TW201428891A - Gray level detection method for wafer non-breakage - Google Patents

Abstract

The present invention provides a gray level detection method for wafer non-breakage, which is used for determining if the wafer is broken off after cleavage. The method includes the following steps: (S1) obtaining the wafer image before cleavage; (S2) applying try cleavage to obtain the threshold value of gray level variation; (S3) proceeding automatic production and turning on the light source; (S4) applying wafer cleavage; (S5) obtaining the wafer image after cleavage; and, (S6) determining whether the wafer is broken off. The present invention determines whether the wafer is broken off to determine if the variation of the average gray level of the wafer images before and after wafer cleavage exceeds the threshold value. Due to the fast reaction characteristic of light, the present invention may shorten the time required for determining whether the wafer is broken off, and does not require high resolution for the image capturing device, so as to have low configuration cost and thus satisfy the usage requirement.

Description

Wafer unbroken gray level detection method

The present invention relates to a method of splitting a wafer, and more particularly to a method of detecting whether a wafer is broken after splitting.

Referring to FIG. 1 and FIG. 2, the wafer splitting machine 1 is used to split the wafer 2 into a grain of grains for subsequent packaging operations, and the wafer 2 is undergoing splitting. Previously, the transverse and longitudinal pre-cut lines 3 are first cut by laser, and then the wafer 2 is attached with a blue film 4 and fixed, and then sent to a wafer splitting machine 1 by a fixing jig 5 for cleaving operation. .
The wafer splitting machine 1 comprises a wafer holder 6, a cantilever rod 7, a file 8, an image capturing member 9, a pressing member 10 and an illumination source 11, the wafer holder 6 sandwiching the wafer holder 6 Fixing the clamp 5 and moving and rotating in a plane direction, the boring tool 8 is fixed on the cantilever rod 7 and disposed above the wafer 2 for performing a splitting operation, and when performing a splitting operation, The pressing member 10 is pressed against the wafer 2 to prevent the wafer 2 from being warped, and the illumination source 11 is integrally connected to the pressing member 10 for generating an imaging light source (not shown) to illuminate the wafer 2 The wafer 2 is penetrated by the wafer 2, and the image capturing element 9 is disposed under the wafer 2, and is used for capturing the image of the wafer 2 to know the position of the wafer 2.
Therefore, the wafer 2 can be positioned by the displacement of the wafer holder 6 and the image capturing component 9 to take image of the wafer 2, and after the positioning is completed, the upper and lower sides of the file 8 can be borrowed. The displacement and the quantitative displacement of the wafer holder 6 continuously chop the plurality of pre-cut lines 3, and the image capturing element 9 continuously monitors whether the positioning of the pre-cut line 3 is offset during the continuous cleaving process. The positioning is compensated by the degree of visual offset, and the splitting operation is completed when both the lateral and longitudinal pre-cut lines 3 are cleaved.
When the process deviation causes the wafer 2 to be unbroken, it is customary to re-align the position to find the unbroken position and manually re-twist, which is not only time-consuming but also causes the wafer 2 to collapse when the alignment is not accurate. The rate of decline. Therefore, the current practice is to detect whether the wafer 2 is broken after the splitting, and if it is not disconnected, the tool position is deepened and the wafer is directly re-cut until the wafer 2 is disconnected.
It is known that the wafer 2 is broken after the splitting. The patent publication No. I351069 of the Japanese Patent Publication No. I351069 discloses a method for optically detecting a wafer split by using a wafer image before and after splitting and utilizing The computer analyzes the difference in the width of the scribe line in the image before and after the wafer cleavage, and then determines whether the wafer impact lobes are correct, that is, whether the wafer is actually disconnected.
However, in this conventional method, to calculate the change in the width dimension, it is first necessary to know the width of the scribe line before and after the splitting, but due to the different process characteristics of the wafer, the variation of the width of the scribe line varies or does not change, and the false positive rate is high, and Let the image judge the width, the file must maintain the lower position to maintain the width change, and the speed is slow. Therefore, the conventional optical detection method not only has a high false positive rate, but also has a slow speed and cannot meet the demand for use.

