JPH10340935A - Checking apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a checking apparatus which automatically checks defects without fail and with efficiency. SOLUTION: A wafer 2 which has defects and for which the approximate positions of the defects which have been grasped by a preliminary checking machine 8, is placed on a stage 1. A learning unit 11c commands a stage controller 6 to have the wafer 2 moved, so that the defect is brought into focus by a camera 3 for picking up the defects. The move and the magnification increase of the camera 3 are performed alternately to correct the position of the defect with accuracy. With respect to the defects whose positions have been corrected with high accuracy, analysis of components is conducted by an analyzer 12, and laser works are conducted by a processor 16. The images obtained by the camera 3, the results of the component analysis and the effects, if any of the laser processes, are correspondingly learned by the learning unit 11c. Based on the results of the learning, the learning unit 11c determines whether to perform the component analysis and the laser works with respect to new defect that exists on the wafer 2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for inspecting a defect on a semiconductor substrate.

[0002]

2. Description of the Related Art Before inspecting a defect in a semiconductor device or a display substrate, it is necessary to know a position where the defect exists. Defect inspection is performed by performing observation, analysis, cross section processing, or the like based on this position.

The position of a defect is searched for by a conventional inspection device and output as information. When processing a cross section or the like, information on the position of a defect must be accurate. However, information obtained by the conventional inspection apparatus lacks accuracy. Therefore, it was necessary to manually find the position of the actual defect around the position indicated by the information.

[0004]

Searching for the position of a defect manually is a very laborious and time-consuming operation, and has the problem that the inspection efficiency is reduced. In addition, the number of products that can be inspected is limited due to poor inspection efficiency. As a result, only products that are particularly important, such as development prototypes or manufacturing jigs, can be inspected, and defects cannot be inspected for products mass-produced for sale or use. There is a problem of being restricted.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to provide an inspection apparatus for automatically, reliably and efficiently inspecting for defects.

[0006]

According to a first aspect of the present invention, there is provided an inspection apparatus for extracting a position of a defect of a semiconductor substrate to be inspected, an observation unit for enlarging and observing the defect, Control means for performing, in parallel with the enlargement, a housing operation for storing the extracted position in the observation field of view of the observation means by relative movement between the semiconductor substrate and the observation means; And learning means for instructing the control means to reproduce the learned relative movement for a new defect.

According to a second aspect of the present invention, in the inspection apparatus according to the first aspect, the learning is performed by storing a direction and an amount of the relative movement.

According to a third aspect of the present invention, in the inspection apparatus according to the second aspect, the orientation and the amount are continuously stored even when the power of the inspection apparatus is turned off. .

[0009] The inspection apparatus according to the fourth aspect is characterized in that:
4. The inspection apparatus according to claim 2, further comprising an analysis unit that analyzes a component of the defect to obtain an analysis result, wherein the observation unit captures image information of the defect, and the learning unit. Determines the type of the new defect that provides new image information based on learning by associating the analysis result with the image information, and determines whether the analysis unit operates on the new defect. To determine.

The inspection apparatus according to a fifth aspect is the inspection apparatus according to the fourth aspect, further comprising repair means for repairing the new defect, wherein the learning means is determined by the association. It is determined whether or not the repair means operates according to the type.

According to a sixth aspect of the present invention, in the inspection apparatus of the fifth aspect, the semiconductor substrate is a display substrate.

According to a seventh aspect of the present invention, in the inspection apparatus according to the first aspect, the extracting means and the observing means are separate from each other, and the semiconductor substrate is a counterpart of the semiconductor substrate. 1st and 2nd which become the reference of the general position
A stage whose movement is controlled by the control means after the first and second positions, which are the respective positions of the pointers, are extracted by the extraction means, and the first and second positions are extracted by the extraction means. Mounted on the stage, the observation unit extracts the first and second positions with respect to the semiconductor substrate mounted on the stage, and the learning unit corrects the semiconductor substrate on the stage. The virtual first and second positions when mounted, and the first and second positions extracted by the observation means.
By comparing the positions with each other, a shift in the positional relationship between the semiconductor substrate and the stage is recognized, and the control unit is instructed to move the stage in which the shift is eliminated.

An inspection apparatus according to an eighth aspect of the present invention is the inspection apparatus according to the first aspect, wherein the learning unit detects that the defect cannot be observed in the observation visual field. The observation means is instructed to reduce the magnification, and the control means is instructed that a pointer serving as a reference for a relative position on the semiconductor substrate is stored in the observation field of view after the reduction, Extracts a reference position that is the position of the pointer stored in the observation field of view, and the learning unit sets a relative position between the position of the pointer and the position of the defect obtained by the extraction unit. The control unit is instructed to move the position by the reference position starting from the reference position.

According to a ninth aspect of the present invention, in the inspection apparatus according to the eighth aspect, the learning means is configured to extract the reference position after the reference position is extracted, and to set the relative position starting from the reference position. Before instructing to perform a movement of a certain position, the observation means is instructed to increase the magnification.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION

Embodiment 1 FIG. In the present embodiment, a configuration of an inspection apparatus for automatically inspecting for defects is shown. FIG. 1 is a block diagram illustrating a configuration of an inspection device according to the present embodiment.
The inspection apparatus of the present embodiment is an apparatus that inspects the presence or absence of a superficial or internal defect of the wafer 2 using scattered light. FIG. 2 is a flowchart sequentially illustrating processing operations of the inspection device of the present embodiment. The defect inspection is started in step S1.

Next, in step S2, coordinates representing the position of a defect on the wafer 2 as an object to be inspected are extracted. As illustrated in FIG. 1, the inspection apparatus according to the present embodiment includes a preliminary inspection machine 8 that extracts coordinates relating to the position of a defect on the wafer 2. First, the inspector himself has the wafer 2
To the preliminary inspection machine 8. The preliminary inspection machine 8 has a well-known configuration, and automatically searches for a defect existing in the brought wafer 2 and extracts the coordinates of the defect. Even when there are a plurality of defects, coordinates are extracted for all the defects.

The extracted coordinates are transmitted to the learning unit 11 via the coordinate signal line 9 and are stored in a storage circuit (not shown) provided in the learning unit 11. After the coordinate extraction by the preliminary inspection device 8 is completed, the inspector moves the wafer 2 from the preliminary inspection device 8 and the XY stage 1
Place on top. The XY stage 1 includes a stage control unit 6
The movement is controlled in the plane direction including the X direction and the Y direction orthogonal to each other in the plane on which the wafer 2 is placed, and the camera approaches the camera 3 for observing the wafer 2 orthogonal to the plane direction. Movement in the Z direction, which is the vertical direction in which the XY stage moves up and down, movement that rotates around the Z axis, and movement that changes the angle between the XY stage and the Z axis (movement that tilts the XY stage) Is possible.

