JPH1187203A - Method for bonding substrates - Google Patents

Abstract

PROBLEM TO BE SOLVED: To form a bonded substrate with a narrow non-contact area at the peripheral portion thereof, which can be readily treated. SOLUTION: After having a pair of Si wafers 11, 12 make contact with each other for bonding through a suction force, at least one of the Si wafers 11, 12 is pressured at the periphery thereof, to bring the peripheries of both of the wafers into contact with each other for further bonding. In the case where there exist large spaces between the peripheries of the Si wafers 11, 12 such that they do not cause a suction force to be generated for bonding, owing to the forced contact of the peripheries by pressure for bonding, the width of the bonding portion at the peripheral portion can be rendered narrow, and thus peeling or the like is hardly caused in the treatment thereafter.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of bonding substrates for bonding a pair of substrates together to form a bonded substrate.

[0002]

2. Description of the Related Art For example, as one method for manufacturing a semiconductor device having an SOI structure, an insulating film and a planarizing film are sequentially formed on the surface of a first semiconductor wafer, and the planarizing film is formed on the insulating film. The first semiconductor wafer is bonded to the second semiconductor wafer by grinding, and the first semiconductor wafer is ground from the back side, and the first semiconductor wafer is formed on the surface of the insulating film opposite to the second semiconductor wafer. There is a method of forming a bonded SOI wafer having a semiconductor thin film formed from a wafer.

FIG. 2 shows a general method for bonding a pair of Si wafers together. in this way,
As shown in FIG. 2A, a SiO ₂ film (not shown) as an insulating film and a polycrystalline Si film (not shown) as a planarization film are sequentially formed on the surface, and a bond wafer is formed. An Si wafer 11 referred to as a base wafer and a Si wafer 12 referred to as a base wafer to be bonded to the Si wafer 11 are prepared.

[0004] Since this bonding is considered to be achieved by hydrogen bonding due to the action of water or a hydroxyl group interposed on the bonding surfaces of the Si wafers 11 and 12, the bonding is performed in order to obtain a sufficient hydrogen bonding force. Prior to this, the bonding surfaces of the Si wafers 11 and 12 are subjected to a hydrophilic treatment and dust is removed.

As the hydrophilic treatment, treatment with hydrofluoric acid and treatment with a mixed solution of aqueous ammonia and hydrogen peroxide are continuously performed. By this hydrophilic treatment, hydroxyl groups on the bonding surfaces of the Si wafers 11 and 12 increase. In addition, only the treatment with hydrofluoric acid may be performed as the hydrophilic treatment.

After the hydrophilic treatment, the Si wafer 11 is sucked by the flat porous member 14 of the vacuum chuck 13 and the Si wafer 11 is sucked.
The Si wafer 12 is positioned and placed on the wafer 11. However, in this state, an interval of several μm is generated between the Si wafers 11 and 12 due to viscous resistance of air and the like, and the Si wafers 11 and 12 do not adhere to each other.

[0007] Therefore, as shown in FIG.
In step 5, pressure is applied to the center of the Si wafer 12 to bring the centers of the Si wafers 11 and 12 into contact with each other. Then, the contact portions adhere to each other by the above-described hydrogen bonding force,
As shown in (c), self-adhesion occurs in which the adhesion due to the hydrogen bonding force proceeds to the peripheral portions of the Si wafers 11 and 12. As a result, as shown in FIG. 3A, a bonded Si wafer 16 in which the Si wafers 11 and 12 are bonded to each other is formed.

By the way, the Si wafers 11 and 12 formed by cutting from the ingot are subjected to polishing and chamfering. For this polishing and the like, as shown in FIG.
0.15 to 0.5 m around the Si wafers 11 and 12
In addition to the chamfered portion having a width A of about m, a slightly thinner portion is formed with a width B of about 1.0 to 2.0 mm.

In the thin portion, the Si wafer 11,
Since the interval between the 12 is too wide and self-adhesion does not occur, a thin portion becomes a non-adhesion portion. If the non-bonded portion is left on the bonded Si wafer 16, the non-bonded portion may be peeled off in a subsequent process for manufacturing a semiconductor device having an SOI structure, and the yield of the semiconductor device is reduced. .

Therefore, the bonded Si wafer 16 is turned upside down from the state shown in FIG. 3A, and as shown by the dashed line in FIG. A width C of about 1.0 mm with respect to the peripheral portion of the Si wafer 11
Has been chamfered. The reason why the chamfering is mainly performed on the Si wafer 11 is to keep the diameter of the bonded Si wafer 16 constant.

Since the width B of the non-adhered portion is more likely to be wider than the width C of the chamfer in the peripheral portion of the bonded Si wafer 16, the Si wafer 1 shown in FIG.
In order to prevent the Si thin film formed by grinding No. 1 from peeling in a subsequent step, the non-bonded portion of the Si thin film is forcibly peeled in advance.

[0012]

However, in the conventional example of the present invention shown in FIGS. 2 and 3, since the width B of the non-adhesive portion in the peripheral portion of the bonded Si wafer 16 is large, the non-adhesive portion is The width C of the chamfer for removal was also wide. Therefore, there is a high possibility that a defect such as chipping occurs in the bonded Si wafer 16 in the chamfering step, and the yield of the semiconductor device is low.

On the other hand, a wafer number for identifying the Si wafers 11 and 12 is laser-engraved on a portion of the Si wafers 11 and 12 where no semiconductor device is formed, that is, a peripheral portion. However, there is a high possibility that the wafer number becomes unreadable due to the above-mentioned chipping, and the production of the semiconductor device is likely to be hindered.