Accordingly, the main object of the present invention is to provide a method for detecting an unbroken gray level of a wafer, which can quickly determine whether a wafer is broken by a change in an average gray scale value of an optical image.
The present invention is a method for detecting an unbroken gray level of a wafer for determining whether a wafer is broken after a wafer is cleaved, and the method includes the step S1: obtaining a wafer image before the splitting, and step S2: testing The crack obtains the gray scale change critical value, step S3: performs automatic production and turns on the light source, step S4: splits the wafer, step S5: obtains the split wafer image and step S6: determines whether the wafer is disconnected.
Step S1: obtaining the image of the wafer before the splitting, and capturing the image of the wafer before the splitting of the wafer through an image picker; Step S2: obtaining a critical value of the gray scale change by trial splitting, by manually testing The method of finding the gray scale change threshold value; step S3: performing automatic production and turning on the light source, in order to start the automatic operation, and irradiating the wafer through a light source; step S4: splitting the wafer to smash the wafer Step S5: obtaining the image of the wafer after the splitting, and capturing the image of the wafer after the splitting of the wafer through the image picker; Step S6: determining whether the wafer is disconnected, by transmitting an image processing unit Whether or not the wafer is disconnected is determined according to whether the wafer image before the wafer split and the average gray scale value of the wafer image after the wafer splitting are changed.
Accordingly, the present invention predicts whether or not the wafer image before the splitting of the wafer and the average gray scale value of the wafer image after the wafer is split by obtaining the characteristic of rapid response of the light, and estimating Whether the wafer is disconnected, the reaction is fast and accurate, and the computational burden on image processing is not large, and the determination result can be quickly made to meet the needs of use.

Conventional knowledge

1. . . Wafer splitting machine

2. . . Wafer

3. . . Pre-cut line

4. . . Blue film

5. . . Fixing fixture

6. . . Wafer mount

7. . . Cantilever rod

8. . . chopper

9. . . Image capture component

10. . . Pressing element

11. . . Illumination source

this invention

S1. . . Obtaining a pre-crack wafer image

S1-1. . . Establish comparison area

S2. . . Triggering to obtain the critical value of gray scale change

S3. . . Perform eye production and turn on the light source

S4. . . Splitting wafer

S5. . . Obtaining a wafer image after splitting

S6. . . Determine if the wafer is disconnected

S7. . . Deepen the knife position

S8. . . Complete the splitting process

20. . . Wafer

30. . . Capture area

31. . . Alignment area

Figure 1 is a structural diagram of a conventional wafer.
Figure 2 is a structural diagram of a conventional wafer splitting machine.
Figure 3 is a diagram showing the steps of the judgment method of the present invention.
Fig. 4 is a view showing the unbroken image of the wafer of the present invention.
Figure 5 is a diagram showing the wafer break image of the present invention.
Figure 6 is a schematic diagram showing the selection of the alignment area of the present invention.

