JP6765645B2 - How to inspect the brazing material pattern - Google Patents

Description

The present invention relates to a method for inspecting a brazing material pattern, particularly a brazing material pattern reflowed into an airtight sealing lid.

When joining metal members or metal members to ceramics, a brazing material may be formed on one of the members, and the metal members may be joined to each other via the brazing material. For example, in an electronic component storage package, a lid for airtight sealing (hereinafter, may be simply referred to as a lid) and a package housing (hereinafter, may be simply referred to as a housing) are joined via a brazing material. , A brazing material is formed in an annular shape on the back surface of the lid in advance, and is used for joining with a housing (for example, Patent Document 1).
Here, the back surface of the lid is the surface on the side where the lid and the housing are joined.

In the electronic component storage package, if the shape of the brazing material deviates from the design shape, the airtightness of the package after joining may deteriorate. Therefore, it is necessary to form a brazing material on the back surface of the lid at an accurate position without any shape defect. On the other hand, in the case of a square housing, the brazing material is formed in a shape including a linear portion corresponding to the end face, but as a method of inspecting the linear brazing material pattern, the width of the brazing material pattern is used. There is known a method of measuring at a plurality of locations at predetermined intervals and evaluating whether or not these pattern widths are within the design values. At this time, the edge detection method is generally adopted for the pattern width measurement. (For example, Patent Document 2 and Patent Document 3)

Japanese Unexamined Patent Publication No. 2011-228353 Japanese Unexamined Patent Publication No. 2016-156746 Japanese Unexamined Patent Publication No. 2018-072115

Electronic component storage packages are electronic components, and manufacturing lines are required to efficiently inspect a large number of components. That is, there is a demand for a method that can accurately inspect whether or not the shape of the brazing material is normal for a plurality of lids in a short time.

In the inspection method based on pattern width measurement, the inspection time can be shortened by widening the measurement interval and reducing the number of measurements. However, it has been difficult to use this means because the measurement accuracy cannot be maintained.

Therefore, the present invention provides a method for inspecting a brazing material pattern formed on a substrate, which can detect a shape abnormality more accurately than a conventional method.

The present invention is a method for inspecting an annular brazing material pattern including a straight portion, which is reflowed on a substrate. The brazing material pattern is imaged, and light / dark data of the brazing material pattern is obtained from the obtained image data. The first step of extraction and the extracted light / dark data are divided into a plurality of strip-shaped data parallel to the stretching direction and in the width direction of the brazing material with respect to the linear pattern of the brazing material having a predetermined width. It has two steps and a third step of evaluating the brightness of the strip-shaped data and determining that the brazing material pattern has a shape abnormality when it is recognized that there is a change in the brightness.

In the present invention, the light and dark data extracted from the image data of the brazing material pattern is divided in parallel in the brazing material pattern direction to obtain band-shaped data, and the brightness of the band-shaped data is evaluated. Here, since the change in the brightness of the brazing material pattern is detected, it is not necessary to measure the pattern width at a plurality of places at predetermined intervals, and it is possible to detect the shape abnormality of the brazing material pattern in a short time. .. Further, even if the pattern width of the brazing material pattern is normal, but there is an abnormality inside the brazing material pattern and the brazing material pattern is not formed according to the design shape, it becomes possible to determine the quality of the shape.

Here, in the present invention, in the third step, it is preferable that the strip-shaped data is divided by the pixels constituting the image data, and the change in brightness is evaluated by the difference in brightness of the adjacent pixels.

Further, in the present invention, in the second step, it is preferable to divide the strip-shaped data at intervals of the pixels.

Further, in the present invention, in the third step, when the change in brightness changes from a dark pixel to a bright pixel and returns to a dark pixel, or a change from a bright pixel to a dark pixel and returns to a bright pixel, It is preferable to recognize that there is a change in the brightness.

Further, in the present invention, it is preferable to recognize that there is a change in the brightness when the number of dark pixels sandwiched between the bright pixels or the number of bright pixels sandwiched between the dark pixels is within a predetermined number. ..

Further, in the present invention, the substrate can be used as a lid for airtight sealing of an electronic component storage package.

