JPH02290040A - Alignment system - Google Patents

Abstract

PURPOSE:To contrive to accomplish high-speed alignment by a method wherein alignment adjustment is conducted by sensing the surface of the material to be processed by three or more image sensing means. CONSTITUTION:A lattice-like pattern and the like, with which an orientation flat 3 and a chip will be formed on the frame 5 to be attached to a chuck 10 through the intermediary of a film 6, is provided and a wafer 1, which is the material to be processed by a blade 11 and the like, is prealigned and set utilizing a flat 3 and the guide part, etc., on both ends of a reference end part 7 and also using a bonding agent and the like. Subsequently, the surface of the wafer 1 is sensed at least by means of three or more cameras such as low magnification cameras 17 and 18, and high magnification cameras 19, 20 and the like of an optical means 16, the result of the image sensing is used in different ways; a precise alignment is conducted subsequent to a coarse alignment, and the adjustment of optical alignment is conducted at high speed without changing magnification of the cameras used.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to a so-called alignment system that accurately aligns a workpiece, such as a semiconductor wafer 8, having a predetermined circuit pattern on its surface to a predetermined position for processing or cutting.

[Prior Art 1] As this type of alignment means,
JP-A-6 0-2 4 4. The inventions disclosed in Japanese Patent Application Laid-open No. 803 and JP-A-61-14820 are well known as conventional examples. In each of the above-mentioned inventions, a holding means for placing the workpiece and two cameras (
a microscope), a means for optically switching the camera between low magnification and high magnification, an imaging means, a magnification converting means, and an A/
It has a D conversion means, an image frame memory, a key pattern memory, a pattern matching means, a display means, and a central processing system, the central processing system is a system for controlling each of the above means, and has two cameras. The captured image is displayed on a display means, a specific area is designated on the display screen and a key is pressed, pattern matching is performed based on the key pattern, and positioning is performed by pattern matching. And, the invention of ■ is [X, Y
This is intended to clarify the positional relationship of key patterns on the axis coordinates and to smoothly perform X-axis alignment and Y-axis alignment. [Problems to be Solved by the Invention] In the conventional example described above, since the alignment system is composed of two cameras, the movement of the wafer is reduced compared to the alignment system composed of one camera, and the alignment can be improved. This allows for faster speeds. However, at least one camera must have both low magnification and high magnification functions, which is optically complex. In addition, if the images taken at low magnification and the images taken at high magnification are different, the wafer must be moved when switching from low magnification to high magnification, making it difficult to achieve faster alignment. There are many unresolved issues.

[Means to solve the problem]

As a specific means for solving the problems of the conventional example, the present invention provides a holding means for holding a workpiece, an optical means for conveying an image on the surface of the workpiece held by the holding means, and the holding means. An alignment system comprising an operating element that acts on a workpiece held in a machine and performs processing, the optical means comprising a camera means of 3g or more. The optical means is composed of three or four camera means, and one camera means or one camera means is configured to take an image of the surface of the workpiece held by the holding means at a relatively low magnification. An optical means is constituted by two low-magnification camera means and two high-magnification camera means for imaging the surface of the workpiece at a relatively high magnification, and the workpiece is held by the holding means. with respect to the working element, the coarse alignment is performed by the low magnification camera means, and then the fine alignment is performed by the high magnification camera means, so that the coarse alignment in the so-called alignment system is performed. Alignment and precision alignment are automatically performed by pattern matching means, and the distance between the two high-magnification camera means is variable depending on the type of workpiece, that is, the chip size. It is an alignment system that can

[Example 1] Next, the present invention will be explained in more detail with reference to an illustrated example. As shown in FIGS. 1 and 2, a workpiece 1 requiring alignment 1 is, for example, A semiconductor wafer 8 is provided with a lattice-like pattern 2 on the surface of the wafer 1, and the lattice-like pattern 2 is arranged in a horizontal direction substantially parallel to the orientation flat 3 of the wafer 1. It is composed of a plurality of linear regions 2a in the direction and a plurality of substantially vertical linear regions 2b in the vertical direction, and both of these linear regions are usually called streets, have a predetermined width W, and have a predetermined width W. are formed with an interval d of . In this case, the width W of the straight region is generally about several tens of μm, and the distance d between the horizontal straight region 2a and the vertical straight region 2b does not necessarily have to be the same. There are a plurality of rectangular areas 4 of substantially the same shape (this area becomes a chip, hereinafter referred to as a chip) separated by these vertical and horizontal straight line areas, and each of these rectangular areas 4 has The same circuit pattern is applied. The workpiece 1 configured as described above, that is, a semiconductor wafer, is mounted on a predetermined frame 5 and subjected to processing. In this case, the mounting on the frame 5 is performed by placing the mounting on the frame 5 approximately at the center of the frame by adhesive means or the like through the film material 6, and when mounting the frame 5, the reference end 7 serving as a reference for the Bria alignment in the frame 5 and the above-mentioned It is positioned and mounted so that the orientation flat 3 is substantially parallel to the orientation flat 3. The frame 5 is provided with cutout portions or guide portions 8 and 9 for rear alignment at the reference end portion 7.
