JPS6144304A - Positioning method of semiconductor chip - Google Patents

Abstract

PURPOSE:To perform preparatory positioning highly accurately and automatically, by picking up the image of a semiconductor chip by an optical device, processing the picked up image signal, and detecting the relative amount of devia- tion of the position, where the image of the semiconductor chip is picked up, and the semiconductor chip. CONSTITUTION:Light L is emitted from a light source 64 of an optical source 58. The light L is projected through a condenser lens 62 and the like on a semiconductor chip 3, which is mounted on a tray 18 on the upper surface of an X-Y-theta table 17 of a preparatory positioning mechanism. The light L, which is regularly reflected by the semiconductor chip 3, is inputted to an ITV camera 59 through an objective lens 64. The picked up image signal, which is inputted to the ITV camera 59, is inputted to an image processing part 65. In the image processing part 65, the center of gravity of the chip 3 is computed, and the center S is obtained. The amount of deviation between the center S and the light axis O is computed. The computed value is outputted to a driving part 54 through a computer. Then X-Y tables 52 and 53 are driven, and the center S of the chip is made to agree with the light axis O of the ITV camera 59.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a positioning method for positioning a semiconductor chip before supplying the semiconductor chip to a die bonding process.

[Technical background of the invention and its problems]

Recently, many semiconductor radars have been introduced for use in computer paddle disc players and optical information processing equipment. In this semiconductor radar, a semiconductor chip is attached to a stem having leads. This chip is then enclosed by a cap that is fitted onto the stem.

In the manufacturing process of this type of semiconductor radar, a die bonding device is used to bond the chip to the stem, and the bonding of the semiconductor chip to the stem must be performed with high precision. In other words, the semiconductor chip is generally bonded with a protrusion of several μm from the end face of the stem, but since this chip is heated to about 200°C when excited to output Raded light, if the amount of protrusion is large, it may cause damage to the stem. Durability is reduced due to poor heat conduction. On the other hand, if the amount of protrusion is small, there is a problem that the LED light emitted from the chip is blocked by the stem.

Conventionally, in order to join the tip to the stem with high precision,
Before supplying the above chips to the die bonding process, they are pre-positioned and then punched.

- The above-mentioned chip was adsorbed with a pin or the like and supplied to the bonding process. However, since conventional preliminary positioning was performed visually by the operator, there was a limit to the accuracy, and there was also large variation depending on the operator. Therefore, in the bonding process, it takes time to perform the relative positioning of the stem and the chip with high precision, resulting in a decrease in work efficiency. There were cases where sufficient accuracy could not be obtained.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a method for positioning a semiconductor chip in which preliminary positioning of a semiconductor chip before a die bonding process can be performed automatically and with high precision.

[Summary of the invention]

This invention images a plurality of semiconductor chips placed on a tray individually using an optical device, and performs image processing on the image signals to determine the relative positional deviation between the pick-up position of each semiconductor chip and the semiconductor chip. This is a semiconductor chip positioning method in which each semiconductor chip is positioned by detecting the amount and controlling the position of the tray using this detection signal.

[Embodiments of the invention]

The present invention will be described below based on an embodiment shown in the drawings. Figures 1 to 3 show a die bonding device for semiconductor radar, where 11 is a Yθ table for bonding, θ is located at the upper stage, and i is provided at the center of the captive bonding position so that rotation can be controlled. It will be done. A heater (not shown) is built into the mounting space 12 on the upper surface of the Yθ table 11, and the temperature of this heater is controlled by a heater control device 13. In the mount space 12VC, a mounting portion 14 for positioning and mounting the stem 2 is formed by a recess. Then, the stem 2 is supplied to the mounting portion 14 by a stem clamp/4151C that engages with the pair of notches 2 & of the stem 2, and at the same time, the stem 2 is pressed against the reference surface of the mounting portion 14 by the push rod 16. is adapted to be positioned. On the other hand, 17 in the figure indicates the temperature of the preliminary positioning mechanism described later.
The θ table is provided with, for example, a square tray 18 fixed to a stage 18& on its upper surface.