Further, in the present invention, it is preferable that the strip-shaped data is divided into a side portion and a corner portion of the brazing filler metal pattern.

According to the present invention, it is possible to detect a shape abnormality of the brazing material pattern formed on the substrate more accurately than the conventional method. In particular, the effect can be more exerted in the inspection of the brazing material pattern formed on the lid.

It is a schematic diagram which shows (a) the cross section and (b) the back surface of the lid used in this embodiment. It is a brazing material pattern inspection apparatus used in this embodiment. It is a schematic diagram which shows the relationship between (a) illumination and a brazing material pattern in this embodiment, and is an enlarged view which shows (b) the cross section and (c) the upper surface of the brazing material pattern to be inspected. It is a flow chart for demonstrating the inspection procedure in this embodiment. It is a schematic diagram for demonstrating the inspection procedure in the Example of this Embodiment. It is a conceptual diagram which showed the brightness evaluation method and the conventional pattern width measurement method in this embodiment. It is a schematic diagram of the defect part to be inspected in the Example of this Embodiment.

Hereinafter, the present invention will be described in detail using embodiments. This embodiment is an example of inspecting a shape abnormality of a brazing material pattern formed on the back surface of a lid used for an electronic component storage package.

As shown in FIG. 1, in the lid 1 used for the inspection in the present embodiment, the plating layer is laminated on the lid base material 2 in the order of the Ni plating layer 3 and the Au plating layer 4, and the brazing material 5 is the back surface of the lid 1. A reflow is formed in the form of four linear patterns connected to the outermost circumference of the.

FIG. 2 shows an example of a brazing material pattern inspection apparatus used for the inspection of the present embodiment.
The brazing material pattern inspection device 6 includes a tray 7, a lighting 8, a lens 9, a camera 10, a transport stage 13, a power supply for lighting 14, and a personal computer for inspection 15.

In the brazing material pattern inspection device 6, a tray 7 in which the lids 1 are arranged is installed on the transport stage 13, and a camera 10 to which the illumination 8 and the lens 9 are connected is arranged on the tray 7. The lid 1 can be illuminated with light from above for imaging, and the tray 7 can be transported to the imaging position by the transport stage 13 provided with the drive system motor.

As the illumination 8, it is preferable to use the split dome illumination 11 and the pseudo-coaxial illumination 12 together. As shown in the enlarged view (b) of FIG. 3, the brazing material 5 is melted in a semi-cylindrical shape. Therefore, in the split dome illumination 11, 5b parts are used for 11b parts and 5c parts are used for 11c parts. By irradiating the 5d part with the 11d part and the 5a part with the pseudo-coaxial illumination 12, the entire inspection surface of the brazing material 5 can be irradiated with light, and an image can be taken with uniform brightness in the width direction. I have to.

When irradiating light with the illumination 8, it is preferable to clearly distinguish the brazing material 5 and the Ni plating layer 3, and each part of the divided dome illumination 11 and the pseudo-coaxial illumination 12 are not turned on at the same time. Imaging may be performed with some lights turned off. As a result, the boundary between the brazing material 5 and the Ni plating layer 3 becomes clear, false detection at the boundary can be prevented, and accurate inspection can be performed.

It is preferable to use a telecentric lens for the lens 9. As a result, it is possible to prevent a difference in the images of the lids appearing at the center and the edge of the field of view, and to inspect a plurality of imaged lids under the same conditions.

It is preferable to use an area sensor camera as the camera 10. As a result, a plurality of lids can be imaged at the same time, and the lids can be selected from the obtained images for inspection.

The inspection personal computer 15 has a transfer stage control unit 16, a lighting control unit 17, a camera control unit 18, an arithmetic control unit 19, and a storage device 20. The transfer stage 13 is connected to the transfer stage control unit 16, the split dome illumination 11 and the pseudo coaxial illumination 12 are connected to the illumination control unit 17 via the illumination power supply 14, and the camera 10 is connected to the camera control unit 18. The transfer stage control unit 16, the lighting control unit 17, and the camera control unit 18 are connected to each other and are connected to the arithmetic control unit 19. As a result, the transfer stage control unit 16 controls the drive system motor of the transfer stage 13 so that the transfer stage 13 can be moved.

