JPS5916407B2 - Manufacturing method of semiconductor device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method of manufacturing a semiconductor device, and more particularly to a method of manufacturing a semiconductor device using semiconductor pellets adhered to an adhesive sheet.

Generally, semiconductor pellets such as transistors, diodes, C, etc. are formed by dividing a single semiconductor wafer in which a large number of elements are formed in a lattice shape into individual elements.

1 and 2 show the most common method of dividing a semiconductor wafer into semiconductor pellets by using a stretchable adhesive sheet. In this method, first, as shown in FIG. 1, a semiconductor wafer is placed on an adhesive sheet 1 called a masking sheet or electron tape.
・Paste 2 and form the cut groove 3 with a scniving,
Further, by bending and pressing with a rubber roller, fractured surfaces 4 are formed along the kerfs 3 to form individual semiconductor pellets 5. Thereafter, the adhesive sheet is stretched in all directions by an appropriate method, and the semiconductor pellets 5 are arranged at intervals as shown in FIG. In this state, the semiconductor pellets are aligned in the same direction, making them easy to store, transport, and handle in the next process. 3 to 5 are disclosed, for example, in Japanese Patent Publication No. 46-35215, and show an example of a method for assembling a semiconductor device using semiconductor pellets 5 in the state shown in FIG. ,
In this method, as shown in FIG. 3, an uncured or semi-cured epoxy resin 7 is first applied onto a support substrate 6 on which semiconductor pellets are to be attached, and then an adhesive sheet 1 is pasted onto this. The semiconductor pellets 5 are made to face each other.

Next, as shown in FIG. 4, the semiconductor pellet 5 is pressed down from above the adhesive sheet 1 with a pressing rod 8, and the semiconductor pellet 5 is bonded to the substrate 6 with the epoxy resin 7. As a result, the adhesive force between the semiconductor pellet 5 and the substrate 6 becomes greater than the adhesive force between the adhesive sheet 1 and the semiconductor pellet 5, and the pressing rod 8 is separated from the adhesive sheet 1 as shown in FIG. If the distance between the adhesive sheet 1 and the substrate 6 is widened accordingly, the semiconductor pellet 5 will be peeled off from the adhesive sheet 1. The method described above has the advantage of being able to efficiently and accurately mount semiconductor pellets because it does not require the orientation of the semiconductor pellets, the front or back of the semiconductor pellets, or the alignment work. There are some drawbacks. (a) The adhesive force between the semiconductor pellet 5 and the substrate 6 must be adjusted so that it is greater than the adhesive force between the semiconductor pellet 5 and the adhesive sheet 1, and a wide range of careful considerations are required. Even if this consideration is taken, variations in adhesive strength are unavoidable, and the semiconductor pellets 5 may not be peeled off smoothly from the adhesive sheet 1, and may be lifted while being adhered to the adhesive sheet, or may be mounted on a substrate shifted from a predetermined position. There is.

(b) When it is necessary to heat the substrate, the heat may cause the adhesive sheet to soften and lose its elasticity, causing problems.

Therefore, in order to prevent the adhesive sheet from being heated above a certain temperature, special considerations are required regarding the heating temperature, heating method, time for pressing down the semiconductor pellet with a pressing rod, etc. The case where the semiconductor pellet 5 is bonded to the substrate 6 with the epoxy resin 7 and then peeled off from the adhesive sheet 1 has been described above, but a similar method may be used, for example, in JP-A No. 47-2427.
The method disclosed in Japanese Patent Application No. 1, in which the bumps of a 7-lip chip are brought into close contact with a connecting conductor and the flip chip is peeled off from an adhesive sheet has similar drawbacks.

Also, Special Publication No. 45-4145, Special Publication No. 48-361
No. 10 and Japanese Unexamined Patent Publication No. 49-124970 disclose a method in which an adhesive sheet is pushed up by pressure to make the semiconductor pellet easier to peel off, and the semiconductor pellet is adsorbed with a vacuum suction chuck (tweezers). However, this method has the disadvantage that the number of work steps using a vacuum suction chuck is increased, and when the semiconductor pellet is a 7-lip chip, the semiconductor pellet must be turned upside down. SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a method for manufacturing a semiconductor device that solves the above-mentioned drawbacks.

