JP2017033365A - Detection method, detection device and detection program - Google Patents

Abstract

An appropriate reference position of a substrate is detected.
A pattern portion 25A corresponding to a set pattern portion 21A of a first template is detected from an image of a substrate by image recognition. The image is subjected to matrix division at the first size, and a pattern portion 26A corresponding to the pattern portion 22A of the second template set at the first size is detected from the matrix-divided image by image recognition. It is determined whether or not the arrangement of the pattern portion 25A and the pattern portion 26A in the image matches the arrangement of the set pattern portion 21A and the pattern portion 22A, and the reference position of the substrate is detected and set based on the result. To do.
[Selection] Figure 11

Description

  The present invention relates to a detection method, a detection apparatus, and a detection program.

2. Description of the Related Art A technique is known in which an electrical test and an appearance inspection are performed on a manufacturing process or a substrate including a semiconductor element after manufacturing, and whether the semiconductor element is good or bad is determined based on the result.
The predetermined position of the substrate is used as a reference, processing is performed on a specific region of the substrate based on the reference position, and information on the specific region of the substrate is acquired. Regarding the setting of the reference position of the board, the part where the electrical characteristics are different from the other is detected by an electrical test, the technique using the part as the reference position, the part whose shape is different from the other is detected by image processing, A technique using the portion as a reference position is known.

Japanese Patent Laid-Open No. 2003-7604 JP-A-62-232504

  If the reference position of a substrate including a semiconductor element in the manufacturing process or after manufacturing cannot be detected with high accuracy, a deviation occurs between the position of the semiconductor element on the substrate and the test result or inspection result, so that the non-defective and defective semiconductors There is a risk that the elements will be confused. In this way, when the non-defective product and the defective semiconductor element are confused, the defective product may be mistaken for the good product or the good product may be discarded as a defective product.

  According to one aspect of the present invention, a step of detecting a first pattern portion corresponding to a set first template by image recognition from a first image of a first substrate having a pattern; Dividing the matrix by one size, detecting the second pattern portion corresponding to the second template set at the first size by image recognition from the matrix-divided first image, and the first And a step of determining whether or not a first arrangement of the first pattern portion and the second pattern portion in the image matches a set second arrangement.

  Moreover, according to one viewpoint of this invention, the detection apparatus and detection program which are used for the said detection method are provided.

  According to the disclosed technique, it is possible to properly detect the reference position of the substrate, and it is possible to execute processing on an appropriate region of the substrate and acquire information on the appropriate region.

It is FIG. (1) explaining an example of the detection method of the reference | standard position using an image recognition technique. It is FIG. (2) explaining an example of the detection method of the reference | standard position using an image recognition technique. It is a figure explaining the setting of the 1st template which concerns on 1st Embodiment. It is a figure explaining the setting of the 2nd template which concerns on 1st Embodiment. It is FIG. (1) explaining the setting of the template arrangement | positioning information which concerns on 1st Embodiment. It is FIG. (2) explaining the setting of the template arrangement information according to the first embodiment. It is FIG. (3) explaining the setting of the template arrangement information according to the first embodiment. It is FIG. (1) explaining an example of the reference | standard position detection of the wafer which concerns on 1st Embodiment. It is FIG. (2) explaining an example of the reference | standard position detection of the wafer which concerns on 1st Embodiment. FIG. 6 is a third diagram illustrating an example of wafer reference position detection according to the first embodiment; FIG. 6 is a diagram (part 4) illustrating an example of wafer reference position detection according to the first embodiment; It is a figure which shows the structural example of the detection apparatus which concerns on 1st Embodiment. It is FIG. (1) which shows an example of the detection process which concerns on 1st Embodiment. It is FIG. (2) which shows an example of the detection process which concerns on 1st Embodiment. It is a figure explaining matrix division | segmentation. It is FIG. (1) which shows the example of arrangement | positioning of the 2nd template corresponding | compatible pattern part in the area | region divided into matrix. It is FIG. (2) which shows the example of arrangement | positioning of the 2nd template corresponding | compatible pattern part in the area | region divided into matrix. FIG. 10 is a diagram (part 3) illustrating an arrangement example of the second template corresponding pattern portions in the matrix-divided region. It is a figure which shows an example of the detection process which concerns on 2nd Embodiment. It is a figure explaining the detection of a template corresponding | compatible pattern part. It is a figure explaining an example of pattern part arrangement | positioning information and a coincidence determination using the same. It is a figure explaining another example of pattern part arrangement | positioning information and a coincidence determination using the same. It is a figure which shows another example of the detection process which concerns on 2nd Embodiment. It is a figure which shows an example of the detection process which concerns on 3rd Embodiment. It is a figure which shows the structural example of the hardware of a computer. It is a figure which shows an example of the principal part cross section of a wafer.

  In recent years, due to demands for the spread of portable devices, energy saving, waste reduction, and the like, there is an increasing demand for semiconductor devices incorporating a nonvolatile memory that can rewrite data and retain data even when the power is turned off. Nonvolatile memory includes EEPROM (Electric Erasable Programmable Read Only Memory), flash memory, FeRAM (Ferroelectric Random Access Memory), and the like.

  One of tests performed on a semiconductor element including such a memory is an electrical test (retention test) for checking whether data is lost after heating to a predetermined temperature. This electrical test may be performed after packaging individual semiconductor elements from a wafer (substrate) or after assembling into a product or the like, but is performed for each semiconductor element at the wafer level before individualization. There are many cases. Based on the test result at the wafer level, a semiconductor element (non-defective chip) regarded as a non-defective product is distinguished from a semiconductor element (defective chip) defined as a defective product. The result of the test at the wafer level is, for example, in the form of a wafer map. In the subsequent process, non-defective chips are selected based on the wafer map, and defective chips are discarded.

  However, if the wafer and the test apparatus are not properly aligned, there may be a deviation between the position of the semiconductor element on the wafer and the test result. For example, the test intended for the semiconductor element A is actually performed for the adjacent semiconductor element B. Depending on the test results of the semiconductor elements A and B, the non-defective chip and the defective chip May be confused. If the non-defective chip is confused with the non-defective chip in this manner, the defective chip may be mistaken for the non-defective chip in a later process, or the non-defective chip may be discarded as a defective chip.

  Therefore, in the wafer level test, a reference mark is provided at a specific position on the wafer, the wafer is aligned with the test device using the mark as a mark, and the test target semiconductor element is located on the wafer. Know what it is. The alignment includes alignment of the origin, alignment in the θ direction, alignment of the reference point when setting or changing the magnification of the wafer image, and the like.

  For example, in one shot area at the time of wafer exposure, one chip is crushed to form a pattern to be a mark, or a portion where no pattern is formed on one chip is provided, and the one chip on the wafer is provided. There is a method of detecting a reference position by image recognition technology using a part as a mark. However, in this method, the yield of a semiconductor device that is a product from a single wafer is reduced, the cost is increased and the yield is increased due to the introduction of a process for providing a semiconductor device that is a product and a mark having a different pattern. There is a risk of lowering the level.

  In addition, a semiconductor element that detects a voltage value different from that of a product semiconductor element during an electrical test or a semiconductor element that is imaged with a contrast different from that of a product semiconductor element during image processing is used as a mark on the wafer. There is also a method of forming. However, these methods may also lead to a decrease in yield of semiconductor elements from a single wafer, an increase in cost and a decrease in yield due to the introduction of a process for providing marks.

  Although there is a method of finding a specific semiconductor element by directly counting the semiconductor elements on the wafer, this method may cause an error in counting, and thousands of semiconductor elements are formed on the wafer. Such a counting error is more likely to occur when the scale is large.

As an example, a method for detecting a mark (reference position) using an image recognition technique will be described.
1 and 2 are diagrams for explaining an example of a reference position detection method using an image recognition technique.
For example, an image 100 as shown in FIG. 1A is acquired for a wafer on which a semiconductor element is formed. The image 100 includes a semiconductor element 110, a scribe region 120, and a mark 130.

  The semiconductor elements 110 are arranged in a matrix. The scribe region 120 is provided between the adjacent semiconductor elements 110. The mark 130 includes a mark 131 provided at the end of the semiconductor element 110 and a mark 132 provided at the end of the semiconductor element 110 and the scribe region 120 on the side thereof. The marks 131 and 132 are formed on the image 100 so as to have a contrast different from that of the semiconductor element 110 and the scribe region 120 where the marks 131 and 132 are not provided. Depending on the size of the marks 131 and 132, the scribe area 120 between the adjacent marks 131 and 132 and the scribe area 120 between the adjacent marks 132 may be displayed on the image 100 as in the example of FIG. The contrast is the same as that of the marks 131 and 132.

