JPH11297886A - Forming method of solder bump - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable a solder bump which is uniform in height to be formed on a pad even if the pads are small in pad pitch and diameter and to feed a fine ball of diameter 105 μm or so accurately to a pad. SOLUTION: A solder bump forming method comprises a process A where flux 50 is applied to pads 12 respectively, a process B where a metal mask 52 provided with openings where balls can be each put in and which are formed at positions corresponding to the pads 12 is made to overlap aligning the openings with the pads 12, a process C where many solder balls 64 prescribed in diameter are placed on the metal mask 52, a process D where solder balls 64 are each dropped down into the openings to come into contact with the flux 50, a process E where excessive solder balls 64 located out of the openings are removed by sweeping the surface of the metal mask 52 with a squeegee, a process F where the metal mask 52 is removed, and a process G where the printed wiring board 10 is heated in a reflow oven to melt the solder balls 64 to form a bump on the pads 12 respectively.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for forming a solder bump for forming a salient salient electrode used for mounting a semiconductor chip or the like on a printed wiring board at a high density.

[0002]

2. Description of the Related Art In response to high integration, multi-pin, and high-density semiconductor chips, a package using bumps (salient electrode) instead of a QFP package which is a four-side lead connection type. BGA (Ball Grid)
Array) packages are being used. That is, the BGA package is a package in which a semiconductor chip is mounted on the surface of a printed wiring board (substrate) constituting the package, and molded to form ball bumps (or solder bumps) as connection terminals on the back surface in a grid pattern. .

In this BGA package, a ball bump (or solder bump) on the lower surface is overlapped with a solder bump similarly formed on an electrode (pad) on a printed wiring board on which the package is mounted, and the solder bump is reflowed. They are joined to each other. At this time, due to the self-alignment effect due to the surface tension of the molten solder,
This is for automatically adjusting the displacement between the package and the printed wiring board.

[0004] As a method of forming solder bumps on the package side or the printed wiring board side, a cream solder printing method and a solder ball mounting method are known. FIG. 5 is a diagram illustrating a cream solder printing method. In this figure, reference numeral 10 denotes a printed wiring board (Print
Wiring Board (PWB), 12 is a pad formed on the surface, and the surface other than the pad 12 is covered with a solder resist 14.

[0005] Reference numeral 16 denotes a metal mask, and a large number of openings 18 corresponding to the pads 12 are formed by etching a metal plate from both sides. The metal mask 16 is placed on the printed wiring board 10 with the opening 18 aligned with the pad 12 and fixed (FIG. 5A). A cream solder 20 is supplied onto the metal mask 16, and the surface of the metal mask 16 is moved by applying a squeegee (spatula) 22 to roll the cream solder 20. The solder 20 is filled (FIG. 3B). Here, the cream solder 20 is obtained by mixing a powdery solder with a highly viscous flux to form a cream.

After filling the openings 18 with the cream solder 20, the metal mask 16 is peeled off from the printed wiring board 10 (FIG. 5C). At this time, the opening 18
The cream solder 20 filled in remains on the pad 12. If the cream solder 20 on the pad 12 is melted by heating the printed wiring board 10 in a reflow furnace, the melted solder is automatically adjusted to the correct position on the pad 12 by the surface tension of the solder itself. (Self-alignment effect), and becomes hemispherical. Therefore, if the solder is cooled and solidified, the solder bumps 24 having a constant height and a constant shape are formed (FIG. 5D).

In the solder ball mounting method, a flux is applied to a pad of a printed wiring board, and a metal mask having an opening corresponding to the pad is formed on the printed wiring board. After the solder balls of a predetermined size prepared in advance are supplied to the openings one by one and temporarily fixed with a flux,
Peel off the metal mask and reflow heat. In this case, the supply of the solder balls is performed using a dedicated device (referred to as a ball placer).

This device (ball placer) includes a suction head for sucking one solder ball at a position corresponding to the opening of the metal mask by negative pressure. The suction head is carried on a metal mask in a state where the solder ball is sucked by the suction head, the suction by the negative pressure is stopped, and the solder ball is dropped onto each pad to supply the solder ball.

[0009]

In the above-mentioned cream solder printing method, when the metal mask 16 is peeled from the printed wiring board 10, it is inevitable that some cream solder remains on the inner surface of the opening 18.