Further, the lighting control unit 17 controls the lighting and extinguishing of each lighting unit of the divided dome lighting 11 and the pseudo-coaxial lighting 12.

Further, the camera control unit 18 makes it possible for the camera 10 to take an image and to acquire the captured image data.

Further, the arithmetic control unit 19 enables the camera control unit 18 to control the timing of acquiring an image with the camera 10 and the lighting control unit 17 to control the timing of turning on and off the split dome illumination 11 and the pseudo-coaxial illumination 12. There is.

Further, the arithmetic control unit 19 processes the obtained image data so that the shape of the brazing material pattern can be evaluated.

Next, the inspection method of the lid 1 according to the present embodiment will be described with reference to the flowchart shown in FIG.

(Preparation)
First, after starting the inspection flow, in step S01, the tray 7 in which the plurality of lids 1 are arranged in a grid pattern is arranged on the transport stage 13.

Next, in step S02, the transfer stage 13 is driven by a command from the transfer stage control unit 16, and the tray 7 is conveyed to a position where the lid 1 is within the field of view of the camera 10.

(1st step)
Next, in step S03, the lid 1 is irradiated with light by the split dome illumination 11 and the pseudo-coaxial illumination 12, the brazing material pattern formed on the back surface of the lid 1 is imaged by the camera 10, and the image data obtained by the imaging is obtained. Transfer to the inspection personal computer 15.

Next, in step S04, an area for masking the area where the brazing material 5 exists is obtained. As a means at this time, for example, setting a threshold value capable of extracting the place where the brazing material 5 exists without excess or deficiency and performing binarization can be mentioned.

Next, in step S05, the region where the brazing material 5 exists is defined by using a mask.

(Second step)
Next, in step S06, the masked data is divided in parallel in the longitudinal direction of the brazing filler metal pattern to obtain a plurality of strip-shaped data. Here, the longitudinal direction refers to the stretching direction with respect to the linear pattern of the brazing material (the direction of the arrow shown in FIG. 5B). Each band-shaped data is divided into pixels constituting the image data. Here, it is preferable that the division widths of the strip-shaped data are equal. By doing so, it is possible to eliminate the factor that changes the sensitivity of the reaction for the same level of defects, and to detect an accurate and homogeneous change in brightness.
In addition, it is possible to reduce the inspection time by excluding the strip-shaped data that is not related to the inspection result.

(Third step)
Next, in step S07, the evaluation of the brightness of the strip-shaped data in the longitudinal direction is repeated for each of the divided strip-shaped data. The evaluation is performed on the inspection personal computer 15.

Regarding the evaluation of the brightness, it is preferable to recognize that there is a shape abnormality when the change from the dark pixel to the bright pixel and return to the dark pixel, or the change from the bright pixel to the dark pixel and return to the bright pixel. By doing so, it is possible to prevent erroneous detection of the boundary of the brazing material 5 as a shape abnormality.

Further, it is preferable to set a threshold value for the number of dark pixels sandwiched between bright pixels or the number of bright pixels sandwiched between dark pixels. By doing so, it is recognized that there is a change in brightness when the number of pixels is a predetermined number or more, so that it is possible to perform an inspection according to the limit on the size of the defect.

Further, the inspection time may be shortened by evaluating the brightness of each strip-shaped data without dividing the brazing material pattern into pixels.

Next, in step S08, when it is evaluated that there is a shape abnormality in at least one strip-shaped data, this lid is determined to be a shape abnormality, and NG is sent to the storage device 20 in the inspection personal computer 15 for all the strip-shaped data. When it is evaluated that there is no shape abnormality, it is determined that there is no abnormality in this lid, OK is stored in the storage device 20 in the inspection personal computer 15, and the process returns to step S04 to proceed to the inspection of the next lid.

FIG. 6 is a conceptual diagram showing the brightness evaluation method and the conventional pattern width measurement method in the present embodiment.
As described above, when the conventional pattern width measurement method is used in the evaluation of the brightness in the third step, it is necessary to accurately calculate each pattern width at a plurality of places of the brazing material pattern (FIG. 6 (a)). In the present embodiment, only the change in brightness in the longitudinal direction of each strip-shaped data is evaluated for a plurality of strip-shaped data divided in parallel in the longitudinal direction of the brazing filler metal pattern (FIG. 6B), so that a complicated calculation process is performed. This is no longer necessary, and abnormal shapes of the brazing material pattern can be detected in a short time.