  Now, the shape of the pattern portion of the region surrounded by the frame 200a in FIG. 1B of such an image 100 is set as a template 200 for image recognition. From the image 100 as shown in FIG. 1A, a pattern portion (set) matching the template 200 as shown in FIG. 1B is detected by image recognition.

  For example, as shown in FIG. 2, a stage on which a wafer for capturing an image 100 is placed is moved (step S1), a pattern portion of a predetermined area of the image 100 to be captured, and a template as illustrated in FIG. 200 is compared (step S2). And it is determined whether the pattern part of the predetermined area | region of the image 100 and the template 200 correspond (step S3). If it is determined that they match (step S3), the pattern portion of the predetermined area of the image 100 is detected as a pattern portion that matches the template 200 (step S4), and the image recognition is completed. On the other hand, when it is determined that the pattern portion of the predetermined area of the image 100 does not match the template 200 (step S3), the process returns to step S1 and the subsequent processing is performed.

  By such image recognition, a region including a pattern portion that matches the set template 200 is detected from the wafer image 100, and the detected region (for example, one point such as the center or corner of the semiconductor element in the region) is detected. ) Can be used as a reference for alignment.

  However, the pattern shape of the image recognition template 200 surrounded by the frame 200a in FIG. 1B is surrounded by the frame 210 and the frame 220 in the same image 100 shown in FIG. 1C. Similar to the pattern shape of the region. Therefore, at the time of image recognition, it is assumed that the pattern part of the frame 210 or the frame 220 in FIG. 1C mistakenly matches the template 200 from the image 100 instead of the pattern part of the frame 200a in FIG. It may be detected. When such an erroneous detection occurs, the region used as a reference is shifted, so that proper alignment cannot be performed on the wafer from which the image 100 is acquired.

  To solve such a problem, it is possible to increase the matching rate (matching rate) of image recognition, for example, by crushing a semiconductor element for one chip on the wafer as a mark. However, this method leads to a decrease in the yield of the semiconductor element as described above. Further, in this method, when the semiconductor element on the wafer is miniaturized, the mark becomes small, the matching rate of image recognition does not increase, and the fundamental improvement may not be achieved.

  In addition, it is possible to increase the matching rate of image recognition using the high magnification image 100. However, when a mechanism such as a laser interferometer stage (driving type used in an exposure apparatus) is installed in order to achieve a high magnification that can obtain a sufficient matching rate even if the semiconductor elements on the wafer are miniaturized, In such a case, the cost of the apparatus will increase significantly.

In view of the above points, a detection method as shown in the following embodiment using an image recognition technique is employed to appropriately detect the reference position of the substrate.
First, the first embodiment will be described.

FIG. 3 is a diagram for explaining the setting of the first template according to the first embodiment.
An image 10 as shown in FIG. 3 is acquired as the template setting image for the wafer on which the semiconductor elements are formed. The image 10 includes a semiconductor element 11, a scribe region 12, and a mark 13.

  The semiconductor elements 11 are arranged in a matrix. The scribe region 12 is provided between adjacent semiconductor elements 11. The mark 13 includes a mark 13 a provided at the end of the semiconductor element 11 and a mark 13 bb provided at the end of the semiconductor element 11 and the scribe region 12 on the side thereof. The mark 13a and the mark 13bb are formed on the image 10 so as to have a contrast different from that of the semiconductor element 11 and the scribe region 12 where the mark 13a and the mark 13bb are not provided. Note that the scribe region 12 between adjacent marks 13bb has the same contrast as the mark 13bb on the image 10, and appears as an aggregate like the mark 13b. As an example, FIG. 3 illustrates a case where three marks 13b, which are aggregates of vertically adjacent marks 13bb, are arranged in a diagonal direction.

  Such marks 13a and 13bb (13b) appearing in the image 10 can be obtained by the structure shown in FIG. FIG. 26 shows an example of a cross section of the main part of the wafer. As shown in FIG. 26, the wafer has a semiconductor substrate 1, an insulating layer 2 formed on the semiconductor substrate 1, and a protective layer 3 formed on the insulating layer 2. The semiconductor element 11 includes a device part 4 in which an element such as a transistor is formed and a pad part 5 formed at an end part. A scribe pattern part 6 is provided in a scribe region 12 adjacent to the pad part 5. It is done. For example, metal is used for the pad portion 5 and the scribe pattern portion 6, and the pad portion 5 and the scribe pattern portion 6 are exposed from the insulating layer 2 and the protective layer 3, for example. During acquisition (imaging) of the image 10, the pad portion 5 and the scribe pattern portion 6 are marked 13a, 13bb (13b) due to the difference in reflectance between the insulating layer 2 and the protective layer 3, and the pad portion 5 and the scribe pattern portion 6. Appears as The same applies to an image 10A described later.

  For the template setting image 10 as described above, for example, a pattern shape of a pattern portion including two marks 13b surrounded by a frame 20a as shown in FIG. 3 is set as the first template 21 for image recognition. The

FIG. 4 is a diagram illustrating the setting of the second template according to the first embodiment.
For the template setting image 10, the pattern shape of the pattern portion including one mark 13 b surrounded by a frame 20 b as shown in FIG. 4 together with the first template 21 as described above is the second template for image recognition. 22 is set. The size of the second template 22 is set based on the arrangement pitch of the semiconductor elements 11 including the scribe region 12 in the image 10. The arrangement pitch is a distance from an intermediate point of one scribe region 12 to an intermediate point of another scribe region 12 facing the one scribe region 12 across the semiconductor element 11.

  The information indicating the arrangement of the pattern portion group that matches the set second template 22 in the template setting image 10 as described above is associated with the information indicating the arrangement of the pattern portion that matches the first template 21. And is set as template arrangement information. The setting of such template arrangement information will be further described with reference to FIGS.

5 to 7 are diagrams for explaining setting of template arrangement information according to the first embodiment.
In setting the template arrangement information using the image 10, first, as shown in FIG. 5, the image 10 is divided into matrices by using the size of the second template 22 as a unit division size. Each divided region 14 of the image 10 divided in this way is compared with the set second template 22.

  A part of the matrix-divided image 10 is shown in FIG. 6A, and the second template 22 is shown in FIG. 6B. 6A illustrates six divided regions 14 (14a, 14b, 14c, 14d, 14e, and 14f) among the divided regions 14 generated by matrix division of the image 10. FIG. Each divided region 14 of the matrix-divided image 10 and the set second template 22 are compared by image recognition, and a pattern portion (divided region 14) that matches the second template 22 is detected from the image 10. The In the example of FIG. 6A, the divided area 14d (lower left) matches the second template 22 of FIG. 6B, and the remaining divided areas 14a, 14b, 14c, 14e, excluding the divided area 14d. 14 f does not match the second template 22. Each divided area 14 in the image 10 is compared with the second template 22 by image recognition and matched, and a divided area 14 that matches the second template 22 is detected.

  FIG. 7 shows an example of the result of comparison and matching determination between the divided region 14 and the second template 22. In FIG. 7, a pattern portion (divided region 14) that matches the second template 22 in the matrix-divided image 10 is illustrated by a frame 20 b. In FIG. 7, the pattern portion that matches the first template 21 in the matrix-divided image 10 is also illustrated by a frame 20 a.

  As a result of the comparison between the divided area 14 and the second template 22 and the matching determination, for example, as shown in FIG. 7, three places in the image 10, that is, three places where three marks 13b (FIG. 3) arranged in an oblique direction exist. A pattern portion matching the second template 22 is detected in the divided region 14. Information indicating the arrangement of the three divided regions 14 (pattern portions matching the second template 22) in the image 10 is acquired.

The information indicating the arrangement of the pattern portion (divided region 14) group matching the second template 22 existing in the image 10 to be exemplified includes the following contents, for example.
That is, first, one of the pattern part groups that match the second template 22 that has a fixed arrangement relationship with the pattern part that matches the first template 21, for example, the upper left pattern part 22a is set as the reference point. The content of the positions of the remaining pattern portions 22b and 22c with respect to the reference point pattern portion 22a set according to such a fixed reference point setting rule is included.

  Specifically, when the XY coordinate system having the origin of the reference point pattern portion 22a is set, the pattern portion 22b is divided into two regions 14 in the + X direction from the origin (X = 0, Y = 0), It is located at one point (X = 2, Y = −1) in the divided region 14 in the −Y direction. The pattern portion 22c has four points in the divided region 14 in the + X direction from the origin (X = 0, Y = 0) and two points in the divided region 14 in the -Y direction (X = 4, Y = -2). Located in. Such positional relationships of the three pattern portions 22a, 22b, and 22c are acquired as information indicating the arrangement of the pattern portion group that matches the second template 22 in the image 10.

