includes data on the front surface image, the back surface image, and the bevel image. Hereinafter, the front image, the back image, and the bevel image may be collectively referred to as “inspection image”.

  The inspection DB 221 may store only one type of surface image for one inspection condition, or may store a plurality of types of surface images. For example, the substrate inspection apparatuses 211-1 to 211-n, for example, display both a luminance image obtained by imaging the wafer surface according to a certain inspection condition and a binary image obtained by processing the luminance image. Both may be registered in the inspection DB 221 as a surface image. By storing the luminance image as a surface image in the inspection DB 221, it is possible for the image server 231 or the like to redetect a defect by changing conditions such as a threshold when detecting the defect. The same applies to the back image and the bevel image.

  The inspection result data stored in the inspection DB 221 includes information (for example, information indicating the position, size, type, etc.) of defects detected by the substrate inspection apparatuses 211-1 to 211-n from the inspection image. You may go out.

  The recipe DB 222 manages and stores “recipe” data, which is information for designating inspection conditions such as optical conditions. Each of the substrate inspection apparatuses 211-1 to 211-n acquires the latest recipe from the recipe DB 222 via the network, and performs wafer inspection according to the acquired recipe.

In addition, the recipe may contain one or more of the following information, for example.
Identification information for identifying a manufacturing process for manufacturing a semiconductor device from a wafer Information specifying an observation method (for example, information specifying bright field observation, dark field observation, or observation by diffracted light)
-Information that specifies the irradiation angle for irradiating the wafer with illumination light-Information about whether or not to use a polarization filter or wavelength selection filter-Information that specifies the type of filter when using a filter-First or second filter Information specifying whether to attach to the illumination unit 2 or to the first or second imaging unit ・ In the manufacturing process identified by the identification information, the surface inspection, the back surface inspection, and the bevel inspection The information server group 230 indicating the necessity of each includes an image server 231 and a recipe server 232. Note that both the image server 231 and the recipe server 232 may be a single server, or may be configured to distribute the load to a plurality of servers.

  The image server 231 is connected to the inspection DB 221 via a network. Application software for browsing inspection result data registered in the inspection DB 221 is installed in the image server 231. The image server 231 is also installed with application software for statistically processing inspection result data and re-extracting defects from the inspection image by appropriately setting conditions for detecting defects. Yes.

  The recipe server 232 is connected to the recipe DB 222 via a network. The recipe server 232 is installed with application software for viewing and editing recipes registered in the recipe DB 222 and registering recipes updated by editing in the recipe DB 222.

  In addition, the recipe server 232 supports the inspection DB 221 and the network so that the recipe can be edited while referring to the setting of the existing recipe and the inspection result of the recipe in order to support the setting of an appropriate recipe. Connected through. As described above, the recipe server 232 downloads the inspection image from the inspection DB 221, processes the inspection image while adjusting the conditions for detecting the defect, redetects the defect, and tunes the recipe through trial and error. Provide functionality.

  The image PC group 240 includes m (m ≧ 1) PCs 241-1 to 241-m for browsing inspection images and the like. Each of the PCs 241-1 to 241-m is connected to the image server 231 and the recipe server 232 via a network. Each of the PCs 241-1 to 241-m functions as a client of the recipe server 232, thereby providing an interface for the user to edit (that is, tune) the recipe.

  The client PC 250 is also connected to the image server 231 and the recipe server 232 via a network, and functions as a client of the image server 231 and the recipe server 232. Thereby, the client PC 250 also provides an interface for the user to view and edit inspection images and recipes.

  As described above, the PCs 241-1 to 241-m and the client PC 250 can access the inspection DB 221 and the recipe DB 222 via the inspection DB 221 and the image server 231 and use the information stored in each DB.

Next, an outline of processing in each of the substrate inspection apparatuses 211-1 to 211-n in the substrate inspection system 200 configured as shown in FIG. 2 will be described.
FIG. 3 is a flowchart showing an operation performed by one substrate inspection apparatus 211-j (1 ≦ j ≦ n) for one wafer in one manufacturing process. Each substrate inspection apparatus 211-j sequentially inspects a plurality of wafers. 3 is executed in each of a plurality of manufacturing steps. One board inspection apparatus 211-j may be in charge of only one specific manufacturing process, or may be inspected in a plurality of manufacturing processes. In the following description, the manufacturing process focused on in FIG. 3 is referred to as “target manufacturing process”.

In step S <b> 101, the board inspection apparatus 211-j downloads a recipe for the target manufacturing process from the recipe DB 222.
Next, in step S102, a substrate transfer device (not shown) carries the substrate into the substrate inspection device 211-j. In step S102, the aligner of the substrate inspection apparatus 211-j may align the substrate. Steps S101 and S102 may be executed in reverse order.

  Subsequently, in step S103, the plane inspection unit 212-j of the substrate inspection apparatus 211-j captures the surface of the loaded substrate according to the inspection conditions specified in the recipe, thereby acquiring a surface image. If the inspection conditions specified in the recipe for the surface inspection are A (A ≧ 1), in step S103, the plane inspection unit 212-j images the surface of the substrate according to each inspection condition. A surface image of the sheet is obtained.

  In step S104, the plane inspection unit 212-j performs inspection based on the surface image to extract (that is, detect) surface defects. For example, the planar inspection unit 212-j performs alignment between the reference image and the surface image acquired in step S103 by a method such as template matching, and then compares the luminance difference between the reference image and the surface image with a threshold value. Based on this, surface defects may be extracted. The plane inspection unit 212-j may calculate the size and position of the extracted defect, and may further determine the type of defect.

  When the A inspection conditions are specified in the recipe for the surface inspection, the plane inspection unit 212-j performs the inspection based on each of the A surface images (that is, each of the A inspection conditions in step S104). Perform the corresponding inspection).

  In step S105, the substrate inspection apparatus 211-j displays the inspection result on the display included in the substrate inspection apparatus 211-j, and registers and stores the inspection result in the inspection DB 221 via the network. The inspection result in step S105 includes the surface image obtained in step S103 and information indicating the size, position, type, and the like of the defect obtained in step S104.

  Subsequently, in step S106, the control unit of the substrate inspection apparatus 211-j determines whether to inspect the back surface in the target manufacturing process according to the recipe. If the back side is inspected, the process proceeds to step S107. If the back side is not inspected, the process proceeds to step S110.

  In step S107, the plane inspection unit 212-j acquires a back image by imaging the back surface of the substrate according to the inspection conditions specified in the recipe. If the inspection conditions specified in the recipe for the back surface inspection are B (B ≧ 1), B back surface images are obtained in step S107.

  Then, in step S108, the plane inspection unit 212-j performs inspection based on the back image and extracts back surface defects. The method for extracting the back surface defect may be, for example, the same method as the method for extracting the surface defect. Alternatively, since no circuit pattern is formed on the back surface, the back surface defect is determined based on whether the pixel value of the back image is within a range defined by a predetermined upper limit and lower limit without using a reference image. It can also be extracted. Further, the plane inspection unit 212-j may calculate the size and position of the extracted defect, and may further determine the type of defect.

  When B inspection conditions are specified in the recipe for the back surface inspection, in step S108, the plane inspection unit 212-j performs inspection based on each of the B back images (that is, for each of the B inspection conditions). Perform the corresponding inspection).

  In step S109, the substrate inspection apparatus 211-j displays the inspection result on the display included in the substrate inspection apparatus 211-j, and registers and stores the inspection result in the inspection DB 221 via the network. The inspection result in step S109 includes the back side image obtained in step S107 and information indicating the size, position, type, and the like of the defect obtained in step S108.