Further, in the third step, since the change in brightness of the divided strip-shaped data in the longitudinal direction is evaluated for the brazing filler metal pattern, the pattern width of the brazing filler metal pattern is normal, but an abnormality occurs inside the pattern and the design shape is formed. It is possible to determine the quality of the shape of the brazing material pattern that was not formed as per the street.

(Post-processing)
When the pass / fail judgment in step 08 is completed for the last lid, the result of the pass / fail judgment recorded in the storage device 20 is output, and the process proceeds to step S09.

Next, in step S09, the transport stage 13 is driven by the command of the inspection personal computer 15, moved to the collection position of the tray, and the tray 7 is collected.

Next, examples of the method for inspecting the brazing filler metal pattern of the present invention will be described in detail. Examples will be described below.

In this embodiment, the brazing material formed on the lid used for the electronic component storage package was inspected using the following brazing material pattern inspection device. As shown in FIG. 7, the lid has an inner peripheral shortage defect 21 in which the width of the brazing material 5 is 2/3 or less of the normal width, and a void defect 22 in which the Au plating layer 4 is exposed in the brazing material 5. Was used as a sample containing. Specifically, a lid of sample A consisting of 952 NG products with insufficient inner circumference defects 21 and sample B consisting of 976 NG products with void defects 22 was prepared. Here, these lids were arranged in a grid pattern on a tray and transported to an inspection device.

Regarding the inspection device in this embodiment, a CMOS camera having an effective pixel count of 4 megapixels was used as the camera 10 in the brazing material pattern inspection device, and a telecentric lens having a depth of field of 0.5 mm was used as the lens 9. .. Further, the split dome illumination 11 has a power consumption of 18 W, and the illumination that can be switched in four steps in the height direction is used, and the pseudo coaxial illumination 12 has a power consumption of 10.5 W.

Here, when the illumination condition of the illumination 8 is expressed as "irradiation angle a ° to b °", all directions in which the angle formed by the vertical line drawn from the imaging device to the imaging target is a ° or more and b ° or less. It means to irradiate the light from.

In this embodiment, two types of images taken at an irradiation angle of 0 ° to 10 ° that emphasizes the shape change of the upper surface of the wax and an irradiation angle of 30 ° to 36 ° that emphasizes the shape change of the side surface of the wax are used. , It was decided to perform edge scanning in the longitudinal direction of the brazing material pattern.

First, in this embodiment, as a method of defining the region where the brazing material 5 exists, the brazing material existence region is determined by using a method of detecting edges at four boundaries between the brazing material and the base. The determined brazing material designation area is divided for each pixel of the CMOS camera.

Next, as shown in FIG. 5A, the above-mentioned brazing material designated area was divided in parallel along the longitudinal direction of the brazing material pattern to obtain a plurality of strip-shaped data. At this time, the strip-shaped data was divided into a side portion and a corner portion of the brazing material pattern as shown in FIGS. 5 (b) and 5 (c).
Further, for the side portion, the data distributed for each pixel from the center of the side portion of the brazing material pattern is used as the edge strip-shaped data, and for the corner portion, the radius of the arc forming the corner portion is specified from the drawing dimensions. The obtained data was defined as corner strip data.

Next, the evaluation of brightness was repeated for the obtained strip-shaped data.
Specifically, for each of the straight line portion in the above-mentioned side strip-shaped data and the arc portion in the above-mentioned corner strip-shaped data, the pixel groups constituting the strip-shaped data are arranged in the clockwise forward direction and the counterclockwise reverse direction. Edge detection was performed by scanning in.

For the evaluation of brightness, edge detection using a Shen-Castan filter with a smoothing coefficient set to 50 is adopted, and the threshold value for edge detection is hysteresis for images with an irradiation angle of 0 ° to 10 °. For images with an irradiation angle of 30 ° to 36 ° and 750 or more, it was determined that there was an edge when the hysteresis was 800 or more.