  The information indicating the arrangement of the pattern part group (22a, 22b, 22c) of the frame 20b that matches the second template 22 acquired in this way is the pattern part (22a, 22a, 22c) of the frame 20a that matches the first template 21. 22b) is associated with information indicating the arrangement. The associated information is set as template arrangement information (FIG. 7).

  As described above, for the template setting image 10, first, a template layout showing the settings of the first template 21 and the second template 22, and the layout of the pattern portion group that matches the first template 21 and the second template 22. Information is set.

  When detecting the reference position of a reference position detection target wafer, an image is acquired for the target wafer, and a pattern portion corresponding to the first template 21 and a pattern portion corresponding to the second template 22 are obtained by image recognition. A group is detected. The arrangement of these pattern parts is compared with the set template arrangement information, that is, the arrangement of the pattern part of the frame 20a that matches the first template 21 and the pattern part group of the frame 20b that matches the second template 22. It is determined whether or not they match. Then, any one of the position of the pattern portion matching the first template 21 in the image determined to match, the predetermined position (the center and corner of the semiconductor element) inside the pattern portion group, and the pattern portion group matching the second template 21 One position or the like is set as the reference position of the target wafer.

The detection of the reference position of the wafer will be described more specifically with reference to FIGS.
8 to 11 are diagrams for explaining an example of wafer reference position detection according to the first embodiment.

  For example, it is assumed that an image 10A as shown in FIG. 8A is acquired for the reference position detection target wafer. This image 10A includes the semiconductor element 11, the scribe region 12, and the mark 13 as in the template setting image 10. The mark 13 includes a mark 13 a provided at the end of the semiconductor element 11 and a mark 13 bb provided at the end of the semiconductor element 11 and the scribe region 12 on the side thereof. Three marks 13b, which are aggregates of vertically adjacent marks 13bb, are arranged side by side in an oblique direction.

  In such an image 10A, the pattern portion that matches the first template 21 set as described above and that also matches the arrangement relationship with the pattern portion group that matches the second template 22 is shown in FIG. The region is surrounded by the frame 20Aa in FIG. The pattern portion 21A in the area surrounded by the frame 20Aa is a pattern portion to be detected as matching the first template 21 by image recognition.

  When image recognition is performed on the image 10A, it is only necessary to detect the pattern portion 21A of the frame 20Aa as shown in FIG. 8A as the pattern portion corresponding to the first template 21. However, as a pattern portion corresponding to the first template 21, for example, a pattern portion 23A surrounded by a frame 20Ac as shown in FIG. 8B similar to the pattern portion 21A surrounded by the frame 20Aa is detected. . The pattern portion 23A in FIG. 8B has a pattern shape that matches the first template 21, but a pattern portion whose arrangement relationship with the pattern portion group corresponding to the second template 22 does not match that in FIG. 8A. It is. For this reason, if the pattern portion 23A is detected by image recognition and used for setting the reference position, there is a possibility that the alignment of the wafer is shifted.

  In this method, after the pattern portion corresponding to the first template 21 is detected by image recognition in this way, the image 10A is divided into matrices. This matrix division is performed with the size of the second template 22. Comparison and matching determination with the second template 22 are performed for each divided region 14 of the image 10A divided in this way, and a pattern portion group corresponding to the second template 22 is detected. The arrangement of the pattern portion group detected as corresponding to the first template 21 and the second template 22 in the image 10A is compared with the template arrangement information set based on the first template 21 and the second template 22, and matches. Determined. This point will be described with reference to examples shown in FIGS.

  FIG. 9A is a diagram illustrating an arrangement corresponding to the example of FIG. FIG. 9A illustrates the arrangement of the divided region 14, the pattern portion 21 </ b> A (frame 20 </ b> Aa) that matches the first template 21, and the pattern portion 22 </ b> A group (frame 20 </ b> Ab) that matches the second template 22. . In this example, according to a certain reference point setting rule, the upper left pattern portion 22Aa in the pattern portion 21A is set as a reference point. Information indicating the arrangement of the pattern portion 21A and the pattern portion 22A group as shown in FIG. 9A is set as template arrangement information (FIGS. 5 to 7).

  FIG. 9B is a diagram illustrating an arrangement corresponding to the example of FIG. In FIG. 9B, the divided area 14, the pattern portion 23 A (frame 20 Ac) detected as corresponding to the first template 21, and the pattern portion 24 A group detected as corresponding to the second template 22 ( The arrangement of the frame 20Ad) is illustrated. In the example of FIG. 9B, according to the same reference point setting rules as in FIG. 9A, the upper left pattern portion 24Ab in the pattern portion 23A is set as the reference point. In this case, the arrangement of the pattern portion 23A and the pattern portion 24A group does not match the arrangement of the pattern portion 21A and the pattern portion 22A group in FIG.

  That is, in the set template arrangement information, as shown in FIG. 9A, the upper left pattern portion 22A (22Aa) in the pattern portion 21A is used as a reference point, and two pattern portion 22A groups obliquely downward to the right from there. (22Ab, 22Ac) are arranged. On the other hand, in the example of FIG. 9B, the upper left pattern portion 24A (24Ab) in the pattern portion 23A is set as a reference point, and only one pattern portion 24A (24Ac) is arranged obliquely downward to the right from there. Instead, the pattern portion 24A (24Aa) is arranged obliquely on the left.

  Thus, the arrangement of the pattern portion 23A and the pattern portion 24A group does not match the arrangement of the pattern portion 21A and the pattern portion 22A group of the set template arrangement information. From this, it is found that there is an error in the detection of the pattern portion corresponding to the first template 21 by the image recognition of the image 10A, that is, the detection of the pattern portion 23A of the frame 20Ac.

  On the other hand, unlike the example of FIG. 8B, the frame 20Ae as shown in FIG. 10B is assumed to correspond to the first template 21 by image recognition for the image 10A as shown in FIG. 10A. Assume that the enclosed pattern portion 25A is detected. Also in this case, similarly to the above, the image 10A is divided into matrices, the divided regions 14 and the second template 22 are compared and matched, and a pattern portion group corresponding to the second template 22 is detected.

  FIG. 11A illustrates the arrangement of the divided region 14, the pattern portion 21A that matches the first template 21, and the pattern portion 22A group that matches the second template 22 as in FIG. 9A. Yes. According to a certain reference point setting rule, the upper left pattern portion 22Aa in the pattern portion 21A is set as a reference point. Information indicating the arrangement of the pattern portion 21A and the pattern portion 22A group as shown in FIG. 11A is set as template arrangement information (FIGS. 5 to 7).

  FIG. 11B is a diagram illustrating an arrangement corresponding to the example of FIG. In FIG. 11B, the divided region 14, the pattern portion 25 </ b> A (frame 20 </ b> Ae) detected as corresponding to the first template 21, and the pattern portion 26 </ b> A group detected as corresponding to the second template 22 ( The arrangement of the frame 20Af) is illustrated. In the example of FIG. 11B, according to the same reference point setting rules as in FIG. 11A, the upper left pattern portion 26Aa in the pattern portion 25A is set as the reference point. In this case, the arrangement of the pattern portion 25A and the pattern portion 26A group matches the arrangement of the pattern portion 21A and the pattern portion 22A group in FIG.

  That is, in FIG. 11A, two pattern portions 22A groups (22Ab, 22Ac) are arranged diagonally to the right from the reference point pattern portion 22A (22Aa). Similarly, in FIG. 11B, two pattern portions 26A group (26Ab, 26Ac) are arranged obliquely downward to the right from the reference point pattern portion 26A (26Aa).

  Thus, the arrangement of the pattern portion 25A and the pattern portion 26A group matches the arrangement of the pattern portion 21A and the pattern portion 22A group of the set template arrangement information. From this, it is found that the detection of the pattern portion corresponding to the first template 21 by the image recognition of the image 10A, that is, the detection of the pattern portion 25A of the frame 20Ae was appropriate.

When the image recognition is properly performed, the reference position of the target wafer can be properly detected and set.
Subsequently, an example (first example) of a detection apparatus and its processing used for the reference position detection as described above will be described.

FIG. 12 is a diagram illustrating a configuration example of the detection apparatus according to the first embodiment.
12 includes an acquisition unit 31, a first setting unit 32, a second setting unit 33, a dividing unit 34, a first generation unit 35, a first detection unit 36, a second detection unit 37, and a second generation. A unit 38, a determination unit 39, and a drive unit 40; The detection device 30 further includes a storage unit 41, an input unit 42, and an output unit 43.

