JPS6143726B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a method for detecting positional deviation of a minute member such as a semiconductor pellet with high precision, and a method for detecting positional deviation of a member such that accurate wire bonding can be performed by compensating for the positional deviation, and the use thereof. The present invention relates to a wire bonding device.

When manufacturing semiconductor devices, wire bonding is an indispensable technique for wiring between the device and the electrodes of the package. In order to automate this wire bonding process, various wire bonding devices have been developed. On the other hand, with recent improvements in semiconductor manufacturing technology, semiconductor patterns have become increasingly fine and complex. At the same time, with the increase in integration, it has become necessary to perform many wiring lines (wire bonding) on microscopic semiconductor devices. In order to perform such fine wire bonding with high precision, accurate position detection of the semiconductor devices and the like is an important point. Therefore, there are many position detection methods,
Further, wire bonding devices using this method have been proposed, but each has various problems, for example, it becomes a complicated and large-scale device and becomes expensive. In addition, it takes a long time to detect the position, resulting in poor productivity.

The present invention has been made in consideration of these circumstances, and its purpose is to provide small parts such as semiconductor pellets with fine patterns.
It is possible to detect positional deviations with high precision and in a short time by excluding the adverse effects of defects such as defects that occur on the scribe surface of semiconductor pellets, and to correct the bonding position based on the detection results to ensure accurate wire bonding. The object of the present invention is to provide a method for detecting positional deviation of a member, which can simplify the device, automate it, save labor, and contribute to improving productivity, and a wire bonding device using the method.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A positional deviation detection method and a wire bonding apparatus using the method will be described in detail below based on an embodiment of the present invention.

FIG. 1 is a schematic diagram of a wire bonding apparatus. An XY table 2 is provided on the base 1. This XY table 2 includes a Y table 2y whose movement is controlled and positioned in the Y direction in the figure by a pulse motor 3, and a Y table 2y provided on this Y table 2y, whose movement is controlled in the X direction in the figure by a pulse motor 4. The X-table 2x is configured by an X-table 2x that is positioned by On the XY table 2, an arm 6 is provided which is rotatable in the vertical direction by a support base 5. A cam follower 7 consisting of a rotating body is provided at the rear end of this arm 6.
is installed. This cam follower 7 is
The arm 6 is moved up and down in contact with a cam 9 rotated by a motor 8 fixed on the XY table 2, and therefore the arm 6 moves up and down around the support shaft of the support base 5. The front end of this arm 6 protrudes from the surface of the XY table 2, and a bonding capillary 10 is attached to the end. This bonding capillary 10
is the bobbin 11 fixed to the XY table 2
A wire 12 wound around the wire 12 is supplied to perform wire bonding of a semiconductor pellet, which is a position detection member to be described later and also serves as a workpiece to be bonded. Further, a support member 13 is fixed to the XY table 2. This support member 13 holds a light projector 14 and a photodetector 15 composed of a photoelectric conversion element or the like. This floodlight 1
4 and the photodetector 15 will be explained later.

On the other hand, a frame pedestal 16 is installed on the side of the base 1, that is, below the bonding capillary 10 that is attached to the arm 6 and protrudes from the surface of the XY table 2. This frame stand 16
A lead frame 18 having a predetermined number (six in this case) of the semiconductor pellets 17 arranged at equal intervals is fed onto the lead frame 18 in a stepwise manner. This frame stand 16
Wire bonding is performed on the semiconductor pellet 17 sent upward by the bonding capillary 10. That is, semiconductor pellet 1
7 is mounted on a lead frame 18, and sent one frame at a time onto the frame stand 16 in a stepwise manner. The position of the bonding capillary 10 facing the sent semiconductor pellets 17 is controlled by the movement of the XY table 2, and each semiconductor pellet 17 is sequentially wire bonded. After wire bonding has been performed on all semiconductor pellets 17, a lead frame 18 carrying a new semiconductor pellet 17 is sent, and wire bonding is performed in the same manner.