Next, in step S3 in FIG. 2, the learning unit 11 reads one of the coordinates of the defect stored in its own storage circuit. In a succeeding step S4, the learning unit 11 controls the stage control unit via the moving signal line 10b so that the position represented by the read coordinates coincides with the focus of the camera 3 that observes a defect on the wafer 2. 6 is instructed to move the XY stage 1. The position of the XY stage 1 after the movement is determined by the learning unit 11.
Are stored in the storage circuit provided in the storage device. The focal point of the camera 3 is located at the center of the field of view of the camera 3 in the plane direction of the XY stage 1 (the direction including the X direction and the Y direction).

By the above movement, the focal point of the camera 3 and the position where the defect exists should coincide with each other. However, the accuracy of the coordinates extracted by the preliminary inspection machine 8 is not so high that the position of the defect and the focus completely match by the movement in step S4. In addition, there is an inherent deviation between the preliminary inspection machine 8 and the stage control unit 6. For these reasons, the focal point and the actual location of the defect will inevitably deviate. Therefore, if the camera 3 is set to a high magnification in order to observe the defect immediately after the movement of the XY stage 1 in step S4, the defect will be out of the field of view of the camera 3 due to the displacement.

Therefore, in order to correct the deviation, the camera 3 takes an image of a defect in the subsequent step S5. The magnification of the camera 3 at the time of imaging is set low enough that a defect shifted by the expected deviation from the center of the field of view enters the field of view in order to ensure that the defect is within the field of view.

The image obtained by the imaging is sent to the image processing section 4 via the image signal line 5 in FIG. In the image processing unit 4, as the operation in step S6, the received image is processed and a defect is extracted by separating the image from the background. Then, with respect to the extracted defect, information such as the coordinates, silhouette, size, surface roughness, and contrast with respect to the background of the defect is collected. The coordinates of the defect are sent to the learning unit 11 via the coordinate signal line 10a.

The accuracy of the camera 3 is higher than that of the preliminary inspection machine 8, and the coordinates of the defect obtained in step S6 are more accurate than the coordinates obtained in step S2. It should be noted that the information such as the silhouette of the defect obtained in step S6 is insufficient for grasping the type of the defect due to the insufficient magnification of the camera 3 and the mismatch between the focus of the camera 3 and the position of the defect.

Therefore, the learning unit 11 having inputted the coordinates of the defect so that the coordinates of the defect obtained in the step S6 and the focus of the camera 3 coincide with each other in order to image the defect at a high magnification, performs the steps S4 and S4. It instructs the stage control unit 6 to perform the same movement of the XY stage 1 again. Due to the highly accurate correction of the position of the defect achieved by this movement (step S7), the defect occurs when the magnification of the camera 3 is raised, which is caused by an inherent deviation between the camera 3 and the stage control unit 6. The problem of being out of view is eliminated.

Next, in step S8, it is determined whether or not the magnification is sufficiently high. If "NO" is determined, the magnification of the camera 3 is increased in step S9, and thereafter, the process returns to step S5. With such a configuration, the magnification increase in step S9 and the highly accurate correction of the defect position in step S7 are alternately repeated until the magnification becomes sufficiently high.
Therefore, the defect does not fall out of the field of view before the magnification at which the information on the defect can be obtained in sufficient detail.

If it is determined in step S8 that the magnification is sufficiently high and "YES" is determined, the process proceeds to step S10.
Move to In step S10, it is determined whether a flag is set. The setting of the flag will be described in step S15, but the flag is not set at this time, and it is determined as “NO”.

Next, in step S11, the processing is repeated a plurality of times in step S7 in response to the processing being repeated a plurality of times in steps S5 to S9.
The integration of the movement of the XY stage 1 for performing high-precision correction is performed. The integration of the movement is performed by obtaining the difference in coordinates between the position of the XY stage 1 after the movement in step S4 and the position of the XY stage 1 at the time of step S11 by subtraction. At the time of integration, subtraction is performed for each of the three component coordinates (X coordinate, Y coordinate, and Z coordinate).

The movement of the XY stage 1 performed a plurality of times for high-precision correction in step S7 reduces the inherent deviation between each of the preliminary inspection machine 8 and the camera 3 and the stage control unit 6. The correction is performed by repeating the increase of the magnification in S9 and the highly accurate correction in step S7. Therefore, the integrated value obtained by the integration in step S11 means that each of the inspection device 8 and the camera 3 and the stage control unit 6
Represents the inherent deviation itself from. The inherent deviation learned as an integral value is held in a storage circuit provided in the learning unit 11.

Next, in step S12, a defect is observed. Observation of a defect is performed based on information such as coordinates, a silhouette, a size, a surface roughness, and a contrast with respect to the background of the defect, which are collected by the image processing unit 4 illustrated in FIG. A monitor (not shown) is connected to the image processing unit 4 so that the inspector can check defect information on the monitor.

Next, in step S13, it is determined whether or not the defect not observed in step S12 remains in the image processing unit 4, that is, whether or not there is a defect to be observed. When it is determined that there is no defect and “NO”, the inspection of the defect is ended in step S14. If "YES" is determined, the process proceeds to step S15.
Move to

In step S15, step S1
At 2, a flag is set as a signal that the defect has been observed at least once. The flag is set to indicate that the inherent deviation between each of the inspection machine 8 and the camera 3 and the stage control unit 6 has been learned by integrating the movement in step S11.

In order to observe other defects that have not been observed, the magnification of the camera 3 illustrated in FIG. 1 must be reduced. Therefore, the magnification is reduced in step S16. However, the magnification after the reduction may be higher than the magnification when the coordinates of the defect are read in step S3. The reason why the magnification can be set relatively high will be described later.

Next, the same reading as the reading of the coordinates of the defect in step S3 is performed in step S3a for the defect that has not been observed. Subsequently, in step S4a, the same movement as the movement of the XY stage 1 in step S4 is performed on the defect whose coordinates have been newly read. However, although similar,
The movement of the XY stage 1 at this time is a movement in which the read coordinates are located at the focal point of the camera 3 and further a step S1.
1 is the sum of the integral values of the movement obtained in step 1. Such a movement is performed by the learning unit 11 by the stage control unit 6.
This is done by giving instructions.

As described above, the integral value of the movement is determined by the inspection machine 8.
And the inherent deviation between each of the cameras 3 and the stage control unit 6. Therefore, by moving the XY stage 1 by adding the integral value obtained in step S11, the correction for the inherent deviation already performed is reproduced. As a result, in parallel with the movement for positioning the coordinates at the focal point, the correction for the inherent deviation is performed, and a new defect is located at the focal point of the camera 3 in a state where the deviation has been corrected. Become.