on both sides. In this way, the workpiece 1 is mounted on a predetermined frame 5 and placed on a holding means 10 such as a chuck table for processing or machining, and while being held by the holding means, for example, a diamond Cutting is performed by an active element such as a rotating blade 11 made of abrasive grains. This holding means is provided with a reduced pressure suction means so that the mounted frame 5 is stably held, and the frame 5 is immediately suction-held in the state where the frame 5 is supplied and placed by an appropriate supply means. is configured to do so. When the frame 5 is supplied, the guide parts 8, 9 and the reference end part 7 provided on the frame 5 are used to pre-align the frame 5, which is pre-aligned when the frame 5 is supplied. In other words, in the cutting process by the blade 11, the angular error between the straight line area 2a in the horizontal direction with respect to the X-axis direction and the straight line area 2b in the vertical direction with respect to the Y-axis direction, which is a predetermined reference, is ±1. .5～
3. It is placed within a range of about 0°. The holding means 10 must correct the above-mentioned angular errors and adjust the alignment with the cutting line, that is, the straight line area, by the blade 11. For this purpose, the holding means 10 is supported by a suitable support mechanism 12, and the support mechanism 1
2 is the drive for changing the angle in the rotation direction θ? ! ! 13, a drive source 14 for moving in the X-axis direction, and a drive source 15 for moving in the Y-axis direction. For example, a pulse motor is used as each of these driving sources. An optical means 16 is arranged on the upper part of the holding means 10 as a detection means, ie, an alignment means, for correcting the angular error and aligning the straight line area with respect to the cutting line. In the first embodiment, this optical means includes a pair of low magnification (macro) cameras 17. 18 and
A pair of high magnification (micro) cameras 19. This system employs a so-called four-eye alignment system in which the two eyes are installed in close proximity to each other. As shown in Fig. 3, each pair of cameras is installed parallel to the X-axis direction at a predetermined interval L (L = about 40 to 55 seconds).
furthermore, the predetermined spacing is adjusted by adjusting means 2.
1 so that it can be adjusted as appropriate. The low magnification is set to approximately 1 to 5 times, and the high magnification is set to approximately 10 to 30 times, such that either pair of cameras photographs the surface of the workpiece 1 placed on the holding means 10. has become,
It is possible to switch from low magnification to high magnification or vice versa as appropriate. As shown in the block diagram shown in FIG. 4, the images captured by the cameras are processed through appropriate optical paths, and then displayed as a monitor so that they can be viewed on a display means such as a television screen equipped with a cathode ray tube (described later). Is displayed. That is, each camera 17 to 20 is equipped with a capacitively coupled device (COD).
) and the like are connected to each other, and electrical signals (low-magnification electrical signals and high-magnification electrical signals) outputted from these imaging means are connected to a video channel selector 26. Then, either a low magnification electrical signal or a high magnification electrical signal is selected, and the selected electrical signal is connected to the central processing unit 30 via, for example, an 8-bit A/D converter 28 and an image frame memory 29. has been done. In this case, the A/D converter 28 and the image frame memory 29 may be provided separately for the low magnification electrical signal and the high magnification electrical signal in the video channel selector 26, respectively. The central processing unit 30 is connected with a key pattern memory 31, a pattern matching means 32, and a display 33 such as a television screen equipped with a cathode ray tube, and further drives the drive sources 13 to 15 of the holding means 10. A signal line 34 for driving the optical means 16 and a signal line 35 for driving the optical means 16 are provided. Image processing will be explained using the block diagram shown in FIG. 4. A low magnification camera 17. 18, the state of the surface is photographed as a split image at a magnification of 1 to 5 times. The images captured by these cameras are sent to the lm means 22. 23 outputs the brightness and darkness of the image as analog signals, and the analog signals are converted to digital signals by the A/D converter 28 via the video channel selector 26 and output, and are temporarily stored in the image frame memory 29. Ru. In this case,
The imaging means 22. 23 Ha, each COD, i.e. x-
A plurality of image sensors arranged in a y matrix (for example,
256 x 256 elements), each image removal element outputs a voltage according to the density of the image as an analog signal, and this analog signal is sent to the video channel selector.
The signal is inputted to the A/D converter 28 via the A/D converter 26, where it is converted into an 8-bit multivalued digital signal, output, and temporarily stored in the image frame memory 29. Therefore, also in this image frame memory 29, both the image pickup means 22. 2 corresponding to 23 image sensors
It has a built-in RAM that can store 56 x 256 x 8-bit multi-value digital signals. In addition, a high magnification camera 19. The IC image captured by the imaging means 20 is also captured by the imaging means 24 in the same manner as described above. 25 outputs an analog signal based on its brightness, and outputs an analog signal from A through the video channel selector 26.
