JP2012069732A - Ball bond inspection device and ball bond inspection method used for the ball bond inspection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a ball bond inspection device and a ball bond inspection method for detecting a feature amount of ball bond from image data created by taking in reflection light from the ball bond without being influenced by a substrate pattern.SOLUTION: Illumination light (incident light i) is irradiated to ball bond 11 by light source of illumination 12, and reflection light j reflected by the ball bond 11 is taken in by a camera 13 so that shading image data vd is generated. At an evaluation value calculation part 15, an evaluation value is calculated from pixel information in a circular or ellipsoidal processing region that is predetermined as having a size less than that of an inspection target. At an evaluation image creation part 16, evaluation value calculation means is executed on any coordinate on the shading image data, so that an evaluation value is arranged to correspond to a coordinate of the processing region for creating an evaluation image. At a black area extraction part 17, continuous black regions with a gray value of a threshold or less are extracted from the evaluation image ve. At a ball feature amount detection part 18, a feature amount of the ball is detected from the black region.

Description

  The present invention relates to a ball bond inspection apparatus and a ball bond inspection method used in the ball bond inspection apparatus, and relates to a ball bond formed in a device manufacturing process, for example, wire bonding or bump bonding in a semiconductor component. The present invention relates to a ball bond inspection apparatus suitable for inspection of a ball bond and a ball bond inspection method used for the ball bond inspection apparatus.

  Among semiconductor components, particularly in semiconductor components in which a semiconductor chip and an inner lead are electrically connected by wire bonding or bump bonding, it is necessary to inspect the state of the joint between the ball bond and the pad in the manufacturing process. is there. For this reason, an apparatus for inspecting the shape of a ball bond has been manufactured. In this apparatus, a processing area of a circle or an ellipse having a size equal to or smaller than the inspection object is scanned with respect to an inspection image obtained after the formation of the ball bond. An evaluation image is created according to the shading contained in the region, and the shape of the ball bond is detected from the evaluation image.

  As this type of related technology, for example, there is a ball bond inspection system described in Patent Document 1.

  The ball bond inspection system includes the steps of constructing a polar coordinate transformation image, an array for storing the results, designating a plurality of expected ball center locations, and the ball bond during the bonding process. Acquiring a digitized image; storing the digitized image as a luminance image; receiving a ball center location from a machine vision system; registering the polar coordinate image and the ball center location; A step of generating a position dimension histogram, applying an edge filter, repeating the above until there are no more expected ball center locations, and depending on the peak of the histogram, the center of the ball bond Half Determining and notifying the machine vision system, wherein the step of obtaining the digitized image includes a difference image or subtracted image before and after the bond to find a bond in the digitized image. Is supposed to be created.

  FIG. 5 is a diagram showing a subtractive image 44 of the bond 09 formed on the pad in the ball bond inspection system described in Patent Document 1. As shown in FIG. Basically, the image of the pad 40 before bonding (see FIG. 5A) is saved and subtracted from the image of the pad 40 after bonding (see FIG. 5B). . As a result, a subtractive image 44 is created as shown in FIG.

JP 09-504401 (Abstract, Fig. 5)

  However, the techniques described in the above documents have the following problems.

  That is, in the method for creating the difference image before or after the bond or the subtractive image in the ball bond inspection system described in Patent Document 1, the ball bond is formed due to the spatial restriction of installing the imaging device and the restriction of the inspection time. It may be difficult to acquire images both before and after being done.

  The present invention has been made in view of the above-described circumstances, and an object thereof is to provide a ball bond inspection apparatus capable of accurately detecting a ball bond and a ball bond inspection method used in the ball bond inspection apparatus.

  In order to solve the above problems, the configuration of the present invention relates to an apparatus for inspecting a two-dimensional feature amount of a ball bond, irradiating an inspection object with illumination light, capturing reflected light, and generating grayscale image data. The evaluation value is calculated from the pixel information in the processing area of the circle or ellipse shape, and the evaluation value is arranged so as to correspond to the coordinates of the processing area by executing the evaluation value calculating means at arbitrary coordinates on the grayscale image data. Thus, an evaluation image is created, a continuous black region having a gray value of a threshold value or less is extracted from the evaluation image, and a feature amount of a ball bond is detected from the black region.