As the filter for edge detection, the Shen-Castan filter was used this time, but a Canny filter, a Sobel filter, or the like may be used. Further, it is desirable to change the above hysteresis to an optimum value according to the filter adopted for edge detection and the lighting conditions.

Regarding the edge processing detected by the evaluation of brightness, the number of dark pixels sandwiched between bright pixels and the number of bright pixels sandwiched between dark pixels in the scanning direction is determined to be NG if it is within 165 μm. And said.
As described above, the upper limit threshold value set in this embodiment is a sufficiently large value with respect to the frequently occurring defect size, and overdetection can be reduced by setting the threshold value in this way. Further, by setting only the upper limit threshold value, it tends to appear as a small shape defect, so that a defect such as the above-mentioned defect void 22 can be detected without omission.

The threshold value set when evaluating the brightness may be changed depending on the product, the type of defect, and the calibration of the inspection algorithm.

Table 1 shows the results of inspection determination for the lid prepared in this embodiment. Here, in Table 1, the inspection method by the binarization process, which is generally used as the abnormality detection means, is referred to as M0, and the inspection method using the brightness evaluation process according to the present invention is referred to as M1.

In the inspection using the treatment according to the present invention, it was found that the oversight rate of the NG product was greatly improved regardless of the type of defect.

From the above, it was found that, according to the present invention, since the defects of the brazing material pattern can be accurately discriminated, the NG product of the brazing material pattern formed on the lid used for the electronic component storage package can be inspected without being overlooked. Further, since the band-shaped data along the brazing material pattern is processed, the processing time can be expected to be shortened as compared with the conventional inspection method by pattern width measurement.

The present invention is not limited to the above embodiment. It is possible to change the content within the scope of the claims.
For example, in the above embodiment, the inspection of the lids arranged in the tray has been described, but the method is not limited thereto.

Further, the present invention is not limited to the inspection of the brazing material pattern formed on the lid.

1 lid 2 lid base material 3 Ni plating layer 4 Au plating layer 5 brazing material 6 brazing material pattern inspection device 7 tray 8 lighting 9 lens 10 camera 11 division dome lighting 12 pseudo coaxial lighting 13 transport stage 14 lighting power supply 15 inspection personal computer 16 transport Stage control unit 17 Lighting control unit 18 Camera control unit 19 Arithmetic control unit 20 Storage device 21 Inner circumference shortage defect 22 Void defect

Claims (7)

This is a method for inspecting an annular brazing material pattern including a straight portion, which is reflowed on a substrate.
The first step of imaging the brazing material pattern and extracting the light and dark data of the brazing material pattern from the obtained image data,
The second step of dividing the extracted light and dark data into strip-shaped data by dividing the extracted linear pattern of the brazing material into a plurality of pieces parallel to the stretching direction and in the width direction of the brazing material.
The third step of evaluating the brightness of the strip-shaped data and determining that the brazing material pattern has a shape abnormality when it is recognized that there is a change in the brightness,
A method for inspecting a brazing material pattern, which comprises having. In the third step,
The strip-shaped data
Divide by the pixels that make up the image data
The method for inspecting a brazing material pattern according to claim 1, wherein the change in brightness is evaluated by a difference in brightness between adjacent pixels. In the second step,
The method for inspecting a brazing material pattern according to claim 2, wherein the strip-shaped data is divided at intervals of the pixels. In the third step,
When the change in brightness changes from a dark pixel to a bright pixel and returns to a dark pixel, or a change from a bright pixel to a dark pixel and returns to a bright pixel, it is recognized that there is a change in the brightness. The method for inspecting a brazing material pattern according to claim 2 or 3. When the number of dark pixels sandwiched between the bright pixels or the number of bright pixels sandwiched between the dark pixels is within a predetermined number.
The method for inspecting a brazing material pattern according to claim 4, wherein it is recognized that there is a change in brightness. The substrate,
The method for inspecting a brazing material pattern according to any one of claims 1 to 5, wherein the lid is used for airtight sealing of an electronic component storage package. The strip-shaped data is
The method for inspecting a brazing material pattern according to any one of claims 1 to 6, wherein the brazing material pattern is divided into a side portion and a corner portion.