JP2007059858A - Chip image inspection method and its system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a chip image inspection method and its system for accurately, quickly and simply deciding the presence/absence of the ink mark of a chip, and for efficiently performing a packaging process, and for improving the efficiency of the packaging process and the yield modulus of an integrated circuit, and for simplifying the overall production flow. <P>SOLUTION: A default value is set, and the image of a chip surface is grasped, and the degree of contrast of the grasped image is increased. Continuously, the number of pixels of the increased image (the number of black pixels or the like) is counted, and the number of pixels is compared with one default value. When the number of pixels is larger than the default value, the ink mark of the chip surface is decided, and when the number of pixels is smaller than the default value, it is decided that there is no ink mark on the chip surface. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a wafer image inspection method and system. In particular, the surface ink mark (ink) is determined by image processing with increased contrast.

  FIG. 1 shows a main production process of a general integrated circuit. After the wafer production 10 is completed at the wafer factory (Fab), the wafer is subjected to a probing test 11 at the inspection factory to inspect the quality of chips (die / chip) on the wafer by electricity. Ink 12 is applied to the surface of the chip that cannot pass the inspection. This is ink because it is marked with colored ink. Subsequently, the wafer is sent to a packaging factory to perform a packaging process (packaging). First, the wafer is cut (saw) 13. Some processes re-inspect and mark defective products with ink 14 (although this procedure is not essential) because cutting causes defects in some chips. When formal packaging is performed by wafer preparation or die bonding 16, the ink (procedure 15) mark inspection is always performed before the chips are sucked. Stop sucking the tip with the ink mark and continue to inspect the next tip. When there is no ink mark on the chip, the chip passes the probe test 11, which means that it is a functioning chip and can be used in a subsequent packaging process. Finally, after completion of chip packaging, the packaged integrated circuit is sent to a test factory for final inspection 17.

  Ink inspection in step 15 above, usually by visual or mechanical (full or semi-automatic) method. However, it goes without saying that visual inspection is prone to error due to individual differences. Conventional machine (full or semi-automatic) inspection uses conventional analog / digital image processing / image processing techniques. The principle and technique captures a chip-like ink image and compares this captured image with a standard ink image in a database. When both the positions, brightness and shape match, it is determined that there is an ink mark (that is, a defective chip). Otherwise, it is determined that there is no ink mark (that is, a non-defective chip).

  However, in the conventional inspection method using a machine, there are many erroneous determinations such as a chip with an ink mark having no ink mark. The upper part of FIG. 2 is a chip capture image 20 and the lower part is a standard image 21 of a database. Since there was a deviation (to the right) in the original ink mark, the captured image 20 and the standard image 21 do not match when compared with the conventional image method, so it is determined (incorrectly) that the chip has no ink mark. Yes. Although the ink position to be applied to the chip in FIG. 3 is accurate, the captured image 22 of the chip is smaller than the database standard image 23 due to the small amount of ink mark, lack of ink, or out of ink. It is determined that the chip has no ink mark. 4 is accurate in both the ink position and size of the chip, but because the color of the ink mark is shallow, the captured image 24 of the chip has a large gray scale difference from the database standard image 23, so (incorrectly) It is determined that the chip has no ink mark. In FIG. 5, the ink color of the chip capture image 26 and the database standard image 27 is different. Usually, the ink colors used for marking are red, blue, and black. In order to match these ink colors, a standard ink image for each color must be stored in the database. This enlarges the database storage space and gives management and execution complexity. Furthermore, every time an ink mark is made, there is a possibility that the color of the ink will be different, so an erroneous determination as shown in FIG. 4 occurs. The shape shown in FIG. 6 is different in the shape used for the chip capture image 28 and the database standard image 29. In this respect, there is no standard format in the industry, which leads to erroneous determination of chip ink.

  The above-mentioned conventional machine inspection (such as misjudging a non-defective chip as a defective chip) wastes the chip, or misdetermines a non-defective chip as a non-defective chip. This is a burden and reduces the efficiency of packaging and inspection.

