JP3996557B2 - Manufacturing method of semiconductor junction wafer - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method of manufacturing a semiconductor bonded wafer by bonding two semiconductor wafer substrates to each other, and more particularly, reducing unbonded portions on the outer periphery of the wafer to make a so-called terrace portion very small or not at all. The present invention relates to a method for manufacturing a semiconductor junction wafer.
[0002]
[Prior art]
A conventional method for manufacturing a semiconductor junction wafer will be described with reference to FIG. First, after the bond wafer (first silicon substrate) 10 and the base wafer (second silicon substrate) 20 are brought into close contact with each other (FIG. 3A), heat treatment is usually performed at a temperature of 1000 to 1200 ° C. (hereinafter referred to as fixing heat treatment). In general, the bonding strength is increased by applying (described below) (FIG. 3B). In this case, as shown in FIG. 7, the thickness (T1) of the two silicon substrates 10 and 20 is generally the same thickness of about 500 μm although it depends on the diameter. The chamfer width (X1) is about 350 to 500 μm.
[0003]
As shown in FIG. 3, in the process after the two wafers are bonded in this way, surface grinding is performed in order to obtain a predetermined thickness for the bond wafer 10 on the side where the device is manufactured (FIG. 3). (D)) Further, the surface ground surface 10a is mirror-polished to manufacture a semiconductor bonded wafer (FIG. 3E). However, as shown in FIG. 8, a gap 4 is formed over the width (W1) of about 1 to 2 mm on the outer periphery of the two wafers joined in this manner, and an unjoined portion 10c is generated.
[0004]
As shown in FIG. 7, the unbonded portion 10c has a shape in which the outer peripheral portions of the two silicon substrates 10 and 20 to be bonded are distorted by mirror polishing, so that the portion of the unbonded portion 10c and the gap 4 is bonded at the time of bonding. It happens. The width (W1) of the unbonded portion 10c includes a general chamfering width (X1) processed to prevent chipping of the silicon substrate.
[0005]
When there is an unbonded portion on the outer periphery in this way, there arises a problem that this portion is peeled off in the process of grinding and polishing the bond wafer after bonding the two wafers or in the device manufacturing process. It is also conceivable that the device characteristics are deteriorated by the unjoined portion. Therefore, it is necessary to remove this unjoined portion in advance.
[0006]
Therefore, as shown in FIG. 3C and FIG. 9, the bonding wafer 10 and the base wafer 20 serving as a support substrate are bonded to the unbonded portion 10c of about 1 to 2 mm width (W1) on the outer periphery of the bond wafer. A terrace area in which the surface of the bond wafer is ground and polished to a predetermined thickness after grinding or etching to remove the unbonded part until the interface 31 is reached, and a step shape is provided on the outer periphery of the wafer. 30 is formed.
[0007]
However, such a method for increasing the terrace area complicates the wafer manufacturing process, lowers productivity, and increases cost. If a large terrace region is provided, the diameter of the bond wafer, which is the substrate on the device fabrication side, is reduced, and the device fabrication area is reduced. Furthermore, if the terrace area is large, the step shape may interfere with handling such as wafer conveyance of the manufacturing apparatus in the subsequent device process or the like.
[0008]
In order to reduce such an unbonded portion generated during the manufacture of a semiconductor bonded wafer, it has been proposed to make the bond wafer to be bonded thinner than the base wafer (see, for example, Patent Document 1). This is because by making the bond wafer thinner than usual, the bond wafer can easily stick to the base wafer, and the unjoined portion can be made smaller.
[0009]
However, this method increases the thickness of the bond wafer to 300 μm or more so as not to cause cracking or chipping of the semiconductor wafer. As a result, an unjoined portion of 1 mm or more remains on the outer periphery of the wafer. In such a method, the terrace region for removing the unjoined portion still needs to be enlarged, and the device-manufacturable area becomes small, which may hinder wafer conveyance in the device-manufacturing process.
[0010]
[Patent Document 1]
Japanese Patent Laid-Open No. 11-26337
[Problems to be solved by the invention]
The present invention has been made to solve the above-described problems, and the present invention can further reduce the unjoined portion of the wafer outer periphery generated by such a conventional manufacturing method, thereby making the terrace area extremely small, or the terrace area. It is an object of the present invention to provide a method for manufacturing a semiconductor junction wafer that does not need to be provided at all.