  According to the configuration of the present invention, a stable ball bond inspection using the density of the ball bond can be performed.

It is a block diagram which shows the electric constitution of the principal part of the ball bond inspection apparatus which is embodiment of this invention. It is a figure explaining the incident / reflection of the light with respect to the test | inspection object and the acquired image in the ball bond test | inspection apparatus which is embodiment of this invention. It is a figure explaining the scanning on the image of the process area | region in the ball bond inspection apparatus which is embodiment of this invention, and preparation of an evaluation image. It is a figure explaining a mode that the external shape of a ball bond is determined from the evaluation image in the ball bond inspection apparatus which is embodiment of this invention. It is a figure explaining the process performed with the ball bond test | inspection apparatus described in patent document 1. FIG.

  FIG. 1 is a block diagram showing an electrical configuration of a main part of a ball bond inspection apparatus according to an embodiment of the present invention.

  As shown in the figure, the ball bond inspection apparatus of this example is an apparatus for detecting a ball bond (inspection object) 11 and includes an illumination (a part of a light source) 12, a camera (imaging means) 13, and a half. It comprises a mirror (part of a light source) 14, an evaluation value calculation unit 15, an evaluation image creation unit 16, a black area extraction unit 17, and a ball bond feature amount detection unit 18. The ball bond 11 is an object to be inspected, for example, an electrical joint formed on a semiconductor component. The illumination 12 is configured as a light source such as an LED (Light Emitting Diode) or a halogen lamp, and irradiates the ball bond 11 with illumination light.

  The half mirror 14 reflects the light emitted from the illumination 12 to irradiate the ball bond 11 with the incident light i, and transmits the reflected light j to the camera 13. In particular, in this embodiment, the irradiation 12 and the half mirror 14 are configured so that the axes of the incident light i and the reflected light j with respect to the ball bond 11 substantially coincide. The camera 13 is for reflecting the reflected light j reflected by the ball bond 11 and generating grayscale image data vd.

  The evaluation value calculation unit 15 calculates an evaluation value from shading information in a circle or ellipse processing region. The evaluation image creating unit 16 executes the evaluation value calculation means 15 at arbitrary coordinates on the grayscale image data vd, and creates evaluation images ve by arranging evaluation values so as to correspond to the coordinates of the processing area. The black area extraction unit 17 extracts a continuous black area whose gray value is equal to or less than a threshold value from the evaluation image ve. The ball bond feature amount detection unit 18 detects the coordinates and outer shape of the ball bond 11 from the black region.

  FIG. 2 is a diagram for explaining the incident / reflected light of the inspection object and the acquired image in the ball bond inspection apparatus according to the embodiment of the present invention. FIG. 3 is a scan and evaluation image on the processing region image. FIG. 4 is a diagram for explaining how to determine the outer shape of the ball bond from the evaluation image.

  With reference to these drawings, processing contents of the ball bond inspection method used in the ball bond inspection apparatus of this example will be described.

  In wire bonding in a semiconductor manufacturing process, for example, a ball bond 11 is formed in a hemispherical shape on a flat pad, and a linear wire is bonded near the upper center of the ball. FIG. 2A is a diagram showing a positional relationship between the ball bond 11 to be inspected and the imaging optical system from the side. When the camera 13 is installed in a direction substantially coincident with the optical axis of the incident light as in the present embodiment, the light is incident in the same direction as the incident angle of light closer to 0 degree (i1) (j1). ) And the angle of incidence increases (i2), and the reflected light deviates from the direction of the camera (j2) and becomes difficult to be captured by the camera.