  The acquisition unit 31 acquires an image of a wafer on which a semiconductor element is formed. The acquisition unit 31 acquires, as an image, a template setting image 10 used for setting the first template 21 and the second template 22 or a reference position detection target wafer image 10A. The acquisition unit 31 includes, for example, an imaging device, and acquires an image by imaging a predetermined area of a predetermined wafer using the imaging device. Note that the acquisition unit 31 may acquire an image of a predetermined wafer by receiving data of an image captured by an imaging device outside the detection device 30.

  The first setting unit 32 sets the first template 21 for image recognition using the template setting image 10 among the images acquired by the acquiring unit 31. For example, as shown in FIG. 3, the first setting unit 32 sets the pattern portion surrounded by the frame 20 a in the first template 21. The first setting unit 32 receives data of the first template 21 set using the template setting image 10 outside the detection device 30 and sets it as the first template 21 used in the detection device 30. May be.

  The second setting unit 33 sets the second template 22 having a predetermined size for image recognition using the template setting image 10 among the images acquired by the acquiring unit 31. For example, as illustrated in FIG. 4, the second setting unit 33 sets the pattern portion surrounded by the frame 20 b in the second template 22. The second setting unit 33 receives the data of the second template 22 set using the template setting image 10 outside the detection device 30 and sets it as the second template 22 used by the detection device 30. May be.

  The dividing unit 34 divides the image 10 or the image 10A acquired by the acquiring unit 31 into a matrix with a predetermined division size. For example, as illustrated in FIG. 5, the dividing unit 34 performs matrix division on the image 10 with the size of the second template 22 (frame 20 b) set by the second setting unit 33. Similarly, the dividing unit 34 divides the image 10 </ b> A into a matrix with the set size of the second template 22. The size of the second template 22 is set based on the arrangement pitch of the semiconductor elements 11 including the scribe region 12 in the image 10 or the image 10A. The relationship between the second template, the arrangement pitch, and the matrix division will be described later.

  The first generation unit 35 uses the first template 21 set by the first setting unit 32, the second template 22 set by the second setting unit 33, and the image 10 divided by the dividing unit 34, Generate template placement information. The template arrangement information indicates the arrangement of the pattern portion that matches the first template 21 and the pattern portion group that matches the second template 22 in the matrix-divided image 10. For example, as illustrated in FIG. 6, the first generation unit 35 performs comparison and matching determination between the divided region 14 of the matrix-divided image 10 and the second template 22. Then, for example, as shown in FIG. 7, the first generation unit 35 associates information indicating the arrangement of the pattern part group that matches the second template 22 with information indicating the arrangement of the pattern part that matches the first template 21. This is generated and set as template arrangement information.

  The first detection unit 36 performs image recognition on the pattern portion corresponding to the first template 21 set by the first setting unit 32 from the image 10A of the reference position detection target wafer among the images acquired by the acquisition unit 31. Detect by. The first detection unit 36 corresponds to the first template 21 by image recognition, for example, the pattern portion 23A of the frame 20Ac as shown in FIG. 8B or the frame 20Ae as shown in FIG. 10B. The pattern portion 25A is detected.

  The pattern portion detected by the first detection unit 36 as corresponding to the first template 21 does not necessarily match the first template 21 at the time of detection (image recognition time). The pattern portion detected by the first detection unit 36 may not match the first template 21 depending on the algorithm and accuracy of image recognition employed for detection.

  The second detection unit 37 is a pattern corresponding to the second template 22 set by the second setting unit 33 from the image 10A of the reference position detection target wafer acquired by the acquisition unit 31 and matrix-divided by the division unit 34. The group is detected by image recognition. The second detection unit 37 corresponds to the second template 22 by image recognition. For example, the pattern unit 24A of the frame 20Ad as shown in FIG. 8B or the frame 20Af as shown in FIG. The pattern portion 26A is detected.

  Note that the pattern portion detected by the second detection unit 37 as corresponding to the second template 22 does not necessarily match the second template 22 at the time of detection (image recognition time). The pattern portion detected by the second detection unit 37 may not match the second template 22 depending on the algorithm and accuracy of image recognition employed for detection.

  The second generation unit 38 generates pattern part arrangement information indicating the arrangement of the pattern part group detected by the first detection unit 36 and the second detection unit 37. The pattern portion arrangement information includes a pattern portion detected by the first detector 36 as corresponding to the first template 21, and a pattern portion group detected as corresponding to the second template 22 by the second detector 37. The arrangement of For example, the second generation unit 38 includes information indicating the arrangement of the pattern unit 23A and the pattern unit 24A as illustrated in FIG. 9B, and the pattern unit 25A and the pattern unit 26A as illustrated in FIG. 11B. Is generated and set as pattern portion arrangement information.

  The determination unit 39 compares the template arrangement information generated by the first generation unit 35 with the pattern part arrangement information generated by the second generation unit 38 and determines whether or not they match. The determination unit 39 determines that the arrangement of the corresponding pattern part group of the detected first template 21 and the second template 22 in the pattern part arrangement information is the matching pattern part of the first template 21 and the second template 22 in the template arrangement information. It is determined whether or not the group arrangement is met.

When acquiring the image 10 or the image 10A by imaging, the drive unit 40 drives a stage on which the wafer to be imaged is placed, and adjusts the imaging region (field of view).
The storage unit 41 stores programs and various data used for processing of the detection device 30.

The input unit 42 inputs various conditions and data required for processing of the detection device 30.
The output unit 43 displays various data obtained by the processing of the detection device 30 on a display device or transmits the data to another device.

13 and 14 are diagrams illustrating an example of the detection process according to the first embodiment.
First, setting of the first template 21 and the second template 22 and generation of template arrangement information indicating their arrangement will be described with reference to FIG.

  The detection device 30 acquires the template setting image 10 by the acquisition unit 31 (step S10). The detection device 30 acquires the template setting wafer image 10 including the semiconductor element 11, the scribe region 12, and the marks 13a and 13b (13bb) as shown in FIG.

  The detection device 30 sets the first template 21 by using the acquired image 10 by the first setting unit 32 (step S11). The detection apparatus 30 sets the pattern part including the two marks 13b surrounded by the frame 20a as shown in FIG.

  The detection device 30 sets the second template 22 by using the acquired image 10 by the second setting unit 33 (step S12). The detection device 30 sets, in the second template 22, a pattern portion including one mark 13b surrounded by the frame 20b as shown in FIG.

  The detection device 30 divides the acquired image 10 into a matrix by the dividing unit 34 using the set size of the second template 22 as a unit division size (step S13). In the detection device 30, the division unit 34 divides the image 10 in a matrix with the size of the second template 22 as shown in FIG.

  The detection apparatus 30 uses the first generation unit 35 to display template arrangement information indicating the arrangement of the pattern part that matches the first template 21 and the pattern part group that matches the second template 22 in the matrix-divided image 10. Is generated (step S14). For example, as illustrated in FIG. 6, the detection device 30 performs comparison and matching determination between the divided region 14 of the matrix-divided image 10 and the second template 22 using the first generation unit 35. The detection device 30 further uses the first generation unit 35 to display information indicating the arrangement of the pattern portion group that matches the second template 22 and the arrangement of the pattern portion that matches the first template 21 as shown in FIG. The template placement information is generated in association with the information shown.

Next, detection of the reference position of the target wafer will be described with reference to FIG.
The detection device 30 moves the stage by the driving unit 40 so that a predetermined area of the reference position detection target wafer placed on the stage enters the field of view (step S20). The detection device 30 acquires an image 10A including a predetermined region of the target wafer placed on the stage by the acquisition unit 31 (step S21). The detection apparatus 30 uses the acquisition unit 31 to obtain an image 10A of the target wafer including the semiconductor element 11, the scribe region 12, and the marks 13a and 13b (13bb) as shown in FIG. 8B or 10B, for example. To get.

  The detection device 30 performs image recognition using the first template 21 set by the first setting unit 32 on the image 10A acquired by the first detection unit 36 (step S22). It is determined whether or not a corresponding pattern portion has been detected (step S23).

  When the first detection unit 36 determines that the pattern portion corresponding to the first template 21 is not detected from the image 10A (step S23), the detection device 30 moves the stage (step S20), An image 10A is acquired for another region (step S21). Similarly, image recognition using the first template 21 is performed (step S22). The detection device 30 executes such processing until it is determined that the pattern portion corresponding to the first template 21 has been detected.

  When the detection device 30 determines that the pattern portion corresponding to the first template 21 has been detected from the image 10A (step S23), the image 10A is set by the division unit 34 by the second setting unit 33. The matrix is divided by the size of the second template 22 (step S24).

  The detection device 30 performs image recognition using the set second template 22 on the matrix-divided image 10A by the second detection unit 37 (step S25). The detection device 30 performs image recognition using the second template 22 for all the divided regions 14 of the matrix-divided image 10A by the second detection unit 37 (step S26).

