JP2002318263A - Method of inspecting trace of probing needle - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method of inspecting a trace of probing needle capable of readily and accurately determining only the trace of a probing needle applied on a pad irrespective of the surface condition of the pad. SOLUTION: Image data A, B is obtained before and after the trace 3 of the probing needle is applied on the pad 31 to be inspected. The data A is compared with the data B to obtain image data C of only the trace 3, and then the trace 3 of the probing needle applied on the pad 31 to be inspected is specified.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of measuring electrical characteristics by bringing a measuring needle of a tester into contact with a pad of a semiconductor chip, and the trace of the measuring needle attached to the pad is within an allowable range on the pad. The present invention relates to a needle mark inspection method of a measuring needle for inspecting whether or not there is a needle mark.

[0002]

2. Description of the Related Art As shown in FIG. 3, a large number (several hundreds to thousands) of semiconductor chips T are formed on a semiconductor wafer W. FIG. 4 is an enlarged plan view of the semiconductor chip T. The semiconductor chip T is provided with a plurality of pads 1 with which a measurement needle contacts when measuring the electrical characteristics of the semiconductor chip T. FIG. 5 is an enlarged plan view of the pad 1. The pad 1 is made of, for example, an aluminum material, and a protective film 2 is formed around the pad 1 for the purpose of insulating the other parts from other parts.

The measurement of the electrical characteristics of the semiconductor chip T is performed by bringing a measuring needle attached to a test head of a tester into contact with the pad 1. Measuring needle is pad 1
After the contact, the pad 1 is slid so as to scrape it. As a result, the two are brought into an appropriate pressure contact state, and sufficient electrical conduction between them is ensured. In this state, the tester applies a predetermined electric signal to the semiconductor chip T via the measuring needle and the pad 1.
And receives the return signal from the semiconductor chip T to measure the electrical characteristics of the semiconductor chip T. When the measurement is completed, the measuring needle separates from the pad 1, but the needle mark 3 of the measuring needle remains on the pad 1.

After the measurement of the electrical characteristics, an inspection of the needle mark 3 is performed. This is because if the measuring needle comes into pressure contact with a position shifted from the pad 1 (the needle mark 3 at this time is shown in FIG. 4).
In this case, sufficient electrical continuity is not ensured, resulting in inaccurate measurement. In addition, the measurement needle penetrates the protective film 2 around the pad 1 or cracks 4 are formed on the protective film 2. Inspection is performed to confirm the position of the needle mark 3 attached to the pad 1 because it causes the product to be defective.

In order to perform the needle mark inspection, it is necessary to specify which needle mark 3 is on the pad 1. Conventionally,
As one method of specifying the needle mark 3, there is a method of comparing a reference pad (a state before the contact measurement with the measuring needle) with a state after the contact measurement of the inspection target pad with the measuring needle is performed.

This will be described with reference to FIG.
First, one reference pad 11 is selected from a plurality of pads, the state of the reference pad 11 before contact measurement by a measuring needle is photographed by a CCD camera, and the image data A 'is stored in a memory. As the reference pad 11, a pad having a relatively good surface condition without stains or dust on the pad surface is selected. The image data is processed as binarized image data in which the shading is white (0 as an assigned integer) and black (0 as an assigned integer). That is, each pixel is represented as a white pixel “0” and a black pixel “1” with a certain density value as a threshold. Binarization reduces the amount of image data, simplifies processing, and saves memory. In this case, a portion without a needle mark or a stain on the pad surface is represented as a background of a white pixel “0”, and a needle mark or a stain is represented as a black pixel “1”.

Then, after measuring the electrical characteristics by bringing the measuring needle into contact with the pad 21 to be inspected, the pad 21 after the contact measurement is photographed with a CCD camera, and the image data B 'is stored in the memory. I do.

Next, the image data A 'of the reference pad 11 is
The needle mark 3 is identified by comparing the image data B ′ of the inspection target pad 21 with the image data B ′ and detecting these differences. Specifically, the CPU reads both image data A ′ and B ′ from the memory and performs a logical operation (EXOR operation) for each corresponding pixel. That is, a certain pixel (pixel in the m-th row and the n-th column) corresponding between both image data A ′ and B ′ is a white pixel “0” and a white pixel “0”, or a black pixel “1” and a black pixel “
If they are 1 ", the pixel in the m-th row and the n-th column is set to a white pixel" 0 ", and if each corresponding pixel is different such as a white pixel" 0 "and a black pixel" 1 ", The pixel at the m-th row and the n-th column is processed as a black pixel “1” to obtain image data C ′, that is, the image data C ′ does not appear in the image data A ′, but appears in the image data B ′. Is the image data extracted as the black pixel “1.” Since the difference between the two image data A ′ and B ′ before and after the contact measurement of the measuring needle is only the needle mark 3, the image data C ′ In the above, the mark represented as a black pixel "1"
Can be recognized as

Next, image data D 'is obtained by overlaying (synthesizing) the image data C' thus obtained with a predetermined allowable range L. The permissible range L is synthesized with the image data C ′ with the outline as a black pixel “1”. Then, for example, the operator looks at the image data D ′ with a microscope, and if the needle mark 3 is within the allowable range L as shown in the figure, it is determined that the measuring needle has been correctly contacted with the pad 21 without causing a positional shift.