  In response to the above-mentioned various disadvantages in conventional image inspection, a new chip image inspection method and system are submitted, and the presence or absence of ink marks on the chip is accurately, quickly and easily performed, and the packaging process can be performed efficiently. The efficiency of the packaging process can be improved, the yield rate of integrated circuits can be improved, and the overall production flow can be simplified.

  One object of the present invention is to provide a new chip image inspection method and system, accurately and quickly and easily determine the presence or absence of an ink mark on a chip, and can efficiently perform the packaging process, thereby improving the efficiency of the packaging process. Can improve the yield rate of integrated circuits and simplify the overall production flow.

  Another object of the invention is not to worry about different ink colors without storing many standard ink images as in conventional image inspection systems. Incorrect determination due to ink mark deviation, ink dryness, ink density, and ink shape can be eliminated.

  In order to achieve the above object, the present invention provides a chip image inspection method. Set a default value to capture an image of the crystal grain surface and increase the contrast of the captured image. Subsequently, the number of pixels of the increased image (such as counting the number of black pixels) is counted, and the number of pixels is compared with one predetermined value. If the number of pixels is larger than the predetermined value, it is determined that there is an ink mark on the crystal particle surface, and if it is smaller, it is determined that there is no ink mark on the crystal particle surface.

  In another embodiment of the present invention, a chip image inspection method is provided, a capture device is provided, and a surface image of a chip is captured. An image processing unit is provided to increase the contrast of captured images. A counting unit is provided to count the number of pixels in the increased image. A comparison unit is provided, and the number of counted pixels is compared with one predetermined value. When the number of pixels is larger than the predetermined value, the chip is determined to have an ink mark.

  The present invention is implemented by software. In a chip image inspection system stored on a computer-readable medium, the above-mentioned procedure or function is executed, chip surface image processing captured from the system is performed, and the presence or absence of an ink (ink) mark is determined on the chip surface. .

  The present invention can be implemented by a programmed computer. A software program is attached, and the captured chip surface image is processed by executing this software program, and the presence or absence of an ink (ink) mark is determined on the chip surface.

The invention of claim 1 is the chip image inspection method,
Capture one chip surface image,
Increase the contrast of the captured chip image,
After counting the number of pixels of the increased image,
A chip image inspection method comprising: comparing a count pixel number with a predetermined value and determining that an ink mark is attached to the chip surface when the count pixel number is larger than the default value It is said.
According to a second aspect of the present invention, there is provided the chip image inspection method according to the first aspect, wherein the image capturing is performed by a photographing machine, a camera, or a microscope.
A third aspect of the present invention is the chip image inspection method according to the first aspect, wherein a part of the captured image is selected as an effective pixel.
In the invention of claim 4, the contrast adjustment procedure described above processes pixels of the captured image, and outputs one high grayscale value when the grayscale value of the pixel exceeds one logical limit value. 2. The chip image inspection method according to claim 1, wherein one low output gray scale value is output when one gray scale value is less than one logical limit value.
The invention according to claim 5 is the chip image inspection method according to claim 4, wherein the number of pixels counting step includes the number of pixels of the low gray scale value in the augmented image.
The invention according to claim 6 is the chip image inspection method according to claim 1, wherein the comparison result is stored in one storage device.

  The present invention newly submits a chip image inspection method and system, accurately and quickly and easily determines the presence or absence of ink marks on the chip, can efficiently perform the packaging process, improve the efficiency of the packaging process, and integrate The yield rate of the circuit can be improved and the overall production flow can be simplified.

  7 is the chip image inspection system of the present invention. Please refer to the chip image inspection method of FIG. In FIG. 7, a wafer 31 surface image is captured by the image capturing device 32 in the system block diagram 30 of the present invention (procedure 41). The image capturing device 32 here may be a camera, a photographing machine, or a scanning device. According to the working method and convenience of the packaging factory, there are several methods for capturing images by the image capturing device 32. An image of the entire chip / wafer 31 is captured and stored in the memory device 34. When an image of a certain chip (die / chip) is necessary, the position image of the chip is taken out alone. Another image capture method captures images on only one chip 312 at a time. In any of the image capturing methods, the number of pixels (ie, resolution) of the chip image is determined by the number of sensors (sensors, CCD, etc.) of the image capturing device 32. Of course, the present invention can use (or store) only a part of the images in order to increase the calculation speed of the post-procedure or reduce the burden on the storage device. In the chip image pixel arrangement shown in FIG. 9, a part of pixels (portion indicated by hatching) is taken in as effective pixels.