  The detection device 30 executes processes such as steps S24 to S26, and the second detection unit 37 detects the pattern unit group corresponding to the second template 22 from the matrix-divided image 10A. In the detection device 30, the arrangement of the pattern unit group detected by the second generation unit 38 as corresponding to the second template 22 by the second detection unit 37 is assumed to correspond to the first template 21 by the first detection unit 36. The pattern portion arrangement information is generated in association with the detected arrangement of the pattern portion (step S27).

  The detection device 30 is detected by the second generation unit 38 as corresponding to the pattern portion 23A detected as corresponding to the first template 21 and the second template 22 as shown in FIG. 9B, for example. Pattern portion arrangement information indicating the arrangement with the pattern portion 24A group is generated. The detection device 30 is detected by the second generation unit 38 as corresponding to the second template 22 and the pattern unit 25A detected as corresponding to the first template 21 as shown in FIG. Pattern portion arrangement information indicating the arrangement with the pattern portion 26A group is generated.

  In the detection device 30, the determination unit 39 compares the pattern portion arrangement information (step S27) generated by the second generation unit 38 with the template arrangement information (step S14) generated by the first generation unit 35 (step S14). Step S28), it is determined whether or not they match (step S29). The detection device 30 includes, for example, pattern portion arrangement information (pattern portion 23A and pattern portion 24A group) as shown in FIG. 9B and template arrangement information (pattern portion 21A and pattern portion as shown in FIG. 9A). 22A group) and a match determination is performed. The detection device 30 includes, for example, pattern portion arrangement information (pattern portion 25A and pattern portion 26A group) as shown in FIG. 11B and template arrangement information (pattern portion 21A and pattern portion as shown in FIG. 11A). 22A group) and a match determination is performed.

  The detection device 30 determines that the template arrangement information (FIG. 9A) and the pattern portion arrangement information (FIG. 9B) do not match, for example, as shown in FIGS. 9A and 9B. If so (step S29), the process returns to step S20.

  That is, in the case of not matching in step S29, the pattern portion 23A detected as corresponding to the first template 21 in the initial processing of steps S20 to S23 is the pattern portion 21A to be used for setting the reference position of the target wafer. This is the case. Alternatively, the pattern portion 24A group detected as corresponding to the second template 22 in the initial processing of steps S24 to S26 may not be the pattern portion 22A group to be used for setting the reference position of the target wafer. In such a case, the detection apparatus 30 moves the stage (step S20), and acquires the image 10A again for another region of the target wafer (step S21). Then, the pattern portion corresponding to the first template 21 is detected from the image 10A (steps S22 and S23), the pattern portion group corresponding to the second template 22 is detected (steps S24 to S26), and thereafter Similar processing is executed (steps S27 to S29).

  On the other hand, for example, as shown in FIGS. 11A and 11B, the detection apparatus 30 matches the template arrangement information (FIG. 11A) and the pattern portion arrangement information (FIG. 11B). If it is determined (step S29), the reference position of the target wafer is set (step S30).

  That is, in the case of matching in step S29, the pattern portion 25A detected as corresponding to the first template 21 in the initial processing in steps S20 to S23 is the pattern portion 21A to be used for setting the reference position of the target wafer. This is the case. In addition, the pattern portion 26A group detected as corresponding to the second template 22 in the initial processing of steps S24 to S26 is the pattern portion 22A group to be used for setting the reference position of the target wafer. The detection device 30 can detect any of the position of the pattern portion 25A that matches the first template 21 in the image 10A, the position inside the pattern portion 25A (the center and corner of the semiconductor element 11), and the pattern portion 26A group that matches the second template 21. One position or the like is set as a reference position of the target wafer.

  The detection device 30 stores programs and various data used in the processing by the storage unit 41. In addition, the detection device 30 inputs various conditions and data required for processing through the input unit 42. The detection device 30 displays various data obtained by the processing on the display device by the output unit 43 or transmits the data to another device.

  The detection device 30 performs image recognition on the target wafer by the processing as described above, and sets an appropriate reference position based on the result. By setting an appropriate reference position for the target wafer, the alignment can be performed with high accuracy, and mixing of non-defective and defective semiconductor elements 11 can be suppressed. Further, since it is not necessary to crush the semiconductor element 11 on the wafer, it is possible to suppress a decrease in yield of the semiconductor element, an increase in cost and a decrease in yield due to the introduction of a process for providing the mark. It becomes possible. Furthermore, the detection apparatus 30 that sets the reference position of the target wafer by the above processing can set an appropriate reference position without necessarily using a mechanism such as a laser interferometer stage. The increase can be suppressed.

  By the way, in the matrix division processing executed by the detection device 30 in steps S13 and S24, the images 10 and 10A are divided into matrices using the set size of the second template 22 as a unit division size. Here, such matrix division will be described.

  FIG. 15 is a diagram for explaining matrix division. FIG. 15A shows an example of layout information on a wafer, FIG. 15B shows an example of matrix division, and FIG. 15C shows another example of matrix division.

The size of the second template 22, which is a unit division size when the image 10 or the image 10A is divided into matrixes, is based on layout information of the semiconductor elements 11 on the wafer as shown in FIG. Set. The layout information includes information regarding the arrangement pitches P _X and P _Y of the semiconductor elements 11 including the scribe region 12 on the wafer as seen in the image 10 and the image 10A. The arrangement pitches P _X and P _Y are distances from an intermediate point of one scribe region 12 to an intermediate point of another scribe region 12 facing the one scribe region 12 across the semiconductor element 11.

For example, the size of the second template 22 is a value obtained by dividing the length S _X of the side in the X direction by the natural number of the arrangement pitch P _X in the X direction of the layout information, and the length S _Y of the side in the Y direction is The layout information is a value obtained by multiplying the arrangement pitch P _Y in the Y direction by a natural number. The size of the second template 22 set in this way, that is, the lengths S _X and S _Y are set as unit division sizes, and matrix division is performed. An example is shown in FIG. For example, when the arrangement pitch P _X = 6 mm and P _Y = 4 mm, the length S _X of the second template 22 is set to 3 mm which is a value obtained by dividing the arrangement pitch P _X by 2, and the length S _Y is A value obtained by multiplying the arrangement pitch P _Y by 2 is 8 mm. Matrix division is performed with such a size.

Alternatively, the size of the second template 22 is set such that the side length S _X in the X direction is a value obtained by multiplying the arrangement pitch P _X in the X direction of the layout information by a natural number, and the length S _Y of the side in the Y direction is set as the layout. A value obtained by dividing the arrangement pitch P _Y in the Y direction of information by a natural number. The size of the second template 22 set in this way, that is, the lengths S _X and S _Y are set as unit division sizes, and matrix division is performed. An example is shown in FIG. For example, when the arrangement pitch P _X = 6 mm and P _Y = 4 mm, the length S _X of the second template 22 is set to 6 mm, which is a value obtained by multiplying the arrangement pitch P _X by 1, and the length S _Y is The arrangement pitch P _{Y is set} to 2 mm, which is a value obtained by dividing by 2. Matrix division is performed with such a size.

  On the wafer, the semiconductor element 11 and the scribe region 12 are repeatedly arranged two-dimensionally. Therefore, by setting the size of the second template 22 that is the unit division size when the matrix is divided according to the certain rule as described above, the pattern portion corresponding to or matching the second template 22 can be obtained by image recognition. It becomes possible to detect with high accuracy.

Note that the arrangement of the pattern portion corresponding to the second template 22 and the matrix division of the wafer on the reference position detection target wafer are noted as follows.
FIGS. 16 to 18 are diagrams showing examples of arrangement of the second template corresponding pattern portions in the matrix-divided region.

  For example, as shown in FIG. 16A, when the matrix division of an image 50 is performed in two divisions in the X direction and two divisions in the Y direction, the diagonal direction of the 2 × 2 matrix division region In addition, it is assumed that the pattern portions Z corresponding to the second template 22 are continuously arranged. If the pattern of the image 50 (matrix division region) as shown in FIG. 16A is repeatedly arranged in the X direction and the Y direction on the reference position detection target wafer, the wafer pattern is: The arrangement is as shown in FIG. In FIG. 16B, the matrix division region 50b is disposed adjacent to the matrix division region 50a in the X direction, the matrix division region 50c is disposed adjacent to the Y direction, and is adjacent to the diagonal direction. The case where the matrix division | segmentation area | region 50d is arrange | positioned is shown.