Now, the light projector 14 irradiates light onto the semiconductor pellet 17 located directly below the capillary 10 from diagonally above. As shown in FIG. 2, this projector 14 transmits light from a light emitter (lamp) 14a through a slit hole 14b and a lens 14c.
The light is focused and irradiated by the irradiation method. This irradiated light is focused on a flat piece by the action of the slit hole 14b and the lens 14c, and in FIG. 2, it is flattened in the direction of arrow A and elongated in the direction of the front and back of the page. ing. When the semiconductor pellet 17 is irradiated with such light, the light irradiated onto the pellet 17 is reflected at a point P in the figure.
Further, the light that has left the pellet 17 is reflected by the surface Q of the lead frame 18 on which the pellet 17 is placed, and takes an optical path as shown by the broken line in the figure. Therefore, the area irradiated with the light is shifted due to the light cutting effect, as shown, for example, in the shaded area B in FIG. 3a. The photodetector 15 scans and detects the image of the irradiated area B(p) on the semiconductor pellet 17 caused by this light cutting effect and the area including this, and detects the deviation of the semiconductor pellet 17 from its normal position. To detect. That is, the photodetector 15 includes a lens 15a that collects reflected light from the irradiation area B, and a one-dimensionally arranged light receiving element array 15 that detects the collected light.
It consists of b. Then, the light intensity between specific widths l and m including the irradiation area B of the pellet 17 is detected as shown in c in FIG. Thus, the light intensity at the detection position (between l and m in the figure) detected by the photodetector 15 is as shown in FIG. 3b, for example. Returning to FIG. 1, this signal is input to the position detector 21. Based on this position detection result, the bonding capillary 10 is positioned relative to the pellet 17 (correction of deviation), and the bonding described above is performed.

Next, a method for detecting the position of the pellet 17 by the above-mentioned optical detection, that is, a method for detecting the amount of deviation from the normal position, will be explained.

The semiconductor pellet 17 has a predetermined pattern 17a formed thereon as shown in FIG. 3a. The reflectance of the region 17b excluding the pattern 17a is different from that of the pattern 17a, and therefore can be easily identified from the difference in the amount of detected light. Therefore, the scanning range (between l and m) is now set so that each side of the pellet 17 is scanned at equal angles. At this time, from the detection signal level of the photodetector 15, as shown in FIG. Distance can be easily determined. This can be achieved, for example, by scanning the scan widths l, m at a constant speed with the position detector 21, and counting high-speed clock pulses each time the detection level changes due to a pattern edge. Thereafter, the length information a, b, c is input to the bonding control device 23 via the interface 22. This control device 23 is composed of a microcomputer or the like equipped with an arithmetic circuit, a comparator circuit, etc., and is equipped with an external storage circuit 24 to perform predetermined arithmetic processing. That is, the scanning ranges l and m are
3, when the pellet 17 is in the normal position, the lengths a and c match, and the length b
is set so that it becomes a predetermined value b _p . Therefore, by performing predetermined arithmetic processing from the length information a, b, and c obtained by detecting the position of the pellet 17, the deviation amounts ΔX and ΔY can be calculated. The above-mentioned positional deviation amounts ΔX and ΔY are respectively expressed by the following equations.

ΔX=a−c/2・α……(1) ΔY=b _p −b/2・tanθ・α……(2) However, α depends on the quantization accuracy of the detection position by the photodetector 14. θ indicates the angle (angle at which each pattern edge intersects) with the scanning direction. That is, the amount of deviation in the X direction from the center position (regular position) is calculated using equation (1), and the amount of deviation in the Y direction is calculated using equation (2).