The signal is converted into a digital signal by the /D converter 28 and temporarily stored in the image frame memory 29 at high magnification. The central processing system 30 uses, for example, a microprocessor incorporating a plurality of RAMs, and the microprocessor is equipped with an operating section 30a. The image corresponding to the image or the signal stored in the key pattern memory 31 can be selectively displayed on the display means 33, and the pattern matching means 32
Moreover, each of the drive sources 13 to 15 of the holding means 10 can be selectively driven. The key pattern memory 31 has a built-in RAM, and a low magnification camera 17. a pattern in at least one specific area on the substrate captured in step 18, that is, a low magnification key pattern and a signal indicating the key pattern position;
In addition, patterns captured by the high-magnification cameras 19 and 20 falling in at least one specific area on the wafer, that is, high-magnification key patterns and signals indicating the positions of the key patterns are respectively stored. The storage input means in this case will be explained. Each of the cameras used has a central axis passing through its center, that is, a hairline a, and the hairline a is displayed at the center of the front display means 33 as a kind of reference line. First, as shown in FIGS. 5 to 9, the dummy wafer 1a mounted on the frame 5 is placed on the holding means 10 at a predetermined position, and the approximately central portion of the dummy wafer 1a is cut along the X axis, cutting line b. to be engraved. The cutting line b along the X-axis engraved in this way becomes a reference line for cutting the wafer. Although the dummy wafer 1a is not equipped with various electronic components and circuits, it has the same size and shape as the wafer 1 to be cut, and is mounted on a frame using the same supply means as the wafer 1. This is what was installed on the 5. The mounting means on the holding means 10 is also mounted by the same means. Therefore, in order to use this cutting line b as a reference line, since the cutting direction of the blade 11 is set, the position of the camera for alignment must be set in accordance with this. Therefore, cutting line b is blade 1
The low magnification camera 17. is set in a direction parallel to the cutting direction of 1. The surface of the dummy wafer 1a is photographed by the imaging means 22. 23, a video channel selector 26, an A/D converter 28, and an image frame memory 29, the central processing unit 30 displays the image on a display means 33. At this time, blade 1
If the set position of the camera is misaligned with respect to the cutting direction 1, as shown in FIG. lea are displayed in a relatively shifted state on the display means 33, and the hairline a obtained by both cameras is also shifted. This shifted state can be adjusted by operating the operating section 30a of the central processing unit 30 to adjust the low magnification camera 17. Adjust the setting position of 18, and as shown in Figure 8, image 1 on the right
The hairline a on 7a and the hairline a on the left image 18a are made to form a single line on the screen of the display means 33, and the cutting line b on both images is made to match the hairline a on the screen. Camera 17, i
Adjust o. Subsequently, the table 10 is moved a required distance in the Y-axis direction so that the cutting line b is directly below the high-magnification camera. And high magnification camera 19. At step 20, the same location as above is photographed, and similarly to the above, it is sent to the central processing unit 30 via the imaging means 24, 25, the video channel selector 26, the A/D converter 28, and the image frame memory 29.
The image is displayed on the display means 33, and if the cutting line b on the right image and the cutting line b on the left image in the display 33 are misaligned, the central processing unit 30 displays the image in the same manner as described above. The user operates the operation unit 30a to adjust the setting position of the camera. In most cases, positioning the low magnification camera will also allow you to position the high magnification camera. Suppose that a high magnification camera 19. In the image taken in step 20, if both the hairline a and the cutting line b are out of alignment as shown in FIG. Match. Normally, the positions of the hairline "a" and the cutting line "b" are adjusted using a low-magnification camera, so in most cases, angular deviations like those shown in Figure 7 are not observed. As shown in Figure 9,
Positional deviations in the vertical direction may be observed. In this case, the operating section 30a of the central processing unit 30 is operated to adjust the setting position of the camera, so that the hair line a on the display means 33 and the cutting line b match as shown in FIG. Adjust accordingly. Next, a sample wafer 1b, which will become the workpiece 1 on which various electronic components and circuits are mounted, is placed on the holding means 10 after being aligned by a predetermined supply means (same as the actual supply means of the workpiece). and low magnification camera 17. 18
Photograph the surface with am means 22. 23, video channel selector 26, A/D converter 28, and image frame memory 29, the central processing unit 30 displays the image on display means 33. At this time, approximately the center of the sample wafer 1b is photographed and displayed, but the sample wafer 1b is not placed in the cutting direction of the blade 11 (the If the positions are misaligned, the image 17a captured by the right low magnification camera 17 and the image 18a captured by the left low magnification camera 18 will be misaligned on the display screen of the display means 33, as shown in FIG. displayed. This shifted state is controlled by operating the operating section 30a of the central processing unit 30, and driving sources 13 to 15 of the holding means 10 are moved so that the straight line area 2a in the center is located on the hairline a in the 1,000-stage display 33. is driven and adjusted to match both images to the state shown in FIG. By aligning in this way, the cutting line b by the blade 11 and the straight line area 2a at the center substantially accurately match. After this alignment is completed, manually move the cursor C on the screen to a specific area that will become a key pattern for alignment ( Specify feature points) spot-wise. In this case, the display screen is the image 17a on the right.
and image 18a on the left, and image 17 on the right.