[0010]

In the above method, the reference pad 11 to be compared with the pad 21 to be inspected is arbitrarily selected, and therefore the two to be compared are not the same pad. Accordingly, the inspection target pad 21 has stains and dust 5
a to 5c (shown by dashed lines, but actually
(Appearing as black pixels above) are also extracted as differences between the image data A ′ and B ′. That is, spots and dust 5a to 5c also appear as black pixels in the image data C 'and are recognized as needle marks. Since the stains and dust 5a to 5c are outside the allowable range L in the image data D ', the original needle trace 3 is outside the allowable range L even though the original needle trace 3 is within the allowable range L. Is determined to be defective.

Further, there is a method of performing a filtering process on the image data B 'based on the difference between the density values of the needle mark 3 and the stains or dust 5a to 5c to make the objects other than the needle mark 3 difficult to see. However, in the image data subjected to the binarization (white / black) processing, it is difficult to distinguish light and dark spots and dusts 5a to 5c close to the needle mark 3 from the needle mark 3 no matter how much the filter processing is performed. In many cases, it cannot be determined. In order to make this filter processing effective, the number of gradations (density resolution) of the image data may be increased. However, in this case, the amount of information of the image is increased, which hinders the speeding up of the processing, and also reduces the memory capacity. It also takes up a lot of storage capacity.

Therefore, conventionally, the inspection target pad 2
Only the image data B 'after the first contact with the measuring needle is acquired, and the image data B' is visually inspected by an operator using a microscope. That is, the needle mark 3 is specified and distinguished from spots, dusts 5a to 5c, other surface abnormalities, and the like based on the shape and size, and depends on the experience and ability of the operator. This places a burden on the operator and also takes time. Further, for example, if there is a stain formed so as to be connected to the needle mark 3 or a stain having the same shape and the same size as the needle mark 3, it is difficult to identify which is the needle mark.

The present invention has been made in view of the above-described problems, and provides a needle mark inspection method of a measuring needle which can easily and accurately specify only a needle mark attached to a pad without being affected by the surface condition of the pad. That is the task.

[0014]

According to the present invention, image data before and after a trace of a measuring needle is attached to a pad to be inspected is acquired, and these image data are compared to obtain a pad attached to the pad to be inspected. Identify needle traces.

With this, even if the surface of the pad after the contact measurement with the measuring needle has stains or dust, surface irregularities, etc.
Since these are present in the inspection target pad even before the contact measurement, they are canceled by taking the difference between the two image data, and as a result, only the needle trace can be extracted.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A needle mark inspection method according to an embodiment of the present invention will be described below with reference to FIGS.

First, the measurement needle is positioned in a non-contact state with respect to the inspection target pad (step S1 in FIG. 1).

Next, as step S2, image data A shown in FIG. 2 is obtained. This is image data obtained by photographing the inspection target pad 31 before contact with the measuring needle with, for example, a CCD camera. As the image data, each pixel is processed as binarized image data of a white pixel “0” and a black pixel “1” as in the related art. Also in this case, a portion without a needle mark or a stain on the pad surface is expressed as a background of a white pixel “0”, and a needle mark or a stain 5d is expressed as a black pixel “1”. Then, the image data A is stored in the memory.

Next, the measurement needle is brought into contact with the pad 31 to be inspected to measure the electrical characteristics (Step S3).
The pad 31 after the contact measurement is photographed with a CCD camera,
This image data B is stored in the memory (step S
4).

Next, in step S5, the CPU as the image processing means reads both image data A and B from the memory and performs a logical operation (EXOR operation) for each corresponding pixel. That is, certain pixels (pixels in the m-th row and the n-th column) corresponding to both image data A and B are white pixels “0” and white pixels “0”, or black pixels “1” and black pixels “1”.
If they are different, the pixel in the m-th row and the n-th column is set to a white pixel “0”, and if the corresponding pixels are different such as a white pixel “0” and a black pixel “1”, The image data C is obtained by processing the pixels in the n-th column as black pixels “1”. The image data C thus obtained is stored in the memory.

By the above processing, the difference between the two image data A and B appears in the image data C. That is, since the stain 5d appears in both the image data A and B before and after the contact with the measuring needle, it is represented as a background of the white pixel “0” on the image data C by the above-described processing.
Needle mark 3 not shown in the image data B but shown in the image data B
Only the black pixel “1” is extracted on the image data C. Thus, the needle mark 3 attached to the pad 31 can be accurately and quickly specified without being affected by the surface condition of the inspection target pad 31.