  Subsequently, in step S110, the control unit of the substrate inspection apparatus 211-j determines whether or not to inspect the bevel in the target manufacturing process according to the recipe. When the bevel inspection is performed, the process proceeds to step S111. When the bevel inspection is not performed, the process proceeds to step S114.

  In step S111, the substrate inspection apparatus 211-j moves the substrate from the planar inspection unit 212-j to the bevel inspection unit 213-j. And the bevel inspection part 213-j acquires a bevel image by imaging the bevel of a board | substrate according to the inspection conditions designated by the recipe.

As will be described later, when the bevel is classified into C parts (C> 1), the bevel inspection unit 213-j individually captures each part and acquires a bevel image corresponding to each part. May be. Further, in this case, assuming that there are D _k inspection conditions (D _k ≧ 1) for the inspection of the k th part (1 ≦ k ≦ C), in step S111, the expression (1) E bevel images defined in (1) are obtained.

In step S112, the bevel inspection unit 213-j executes an inspection based on the bevel image and extracts a bevel defect. For example, the bevel inspection unit 213-j may extract a bevel defect based on whether or not the pixel value of the bevel image is within a range defined by a predetermined upper limit and lower limit. Since the notch of the wafer appears dark in the bevel image, the bevel inspection unit 213-j may change the upper and lower limit values in the notch and portions other than the notch. Further, the bevel inspection unit 213-j may calculate the size and position of the extracted defect, and may further determine the type of defect.

When the bevel is classified into C parts and D _k inspection conditions are specified in the recipe for the inspection of the k-th part, the bevel inspection unit 213-j performs E bevel images in step S112. (I.e., E inspections corresponding to E combinations of the part and the inspection condition) are performed.

  In step S113, the board inspection apparatus 211-j displays the inspection result on the display included in the board inspection apparatus 211-j, and registers the inspection result in the inspection DB 221 via the network. The inspection result in step S113 includes the bevel image obtained in step S111 and information indicating the size, position, type, and the like of the defect obtained in step S112.

Finally, in step S114, the substrate transfer apparatus carries the substrate out of the substrate inspection apparatus 211-j, and the processing in FIG.
As described above, in this embodiment, the back surface and the bevel on which the circuit pattern is not formed are inspected only in the necessary manufacturing process according to the recipe.

  FIG. 4 is a diagram for explaining the wafer and the coordinate system. FIG. 4A is a diagram for explaining the wafer 300 and the xyz coordinate system (that is, a three-dimensional orthogonal coordinate system also called Cartesian coordinate system). FIG. 4B is a diagram illustrating an rθ coordinate system (that is, a polar coordinate system) used to indicate the position of the bevel defect. FIG. 4C is a cross-sectional view of the wafer 300 in the vicinity of the bevel.

  FIG. 4A shows a perspective view of the wafer 300 with the surface 301 facing upward, with the x-axis, y-axis, and z-axis superimposed. The back surface 302 of the wafer 300 is hidden in FIG. For convenience of illustration, the notch 303 is highlighted in FIG.

  In the following, it is assumed that the surface 301 is on the xy plane, the thickness direction of the wafer 300 is the z direction, and the surface 301 side is the positive side of the z axis. The direction connecting the center point of the surface 301 and the notch 303 is the y direction, and the notch 303 is on the negative side of the y axis.

  FIG. 4B shows the relationship between the xy plane and the rθ coordinate system. In the following, it is assumed that the θ coordinate in the 6 o'clock direction with the notch 303 is 0, and the θ coordinate increases counterclockwise. The r coordinate indicates the distance from the center point defined as (x, y) = (0, 0). Therefore, the following expressions (2) to (5) are satisfied. Note that the unit of θ in the equations (2) to (5) is radians, and using the equations (2) to (5), between the xyz coordinate system that is an orthogonal coordinate system and the rθz coordinate system that is a cylindrical coordinate system. Can be converted.

x = r · cos (θ−π / 2) (2)
y = r · sin (θ−π / 2) (3)

θ = arctan (y / x) + π / 2 (5)
For example, the plane inspection unit 212-j represents the position of the detected front surface defect and back surface defect with xy coordinates, and registers them in the inspection DB 221 as a part of the inspection result. Further, the bevel inspection unit 213-j represents the position of the detected bevel defect with the θ coordinate and registers it in the inspection DB 221 as a part of the inspection result.

  In FIG. 4A, for convenience of illustration, the z direction (that is, the thickness direction) is emphasized. As shown in FIG. 4A, the bevel 304 of the wafer 300 of this embodiment has a shape such that the distance from the z-axis (that is, the moving radius r) is the largest in the portion between the front surface 301 and the back surface 302. is there. A specific cross-sectional shape is as shown in FIG. FIG. 4C is a part of a sectional view of the wafer 300 taken along a plane passing through the z axis.

  As shown in FIG. 4C, the bevel 304 is classified into five parts according to the position in the thickness direction of the wafer 300. Each part is referred to as Near Edge Front 305, Upper Bevel 306, Apex 307, Lower Bevel 308, and Near Edge Back 309 in order from the positive side of the z-axis. For example, the standard of SEMI (Semiconductor Equipment and Materials International) may be adopted to determine the range of each part.

  In this embodiment, the bevel inspection units 213-1 to 213-n in FIG. 2 are specified so that the cause of the defect can be specified more accurately, as clearly shown in FIG. 5 and FIG. An examination focusing on each part of the bevel is performed. Then, the image server 231 comprehensively determines the examination results for each of the five parts.

  FIG. 5 is a diagram for describing imaging of a bevel image. When the bevel 304 of the wafer 300 is classified into five parts as shown in FIG. 4C (that is, when C = 5 in the description of FIG. 3), the bevel image corresponding to each part is displayed as the bevel inspection unit 213. FIG. 5 shows the imaging optical axis when -j is imaging.

  In the example of FIG. 5, the imaging optical axis 310 when imaging a bevel image corresponding to the Near Edge Front 305 is perpendicular to the plane of the Near Edge Front 305. The imaging optical axis 311 when imaging a bevel image corresponding to Upper Bevel 306 is perpendicular to the surface of Upper Bevel 306, and the imaging optical axis 312 when imaging a bevel image corresponding to Apex 307 is perpendicular to the surface of Apex 307. is there. Similarly, the imaging optical axis 313 when imaging a bevel image corresponding to the Lower Level 308 is perpendicular to the surface of the Lower Level 308. The imaging optical axis 314 when imaging a bevel image corresponding to the Near Edge Back 309 is perpendicular to the surface of the Near Edge Back 309.

  For example, the second imaging unit included in the bevel inspection unit 213-j may include five fixed image sensors, and the imaging optical axes of the image sensors may be the imaging optical axes 310 to 314 in FIG. Alternatively, the second imaging unit may include one movable image sensor that can move at least to five positions such that the imaging optical axis coincides with each of the imaging optical axes 310 to 314.

  Similarly, the 2nd illumination part with which the bevel test | inspection part 213-j is provided may also contain 5 sets of illumination sources corresponding to 5 site | parts. Each set of illumination sources may be composed of one illumination source in some embodiments, or may be composed of a plurality of illumination sources having different wavelengths or the like. In addition, the second illumination unit may include a set of illumination sources configured to be movable so that the five parts can be illuminated in order. Note that the illumination optical axis when coaxial illumination is specified in the recipe matches the imaging optical axes 310 to 314 of FIG.