For example, the stage 18& is fixed with a structure in which a rectangular receiving groove for the tray 18 is provided, and an open groove is provided on one side of the groove to facilitate insertion of the tray 18. The tray 18VC is provided with, for example, a rectangular accommodation recess for accommodating a large number of semiconductor chips 3, and semiconductor chips and chips 3 are arranged and placed in each of the recesses.

A chip supply mechanism 2Q is provided between the tray 18 and the placement section 14 for picking up (transporting and supplying) the chips 3 in the tray 18 to the placement section 14. To explain this chip supply mechanism 20, 21 is a guide rail.
2 is provided so as to be able to freely reciprocate. A chip holder 24 is attached to this conveying member 22 via a vertical movement mechanism 23, and a parquet bin 25 is provided on this chip holder 24. Then, as will be described later, the πθ table 17 of the preliminary positioning mechanism detects the amount of positional deviation between the pickup position and the semiconductor chip 3 for each semiconductor chip 3, and the step 3 in the tray 18, which has been preliminary positioned, is After being positioned by rotating by Ω' and θ every time, it is vacuum suctioned by the vacuum pin 25tlC and supplied to the mounting portion J4 of the mount space 12. Therefore, it is necessary to form the concave structure of the tray 18 so that the θ rotation is suppressed to a constant angle. For example, the size of the recess is approximated to the size of the semiconductor chip 3. Further, a nozzle 41 is provided near the mount space 12 for spraying cooling gas onto the joint between the stem 2 and the chip 3.

Further, a position measuring device 27 based on a light sectioning method is installed on the n' table 26 provided on the side of the Yθ table 11, facing the mount base 12Fc. and,
The relative position of the chip 3 attracted by the vacuum pin 25 and the stem 2 placed on the mounting portion 14Vc is optically measured, and the Yθ table 1 is adjusted based on the measurement signal.
1 to position the stem 2 and tip 3 at predetermined positions.

Rigini 1 Yθ table 11 and Y table 260
The control system will be explained. First, each table of the Yθ table 1 is provided with a motor 11bzlle for driving it, and each of these motors 11b and 11c is connected to a driver 28.
Connected to Combi-Y29 via. In addition, each table of the correction table 26 has a motor 2 for driving it.
6 motors 26m, 26b are provided, and these motors 26m, 26b
is connected to the combination 29 via the driver 3o. Furthermore, the position measuring device 27 is configured in Km as shown in FIGS. 4 to 6. That is, 31 is a light source, and the light from this light source 31 is passed through a condensing lens 32 and a slit JJK into a ribbon-shaped slit.

Light. This slit light is connected to the stem 2 and chip 3.
, and the optical images of the projection areas A and B are sent to the lens 34)
When the images are formed on the focus plate 35, two images a and b appear to be separated from each other as shown in FIG. As shown in FIG. 6, this amount of deviation Δd is proportional to the amount of relative positional deviation ΔtK between stem 2, chi, and f3, so the relative positional deviation Δt can be found from the amount of deviation Δd of images A and B. If the one-light cutting method is used in this way, stem 2 and chi.

It is possible to measure the relative position (Y direction) with The inclination of the stem 2 and the tip 3 in the θ direction was measured using the optical cutting method.
It can be found by trigonometry by repeating the process in more than one place. Also,
Said video 1. b is imaged by the ITV camera 37, and this image is transmitted as an electric signal to the combination via the transducer 38, frame memory 39 and interface 4Q.
- Sent on 29th VC in the evening. Then, the computer 29 calculates the relative positional deviation amount Δt, and based on this control signal, the Yθ table 11 is driven to correct Δt. That is, as shown in FIG.
The slit light is recognized by capturing the V camera image, and the position of this slit light is measured. Then, the relative positions of the tip 3 and the stem 3 relative to the slit tip tK are calculated. Here, it is determined whether the position is the set value or not, and Y
In the case of FJS, END is determined, and in the case of NO, the feedback amount of the Yθ table 11 is determined. And Yθ
The table 11 is driven and the above operation is repeated again.