For example, a specific area 2r serving as a key pattern is specified on the image 18a on the left, and a specific area 2p is specified using the key pattern 0 on the left image 18a. Each of these designated specific areas has a characteristic as a key pattern, and the pixels within the range designated by the cursor C are, for example, each of the imaging means 22. 23 pixels (256 x 25 each
6), the pixels corresponding to the range of 32×32=1024, and the multi-valued digital signals of the low magnification key pattern received from these pixels are stored in the key pattern memory 31, respectively, and at the same time, the keys Each specific area 2r. Signals specifying the 2β position, that is, the position on the X-axis and Y-axis coordinates, are also stored. In this case, in the image 17a on the right and the image 18a on the left, the designated specific areas are areas each having remarkable characteristics, and even if they have different patterns from each other,
Alternatively, they may have the same pattern. The specification of the specific area at this low magnification is completed! After c, when the camera is switched from low magnification to high magnification, a display screen as shown in FIG. 12 is displayed on the display means 33. In this case, with the alignment using the previous low magnification,
Although the image 19b on the right and the image 20b on the left are displayed almost aligned, they may still be displayed slightly shifted. Therefore, the left and right images are aligned by looking at the display screen, and the force-sol C is moved on the display screen in the same manner as described above, so that the right image 19b and the left image 20b have remarkable characteristics, respectively. Specify the area as a key pattern in the same way as above, and specify each specific area 2.
R. Specify 2L. The multivalued digital signals of the high magnification key pattern received from these specified areas and the signal specifying the position on the X-Y coordinates are stored in the key pattern memory 3.
1 is stored. In addition, the specific area 2r that becomes the key pattern
, 2J2, 2R, and 2L, the selection is done manually, but the key pattern automatic setting means disclosed in Japanese Unexamined Patent Application Publication No. 61-204716 by the same applicant automatically selects a specific area as a key pattern. You may choose. Next, the holding means 10 is rotated by 90 degrees so that the holding means 10
A specific area (key pattern) for alignment for cutting in the Y-axis direction is selected. A high magnification display screen as shown in FIG. 13 is displayed on the display means 33. In this case, since the X-axis alignment has already been done, low-magnification alignment is not necessary. The image 19b on the right and the image 20b on the left are displayed almost aligned, but if they are still displayed slightly shifted, they are aligned and adjusted in the same way as described above. Then, by moving the cursor C on the display screen, between the right image 19b and the left image 20b,
The specific areas 2R-1 are selected using areas each having remarkable characteristics as key patterns. 2L is selected, and the multi-level digital signal of the high magnification key pattern received from these selected areas and the X-Y
A signal specifying the position on the coordinates and a signal specifying the position on the coordinates are respectively stored in the key pattern memory 31. With the above steps, preparations for performing automatic alignment are completed. Next, the pattern matching means 32 will be explained. This pattern matching means matches each low magnification key pattern and high magnification key pattern stored in the key pattern memory 31 described above with the pattern of the corresponding area on the wafer 1 to be actually cut placed on the holding stage 10. This is a means for preventing the cutting of the wafer 1, thereby determining whether the linear areas tl2a, 2b of the wafer 1 to be cut are set at a position that coincides with the cutting area (hairline a on the camera) by the blade 11 during cutting. At the same time as detection, if there is a discrepancy, it works to make them match. An example of this pattern matching effect will be described with reference to the 7-row chart in FIG. Note that this pattern matching action is performed for each left and right screen on the display screen, and for each low magnification key pattern and high magnification single key pattern, but the effect is substantially the same. As an example, a pattern matching operation based on a low-magnification key pattern in imaging with the low-magnification camera 17 will be described. The surface of the wafer 1, which is the object to be added, placed on the holding means 10 is observed with a low magnification camera 17.
The captured image is sent to the display means 3 via the imaging means 22, video channel selector 26, A/D converter 28, image frame 29 and central processing unit 30 as described above.
Image 17a on the right half of 3 (see Figures 10 and 11)
will be displayed, and pattern matching will begin immediately. That is, in steps n-1, the cursor C searches the right half of the image 17a' for a specific area that is the target corresponding to the key pattern 2r. In this case, when a key pattern is specified in the key pattern memory 31, the key pattern position is stored on the The cursor C is sequentially moved to search for the target, and in step n-2, 1