In order to give your members a deeper understanding and recognition of the features, purposes and effects of the present invention, the preferred embodiments are illustrated with the following description:
Referring to FIG. 3, FIG. 4 and FIG. 5, the present invention is a wafer unbroken gray level detecting method for determining whether a wafer 20 is cleaved after the wafer 20 is cleaved. Disconnected, comprising the step S1: obtaining the image of the wafer before the splitting, step S2: obtaining the threshold value of the gray scale change by the test splitting, step S3: performing automatic production and turning on the light source, step S4: splitting the wafer, step S5: Obtaining the split wafer image and step S6: determining whether the wafer is disconnected.
Step S1: obtaining the image of the wafer before the splitting, and capturing the image of the wafer before the splitting of the wafer through an image capture device.
Step S2: The threshold value of the gray scale change is obtained by trial splitting, and the threshold value of the gray scale change is found through a manual test. For the general wafer 20, the wafer 20 is usually low in brightness before being disconnected. (As in Figure 4), it has a lower grayscale value, and when the wafer 20 is disconnected, the brightness increases (as in Figure 5) and has a higher grayscale value. Therefore, the present invention can select any one of the entire batches of wafers 20, perform manual testing, and observe changes in grayscale values. For example, if the average grayscale value is increased from 50 to 150, the median value can be taken as an average. The critical point at which the grayscale value changes, that is, the critical point is 100.
Step S3: performing automatic production and turning on the light source, in order to start the automated operation, and irradiating the wafer 20 through a light source, the main purpose is to increase the brightness of the wafer 20.
Step S4: Splitting the wafer to perform the splitting operation on the wafer 20, which is to split the wafer 20 by a conventional splitting machine.
Step S5: Obtaining the image of the wafer after the splitting, and capturing the image of the wafer after the splitting of the wafer 20 through the image picker.
Step S6: determining whether the wafer is disconnected, and determining whether the average gray level value of the wafer image after the cracking of the wafer 20 and the wafer image after the wafer 20 is cleaved by an image processing unit is determined by an image processing unit. Whether the wafer 20 is broken. That is, for the other wafers 20, when the average gray scale value of the wafer image after the crack of the wafer 20 is set to exceed 100 (critical value), it is determined that the wafer is off, and if not, it is not turned off.
According to the determination result of step S6, the present invention may further perform step S7: deepen the location, or step S8: complete the splitting process. Step S7: deepening the tool position, when it is determined that the wafer 20 is not disconnected, deepening the tool depth of the cleaving action, and returning to step S4, the depth of the tool bit deepening the cleaving action (ie, increasing the cleaving) The depth of the way, while the wafer 20 is indeed disconnected. Step S8: The cleaving process is completed. When it is determined that the wafer 20 is disconnected, the wafer 20 is displaced to the next knife position to perform the splitting of the next knife to complete the cleaving process.
Referring to FIG. 6 again, the present invention may further include a step S1-1: establishing a comparison area, which may be used by the user in the capture area 30 of the image capture device before performing step S2. For the wafer image before the cleaving of the wafer 20 and the wafer image after the cleaving of the wafer 20, it is expected that there will be a significant change, and a comparison region 31 is selected, and the image processing unit according to the ratio Whether or not the wafer 20 is disconnected is determined as to whether or not the average grayscale value of the region 31 changes. Therefore, through the use of the comparison area 31, the difference between the average gray level value of the wafer image before the crack of the wafer 20 and the wafer image after the wafer 20 is split can be increased to speed up the judgment and the accuracy. .
As described above, the present invention observes whether the wafer has a rapid response characteristic, and obtains an average gray scale value of the wafer image before the wafer is split and the wafer is cleaved to observe whether the wafer is changed. It is presumed that the wafer is disconnected, the response is fast, and the image resolution is not high, so the installation cost is low, and the computational burden on image processing is not large, and the determination result can be quickly made to meet the needs of use.
The above is only the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention, that is, the equivalent changes and modifications of the content of the patent application of the present invention should be the technology of the present invention. category.

S1. . . Obtaining a pre-crack wafer image

S1-1. . . Establish comparison area

S2. . . Triggering to obtain the critical value of gray scale change

S3. . . Automated production and turn on the light source

S4. . . Splitting wafer

S5. . . Obtaining a wafer image after splitting

S6. . . Determine if the wafer is disconnected

S7. . . Deepen the knife position

S8. . . Complete the splitting process

Claims (4)

A method for detecting a gray level of a wafer for determining whether a wafer is broken after a wafer is cleaved, comprising:
Step S1: obtaining a pre-crack wafer image, and capturing an image of the wafer before the cleaving of the wafer through an image capture device;
Step S2: obtaining a critical value of the gray scale change by trial splitting, and finding a critical value of the gray scale change by means of manual test;
Step S3: performing automatic production and turning on the light source, in order to start the automatic operation, and irradiating the wafer through a light source;
Step S4: splitting the wafer to perform a splitting operation on the wafer;
Step S5: obtaining the image of the wafer after the splitting, and capturing the image of the wafer after the splitting of the wafer through the image picker;
Step S6: determining whether the wafer is disconnected, and determining whether the wafer is determined by an image processing unit according to whether the wafer image before the splitting of the wafer and the average gray scale value of the wafer image after the wafer splitting are changed. Whether the circle is broken. The wafer unbroken gray level detecting method according to the first aspect of the patent application, further comprising a step S7: deepening the tool position, in order to deepen the tool bit depth when the wafer is not disconnected And return to step S4. The wafer unbroken gray level detecting method according to claim 1, further comprising a step S8: completing the splitting process, wherein when the wafer is determined to be disconnected, the wafer is subjected to the next knife. Splitting to complete the cleft palate process. The method for detecting a wafer unbroken gray level according to claim 1, further comprising a step S1-1: establishing a comparison area, wherein the user is allowed to use the image capture device before performing step S2. In the captured area, a comparison area is circled, and the image processing unit determines whether the wafer is disconnected according to whether the average gray level value of the comparison area changes.