  However, in the arrangement of FIG. 16B, for example, the following can occur when detecting the upper left pattern portion Z (Za) of the matrix division region 50a as the reference position. In other words, it may happen that the upper left pattern portion Z (Zb) of the matrix division region 50e (dotted line) where a part of the pattern portion Z overlaps and has the same pattern configuration is erroneously detected. When such a detection occurs, a reference position shift occurs in the target wafer, and proper alignment cannot be performed.

  This is not limited to the case as shown in FIG. 16. For example, when the pattern part Z continues in the X direction of the 2 × 3 matrix division region as in the image 51 of FIG. This is the same when the pattern portion Z continues in the Y direction. Further, when the pattern portion Z continues in the Y direction of the 3 × 2 matrix division region as in the image 52 of FIG. 17B, the same erroneous detection of the pattern portion Z may occur. The same applies when Z continues in the X direction. Further, as in the images 53, 54, and 55 in FIGS. 18A, 18B, and 18C, the X direction, the Y direction, and the diagonal direction of the 3 × 3 matrix division region. In the case where the pattern portion Z continues, the same erroneous detection of the pattern portion Z may occur.

  That is, the pattern portion Z is continuous in the X direction, the Y direction, and the diagonal direction of the matrix division region, and such a pattern of the matrix division region is repeatedly arranged in the X direction and the Y direction on the target wafer. In such a case, erroneous detection and deviation of the reference position of the target wafer may occur.

  The arrangement of the pattern portion Z on the target wafer should be set so that the pattern portions Z corresponding to the second template 22 are not arranged in a row in the X direction, the Y direction, and the diagonal direction. Is desirable. Further, if possible in terms of the pattern configuration of the pattern portion Z, the pattern portion Z is not arranged in a row in the X direction, the Y direction, and the diagonal direction by changing the number of matrix divisions. It is desirable to make it.

Next, a second embodiment will be described.
Here, a second example of the detection process will be described.
FIG. 19 is a diagram illustrating an example of detection processing according to the second embodiment.

  In the second embodiment, first, in accordance with the example of steps S10 and S12 to S14 described in FIG. 13, the setting of the second template 22 and the generation of the template arrangement information indicating the arrangement of the second template 22 are performed. Thereafter, the reference position of the target wafer is detected by a process as illustrated in FIG.

  The detection apparatus 30 moves the stage by the drive unit 40 (step S40), and the acquisition unit 31 performs an image 10A including a predetermined region of the reference position detection target wafer on the stage under relatively high magnification conditions (step S41). Is acquired (step S42). The detection device 30 performs image recognition using the second template 22 set by the second setting unit 33 on the high-magnification image 10A acquired by the second detection unit 37 (step S43). It is determined whether or not a pattern portion corresponding to the template 22 has been detected (step S44).

  FIG. 20 is a diagram for explaining the detection of the template corresponding pattern portion. Here, FIG. 20A is a diagram illustrating an example of a high-magnification image. FIG. 20B is a diagram illustrating an example of the second template.

  As shown in FIG. 20A, the acquired high-magnification image 10A (steps S40 to S42) includes the semiconductor element 11, the scribe region 12, and the marks 13a and 13b (13bb) of the target wafer. Image recognition using the second template 22 as shown in FIG. 20B is performed on such a high-magnification image 10A, and the corresponding pattern portion, in this example, one of the pattern portions 61 of the two marks 13b. Alternatively, it is determined whether or not the pattern portion 62 is detected (steps S43 and S44).

  When the second detection unit 37 determines that the pattern unit 61 or the pattern unit 62 corresponding to the second template 22 is not detected from the high-magnification image 10A (step S44), the detection device 30 returns to step S40. The subsequent processing is executed.

  When the detection device 30 determines that the pattern portion 61 or the pattern portion 62 corresponding to the second template 22 has been detected from the high-magnification image 10A (step S44), the image 10A is subjected to a relatively low magnification condition. (Step S45). The detection device 30 divides the image 10A, which has been changed to a low magnification by the dividing unit 34, into a matrix with the set size of the second template 22 (step S46), and the second detecting unit 37 for all the divided regions 14. Then, image recognition using the second template 22 is performed (steps S47 and S48).

  The detection device 30 executes processes such as steps S45 to S48, and the second detection unit 37 detects a pattern unit group corresponding to the second template 22 from the low-magnification image 10A divided into matrices. The detection device 30 generates the arrangement of the pattern part group detected by the second generation unit 38 as corresponding to the second template 22 by the second generation unit 38 as the pattern part arrangement information (step S49).

  The detection device 30 further includes a template arrangement indicating the arrangement of the second template 22 generated by the first generation unit 35 and the pattern portion arrangement information (step S49) generated by the second generation unit 38 by the determination unit 39. The information (step S14) is compared (step S50). Then, the detection device 30 determines whether or not the pattern portion arrangement information matches the arrangement of the pattern portion group of the template arrangement information (step S51). Further, it is determined whether or not the pattern part (steps S41 to S44) detected under the high magnification condition in the pattern part arrangement information matches the pattern part of the reference point in the template arrangement information (step S51).

  FIG. 21 is a diagram for explaining an example of pattern portion arrangement information and match determination using the pattern portion arrangement information. Here, FIG. 21A is a diagram illustrating an example of template arrangement information indicating the arrangement of the set second template. FIG. 21B is a diagram illustrating an example of pattern portion arrangement information indicating the arrangement of the pattern portion group corresponding to the second template obtained from the low-magnification image.

  The template arrangement information in FIG. 21A includes information indicating the arrangement of the three pattern portions 22A group (22Aa, 22Ab, 22Ac) aligned in the oblique direction that matches the second template 22. According to a certain reference point setting rule, the pattern portion 22Aa at the upper left of the group of three pattern portions 22A is set as a reference point, and the two pattern portions 22Ab and 22Ac are arranged side by side obliquely downward from there. Yes.

  The pattern part arrangement information in FIG. 21B includes information indicating the arrangement of three pattern parts 63 arranged in an oblique direction, corresponding to the second template 22, obtained from the low-magnification image 10A. .

  Now, in steps S41 to S44, the pattern portion 61 (FIG. 20A) is detected from the high-magnification image 10A, and the pattern portion 61 is a pattern portion 63 group (63a, 63b) indicated by the pattern portion arrangement information. , 63c), it is assumed that the upper left pattern portion 63a. The arrangement of the pattern portions 63a in the pattern portion 63 group matches the arrangement of the reference point pattern portions 22Aa in the pattern portion 22A group of the template arrangement information.

  The detection device 30 compares the pattern portion arrangement information (FIG. 21B) and the template arrangement information (FIG. 21A) (step S50). First, the arrangement of the pattern portion 63 group of the pattern portion arrangement information is compared with the pattern portion arrangement information. Then, it is determined whether the arrangement of the pattern portion 22A group of the template arrangement information matches (step S51). Then, after determining that they match, the arrangement of the pattern portion 63a (the pattern portion 61 in FIG. 20A) in the pattern portion 63 group of the pattern portion arrangement information is the pattern of the reference points in the pattern portion 22A group of the template arrangement information. It is determined whether it matches the arrangement of the part 22Aa (step S51).

  In the example of FIG. 21, since both match, the detection apparatus 30 uses the pattern portion 63a, that is, the pattern portion 61 detected from the high-magnification image 10A as a reference point, and the reference position of the target wafer based on this reference point. Is set (step S52). For example, the detection apparatus 30 sets the pattern unit 63a and the pattern unit 61 to the reference position of the target wafer. As a result, the detection process ends.

On the other hand, if the detection device 30 does not determine that they match in step S51, the detection device 30 performs the following processing.
FIG. 22 is a diagram for explaining another example of pattern portion arrangement information and match determination using the pattern portion arrangement information. Here, FIG. 22A is a diagram illustrating an example of template arrangement information indicating the arrangement of the set second template. FIG. 22B is a diagram illustrating an example of pattern portion arrangement information indicating the arrangement of the pattern portion group corresponding to the second template, which is obtained from the low-magnification image. FIG. 22C is a diagram illustrating an example of deviation correction.

22A and 22B show examples of template arrangement information and pattern portion arrangement information, respectively, as in FIGS. 21A and 21B.
The template arrangement information in FIG. 22A includes information indicating the arrangement of the three pattern portions 22A group (22Aa, 22Ab, 22Ac) aligned in the oblique direction that matches the second template 22, and the upper left pattern. The portion 22Aa is set as a reference point.

  The pattern portion arrangement information in FIG. 22B includes information indicating the arrangement of the three pattern portion 63 groups arranged in the diagonal direction corresponding to the second template 22 obtained from the low-magnification image 10A. .