FIG. 6 is a view showing a state in which the cream solder remains on the metal mask 16 thus peeled off. This figure 6
Reference numeral 26 denotes cream solder remaining in the opening 18.
Since the amount of the residual cream solder 26 is not constant and non-uniform due to the openings 18, the amount of the cream solder 20 placed on the pads 12 becomes non-uniform, and the height of the solder bumps 24 becomes uneven. Occurs.

In order to solve this problem, it is conceivable to optimize the material properties such as the adhesive index of the cream solder and the separation speed of the printing press (peeling speed of the metal mask), the optimization of equipment conditions, and the stricter management. The solution was difficult. Such irregularities in the height of the solder bumps 24 may cause poor connection during mounting.

On the other hand, according to the solder ball mounting method, one solder ball is supplied to each pad, so that the amount of solder supplied to each pad can be made uniform. However, in order to supply one solder ball to each pad, a dedicated device (ball placer) is required, and there is a problem that this device is complicated and expensive.

On the other hand, with the miniaturization and higher density of electronic circuits, it is required to reduce the pad pitch and the pad diameter. For this reason, it is required that the amount of solder supplied to the pads be more strictly controlled. However, in the cream solder printing method, since the diameter of the opening of the metal mask is further reduced, it is more difficult to control the supply amount of the cream solder, and it is impossible to form a solder bump having a uniform height.

In the above-described solder ball mounting method, it is necessary to reduce the diameter of a solder ball to be used as the amount of solder supplied to a pad is reduced. It is impossible to suck small diameter solder balls.

For example, a conventional ball placer has a diameter of about 30.
Although a device capable of sucking a solder ball having a diameter of 0 μm (0.3 mm) was possible, for a solder ball having a diameter of about 105 μm, it was necessary to reduce the diameter of the hole for sucking the solder ball. However, it is extremely difficult to machine such small holes, making it impossible to fabricate such a ball placer.

[0016]

SUMMARY OF THE INVENTION The present invention has been made in view of such circumstances, and it is possible to form a solder bump having a uniform height when a pad pitch and a pad diameter are small. It is an object of the present invention to provide a method for forming a solder bump that can accurately supply a solder ball to a pad.

[0017]

According to the present invention, there is provided a solder bump forming method for forming a solder bump on each pad by supplying and reflowing one solder ball on each pad of a printed wiring board. a) a flux is applied to each pad; b) a metal mask having an opening for receiving one solder ball at a position corresponding to each pad is superimposed to align the opening with each pad; and c) a predetermined diameter. A) A large number of solder balls are placed on the metal mask. D) One solder ball is dropped into each of the openings and adhered to the flux. E) The surface of the metal mask is rubbed with a squeegee to enter each of the openings. Remove the extra solder balls, f) remove the metal mask, g)
This is achieved by a solder bump forming method comprising the steps of heating a printed wiring board in a reflow furnace and melting solder balls to form solder bumps on each pad.

A substantially frustoconical opening is formed in the metal mask used here, and the surface of the metal mask having an opening on the small diameter side of the opening is brought into close contact with the printed wiring board, and the opening on the large diameter side of the opening is formed. It is desirable to supply solder balls from the openings. A metal mask having such a truncated conical opening can be made as follows.

That is, a photo-curable (eg, ultraviolet-curable) resist is applied to a substrate such as a stainless steel plate, and the pad position is exposed and cured to form a resist column.
A metal mask is formed on the substrate by forming a metal plating layer by thickening electrolytic nickel plating, eluting and removing the columns of the resist, and removing and removing the substrate.

In this case, the portion which is cured into a column shape by light (ultraviolet) exposure has a large diameter on the front side and a back side (a small diameter on the side in close contact with the substrate).

This metal mask has a thickness of about 120 μm (0.12 μm) for a solder ball having a diameter of 105 μm.
0 mm), and the opening diameter of the opening is about 130 μm on the small diameter side.
(0.130 mm), about 160 μm (0.160 mm) on the large diameter side
mm).

[0022]

[Function] A solder ball controlled to a predetermined diameter is supplied to a metal mask on a metal mask, and if necessary, horizontal or vertical vibration is applied to the solder ball controlled to a predetermined diameter. Drop them one by one into the controlled openings and attach them to the flux and temporarily fix them.