As described above, the non-bonded portion of the Si thin film formed by grinding the Si wafer 11 is forcibly peeled in advance, but the width B of the non-bonded portion is large. Therefore, the time required for removing the Si thin film is long, and it has been difficult to manufacture a semiconductor device with high throughput.

That is, in the bonded Si wafer 16 formed in the conventional example of the present invention shown in FIGS. 2 and 3, since the width of the non-bonded portion in the peripheral portion is wide, the subsequent processing is not easy. Accordingly, an object of the present invention is to provide a method capable of forming a bonded substrate in which the width of a non-adhesive portion in a peripheral portion is small and the subsequent processing is easy.

[0016]

According to the first aspect of the present invention, even when a large gap exists between the peripheral portions of a pair of substrates and adhesion by an attraction force does not occur in the peripheral portions, the method for bonding substrates can be performed in a peripheral region. Since the parts are forcibly brought into contact with each other by pressure and bonded, the width of the non-bonded part in the peripheral part is narrow.

In the method for bonding substrates according to the second aspect, since the adhesion proceeds from the central portion to a portion other than the central portion, that is, the peripheral portion, for example, the adhesive proceeds from a part of the peripheral portion to the peripheral portion on the opposite side. The time required for the adhesion to proceed is shorter than in the case. Moreover, even if you start bonding from the center,
Since the peripheral portions of the pair of substrates are forcibly brought into contact with each other by pressing and bonded, the width of the non-bonded portion in the peripheral portions is narrow.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention applied to bonding a pair of Si wafers will be described below with reference to FIG. Also in the present embodiment, steps substantially the same as those in the conventional example shown in FIGS. 2 and 3 until the bonding by the hydrogen bonding force causes self-adhesion that progresses from the center of the Si wafers 11 and 12 to the periphery. Execute

However, in the present embodiment, immediately after the self-adhesion is completed as shown in FIG. 3A, as shown in FIG. Pressure is applied to the peripheral portion with a ring-shaped jig 17 and the Si wafer 1
The peripheral portions 1 and 12 are elastically deformed and brought into contact with each other, and the peripheral portions also cause adhesion by hydrogen bonding force.

When the peripheral portions of the Si wafers 11 and 12 are adhered to each other, even if the pressure applied to the peripheral portions of the Si wafers 11 and 12 by the jig 17 is released, the adhesive force of the Si wafers 11 and 12 is smaller than the elastic stress of the Si wafers 11 and 12. If it is stronger, Si wafers 11 and 12
The adhesion between the peripheral portions is maintained.

Therefore, the width of the non-bonded portion in the peripheral portion of the bonded Si wafer 18 formed in the above embodiment is
The width B of the non-bonded portion in the peripheral portion of the bonded Si wafer 16 formed in the conventional example shown in FIGS.
For this reason, the width of the chamfer for removing the non-adhered portion may be narrow, and the possibility of occurrence of defects such as chipping in the bonded Si wafer 18 in the chamfering step is low, and the yield of the semiconductor device is high.

In addition, if the width of the non-adhesive portion is narrow as described above,
Since there is a low possibility that the portion inside the chamfered portion having a width of 1 mm is peeled off, the possibility that the wafer numbers imprinted on the Si wafers 11 and 12 become unreadable is also low, and it is unlikely to hinder the manufacture of the semiconductor device. . Further, since the width of the non-bonded portion is narrow as described above, the time required for forcibly peeling off the non-bonded portion in advance in the Si thin film formed by the subsequent grinding of the Si wafer 11 is also short. In addition, a semiconductor device can be manufactured with high throughput.

In the above embodiment, the jig 17 is used only for the Si wafer 12 of the pair of Si wafers 11 and 12.
Is applied, but pressure may be applied to both the pair of Si wafers 11 and 12. Further, in the above-described embodiment, the pressure is applied to the central portion of the Si wafer 12 by the push rod 15 to cause self-adhesion to progress from the central portion to the peripheral portion. May be caused.

In the above embodiment, the present invention is applied to the bonding of a pair of Si wafers 11 and 12, but the bonding of the semiconductor wafers other than the Si wafers 11 and 12 and the bonding of the semiconductor wafers other than the semiconductor wafer are performed. The present invention can be applied to the bonding of the substrates.

[0025]

According to the method for bonding substrates according to the first aspect, since the width of the non-bonded portion in the peripheral portion is narrow, a bonded substrate that can be easily processed thereafter can be formed.

In the substrate bonding method according to the second aspect, the time required for the progress of bonding is short, and the width of the non-bonded portion in the peripheral portion is narrow, so that the bonded substrate that can be easily processed thereafter has a high throughput. Can be formed.

[Brief description of the drawings]

FIG. 1 is a side sectional view showing a final step of an embodiment of the present invention.

FIG. 2 is a side sectional view showing a general method for bonding substrates in the order of steps.

FIG. 3A is a side sectional view showing a final step of a conventional example of the invention of the present application, and FIG. 3B is a side sectional view of the vicinity of a peripheral portion of a bonded substrate formed in the conventional example.

[Explanation of symbols]

11, 12: Si wafer (substrate), 18: Bonded S
i wafer

Claims (2)

[Claims] 1. A method in which a part of a pair of substrates is brought into contact with each other, and the adhesion of the pair of substrates by an attraction force is advanced from the contact portion to a portion other than the contact portion to bond the pair of substrates. In the method of bonding substrates, after the bonding is completed, pressure is applied to at least one peripheral portion of the pair of substrates, and the peripheral portions of the pair of substrates are brought into contact with each other and bonded. Method. 2. The method according to claim 1, wherein the center of each of the pair of substrates is set to the part.