The degree of matching P is calculated sequentially as the pixels move pixel by pixel. In step n-3, the calculated matching degree P is compared with the threshold value, and if the matching IIP is greater than or equal to the threshold value, the process proceeds to step n-6, where the setting IIP and its coordinate positions are listed and stored in the memory. Thereafter, the process proceeds to step n-4, where it is determined whether or not the entire area has been searched. If the search has not been performed, the process proceeds to step n-5, where the cursor C is moved by one pixel, and step n-2
The matching degree P is calculated, and the same operations as described above are repeated. If the matching degree P is smaller than the threshold in step ri-3, the process proceeds to step n-4, and it is determined whether the search has been performed in the entire area. If not, in step n5, the cursor C is moved by one pixel. The matching degree P is calculated in step n-2, and the same operation is repeated thereafter. ``After completing the search for all areas, proceed from step n-4 to step n-7, where the largest matching P of the listed areas is determined. Since the area with the matching degree P matches the key pattern 2r, it is recognized as the area, that is, the target, and serves as the reference for alignment.The pattern matching operation is performed on the left screen of the display means and the display screen at high magnification. , furthermore, the holding means 10 is
When the wafer is rotated by 0°, the same process is performed to select a region on the wafer that serves as a reference for alignment, that is, a target. 'Next, actual alignment and cutting will be explained with reference to FIG. 4 and the flowchart of FIG. 15. First, after all camera settings and a specific area serving as a key pattern are set and stored using a dummy wafer and a sample wafer, the actual processing is started. Therefore, the frame 5 on which the wafer 1, which is the workpiece, is mounted is pre-aligned by an appropriate supply means and then supplied to the holding stage 10, and the wafer is held in a stable state by the holding means, for example, by suction means. Retained. Low magnification cameras 17,' 18
The image of the surface is captured by the central processing unit 3o through a predetermined conversion means, and the image is displayed on the display means 33 on the left and right sides. A specific area is searched for as a target corresponding to the low magnification, that is, macro key pattern > 2r, 2J2 stored in the pattern memory 31. Next, in step 2, the target key pattern 2r selected in the previous step 2 is searched. ,2a
Calculate the position on the X-Y coordinates of
Drive the drive source 13 of the holding means 10 based on the X-'Y coordinate value of 2ρ to correct the angle θ, and proceed to step - Confirm on the Y coordinate, and in step (2), the parallelism between the center/ring of the cameras 17 and 18, that is, the hairline a appearing on the image, and the straight line area 2a along the X axis is within the acceptable range. If the value is not within the allowable range, the process returns to step (3) again to correct the angle θ, and confirm the corrected positional relationship. If the parallelism between the hairline a and the linear region 2a along the X-axis is within the permissible range, the process proceeds to step (2), and the micro key pattern 2R is created based on the X-Y coordinate values. Targets 2R and 2L on the wafer corresponding to 2L are the cameras 19. Drive source 14. 15 to correct the holding means 10 in the X-axis or Y-axis direction, and the rough alignment is completed. Next, in step (3), the high magnification camera 19. 20, and the enlarged image is displayed on the 1,000-stage display 33 via a predetermined conversion means in the same manner as described above. Then, by the pattern matching operation, the micro key pattern 2R. Target 2R. which is a specific area corresponding to 2L. Search for 2L. Then, the process proceeds to step (2), and the target 2R. Calculate the position of 2L using X-Y coordinates. After that, proceed to step (3), and use the calculated X-Y coordinates and the micro key pattern 2R. stored in the key pattern memory 31. The deviation of the angle θ is corrected based on the X-Y coordinate values of 2L, and the corrected positional relationship is confirmed based on the XY coordinates in step [STEP]. ．． 20, that is, the parallelism between the hairline a and the straight line region 2a along the Return to (2), correct the angle θ, and confirm the corrected positional relationship based on the X-Y coordinates. If the parallelism between the hairline a and the straight line area 2a along the Move the holding stage 10 in the Y-axis direction. Then, the position of the holding means 10 is stored in the central processing unit 30,
The close alignment in the X-axis direction is completed. After the correction to reach the high magnification, that is, the X-axis alignment, the holding means 10 is rotated by 90 degrees in step O, and the high magnification camera 19 is rotated in step [phase].
20 causes the display means 33 to display an image of the surface. In this case, the wafer 1 placed on the holding means 10 is almost at an appropriate position due to the correction using the low-magnification and high-magnification images, that is, the alignment, in steps ① to ① performed previously, so step [Phase] and below do not require alignment twice at low magnification and high magnification, so one close alignment using only the high magnification camera is sufficient. Continuing the explanation based on the flowchart shown in FIG. 15, in step [phase], the micro key patterns 2R',
Targets 1 to 2R' on the wafer corresponding to 2m-. 2
Search for L1. Next, in step [phase], target 2R'. selected in step [phase]. Calculate the position of 2L-1 using X-Y coordinates. After that, proceed to step [phase] and calculate the calculated X-Y coordinates and the micro key pattern 2R-21- stored in the key pattern memory 31.