The upper surface of the metal mask is rubbed with a squeegee to remove excess solder balls. If the solder ball is reflowed after the metal mask is removed, the molten solder rises in a hemispherical shape and can be cooled to form a solder bump. Here, since the diameter of the solder ball supplied to each pad is constant, the height of the solder bump is also uniform.

[0024]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is an explanatory view of the first half of a process according to an embodiment of the present invention, FIG. 2 is an explanatory view of the latter half of the process, FIG. 3 is an enlarged view of a portion indicated by an arrow III in FIG. It is a manufacturing process drawing. In FIG. 1, reference numeral 10 denotes a printed wiring board,
12 is a pad, and 14 is a solder resist. This printed wiring board 10 may be a printed wiring board that is a part of a package for accommodating a semiconductor bare chip, or a printed wiring board for mounting an IC package having BGA terminals on its lower surface. Good.

A flux 50 is applied to the surface of the printed wiring board 10 and is collected on the pad 12 without the resist 14, that is, on the pad 12 by utilizing the surface tension (FIG. 1A). A metal mask 52 is overlaid on the printed wiring board 10 (FIG. 1B). This metal mask 52 is made by the method shown in FIG. In FIG. 4, reference numeral 54 denotes a conductive substrate made of a thin stainless steel plate, and one surface thereof is coated with an ultraviolet curing resist 56. The resist 56 may be formed using a photosensitive dry film.

The resist 56 is exposed to ultraviolet light to cure only the position of the pad 12. For example, a negative film on which the pad 12 is imprinted is overlaid on the resist 56 and exposed from above to cure only the position corresponding to the pad 12. At this time, the resist 56 hardens deeply in the thickness direction from its surface, so that the cross-section is hardened so as to have a substantially inverted truncated conical shape. Therefore, if the uncured resist is removed with a solvent, an inverted truncated conical column 58 is formed (FIG. 1B).

The substrate 54 is placed in a nickel plating bath and subjected to electrolytic nickel plating (FIG. 1C). The plating layer 60 is formed to a thickness corresponding to the thickness of the metal mask 52. That is, about 120 μm (0.120 mm)
It is thickly plated. It is needless to say that the thickness of the resist 56 is applied to a sufficient thickness in consideration of the thickness of the plating layer 60.

After the formation of the plating layer 60, the portions of the resist pillars 58 are eluted by the solvent (FIG. 1D). When the plating layer 60 is peeled off from the substrate 54, the metal mask 52 is completed by the plating layer 60. In the metal mask 52, a truncated conical opening 62 is formed by a cavity remaining after the column 58 is eluted.

In this embodiment, a solder bump having a height of 70 μm is formed on a pad 12 having a pitch of about 0.25 mm and a diameter of about 0.13 mm. In this case, the opening 62
As shown in FIG. 3, the diameter on the large diameter side is 160 μm (0.
16 mm) and the diameter on the small diameter side is 130 μm (0.13 mm).
mm). The thickness of the metal mask 52 is preferably about 120 μm.

After the metal mask 52 thus formed is placed on the printed wiring board 10 and the opening 62 is positioned and fixed to the pad 12 (FIG. 1B), a fixed diameter (105 μm) is formed thereon. A large number of solder balls 64 are placed (FIG. 1C). Then, the whole is vibrated horizontally (FIG. 1 (D)). At this time, some vertical vibration may be applied.

The vibration causes the solder balls 64 to drop into the openings 62 one by one. The dimensions of the opening 62 are determined so that only one solder ball 64 enters and two solder balls 64 do not enter. Since the opening 62 has a large upper opening diameter, the solder ball 64 smoothly enters the opening 62 and hits the pad 12.

Next, a squeegee 66 is placed on the metal mask 52.
Then, the extra solder balls 64 that did not enter the openings 62 are removed from above the metal mask 52 (FIG. 1E). Also, the solder ball 64 in the opening 62
Is temporarily attached to the flux 50 applied to the pad.

Next, without moving the solder balls 64, the metal mask 52 is gently peeled off (FIG. 2 (F)) and heated in a reflow furnace (FIG. 2 (G)). In FIG. 2G, reference numeral 68 denotes an infrared heater of the reflow furnace. By this reflow heating, the solder balls 6
4 melts, and the molten solder rises hemispherically on the pad 12 due to its surface tension. Therefore, if cooled as it is, a hemispherical solder bump 70 is completed. This solder bump 70 has a height of about 70 μm.