That is, the present invention includes a step of releasably adhering a semiconductor pellet to an adhesive sheet, and a step of adhering a semiconductor pellet to the main surface of the semiconductor pellet on the side opposite to the main surface on the adhesive sheet side (in the main text, the electrode portion is also a part of the main surface). ) is in close proximity to or in contact with a placing portion for placing the semiconductor pellet, and the semiconductor pellet is arranged substantially perpendicular to the main surface of the adhesive sheet side of the semiconductor pellet. a step in which the needle body penetrates the adhesive sheet, and further presses the main surface of the semiconductor pellet on the adhesive sheet side with the tip of the needle body to peel the semiconductor pellet from the adhesive sheet; The present invention relates to a method for manufacturing a semiconductor device, which includes a step of separating a needle from the semiconductor pellet and the adhesive sheet, and includes peeling the semiconductor pellet from the adhesive sheet and separating the semiconductor pellet from the adhesive sheet. The present invention relates to a method for easily and accurately placing semiconductor pellets on a placing section. A method of manufacturing a semiconductor device according to an embodiment of the present invention will be described below with reference to the drawings.

FIGS. 6 to 11 show, in order of process, an embodiment in which the present invention is applied to face-down bonding on a substrate of a 7-lip chip.

In this method, first, an adhesive sheet 11 shown in FIG.
A semiconductor pellet, ie, a flip chip 12, is connected to the semiconductor pellet. This can be formed by a method similar to that described in FIGS. 1 and 2. As the adhesive sheet 11, a transparent, single-sided adhesive vinyl chloride resin thin film, such as a masking sheet sold by Tsukasa Shokai Co., Ltd., or one called Electron Tape (trade name) is suitable. This adhesive sheet 11 is attached to the annular frame 13 with a stopper 14.
is fixed. The flip chip 12 is the bump electrode 1
5, and is bonded so that the main surface 16 including the bump electrode 15 is on the lower side. The other main surface 17, which is adhered to the adhesive sheet 11, is an inert surface because the flip chip has a brainer structure. Furthermore, this fringe tip 12 has a size of 1×1×0.3 m, that is, 1 wm square, and has a total thickness of 0.3 including the bump electrode 15! It is formed to be m. The bump electrode 15 has a structure in which a metal layer such as copper is used as a base and solder is applied thereon. In FIG. 6, which schematically shows the manufacturing apparatus, the frame 13 holding the adhesive sheet 11 is vertically movable by a moving mechanism (not shown).

Note that the downward movement of the frame 13 is restricted by a stopper 18, and the upward movement range is determined by the relationship with the needle body 19. The needle body 19 penetrates the adhesive sheet 11 and directly presses one main surface 17 of the seven-ribbon chip 12, and has a tip diameter of 20 to 100 μm and a tip angle of 10 μm.
~30 points and a root diameter of 0.3~1- are suitable.

This needle body 19 is attached to a support body 27 via a lever 26, and is movable vertically by a moving mechanism (not shown). The spring 26 always applies a downward biasing force to the needle body 19, but its downward movement is restricted by a stopper 27a provided on the support body 27. 20 is a holding stand, and flip chip 12
It holds a substrate 21 for mounting the device.

Although this holding stand 20 is configured not to move up and down in the example, it may be configured to move up and down. The substrate 21 is for forming a hybrid integrated circuit device, and has wiring conductors 22 formed in a predetermined pattern on a ceramic board 21a. In this embodiment, the wiring conductor 22 is composed of an Ag--Pd thick film conductor 23, a solder layer 24, and a solder flux 25, as shown in detail in FIG. When the preparations shown in FIG. 6 are completed, the frame body 13 is lowered until it abuts against the stopper 18 as shown in FIG.
Let it be m. This interval is approximately 2 times the thickness of the flip tip 12.
It is desirable that it be at least twice as large. This includes the flip chip (not shown) already attached to the substrate 21 and the adhesive sheet 11.
This is because it is convenient to proceed with the work without contacting the flip chip 12 which is still adhered to the flip chip 12.
However, if this distance is too large, it becomes difficult to align the flip chip 12 with the wiring conductor 22 using a microscope or the like, so it is desirable to limit the distance to about 2 rm at maximum. The positioning of the flip-tip 12 with respect to the wiring conductor 22 is performed by moving the frame 13 and the stopper 18 in the horizontal direction in the state shown in FIG. When the alignment is completed, the needle body 19 is lowered to attach the flip chip 12.
The tip 19a of the needle body 19 is placed on the adhesive sheet 11 corresponding to the approximate center of the main surface 1r of the main surface 1r, and by further lowering the needle body 19, the adhesive sheet 11 is elastically stretched as shown in FIG. Press down on 12.