By correcting the inherent deviation in advance by the addition as described above, the high-precision multiple steps S7 must be performed in parallel with the repetition of the magnification increase in the step S9 from the low magnification. Correction can be omitted. Therefore, step S
The time required for defect inspection is shortened by the amount of time the processing is repeated a plurality of times in Step 5 to Step S9. Further, even if the magnification after the reduction in step S16 is set to be higher than the magnification in step S3, the defect does not go out of the field of view of the camera 3 because the inherent deviation is corrected in advance. .

In the inspection apparatus of this embodiment, the inherent deviation between each of the preliminary inspection machine 8 and the camera 3 and the stage control unit 6 is stored and read out as a difference between coordinates before and after the movement, which is an integrated value of the movement. You. A set of coordinate differences for each component is a vector, which represents the direction and amount of movement. By recognizing the inherent deviation as a value representing the vector, the reproduction of the correction is performed accurately and reliably.

After step S4a, the process proceeds to step S5, and the processes of steps S5 to S9 are repeated until it is determined that the magnification of the camera 3 is sufficiently high (step S8). After it is determined in step S8 that the magnification is sufficiently high, it is determined in step S10 whether a flag is set. If the flag is set, the inherent deviation has already been learned by the integration of the movement in step S11, and the process skips step S11 and proceeds to step S12. If it is determined in step S13 that there is a defect to be observed after the defect has been observed in step S12, the processing from step S15 is continued.

Also in this case, for the above-mentioned reason, the processing can be started from a magnification higher than the magnification in step S3, and further, the integration of the movement in step S11 can be prevented from being performed again. Therefore, the inspection of the defect is performed quickly.

On the other hand, if it is determined in step S13 that all defects have been observed and there is no defect to be observed, the inspection is terminated in step S14.

In the above description, the XY stage 1
The movement of which is controlled by the stage control unit 6 is used. However, a configuration in which the camera 3 is moved with respect to the wafer 2 on the XY stage 1 may be adopted.

The inherent deviation between each of the preliminary inspection machine 8 and the camera 3 and the stage control unit 6 is determined for each of these combinations. Therefore, even when the power of the inspection apparatus of the present embodiment is turned off, the storage circuit provided in the learning unit 11 stores the integrated value obtained in step S11 and the state in which the flag is set in step S15. You can leave it as it is. With such a configuration, it is possible to perform a defect inspection promptly after the power is newly turned on, by making use of the inherent deviation learned before the power is turned off.

In this case, step S17 may be added to the processing operation illustrated in the flowchart of FIG. 2 as illustrated in FIG. Step S17 is processing performed subsequent to step S2. In step S17, it is determined whether or not a flag is set, as in step S10. If the flag is not set, "NO" is determined, and the routine goes to Step S3. In this case, processing after step S3 is performed,
By performing the integration of the movement in step S11, the inherent deviation is grasped.

On the other hand, in step S17, the learning unit 1
"YE" based on the flag held in the storage circuit No. 1
If it is determined to be S ", the process proceeds to step S3a. In this case, the inherent deviation learned before the power is turned off is utilized, and the process is started from a relatively high magnification.

Note that an optical microscope can be used as the camera 3. However, it is also possible to use a laser microscope, an electron microscope or an AFM (atomic force microscope) instead of the camera 3. As the image, a dark field image or a bright field image may be taken.

Embodiment 2 Hereinafter, the same components as those already described are denoted by the same reference numerals, and description thereof will be omitted. In the present embodiment, there is shown an inspection apparatus that includes an analysis unit that analyzes a component of a defect and that quickly observes a defect by associating an analysis result with image information.

FIG. 4 is a block diagram illustrating a configuration of an inspection apparatus according to the present embodiment. In the inspection device shown in FIG. 1, the learning unit 11 provided in the inspection device of the first embodiment illustrated in FIG. 1 is replaced with a learning unit 11a, and a component analysis unit 12 is further attached. The configuration of the other parts is the same. The learning unit 11a has the same functions as those of the learning unit 11 according to the first embodiment, but further includes the functions of the present embodiment.

FIG. 5 is a flowchart illustrating a processing operation performed by the inspection apparatus according to the present embodiment. Steps S18 to S20 shown in the figure are processes inserted between step S8 and step S10 illustrated in FIG. Hereinafter, differences from the first embodiment will be described.

After it is determined in step S8 that the magnification is sufficient, the learning unit 11a in step S18 obtains information such as the silhouette, size, surface roughness, and contrast with the background of the defect collected in step S6 in FIG. Is known, that is, whether or not the defect is known. The determination as to whether the defect is known is made by the learning unit 11a based on whether or not the result of the association in step S20 described later is stored. The determination as to whether or not the defect is known is based on whether or not the type of the defect is known.

If it is determined in step S18 that the defect is not a known defect ("NO"), step S19
Move to In step S19, the component of the defect is analyzed using the component analyzer 12 illustrated in FIG. The configuration of the component analyzer 12 is well known, and for example, an instrument for performing energy dispersive X-ray (EDX) analysis can be used. The result of the component analysis is represented by an analysis result signal line 10c.
Is provided to the learning unit 11a via the.

The learning unit 11a associates the result of the component analysis received in step S20 with the information on the defect and the like regarding the silhouette and the like collected in step S6 in FIG. 2 in FIG. In other words, it learns what kind of component a defect that gives certain information has.
The result of the learning (association) performed in this way is stored in the learning unit 11a and accumulated. When the process in step S20 ends, the process proceeds to step S10 illustrated in FIG. 2, and the above-described process is performed thereafter.
In step S12, the information of the defect and the result of the component analysis are displayed on a monitor not shown in FIG.

Next, it is determined that there is a defect to be observed in step S13 illustrated in FIG.
A case where the subsequent processing is performed on a new defect will be described. This new defect is identified in step S20 of FIG.
It is assumed that the analysis result and the defect information have already been associated with each other, and that the defect is a known defect whose component can be grasped from information such as a silhouette.

When the process proceeds to step S18 in FIG. 5 after the processes in steps S15 to S8 in FIG. 2 are performed, it is determined that the new defect is a known defect that has been learned by the learning unit 11a (“YES”). Is determined. There is no need to learn components and information for learned defects. Therefore, the processes in steps S19 and S20, which are unnecessary because they have been learned, are skipped, and the process immediately proceeds to step S10 in FIG. In step S12, the information of the defect and the result of the component analysis that are associated with the information of the defect are displayed on a monitor not shown in FIG.

By the processing operation as described above, it is possible to observe a defect for a learned defect without performing component analysis and association performed in steps S19 and S20, respectively. by this,
Inspection of defects is performed quickly. Further, since the component analysis is not performed on the learned defect, it is possible to prevent the wafer 2 from being damaged by the energy dispersive X-ray analysis.

In the conventional inspection apparatus, a configuration is adopted in which an inspector previously learns the association by a learning unit before the inspection. However, this conventional inspection apparatus is effective only for the correspondence learned in advance, and for the combination of the result of the component analysis and the information that the inspector did not learn, effective information regarding the defect is provided. Can not provide.

