JPH11233550A - Wire bonding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent erroneous bonding by generating a detecting area from an image inputted into an image processing operation circuit, storing a predetermined identifiable position other than a lead within the detecting area as a reference position, and detecting a first lead position using the reference position as a reference. SOLUTION: A first reference point 15px, which is stored in an image processing operation circuit in advance by photographing a fishing pin 15p serving as a first reference, is detected. With the point 15px as a starting point, the adjacent leads 15px to 15xe are sequentially detected to thereby generate a detecting area. Then, a camera is moved to the next detecting point by driving an XY table, thereby setting the leads 15xe to 15xi as the next detecting area. Other leads are similarly detected sequentially and, using the last detected lead as a new reference point, the leads are detected to thereby detect the positions of all the leads. Therefore, the pin 15p and any position other than that of a lead can be used as a reference as long as they can be discriminated from the adjacent leads on the image, and thus erroneous detections can be prevented.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a wire bonding method.

[0002]

2. Description of the Related Art When a semiconductor integrated circuit (IC) or a large-scale integrated circuit (LSI) is manufactured, as shown in FIG.
After positioning a lead frame LF on which a semiconductor pellet S as a semiconductor component is mounted on a table T as a bonding stage (also referred to as a stage), a bonding tool (not shown) holding a wire W is connected to a lead frame LF as a frame. By relatively displacing the semiconductor pellet S with respect to the semiconductor pellet S, the wire W is guided to a lead post (hereinafter, also referred to as a lead) LP provided on the lead frame LF and a pad (electrode) SP of the semiconductor pellet S to perform bonding. Do. The actual bonding points of the lead frame LF and the semiconductor pellet S positioned on the table T are generally shifted from the normal bonding points previously determined as coordinates on an XY table (not shown).

As a method of detecting and correcting the amount of displacement of the bonding point, two fixed points of the lead post LP are determined in advance, and the actual position of the two fixed points is detected and compared with the predetermined two fixed points. The actual bonding position on the lead post LP is calculated by detecting the shift amount.

In the case of detecting and correcting the amount of deviation between the two fixed points of the lead post LP, variations in the relative positions of the respective lead posts LP are not taken into account, but are determined at predetermined positions using the two fixed points. The position of each lead post LP is detected by adding the shift amount.

[0005]

In the conventional wire bonding method, when the image pickup means moves to a predetermined position of the lead post LP and tries to detect it, the lead post LP to be detected and the adjacent lead post are used. L
Even when P is relatively shifted, an arbitrary position is set as the reference position if the detection position is within the width of the lead post LP. Then, by storing in advance the relative positional relationship between the lead post LP at the reference position and the adjacent lead post, the lead post L continuously adjacent to the lead post LP at the reference position is stored.
The position of P is detected.

However, in recent years, the number of lead posts LP has increased with the miniaturization and high integration of semiconductors, the intervals between the lead posts LP have become narrower, and
Tend to be smaller. Such detection of the position of the lead post LP is performed by digitizing an image obtained from an imaging unit such as a camera and performing image processing. The detection by this image processing is easy to detect when the width of the lead post LP is large, but when the width of each lead post LP becomes narrow, the detection area of the image of each lead post LP obtained by the imaging means becomes small. It may be difficult to detect the position of the post LP.

As a result, when the image pickup means detects between the lead post LP to be detected and an adjacent lead post LP, that is, between the read posts LP, the image processing of the image obtained from the image pickup means is performed. There is a possibility that the reference position of the reference lead post LP cannot be accurately obtained. That is, the reference position of the lead post LP serving as the first reference may be erroneously detected as a position on the adjacent lead post LP being a normal position.

Accordingly, when the adjacent lead post LP is set as a reference position, the lead post adjacent to the lead post LP, and the lead post LP adjacent to the lead post LP are further set. . . . As described above, since the positions of the lead posts LP are sequentially obtained from the immediately preceding position information, if the position of the first lead post LP is erroneously detected, the bonding connection is performed based on the erroneous position information. This results in bonding failure.

An object of the present invention has been made in view of the above-mentioned drawbacks of the prior art, and it is possible to reliably detect the position of each lead of a frame having a narrow lead having multiple pins. It is to provide a wire bonding method.

[0010]

According to the present invention, a frame on which a semiconductor component is mounted is placed on a stage, and a tool for holding the wire is displaced relative to the frame and the semiconductor component. In the wire bonding method of bonding the leads provided on the frame and the electrodes on the semiconductor component, respectively, the image captured by the image capturing means is digitized and input to an image processing circuit, and A detection area is generated from an input image, and a specific part other than the lead in the detection area is stored as a reference position, and a first lead position is detected based on the reference position. It is.

Further, the present invention is to detect adjacent lead positions sequentially based on the first lead position.

The first reference position of the present invention is a fishing pin (tie bar).

Further, the first lead of the present invention is a lead adjacent to the fishing pin (tie bar).