  The shape of the ball bond 11 is inclined with respect to the incident light at many portions thereof. In particular, in the case of the ball bond with the bonding wire connected, the gray value of the portion where the wire in the center of the ball bond exists also becomes small, so the ball bond 11 image acquired by the camera 13 is, for example, as shown in FIG. become. That is, an image having a small gray value is obtained in a ball bond portion and a portion other than the ball bond having an inclination. In this way, it is desirable to acquire an image in which the ball bond portion is occupied by a small gray value, but detection is possible even if there are some pixels with a large gray value in the ball bond portion. For example, depending on how the ball bond is formed in the bonding apparatus, the shape of the ball bond may be frustoconical, and a case where a slight horizontal portion of the ball bond portion appears as a white pixel in the image corresponds to such a case. To do. In this case, it is also effective to reduce the appearance of bright pixels by slightly defocusing the focus. For the purpose of reducing the density value of the ball bond portion and clarifying the density difference between the ball bond and the portion other than the ball bond, the binarization processing is applied to the gray image data vd. It may be used instead.

  Next, the evaluation value calculation unit 15 will be described. The evaluation value calculation unit 15 is used when an evaluation value is calculated by an evaluation image creation unit 16 to be described later. The evaluation value calculation unit 15 determines the shape of the processing area and calculates the evaluation value from the shades included in the processing area. .

  In the present embodiment, the shape of the processing region is determined in advance as a circle that is the same as or smaller than the diameter of the ball bond 11 to be inspected. In particular, when there is variation among products, it is desirable to determine according to the assumed minimum diameter. However, making the diameter too small may lead to erroneous detection of another having a characteristic similar to a ball bond as a ball bond, and thus needs to be determined appropriately. Depending on the ratio parameter described below, detection is possible even if the diameter of the processing region is larger than the ball bond 11. Further, the shape of the processing region may be determined based on the acquired grayscale image data vd without being determined in advance, or the shape may be an ellipse.

  The processing area is used by scanning the grayscale image data vd. At this time, an evaluation value is calculated from the shading contained in the processing area. The evaluation value must be a value that distinguishes between the case where the processing area overlaps the ball bond and the case where it does not. In the present embodiment, the ratio of the number of pixels having a density value equal to or lower than the density threshold included in the processing area to the total number of pixels in the processing area is calculated, the calculated ratio is binarized by the ratio threshold value, and binary Is an evaluation value.

  The light / dark threshold is a value that should be determined so that most of the light / dark values of the ball bond portion in the light / dark image data vd are equal to or smaller than the light / dark threshold, and may be determined in advance as a fixed value, or the light / dark image data vd may be determined as a histogram. There is also a method of determining the threshold value from the graph shape represented by.

  As for the ratio, since the processing area is set smaller than the ball bond 11, the ratio of the pixels having a small gray value included in the processing area is often 100%, but the ball portion on the gray image data vd. For example, the ratio threshold value may be set to 90% in consideration of the variation of the density condition. Lowering the percentage threshold from 100% increases the possibility of avoiding detection failure due to ball bond shading variations in the acquired image due to variations in the three-dimensional shape of the ball bond, and the size of the ball bond is smaller than expected. This has the effect of increasing the possibility of avoiding non-detection caused by the problem. However, if the ratio threshold is lowered too much, it may lead to erroneous detection of another having a characteristic similar to a ball bond as a ball bond. The calculation of the evaluation value is not limited to the above, and may be, for example, the total sum of shades included in the processing region.

  Next, the evaluation image creation unit 16 will be described. The evaluation image creation unit 16 executes the evaluation value calculation unit 15 at arbitrary coordinates of the grayscale image data vd, and arranges the evaluation values by associating the calculated evaluation values with the processing area coordinates at that time, thereby evaluating the evaluation image. Create ve. It should be noted that the coordinates where the processing region protrudes from the grayscale image data ve may be excluded from the processing target, or the evaluation value may be calculated only from the grayscale within the region. If the coordinates of the ball bond 11 are known in advance on the image data vd, the processing range can be limited to increase the processing speed and reduce false detections.

  The grayscale image data acquired by imaging the semiconductor component on which the ball bond is formed includes, for example, protrusion from the ball bond pad, scratches by the probe, and the influence of the substrate pattern as shown in FIG. There is a case. In such a case, stable detection of the ball bond may be difficult in general processing such as binarization, labeling, and pattern matching. In the present embodiment, by giving the above-described characteristics to the processing area to be scanned, it is possible to create an evaluation image ve in which the black area including the ball bond feature amount is isolated as shown in FIG. Contributes to stable detection of feature quantities.