On the other hand, in the inspection apparatus of the present embodiment, component analysis and association are performed for unlearned defects, and the results are stored. That is, the ability of the learning unit 11a improves as the inspection device is used. In addition, there is no need for the inspector to perform learning in the learning section in advance, and the burden on the inspector is reduced.

Embodiment 3 In the present embodiment,
1 shows an inspection apparatus having a function of repairing an inspected defect. FIG. 6 is a block diagram illustrating a configuration of the inspection apparatus according to the present embodiment. In the inspection device shown in FIG. 3, the optical camera 3 provided in the inspection device of the first embodiment illustrated in FIG. 1 is replaced with an electron microscope 14. The learning unit 11 in FIG. 1 is replaced with a learning unit 11b, and a focused ion beam (FIB) optical system 15 is further attached. The configuration of the other parts is the same. The learning unit 11b according to the first embodiment
The function of the present embodiment is added to the function of the learning unit 11 as it is.

FIG. 7 is a flowchart illustrating a process performed by the inspection apparatus according to the present embodiment. Steps S21 to S6a shown in the figure are processes inserted between step S10 or step S11 and step S12 of the first embodiment illustrated in FIG. Hereinafter, differences from the first embodiment will be described.

In step S5 in FIG. 2, the image of the defect is taken by the electron microscope 14 in FIG.
In, the image processing unit 4 extracts a defect from data obtained by imaging. Then, after the processing of step S10 or step S11 exemplified in FIGS. 2 and 7, the processing of the present embodiment is performed.

In step S21, the repair of the defect is performed by the focused ion beam optical system 15 in FIG. 6 based on the information on the defect extracted in step S6 in FIG. Specifically, the learning unit 11b instructs the focused ion beam optical system 15 to irradiate the focused ion beam to the defect whose position has been corrected with high accuracy in step S7 of FIG. Irradiation removes foreign matter, cuts short-circuited portions of wiring, or re-forms lost portions of wiring. Further, the present invention is not limited to such processing, and switching to a spare circuit may be performed while avoiding a portion where a defect occurs.

The inspection apparatus of the present embodiment is similar to that of the first embodiment.
With the same configuration as described above, the position of the defect is accurately grasped, and thereafter the defect is repaired. Therefore, the focused ion beam does not deviate from the defect, and the repair is performed accurately. Also,
It is possible to repair only a portion having a defect which could only be performed by a manual instruction in the past. After the defect has been repaired, the process proceeds to step S5a.

In steps S5a and S6a, as in steps S5 and S6 in FIG.
The repaired defect is imaged, and the defect is extracted based on the information obtained thereby. Continue with step S12
In, the defect is observed. At this time, what is displayed on the monitor is the data on the repaired defect.
The inspector can confirm that the defect has been properly repaired.

Embodiment 4 In the present embodiment,
A configuration in which the respective processes of the second and third embodiments are combined will be described. FIG. 8 is a block diagram illustrating a configuration of the inspection device according to the present embodiment.
The inspection apparatus shown in the figure differs from the inspection apparatus of the third embodiment in that a camera 3 is used as a means for imaging a defect and a laser processing machine 16 is used as a means for repairing the defect. . However, there is no difference in that only the equipment used is different, and it fulfills the function of imaging and the function of repairing. A learning unit 11c having the functions of the learning units 11a and 11b according to the second and third embodiments is provided instead.

FIG. 9 is a flowchart illustrating the processing of the inspection apparatus according to the present embodiment. The processing shown in the figure is a combination of the processing of the second embodiment and the processing of the third embodiment illustrated in FIGS. 5 and 7, respectively, to which the processing of the present embodiment is added. The processing of the present embodiment
Before step S21 and step S6a of the third embodiment
After each is inserted. Depending on the inserted process,
The combination of the processes of the second and third embodiments is effectively performed. This will be described below.

When the processing according to the second embodiment illustrated in FIG. 9 is completed, it is determined that the defect is a known defect in step S18, or the component analysis is performed in step S19. The component of the defective defect is known.

In step S22 of the present embodiment inserted after step S10 or step S11,
Determine the components of the defect. Understanding the components means step S
If determined to be a known defect at 18,
The component of the defect corresponding to the image information of the defect is represented by the learning unit 1 in FIG.
1c is to be read from the storage circuit provided in 1c. on the other hand,
If the component analysis is performed in step S19 as not a known defect, the result of the component analysis is to be kept.

The laser beam machine 16 shown in FIG. 8 provided in the inspection apparatus according to the present embodiment uses laser light emitted from the laser beam machine 16 in the same manner as the focused ion beam optical system 15 according to the third embodiment illustrated in FIG. Perform repairs. However, depending on the component of the defect, it may be impossible or unsuitable to cause the laser beam to evaporate or the like. In this case, even if the defect is irradiated with laser light using the laser processing machine 16, the trouble of irradiation and the time required for the irradiation are wasted.

In step S23 following step S22, it is determined whether the defect can be repaired. The determination is made based on the component of the defect ascertained in step S22. If it is determined that the defect component can be evaporated or the like by the laser beam and can be repaired ("YES"), the process proceeds to step S21. Also, if it is not possible to judge whether the component can be repaired, that is, if it is unknown whether the component is repairable or impossible, it is determined to be “YES” for the time being. After step S21, the processing of the third embodiment, that is, the repair of the defect and the imaging of the defect thereafter are performed.

When the repair in step S21 is performed effectively, that is, when the defect evaporates due to the laser beam, the silhouette etc. of the defect extracted in step S6a are obtained before and after the repair in step S21. Change. On the other hand, when the evaporation or the like by the laser beam does not occur and the effect of the repair is not obtained, the silhouette of the defect does not change before and after the repair in step S21. Therefore, based on the presence or absence of a change in the silhouette or the like of the defect extracted in step S6a,
It is possible to determine whether or not the repair is effective.

Step S24 for performing the processing of this embodiment
In, the learning unit 11c automatically determines whether or not there is a repair effect based on whether or not there is a change in the information obtained in step S6a. The learning unit further associates the determination result with the defect component, and stores the association result in a storage circuit provided therein. The defect component refers to the component identified in step S22, and by this association,
If the components are grasped, the presence or absence of the effect of the repair will be automatically known.

The association performed in step S24 is used for a new defect that causes “YES” to be determined in step S13 illustrated in FIGS. The use will be described in detail. When it is determined whether the new defect can be repaired in step S23 after the components are grasped in step S22 in FIG. 9, this association is stored. That is, it is read out from the circuit and is used as a material for determining whether repair is possible.

If it is determined in step S23 that the defect component is unsuitable for laser beam repair based on the association and the determination is "NO", step S21 is performed.
Steps S24 to S24 in which the process of the third embodiment for repairing a defect and the like and the process of the present embodiment are skipped.
Move to 2. In this manner, it is possible to prevent the repair or the like from being performed on a defect that is known in advance as having no repair effect, thereby shortening the inspection time. If "YES" is determined in the step S23, the processes after the step S21 are performed.