In this case, since the relative position in the X direction has a large allowable value for mounting accuracy, only preliminary positioning is performed and no particular correction is performed. However, if mounting accuracy is required, correction may be performed.

On the other hand, the temperature θ table 17 of the preliminary positioning mechanism is θ
A table 51 is provided at the bottom, and a 1cY table 52 and an X table 53 are sequentially provided above this. Each table has motors 51ht52*, 5 that drive them.
3& are provided, and these motors are connected to the computer 29 via a drive section 54. In addition, the XYθ
A pedestal 55 is erected on the side of the table 17, and a mowing table 56 is provided on this pedestal 55Vc. An arm 57 is fixed to this mushroom table 56, and this arm 5
An optical device 58 is attached to the protruding end of 7. As shown in FIG.
It consists of a 'ff camera 59 and a lens system 60.

The ITV camera 59 is attached to the arm 57 with its optical axis O being vertical. A half mirror 6 is arranged below the ITV camera 59 at an angle of 45 degrees. The optical axis. A pair of condensing lenses 62 are arranged in a direction perpendicular to K, facing the half mirror 6, with a slit plate 63 in between. The illumination light emitted from the light source 64 is made incident on this condensing lens 6xK. The illumination light emitted from the condenser lens 62 is reflected vertically downward by the half mirror 61, passes through a plurality of objective lenses 64, and illuminates the semiconductor chip 3 placed on the tray 18 on the upper surface of the XY19 table 17. The illumination light regularly reflected by the semiconductor chip 3 passes through the objective lens 64 and the half mirror 6 and enters the light receiving surface 59 & of the ITV camera 59 @I'ff camera 59
The imaging signal incident on the image processing unit 6 connected to this
5 (shown in FIG. 1). This image processing section 65
calculates the deviation of the chip 3 in the XYθ directions and outputs the detected signal from the combination. Then,
A drive signal is output from the computer 29 to the drive section 54, and the drive signal causes each of the motors 51h,
52* and 53h are activated, and the θ, y, and x tables 51, 52, and 53 are driven in accordance with the amount of positional deviation of the chip f3.

A method of detecting the amount of positional deviation of each chip 3 will be explained with reference to FIG. 9. First, the correction table 5 on the pedestal 55
6 to set the optical axis O of the rTV camera 59 to the θ table 5.
Align it with the center of rotation of . In this state, when illumination light is emitted from the light source 64 and the light specularly reflected by the chip 3 enters the ITV camera 59VC, an imaging signal of the chip 3 is input from the ITV camera 59 to the image processing section 65. Then, the image processing unit 65 calculates the center of gravity of the chip 3 to find the center S, and calculates the amount of deviation between the center S and the optical axis O.

Further, the relative positions in the ruled direction of a plurality of points on the outer periphery of the tick f3, six points A to F in this embodiment, are recognized. Then, the inclination angles of the line segments AB, BC, DE, and EF are determined by trigonometry, and the rotation angle of the tip f3 is calculated from the average value of these. These calculated values are displayed on the computer 29.
When the signal is output from the drive section 54'FC, the °C table 52.degree. Next, the θ table 5 is driven to rotate so that the rotation angle of the chip 3 matches the reference value. Therefore, the position of the chip 3 in the n'θ direction can be precisely matched with the reference value. At this time, the ITV camera 59 receives the illumination light specularly reflected by fs, so if the above-mentioned f3 is tilted vertically on the tray 18, or if the tabs are not installed horizontally, , IT'/camera 59 cannot capture this reflected illumination light from step 3. Therefore, such a chip 3 can be excluded as a defective workpiece.

