KR20060127529A - Method for inspecting of semiconductor device - Google Patents

Abstract

A method for testing a semiconductor device is provided to improve precision of a defect test in a ball or a printed circuit board of the semiconductor device by obtaining a lighting image and a coaxial image. A light image is obtained as light image information(S10). A coaxial image is obtained as coaxial image information(S20). An annular image is obtained by subtracting the coaxial image information from the lighting image information(S30). A region between inner and outer diameters are uniformly divided(S40). A binary value of the divided region is analyzed(S50). It is judged whether there are pixels having a binary value of 1 in the divided regions(S60). When there is at least one pixel having a binary value of 1, it is judged that a corresponding region is a normal(S70). It is judged whether a rate of the normal region is equal to or greater than a set value(S80). When the rate of the normal region is equal to or greater than a set value, it is judged that a corresponding ball is normal(S90).

Description

Method for Inspecting Of Semiconductor device

1 and 2 are schematic configuration diagrams of a semiconductor device inspection apparatus according to an embodiment of the present invention.

3 is a flowchart illustrating a method of inspecting a semiconductor device according to an embodiment of the present invention.

4 to 6 is an image photograph showing a method of inspecting a semiconductor device according to an embodiment of the present invention.

7 is a flowchart illustrating a method of inspecting a semiconductor device according to another exemplary embodiment of the present disclosure.

8 and 9 are schematic configuration diagrams of a semiconductor device inspection apparatus according to still another embodiment of the present invention.

10 is a flowchart of a semiconductor device inspection method according to still another embodiment of the present invention.

11 to 13 are image photographs showing a method of inspecting a semiconductor device according to still another embodiment of the present invention.

Description of the main parts of the drawing

10: test object

 11: ball

12: printed circuit board

20: Doom lighting unit

21: the first laser

22: second laser

23: third laser

24: fourth laser

25: Auxiliary Doom Lighting

30: beam split

40: camera

50: side lights

100, 200, 300: Defect site

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for inspecting the presence of abnormalities in the appearance of semiconductor devices, and more particularly, to a method for inspecting semiconductor devices for improving the accuracy of inspection of the appearance defects of balls or printed circuit boards of semiconductor devices.

In general, semiconductor devices are important parts used in computers, home appliances, etc., and must be thoroughly inspected before production and shipment. These semiconductor devices require a higher degree of precision than other components. Any defect in the system can have a fatal effect on performance.

Since the external defects of the semiconductor elements, in particular the defects of the leads or the balls, may also occur in the process of assembling the semiconductor elements to the PCB, the inspection of the state of the leads or the balls of the semiconductor elements having the printed circuit board on which the BGA is mounted is very difficult. It is one of the important processes.

Lead or ball defects of the semiconductor devices as described above may appear in various forms. In particular, in the BGA type semiconductor devices, the solder balls on the lower surface form non-uniform surfaces, Alternatively, defects such as uneven pitch, width, offset, shape deformation, and missing may occur.

Typically, external defects of semiconductor devices having a printed circuit board on which the ball grid array is mounted are made by vision inspection using a CCD camera or the like.

However, according to the conventional vision inspection, only the coaxial illumination is used to photograph the semiconductor element and the black pie is judged as damage, so the damage is not recognized only by the coaxial illumination of the defective sample, or the illumination is directed from the side of the inspected object. If there is a problem that the damage is raised white and was not recognized as a damage.

An object of the present invention for solving the problems of the prior art is to obtain a fallen image and the coaxial image of the center portion of the ball brightly formed in the peripheral portion of the ball mounted on the printed circuit board of the semiconductor device, It is to provide a semiconductor device inspection method that can improve the accuracy of the inspection of the defect of the ball by accurately obtaining the image of the weak reflection by the coaxial illumination by scoring the annular image minus the coaxial image.

In addition, the present invention obtains a coaxial image formed on the surface of the printed circuit board and a fallen image formed on the surface of the printed circuit board, and recognizes a black portion in the difference image through the difference image obtained by subtracting the fallen image from the coaxial image. Accordingly, an object of the present invention is to provide a method for inspecting a semiconductor device capable of improving the accuracy of inspection of defects on a printed circuit board by detecting damage that was not detected by a single image of a coaxial image or a fallout image.

The semiconductor device inspection method of the present invention for solving the above technical problem is a semiconductor device inspection method for determining the presence or absence of a defect of a ball grid array mounted on a printed circuit board through an image captured by a camera, the semiconductor device Imaging the circumference of the ball provided on the white to obtain white image information through the captured image; Imaging the central region of the ball provided on the semiconductor device to appear white and acquiring coaxial image information through the captured image; Acquiring annular image information about a ball by subtracting coaxial image information from the fallen image information; And determining whether a ball is defective by comparing the annular image information with preset reference image information.