Normally, the frame body 13 is moved so that the bump electrode 15 of the flip-chip can come into contact with the wiring conductor 22 by lowering the needle body 19.
However, if the bump electrode 15 cannot be contacted, fine adjustment of the stopper 18 or holding base 20 is required.
is finely adjusted so that the bump electrode 15 is in contact with the wiring conductor 22. The pressure on the seven-lip chip 12 by the needle body 09 may be such that the bump electrode 15 is lightly pressed against the wiring conductor 22. 7 lip chips 1 via adhesive sheet 11
2 is pressed by the needle body 19, the spring 26 is compressed, and a pressing force is applied by the spring 26. In the state shown in FIG. 9 in which the bump electrode 15 mentioned above is in contact with the wiring conductor 22, the top layer of the wiring conductor is the flux 25 as shown in FIG. Retained.

Further, in this state, the needle body 19 has not yet penetrated the adhesive sheet 11. After the flip chip 12, which may penetrate through the wiring conductor 22 in this state, has been placed on the wiring conductor 22, the frame 13 and the adhesive sheet 11 are raised as shown in FIG.

That is, the needle body 19 and the substrate 21 are substantially fixed. In other words, the distance between the needle body 19 and the substrate 21 is substantially fixed, and the distance between the substrate 21 and the adhesive sheet 11 is approximately 0.3.
The frame body 13 is raised so as to expand from the key to 3 to 15 rvn. If the needle body 19 is fixed as described above and the adhesive sheet 11 is moved from the tip 19a of the needle body toward the base, the needle body 19 will be in a state where it penetrates the adhesive sheet 11. Furthermore, if the adhesive sheet 11 is moved upward, the 7-lip chip 12 bonded thereto also tends to move upward.
However, the upward movement of the 7 lip chip 12 is restricted by the tip 19a of the needle that penetrates the adhesive sheet 11, and the main surface 17 is pressed by the tip 19a of the needle. The sheet 11 is peeled off, and only the adhesive sheet 11 rises together with the frame 13. At this time, an upward force is applied to the needle body 19 as the adhesive sheet 11 rises, but since the spring 26 is strong enough to overcome this force, the needle body 19 hardly moves upward. do not. If the needle body 19 penetrates the adhesive sheet 11 as shown in FIG.
As shown in the figure, the needle body 19 is again pushed downward by the biasing force of the spring 26, thereby causing the flip-tip 12 to move against the needle body 19.
Pressed by After peeling off the seven-rib chip 12 and placing it on the substrate, the needle body 19 is moved upward as shown in FIG.

Although this creates a void 28 in the adhesive sheet 11, it does not become an obstacle when attaching the remaining flip chips. Even if the flip tip 12 is not pressed by the needle body 19, it is maintained in a temporarily fixed state by the adhesive force of the flux 25. When the needle body 19 is removed from the adhesive sheet 11, the adhesive sheet 11 returns to its flat state in which it is stretched around the frame as shown in FIG. 6 due to its elasticity. The frame 13 raised in FIG. 10 is lowered until it is limited by the stopper 18 for the next flip chip installation. Of course, the needle body 19 may be removed from the adhesive sheet 11 after the frame body 13 is lowered as shown in FIG. When bonding a plurality of flip chips 12 to a single board 21, the above operations are repeated to place the required number of 7 flip chips on the board, and then the board is passed through an electric furnace called a reflow oven. , the solder of the bump electrodes 15 and the wiring conductors 22 is reflowed (remelted) to bond the 7-lip chip 12 to the substrate 21.

The method of the embodiment described above has the following advantages.