Since the metal mask 52 used in this embodiment has a truncated conical opening 62, the solder balls 6
4 is guided to the conical surface of the inner wall of the opening 62 and the pad 1 is smoothly moved.
2, and the solder ball 64 can be stabilized at an accurate position. Since a metal mask which is widely used forms an opening by etching a metal plate from both sides, the opening has a pincushion shape in which the shape of the cavity is wide on both sides and narrow in the middle. As a result, the solder ball entering the opening largely moves on the pad, and the positioning of the solder ball tends to be unstable. However, according to the embodiment shown in FIG. 4, there is no such disadvantage.

[0035]

As described above, according to the first aspect of the present invention, a solder ball is dropped into an opening of a metal mask, an extra solder ball is removed by a squeegee, the metal mask is peeled off, and reflow heating is performed to form a solder bump. Since it is formed, a solder ball having a small diameter can be accurately and efficiently supplied onto each pad. Therefore, even when the pad pitch and the pad diameter are reduced, a solder bump having a uniform height can be formed.

If the opening of the metal mask is formed in the shape of a truncated cone, solder balls can be supplied from the opening on the larger diameter side of the opening to accurately position the pad on the pad. . Such a metal mask is formed by coating a conductive substrate with a photo-curable (for example, ultraviolet-curable) resist, exposing a portion corresponding to a pad and curing the resist to form a resist column, and thickening metal plating. Then, the resist can be removed and the plating layer can be peeled off from the substrate (claim 3).

In this case, the substrate can be a stainless steel plate, and the metal plating can be electrolytic nickel plating. When a solder bump having a height of about 10 μm is formed on a pad having a pitch of about 0.25 mm and a diameter of about 0.13 mm, the diameter of a small-diameter solder ball is set to about 0.105 mm.
Preferably, the opening of the metal mask is frustoconical with a maximum diameter of about 0.16 mm and a minimum diameter of about 0.13 mm, and the thickness of the metal mask is about 120 μm.

[Brief description of the drawings]

FIG. 1 is a diagram showing a first half process of one embodiment of the present invention.

FIG. 2 is a diagram showing the latter half of the process.

FIG. 3 is an enlarged view of a part viewed from an arrow III in FIG. 1;

FIG. 4 is a view showing a manufacturing process of a metal mask.

FIG. 5 is an explanatory view of a conventional cream solder printing method.

FIG. 6 is a diagram for explaining the inconvenience.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Printed wiring board 12 Pad 14 Solder resist 50 Flux 52 Metal mask 54 Substrate 56 Photocurable resist 58 Resist pillar 60 Nickel plating layer 62 Opening 64 Solder ball 66 Squeegee

Claims (5)

[Claims] 1. A solder bump forming method for forming a solder bump on each pad by supplying and reflowing one solder ball on each pad of a printed wiring board, a) applying a flux to each pad B) overlaying a metal mask having an opening for one solder ball at a position corresponding to each pad so that the opening is aligned with each pad; c) applying a large number of solder balls of a predetermined diameter to the metal mask; D) One solder ball is dropped into each opening and adhered to the flux. E) Excess solder balls that do not enter each opening are removed by rubbing the metal mask surface with a squeegee. f) Remove the metal mask; g) Heat the printed wiring board in a reflow furnace to melt the solder balls to form solder bumps on each pad Solder bump forming method characterized by having each of the steps described above. 2. An opening of the metal mask is formed in a substantially frusto-conical shape, and a surface of the metal mask on the smaller diameter side of the opening is brought into close contact with a printed wiring board. The method of claim 1 wherein solder balls are supplied. 3. The metal mask used in the step b) includes: b-1) applying a photocurable resist to the pad positions on the surface of the substrate; and b-2) exposing the pad positions to make the pads slightly smaller than the solder ball diameter. B-3) forming a metal plating layer on the surface of the substrate, b-4) eluting and removing the resist columns, and b-5) removing the substrate from the substrate. The method of claim 2, wherein the solder bump is formed by removing. 4. The substrate is a stainless steel plate, and the metal plating layer is formed by thick electrolytic nickel plating.
Solder bump forming method. 5. The solder bump formation according to claim 4, wherein the solder ball diameter is about 105 μm, the metal mask is about 120 μm in thickness, and the opening has a frusto-conical shape with opening diameters of about 130 μm and about 160 μm. Method.