The determining of the defect of the ball may include binarizing each pixel in the annular image, setting an inner diameter and an outer diameter of the annular shape, dividing an area between the inner diameter and the outer diameter, and dividing the division. And determining that the ball is normal when the ratio of the normal area is greater than or equal to a predetermined value for the area in which at least one pixel having a binary value 1 exists in the predetermined area.

Alternatively, determining whether the ball is defective may include binarizing each pixel in the annular image and setting an inner diameter and an outer diameter of the annular shape, and equally dividing an area between the inner diameter and the outer diameter in a fan shape. Determining a normal region for a region in which the binary value 1 exists on the inner diameter line and the binary value 0 on the outer diameter line among the divided regions; And determining that it is.

In addition, the present invention is a method for inspecting a semiconductor device to determine the presence or absence of defects on a printed circuit board through an image captured by a camera, the surface of the printed circuit board is irradiated by irradiating coaxial direction illumination to the semiconductor device and the image taken Acquiring coaxial image information to appear; Irradiating the semiconductor device with lateral illumination and acquiring fallen image information in which a printed circuit board surface appears black through the captured image; Acquiring difference image information obtained by subtracting falling image information from the coaxial image information and detecting edge pixels of respective directions of a printed circuit board using the difference image information; Setting an angular reference line using an average value of coordinates of each of the edge pixels; And comparing the detected edge pixel with the pixel of the reference line to determine whether there is a defect in the printed circuit board.

The edge pixel is obtained by binarizing the difference image information and detecting the outermost pixels of each row and column among the pixels having a binary value of 1 as the edge pixel.

The step of determining whether the printed circuit board is defective by comparing the reference line and the detected edge includes determining whether an edge pixel having a deviation from the maximum error setting value exists in both sides of the reference line among the edge pixels and setting a maximum error. If at least one pixel is out of the value, the corresponding printed circuit board is judged to be defective, and if there is no edge pixel outside the maximum error setting value in both directions of the reference line, the edge pixel is at both sides of the corresponding reference line. If it is determined whether the minimum error setting value is out of the minimum error setting value, if at least one edge pixel outside the minimum error setting value exists, set the edge pixels outside the reference line as the first damage group, the pixel outside the minimum error setting value The length of the reference line if it does not exist Determining whether the edge pixel is out of the set value continuously for more than the set value, if the continuous deviation is more than the set value, the edge pixels are set as a second damage group, the damage area between the first damage group and the reference line and the second damage group The damage area between the damage group and the reference line is calculated, and when the total area is equal to or larger than a set value, the printed circuit board is determined to be defective.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

1 and 2 are schematic configuration diagrams of a semiconductor device inspection apparatus according to an embodiment of the present invention, Figure 3 is a flow chart showing a method for inspecting a semiconductor device according to an embodiment of the present invention, Figures 4 to 6 Image photographs showing a method for inspecting a semiconductor device according to an embodiment of the present invention.

First, according to an exemplary embodiment of the present invention, as shown in FIG. 1, a semiconductor device (hereinafter referred to as an inspection object: 10) having a printed circuit board on which a ball grid array is mounted is transferred to a lower portion of the dome lighting unit 20. .

Then, among the first to fourth lasers 21, 22, 23, and 24 provided in the dome lighting unit 20, the first laser 21 located in the lateral direction of the object to be inspected is turned on to provide fall lighting. After the irradiation, the image is reflected from the surface of the inspected object 10 and a falling image is obtained from the light focused on the lens of the camera 40 through the beam split 30, and the obtained falling image is detected as falling image information (S10). ).

In the fallen image, strong reflection is formed at two portions of the ball by the fall light provided through the first laser 21, so that the circumference of the ball 11 appears white as shown in FIG.

Subsequently, as shown in FIG. 2, the fourth laser 24 of the dome lighting unit 20 is turned on to irradiate coaxial illumination from an upper direction of the inspected object 10, and then reflected from the surface of the inspected object to be beam split 30. A coaxial image is obtained from the light focused by the lens of the camera 40, and the coaxial image is detected as coaxial image information (S20). In this case, in order to increase the intensity of the coaxial illumination, the side lighting 25 provided around the dome lighting unit 20 is turned on so as to irradiate light onto the object 10 through the beam split 30. In addition, the auxiliary dome lighting 50 provided on the upper portion of the dome lighting unit 30 may be turned on to further illuminate the lighting.

Since the coaxial image strongly irradiates the illumination from the upper part of the inspected object 10, strong reflection is made in the central area of the ball 11 so that the central area of the ball 11 appears white as shown in FIG. 5.

Meanwhile, the annular image information obtained by subtracting the coaxial image information is detected from the obtained fallen image information to acquire the annular image as shown in FIG. 6 (S30).