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a wire bonding method according to the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing a circuit configuration of a wire bonding apparatus using the wire bonding method according to the present invention, and FIGS. 2 and 3 are explanatory diagrams for explaining a detection method of the wire bonding method according to the present invention.

As shown in FIG. 1, a wire bonding apparatus includes a timing circuit 1 and a camera 2 as an image pickup means.
, An amplifier 3, an A / D conversion circuit 4, a monitor 5 as display means, an image processing operation circuit 6, and an XY table drive circuit 7.

The timing circuit 1 is composed of an oscillation circuit 1a for generating a predetermined reference signal and a synchronization signal generation circuit 1b for generating horizontal and vertical synchronization signals. The synchronization signal generation circuit 1b is a reference signal for the oscillation circuit 1a. Is input to the camera 2 as an imaging means. Camera 2 is 2
A lead frame 15 as a frame on a bonding stage (also called a stage) 9 and / or a semiconductor pellet 16 as a semiconductor component is mounted on an XY table (not shown) movable in the dimension direction.

Image information picked up by the camera 2 is amplified by an amplifier 3 and converted from an analog signal to a digital signal by an A / D conversion circuit 4. This A / D conversion circuit 4
As shown in FIG. 4, for example, as shown in FIG. 4, the portion of the lead post LP is lightened and read as a high level value as a digital value or "1" at the time of binarization processing, and the other portions are darkened and digitalized. The value is set to a low level value or “0” during the binarization process.

The A / D conversion circuit 4 inputs the digitized image information directly to the monitor 5 as display means and also to the image processing operation circuit 6. The image processing operation circuit 6 has a built-in memory as a storage means, a CPU as a central processing unit, and the like, and performs pattern matching between image information newly obtained by imaging by the camera 2 and previously set image information. And the like, and control for synchronizing the control of the camera 1 with the synchronization signal from the timing circuit 1 and the like.

Next, the wire bonding method according to the present invention will be described below.

As shown in FIG. 1, a lead frame 15 as a frame on which a semiconductor pellet 16 as a semiconductor component is mounted is automatically transferred by a transfer means (not shown) onto a bonding stage 9 for performing wire bonding and positioned. Then, the XY table driving circuit 7 drives a XY table (not shown) movable in a two-dimensional direction to drive the camera 2 in a memory in the image processing operation circuit 6 at the time of self-teaching to set conditions and the like. Move to a predetermined regular position.

Then, the positions of two predetermined fixed points on the lead frame 15 are detected. The predetermined two fixed points, that is, the normal position and the position newly obtained by the camera 2 are compared with each other, the amount of deviation from the normal position is detected by the image processing operation circuit 6, and the relative position of the lead frame 15 is determined. Perform position correction. In the correction using the two fixed points, the relative position of the lead frame 15 is detected and the entire position is corrected. However, in the correction using the two fixed points, the shift amount of the bonding point for each lead is not corrected.

Next, correction for each lead of the lead frame 15 is performed using the camera 2 or the like as an image pickup means. This correction is referred to as correction by read locating.

As shown in FIG. 2, the width of each of the leads 15xa to 15xe and 15ya to 15ye is digitally processed and obtained as image information. It is designed to detect. Therefore, as shown in FIG. 2, the relocate detects, for example, the position of Oxa, Oxb (Oya, Oyb).

In the detection by the lead locate, that is, the position detection of the bonding point, the position of the predetermined regular lead, which is the first reference, is the lead 1.
Assuming that the actual position detected by moving the camera 2 is, for example, Oxa1 (or Oya1), the image processing operation circuit 6 determines that the actual position is the second lead 15xb (or lead 15ya). 15yb) may be determined to be a predetermined regular lead 15xa (or 15ya) serving as an initial reference.

If the 15xb is erroneously detected as being the first reference point, the read locate is performed for the next lead 15xb.
c, 15xd, and 15xe are sequentially shifted one by one and detected. The same applies when 15yb is erroneously detected as the first reference point.

In such detection by read locating, if it is determined that an erroneous position is a regular bonding point, bonding will result in a bonding failure if bonding is performed as it is.

Therefore, in the wire bonding method according to the present invention, even if the width of the leads and the width between the leads are too small to capture even an image, the bonding position serving as the first reference point is reliably captured and detected. Can be done.

That is, as shown in FIG. 2, the lead frame 15 is mounted on an island 15L on which the semiconductor pellet 16 is mounted, and the island 15L is connected to a fishing pin (tie bar) 15p as a support. Supported. This fishing pin 15p and each lead 15xa to 15x
An opening 18 is formed between e and 15ya to 15ye. Therefore, the fishing pins 15p are usually formed at four corners. This fishing pin 15p
Compared to 5xa to 15xe and 15ya to 15ye, they are not easily deformed, and can be easily identified because they are different in thickness and length from each lead. In addition, even when the detection area is narrowed by increasing the magnification of the camera 2, it is possible to reliably distinguish and distinguish the leads 15xa to 15xe and 15ya to 15ye.