However, in general, the pattern 17a is often close to the end of the pellet 17, and if there is a chipped part at the end due to scribing or the like, or due to the roughness of the scribing, the pattern 17a may be close to the end of the pellet 17. If a
A signal as shown by the dotted line is added. Therefore, errors occur in the detected positional deviation amounts ΔX and ΔY. In other words, detection is performed in a state as shown in FIG. 3c. It is difficult to say that the semiconductor pellet 17 is completely free of the above-mentioned defective portions, and therefore the position detection accuracy is poor.
Performing bonding in this state is undesirable because it leads to the production of defective products. Therefore, in the method of the present invention, the above-mentioned error is eliminated in the following manner by taking advantage of the fact that the scattered light reflection due to defects or convex portions of the pellet ends is localized.

That is, as shown in FIG. 4, the scanning ranges l and m are
It is moved in a direction perpendicular to the scanning detection direction (here, the Y direction), and detection of C ₁ , C ₂ , and C ₃ is performed. Then, the positional deviation amounts ΔX ₁ , ΔY ₁ , ΔX ₂ , respectively detected by the above-mentioned scans C ₁ , C ₂ , C ₃ ,
ΔY ₂ , ΔX ₃ , and ΔY ₃ are stored and held in the memory circuit 24. Based on these deviation amounts, pellet 1
7, correct positional deviation amount detection calculation is performed. now,
Assuming that there is no defect in the pellet 17, the following equation holds true.

ΔX ₁ ΔX ₂ ΔX ₃ (3) ΔY ₁ ΔY ₂ +dyΔY ₃ +2dy (4) Here, the symbol () takes into consideration the error element due to α. When the above equations (3) and (4) hold, for example, the positional deviation amounts ΔX and ΔY are ΔX=ΔX ₁ +ΔX ₂ +ΔX ₃ /3 (5) ΔY=ΔY ₁ +ΔY ₂ +ΔY ₃ +3dy/3 ……
…(6) can be shown as. In this case, now ΔX ₁
When ΔX ₃ \ΔX ₂ or ΔY ₁ ΔY ₃ \ΔY ₂ , assuming that there is an influence of the pellet edge at the C ₂ position,
Using only two detection deviation amounts ΔX ₁ , ΔY ₁ , ΔX ₃ , ΔY ₃ ΔX=ΔX ₁ +ΔX ₃ /2 ………(5)′ ΔY=ΔY ₁ +ΔY ₃ +3dy/2……(6)
′. That is, the average value of two or more sets of detection deviation amounts having approximately equal values is used. In this way, by sequentially repeating scans until at least two or more sets of approximately equal detection deviation amounts are obtained, scans related to missing pellets 17 are excluded from calculations, and therefore errors caused by the above-mentioned defects can be eliminated. Can be removed. Further, errors caused by the quantization accuracy of the photoelectric detection position can also be alleviated by performing the above-mentioned averaging process.

Therefore, the positional deviation amount Δ detected as above
X and ΔY are applied as positional deviation correction amounts to an X-direction position control device 25 and a Y-direction position control device 26, and the pulse motors 3 and 4 are driven to position the XY table 2. At this time, the motor control device 27 is energized by the control device 23, and the motor 8 is rotated, the cam 9, the cam follower 7, and the arm 6 are moved up and down, and the wiper 1
Bonding of the second pellet to the pellet 17 is performed.
When the bonding of the pellets 17 for one frame is completed, the control device 23 drives the frame step feed control device 28 to move the semiconductor pellets 17 placed on the next lead frame 18 onto the frame stand 16. lead to. Then, wire bonding is performed in the same manner.