[0012]
[Means for Solving the Problems]
In order to achieve the above object, the present invention provides a semiconductor bonded wafer in which at least two semiconductor wafers are directly brought into close contact with each other, subjected to heat treatment and bonded, and then an unbonded portion on the outer peripheral portion of the semiconductor wafer is removed. A method for manufacturing a semiconductor junction wafer, wherein a thickness of at least one of the two semiconductor wafers is set to 100 μm or less in advance, and is brought into close contact with the other semiconductor wafer. .
[0013]
As described above, in the method of manufacturing a semiconductor bonded wafer, at least one of the two semiconductor wafers to be bonded has a thickness of 100 μm or less, which is extremely thinner than a conventional wafer, and is in close contact with the other semiconductor wafer. By doing so, the semiconductor wafer becomes very supple and easy to stick, so that the unjoined portion of the outer peripheral portion is significantly reduced, and the terrace region for removing the unjoined portion can be made extremely small.
[0014]
In this case, it is preferable that the chamfered width of the outer peripheral portion of the semiconductor wafer be 100 μm or less to be brought into close contact with each other .
Thus, by setting the chamfering width of the outer peripheral portion of the semiconductor wafer to be bonded to 100 μm or less which is extremely smaller than the conventional chamfering width, the gap when two wafers are bonded can be reduced. Combined with the wafer thickness being 100 μm or less, the unjoined portion can be made smaller.
[0015]
In this case, a wafer having a diameter larger than a standard value is used as the two semiconductor wafers, and after joining the two semiconductor wafers, unbonded portions at the outer peripheral portions of both of the joined semiconductor wafers are used as both wafers. It is preferable to set the diameter of the semiconductor junction wafer manufactured by removing both to the standard value .
[0016]
As described above, a wafer having a diameter larger than the standard value is used as the two semiconductor wafers to be bonded, and after bonding the two semiconductor wafers, only the bond wafer side is removed to provide a terrace region. Instead, by removing both the bonded wafer and the base wafer from the unbonded portion at the outer periphery of both bonded semiconductor wafers, the diameter of the manufactured semiconductor bonded wafer is taken as the standard value, so that the terrace region is completely eliminated. A non-bonded portion can be removed without providing, and a device-manufacturable area is large, and since there is no step shape, a wafer that does not hinder wafer transportation in the device-manufacturing process can be manufactured. In addition, in the present invention, removal of unbonded portions of both the bond wafer and the base wafer can be performed by normal chamfering because one wafer is thin and a chamfer width is small. In addition to the simple processing steps, the outer peripheral shape of the resulting bonded wafer is substantially the same as the chamfered shape of a normal wafer, which is extremely preferable in terms of handling.
[0017]
In this case, it is preferable that the processing width for removing the unbonded portion is at least three times the chamfer width at the outer peripheral portion of the semiconductor wafer before bonding .
Thus, if the processing width for removing the unbonded portion is set to be three times or more the chamfer width at the outer peripheral portion of the semiconductor wafer before bonding, the unbonded portion can be surely eliminated.
[0018]
In this case, a silicon wafer can be used as the semiconductor wafer .
Thus, by using a silicon wafer as a semiconductor wafer to be bonded, semiconductor bonded wafers that can be used for manufacturing various semiconductor devices can be manufactured.
[0019]
Then, the semiconductor junction wafer produced by the method of manufacturing a semiconductor junction wafer according to the present invention, since unwelded portion when the bonding two semiconductor wafers is significantly smaller, was provided for removing this The terrace area can also be very small or have no terrace area. As a result, the device fabrication area of the semiconductor bonded wafer can be made larger than before, and it is possible to prevent an obstacle such as wafer conveyance during device manufacturing.
[0020]
Hereinafter, the present invention will be described in more detail, but the present invention is not limited thereto.
Conventionally, the thickness of a wafer in a semiconductor junction wafer is set according to the standard of the device to be manufactured. Generally, the thickness of a bond wafer on the side on which a device is manufactured is almost the same as that of a base wafer serving as a support substrate. The same thickness of 400-600 μm was used. In addition, the diameters of the bond wafer and the base wafer to be used have been the diameter standards of the bonded wafer that will be the final product.
[0021]
In the method described in Japanese Patent Laid-Open No. 11-26337 described above, a thinner wafer than usual is used for the bond wafer, but a thick wafer having a thickness of 300 μm or more is used in consideration of cracks and chipping of the wafer. Moreover, about the diameter, the thing as a standard value was used.