In the above description, the laser processing machine 1
6 is used to repair a defect. But,
It is of course possible to repair the defect using another configuration. For example, the foreign matter can be removed by using a mechanism for adsorbing or holding the foreign matter, a nozzle for ejecting a gas, a liquid, or a solid toward the foreign matter, a mechanism for irradiating the foreign matter with an electron beam, or the like. These are related to repair for removing foreign substances. For example, repair can be performed by forming a film using a mechanism using energy by irradiation with a laser beam or a microwave in a gas atmosphere.

Embodiment 5 FIG. 10 is a block diagram illustrating a configuration of the inspection apparatus according to the present embodiment. The inspection apparatus shown in FIG. 1 is a first embodiment illustrated in FIG.
The preliminary inspection machine 8 and the coordinate signal line 9 provided in the inspection apparatus of FIG. 1 have been removed, and the monitor connected to the image processing unit 4 is explicitly shown as a monitor 17. Portions other than those relating to this remain unchanged.

The preliminary inspection machine 8 illustrated in FIG. 1 is provided in the inspection machine for extracting the coordinates of the defect in step S2 illustrated in FIG. With the removal of the preliminary inspection machine 8, in the present embodiment, the inspector must find the position of the defect on the wafer 2.

FIG. 11 is a flowchart illustrating an inspection procedure performed by using the inspection apparatus of the present embodiment. In the inspection procedure of the present embodiment, among the processes illustrated in FIG. 2, step S2 of FIG. 2 performed by the preliminary inspection machine 8 of the first embodiment is omitted, and instead, the steps of the present embodiment are performed. S2a is inserted. In step S2a, it is the inspector, the camera 3, and the image processing unit 4 that are involved in extracting and holding the coordinates of the defect.
Hereinafter, portions different from the first embodiment will be described.

In step S2a, the examiner checks the state shown in FIG.
After the wafer 2 is placed on the XY stage 1 illustrated in FIG. 1, the appearance of a defect is waited while the image obtained by the camera 3 and displayed on the monitor 17 is visually observed. While the inspector looks at the monitor 17, the stage control unit 6 moves the XY stage 1 in the X direction or the Y direction, and the surface of the wafer 2 is scanned by the camera 3 by this movement. The camera 3 sequentially captures the surface of the wafer 2 and transmits this image to the image processing unit 4 via the image signal line 5. The image processing unit 4 gives an image from the camera 3 to the monitor 17 via the image signal line 5a. The state of the surface of the wafer 2 displayed on the monitor 17 is constantly updated so that the portion of the wafer 2 captured by the camera 3 does not differ from the portion displayed on the monitor 17.

The inspector issues a command to the image processing section 4 when a defect is displayed on the monitor 17. Upon receiving this command, the image processing unit 4 processes the image at the time the command is issued, extracts a defect from the background, and specifies its coordinates. The coordinates extracted in this way are the coordinate signal line 1
0a is provided to the learning unit 11 and is stored in a storage circuit provided therein.

The extraction and holding of the coordinates as described above are continued until the scanning of the surface of the wafer 2 is completed.
In order to reduce the time required for scanning, it is preferable that the magnification of the camera 3 is set low enough to indicate the presence or absence of a defect. After the process in step S2a ends, the process proceeds to step S3 which is not different from the first embodiment.

In the present embodiment, it is possible to extract and hold the coordinates of a defect using the camera 3 and the monitor 17 without preparing the preliminary inspection machine 8 of the first embodiment illustrated in FIG. Illustrated. Whether or not to use the preliminary inspection machine 8,
Whether the camera 3 or the monitor 17 is used, the result that the coordinates of the defect are held in the learning unit 11 remains the same, and the processing after step S3 is performed as it is in the first embodiment. Although the monitor 17 is used in the above example, an eyepiece can be used instead.

Embodiment 6 FIG. FIG. 12 is a block diagram illustrating a configuration of the inspection apparatus according to the present embodiment. The inspection apparatus shown in the figure is a second embodiment illustrated in FIG.
The monitor 17 connected to the image processing unit 4 is shown as a monitor 17 from which the preliminary inspection machine 8 and the coordinate signal line 9 provided in the inspection apparatus of FIG. Portions other than the portion related to this change remain unchanged. That is, this change
This is the same as the change of the inspection apparatus of the first embodiment illustrated in FIG. 1 to the inspection apparatus of the fifth embodiment illustrated in FIG.

FIG. 13 is a flowchart illustrating an inspection procedure performed using the inspection apparatus according to the present embodiment. In the inspection procedure illustrated in FIG. 11, steps S18 to S20 of the second embodiment illustrated in FIG. 5 are inserted between steps S8 and S10 of the fifth embodiment illustrated in FIG. It was done. Step S10
Illustrations after that are omitted.

As can be understood from the above configuration, the difference between the configuration of the fifth embodiment and the configuration of the present embodiment is the difference between the configuration of the first embodiment and the configuration of the second embodiment. Is the same as That is, step S2 of the second embodiment illustrated in FIG. 2 is simply replaced with step S2a of the fifth embodiment illustrated in FIG. The effects obtained particularly in the present embodiment with respect to the fifth embodiment are the same as the effects obtained particularly in the second embodiment with respect to the first embodiment.

According to the structure of the present embodiment, similar to the fifth embodiment, the preliminary inspection machine 8 of the second embodiment illustrated in FIG.
It is understood that the effect of Embodiment 2 can be obtained even if the coordinates of the defect are extracted and held using the camera 3 and the monitor 17 without preparing.

Embodiment 7 FIG. FIG. 14 is a block diagram illustrating a configuration of an inspection device according to the present embodiment. FIG.
Camera 3 provided in the inspection device of Embodiment 4 exemplified in FIG.
And the laser beam machine 16 are replaced by an electron microscope 14 and a focused ion beam optical system 15, respectively.
The function performed is the same, with no substantial changes. Further, the preliminary inspection device 8 and the coordinate signal line 9 illustrated in FIG. 8 are omitted as in the case of the fifth embodiment.
The monitor 17 is illustrated as in the case of the fifth embodiment.

FIGS. 15 and 16 are diagrams integrally showing one flowchart exemplifying a procedure of an inspection performed by using the inspection apparatus of the present embodiment. In both figures, "A" is connected at the portion of the symbol surrounded by a circle. Specifically, subsequent to step S20 illustrated in FIG. 15, the process of step S10 illustrated in FIG. 16 is performed.

The flowcharts illustrated in FIGS. 15 and 16 are the same as those of the fourth embodiment except that step S2 of the fourth embodiment is replaced with step S2b of the present embodiment. Is the same as Step S2b is a step of performing the same processing as step S2a of the fifth embodiment illustrated in FIG. 11, and the function performed by the camera 3 in step S2a of the fifth embodiment is performed in step S2b of the present embodiment. The only difference is that the electron microscope 14 plays instead.