  Of course, reflective lighting may be specified in certain recipes. In this case, a non-perpendicular angle may be designated as the imaging optical axis so that imaging is performed from an angle that is not perpendicular to the surface of each part of the bevel 304 (for example, the surface of Apex 307). In any case, the bevel inspection unit 213-j performs imaging according to the recipe.

  FIG. 6 is a diagram illustrating an example of the front image, the back image, and the bevel image that are the basis of the composite image 100 in FIG. 1. As described with reference to FIG. 2, the front image, the back image, and the bevel image may be images obtained indirectly from imaging.

  Hereinafter, for convenience of explanation, the image server 231 generates the composite image 100 of FIG. 1 by application software installed in the image server 231, and various types of information including the composite image 100 are displayed on the screen of the display provided in the image server 231. Shall be displayed. However, various types of information including the composite image 100 may be displayed on a screen of a display included in the PCs 241-1 to 241-m or the client PC 250 that accesses the image server 231 via the network. Further, the PCs 241-1 to 241-m or the client PC 250 may generate the composite image 100.

  In the example of FIG. 6, the surface image 400 is a binary image obtained by the plane inspection unit 212-j of the board inspection apparatus 211-j by the processing in steps S <b> 103 and S <b> 104 in FIG. 3. That is, the plane inspection unit 212-j images the surface 301 of the wafer 300 in step S103, and processes the image obtained by the imaging by binarization or the like for defect extraction in step S104. As a result, a surface image 400 is obtained.

  Similarly, the back surface image 410 is obtained by the plane inspection unit 212-j imaging the back surface 302 of the wafer 300 in step S107, and processing the image obtained by the imaging by binarization or the like for defect extraction in step S108. It is the binary image obtained by the above.

  Further, the bevel image 420 is processed by the bevel inspection unit 213-j imaging a part of the wafer 300 where the bevel 304 is located in step S111, and binarizing and the like for defect extraction in step S112. It is the binary image obtained by doing.

In the front surface image 400, the back surface image 410, and the bevel image 420, which are binary images obtained through processing in this way, the defect area is black and the area other than the defect is white.
That is, in the surface image 400, the areas of the surface defects 401a to 401d corresponding to the surface defects 103a to 103d in FIG. 1 are shown in black.

  Moreover, in the back surface image 410, the linear area | region of the back surface defect 411 corresponding to the back surface defect 104 of FIG. 1 is represented black. As can be seen from a comparison between FIG. 1 and FIG. 6, the position and shape of the back surface defect 104 in the composite image 100 in FIG. 1 can be obtained by the image server 231 performing mirror image conversion of the back surface defect 411 in the back surface image 410.

In the belt-like (that is, rectangular) bevel image 420, the areas of the bevel defects 421a to 421f are shown in black. Here, the left end of the bevel image 420 corresponds to the point where the value of the θ coordinate defined in FIG. 4B is 0 °, and the right end of the bevel image 420 is the point where the value of the θ coordinate is 360 °. Correspond. Therefore, when the value of theta coordinates representing respective positions of the bevel defects 421a~421f and theta _a through? _F, from the bevel defects 421a~421f respective horizontal coordinates in the rectangular bevel image 420, the theta _a through? _F It can be calculated. For example, when the horizontal width of the bevel image 420 is H, the horizontal coordinate value of the bevel defect 421a is h _a , and the unit is expressed in radians, θ _a is calculated by Expression (6).

θ _a = 2πh _a / H (6)
That is, the image server 231 calculates the horizontal direction of the projection point 105a on the composite image 100 from θ _a calculated using Expression (6), the radius of the wafer in the surface image 400, Expression (2), and Expression (3). And the vertical coordinate can be calculated. Therefore, the image server 231 projects the bevel defect 421a in that the contour of theta = theta _a surface image 400 according to the calculated value of the theta _a, expressed as a projection point 105a in the composite image 100 of FIG. 1 can do.

The relationship between the other bevel defects 421b to 421f and the projection points 105b to 105f is the same as described above.
Note that the above image processing is conceptually a surface representing the xy plane by projecting the position of the bevel defect 421a shown in the bevel image 420 in the xyz space onto the same xy plane as the surface 301 of the wafer 300. It can also be regarded as a process of superimposing on the image 400.

  As described above, the image server 231 converts the back image 410 into a mirror image, superimposes it on the front image 400, and projects the positions of the bevel defects 421 a to 421 f of the bevel image 420 onto the front image 400. A composite image 100 is generated. In the present embodiment, the image server 231 further executes processing for forming the line segments 106 a to 106 f in FIG. 1 on the composite image 100.

The method for obtaining the composite image 100 of FIG. 1 has been described above with reference to FIGS.
Next, an example of a screen displayed in combination with the composite image 100 in FIG. 1 will be described with reference to FIGS. The screens described in FIGS. 7 to 11 also provide a GUI for allowing the user to select an inspection image that is the basis of the composite image 100. In addition, the numerical example in the following description is for convenience of description, and does not indicate an actual numerical tendency or a preferable number of significant figures.

  FIG. 7 is a diagram illustrating an example of a screen that displays an inspection result. A screen 500 in FIG. 7 is displayed on the display of the image server 231 or the PCs 241-1 to 241-m or the client PC 250 that accesses the image server 231 as a client. In addition to the control for displaying the screen 500, the image server 231 further displays a GUI (Graphical User Interface) such as a menu for selecting the target data to be displayed on the screen 500 from the examination DB 221. Control for displaying on the screen may be performed.

  In FIG. 7, a screen 500 includes a wafer information display area 501, an inspection number selector 502, a save button 503, a histogram display area 504, a defect display area 505, a bevel defect summary display area 506, a surface defect summary display area 507, and a back surface defect summary display. An area 508, the same chip defect information display area 509, an overlay button 510, a front surface inspection information display area 511, a back surface inspection information display area 512, a bevel inspection information display area 513, and a composite image display area 514 are included.

The wafer information display area 501 is an area for displaying wafer information, which is a kind of header information, and includes, for example, a table as shown in FIG.
FIG. 8 is a diagram illustrating an example of a table displayed in the wafer information display area 501. The first line of the table is a heading line in which “wafer information” is written. The items in the table of FIG. 8 represent the following contents.

-Product name The product name of the wafer 300 to be inspected (that is, identification information for identifying the type of the semiconductor device manufactured from the wafer 300) is "AAA-012."

・ Lot ID
The ID (identifier) of the currently focused lot (that is, the lot whose information is displayed on the screen 500) is “345-BBB-678”.

Attribute name, date, and time In the histogram display area 504 and the defect display area 505, the inspection result of the bevel 304 performed on April 2, 2008 at 13:52:57 is displayed.

Inspection device name, inspection recipe version The screen 500 has a version of “1” by an inspection device named “CCC-901” (for example, one of the substrate inspection devices 211-1 to 211-n in FIG. 2). The inspection result according to the inspection recipe is displayed.

Process name The name of the target manufacturing process is “PROC-3”.
・ Exposure device name, coater / developer name In the manufacturing process “PROC-3”, a coater / developer named “EEE-567” applies a resist to the wafer 300, and an exposure device named “DDD-234” An exposure process is performed, and the coater / developer performs development.

・ Slot number, wafer ID
In the composite image display area 514, the composite image 100 related to the wafer 300 having the slot number “12” with the ID “Loader_1.12” is displayed. The slot number is a number for identifying individual wafers in one lot composed of a plurality of wafers. In the present embodiment, since one lot consists of 25 wafers, numerical values of 1 to 25 are used as slot numbers.