The deviation of the angle θ is corrected based on the X-Y coordinate values of the camera 19. Then, in step O, the corrected positional relationship is confirmed based on the X-Y coordinates, and in step [phase], the camera 19. 20, that is, the parallelism between the hairline a appearing on the image and the straight line area 2b along the Y axis, is within the permissible range, and if the value is not within the permissible range, Then, go back to step (2) above, correct the angle θ, and convert the corrected positional relationship to
- Confirm based on Y coordinate. And hairline a
If the parallelism between and the straight line area 2b along the Y-axis is within the allowable range, in step 0, the parallelism is adjusted in the Y-axis direction based on the X-Y coordinate values so that the straight line area 2b coincides with the hair line a. Moving the holding means 10. Then, the position of the holding means 10 is stored in the central processing unit 30, and the Y-axis alignment is completed. In the manner described above, alignment of the linear regions 2a and 2b on the wafer 1 is performed. In addition, the time required for the entire alignment process above is approximately 1
It is about 0 seconds. In this way, immediately after the alignment of the wafer 1 to be processed in the X-axis direction (linear area 2a) and Y-axis direction (linear area 2b) is completed, the cutting operation by the blade 11 is performed. When the above-mentioned alignment is properly performed, cutting is performed through the center of the straight line regions 2a and 2b, making it possible to perform accurate cutting at predetermined positions. In the first embodiment described above, a pair of low-magnification cameras and a pair of high-magnification cameras are used, but for example, only one low-magnification camera may be used for rough alignment. That is, in the second embodiment, one low magnification camera 17- and two high magnification cameras 19-, 20- are placed close to the optical means 16 disposed on the upper part of the holding means 10. It is equipped with a so-called trinocular alignment system. Regarding the installation of each of these cameras, the low magnification camera 17' is arranged corresponding to the high magnification camera 19-,
A pair of high-magnification cameras 19' and 20- are arranged parallel to the The spacing between the wafers and the wafers is configured to be appropriately adjusted by adjusting means according to the type of semiconductor wafer, that is, the chip size. Then, the low magnification is set to approximately 1 to 5 times as described above, and the high magnification is also set to approximately 10 to 30 times, and the camera of either magnification is set to a position such as a semiconductor wafer placed on the holding means 10. The surface of the workpiece 1 is photographed, and the magnification can be switched from low magnification to high magnification or vice versa as appropriate. Furthermore, as shown in the block diagram shown in FIG. 16, the images captured by the camera are processed through appropriate optical paths and displayed as a monitor on a display means such as a television screen equipped with a cathode ray tube (described later). Ru. In this case as well, image processing is performed in the same manner as in the first embodiment, so the same parts as in FIG. 4 are given the same reference numerals and details thereof will be omitted. That is,
The low magnification camera 17- has an imaging means 22, and the high magnification camera 19'20' has an imaging means 24. 25 are connected to each other, and these imaging means 22 and 24 . 25 is connected to a video channel selector 26. The video channel selector 26 is connected to the central processing unit 30 via an A/D converter 28 and an image frame memory 29.
It is connected to the. Furthermore, the central processing unit 30 is connected with a pattern matching means 32, a key pattern memory 31, and a display stage 33 such as a television screen equipped with a cathode ray tube.
A signal line 34 for driving the drive Iiii 13 to 15 (see FIG. 1) and a signal line 35 for driving the optical means 16
They are also the same in that they are arranged respectively. Next, image processing will be explained using the block diagram shown in FIG. The captured image is output as an analog signal indicating the brightness of the image by the imaging means 22, and the analog signal is converted into a digital signal by the A/D converter 28 via the video channel selector 26 and output, and is temporarily stored in the image frame memory 29. is memorized. In this case, the image pickup means 22 has a plurality of image pickup elements (for example, 256×256) arranged in an x-y matrix, and each image pickup element outputs a voltage according to the density of the image as an analog signal. , this analog signal is magnified 1 to 5 times by a video channel selector 26 and inputted to an A/D converter 28, where it is converted into an 8-bit multilevel digital signal and output, and is sent to an image frame memory 29. is temporarily stored. Note that the images captured by the high-magnification cameras 19 and 20 are also captured by the imaging means 24 in the same manner as described above. The analog signal is output based on the brightness of the 2Stfi, which is converted into a digital signal by the A/D converter 28 via the video channel selector 26, and temporarily stored in the image frame memory 29 at a high magnification. This is the same as the first embodiment. The central processing system 30, the operating section 30a, the key pattern memory 31, the 1,000-stage pattern matching 32, the display means 33, the signal line 34. 35 and their respective operations and functions are substantially the same as those of the first embodiment. The image taken by the low-magnification camera 17 is different from that of the first embodiment in that only one image is displayed on the display means 33, and the image taken by the low magnification camera 17 is different from that of the first embodiment in that only one image is displayed on the display means 33. There are only some differences in settings and pattern matching means. Therefore, regarding this trinocular alignment system, the first
This will be explained with reference to the flowchart shown in FIG. First, an image captured by the low magnification camera 17- is subjected to predetermined image processing and displayed on the display means 33, and in step (1), a target corresponding to the macro key pattern 3r stored in the key pattern memory 31 is displayed. Search for a specific area. In this case, as in the first embodiment, pattern matching is performed on the entire captured image or by moving pixel by pixel and calculating the matching degree P for each pixel in a certain masked area.