In FIG. 16, the processing after step S13 is the same as in FIG. 1, and is not shown. As is clear from the description of the fifth and sixth embodiments, the effects of the fourth embodiment can be obtained without using the preliminary inspection machine 8.

In FIG. 14, a thin film transistor (TFT) substrate 2a for a display is X
-It is mounted on the Y stage 1. The TFT substrate 2a is inspected according to the inspection procedure illustrated in FIGS. 15 and 16, and repair of the defect is performed with high accuracy in step S21 as in the fourth embodiment.

In general, a substrate for a display needs a large display area. The larger the area of a substrate having a large area, the higher the probability that a defect exists in one substrate. Then, for example, the probability that the display substrate is determined to be defective due to only one defect increases.

However, since the inspection apparatus of this embodiment repairs a defect with high accuracy in step S21, the display substrate is reliably rescued and used effectively.

Embodiment 8 FIG. In the inspection apparatus described in the first to seventh embodiments, problems that cannot be solved remain. In the present embodiment, a configuration of an inspection device that solves such a problem will be described below.

First, the first unsolved problem will be described. When the wafer 2 is moved from the preliminary inspection machine 8 of the first embodiment illustrated in FIG. 1 onto the XY stage 1, the XY stage 1 and the wafer 2 are inevitably mounted. The position shift occurs between them. The learning unit 11 uses the coordinates extracted by the preliminary inspection machine 8 to execute X
Since the -Y stage 1 is moved and the camera 3 takes an image of a defect, there is an inconvenience that the defect does not fit within the field of view of the camera 3 while leaving such an initial deviation. Therefore, it is desirable to eliminate such initial deviation in advance in defect inspection. For this purpose, the present embodiment uses the configuration described below.

FIG. 17 is a block diagram illustrating the configuration of the inspection apparatus according to the present embodiment. In the inspection device illustrated in FIG. 3, the learning unit 11 included in the inspection device of the first embodiment illustrated in FIG. 1 is replaced with a learning unit 11d. The configuration of the other parts is the same. Learning unit 1
1d is a function obtained by adding the function of the present embodiment while keeping the function of the learning unit 11 of the first embodiment.

FIG. 18 is a flowchart illustrating a processing operation performed by the inspection apparatus according to the present embodiment. In the processing operation according to the present embodiment, new processing is added between step S2 and step S3 of the first embodiment illustrated in FIG. For simplicity of illustration, FIG. 18 illustrates a process only between steps S2 and S3. The newly inserted processing is performed as a function unique to the learning unit 11d according to the present embodiment. Hereinafter, differences from the first embodiment will be described.

In step S2, the coordinates of the defect are extracted and held by the preliminary inspection machine 8, and the wafer 2
After being placed on the Y stage 1, in step S25, processing relating to the origin mark is performed.

The origin mark is a point which is a reference for a relative position on the surface of the wafer 2. A plurality of chips arranged on the wafer 2 may have a part having a specific shape or a mark indicating a specific position in the chip, depending on the type of the integrated element. Or The coordinates of the origin mark, which is a specific location serving as a mark of the position in the wafer 2 (chip), are extracted and held during the processing in step S2.

In step S25, step S2
The XY stage 1 is moved based on the coordinates of the first origin mark, which is one of the plurality of origin marks extracted in the above, and the first origin mark is set within the field of view of the camera 3. Such movement of the XY stage 1 for imaging the first origin mark is instructed by the learning unit 11d to the stage control unit 6. Subsequent step S
In 26, the first origin mark is imaged using the camera 3, and the coordinates of the origin mark on the XY stage 1 are extracted and held.

Next, in step S27, based on the coordinates of the second origin mark, which is one of the remaining origin marks, the second origin mark is positioned within the visual field in the same manner as in the case of the first origin mark. The XY stage 1 is moved so as to fall within the range. Then, similarly to step S26, in step S28, imaging of the second origin mark, extraction of coordinates, and holding are performed.

In step S29, step S2
6 and the first extracted in step S28, respectively.
And the coordinates of the second origin mark and the virtual first and second positions when the wafer 2 is placed at the expected position when the wafer 2 is moved from the preliminary inspection machine 8 to the XY stage 1.
Are compared with the coordinates of the origin mark. And
From the result of this comparison, the positional deviation between the actual wafer 2 and the virtual wafer 2 is calculated.

The learning unit 11d instructs the stage control unit 6 to move the XY stage 1 by the amount by which the displacement amount obtained by the calculation is eliminated. The displacement is corrected by such movement of the XY stage 1, and the defect inspection in step S3 and subsequent steps is performed satisfactorily without initial position displacement.

The reason why two origin marks are used will be described. When only one origin mark is used, even if the XY stage 1 is moved so that the actual position of the origin mark coincides with the virtual position, the wafer 2 is centered on the origin mark. Is still allowed. Therefore, by determining the position of the other origin mark, such a deviation of the wafer 2 is completely shut out, and the reliability of the correction is improved.

In the above description, an example is shown in which a mark or the like in a chip is used as an origin mark. However, the above-described correction can be performed by using the coordinates of each of the two defects extracted in step S2 as the origin mark.

Next, a second unsolved problem will be described. The direction and amount of movement of the XY stage 1 instructed from the learning unit 11d to the stage control unit 6 illustrated in FIG. 17 and the direction and amount of movement of the XY stage 1 actually Inevitably, an error occurs. When repeating the movement of the XY stage 1 as high-precision correction in step S7 in order to inspect for defects one after another,
Such errors are accumulated and result in a large deviation, and as a result, a defect that should be contained in the field of view of the camera 3 may be out of the field of view. In this case, the imaging in step S5 cannot be performed, and the defect inspection cannot be performed properly.

Further, since the preliminary inspection device 8 is a device for detecting the scattered light and extracting the coordinates of the defect, even in the case of a minute defect, the preliminary inspection device 8 has a sufficient effect in step S2 in FIG. Can be detected. On the other hand, when the optical means such as the camera 3 is used, the above-described enlargement effect is not obtained, and the magnification in step S6 must be high in order to extract the coordinates of the defect. This applies to the case where the electron microscope 14 illustrated in FIG. 6 is used. However, if the magnification in step S6 is set to be high, the probability that the defect will be out of the field of view will inevitably increase.

Due to the dilemma described above, there is a gap between the magnification at which a defect is accommodated in the field of view and the magnification at which a minute defect is imaged reliably. It is impossible with the configuration of FIG.