・ Wafer level (maximum value)
For the wafer of slot number “12”, the maximum value among the values obtained by the image server 231 calculating the level indicating the degree of surface unevenness according to each inspection condition is “4.5”.

-Defect area, number of defects, number of defective chips, inspection results The area of defects (including all surface defects, back surface defects, and bevel defects) detected as a result of the inspection relating to the manufacturing process "PROC-3" is 7341 mm ² There are 8363. There are 221 defective chips. Since these results do not satisfy a predetermined standard, the wafer 300 was determined to be “failed” by the inspection device “CCC-901” in the manufacturing process “PROC-3”.

Depending on the embodiment, some of the items exemplified above may be omitted, and other items may be further displayed in the wafer information display area 501.
Returning to the description of FIG. 7, the inspection number selector 502 is a GUI for selecting an inspection number for identifying an inspection condition and switching the display. When the inspection number is selected by the inspection number selector 502, the image server 231 switches part or all of the display contents of the screen 500 according to the selection result.

  A save button 503 is a GUI for saving display contents collected as a report on the screen 500 in a file. For example, when a user viewing the screen 500 via the PC 241-j clicks the save button 503 with the mouse, the PC 241-j transmits a save request to the image server 231. Then, the image server 231 transmits a file including the contents of the screen 500 to the PC 241-j, and the PC 241-j stores the received file.

  In the histogram display area 504, a histogram representing the defect classification result is displayed. For example, when the bevel defect is classified into three types of a line defect, a surface defect, and a point defect by the inspection apparatus, the image server 231 calculates the area of each of the line defect, the sheet defect, and the point defect. Originally, a histogram indicating the ratio of three types of defects is generated.

  By displaying the histogram, it becomes possible for the user to intuitively grasp what kind of defects are many by looking at the screen 500. Note that the histogram may be based on the number instead of the area.

  The image server 231 receives a command for instructing to display a histogram related to the front surface defect or the back surface defect in the histogram display area 504 via a menu (not shown) or another screen. Then, the image server 231 switches display contents of the histogram display area 504 according to the command.

  The defect display area 505 is an area showing an image of a defect actually detected. For example, the image server 231 discriminates three defects having the top first to third areas in each classification among the three types of bevel defects classified into linear defects, planar defects, and point defects. To do. For the determination, the image server 231 may refer to data regarding the area of the defect among the inspection result data stored in the inspection DB 221. The image server 231 cuts out the determined nine (= 3 × 3 types) defect portions from the bevel image stored in the inspection DB 221 and displays the cut-out image in the defect display area 505.

  Note that the image server 231 receives a command for instructing to display the image of the front surface defect or the rear surface defect cut out from the front surface image or the back surface image in the defect display area 505 via a menu (not shown) or another screen. The image server 231 switches display contents of the defect display area 505 according to the command.

  As described above, the image server 231 cuts out and displays only the part determined to be defective by the board inspection apparatus 211-j from the inspection image, so that the user can analyze the inspection result and investigate the cause of the defect more efficiently. So that the user is supported.

  The bevel defect summary display area 506, the front surface defect summary display area 507, and the back surface defect summary display area 508 are areas for displaying a summary of statistical information about defects, and include, for example, a table as shown in FIG.

  FIG. 9 is a diagram illustrating examples of tables displayed in the bevel defect summary display area 506, the surface defect summary display area 507, and the back surface defect summary display area 508, respectively. The user can grasp the correlation between the bevel defect, the front surface defect, and the back surface defect from the table of FIG. For example, the table of FIG. 9 is effective for allowing the user to grasp the tendency that a wafer with few bevel defects has few surface defects or back surface defects.

  The formats of the three tables are the same, and the first line is a heading line in which “defect list” is written. The heading row further indicates which inspection result is a bevel inspection, a front surface inspection, and a back surface inspection. In this embodiment, the results of the inspection under the inspection condition “1” selected in FIG. 7 in the manufacturing process “PROC-3” shown in FIG. 8 are displayed in the three tables of FIG. It is assumed that

Each item in the three tables of FIG. 9 represents the following contents.
Defect type number and defect type Bevel defects are classified into three types, “linear”, “planar”, and “dot”, depending on their shapes. In addition, surface defects and back surface defects are classified into two types depending on the shape: linear “comet” with a thickness like a comet and extremely thin linear “scratch” with a scratch. . The defect type number and the defect type indicate the number and name of each classification.

    The board inspection apparatuses 211-1 to 211-n classify the defects based on the inspection image, register the classified result in the inspection DB 221 as a part of the inspection result data, and the image server 231 performs the classification from the inspection DB 221. The table of FIG. 9 may be created by reading the results.

・ Defect number Indicates the number of defects classified into each defect type. According to FIG. 9, the total number of bevel defects is 8000 (= 1424 + 1244 + 5332), the total number of surface defects is 178 (= 62 + 116), and the total number of back surface defects is 185 (= 13 + 172). Note that the “number of defects” in FIG. 8 is the sum of defect numbers 8363 (= 8000 + 178 + 185) in FIG.

・ Total defect area (mm ² )
The sum of the areas of defects classified into each defect type is shown. The area of each defect is calculated, for example, by the substrate inspection apparatuses 211-1 to 211-n based on the inspection image, and is registered in the inspection DB 221 as a part of inspection result data. The image server 231 can create the table of FIG. 9 by reading the classification result from the inspection DB 221.

According to FIG. 9, the total area of bevel defects is 3204 (= 2461 + 503 + 240) mm ² , the total area of surface defects is 3952 (= 2885 + 1067) mm ² , and the total area of backside defects is 185 (= 63 + 122) mm ² . Note that the “defect area” in FIG. 8 is the sum of defect total areas 7341 in FIG. 9 (= 3204 + 3952 + 185) mm ² .

・ Defect area (%)
The ratio of the area for each defect type is shown. For example, the ratio of the area of the bevel defect of the type “linear” to the total area of the bevel defect is 77% (= 2461/3204 × 100).

  Depending on the embodiment, some of the items exemplified above may be omitted, and items other than the above may be further added to the bevel defect summary display area 506, the surface defect summary display area 507, and the back surface defect summary display area 508. May be displayed.

Next, the same chip defect information display area 509 in FIG. 7 will be described with reference to FIG.
FIG. 10 is a diagram showing an example of a table displayed in the same chip defect information display area 509. A circuit pattern of each chip is formed in a two-dimensional array on the surface of the wafer, and each chip is identified by a set of an index in the x direction and an index in the y direction (hereinafter referred to as “chip index”).

  When a defect is detected with a chip having the same chip index in different wafers, there may be a cause in a semiconductor manufacturing apparatus, a substrate transfer apparatus, or the like. The table of FIG. 10 is a table that assists the user in finding the cause of the occurrence of defects common to different wafers.

  The table of FIG. 10 shows statistical information on defects detected by chips having the same chip index in units of one lot of 25 wafers. That is, the table of FIG. 10 represents the following.

  -As a result of inspecting one lot (that is, 25 wafers) under the inspection condition of inspection number "1", there were three wafers in which defects were detected in the chip represented by the chip index (15, 7). . The three wafers are specifically wafers with slot numbers 4, 8, and 20.

  -As a result of inspecting one lot (that is, 25 wafers) under the inspection condition of inspection number "2", there were two wafers in which defects were detected in the chip represented by the chip index (4, 36). . The two sheets are specifically wafers with slot numbers 9 and 17.