Next, a method of bonding the semiconductor chip 3 to the stem 2 using the bonding apparatus configured as described above will be described. First, stem clan/f
The stem 2 flanged by the bushing rod 15 is placed on the mounting part 14 of the mount base 12, and at the same time the stem 2 is flanged by the bushing rod 1.
6 against the reference surface of the mounting section 14.

Therefore, the stem 2 is set at a predetermined position on the mounting section 14. On the other hand, at this time, the mount base 12 is heated to some extent by the heater, the stem 2 is heated, and the adhesive such as indium on the surface of the chip mounting portion is melted. On the other hand, one chip 3 on the tray 18 that is bonded to the stem 2 is in a state where the positional deviation in the n'θ direction has been corrected as described above.

In this state, first, the conveying member 22 of the chip feeder groove 20 is conveyed onto the tray 18, and the chip 3 whose positional deviation has been corrected is vacuum-sucked from the tray 18 by the vacuum pin 25. Next, the conveying member 2
The chip 3, which is driven in the opposite direction and attracted to the vacuum pin 25, stops at a position facing the stem 2 on the mount base 12. Then, the tip holder 24 is lowered by the vertical movement mechanism 23 and is held in a state where the tip 3 is slightly spaced from the upper surface of the stem 2. Here, the above-described optical cutting method using the position measuring device 27 automatically aligns the stem 2 and the chip 3 so that they are in the required position at the time of mounting. After positioning the stem 2 and chip f3 at predetermined positions in this way, when the chip holder 24 is lowered by the vertical movement mechanism 23, the chip 3 attracted to the maculum bin 25 is placed on the stem 2. .

At this time, since the indium and oxide on the surface of the stem 2 are melted, the chip 3 is adhered onto the stem 2. After maintaining this state for a certain period of time, the heater is turned off by the heater control device 13, and cooling gas is spouted from the nozzle 41.
When sprayed onto the joint between the stem 2 and the chip 3, the molten indium and indium are cooled and solidified, completing the bonding.

〔Effect of the invention〕

As described above, according to the present invention, the semiconductor chips 7'' can be positioned automatically and with high precision before the semiconductor chips are transferred to the dumping process.

Since the receiver can be transferred to the dumping process with the rod positioned with high precision, it is possible to improve the efficiency and accuracy of the bending process.

[Brief explanation of the drawing]

The drawings show an embodiment of the method of the invention, and FIG.
2 is a plan view of the part in the mending process, 3 is a perspective view of the relationship between the stem and the tip, and 4 to 6 are the functions of the position measuring device. Explanatory diagram, Fig. 7 is a flowchart of the position measurement method, Fig. 8
The figure is a configuration diagram of the optical system of the preliminary positioning mechanism, and FIG. 9 is an explanatory diagram for calculating the amount of positional deviation of the chip by the optical system. 3... Semiconductor Chi, f, 17... Bow αθ Cheagle, 54
...Drive unit, 58...Optical device, 65...Image processing unit Fig. 3 Fig. 4 Fig. 5 Fig. 7, Fig. 8 Fig. 9

Claims (3)

[Claims] (1) A step of individually and sequentially imaging a plurality of semiconductor chips placed on a tray using an optical device, and a relative position between the pickup position of the semiconductor chip and the semiconductor chip for each imaging signal from the optical device. A method for positioning a semiconductor chip, comprising: detecting the amount of deviation by an image processing section; and controlling the position of the tray based on a signal from the image processing section to position each semiconductor chip. . (2) The method for positioning a semiconductor chip according to claim 1, wherein the tray has a semiconductor chip accommodating recess that suppresses a rotation angle during positioning of the semiconductor chip within a certain range. (3) The tray has a recess for accommodating semiconductor chips such that the semiconductor chips are arranged in a matrix, and the
2. The semiconductor chip positioning method according to claim 1, wherein each semiconductor chip is positioned by moving in a direction coincident with the Y direction.