[0022]
However, since the thickness of the bond wafer for finally producing the device is processed to a predetermined thickness (generally 100 μm or less) after bonding, a thin wafer having a thickness of 100 μm or less is bonded in advance. Also good. As a result of investigations and examinations by the present inventors, when the thickness of at least one wafer is set to 100 μm or less, the wafer becomes remarkably more flexible than the wafer having a thickness exceeding 100 μm, and the wafer is superposed on the other substrate. It was found that the unbonded portion can be greatly reduced. On the other hand, due to improvements in wafer handling technology in recent device fabrication processes, there is less risk of wafer cracking and chipping even when a wafer having a thickness of 100 μm or less is used in the manufacture of a bonded wafer in advance. Therefore, the present inventor decided to use a thin wafer having a thickness of 100 μm or less for the production of a semiconductor junction wafer. In addition, by using a wafer larger than the standard value in advance and using a thin wafer, it is possible to produce a bonded wafer without a terrace by removing both the bonded wafer and the base wafer together with the unbonded portion that has become smaller. I found.
The present invention has been completed as a result of studying various conditions based on such a basic idea.
[0023]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be specifically described with reference to the drawings. However, the present invention is not limited to these embodiments.
First, as shown in FIGS. 1A and 4, two semiconductor substrates to be bonded to each other, a bond wafer 1 on which a device is to be manufactured, and a base wafer 2 to be a support substrate are prepared. As shown in FIG. 1A, the diameters of the bond wafer 1 and the base wafer 2 are φ, which is a standard value of the diameter of the bonded wafer to be manufactured. As shown in FIG. 4, in this example, the thickness (T1) of the base wafer 2 is 400 to 600 μm, whereas the thickness (T2) of the bond wafer 1 is as thin as 100 μm or less. Further, the chamfering width (X2) of both the wafers 1 and 2 is set to 100 μm or less, which is smaller than the conventional chamfering width of 350 to 500 μm in order to avoid physical gaps after bonding. With such a chamfering width, the unjoined portion can be made smaller, and sufficient mechanical processing accuracy can be obtained.
[0024]
Next, the bond wafer 1 and the base wafer 2 are brought into close contact with each other in a clean atmosphere. Thereafter, the bonded bond wafer 1 and the base wafer 2 are bonded to each other by applying a fixing heat treatment in an oxidizing atmosphere, for example (FIG. 1B). As shown in FIG. 5, since the bond wafer 1 is as thin as 100 μm or less and easily sticks to the base wafer, the width (W2) of the non-bonded portion 1c of the bond wafer 1 at this time is the conventional non-bonded as shown in FIG. It is significantly smaller than the width (W1) of the joint 10c.
[0025]
The unbonded portion 1c of the bond wafer 1 is removed to perform a process for forming the terrace region 3 (FIG. 1 (c)). In this case, unlike the conventional bonded wafer, the unbonded portion 1c is extremely small, so that the shaving margin can be reduced and the step of forming the terrace region can be easily performed. As a result of investigation and examination by the present inventor, it was found that the width (W2) of the unjoined portion is about three times the chamfering width (X2) of the wafers 1 and 2 before joining. That is, for example, when the chamfer width (X2) is 100 μm, the unjoined portion can be removed almost completely by removing about 300 μm, which is about three times as much. Therefore, the removal amount is much smaller than that of the conventional bonding wafer where the terrace region is formed by removing 1 mm or more.
[0026]
Then, in order to make the bond wafer 1 have a predetermined thickness, surface grinding is performed (FIG. 1 (d)), and the surface ground surface 1a is mirror-polished to a mirror-polished surface 1b (FIG. 1 (e)). )), And complete the semiconductor junction wafer. Since the wafer of the present invention has a small terrace area, the device-manufacturable area can be increased. Furthermore, since the steps in the terrace area are small, it is less likely to be an obstacle to wafer conveyance. Further, in this case, since a thin bond wafer having a thickness of 100 μm or less is used in advance, there is an advantage that a margin for grinding can be reduced and both cost and time can be saved.
[0027]
In the example shown in FIG. 1, a thin wafer having a thickness of 100 μm or less is used as the bond wafer. However, both the bond wafer and the base wafer may be thinned, or only the base wafer may be thinned.
[0028]
As a specific method for manufacturing such a wafer having a thickness of 100 μm or less, for example, the etching amount of the wafer is increased more than usual, and the etching amount of about 30 to 40 μm on both sides of the normal wafer is, for example, about 100 μm. To do.