  The chip image captured in step 41 is subjected to specific image processing (image data processing / image processing) 332 (step 42). In an embodiment of the present invention, the contrast of the image obtained from the previous procedure is increased (also referred to as image enhancement, contrast enhancement). That is, pixels close to black are made black, and pixels close to white are made white. The gray scale value in the 8-bit digital image technology is divided into a total of 256 gradations from 0 to 255, of which the gray scale value of the black pixel is 0 and the white gray scale value is 255. FIG. 10 is an input / output relationship diagram of the image processing unit 332 of the present invention. Among them, the horizontal axis indicates the gray scale value of the input image, and the vertical axis indicates the gray scale value of the image output after processing. When the input grayscale exceeds a predetermined logic threshold (threshold), a fixed high grayscale value (such as 255 for white) is output. When the input grayscale value is smaller than a predetermined threshold (threshold), one fixed low grayscale value (such as 0 for black). Since logical limit values are used, this type of image processing technique with increased contrast is called thresholding. FIG. 11 shows an example of a gray scale histogram of a chip image. The horizontal axis indicates each gray scale value, and the vertical axis indicates the number of pixels. The image of FIG. 11 is subjected to the processing of FIG. 10, and then the output result of FIG. 12 or FIG. 13 is obtained.

  After the contrast in the procedure 42 is increased, the number of pixels of the processed image (such as FIG. 12 or FIG. 13) is counted (procedure 43 and block 334). In an embodiment of the present invention, the number of black (grayscale value 0) pixels is counted. Then, the counted number of pixels is compared with the default value N (procedure 44 and block 336). If the number of black (grayscale value 0) pixels is less than (or equal to) the default value N, the result is shown in FIG. As described above, it is determined that the chip has no ink mark (procedure 45). If the number of black (grayscale value 0) pixels is larger than the predetermined value N, the chip is determined to be a defective chip with an ink mark as shown in FIG. 13 (procedure 46).

  By the above-described chip image capture, image processing, pixel count and numerical comparison in the embodiment of the present invention, the presence or absence of an ink mark can be determined accurately, quickly and easily, and the subsequent packaging process can be performed efficiently. It can improve the yield rate of integrated circuits and simplify the entire production process flow. The present invention does not worry about different ink colors without storing many standard ink images as in conventional image inspection systems. Incorrect determination due to ink mark deviation, ink dryness, ink density, and ink shape can be eliminated.

  However, the above description is a preferred embodiment of the present invention and does not limit the scope of the present invention. Accordingly, all that is similar or similar to the structure, device, and features of the present invention shall fall within the spirit of the present invention and the scope of the claims. For example, the known technique indicates a defective chip with an ink mark, and the non-ink mark is identified by a non-defective chip. You can also

  Furthermore, in the embodiment, the input and output relationship of the image processing is shown in FIG. However, the present invention uses a completely opposite input / output relationship, and the input / output relationship shown in FIG. That is, when the input gray scale exceeds a predetermined logical threshold value (threshold), one fixed high gray scale value (such as 255 representing white) is output. When the input grayscale value is smaller than a predetermined threshold (threshold), one fixed low grayscale value (such as 0 for black).

  Further, the high output gray scale value (white) in FIG. 10 is not limited to 255, and the low output gray scale value (black) is not limited to 0. In addition, the logical limit value of FIG. 10 may not be fixed, but may be a variable method or an adaptive method.