Select target 3r. Next, in step 2', the position of the selected target 3r is calculated in X-Y coordinates. After that, proceed to step (1), and use the calculated X-Y coordinates, the X-Y coordinate values of the macro key pattern 3r stored in the key pattern memory 31, and the X-Y coordinate values of the micro key pattern 3R. Based on the target 3R corresponding to the micro key pattern 3R, the high magnification camera 19
The drive sources 14 and 15 (see FIG. 1) are driven to move the holding means 10 in the X and Y axis directions so that the holding means 10 is positioned approximately at the center of the holding means 10. Then, in step (1), a target 3R corresponding to the micro key pattern 3R is searched for by a pattern matching operation. Next, proceed to step 1, calculate the position of 3R to target 1 on the X-Y coordinates,
Furthermore, in step 1, wafer 1 is divided into 1 chip
move in the X-axis direction by a rectangular area divided by a and 2b),
In step ■', search for targets 1-3R similar to the previous one,
In step 1, set the position of the target 3R to
Calculate using Y coordinate. Then, in step (2), the drive source 13 of the holding means 10 is driven to correct the angle θ based on the coordinate values calculated in the preceding step (1) and step (2), and the rough alignment is completed. Next, the camera 20 is switched to a high magnification camera 20, and in step 0', a specific area that is also the target 3L is searched for, and in step 0', the position of the target 3L is calculated in X-Y coordinates. Then, in step 0-, the calculated XY coordinate values of the target 3R and the XY coordinate values of the target 3L are
-Y coordinate value and X of micro key patterns 3R and '3L stored in key pattern memory 31 = Y
The angle θ is corrected by driving the drive source 13 of the holding means 10 based on the coordinate values. In step 0', the corrected positional relationship is confirmed using the X-Y coordinate values, and in step [phase] 1, the central axes of the cameras 19-, 20-,
That is, it is detected whether the parallelism between the hairline a appearing on the image and the straight line area 2a along the X axis is within the permissible range, and if the value is not within the permissible range, step C ', the X-Y generated in step 0-
Coordinate values and the micro key pattern 3R. stored in the key pattern memory 31. The angle θ is corrected again based on the X-Y coordinate values of 3L, and the process returns to step 0' to confirm the corrected position relationship. If the parallelism between the hairline a and the straight line area 2a along the X-axis is within the allowable range, in step [phase] 1, the hairline a and the straight line area 2a along the X-axis are aligned. The drive source 15 is driven to perform correction in the Y-axis direction. Then, the position of the holding stage 10 is stored in the central processing unit 30, and the close alignment in the X-axis direction is completed. After the correction, that is, the alignment in the X-axis direction, the holding means 10 is rotated by 90 degrees in step 1, and the surface thereof is displayed as an image on the display means 33 by the high magnification cameras 19 and 20 in step 0-. . In this case, since the wafer 1 placed on the holding means 10 is almost at the appropriate position due to the correction in the previous steps 1 to [phase]', that is, the alignment, the wafer 1 placed on the holding means 10 is almost at the appropriate position, so in steps Since there is no need for two alignments at magnification and high magnification, a single close alignment using a high magnification camera is sufficient. That is, in step [phase] 1, the micro key patterns 3R-. Targets 3R1 and 3L1 on the wafer corresponding to 3L1 are searched. Next, proceed to step 0', and calculate the position of the target 3R''.
Micro key pattern 3R-1 stored in 1. 3L
]-Y coordinate value and target 3 R.
7. 3 Correct the deviation of the angle θ based on the X-Y coordinate values of 1', and in step [phase] 1, Q the corrected positional relationship with the , central axes of cameras 19'-, 20',
That is, it is detected whether the parallelism between the hairline a shown on the image and the straight line area 12b along the Y axis is within the allowable range, and if the value is not within the allowable range, the parallelism is detected. Then, return to step 1 again to correct the angle θ, and confirm the corrected positional relationship. If the parallelism between the hairline a and the straight line area 2b along the Y axis is within the allowable range, in step 0', the straight line area 2b is
The drive source 15 is driven so that b coincides with the central axis of the cameras 19-, 20-, that is, the hairline a, and the holding means 10 is moved in the Y-axis direction to perform correction. The X-Y coordinate position of the holding means 10 is then stored in the central processing unit 30. In this way, alignment is completed. Even in this case, the time required for the entire alignment process is about 10 seconds. In this way, after the alignment of the wafer 1 to be processed in the X-axis direction and the Y-axis direction is completed, the cutting operation by the blade 11 is immediately performed in the same manner as described above, but if the above-mentioned alignment is properly performed. As a result, cutting passes through the center of the straight line areas 2a and 2b. This makes it possible to perform accurate cutting at a predetermined position, and also allows sequentially supplied workpieces to be sequentially aligned and subjected to cutting using the same means. Effects of the Invention As explained above, the alignment system of the present invention includes a holding means for holding a workpiece, an optical means for imaging the surface of the workpiece held by the holding means, and an optical means for capturing an image of the surface of the workpiece held by the holding means. A working element that acts on the workpiece held by the holding means and performs processing. ．． In the implant system, the optical means is composed of three or more camera means, so that each camera means sets a predetermined range of the surface of the workpiece with respect to the surface image of the workpiece. It is possible to form an image by capturing images under the specified conditions and magnification, and it is easy to set key patterns for alignment.
In addition, since there is no need to change conditions or magnification during the optical path, an excellent effect is achieved in that alignment work can be performed quickly and accurately. The optical means may include three or four camera means, and the optical means may include one or two camera means for taking an image of the surface of the workpiece held by the holding means at a relatively low magnification. Since it is composed of two low-magnification camera means and two high-magnification camera means that image the surface of the workpiece at a relatively high magnification, the low-magnification camera means performs rough alignment, and the high-magnification camera means performs rough alignment. Precise alignment is performed by the magnification camera means, without the need for switching in the optical path, and each camera means performs its own alignment, thereby reducing alignment time and reducing the effect of blades, etc. on the workpiece. This has the excellent effect of quickly and accurately aligning the elements with each other. Furthermore, coarse alignment is performed by the low magnification camera means, and then fine alignment is performed by the high magnification camera means, and the coarse alignment and fine alignment are automatically performed by the pattern matching means. This also brings about the excellent effect that alignment work can be performed even more quickly and accurately. Furthermore, since the distance between the two high-magnification camera means can be changed depending on the type of workpiece, that is, the chip size, the processing position for processing or cutting the workpiece can be adjusted to
Even if the axis and the Y axis are different, it can be quickly dealt with, and there is also the effect of being excellent in versatility.