[0029]
The wafer polishing method is not to attach the wafer to be polished directly to the polishing glass plate as in the prior art, but first, the wafer to be polished is bonded to a mirror-polished wafer as a dummy wafer, and the dummy wafer is polished. The wafer to be polished through the dummy wafer is fixed to the plate and then polished under normal mirror polishing conditions. After polishing, the dummy wafer is separated to obtain a wafer having a thickness of 100 μm or less. Can be manufactured.
Alternatively, a method may be used in which a film is pasted on the back surface of a wafer to be polished, the wafer to be polished is fixed to a plate through the film, the polishing is performed, and the film is removed after polishing.
[0030]
On the other hand, in the present invention, since the unjoined portion is small, it is also possible to use a method of removing all unjoined portions without providing a terrace region by sticking a wafer having a diameter slightly larger than the standard only to the unjoined portion in advance. it can. Hereinafter, another embodiment of the present invention will be described.
FIG. 2 is a flowchart showing another example of a method for manufacturing a semiconductor junction wafer according to the present invention. In this embodiment, first, in addition to the thickness of the bond wafer 1 ′ being 100 μm or less, a bond wafer 1 ′ and a base wafer 2 ′ having a diameter larger than the standard value φ of the bonded wafer diameter manufactured in advance are prepared. (FIG. 2A).
[0031]
For example, if the width of the unjoined portion when the wafers 1 and 2 are bonded and joined is W2, the diameter of the wafer is φ + (W2 × 2), which is a diameter that is twice as large as the unjoined portion width W2. And For example, if the standard value φ of the wafer diameter is 150 mm and the chamfer width X2 of the bond wafer 1 ′ and the base wafer 2 ′ before bonding is 100 μm, the unbonded portion width W2 is equal to the chamfer width X2 as described above. Since it is about 3 times, W2 is about 300 μm (0.3 mm), and a bond wafer 1 having a diameter of 150.6 mm, which is about 0.6 mm, which is twice as large as 0.3 mm from the standard value 150 mm of the manufactured bonded wafer diameter. You can prepare 'and base wafer 2'. Manufacturing a wafer having a diameter approximately 0.6 mm larger than the diameter size of such a general silicon wafer can be performed without any particular trouble in the wafer manufacturing process.
[0032]
Next, the bond wafer 1 ′ and the base wafer 2 ′ are brought into close contact as in the above-described embodiment of FIG. 1 and bonded by a fixing heat treatment (FIG. 2B).
Next, the terrace region is not formed by removing only the unbonded portion of the bond wafer 1 ′ as in the embodiment of FIG. 1, but the unbonded portions on the outer peripheral portions of both the bond wafer 1 ′ and the base wafer 2 ′. Both parts 32 are removed, and the wafer diameter is set to the standard value (FIG. 2 (c), FIG. 6). The unbonded portion 32 can be easily removed by removing only the outer periphery of the bonded wafer W2 as described above. In this case, for example, since the above-mentioned unjoined portion width W2 = 0.3 μm, it is within a range in which mechanical machining accuracy can be obtained by normal chamfering, and therefore, the chamfering in which unjoined portion is generally removed It can be done easily by processing. Therefore, not only is it very preferable in the processing step, but the outer peripheral shape of the bonded wafer after processing can be made substantially the same as the chamfered shape of a normal wafer, so that no problem occurs in wafer handling and the like.
[0033]
Then, as in the embodiment of FIG. 1, in order to make the bond wafer 1 ′ to a predetermined thickness, surface grinding is performed (FIG. 2 (d)), and the surface ground surface 1a ′ is further mirror-polished to give a mirror surface. A polished surface 1b ′ is formed (FIG. 2E) to complete the semiconductor bonding wafer. Also in this case, since a thin bond wafer is used in advance, the grinding margin can be reduced. Depending on the standard on the bond wafer side, this grinding step can be omitted.
In this embodiment, the bonded wafer does not have a terrace portion, and a wafer having a standard diameter can be obtained, so that the device manufacturing area can be maximized. If this method is adopted, a semiconductor bonded wafer having almost the same shape as that of a normal silicon wafer that is not a bonded wafer can be manufactured, and can be put into a device manufacturing process in the same manner as a normal silicon wafer.
[0034]
【Example】
EXAMPLES Hereinafter, although an Example and a comparative example are shown and this invention is demonstrated more concretely, this invention is not limited to these.