Thus, according to the device configured as described above, the amount of positional deviation of the semiconductor pellet 17 can be easily and accurately detected by detecting the irradiated image formed on the semiconductor pellet 17 by the light projector 14 due to the light cutting effect. be able to. Moreover, at this time,
Based on the magnitude of the detected value b, a rough positional deviation correction of the semiconductor pellet 17 is performed, and then accurate positional deviation correction can be performed by calculating the above-mentioned equations (1) and (2). In addition, the above calculation is performed multiple times by shifting the position for scanning detection by dy in the direction perpendicular to the scanning direction, and the positional deviation amounts ΔX, ΔY calculated from those detected values are calculated within a certain specified range. Only those within the range are selected and used as correction amounts. Therefore, error factors such as pattern defects can be removed from the correction amount, and accurate positioning can always be performed. Furthermore, as described above, this detection method is simple and accurate, and can be measured in a short time. Moreover, it can be achieved with a simplified measurement system and can be realized at low cost. Therefore, a series of processes such as positional deviation amount detection, positional deviation amount correction, and wire bonding can be automatically performed, thereby increasing productivity and saving labor.

In addition, the positional deviation detection method itself projects clear linear light from diagonally above to detect the detected member (pellet 1).
7) Formed on the light irradiation surface with a light cutting effect,
This is an extremely simple method of detecting by scanning in a straight line. Furthermore, since the scanning is performed by shifting a plurality of times in a direction perpendicular to the scanning direction, detection errors due to pellet (pattern) defects can be eliminated, and the effect is tremendous.

Note that the present invention is not limited to the above embodiments. For example, the photodetector 15 may use a linearly arranged photo array to detect the amount of light received at each array position, and scan by moving one photoreceiver and an optical system that guides the light to this photoreceiver. It may be configured as follows. Further, a device such as a microcomputer may be used as the arithmetic processing system. It goes without saying that rotational deviation of the detected member (pellet 17) may also be detected and corrected. In short, the present invention can be implemented with various modifications without departing from the gist thereof.

[Brief explanation of the drawing]

FIG. 1 is a schematic configuration diagram of a wire bonding apparatus showing an embodiment of the present invention, and FIGS. 2 to 4 are
FIG. 6 is a diagram for explaining a method for detecting a positional deviation of pellets in the apparatus of the same embodiment. 2...XY table, 10...Bonding/
Capillary, 12... wire, 14... floodlight,
15...Photodetector, 17...Semiconductor pellet, 1
8... Lead frame, 21... Position detector, 2
3... Bonding control device (arithmetic circuit, comparison circuit), 24... Memory circuit, 25, 26... Position control device, a, b, c... Detection value, ΔX, ΔY...
Amount of positional deviation.

Claims (1)

[Claims] 1. Light is irradiated from diagonally above to form a light irradiation surface due to a light cutting effect on the member to be detected, and a specific width including this light irradiation surface is linearly scanned to perform light detection. In the method of detecting the amount of deviation of the member from a predetermined position, the scanning is performed multiple times by shifting the scanning position in a direction perpendicular to the scanning direction until at least two or more substantially equal positional deviation detection values are obtained. , the average value of the positional deviation detection values that are substantially equal to each other obtained by these scans or the positional deviation detection value of any one of them is used as the positional deviation amount of the member from the predetermined position. Method for detecting positional deviation of members. 2. A mounting table on which a member to be detected is placed, an XY table arranged on the side of this mounting table, and a device provided on the XY table, which irradiates the member to be detected with light from diagonally above to detect the member. a light projector that forms a light irradiation surface by a light cutting effect on the top; a photoelectric detector that is installed on the XY table and linearly scans a specific width including the light irradiation surface; scanning position setting means for setting the movement in a direction perpendicular to the scanning position; an arithmetic circuit that calculates the amount of positional deviation of the member from the detection results of the photoelectric detector at each scanning position set by the means; and this arithmetic circuit. The scanning position is moved by the scanning position setting means and scanning is performed by the photoelectric detector until at least two or more substantially equal positional deviation amounts are detected by comparing the calculated positional deviation amounts with each other. and a table position control device that moves and sets the XY table position by inputting the average value of two or more substantially equal positional deviation amounts detected by the comparison circuit, or the positional deviation amount of any one of them. and a wire bonding mechanism provided on the XY table and wire bonding to the member at a position determined by the table position control device.