(a) Since the pellet, that is, the flip chip 12 is forcibly peeled off from the adhesive sheet 11 by the needle body 19, the 7-rib chip 12, the adhesive sheet 11, and the flip chip 12 are separated from each other.
There is no need to consider the magnitude of the adhesive force between the substrate 21 and the substrate 21. That is, the adhesive force between the 7-lip chip 12 and the substrate 21 may be very small, and it is also possible to carry out the method even when the flip chip is simply placed on the substrate and no adhesive force acts between them. In addition, there is a large tolerance for variations in the adhesive strength between the 7-lip chip and the adhesive sheet, so
It is easy to select the adhesive for the adhesive sheet. (5) There are fewer restrictions on the flatness of the substrate 21 and the parallelism between the substrate 21 and the adhesive sheet 11 or chip 12.

That is, since the needle body 19 presses the tip 12 from above the stretchable adhesive sheet 11, even if the substrate 21 is slightly inclined, the tip 12 can be inclined to match the inclination. (c) Even when it is necessary to heat the substrate 21, there are fewer restrictions on the heating temperature, heating method, etc.

That is, the tip 12 is peeled off from the adhesive sheet 11 and the needle body 19
state in which the tip is brought into contact with the substrate (first
Even if the substrate is heated in the state shown in Figure 0), the adhesive sheet 1
Since it is unnecessary to prolong the period in which the adhesive sheet is in contact with the substrate by the X needle while being adhered to the substrate 1, that is, in a state where the adhesive sheet shown in FIG. 9 is easily affected by the heat of the substrate, The negative effects of this can be minimized. Of course, since the adhesive force between the chip and the substrate may be small, it is often sufficient to heat the substrate at a lower temperature than in the past. Furthermore, the step of clamping the chip to the substrate and the heating step of firmly bonding the chip to the substrate can be performed separately, and by using this method, restrictions on heating are substantially lifted. (6) From the above, the chip adhered to the adhesive sheet can be reliably and easily attached to the substrate.

(6) In addition, in this embodiment, the biasing force is applied to the needle body 19 by the spring 26, so if the amount of descent of the needle body 19 by the moving mechanism becomes larger than the desired value, this is absorbed by the spring 26. This prevents the flip tip 12 from being destroyed.

Although one embodiment of the present invention has been described above, the present invention is not limited to this embodiment and can be further modified.

For example, in the step shown in FIG. 9, instead of lowering the needle body 19, the substrate 21 is raised on the holding table 20 to bring the flip chip 12 into contact with the substrate 21, and the needle body 19 is moved through the adhesive sheet 11. Alternatively, the flip chip 12 may be pressed. Further, in the step shown in FIG. 10, the needle body 19 and the substrate 21 may be lowered instead of raising the adhesive sheet 11.

Further, the needle body 19 may be moved only by the moving mechanism without applying the biasing force by the bar 26.

In such a case, after the tip 12 is peeled off from the adhesive sheet 11, the needle body 19 will not follow the tip 12 and descend;
Since the amount of descent of the chip 12 is less than the thickness of the adhesive sheet 11, no problem arises in practice. Further, in the process shown in FIG. 10, a stopper is provided to prevent the needle body 19 from moving upward, and this stopper is activated only when the adhesive sheet 11 is raised as shown in FIG. It may be possible to prevent the needle body 19 from following the upward movement by moving to .

Furthermore, it is possible to obtain a product by simply bonding the flip chips 12 to the adhesive sheet 11 without using the method shown in FIGS. It's okay.

Further, the adhesion of the pellets to the adhesive sheet is not limited to adhesion using an adhesive, and may be electrostatic adhesion or other adhesion.

Further, the number of needle bodies 19 is not limited to one, but may be a plurality of needles for one pellet.

When using a plurality of needle bodies, it is preferable to use three needle bodies from the viewpoint of stable pressing. Further, the tip shape of the needle body 19 may simply be a general needle tip shape having a predetermined radius of curvature;
For example, the tip may have a ten-shaped tip shape so that the same effect as pressing by a plurality of needle bodies can be obtained. Moreover, it is applicable not only to attaching flip chips but also to die bonding of semiconductor pellets such as general diodes and transistors.

Further, the substrate 21 may be replaced by an insulating substrate or a metal substrate without the wiring conductor 22, or a metal plate that also serves as a lead electrode.