(Example 1, Example 2)
A semiconductor junction wafer was manufactured by the method shown in FIG. First, as Example 1, as shown in FIG. 1A, two silicon substrates to be bonded (diameter 6 inches (150 mm), crystal orientation <100>) were prepared. The thickness of the bond wafer was 100 μm, and the thickness of the base wafer was 550 μm. Further, the chamfer widths of the outer peripheral portions of the bond wafer and the base wafer were 100 μm, respectively.
[0035]
Next, as shown in FIG. 1B, the bond wafer and the base wafer were brought into close contact with each other in a clean atmosphere. Thereafter, the bonded wafer and the base wafer were subjected to fixing heat treatment at 1100 ° C. for 2 hours in an oxidizing atmosphere.
[0036]
Here, in the normal manufacturing method, as shown in FIG. 1C, processing for forming a terrace region is performed to remove the unjoined portion on the outer periphery of the bond wafer. In Example 1, the size of the unjoined portion is used. This process was omitted to investigate. Then, as shown in FIG. 1 (f), the bond wafer was surface ground. In the surface grinding, a surface grinding machine 8 as shown in FIG. 10 is used, the bonding wafer 6 is fixed to the rotary table 7 and rotated, and the bonded grinding wafer 5 is rotated to a thickness t = Surface grinding was performed thinly to 10 μm or less. And the width W2 of the unjoined part of the wafer outer peripheral part as shown in FIG. 11 was measured. At this time, as shown in FIG. 1 (g) and FIG. 12, the outer peripheral area of the wafer was divided into four, and the length that was the widest and unbonded in each area was measured.
[0037]
As a result, as shown in FIG. 12, the width W2 of the unbonded portion of the wafer of Example 1 is 0.3 mm (300 μm), which is three times the chamfered width of the silicon wafer before bonding, even in the maximum region. It turns out that it is a small thing of the grade or less.
[0038]
Further, as Example 2, a semiconductor bonded wafer was manufactured in the same manner as in Example 1 except that the thickness of the bond wafer was set to 70 μm, and the width W2 of the unbonded portion was measured. As a result, as shown in FIG. 12, it can be seen that the width of the unjoined portion is smaller than that of the first embodiment.
[0039]
(Comparative Example 1 and Comparative Example 2)
In the same manner as in Example 1, a semiconductor bonded wafer was manufactured, and the width of the unbonded portion was measured. However, in Comparative Example 1, the thickness of the bond wafer was 550 μm, and in Comparative Example 2, the thickness of the bond wafer was 150 μm. After bonding and bonding the bond wafer and the base wafer, the wafer was subjected to surface grinding in the same manner as in Example 1 without forming the terrace region, the width of the unbonded portion was measured, and the measurement results are shown in FIG. did.
[0040]
From FIG. 12, Comparative Example 1 in which the thickness of the bond wafer and the base wafer is the same 550 μm is about 0.95 mm (950 μm) in the region where the width of the unbonded portion is the maximum, which is more than three times that of Example 1. You can see that In Comparative Example 2, although the thin wafer having a thickness of 150 μm was used as the bond wafer, the width of the unjoined portion was about 0.6 mm (600 μm) in the maximum region, which was not joined as compared with Example 1. It can be seen that the width of the part is almost twice as large.
[0041]
(Example 3)
A semiconductor bonding wafer having a diameter of 150 mm was manufactured by the method shown in FIG. First, as shown in FIG. 2A, two silicon substrates (crystal orientation <100>) to be bonded were prepared. The thickness of the bond wafer was 100 μm, and the thickness of the base wafer was 550 μm. The chamfer width of the outer peripheral portion of the bond wafer and the base wafer was 100 μm.
[0042]
Here, since the chamfered width of the bond wafer and the base wafer is 100 μm, it can be estimated that the width of the unjoined portion when the two are bonded is about 300 μm (0.3 mm) at the maximum. Therefore, in Example 3, a bond wafer and a base wafer having a diameter of 150 mm + (0.3 mm × 2) = 150.6 mm were prepared in order to manufacture a bonded wafer having a standard value of 150 mm.
[0043]
Then, both wafers are brought into intimate contact with each other in the same manner as in Example 1 and bonded and heat treated (FIG. 2 (b)), and then both unbonded portions at the outer peripheral portions of both bonded wafers are chamfered. The wafer was also removed (FIG. 2 (c)). As for the removal amount, 0.3 mm of the width of the unjoined portion was removed from the outer periphery of the wafer, and the wafer diameter was set to 150 mm which is a standard value.
