Patent Application
20100072262
1. Field of the Invention
The present invention relates to wire bonding methods that optically recognize a position of a bonding point on a semiconductor device and bond a wire to the bonding point.
2. Description of Related Art
A conventional wire bonding method will be described with reference to
Typically, in such a wire bonding method, first, an image pick-up device 11 detects amounts of displacement from regular positions of at least two fixed points on the semiconductor chip 2 and at least two fixed points on the lead frame 1, and then a calculation unit corrects prestored bonding coordinates based on the detected amounts of displacement. In a case in which the detection is carried out by the image pick-up device 11, an X motor 12 and a Y motor 13 are driven so that the central axis 11a of the image pick-up device 11 comes exactly above a point of measurement. After the bonding coordinates are corrected as described above, the capillary 15 is moved in an X-Y direction and a Z direction, and the wire 4 inserted through the capillary 15 is bonded between the pads P1, P2, . . . of the semiconductor chip 2 and the corresponding respective lead portions L1, L2, . . . of the lead frame 1.
In this case, since the central axis 11a of the image pick-up device 11 and the central axis 15a of the capillary 15 are offset by a distance W, the bonding coordinates are corrected after the amounts of displacement from fixed points are detected by the image pick-up device 11, and then, the X motor 12 and the Y motor 13 move the X-Y table 16 by the distance W, so that the capillary 15 is positioned above a first bonding position of the workpiece 3. Subsequently, based on the movement of the X-Y table 16 in the X-Y direction by the X motor 12 and the Y motor 13 and on the movement of the capillary 15 in the Z direction by the capillary arm 17 being moved up and down (or caused to swing) by a Z motor 14, the wire 4 is bonded at the corrected bonding coordinates. In
The above wire bonding method includes capturing an image of the bonding point and recognizing the position of the bonding point. However, in the wire bonding apparatus 10, changes in the ambient temperature caused by heat sources as well as operating heat generation produce a difference between the thermal expansion of the capillary arm 17 and of the pick-up device holding arm 19 that holds the image pick-up device 11. Consequently, the amount of offset W between the central axis 11a of the image pick-up device 11 and the central axis 15a of the capillary 15 varies. An error due to this variation is expressed as displacement of the bonding position. Generally, the detection of the displacement of the bonding position is realized by detecting central positions or displacement of balls bonded to the pads P1, P2, . . . (referred to as pressure-bonded balls) using the image pick-up device 11.
Conventionally, a wire bonding method including recognizing the position of the bonding point as described above and recognizing a post-bonding image after bonding is carried out according to the steps shown in the flowchart in
Subsequently, a semiconductor chip 2B that is a second chip on the lead frame 1 is transferred to the bonding center (the central axis 11a of the image pick-up device 11) in Transfer Step 30 and treated in the same manner as the semiconductor chip 2A. Specifically, the semiconductor chip 2B goes through Bonding Point Image Capture Step 31, Bonding Position Recognize Step 32, Recognition Complete Step 33, and then Wire Bonding Step 34. Thereafter, a post-bonding image of a pressure-bonded ball formed on the semiconductor chip 2B after bonding is captured in Post-Bonding Image Capture Step 35, and then an amount of displacement in the post-bonding image is recognized in Post-Bonding Image Recognize Step 36. After this post-bonding image recognition is completed, at Recognition Complete Step 37, if any displacement is detected, the amount of offset W described above is corrected. Examples of this type of wire bonding are disclosed in Japanese Patent Application Unexamined Publication Disclosure No. 8-31863 (Japanese Patent No. 3235008) and Japanese Patent Application Unexamined Publication Disclosure No. 9-306939 (Japanese Patent No. 3560731). As shown by the two-dot chain lines in
Another example of a technique that realizes faster wire bonding is Japanese Patent Application Unexamined Publication Disclosure No. 9-36164. In the technique disclosed in Japanese Patent Application Unexamined Publication Disclosure No. 9-36164, an amount of offset between a capillary and a lighting device having a camera (image pick-up device) is equal to the integral multiple of a pitch between semiconductor chips. In order to set the amount of offset between the capillary and the image pick-up device to the integral multiple of the pitch between semiconductor chips, the image pick-up device is attached to a bonding head such that the image pick-up device is allowed to move along the X direction. Here, when the amount of offset is, for example, equal to the pitch between semiconductor chips, capturing and recognizing an image of a first semiconductor chip that has already been wire bonded is carried out during the wire bonding to a second semiconductor chip.