  In steps S41 to S44, the pattern portion 62 (FIG. 20A) is detected from the high-magnification image 10A, and the pattern portion 62 is a pattern portion 63 group (63a, 63b, 63c) indicated by the pattern portion arrangement information. ) Is the central pattern portion 63b. The arrangement of the pattern parts 63b in the pattern part 63 group does not match the arrangement of the reference point pattern parts 22Aa in the pattern part 22A group of the template arrangement information.

  The detection device 30 compares the pattern portion arrangement information (FIG. 22B) and the template arrangement information (FIG. 22A) (step S50). First, the arrangement of the pattern portion 63 group of the pattern portion arrangement information is compared with the pattern portion arrangement information. Then, it is determined whether the arrangement of the pattern portion 22A group of the template arrangement information matches (step S51). Then, after determining that they match, the arrangement of the pattern part 63b in the pattern part arrangement information corresponding to the pattern part 62 (FIG. 20A) detected from the high-magnification image 10A is the reference point in the template arrangement information. It is determined whether or not the pattern portion 22Aa matches the arrangement (step S51).

  In the example of FIG. 22, the arrangement of the pattern part 63b in the pattern part arrangement information does not match the arrangement of the reference point pattern part 22Aa in the template arrangement information. Therefore, it can be seen that the pattern portion 63b, that is, the pattern portion 62 detected from the high-magnification image 10A should not be detected as the reference point. The pattern portion 63a in the pattern portion arrangement information matches the arrangement of the reference point pattern portion 22Aa in the template arrangement information, and is the pattern portion 61 in the high-magnification image 10A.

  As illustrated by an arrow in FIG. 22C, the detection device 30 corrects a deviation between the pattern portion 63b and the pattern portion 63a that should have been detected as the reference point (step S53), and the pattern portion. A reference point is set at the position 63a. In this way, the detection apparatus 30 sets the reference point at the position of the pattern unit 63a, the position of the pattern unit 61 in the high-magnification image 10A, and sets the reference position of the target wafer based on the reference point (step S52). ). As a result, the detection process ends.

  In the second embodiment, the reference position of the target wafer is appropriately set by using one type of template, and in the above example, the second template 22. Specifically, first, the pattern portion 61 or the pattern portion 62 corresponding to the second template 22 is detected from the high-magnification image 10A of the target wafer, and then the image 10A is changed to the low magnification to change the second template 22. The pattern unit 63 group corresponding to is detected. A template in which the arrangement of the detected pattern portion 63 group and one pattern portion in the pattern portion 63 group, that is, the pattern portion 61 or the pattern portion 62 detected in the high-magnification image 10A, is set for the second template 22. A match is determined by comparison with the arrangement of the arrangement information. In the second embodiment, the reference position of the target wafer is appropriately set by such a method.

  By appropriately setting the reference position of the target wafer, the alignment can be performed with high accuracy, and mixing of non-defective and defective semiconductor elements 11 can be suppressed. Further, since it is not necessary to crush the semiconductor element 11 on the wafer, it is possible to suppress a decrease in yield of the semiconductor element, an increase in cost and a decrease in yield due to the introduction of a process for providing the mark. It becomes possible. Furthermore, even if a mechanism such as a laser interferometer stage or the like is not necessarily used, an appropriate reference position can be set, so that an increase in apparatus cost can be suppressed.

  When using the method described in the second embodiment, the first setting unit 32 for setting the first template 21 as shown in FIG. 12 and the pattern unit corresponding to the first template 21 are used. The detection device 30 that omits the processing function of the first detection unit 36 that detects the above may be used.

Further, after the setting of the reference position of the target wafer in step S52, the following processing as shown in FIG. 23 may be further performed.
FIG. 23 is a diagram illustrating another example of the detection process according to the second embodiment.

  In this example, after setting the reference position of the target wafer as described above (step S52), as shown in FIG. 23, the detection apparatus 30 first changes the magnification of the image 10A on which the reference position is set to a high magnification again. (Step S54). Thereby, for example, an image 10A having a magnification as shown in FIG. In the detection device 30, the second detection unit 37 performs image recognition using the set second template 22 on the high-magnification image 10A (step S55). The detection device 30 detects the pattern portion 61 and the pattern portion 62 corresponding to the second template 22 by image recognition on the high-magnification image 10A.

  The detection device 30 determines whether or not the pattern shapes of the pattern portion 61 and the pattern portion 62 detected as corresponding to the second template 22 match the pattern shapes of the pattern portion 22Aa and the pattern portion 22Ab included in the template arrangement information. Is determined (step S56). The detection apparatus 30 further includes the arrangement (coordinates) of the pattern unit 61 and the pattern unit 62 detected as corresponding to the second template 22 and the arrangement (coordinates) of the pattern unit 22Aa and the pattern unit 22Ab included in the template arrangement information. (Step S57).

If the detection device 30 determines that both the pattern shape and the coordinates match (steps S56 and S57), the detection process ends.
When the detection device 30 determines that the pattern shapes do not match (step S56) or determines that the coordinates do not match (step S57), the information indicating that an error has occurred is displayed by means such as display. Output (step S58), and the detection process ends.

By performing the processing as shown in FIG. 23, it is possible to set an appropriate reference position of the target wafer with high accuracy.
Next, a third embodiment will be described.

Here, a third example of the detection process will be described.
FIG. 24 is a diagram illustrating an example of detection processing according to the third embodiment.
In the third embodiment, as in the second embodiment, first, the setting of the second template 22 and the arrangement of the second template 22 are performed according to the example of steps S10 and S12 to S14 described in FIG. The template arrangement information indicating is generated. Thereafter, the reference position of the target wafer is detected by a process as illustrated in FIG.

  The detection device 30 moves the stage by the drive unit 40 (step S60), and the acquisition unit 31 performs an image 10A including a predetermined region of the reference position detection target wafer on the stage under relatively low magnification conditions (step S61). Is acquired (step S62). In the detection device 30, the dividing unit 34 divides the acquired low-magnification image 10A into a matrix with the set size of the second template 22 (step S63). And the detection apparatus 30 performs image recognition using the 2nd template 22 by the 2nd detection part 37 about all the division | segmentation area | regions 14 of the image 10A of the low magnification divided by the matrix (step S64, S65). The detection device 30 determines whether or not the pattern portion group corresponding to the second template 22 is detected by the image recognition (step S66).

  When the second detection unit 37 determines that the pattern unit group corresponding to the second template 22 is not detected from the low-magnification image 10A obtained by dividing the matrix (step S66), the detection device 30 returns to step S60. The subsequent processing is executed.

  For example, the pattern portion 63 group (63a, 63b, 63c) corresponding to the second template 22 as shown in the pattern portion arrangement information in FIG. Suppose. In this case, the detection apparatus 30 causes the second generation unit 38 to generate pattern part arrangement information indicating the arrangement of the pattern part 63 group. One of the pattern parts 63 group, for example, the upper left pattern part 63a is set as a reference point.

  When it is determined that such a pattern unit 63 group has been detected (step S66), the detection device 30 changes the image 10A to a relatively high magnification condition (step S67). Thereby, for example, an image 10A having a magnification as shown in FIG.

  The detection device 30 performs image recognition using the set second template 22 on the image 10A changed to the high magnification by the second detection unit 37 (step S68). The detection device 30 detects the pattern portion corresponding to the second template 22, in the example of FIG. 20A, the pattern portion 61 and the pattern portion 62 by image recognition on the high-magnification image 10 </ b> A. Of the detected pattern portion 61 and pattern portion 62, the pattern portion 61 corresponds to the reference point pattern portion 63a of the pattern portion 63 group obtained from the low-magnification image 10A.

  The detection device 30 determines whether or not the pattern shapes of the pattern portion 61 and the pattern portion 62 detected as corresponding to the second template 22 match the pattern shapes of the pattern portion 22Aa and the pattern portion 22Ab included in the template arrangement information. Is determined (step S69). The detection apparatus 30 further includes the arrangement (coordinates) of the pattern unit 61 and the pattern unit 62 detected as corresponding to the second template 22 and the arrangement (coordinates) of the pattern unit 22Aa and the pattern unit 22Ab included in the template arrangement information. (Step S70).

  When the detection device 30 determines that the pattern shape and the coordinates are both matched (steps S69 and S70), the pattern unit 61 of the high-magnification image 10A (the pattern unit 63a in the low-magnification image 10A) is set as a reference point. Then, based on the reference point, the reference position of the target wafer is set (step S71). As a result, the detection process ends.

  When the detection device 30 determines that the pattern shapes do not match (step S69) or determines that the coordinates do not match (step S70), the information indicating that an error has occurred is displayed by means such as display. Output (step S72), and the detection process ends.