In more detail, the fallen image information and the coaxial image information have a value of 0 to 255. The fallen image has a large value of 255 especially at the circumference of the ball, and the coaxial image information is larger than the circumference of the ball. You will have a lot of 255 in the center area and 0 in the circumference of the ball.

Accordingly, if the coaxial image information is subtracted from the fallen image information, 255 values are distributed in the periphery of the ball and negative (-) image values exist in the central area of the ball. The central region of the ball has an image information value of 0, and when the image is imaged, an annular image is obtained in which the circumferential portion of the ball is white and the central region is black.

When the annular image information is obtained, it is determined whether the outline of the ball has a normal shape by comparing with the preset reference image information.

To this end, first, image information of each pixel in the annular image is binarized, an inner diameter and an outer diameter of the annular shape are set, and the area between the inner diameter and the outer diameter is equally divided into a fan shape (S40). In the binarization of the image information, 255, which is the brightest image value, is assigned to binary value 1, and 0, which is the darkest image value, is assigned to binary value 1.

After binarization is performed on the pixels in each of the divided regions, a binary value in the divided region is analyzed (S50), and whether there is a pixel having a binary value 1 therein is determined (S60). If at least one pixel is present, the corresponding area is determined to be a normal area (S70), and if the ratio of the normal area is greater than or equal to the set value (S80), if it is greater than or equal to the set value, the ball is determined to be normal (S90).

7 is a flowchart illustrating a method of inspecting a semiconductor device according to another embodiment of the present invention, and descriptions of the same parts as the method of inspecting a semiconductor device according to an embodiment of the present invention will be omitted.

To this end, as in the above embodiment, the binary value in the divided region is analyzed (S50), and whether the binary value 1 exists on the inner diameter line of the divided region and the binary value 0 exists on the outer diameter line (S61). Only the region where the binary value 1 exists on the inner diameter line and the binary value 0 on the outer diameter line is determined as a normal region (S73), and it is determined whether the ratio of the normal region is equal to or larger than a set value (S81). If it is more than the set value it is determined that the ball is normal (S92).

8 and 9 are schematic configuration diagrams of a semiconductor device inspection apparatus according to still another embodiment of the present invention, and FIG. 10 is a flowchart of a semiconductor device inspection method according to another embodiment of the present invention, and FIGS. 11 to FIG. 13 is an image photograph illustrating a method of inspecting a semiconductor device according to still another embodiment of the present invention.

First, as shown in FIG. 8, the test object 10 is transferred to the lower part of the dome lighting unit 20, and then the auxiliary dome light 25 provided on the upper part of the test object 10 is turned on to coaxially illuminate the test object. And a coaxial image in which the surface of the printed circuit board 12 is white as shown in FIG. 11 from the light reflected from the inspected object 10 and focused on the camera 40 lens through the beam split 30. In operation S10, the coaxial image is detected as coaxial image information. In this case, in order to increase the intensity of the coaxial illumination, the side lighting 50 provided around the dome lighting unit may be turned on to irradiate light onto the inspection object 10 through the beam split 30.

Subsequently, as shown in FIG. 9, the first laser 21 and the second laser 22 positioned in the lateral direction of the inspected object 10 are turned on to illuminate the inspected object 10 with fall lights. From the light reflected from the inspected object 10 and focused on the lens of the camera 40 through the beam split 20, a fallen image of a black surface of the printed circuit board 12 is obtained as shown in FIG. Is detected as fallen image information (S20).

Subsequently, subtraction image information is subtracted from the coaxial image information to obtain difference image information as shown in FIG. 13, and binarization is performed on pixels in the obtained difference image information, and pixels having a binary value of 1 as a result of binarization. The outermost pixels of each row and column are detected as edge pixels (S30).

When the edge pixel is detected, the reference line in each direction is set using the coordinate average value for each edge pixel (S40).

Then, the pixels on the reference line and the edge pixel are compared with each other in the respective directions (S50), and after determining whether an edge pixel outside the maximum error tolerance setting value exists in both directions of the reference line among the edge pixels ( S51) If at least one pixel deviates from the maximum error setting value, the corresponding printed circuit board is determined as defective (S52).

If there is no edge pixel outside the maximum error tolerance set value in both directions of the reference line, it is determined whether the edge pixel deviates from the minimum error set value on both sides of the reference line in each direction (S61). If at least one edge pixel outside the error setting value exists, the edge pixels outside the reference line are set as the first damage group (S62).

In addition, when there is no pixel outside the minimum error setting value, it is determined whether the edge pixels deviating from the reference line are continuously distributed over the set value in the length direction of the corresponding reference line (S63), and when the pixels are continuously distributed over the set value. The edge pixels are set as the second damage group (S64).

The damage area between the first damage group and the corresponding reference line and the damage area between the second damage group and the corresponding reference line are calculated (S70).

When the damage area calculation is completed in each direction, the total sum of the damage areas is calculated (S71), and the damage area total is determined to be greater than or equal to the set value (S72). It judges that it is defective (S80).