  Next, the black area extraction unit 17 will be described. As shown in FIG. 3B, the evaluation image ve created by the evaluation image creation unit 16 is a black region in which lines such as a substrate pattern and noise of a small area are removed from the grayscale image data vd and the ball feature amount is included. Other than that, a continuous black region (2) having a large area, for example, where the processing region is completely filled, remains as shading. In this embodiment, since the evaluation value is a binary image, a black region is extracted by performing a labeling process on black pixels. If the evaluation image ve created by the evaluation image creating unit 16 is a multi-tone grayscale image, a black region is extracted by performing a labeling process on grayscale pixels below the threshold value using the grayscale value as a threshold value.

  The ball bond feature quantity detection unit 18 calculates the area of the black region, the length of the circumscribed rectangle, and the roundness, and determines whether or not these values fall within the range, thereby determining the black that includes the ball bond feature quantity. Region (1) can be identified. The range may be determined in advance, or may be determined based on the black region when the processing region is applied to the standard size ball bond 11 under the same conditions.

  The feature amount of the ball can be calculated from the black region (1) including the ball bond feature amount. For example, the center of gravity of the black area can be calculated and used as ball bond coordinates. Further, for example, the center of the circumscribed rectangle of the black area can be set as the ball bond coordinates. Further, for example, a shape that is extended outward by the radius r of the circular filter with respect to the outer shape of the black region can be a ball bond outer shape.

  As described above, in this embodiment, the illumination light is applied to the ball bond 11 by the light source of the illumination 12, and the reflected light reflected by the ball bond 11 is captured by the camera 13 to generate grayscale image data vd. An evaluation value is calculated from pixel information in a circle or ellipse processing area determined in advance as a size equal to or smaller than the object, and evaluation value calculating means is executed at arbitrary coordinates on the grayscale image data to correspond to the coordinates of the processing area. Thus, the evaluation image ve is created by arranging the evaluation values, and the ball feature amount is detected from the feature amount of the continuous black region of the evaluation image. Therefore, stable ball bond inspection using the density of the ball bond can be performed.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to the embodiment, and even if there is a design change without departing from the gist of the present invention. Are included in the present invention. For example, in the ball bond detection device of FIG. 1, coaxial illumination may be disposed next to the camera 13 without using the half mirror 14 and the ball bond may be irradiated with light from above. In addition, for example, when illumination other than coaxial illumination is used, the density of the acquired image may differ, but by appropriately calculating the shape of the processing region and the evaluation value according to the characteristics of the shade, the ball bond Detection is possible.