  Therefore, the table of FIG. 10 shows that the semiconductor manufacturing apparatus or the substrate transfer apparatus has a defect that causes a defect at a specific chip position represented by the chip index of (15, 7) and (4, 36). Suggests the possibility to do. In other words, the table of FIG. 10 provides a judgment material for narrowing down a portion to be checked intensively when the user checks whether there is a defect in the semiconductor manufacturing apparatus and the substrate transfer apparatus. Since the inspection number is also displayed in the table of FIG. 10, the table of FIG. 10 can also be used for comparison of defect detection sensitivity for each inspection condition.

  Note that the conversion formula for converting the xy coordinates on the surface of the wafer into a chip index is uniquely determined from, for example, design data of individual types of wafers. Therefore, the image server 231 can create the table of FIG. 10 by reading the inspection result data from the inspection DB 221, recognizing the xy coordinates of the position of each defect, and calculating the chip index from the above conversion formula. . Depending on the embodiment, some of the items exemplified above may be omitted, and other items may be further displayed in the same chip defect information display area 509.

  Returning to the description of FIG. The overlay button 510 is a GUI for instructing the image server 231 to generate a composite image and display it in the composite image display area 514. It should be noted that from which inspection image the composite image should be generated is designated in the front surface inspection information display region 511, the back surface inspection information display region 512, and the bevel inspection information display region 513. For example, the image server 231 detects an operation in which the user clicks the overlay button 510, and generates and displays a composite image in response to detection of the click operation.

Next, the front surface inspection information display area 511, the back surface inspection information display area 512, and the bevel inspection information display area 513 will be described with reference to FIG.
FIG. 11 is a diagram illustrating examples of tables displayed in the front surface inspection information display area 511, the back surface inspection information display area 512, and the bevel inspection information display area 513. These three tables list all the inspections that have passed through the wafer and include a GUI for instructing the image server 231 as an input to use any combination of inspection images to generate a composite image.

  In the surface inspection information display area 511, a table regarding the surface 301 of the wafer 300 is displayed. In this table, one row corresponds to the inspection result in one manufacturing process, and the line corresponding to the latest manufacturing process is displayed higher so that the user can see the transition of the inspection result in each manufacturing process from the comparison of each row. The Each line includes the following items.

Check box The check box is a GUI for allowing the user to select whether or not to use the inspection result represented by the row for generating a composite image. In FIG. 11, among the three lines corresponding to the three manufacturing processes, only the check box in the first line is checked, and only the inspection result of the manufacturing process corresponding to the first line is selected as the input of the composite image. ing.

Process Color This item indicates what color the inspection result represented by the row is represented in the composite image when the inspection result represented by the row is used to generate a composite image. The color may be indicated by letters or the actual color may be displayed. According to FIG. 11, the image server 231 respectively detects defects in the inspection images of the processes “PROC-1”, “PROC-2”, and “PROC-3” (if selected as input for synthesis). Represented in orange, yellow, and red, a composite image 100 is generated.

-Process This item indicates the name of the manufacturing process.
Test Result This item represents the test result represented by the row by, for example, “pass” or “fail”.

・ Number of defects (maximum value)
This item is the number of surface defects detected under the inspection condition in which the most surface defects are detected among the inspections performed according to one or more inspection conditions regarding the manufacturing process represented by the row.

    For example, regarding the latest manufacturing process “PROC-3” shown in FIG. 8, the number of surface defects detected under the inspection condition of the inspection number “1” selected by the inspection number selector 502 in FIG. 9 is 178 (= 62 + 116). According to FIG. 11, the value of the item “number of defects (maximum value)” in the row of the manufacturing process “PROC-3” in the table of the surface inspection information display area 511 is also 178. That is, regarding the manufacturing process “PROC-3”, the most surface defects are detected under the inspection condition of the inspection number “1”.

-Number of defective chips This item indicates the number of chips determined to be defective as a result of the inspection relating to the manufacturing process represented by the row. The number of defective chips of 221 in the manufacturing process row of “PROC-3” matches the “number of defective chips” in FIG.

Defect area maximum value This item indicates the maximum value of the surface defect area detected in the inspection relating to the manufacturing process represented by the row. The image server 231 can recognize the maximum defect area value, for example, by reading inspection result data from the inspection DB 221.

In the back surface inspection information display area 512, a table related to the back surface 302 of the wafer 300 is displayed. The format of this table is the same as the table displayed in the surface inspection information display area 511.
In the bevel inspection information display area 513, a table related to the bevel 304 of the wafer 300 is displayed. The format of this table differs from the above two tables by one column, and there is a column of “total defect area” instead of the column of “number of defective chips”. In the example of FIG. 11, the result of performing the bevel inspection in the manufacturing process “PROC-3” according to one or more inspection conditions is displayed. Specifically, the following is displayed.

  In order to instruct the image server 231 to generate a composite image including the result of the bevel inspection in the manufacturing process “PROC-3”, a check mark is added to the check box.

The result of the bevel inspection in the manufacturing process “PROC-3” is “pass”.
The number of bevel defects detected in the inspection condition where the most bevel defects are detected is 8301.

· The total area of the determination portion to be defective in at least one or more test conditions is 3204mm ^2. This area corresponds to the sum of the total defect areas in the table displayed in the bevel defect summary display area 506 in FIG.

The area of the largest detected bevel defect is 2.7840 mm ²
As described above, according to the GUI shown in FIG. 11, the user can click on the overlay button 510 in FIG. The displayed composite image can be switched. Therefore, the GUI of FIG. 11 and the overlay button 510 of FIG. 7 allow the user to intuitively obtain knowledge that, for example, if a defect occurs in the bevel in the first step, the number of surface defects is likely to increase after the second step. Make it possible.

  Depending on the embodiment, some of the items exemplified above may be omitted, and items other than the above may be further added to the front surface inspection information display region 511, the back surface inspection information display region 512, and the bevel inspection information display region 513. May be displayed.

  Returning to the description of FIG. 7, the composite image display area 514 is an area for displaying a composite image such as the composite image 100 of FIG. 1, and is displayed whenever the overlay button 510 is clicked as described above. Is switched.

The example of the screen 500 has been described above with reference to FIGS.
Subsequently, in addition to the screen 500 in order to show the relationship between defects, examples of other information supplementarily displayed by the image server 231, the PCs 241-1 to 241-m, or the client PC 250 will be described with reference to FIGS. Will be described with reference to FIG.

  FIG. 12 is a diagram illustrating an example of a bevel image of an entire lot. For example, one lot consists of 25 wafers. In FIG. 12, 25 belt-like bevel images 420-1 to 420-25 corresponding to 25 wafers are displayed side by side so that portions corresponding to the same angle are aligned vertically. That is, in FIG. 12, the θ axis is a horizontal coordinate axis.

  The display as shown in FIG. 12 is effective for assisting the user in finding an angle at which a defect is likely to occur regardless of the slot. When there is a specific angle at which a defect is likely to occur regardless of the slot, there is a high possibility that the cause of the defect is hidden in a semiconductor manufacturing apparatus, a substrate transfer apparatus, or the like.

  FIG. 13 is an example of a graph showing the distribution of bevel defects in one lot. In the example of FIG. 13, the bevel 360 ° is divided into nine sections each of 40 °, and a polyline plotting the number of defects detected in each section added to one lot (ie, 25 wafers). The graph is shown. 13A and 13B, the horizontal axis is the θ axis, and the vertical axis is the number of defects.

Specifically, FIG. 13A is a summary of the number of defects detected in one lot of wafers for each part, and is compiled into five line graphs corresponding to the five parts. Note that although there are _Dk inspection conditions (1 ≦ k ≦ C = 5 and _Dk ≧ 1) defined for each part, each line graph in FIG. It is a graph about one inspection condition in _Dk ways.