Therefore, the learning section 11d of the present embodiment
Processing unique to the present embodiment, as illustrated in FIG. 19, is performed. Hereinafter, these processes will be described. FIG.
5 is a flowchart illustrating a processing operation performed by the inspection device of the present embodiment. In the processing operation according to the present embodiment, a new process is added between steps S6 and S7 of the first embodiment illustrated in FIG. For simplicity of illustration, steps S4 and S4
FIG. 19 exemplifies the processing only between the steps 8 and 8.

The fact that the defect could not be extracted in step S6 due to either the defect being out of the image pickup processing at the time of imaging in step S5 or the defect being too small. It is assumed that extraction of a defect has failed.

In step S30 following step S6, it is determined whether or not the defect extraction in step S6 is successful. If “YES” is determined, it is determined that there is no need to perform a process specific to the present embodiment, and step S7 is performed.
Then, the processing of the first embodiment is performed as it is.

On the other hand, if "NO" is determined, the flow shifts to step S31. As described above, there are two causes for the failure of the defect extraction: protruding from the imaging field of view and minuteness. In steps S31 to S33, first, a measure is taken when a failure occurs due to the minuteness. . If a defect cannot be extracted by the measures in steps S31 to S33, the failure is caused by protruding, and the processing from step S36 is performed. By thus providing the processing in two stages,
It is possible to eliminate any cause of failure.

In step S31, step S9
The magnification is increased in the same manner as described above. And step S
32 and step S33, step S
Image pickup and extraction of defects are performed in the same manner as in step 5 and step S6. If the cause of the extraction failure in step S6 is minute, the extraction of the defect in step S33 should succeed by increasing the magnification in step S31.

In step S34, step S30
Similarly to the above, it is determined whether or not the defect extraction in step S33 has succeeded. If "YES" is determined, the process shifts to step S7, and the process of the first embodiment is performed as it is.

If "NO" is determined, the cause of the failure is not minute but protruding, and the flow shifts to step S35. In step S35,
Among the origin marks extracted by the preliminary inspection machine 8 in step S2 of FIG. 18, the one closest to the defect to be inspected is picked up. Then, based on the coordinates of the origin mark, the learning unit 11d instructs the stage control unit 6 to move the origin mark into the field of view of the camera 3.

In step S36, step S3
Prior to the imaging of the origin mark in step 7, the magnification of the camera 3 is reduced so that the origin mark is surely contained within the field of view. Thereby, the substantial visual field of the camera 3 is expanded.

In step S38, step S3
In step 7, the origin mark imaged is extracted, and its actual coordinates become known. Then, the difference between the position of the defect to be inspected and the origin mark is known by the processing in step S2 in FIG. 18, and the relative difference in position with the actual coordinates of the origin mark as the starting point. When the X-Y stage 1 is moved by the amount of, the defect to be inspected should exactly fall within the field of view of the camera 3. Therefore, in step S39, the XY stage 1 is caused to perform a relative movement starting from the actual coordinates of the origin mark.

When the deviation of the defect with respect to the visual field is corrected by the relative movement in step S39, in step S6 after the imaging in step S5,
The defect should be extracted. Step S4 for increasing the magnification between step S39 and step S5.
By sandwiching 0, it is possible to avoid the inconvenience that the defect must be inspected from the magnification reduced in step S36. As the magnification after being raised in step S40, it is sufficient that the magnification immediately before being reduced in step S36 is stored in the storage circuit of the learning unit 11d and used.

With the above configuration, step S6
In the case where the extraction of the defect has failed, the cause of the failure can be compensated for and the defect can be appropriately extracted again in step S6. Thereby, the reliability of defect inspection is improved as compared with the case where the configuration described in the first to seventh embodiments is used.

If the extraction in step S6 fails even after the relative movement in step S39, the processing is continued indefinitely for one defect by a loop through steps S5, S34 and S35. May occur. Therefore, the number of times this loop process is performed for a certain defect is counted, and when this count value reaches a certain set value, the observation of the defect is abandoned, and step S3 illustrated in FIG.
A configuration may be adopted in which the process shifts to a and the next defect is observed. In addition to the configuration for counting the number of times the loop is processed, a configuration may be used in which the processing shifts to step S3a when the processing time required for one defect exceeds a certain set time.

In this case, a defect that cannot be automatically inspected must be manually inspected later. Therefore, it is preferable to adopt a configuration in which the coordinates of defects that could not be inspected automatically are stored in, for example, a dedicated file. By adopting such a configuration, it is possible to know only the defects that need to be manually inspected, which is efficient.

By using the inspection apparatus of the present embodiment,
For a defect that could not be inspected in the first imaging, imaging is performed several times with the conditions of magnification and imaging position changed. This allows for more reliable defect extraction,
Inspection reliability is improved.

Such an effect is obtained particularly when an electron microscope or the like is used, that is, when the XY stage 1 is used for observation.
This is noticeable when observation is performed while changing the angle with respect to the plane. This is because when the angle of the XY stage 1 is changed, the XY stage 1 is rotated about a certain axis in the XY stage 1, so that the position of the XY stage 1 is the same. This is also because the amount of movement during rotation differs depending on the distance from the central axis.

If the defect is located near the central axis, even if there is a slight error in the rotation angle, it can be settled within the field of view of the image. It will not fit. However, the configuration according to the present embodiment is useful for capturing an image of a defect that has ceased to fit within the field of view and re-imaging the image, and is useful. For the same reason, when the XY stage 1 is rotated on the XY plane (with the Z axis as the center) for observation, a defect located at a position distant from the center of rotation increases the displacement. It is possible to deal effectively with the configuration of the embodiment.

For the reasons described above, the observation conditions are changed before and after the XY stage 1 is rotated. Therefore, using the integrated value of the movement of the XY stage 1 before the change obtained in step S11 illustrated in FIG. 2 also after the change, the “XY stage to which the integrated value of the movement is added” in step S4a is used. It is inappropriate to do "1 move". To address these inconveniences,
The movement amount from the state in step S34 to the state after the relative movement in step S39 is integrated, and this integrated value may be stored instead of the integrated value in step S11. With such a configuration, an integrated value suitable for the changed condition is used in step S4a.

[0122]

According to the first aspect of the present invention, the relative movement habit of each control means is grasped by learning, and the relative movement habit is determined when a new defect is moved. Will be reflected. This makes it possible to omit the repetition of enlargement and relative movement in order to prevent the defect from deviating from the observation visual field. That is, the time required for the inspection is shortened, and it becomes possible to inspect not only the development prototype and the manufacturing jig, but also products mass-produced for the purpose of sale or use.

According to the configuration of the second aspect, the habit of the control means is grasped by the two parameters of the direction and the amount. These parameters can be stored and read with their values as they are. Therefore, the relative movement performed based on the orientation and the amount is highly reproducible, and the defect remains properly located in the observation field of view.

According to the configuration of the third aspect, it is not necessary to learn the relative movement every time the power of the inspection apparatus is turned on. As a result, the time required for the learning becomes unnecessary, and the time required for the inspection is further reduced in addition to the effect of the configuration according to the first aspect.