For semiconductor devices for an automobile, for example, an one hundred percent inspection for pressure-bonded balls after wire bonding is a must in order to improve the reliability in the manufacturing process. However, when Post-Bonding Image Recognize Step 26 for recognizing the post-bonding image of the semiconductor chip 2A is carried out using the method shown in
In contrast, the technique disclosed in Japanese Patent Application Unexamined Publication Disclosure No. 9-36164 recognizes an image of the semiconductor chip after wire bonding during wire bonding to a second semiconductor chip. Consequently, it is possible to eliminate waste in time to a large extent, thereby improving the productivity. However, the capturing of an image after wire bonding is carried out after the next semiconductor chip is transferred to the bonding center, that is, the image of is captured after the semiconductor chip that has been wire bonded is transferred by a distance equal to the integral multiple of the pitch. This means that the position (stage) for wire bonding is different from the position (stage) for capturing of an image after wire bonding, and thus resulting in problems as described below.
Firstly, in a case in which a semiconductor chip that has been wire bonded at a wire bonding position is transferred to an image capturing position, it is practically impossible to place the semiconductor chip in exactly the same condition at the two different positions (stations) on the order of a micrometer, and it is also impossible to place a lead frame in exactly the same condition at the two different positions. Therefore, measuring the semiconductor chip at a different position from the wire bonding position (station) and correcting (feeding back to the capillary) the amount of offset based on the result of the measurement cannot provide an appropriate amount of offset, and thus failing to carry out the wire bonding accurately.
Secondly, detection conditions such as temperatures and lighting are slightly different between the wire bonding position and the image capturing position. In addition, the detection conditions can change over time slightly because the image after wire bonding is captured after the semiconductor chip that has been wire bonded is transferred. Therefore, similarly to the first problem, correcting (feeding back to the capillary) the amount of offset based on the result of the measurement cannot provide an appropriate amount of offset due to the differences in the detection conditions as described above, and thus failing to carry out the wire bonding accurately.
Thus, an object of the present invention is to provide a wire bonding method capable of further improving the accuracy in wire bonding and realizing faster wire bonding.
In order to achieve the above described object, the present invention provides a wire bonding method including the steps, sequentially carried out using a wire bonding apparatus provided with a capillary through which a wire is inserted and an image pick-up device that is offset from the capillary, of:
the step of recognizing the amount of displacement in the post-bonding image of the first semiconductor chip is carried out during wire bonding of the wire from a first bonding point to a second bonding point of a second semiconductor chip, the wire bonding being performed from a first bonding point to a second bonding point of the second semiconductor that has been corrected based on the amount of displacement for the second semiconductor, following the steps, sequentially carried out after capturing the post-bonding image of the first semiconductor chip after bonding, of:
According to the wire bonding method of the present invention, the step of recognizing the amount of displacement in the post-bonding image of the first semiconductor chip is carried out during wire bonding of the wire from a first bonding point to a second bonding point of a second semiconductor chip, this wire bonding being performed from a first bonding point to a second bonding point of the second semiconductor that has been corrected based on the amount of displacement for the second semiconductor, following the steps, sequentially carried out after capturing the post-bonding image of the first semiconductor chip after bonding, of: transferring the second semiconductor chip to the bonding center; capturing an image, using the image pick-up device, of the second bonding point on the second semiconductor chip; and recognizing a position of the second bonding point to calculate an amount of displacement of a position of the second bonding point. Accordingly, accuracy in image capturing does not decrease and faster wire bonding can be realized.
a)-2(f) are diagrams illustrating an exemplary embodiment of a flow of image processing according to the present invention;
One embodiment of the wire bonding method according to the present invention will be described below with reference to
Next, the embodiment will be described based on
Subsequently, a semiconductor chip 2B that is a second chip on the lead frame 1 is transferred to the bonding center 50 (see
Then, a semiconductor chip 2C that is a third chip on the lead frame 1 is transferred to the bonding center 50 (see
As described above, the capturing of a post-bonding image is carried out at a position at which the wire bonding to the semiconductor chips 2A, 2B, . . . is performed in Wire Bonding Steps 24, 34, . . . , respectively. Therefore, either of the first and second problems of Japanese Patent Application Unexamined Publication Disclosure No. 9-36164 as listed above does not occur. Further, Post-Bonding Image Recognize Steps 26, 36, . . . for recognizing the post-bonding images of the semiconductor chips 2A, 2B, . . . are respectively carried out during Wire Bonding Steps 34, 44, . . . to the semiconductor chips 2B, 2C, and thus faster wire bonding can be realized.