  Also in the third embodiment, similarly to the second embodiment, the reference position of the target wafer is appropriately set using one type of template, in the above example, the second template 22. Specifically, first, the pattern portion 63 group corresponding to the second template 22 is detected from the low-magnification image 10A of the target wafer, and then the image 10A is changed to a high magnification to correspond to the second template 22. The pattern part 61 and the pattern part 62 are detected. Then, the arrangement of the pattern portion 63 group, the pattern portion 61 and the pattern portion 62, the arrangement of the reference point pattern portion 63a or the pattern portion 61 is compared with the arrangement of the template arrangement information set for the second template 22, and a match is determined. In the third embodiment, the reference position of the target wafer is appropriately set by such a method.

  By appropriately setting the reference position of the target wafer, the alignment can be performed with high accuracy, and mixing of non-defective and defective semiconductor elements 11 can be suppressed. Further, since it is not necessary to crush the semiconductor element 11 on the wafer, it is possible to suppress a decrease in yield of the semiconductor element, an increase in cost and a decrease in yield due to the introduction of a process for providing the mark. It becomes possible. Furthermore, since an appropriate reference position can be set without necessarily using a mechanism such as a laser interferometer stage, an increase in apparatus cost can be suppressed.

  When using the method described in the third embodiment, the first setting unit 32 for setting the first template 21 as shown in FIG. 12 and the pattern unit corresponding to the first template 21 are used. The detection device 30 that omits the processing function of the first detection unit 36 that detects the above may be used.

The processing function of the detection apparatus 30 that can be used in the methods described in the first to third embodiments above can be realized using a computer.
FIG. 25 is a diagram illustrating a configuration example of computer hardware.

  The computer 300 is controlled by the processor 301. The processor 301 is connected to the RAM 302 and a plurality of peripheral devices via a bus 309. The processor 301 may be a multiprocessor. The processor 301 is, for example, a central processing unit (CPU), a micro processing unit (MPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), or a programmable logic device (PLD). Further, the processor 301 may be a combination of two or more elements among a CPU, MPU, DSP, ASIC, and PLD.

  The RAM 302 is used as a main storage device of the computer 300. The RAM 302 temporarily stores at least part of an OS (Operating System) program and application programs to be executed by the processor 301. The RAM 302 stores various data necessary for processing by the processor 301.

  Peripheral devices connected to the bus 309 include an HDD (Hard Disk Drive) 303, a graphic processing device 304, an input interface 305, an optical drive device 306, a device connection interface 307, and a network interface 308.

  The HDD 303 magnetically writes data to and reads data from a built-in disk. The HDD 303 is used as an auxiliary storage device for the computer 300. The HDD 303 stores an OS program, application programs, and various data. A semiconductor storage device such as a flash memory can be used as the auxiliary storage device.

  A monitor 311 is connected to the graphic processing device 304. The graphic processing device 304 displays an image on the screen of the monitor 311 in accordance with an instruction from the processor 301. Examples of the monitor 311 include a display device using a CRT (Cathode Ray Tube) and a liquid crystal display device.

  A keyboard 312 and a mouse 313 are connected to the input interface 305. The input interface 305 transmits a signal transmitted from the keyboard 312 or the mouse 313 to the processor 301. The mouse 313 is an example of a pointing device, and other pointing devices can be used. Examples of other pointing devices include a touch panel, a tablet, a touch pad, and a trackball.

  The optical drive device 306 reads data recorded on the optical disk 314 using laser light or the like. The optical disk 314 is a portable recording medium on which data is recorded so that it can be read by reflection of light. The optical disk 314 includes a DVD (Digital Versatile Disc), a DVD-RAM, a CD-ROM (Compact Disc Read Only Memory), a CD-R (Recordable) / RW (ReWritable), and the like.

  The device connection interface 307 is a communication interface for connecting peripheral devices to the computer 300. For example, a memory device 315 or a memory reader / writer 316 can be connected to the device connection interface 307. The memory device 315 is a recording medium equipped with a communication function with the device connection interface 307. The memory reader / writer 316 is a device that writes data to the memory card 317 or reads data from the memory card 317. The memory card 317 is a card type recording medium.

  The network interface 308 is connected to the network 310. The network interface 308 transmits and receives data to and from other computers or communication devices via the network 310.

With the hardware configuration described above, the processing function of the detection device 30 can be realized.
The computer 300 realizes the processing function of the detection device 30 by executing a program recorded on a computer-readable recording medium, for example. A program describing the processing contents to be executed by the computer 300 can be recorded in various recording media. For example, a program to be executed by the computer 300 can be stored in the HDD 303. The processor 301 loads at least a part of the program in the HDD 303 into the RAM 302 and executes the program. In addition, a program to be executed by the computer 300 can be recorded on a portable recording medium such as the optical disk 314, the memory device 315, and the memory card 317. The program stored in the portable recording medium becomes executable after being installed in the HDD 303 under the control of the processor 301, for example. Further, the processor 301 can also read and execute the program directly from the portable recording medium.

  Further, the detection apparatus 30 as described above is used in semiconductor manufacturing such as a test apparatus that performs an electrical test of a wafer on which a semiconductor element is formed, an inspection apparatus that performs an appearance inspection, and a film forming apparatus and an etching apparatus. It can be employed in various manufacturing apparatuses.

  In the above description, the detection and setting of the reference position of the wafer is taken as an example. However, the method as described in the first to third embodiments is not limited to the wafer, but the reference position of various substrates. It can be used for detection and setting.

DESCRIPTION OF SYMBOLS 1 Semiconductor substrate 2 Insulating layer 3 Protective layer 4 Device part 5 Pad part 6 Scribe pattern part 10,10A, 50,51,52,53,54,55,100 Image 11,110 Semiconductor element 12,120 Scribe area | region 13,13a , 13b, 13bb, 130, 131, 132 Marks 14, 14a, 14b, 14c, 14d, 14e, 14f Divided regions 20a, 20b, 20Aa, 20Ab, 20Ac, 20Ad, 20Ae, 20Af, 200a, 210, 220 Frame 21 1 template 22 2nd template 22a, 22b, 22c, 21A, 22A, 22Aa, 22Ab, 22Ac, 23A, 24A, 24Aa, 24Ab, 24Ac, 25A, 26A, 26Aa, 26Ab, 26Ac, 61, 62, 63, 63a, 63b, 63c Unit 30 detecting device 31 obtaining unit 32 first setting unit 33 second setting unit 34 dividing unit 35 first generation unit 36 first detection unit 37 second detection unit 38 second generation unit 39 determination unit 40 drive unit 41 storage Unit 42 Input unit 43 Output unit 50a, 50b, 50c, 50d, 50e Matrix division region 200 Template 300 Computer 301 Processor 302 RAM
303 HDD
304 graphic processing device 305 input interface 306 optical drive device 307 device connection interface 308 network interface 309 bus 310 network 311 monitor 312 keyboard 313 mouse 314 optical disk 315 memory device 316 memory reader / writer 317 memory card

Claims (6)

Detecting a first pattern portion corresponding to a set first template by image recognition from a first image of a first substrate having a pattern;
Dividing the first image by a first size into a matrix;
Detecting a second pattern portion corresponding to a second template set at the first size by image recognition from the matrix-divided first image;
Determining whether or not the first arrangement of the first pattern portion and the second pattern portion in the first image matches the set second arrangement. Using the second image of the second substrate having the pattern to set the first template;
Dividing the second image by the first size in a matrix;
Setting the second template of the first size using the matrix-divided second image;
Generating information indicating an arrangement of a third pattern portion matching the first template and a fourth pattern portion matching the second template in the second image, and setting the second arrangement as the second arrangement; The detection method according to claim 1, further comprising: The second pattern portion includes a plurality of pattern portions corresponding to the second template,
The detection method according to claim 2, wherein the fourth pattern portion includes a plurality of pattern portions that match the second template.   4. The detection method according to claim 1, wherein at the time of image recognition for detecting the second pattern portion, the field-of-view magnification is set lower than at the time of image recognition for detecting the first pattern portion. A first detector for detecting a first pattern corresponding to the set first template by image recognition from a first image of a first substrate having a pattern;
A dividing unit for dividing the first image by a first size in a matrix;
A second detector for detecting a second pattern corresponding to the second template set at the first size by image recognition from the matrix-divided first image;
And a determination unit that determines whether or not the first arrangement of the first pattern portion and the second pattern portion in the first image matches the set second arrangement. On the computer,
A first pattern portion corresponding to the set first template is detected by image recognition from a first image of a substrate having a pattern;
Dividing the first image by a first size in a matrix;
A second pattern portion corresponding to a second template set in the first size is detected by image recognition from the matrix-divided first image;
A detection program for executing a process of determining whether or not the first arrangement of the first pattern portion and the second pattern portion in the first image matches a set second arrangement.

Images