A part or all of the above-described embodiment can be described as in the following supplementary notes, but is not limited thereto.
(Appendix 1)
An apparatus for inspecting a two-dimensional feature of a ball bond,
A light source for irradiating the inspection object with irradiation light;
An imaging unit that captures reflected light from the inspection object and generates grayscale image data;
Evaluation value calculation means for calculating an evaluation value from pixel information in a circle or ellipse processing region;
Evaluation image creating means for creating an evaluation image by executing the evaluation value calculating means at arbitrary coordinates on the grayscale image data and arranging evaluation values so as to correspond to the coordinates of the processing region;
Black area extraction means for extracting a continuous black area having a gray value equal to or less than a threshold value from the evaluation image created by the evaluation image creation means;
Ball bond feature quantity detecting means for detecting a ball bond feature quantity from the black area extracted by the black area extracting means;
A ball bond inspection apparatus characterized by comprising:
(Appendix 2)
The evaluation value calculation means includes
2. The ball bond detection device according to claim 1, wherein an evaluation value is calculated from pixel information in a circle or ellipse processing area determined as a size smaller than the inspection object.
(Appendix 3)
The evaluation value calculation means includes
3. The ball bond detection device according to claim 1, wherein a ratio of the number of pixels having a gray value range specified in advance from the image information in the processing area to the total number of pixels in the area is calculated as an evaluation value.
(Appendix 4)
The evaluation value calculation means includes
The ball bond detection device according to appendix 1 or 2, wherein the sum of the gray values is calculated from the image information in the processing area to obtain an evaluation value.
(Appendix 5)
The evaluation image creating means includes
The evaluation value calculation unit is executed at arbitrary coordinates on the binary image created by performing binarization processing on the grayscale image data generated by the imaging unit in advance, and the evaluation value corresponds to the coordinates of the processing region. The ball bond detection device according to appendix 1, 2, 3 or 4, wherein an evaluation image is created by arranging
(Appendix 6)
The ball bond feature amount detection means includes:
The ball bond according to appendix 1, 2, 3, 4 or 5, wherein the center of gravity of the black region satisfying at least one of an area range, a circumscribed rectangle length range and a roundness is a ball bond coordinate Detection device.
(Appendix 7)
The ball bond feature amount detection means includes:
Additional remark 1, 2, 3, 4 or 5 characterized in that the center of the circumscribed rectangle of the black region satisfying at least one of the area range, circumscribed rectangle length range and roundness is the ball bond coordinate. Ball bond detector.
(Appendix 8)
The ball bond feature amount detection means includes:
The ball bond outer shape is a shape that extends outwardly by the radial length of the circular filter with respect to the outer shape of the black region that satisfies at least one of the area range, the circumscribed rectangle length range, and the roundness. The ball bond detection device according to appendix 1, 2, 3, 4 or 5.
(Appendix 9)
A method for inspecting a two-dimensional feature of a ball bond,
An irradiation process for irradiating the inspection object with irradiation light;
An imaging process that captures reflected light from the inspection object and generates grayscale image data;
An evaluation value calculation process for calculating an evaluation value from pixel information in a processing area of a circle or an ellipse;
An evaluation image creating process for creating an evaluation image by executing the evaluation value calculation process at an arbitrary coordinate on the grayscale image data and arranging the evaluation value so as to correspond to the coordinates of the processing region;
A black area extraction process for extracting a continuous black area whose gray value is equal to or less than a threshold value from the evaluation image created in the evaluation image creation process;
A ball bond feature amount detection process for detecting a ball bond feature amount from the black region extracted in the black region extraction process;
A ball bond inspection method comprising:
(Appendix 10)
The evaluation value calculation process includes:
The ball bond detection method according to appendix 9, wherein an evaluation value is calculated from pixel information in a circle or ellipse processing area determined as a size smaller than the inspection object.
(Appendix 11)
The evaluation value calculation process includes:
The ball bond detection method according to appendix 9 or 10, wherein a ratio of the number of pixels having a gray value range designated in advance from the image information in the processing area to the total number of pixels in the area is calculated as an evaluation value.
(Appendix 12)
The evaluation value calculation process includes:
The ball bond detection method according to appendix 9 or 10, wherein the sum of the gray values is calculated from the image information in the processing area to obtain an evaluation value.
(Appendix 13)
The evaluation image creation process includes:
The evaluation value is calculated so as to correspond to the coordinates of the processing region by executing the evaluation value calculation processing at arbitrary coordinates on the binary image created by performing binarization processing on the grayscale image data generated by the imaging processing in advance. The ball bond detection method according to appendix 9, 10, 11 or 12, wherein an evaluation image is created by arranging
(Appendix 14)
The ball feature amount detection process includes:
14. The ball bond detection method according to appendix 9, 10, 11, 12 or 13, wherein the center of gravity of the black region satisfying an area range and / or a circumscribed rectangle length range is defined as a ball bond coordinate.
(Appendix 15)
The ball feature amount detection process includes:
14. The ball detection method according to appendix 9, 10, 11, 12, or 13, wherein the center of the circumscribed rectangle of the black region that satisfies the area range or the circumscribed rectangle length range or both is used as a ball bond coordinate.
(Appendix 16)
The ball feature amount detection process includes:
Appendices 9, 10 characterized in that a ball bond outer shape is formed by extending the outer shape of the black region that satisfies the area range, the circumscribed rectangle length range, or both to the outside by the radial length of the circular filter. The ball bond detection method according to 11, 12, or 13.