In short, any type of mounting section may be used in the present invention. Therefore, it can be applied not only to the ponding process but also to other processes. For example, it can be applied to a process where semiconductor pellets are simply peeled off and arranged from an adhesive sheet. In such a case, the substrate 21 in the embodiment serves as a mounting plate or arrangement plate for the semiconductor pellets. Further, in the embodiment, when peeling the chip 12 from the adhesive sheet 11, the chip 12 is peeled off with the chip 12 in contact with the substrate 20, but when the chip 12 is peeled off from the adhesive sheet 11, the bump electrode 15 is separated from the wiring conductor 22, the adhesive sheet 11 is raised or the needle body 19 is lowered so that the needle body 19 penetrates the adhesive sheet, and then the needle body 19 penetrates the main surface 17 of the chip. Press directly to insert tip 12.
may be peeled off from the adhesive sheet 11. In this case, after peeling, the chip 12 falls under its own weight, but the distance it falls is so small that no problem occurs.Also, in the several modifications described above, it is possible to obtain almost the same effect as the above-mentioned embodiment. I can do it.

[Brief explanation of the drawing]

FIG. 1 is a side view showing a state in which a semiconductor wafer is divided into semiconductor pellets, FIG. 2 is an explanatory side view showing a state in which the adhesive sheet shown in FIG. 1 is stretched, and FIGS. Fifth
The figures are side views showing the conventional semiconductor pellet mounting method in the order of steps;
FIG. 1 is a partial longitudinal cross-sectional view showing a method for installing a seven-lip chip according to an embodiment of the present invention in the order of steps. In the symbols used in the drawings, 11 is an adhesive sheet, 12 is a 7-lip chip, 13 is a frame, 15 is a bump electrode, 16 and 17 are main surfaces, 18 is a stopper, 19 is a needle body, and 20 is a A holding stand, 21 a substrate, and 22 a wiring conductor.

Claims (1)

[Scope of Claims] 1. A step of releasably adhering a semiconductor pellet to an adhesive sheet, and a mounting portion on which the main surface of the semiconductor pellet on the side opposite to the main surface on the adhesive sheet side is on which the semiconductor pellet is placed. a state in which a needle body arranged substantially perpendicular to the main surface of the semiconductor pellet on the adhesive sheet side penetrates the adhesive sheet; a step of peeling the semiconductor pellet from the adhesive sheet by pressing the main surface on the side of the adhesive sheet with a tip of the semiconductor pellet, and a step of separating the needle from the semiconductor pellet and the adhesive sheet after the peeling. A method for manufacturing a semiconductor device, the method comprising manufacturing the device. 2. The method of manufacturing a semiconductor device according to claim 1, wherein the semiconductor pellet is a flip-chip type semiconductor element. 3. The method of manufacturing a semiconductor device according to claim 1 or 2, wherein the mounting portion is a connecting conductor for the semiconductor pellet. 4. The method of manufacturing a semiconductor device according to claim 3, wherein the connecting conductor is a conductor having a solder layer on which flux for solder bonding is applied. 5. The method of manufacturing a semiconductor device according to claim 1, wherein the placing part is a part to which the semiconductor pellet is fixed, which is made of a metal substrate or an insulating substrate. 6. The step of peeling off the semiconductor pellet is performed by moving the adhesive sheet in a direction away from the placement portion while keeping the distance between the needle and the placement portion substantially constant, so that the needle is removed from the adhesion. The semiconductor according to claim 1, which is a step of penetrating the sheet and further pressing the main surface of the semiconductor pellet on the adhesive sheet side with the tip of the needle body to peel the semiconductor pellet from the adhesive sheet. Method of manufacturing the device. 7. The step of peeling off the semiconductor pellet is performed by keeping the distance between the needle body and the placement part substantially constant and moving the tip of the placement part and the needle body in a direction away from the adhesive sheet. The step of peeling off the semiconductor pellet from the adhesive sheet by putting the needle through the adhesive sheet and pressing the main surface of the semiconductor pellet on the adhesive sheet side with the tip of the needle 2. A method for manufacturing a semiconductor device according to item 1. 8. The step of bringing the main surface of the semiconductor pellet opposite to the main surface on the adhesive sheet side into a state in which the main surface of the semiconductor pellet is in proximity to or in contact with the placing part for placing the semiconductor pellet is the step of placing the adhesive sheet with the needle body. 2. The method of manufacturing a semiconductor device according to claim 1, further comprising a step of pressing the semiconductor pellet through a holder so that the main surface of the semiconductor pellet approaches or contacts the mounting portion.