  The graph of FIG. 13A is useful for assisting the user in registering, for example, an angle at which many defects occur regardless of the part or a specific angle at which many defects occur at a specific part. The graph of FIG. 13A is effective for assisting the user in understanding which part of the bevel 304 is likely to have a defect and investigating the cause of the defect.

  FIG. 13B shows the total number of defects detected by Apex 307 of one lot of wafers for each inspection condition and corresponds to five inspection conditions written as “# 1” to “# 5”. This is a summary of five line graphs. In this example, the line graph of the inspection condition “# 5” matches the line graph of “Apex” shown in FIG.

The image server 231 can also create and display a line graph similar to that in FIG. 13B for four parts other than the Apex 307.
The graph of FIG. 13B helps the user narrow down inspection conditions effective for detecting defects and assists the user in reviewing the inspection conditions. As a result, the wafer can be efficiently processed with a small number of inspection conditions. It helps to set up recipes to inspect and reduce inspection time.

  Subsequently, the second embodiment will be described with reference to FIGS. 14 and 15. In the second embodiment, the composite image format is different from that of the first embodiment. Description of points common to the first embodiment is omitted.

  FIG. 14 is a diagram illustrating an example of a composite image in the second embodiment. The composite image 600 of FIG. 14 includes a wafer surface region 601 expressed as a circle representing the edge of the surface 301 of the wafer 300. In FIG. 14, the shape of the notch 303 is omitted, and the wafer surface region 601 is shown in a circular shape. Further, five circles are formed concentrically with the wafer surface region 601 in the composite image 600, and ring-shaped regions between two adjacent concentric circles correspond to the five portions of the bevel 304, respectively.

  In the example of FIG. 14, a Near Edge Front region 602, an Upper Bevel region 603, an Apex region 604, a Lower Bevel region 605, and a Near Edge Back region 606 are sequentially arranged from the inside of the five ring-shaped regions. That is, in FIG. 4C, the portion with a larger z coordinate is closer to the surface 301 of the wafer 300, and therefore, in FIG.

  The composite image 600 represents a large number of defects. For example, a plurality of defects including the surface defect 607 (specifically, the surface defect and the mirror image-converted back surface defect) are expressed in the wafer surface region 601. In the Near Edge Front area 602, a plurality of bevel defects including the bevel defect 608 (more specifically, bevel defects generated in the Near Edge Front 305) are displayed. Further, a plurality of bevel defects including the bevel defect 609 (more specifically, bevel defects generated in the lower bevel 308) are displayed in the lower bevel area 605.

  In the second embodiment, each bevel defect (for example, bevel defects 608 and 609) may be represented by a point in the composite image 600, or may be represented by a predetermined two-dimensional shape such as a circle or a rectangle. Alternatively, the image server 231 may represent each bevel defect in the composite image 600 as a projection area obtained by projecting the bevel defect area in the bevel image stored in the inspection DB 221 onto the xy plane. In this case, each bevel defect is not a uniform shape, but is represented by an individual two-dimensional shape corresponding to the shape of each bevel defect in the bevel image, and thus the bevel defect type is highly visible.

  Similar to the first embodiment, for example, for the lower bevel 308 in FIG. 4C, the second imaging unit included in the bevel inspection unit 213-j matches the optical axis of the imaging optical axis 313 in FIG. 5 with the wafer 300. Image. As a result, a rectangular belt-like bevel image is obtained in the same manner as the bevel image 420 in FIG. The image server 231 converts the obtained rectangular bevel image into a ring-shaped Lower Bevel area 605 having a width as shown in FIG. This conversion can be conceptually regarded as an operation of projecting the position of the defect on the lower bevel 308 onto the xy plane (that is, the same plane as the surface 301 of the wafer 300) in a three-dimensional space.

  The image server 231 generates an image of a ring-shaped region in the same manner for the other four parts of the bevel 304. The image server 231 then superimposes and combines the front image, the mirror image-converted back image, and the five ring-shaped images obtained as described above, thereby forming a composite image that is a two-dimensional image. 600 is generated.

The image server 231 may further form several line segments on the composite image 600. For example, the composite image 600 of FIG. 14 includes a line segment 610 indicating θ = 0.
Thus, in the composite image 600, the front surface defect, the back surface defect, and the bevel defect are expressed in association with each other on the two-dimensional plane. Furthermore, in the second embodiment, the image server 231 also provides a function for assisting the user in finding a defect group having a high positional relationship. Specifically, the image server 231 generates the composite image 600 in which defects with high positional relevance are highlighted by executing the processing of FIG.

FIG. 15 is a flowchart of processing for highlighting a bevel defect in association with a front surface defect or a back surface defect.
In step S201, the image server 231 receives input designating both ends of the range of the θ coordinate, and recognizes the designated both ends, that is, the angles θ ₁ and θ ₂ (θ ₁ <θ ₂ ). In the following description, it is assumed that the inputs specifying the angles θ ₁ and θ ₂ are given by the user via the input device of the image server 231 for convenience of explanation. However, as a matter of course, the input may be given via an input device included in the PC 241 -k or the client PC 250 and a network.

The inputs specifying the angles θ ₁ and θ ₂ are given through a pointing device such as a mouse as inputs specifying two points on the display of the image server 231 that displays the composite image 600, for example. The image server 231 may recognize the angles θ ₁ and θ ₂ by converting the coordinates of the two specified points into the rθ coordinate system.

Upon recognizing the angles θ ₁ and θ ₂ , the image server 231 may form a line segment 611 indicating θ = θ ₁ and a line segment 612 indicating θ = θ ₂ on the composite image 600 as shown in FIG. Good. By forming the line segments 611 and 612, the visibility of a range having a spread of Θ (= θ ₂ −θ ₁ ) is increased.

Subsequently, in step S202, the image server 231 searches for defects having θ coordinates satisfying θ ₁ ≦ θ ≦ θ ₂ among defects in each of the five parts of the bevel 304, and displays them in the same color. Since the θ coordinate indicating the position of each bevel defect is included in the inspection result data stored in step S113 in FIG. 3, the image server 231 refers to the inspection result data to identify each bevel defect. It can be determined whether or not θ ₁ ≦ θ ≦ θ ₂ is satisfied.

As will be described later, since the process of FIG. 15 includes a repeated loop, step S202 may be executed a plurality of times. Therefore, more precisely, when executing step S202 at the p-th time (p ≧ 1), the image server 231 selects all the ranges in the composite image 600 corresponding to each bevel defect that satisfies θ ₁ ≦ θ ≦ θ _2. Displayed in the pth color. If p ≠ q, the p-th color and the q-th color are different from each other.

  Further, for example, as shown in FIG. 11 of the first embodiment, when the composite image 600 is first generated by color-coding for each manufacturing process, the pth color is a color different from the color of any manufacturing process. It is preferable.

In step S203, the image server 231 searches for defects on the front surface 301 and the back surface 302 that have θ coordinates satisfying θ ₁ ≦ θ ≦ θ ₂ and displays them in the same color.

Here, the xy coordinates indicating the position of each surface defect are included in the inspection result data stored in step S105 of FIG. Further, the xy coordinates indicating the position of each back surface defect are included in the inspection result data stored in step S109 of FIG. Therefore, the image server 231 can calculate the θ coordinate of each front surface defect and each back surface defect using Expression (5), and whether each front surface defect and each back surface defect satisfies θ ₁ ≦ θ ≦ θ _2. Can be determined.