  The present invention can be applied not only to a ball bond inspection apparatus for bonding wires but also to all apparatuses for inspecting the ball shape in device assembly.

11 Ball bond (inspection object)
12 Illumination (part of light source)
13 Camera (imaging means)
14 Half mirror (part of light source)
DESCRIPTION OF SYMBOLS 15 Evaluation value calculation part 16 Evaluation image creation part 17 Black area extraction part 18 Ball bond feature-value detection part

Claims (10)

An apparatus for inspecting a two-dimensional feature of a ball bond,
A light source for irradiating the inspection object with irradiation light;
An imaging unit that captures reflected light from the inspection object and generates grayscale image data;
Evaluation value calculation means for calculating an evaluation value from pixel information in a circle or ellipse processing region;
Evaluation image creating means for creating an evaluation image by executing the evaluation value calculating means at arbitrary coordinates on the grayscale image data and arranging evaluation values so as to correspond to the coordinates of the processing region;
Black area extraction means for extracting a continuous black area having a gray value equal to or less than a threshold value from the evaluation image created by the evaluation image creation means;
Ball bond feature quantity detecting means for detecting a ball bond feature quantity from the black area extracted by the black area extracting means;
A ball bond inspection apparatus characterized by comprising: The evaluation value calculation means includes
The ball bond detection device according to claim 1, wherein an evaluation value is calculated from pixel information in a processing area of a circle or an ellipse determined as a size smaller than the inspection object. The evaluation value calculation means includes
3. The ball bond detection device according to claim 1, wherein a ratio of the number of pixels having a gray value range designated in advance from the image information in the processing area to the total number of pixels in the area is calculated as an evaluation value. . The evaluation image creating means includes
The evaluation value calculation unit is executed at arbitrary coordinates on the binary image created by performing binarization processing on the grayscale image data generated by the imaging unit in advance, and the evaluation value corresponds to the coordinates of the processing region. The ball bond detection device according to claim 1, wherein an evaluation image is created by arranging the two. The ball bond feature amount detection means includes:
5. The ball bond detection according to claim 1, wherein a center of gravity of the black region satisfying at least one of an area range, a circumscribed rectangle length range, and a roundness is set as a ball bond coordinate. apparatus. A method for inspecting a two-dimensional feature of a ball bond,
An irradiation process for irradiating the inspection object with irradiation light;
An imaging process that captures reflected light from the inspection object and generates grayscale image data;
An evaluation value calculation process for calculating an evaluation value from pixel information in a processing area of a circle or an ellipse;
An evaluation image creating process for creating an evaluation image by executing the evaluation value calculation process at an arbitrary coordinate on the grayscale image data and arranging the evaluation value so as to correspond to the coordinates of the processing region;
A black area extraction process for extracting a continuous black area whose gray value is equal to or less than a threshold value from the evaluation image created in the evaluation image creation process;
A ball bond feature amount detection process for detecting a ball bond feature amount from the black region extracted in the black region extraction process;
A ball bond inspection method comprising: The evaluation value calculation process includes:
The ball bond detection method according to claim 6, wherein an evaluation value is calculated from pixel information in a processing area of a circle or an ellipse determined as a size smaller than the inspection object. The evaluation value calculation process includes:
8. The ball bond detection method according to claim 6, wherein a ratio of the number of pixels having a gray value range specified in advance from the image information in the processing area to the total number of pixels in the area is calculated as an evaluation value. . The evaluation image creation process includes:
The evaluation value is calculated so as to correspond to the coordinates of the processing region by executing the evaluation value calculation processing at arbitrary coordinates on the binary image created by performing binarization processing on the grayscale image data generated by the imaging processing in advance. The ball bond detection method according to claim 6, wherein an evaluation image is created by arranging the positions. The ball feature amount detection process includes:
10. The ball bond detection according to claim 6, 7, 8 or 9, wherein the center of gravity of the black region satisfying at least one of an area range, a circumscribed rectangle length range, and a roundness is defined as a ball bond coordinate. Method.

Images