  In the embodiment of the present invention, when the wafer is charged or die bonded (16 in FIG. 1), the above-described image capture 41, contrast adjustment 42, pixel count 43, and comparison with the default value N are compared. Perform 44 procedures. However, in other embodiments of the present invention, all the chips on the wafer chip are processed in a batch after the procedures of the image capture 41, the contrast adjustment 42, the pixel count 43 and the comparison 44 with the predetermined value N are collectively processed. The result (that is, a good chip or a defective chip) is stored in the memory device 34. When the wafer is charged or die bonded, the chip sucked out by reading from the memory device 34 is a good chip or a defective chip. I understand the tip.

  Image processing (increased contrast) 332, pixel count 334, and comparison 336 with default values in the previous embodiment. These three parties execute a software program together with a central processing unit (CPU) 35. However, these three (or a part thereof) can also be implemented by an electronic hardware system. Regardless of the software or hardware system, the individual parts can be executed by those who are familiar with the electronics and software fields, and therefore the description thereof is omitted here. The above-described software program is stored in a computer-readable medium such as a CD-ROM or a storage device. After these software are loaded on the computer, the central processing unit (CPU) 35 executes the command.

It is a figure which shows the main production processes of a general integrated circuit. It is the fault and illustration of the ink mark test | inspection by a well-known technique. It is the fault and illustration of the ink mark test | inspection by a well-known technique. It is the fault and illustration of the ink mark test | inspection by a well-known technique. It is the fault and illustration of the ink mark test | inspection by a well-known technique. It is the fault and illustration of the ink mark test | inspection by a well-known technique. It is a system block diagram of wafer image inspection of the example of the present invention. It is a system flow figure of wafer image inspection of the example of the present invention. FIG. 6 is an arrangement of effective pixels of a chip image according to another embodiment of the present invention. FIG. It is an input / output relationship diagram of the image processing unit in the embodiment of the present invention. It is an example of a gray scale histogram (histogram) of one chip image. 4 is a grayscale histogram output result obtained after image processing according to the present invention. 4 is a grayscale histogram output result obtained after image processing according to the present invention. It is an input / output relationship diagram of an image processing unit in another embodiment of the present invention.

Explanation of symbols

10 Wafer production (Fab)
11 Wafer probing test (Probe test)
12 Ink mark 13 Wafer cut (saw)
14 Marking 15 Ink Inspection 16 Wafer Preparation 17 Final Inspection 20 Chip Image Capture 21 Database Standard Image 22 Chip Image Capture 23 Database Standard Image 24 Chip Image Capture 25 Database Standard Image 26 Chip Image Capture 27 Database Standard Image 28 Chip Image Capture 29 Standard image of database 30 System block diagram of the present invention 31 Chip / wafer 312 Chip 32 Image capture device 33 Image processing, pixel count, comparison with default value 332 Image processing (increased contrast) unit 334 Pixel count unit 336 Unit 34 to be compared with predetermined value N Memory device 35 Central processing unit (CPU)
41 Image Capture 42 Contrast Adjustment 43 Pixel Counting Procedure 44 Comparison of Number of Black Pixels with Default Value 45 Determining Good Chip without Ink Mark 46 Determining Bad Chip with Ink Mark

Claims (6)

In the chip image inspection method,
Capture one chip surface image,
Increase the contrast of the captured chip image,
After counting the number of pixels of the increased image,
A chip image inspection method comprising: comparing a count pixel number with a predetermined value and determining that an ink mark is attached to the chip surface when the count pixel number is larger than the default value .   The chip image inspection method according to claim 1, wherein the image capturing is performed by a photographing machine, a camera, or a microscope.   2. The chip image inspection method according to claim 1, wherein a part of the captured image is selected as an effective pixel.   The contrast adjustment procedure described above processes the captured image pixel and outputs one high grayscale value when the pixel grayscale value exceeds one logical limit value. 2. The chip image inspection method according to claim 1, wherein when the value falls below a logical limit value, one low output gray scale value is output.   5. The chip image inspection method according to claim 4, wherein the number of pixels counting step includes the number of pixels of the low gray scale value in the augmented image. 2. The chip image inspection method according to claim 1, wherein the comparison result is stored in one storage device.

Images