More precisely, when the image server 231 executes step S203 for the p-th time, all the ranges in the composite image 600 of the surface defects and the back surface defects that satisfy θ ₁ ≦ θ ≦ θ ₂ are all in the p-th color. Is displayed.

  By executing the above steps S202 and S203, in the composite image 600, the defect included in the sector surrounded by the line segments 611 and 612 in FIG. 14 is expressed in the same pth color. That is, the defects that are included in the set spread of Θ and are presumed to have a high positional relationship with each other are expressed in the same color by the image server 231 so that the defects can be easily recognized by the user. The positional relevance of is emphasized. Therefore, the highlighting in steps S202 and S203 is useful for supporting an analysis for investigating the cause of the defect from the positional relationship between the defects. Note that steps S202 and S203 may be executed in reverse order.

  Thereafter, in step S204, the image server 231 determines whether a defect group to be highlighted still remains. If there are still defect groups to be highlighted, the process returns to step S201, and if there are no more defect groups to be highlighted, the process of FIG. Note that the determination in step S204 is based on whether a specific input is given from the user via the input device of the image server 231, for example.

  Next, a third embodiment will be described with reference to FIG. In the third embodiment, the image server 231 performs highlight display based on a predetermined constant φ. Description of points common to the second embodiment will be omitted as appropriate.

  FIG. 16 is a diagram illustrating an example of a composite image in the third embodiment. The composite image 600b in FIG. 16 is similar to the composite image 600 in FIG. 14, with the wafer surface region 601 as the center, the near edge front region 602, the upper bevel region 603, the apex region 604, the lower bevel region 605, and the near edge back. Region 606 is included.

Further, for example, φ = 36 ° is determined in advance, and the composite image 600b includes 10 line segments that equally divide 360 ° by the angle φ.
In the example of FIG. 16, a first sector sandwiched between a line segment 613 indicating θ = φ and a line segment 614 indicating θ = 2φ, a line segment 617 indicating θ = 6φ, and a line segment 618 indicating θ = 7φ. There are many defects in the second sector sandwiched between. For example, the first sector includes a bevel defect 615 in the Near Edge Back region 606, a surface defect 616 in the wafer surface region 601 and the like. The second sector includes a bevel defect 619 in the Apex region 604, a surface defect 620 in the wafer surface region 601 and the like.

  In this manner, the image server 231 forms a plurality of line segments on the composite image 600b according to a predetermined constant φ, and further, all the images included in the range φ · i ≦ θ <φ · (i + 1). The surface defect, the back surface defect, and the bevel defect are expressed by the same i th color. For simplification, if the unit of φ is degrees and φ is a divisor of 360, 0 ≦ i <360 / φ, and at least the i-th color and the (i + 1) -th color are different.

  Thus, according to the third embodiment, defects that are close in position are expressed in the same color without intervention from the user, so that the user can quickly visually recognize the positional relationship between the defects. . Therefore, the third embodiment is also useful for supporting the analysis for pursuing the cause of the defect from the positional relationship between the defects.

The present invention is not limited to the above-described embodiment, and can be variously modified. Some examples are described below.
The information displayed on the display is not limited to the information exemplified above. For example, coordinate values of individual defects may be further displayed. The image server 231 may label each defect and include a label in the composite image.

  The format of the composite image is not limited to that exemplified in the above embodiment. For example, in FIG. 1, the projection points 105a to 105f can be omitted, and the position of the bevel defect can be expressed only by the line segments 106a to 106f, and vice versa. The image server 231 may further provide a GUI for independently switching display / non-display of the projection points 105a to 105f and the line segments 106a to 106f.

  Further, the image server 231 may represent the front surface defect and the back surface defect in the composite image 100 according to the actual defect shape as shown in FIG. 1, and the actual defect shape may be a predetermined shape such as a circle or a rectangle. May be replaced and expressed.

An arbitrary method may be used for the image server 231 to emphasize defects having a strong positional relationship.
For example, the image server 231 may receive input specifying a surface defect via an input device. Then, the image server 231 may highlight the back surface defect whose distance from the designated front surface defect is equal to or less than the threshold value by a method such as expressing it in the same color in the composite image. The image server 231 further highlights a projection point, a projection area, a predetermined shape, or a line segment that represents a bevel defect whose proximity to the specified surface defect satisfies the criterion by a method such as expressing the same color in the composite image. May be.

Here, the image server 231 determines whether or not the proximity between the designated surface defect and the bevel defect satisfies the criterion, the position (r _t , θ _t ) of the surface defect expressed in the rθ coordinate system, and the bevel defect. it can be determined based on the theta coordinate theta _b.

For example, the image server 231 may determine that “a surface defect and a bevel defect are close if f (r _t , θ _t , θ _b ) ≦ T” using an appropriate function f and an appropriate threshold value T. . The function f (r _t , θ _t , θ _b ) may be, for example, a function that decreases as r _t increases when θ _t and θ _b are fixed. Further, the image server 231 may use different functions or different threshold values for different parts of the bevel 304.

Of course, the image server 231 may make a determination independent of the value r _t of the r coordinate of the surface defect. For example, the image server 231 determines that the proximity between the surface defect and the bevel defect is “if θ ₁ ≦ θ _t ≦ θ ₂ and θ ₁ ≦ θ _b ≦ θ ₂ ” as in the second embodiment. Also good. Alternatively, the image server 231 determines the proximity between the surface defect and the bevel defect as “φ · i ≦ θ _t <φ · (i + 1) and φ · i ≦ θ _b <φ · (i + 1) as in the third embodiment. If so, it may be determined.

  Further, the bevel image 420 in FIG. 6 and the bevel images 420-1 to 420-25 in FIG. 12 are images obtained by imaging one part of the bevel 304. However, in the first embodiment, the image server 231 may generate a new bevel image by vertically arranging five belt-shaped bevel images obtained from imaging of five parts and combining them into one sheet. . Then, the image server 231 may use the generated new bevel image instead of the bevel image 420 or 420-1 to 420-25.

  Of course, the image server 231 may calculate the position of the projection point using each of the five bevel images corresponding to the five parts of the bevel 304 individually. In the composite image 100, for example, the image server 231 represents, for example, a projection point corresponding to a bevel defect on the Near Edge Front 305 with a first color, and represents a projection point corresponding to the bevel defect on the Upper Bevel 306 with a second color. May be. Similarly, the image server 231 may represent the five parts by distinguishing them with five colors for the other parts.

  In this way, the image server 231 visually associates different colors for each part, so that each part of the bevel 304 having a three-dimensional spread in the composite image 100 that is one two-dimensional image is visually displayed. Distinction can be realized.

  Alternatively, the visual distinction of each part may be a distinction not by color but by shape as long as the user can recognize the difference in the display form of the projection points. For example, the image server 231 may use a method in which projection points corresponding to bevel defects on the Near Edge Front 305 are represented by circles, and projection points corresponding to bevel defects on the Upper Bevel 306 are represented by triangles.

  Further, the image server 231 may change the display form of the line segment connecting the center point of the wafer 101 and the projection point in the composite image 100 for each part. The display form of the line segment can be, for example, a color, a thickness, a broken line pattern, a length from the center point, and a combination thereof.

  When overviewing the above-described embodiment, an information processing apparatus such as the image server 231, the PCs 241-1 to 241-m, or the client PC 250 can be used alone or in cooperation with other information processing apparatuses. Function as.

  For example, the image server 231 accepts an instruction to select at least two of the plurality of inspection images obtained by imaging the front surface, the back surface, and the bevel of the substrate in one or more manufacturing processes of the semiconductor device. It functions as a selection instruction means.