[Brief explanation of drawings]

FIG. 1 is a perspective view schematically showing the main parts of an apparatus to which an alignment system according to the present invention is applied, and FIG. 2 is an enlarged view of a part of the object to be processed by the alignment system. A plan view, FIG. 3 is a theoretical diagram showing adjustment of the set positions of low-magnification and high-magnification cameras applied to the alignment system, and FIG. 4 is a block diagram of the four-eye alignment system according to the first embodiment of the present invention. Figure 5 is a schematic perspective view of the holding means for explaining the position setting of the camera by the dummy way 8 which is a part of the alignment system, Figure 6 is a plan view of the dummy way 8, and Figure 7 is a An explanatory diagram of an example in which the dummy wafer is imaged by a camera and displayed on a display means, FIG. 8 is an explanatory diagram of a state in which positioning has been performed from the display image of FIG. An explanatory diagram for explaining the deviation between the hair line and the test cutting line; FIG. 10 is an explanatory diagram showing an image of a sample wafer captured by a low magnification camera displayed on a display means;
Fig. 11 is an explanatory diagram showing a state in which the position has been aligned on the X-Y coordinates from the shifted position in Fig. 10, and a state in which a specific area to be a key pattern is specified, and Fig. 12 is an explanatory diagram showing the position in Fig. 11. An explanatory diagram showing how to switch to a high-magnification camera, adjust the position on the X-Y coordinates, and specify a specific area that will become a key pattern. An explanatory diagram showing a state in which an image retracted by a high-magnification camera is displayed on a display means, the position is adjusted on the X-Y seat ridge, and a specific area that becomes a key pattern is designated. Figure 14 is a pattern matching in the alignment system. 15 is a flowchart showing the alignment process in the first embodiment, FIG. 16 is a block diagram of the trinocular alignment system 1 of the second embodiment, and FIG. 17 is the second embodiment. FIG. 3 is a flowchart diagram illustrating an example alignment process. 1... Workpiece (wafer) 1a... Dummy wafer 1b... Sample wafer 2... Pattern 2a... Horizontal straight line area 2b... Vertical straight line Region 2r, 2°C, 2r-, 2ρ-, 3r. 3ρ3r1. 3bu...low magnification specific areas 2R, 2L, 2R", 21'
．． 3R. 3L. 3R".3L-...High magnification specific area 3...Orientation flat 4...
...Rectangular area 5...Frame 6...
...Film 7...Reference end portions 8, 9...
・Guide part 10... Holding means 11...
...Blade 12...The support mechanism 13~
15... Drive source 16... Optical means 1
7. 17=, 18...low magnification camera 17a,
18a...Images 19b and 20 captured with a low magnification camera
b...Images 19 and 20 captured with a high-magnification camera. 1
9", 20'...High magnification camera 21...
Adjustment means 22 to 25...Imaging means 26...
...Video channel selector-28...A/
D converter 29...Memory 30...
Central processing unit 30a...operation section 31...
Key pattern memory 32... Pattern matching means 3233... Display means 34.
35...Signal line a...Hair line b...
... Cutting line C ... Cursor patent applicant DISCO Co., Ltd.

Claims (9)

[Claims] (1) A holding means for holding a workpiece, an optical means for imaging the surface of the workpiece held by the holding means, and an optical means that acts on the workpiece held by the holding means to process the workpiece. an alignment system comprising an operating element for carrying out the operation, wherein said optical means is constituted by three or more camera means. (2) The alignment system according to claim (1), wherein the optical means comprises four camera means. (3) The alignment system according to claim (1), wherein the optical means comprises three camera means. (4) Two low-magnification camera means that image the surface of the workpiece held by the holding means at a relatively low magnification, and two high-magnification camera means that image the surface of the workpiece at a relatively high magnification. 3. The alignment system according to claim 2, wherein the optical means includes a magnification camera means. (5) One low magnification camera means that images the surface of the workpiece held by the holding means at relatively low magnification, and two high magnification camera means that images the surface of the workpiece at relatively high magnification. 4. The alignment system according to claim 3, wherein the optical means includes a magnification camera means. (6) Alignment in which the workpiece held by the holding means is aligned in a desired positional relationship with respect to the operating element, in which rough alignment is performed by a low-magnification camera means, and then fine alignment is performed by a high-magnification camera means. The alignment system according to claims (4) and (5), which performs the following. (7) The coarse alignment and fine alignment are automatically performed by pattern matching means (6).
) described alignment system. (8) The alignment system according to any one of claims (4) to (7), wherein the low magnification camera means has an optically low magnification, and the high magnification camera means has an optically high magnification. (9) The alignment system according to any one of claims (4) to (8), wherein the distance between the two high-magnification camera means can be varied depending on the type of workpiece.