According to the fourth aspect of the present invention, only the defect types in which the image information and the analysis result are not associated with each other are the targets of the analysis means. This eliminates the waste of performing the analysis for the type to which the association has been made, and reduces the time required for the inspection.

Further, the association with the unlearned image information is successively performed, so that the types that have been learned increase. Such a configuration is different from a conventional configuration in which learning is not performed during use, in which the type is determined only for defects for which learning of association has been performed in advance before performing inspection.
The performance of the inspection device is improved in parallel with use without taking the trouble of the user.

According to the fifth aspect of the present invention, it is possible to eliminate the waste that the repairing means operates for a defect of a type that cannot be repaired by the repairing means.

According to the structure of claim 6, a display substrate which is determined to be defective due to only one defect existing in a wide area is inspected and repaired by the structure of claim 5. Is done. A display substrate which has been wasted because inspection and repair of a defect in a mass-produced product could not be performed in the past can be utilized by being relieved by the configuration according to claim 5, and the display productivity can be improved. Is improved.

According to the structure of the seventh aspect, the semiconductor substrate is correctly mounted on the stage. As a result, the storage operation by the control means is performed without failure, and it is possible to reliably observe the defect.

According to the configuration of the eighth aspect, a defect which has deviated from the observation visual field is stored in the observation visual field.
As a result, it is possible to reliably perform the inspection for the new defect.

According to the configuration of the ninth aspect, the trouble of having to start observing the defect from the magnification after the reduction is eliminated. As a result, the time for several stages of defect enlargement and observation performed by the observation means is reduced, and the inspection efficiency is improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an example of a configuration of an inspection device according to a first embodiment.

FIG. 2 is a flowchart showing an example of a processing procedure when using the inspection device according to the first embodiment.

FIG. 3 is a flowchart showing another example of the processing procedure when using the inspection device according to the first embodiment.

FIG. 4 is a block diagram showing an example of a configuration of an inspection device according to a second embodiment.

FIG. 5 is a flowchart showing an example of a processing procedure when using the inspection device according to the second embodiment.

FIG. 6 is a block diagram showing an example of a configuration of an inspection device according to a third embodiment.

FIG. 7 is a flowchart showing an example of a processing procedure when using the inspection device according to the third embodiment.

FIG. 8 is a block diagram showing an example of a configuration of an inspection device according to a fourth embodiment.

FIG. 9 is a flowchart showing an example of a processing procedure when using the inspection device according to the fourth embodiment.

FIG. 10 is a block diagram showing an example of a configuration of an inspection device according to a fifth embodiment.

FIG. 11 is a flowchart showing an example of a processing procedure when using the inspection device according to the fifth embodiment.

FIG. 12 is a block diagram showing an example of a configuration of an inspection device according to a sixth embodiment.

FIG. 13 is a flowchart showing an example of a processing procedure when using the inspection device according to the sixth embodiment.

FIG. 14 is a block diagram showing an example of a configuration of an inspection device according to a seventh embodiment.

FIG. 15 is a flowchart showing an example of a processing procedure when using the inspection device according to the seventh embodiment.

FIG. 16 is a flowchart showing an example of a processing procedure when using the inspection device according to the seventh embodiment.

FIG. 17 is a block diagram showing an example of a configuration of an inspection device according to an eighth embodiment.

FIG. 18 is a flowchart showing an example of a processing procedure when using the inspection device according to the eighth embodiment.

FIG. 19 is a flowchart showing another example of the processing procedure when using the inspection device according to the eighth embodiment.

[Explanation of symbols]

1 XY stage, 2 wafers, 2a thin film transistor (TFT) substrate, 3 cameras, 4 image processing units,
5 image signal line, 5a image signal line, 6 stage control section, 7 control signal line, 8 preliminary inspection machine, 9, 10a coordinate signal line, 10b movement signal line, 10c analysis result signal line, 10d repair signal Lines, 11, 11a-11
d Learning unit, 12 component analysis unit, 14 electron microscope, 1
5 focused ion beam optical system, 16 laser beam machine,
17 Monitor.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code FI G06T 7/00 G06F 15/62 405A

Claims (9)

[Claims] 1. An extracting means for extracting a position of a defect of a semiconductor substrate to be inspected, an observing means for enlarging and observing the defect, and the extracted position is contained in an observation field of view of the observing means. Control means for performing the storing operation in parallel with the enlargement by relative movement between the semiconductor substrate and the observation means; learning of the relative movement; An inspection apparatus comprising: a learning unit that instructs the control unit to reproduce a new defect. 2. The inspection apparatus according to claim 1, wherein the learning is performed by storing a direction and an amount of the relative movement. 3. The inspection apparatus according to claim 2, wherein the orientation and the amount are continuously stored even when the power of the inspection apparatus is turned off. 4. The inspection apparatus according to claim 1, further comprising an analysis unit for analyzing a component of the defect to obtain an analysis result, wherein the observation unit is configured to detect the defect. The learning means determines the type of the new defect that gives new image information based on learning by associating the analysis result with the image information. An inspection device that determines whether the analysis unit operates. 5. The inspection device according to claim 4, further comprising a repair unit that repairs the new defect, wherein the learning unit is configured to perform the repair in accordance with the type determined by the association. An inspection device that determines whether the repair means operates. 6. The inspection apparatus according to claim 5, wherein the semiconductor substrate is a display substrate. 7. The inspection device according to claim 1, wherein the extraction unit and the observation unit are separate from each other, and the semiconductor substrate is a reference for a relative position on the semiconductor substrate. 1
The first and second pointers,
Is extracted by the extraction unit, and after the first and second positions are extracted by the extraction unit, the extraction unit is placed on a stage whose movement is controlled by the control unit. With respect to the semiconductor substrate mounted on a stage, the first and second positions are extracted, and the learning unit is configured to perform the virtual second processing when the semiconductor substrate is correctly mounted on the stage. By comparing the first and second positions with the first and second positions extracted by the observation means, respectively, the positional relationship between the semiconductor substrate and the stage is recognized, and An inspection apparatus for instructing the control means to move the stage to be canceled. 8. The inspection apparatus according to claim 1, wherein the learning unit, when detecting that the defect is not observable in the observation visual field, reduces a magnification of the observation unit. The control means instructs that a pointer serving as a reference for a relative position on the semiconductor substrate is stored in the observation field of view after the pull-down, and the observation means is stored in the observation field of view. Extracting a reference position, which is the position of the pointer, wherein the learning unit moves the relative position between the position of the pointer and the position of the defect obtained by the extraction unit, An inspection apparatus for instructing the control unit to perform the process with a reference position as a starting point. 9. The inspection device according to claim 8, wherein the learning unit is after the reference position is extracted,
An inspection apparatus for instructing the observation means to increase the magnification before instructing to move the relative position starting from the reference position.