  The instruction may be given from an input device included in the image server 231 or may be given from another information processing device such as the PC 241-1 via a network. Further, the instruction is given as, for example, a combination of instructions for turning on or off each check box in FIG. 11 and a click operation of the overlay button 510 in FIG.

  Furthermore, for example, the image server 231 is based on the defect position information input as information indicating the position of the defect in each of the two or more inspection images selected by the instruction received by the selection instruction unit. Combining means for associating the defects in the two or more selected inspection images with each other and generating a composite image that is a two-dimensional image from the two or more selected inspection images. Also works.

Of course, the PCs 241-1 to 241-m or the client PC 250 can also function as the combining unit.
The defect position information may be generated by, for example, the substrate inspection apparatuses 211-1 to 211-n, stored in the inspection DB 221, and input from the inspection DB 221 to the image server 231 as the synthesis unit. Alternatively, the defect position information may be input as information indicating the position on the inspection image via the image server 231 or an input device included in another information processing apparatus connected via a network. Alternatively, the image server 231 executes image processing for defect detection again based on the inspection image and calculates the position of the defect, whereby the defect position information is stored in the image server 231 itself as the synthesis unit. May be given.

For example, the image server 231 as the synthesizing unit generates a synthesized image as illustrated in FIG. 1, FIG. 14, or FIG. 16, or a synthesized image (not shown).
Information processing apparatuses such as the image server 231, the PCs 241-1 to 241-m, and the client PC 250 also function as a display unit that displays the generated composite image. Specifically, the function as the display means is realized by a display device included in the information processing apparatus.

  In the substrate inspection system 200 of FIG. 2, the substrate inspection devices 211-1 to 211-n obtain the inspection image by imaging the front surface, the back surface, and the bevel of the substrate, respectively, and the inspection It is an example of the defect inspection apparatus which detects the position of a defect based on an image, and produces | generates the said defect position information. The inspection DB 221 is an example of a storage device that stores the plurality of inspection images obtained by the defect inspection apparatus.

  The defect position information is input from the substrate inspection apparatuses 211-1 to 211-n as the defect inspection apparatus to an information processing apparatus such as the image server 231 as the defect association apparatus. Then, the image server 231 or the like as the synthesizing unit acquires the selected two or more inspection images from the inspection DB 221 as the storage device and generates the composite image.

100, 600, 600b Composite image 101, 300 Wafer 102, 303 Notch 103a-103d, 401a-401d, 607, 616, 620 Surface defect 104, 411 Back surface defect 105a-105f Projection point 106a-106f, 610-614, 617, 618 Line segment 107 Die 108 Shot area 200 Substrate inspection system 210 Substrate inspection device group 211-1 to 211-n Substrate inspection device 212-1 to 212-n Planar inspection unit 213-1 to 213-n Bevel inspection unit 220 DB group 221 Inspection DB
222 Recipe DB
230 server group 231 image server 232 recipe server 240 image PC group 241-1 to 241-m PC
250 client PC
301 Front 302 Back 304 Bevel 305 Near Edge Front
306 Upper Bevel
307 Apex
308 Lower Bevel
309 Near Edge Back
310-314 Imaging optical axis 400 Front image 410 Back image 420, 420-1 to 420-25 Bevel image 421a to 421f, 608, 609, 615, 619 Bevel defect 500 Screen 501 Wafer information display area 502 Inspection number selector 503 Save button 504 Histogram display area 505 Defect display area 506 Bevel defect summary display area 507 Surface defect summary display area 508 Back defect summary display area 509 Same chip defect information display area 510 Overlay button 511 Surface inspection information display area 512 Back surface inspection information display area 513 Bevel Inspection information display area 514 Composite image display area 601 Wafer surface area 602 Near edge front area 603 Upper bevel area 604 Apex area 605 Lower bev l area 606 Near Edge Back area

Claims (10)

Selection instruction means for receiving an instruction to select at least two of a plurality of inspection images obtained by imaging at least two of the front surface, back surface, and bevel of the substrate in one or more manufacturing steps of the semiconductor device When,
Based on the defect position information input as information indicating the position of the defect in each of the two or more inspection images selected by the instruction received by the selection instruction means, the two or more selected sheets Combining means for associating the defects appearing in each inspection image, and generating a composite image that is one two-dimensional image from the two or more selected inspection images;
Display means for displaying the composite image generated by the composite means;
A defect associating apparatus comprising: In the first case, the two or more selected inspection images include a surface image obtained by imaging the surface of the substrate and a bevel image obtained by imaging the bevel of the substrate. The synthesizing means is defined at a projection point or a projection region obtained by projecting the position or region of the first defect reflected in the bevel image on the same plane as the surface of the substrate, or the surface of the substrate. Generating the composite image by superimposing a line segment from the center point of the substrate toward the projection point or the projection area on the surface image;
The defect association apparatus according to claim 1, wherein: In the first case, the synthesizing means,
Determining whether the position of the second defect in the surface image and the position of the projection point or the projection area are close based on a predetermined criterion;
When it is determined that the position of the second defect is close to the position of the projection point or the projection area, the projection point, the projection area, or the line segment is the same as the second defect in the composite image. Represented by color,
The defect associating apparatus according to claim 2, wherein: The synthesis means includes
On the surface of the substrate, a second line connecting the center point and the direction of the first line connecting the projection point or the projection area and the center point and the position of the second defect. When the direction is included within the specified range of the angle specified with respect to the center point, or the angle difference between the direction of the first line and the second line is equal to or less than a threshold value,
Determining that the position of the second defect is close to the position of the projection point or the projection area;
The defect associating apparatus according to claim 3. In the first case, the synthesizing unit has a second defect reflected in the bevel image, depending on which of the plurality of parts of the bevel classified by the position in the thickness direction of the substrate is present. Changing the display form of the line segment to form the line segment;
5. The defect associating apparatus according to claim 2, wherein   The defect association apparatus according to claim 5, wherein the display form of the line segment is based on one or more combinations of a color, a length, a thickness, or a broken line pattern. The display means further displays inspection result information representing a result of performing defect extraction processing based on each of the plurality of inspection images together with a graphical user interface for selecting the plurality of inspection images. ,
The defect associating apparatus according to any one of claims 1 to 6, wherein: In the second case, the two or more selected inspection images include a front image obtained by imaging the front surface of the substrate and a back image obtained by imaging the back surface of the substrate. The synthesizing unit projects and superimposes a defect shown in the back image on the front image,
The defect associating apparatus according to claim 1, wherein The defect association apparatus according to any one of claims 1 to 8,
A defect inspection apparatus that obtains the inspection image by imaging the front surface, the back surface, and the bevel of the substrate, and detects the position of the defect based on the inspection image to generate the defect position information;
A storage device for storing the plurality of inspection images obtained by the defect inspection device;
The defect position information is input from the defect inspection apparatus to the defect association apparatus,
The synthesizing unit obtains the selected two or more inspection images from the storage device and generates the synthesized image;
A board inspection system characterized by that. Information processing device
Receiving an instruction to select at least two of a plurality of inspection images obtained by imaging at least two of the front surface, the back surface, and the bevel of the substrate in one or more manufacturing steps of the semiconductor device;
Based on the defect position information input as information indicating the position of the defect shown in each of the two or more inspection images selected by the received instruction, each of the two or more selected inspection images is shown. Generating a composite image that is a two-dimensional image from the two or more selected inspection images,
Displaying the generated composite image;
A defect association method characterized by the above.