Therefore, in this wire bonding method, an arbitrary position of the fishing pin 15p is set as a first reference point at the time of self-teaching or the like and is stored in the memory in the image processing operation circuit 6. Then, when detecting the lead locate, the camera 2 first detects a predetermined regular position 15px of the fishing pin 15p.

Correspondence between the fishing pin 15px serving as the first reference point and the adjacent lead post 15xa or 15ya,
That is, the adjacent lead post 15xa or 15ya
Is set in advance and stored in a memory in the image processing operation circuit 6 composed of a microcomputer or the like from the fishing pin 15px.

A method of detecting the positions of the leads 15xa to 15xe (or 15ya to 15ye) from the first reference point 15px will be described with reference to FIG.

As shown in FIG. 3A, first, a fishing pin 15p serving as a first reference is imaged by the camera 2 and the first reference point 15p stored in the image processing operation circuit 6 in advance.
Detects px. The fishing pin 15p is different in length and thickness from the leads 15xa to 15xe (or 15ya to 15ye), and can be easily identified by image processing. In addition, since the fishing pin 15p is relatively strong and hardly deformed, it is hard to be erroneously detected, and the leads 15xa to 1x
It is suitable for detecting 5xe (or 15ya to 15ye).

When the first reference point 15px is detected, the image processing operation circuit 6 performs the image processing on the first reference point 15px.
The adjacent leads 15xa to 15xe (or 15ya to 15ye) are sequentially detected starting from 5px.
(B) As shown in FIG.
It is determined that the ranges xa to 15xd are ranges that can be collectively processed on the same screen, and the leads 15xa to 15xd are detected. The range shown in FIG. 3B is called a detection area. This detection area is generated in the image processing circuit 6 from the image input from the camera 2 to the image processing circuit 6.

The positions of the leads 15xa to 15xd detected in the detection area are stored in the image processing operation circuit 6. When the detection of the leads 15xa to 15xd is completed, the image processing operation circuit 6 switches to the XY table driving circuit 7
To drive the XY table (not shown) to move the camera 2 to the next detection point.

As shown in FIG. 3C, the lead 15xe following the lead 15xd is detected, and the leads 15xe to 15xi are set as the next detection screen, that is, the next detection area. The leads 15xe to 15xi in the next detection area are sequentially detected similarly to the leads 15xa to 15xd.

In the same manner, a range that can be collectively processed is set in a range including the last read always detected in the previous detection area, and a read last detected in the previous detection area is replaced with a new one. Each lead is detected as a reference point, and the positions of all leads are detected.

In this embodiment, a fishing pin (tie bar) is used as an initial reference point for detecting a lead. However, if it is possible to distinguish an adjacent lead on an image, any fishing rod other than the fishing pin and the lead may be used. It is also possible to use the position as a reference.

Further, if a plurality of reference points such as a fishing pin (tie bar) used as the first reference point in the present embodiment and which can be distinguished from adjacent leads on the image are provided, the detection can be made more reliable. .

The detection of the read location according to the present embodiment includes a case where the detection is performed manually using a microscope or the like, and a case where the detection is automatically performed using imaging means such as a camera.

[0041]

As described above, according to the present invention,
It is possible to reliably and continuously detect the position of a lead without erroneous detection based on the correspondence between specific position information that can be easily distinguished from others and an adjacent lead. Therefore,
Erroneous bonding can be prevented, and the yield can be improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing a circuit configuration of a wire bonding apparatus using a wire bonding method according to the present invention.

FIG. 2 is an explanatory diagram illustrating a detection method of a wire bonding method according to the present invention.

FIG. 3 is an explanatory diagram illustrating a detection method of a wire bonding method according to the present invention.

FIG. 4 is a diagram illustrating a conventional wire bonding method.

[Explanation of symbols]

 Reference Signs List 1 timing circuit 1a oscillation circuit 1b synchronization signal generation circuit 2 camera 3 amplifier 4 A / D conversion circuit 5 monitor 6 image processing operation circuit 7 XY table drive circuit

Claims (4)

[Claims] 1. A frame on which a semiconductor component is mounted is arranged on a stage, and a tool for holding the wire is displaced relative to the frame and the semiconductor component so that the wire is provided on the frame. In the wire bonding method of guiding and bonding to electrodes on the semiconductor component, an image captured by an imaging unit is digitized and input to an image processing circuit, and a detection area is determined from the image input into the image processing circuit. A wire bonding method comprising: generating and storing a specific part other than the lead in the detection area as a reference position; and detecting a position of a first lead based on the reference position. 2. The wire bonding method according to claim 1, wherein adjacent lead positions are sequentially detected based on the first lead position. 3. The wire bonding method according to claim 1, wherein the first reference position is a fishing pin (tie bar). 4. The wire bonding method according to claim 1, wherein the first lead is a lead adjacent to the fishing pin (tie bar).