Next, in step S6, the CPU reads out the image data C from the memory and
Over the predetermined allowable range L (combined)
Obtain image data D. The permissible range L is the outline of a black pixel.
1 "is synthesized with the image data C. Then, for example, when the operator looks at the image data D with a microscope and the needle mark 3 is within the allowable range L as shown in the figure, the measurement needle does not cause a positional shift and operates normally. It is determined that the pixel has been touched by the pad 31. Alternatively, the CPU may determine whether a pixel located outside the allowable range L has data of a black pixel “1”. Inspection process (Steps S1 to S
Over 6), automation without human intervention can be realized.

Also, the process from acquisition of image data A to contact measurement with a measuring needle to acquisition of image data B after the contact measurement is not performed intermittently for one pad, but the contact measurement of the measuring needle is performed with a certain pad. During this time, if the image data of the pad to be measured next before the contact with the measuring needle is acquired, the inspection of multiple pads can be performed continuously, thus shortening the inspection time. Can be achieved.

Further, the image data A and B are not photographed one pad at a time, but by using a high-resolution CCD camera.
Simultaneously put together multiple pads formed on one chip,
Alternatively, by simultaneously photographing a plurality of pads formed on one wafer, the inspection process can be further shortened. Furthermore, if a comparison between the two image data A and B and a process of determining the limit of the needle mark 3 obtained by this comparison are simultaneously performed for a plurality of pads by using a CPU having a higher processing capability, a further time can be obtained. Shortening can be achieved.

In the above-described binarization processing, white pixels may be assigned to an integer value "1", and black pixels may be assigned to an integer value "0". Alternatively, the image data may be processed as image data having a number of gradations larger than two gradations. In this case, the comparison between the image data A and B in step S5 is performed by simply subtracting the density value (integer value) of each corresponding pixel between the image data A and B. Becomes 0, and only a different part, that is, the needle mark 3 can be extracted as a certain integer value.

The comparison between the image data A and B is performed by the CPU.
Alternatively, the needle trace may be specified by visual comparison of the operator. Even in this case, as shown in FIG. 2, the image data B ′ (FIG. 6) is visually observed and the needle mark 3 is identified by distinguishing it from stains and dust based on the shape and size. The needle mark 3 can be easily and reliably specified by comparing the image data A and B even visually.

[0027]

According to the first aspect of the present invention, the needle trace can be easily specified without being erroneously recognized, and an accurate needle trace inspection can be performed.

According to the second aspect of the present invention, the operation of specifying the needle trace can be automated, and the burden on the operator can be reduced.

According to the third or fourth aspect of the present invention, the time required for the needle mark inspection can be reduced.

[Brief description of the drawings]

FIG. 1 is a flowchart showing a procedure of a needle mark inspection method according to an embodiment of the present invention.

FIG. 2 is a schematic diagram illustrating the concept of a needle mark inspection method according to an embodiment of the present invention.

FIG. 3 is a plan view of a semiconductor wafer.

FIG. 4 is an enlarged plan view of a semiconductor chip.

FIG. 5 is an enlarged plan view of a pad.

FIG. 6 is a schematic view showing the concept of a conventional needle mark inspection method.

[Explanation of symbols]

2 ... protective film, 3 ... needle mark, 31 ... pad, L ... allowable range, T ... semiconductor chip, W ... semiconductor wafer.

Continued on the front page F term (reference) 2G003 AA07 AF02 AF03 AG03 AG12 AG13 AH05 AH07 2G011 AC02 AC14 AE03 2G051 AA51 AB02 CA03 CA04 EA08 EA11 EA14 EB01 EB02 ED07 4M106 AA01 AA20 BA10 CA38 DB04 DB21 DJ18 DJ21 DJ21

Claims (4)

[Claims] 1. A method for inspecting a needle mark of a measuring needle attached to a pad formed on a semiconductor chip, the method comprising: acquiring image data before and after the needle mark is attached to a pad to be inspected. A method of inspecting a needle mark of a measuring needle, wherein the image data is compared to specify the needle mark attached to the pad to be inspected. 2. A method according to claim 1, wherein a difference between the image data before and after the needle mark is added is detected by an image processing means, and image data showing only the needle mark on the pad is obtained based on the detection result. The needle mark inspection method for a measuring needle according to claim 1. 3. When the measurement needle is in contact with the pad to be inspected and the measurement is being performed, image data before contact of the measurement needle with the next pad to be inspected is acquired. The needle mark inspection method for a measuring needle according to claim 1. 4. The method according to claim 1, wherein at least one of the image data before and after the needle mark is attached is acquired simultaneously for a plurality of pads.