As described above, when a defect occurs on the surface of the printed circuit board, as shown in FIG. 6, the defective portion is blackened as shown by reference numeral "100", or as shown by reference numeral "200" as shown in FIG. 7. Defective part comes up white. In this case, when the defective portion is black or white, the difference image of the two images is always black as shown in FIG. 9 even if only one of the two occurs.

Therefore, a video image acquisition error occurs in the coaxial image or the fallen image, or a defect that is not detected by each image alone can be accurately detected through the difference image of the two images.

As described above, the present invention obtains a coaxial image in which a center portion of a ball is brightly formed and a coaxial image in which a circumferential portion of a ball mounted on a printed circuit board is brightly obtained. By accurately acquiring the image of the part with weak reflection by the illumination, the accuracy of the inspection of the defect of the ball can be improved.

In addition, the present invention obtains a coaxial image formed on the surface of the printed circuit board and a fallen image formed on the surface of the printed circuit board, and recognizes a black portion in the difference image through the difference image obtained by subtracting the fallen image from the coaxial image. Therefore, the accuracy of the inspection of defects on the printed circuit board can be improved by detecting damage that was not detected by a single image of the coaxial image or the fallen image.

While the present invention has been described with reference to the accompanying drawings, it will be apparent to those skilled in the art that many various obvious modifications are possible without departing from the scope of the invention from this description. Therefore, the scope of the invention should be construed by the claims described to include many such variations.

Claims (7)

In the inspection method of a semiconductor device for determining the presence or absence of a defect of a ball grid array mounted on a printed circuit board through an image captured by a camera, Capturing the circumference of a ball provided on the semiconductor device in white and obtaining fallen image information through the captured image; Imaging the central region of the ball provided on the semiconductor device to appear white and acquiring coaxial image information through the captured image; Acquiring annular image information about a ball by subtracting coaxial image information from the fallen image information; Determining whether a ball is defective by comparing the annular image information with preset reference image information; Inspection method of a semiconductor device comprising a. The method of claim 1, Determining whether the ball is defective; Binarizing each pixel in the annular image and setting an inner diameter and an outer diameter of the annular image; Evenly dividing an area between the inner diameter and the outer diameter; Determining a normal area with respect to an area in which at least one pixel having a binary value 1 exists in the divided area; And determining that the ball is normal when the ratio of the normal area is greater than or equal to a set value. The method of claim 1, Determining whether the ball is defective; Binarizing each pixel in the annular image and setting an inner diameter and an outer diameter of the annular image; Dividing the region between the inner diameter and the outer diameter evenly in a fan shape; Determining a normal region with respect to a region in which the binary value 1 exists on the inner diameter line and the binary value 0 exists on the outer diameter line among the divided regions; And determining that the ball is normal when the ratio of the normal area is greater than or equal to a set value. In the inspection method of a semiconductor device for determining the presence or absence of a defect of a printed circuit board through an image captured by a camera, Irradiating the semiconductor device with coaxial illumination and acquiring coaxial image information in which a surface of a printed circuit board is white through the captured image; Irradiating the semiconductor device with lateral illumination and acquiring fallen image information in which a printed circuit board surface appears black through the captured image; Acquiring difference image information obtained by subtracting falling image information from the coaxial image information and detecting edge pixels of respective directions of a printed circuit board using the difference image information; Setting an angular reference line using an average value of coordinates of each of the edge pixels; Comparing a pixel of the reference line with the detected edge pixel to determine whether a printed circuit board is defective; Inspection method of a semiconductor device comprising a. The method of claim 4, wherein The detecting of the edge pixel comprises binarizing the difference image information and detecting the outermost pixels of each row and column among the pixels having the binary value 1 as edge pixels as edge pixels. The method according to claim 4 or 5, Comparing the reference line with the detected edge to determine whether a printed circuit board is defective; It is determined whether an edge pixel outside the maximum error setting value exists in both sides of the corresponding reference line among the edge pixels, and if at least one pixel outside the maximum error setting value exists, the corresponding printed circuit board is determined to be defective. An inspection method of a semiconductor device. The method of claim 6, When there is no edge pixel outside the maximum error setting value in both directions of the reference line; It is determined whether the edge pixels deviate from the minimum error setting value on both sides of the corresponding reference line. If at least one edge pixel deviating from the minimum error setting value exists, the edge pixels deviating from the reference line are set as the first damage group. , If no pixels deviate from the minimum error setting value, it is determined whether the edge pixels deviate more than a set value continuously in the longitudinal direction of the reference line. If deviate more than the set value, the corresponding edge pixels are set as the second damage group. and, Calculate a damage area between the first damage group and the reference line and a damage area between the second damage group and the reference line, And determining the printed circuit board as defective if the total area is equal to or larger than a set value.

Images