Thereafter, a surface grinding process and a mirror polishing process were obtained (FIGS. 2D and 2E) to complete a semiconductor bonding wafer.
[0044]
When the manufactured semiconductor bonded wafer was observed and measured, no unbonded portion was found, and it was confirmed that it was bonded to the outermost periphery even though no terrace region was provided. Further, since the diameter of the wafer is a standard value of 150 mm, the wafer can be transported in the device manufacturing process as in the case of a normal silicon wafer that is not a bonded wafer.
[0045]
The present invention is not limited to the above embodiment. The above embodiment is merely an example, and the present invention has the same configuration as that of the technical idea described in the claims of the present invention, and any device that exhibits the same function and effect is the present invention. It is included in the technical scope of the invention.
[0046]
【The invention's effect】
As described above, according to the present invention, a wafer having a small terrace area or a wafer having no terrace area is manufactured because a non-bonded portion is small when two wafers are bonded in manufacturing a semiconductor bonded wafer. be able to. Accordingly, the area where devices can be manufactured can be increased, and wafer transport can be facilitated, so that the manufacturing yield, productivity, and cost of semiconductor bonded wafers and devices can be improved.
[Brief description of the drawings]
FIG. 1 is a flowchart showing an experimental flow for examining a width of an outer peripheral unbonded example in an example of a method for manufacturing a semiconductor bonded wafer according to the present invention and an example.
FIG. 2 is a flowchart showing another example of a method for manufacturing a semiconductor junction wafer according to the present invention.
FIG. 3 is a flowchart showing a conventional method for manufacturing a semiconductor junction wafer.
FIG. 4 is an enlarged view showing a chamfered portion of a wafer outer periphery of a bond wafer and a base wafer in the manufacturing method of the present invention.
FIG. 5 is an enlarged view showing an outer peripheral bonding portion of a semiconductor bonding wafer in the manufacturing method of the present invention.
FIG. 6 is an enlarged view showing a state where an unbonded portion of a semiconductor bonded wafer is removed in the manufacturing method of the present invention.
FIG. 7 is an enlarged view showing a chamfered portion of a wafer outer peripheral portion of a bond wafer and a base wafer in a conventional manufacturing method.
FIG. 8 is an enlarged view showing a peripheral junction portion of a semiconductor junction wafer in a conventional manufacturing method.
FIG. 9 is an enlarged view showing a terrace region of a semiconductor junction wafer in a conventional manufacturing method.
FIG. 10 is a configuration diagram of a surface grinding machine used in an example of the present invention.
FIG. 11 is an enlarged view showing a non-bonded portion of a semiconductor bonded wafer in an embodiment of the present invention.
FIG. 12 is a diagram showing experimental results in Examples and Comparative Examples.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1,1 '... Bond wafer, 1a, 1a' ... Surface grinding surface, 1b, 1b '... Mirror-polished surface, 1c ... Unjoined part, 2, 2' ... Base wafer, 3 ... Terrace area, 4 ... Gap, 5 DESCRIPTION OF SYMBOLS: Grinding wheel, 6 ... Joining wafer, 7 ... Rotary table, 8 ... Surface grinding machine, 10 ... Bond wafer, 10a ... Surface grinding surface, 10b ... Mirror surface polishing surface, 10c ... Unjoined part, 20 ... Base wafer, 30 ... Terrace region, 31 ... bonding interface, 32 ... unbonded part.

Claims (5)

A method for manufacturing a semiconductor bonded wafer, wherein at least two semiconductor wafers are directly brought into close contact with each other, subjected to heat treatment and bonded, and then an unbonded portion on an outer peripheral portion of the semiconductor wafer is removed . A wafer having a diameter larger than a standard value is used as a semiconductor wafer, and at least one of the two semiconductor wafers is made to have a thickness of 100 μm or less in advance to be in close contact with the other semiconductor wafer, and then bonded. the method of manufacturing a semiconductor junction wafer, wherein the standard value of the diameter and to Rukoto semiconductor junction wafer to be manufactured by removing the unbonded portion both of the wafer at the outer peripheral portion of both of the semiconductor wafer.   2. The method of manufacturing a semiconductor bonded wafer according to claim 1, wherein the semiconductor wafer is bonded by being closely contacted with a chamfer width of an outer peripheral portion of the semiconductor wafer being 100 [mu] m or less. 3. The method for manufacturing a semiconductor bonded wafer according to claim 1 , wherein a processing width for removing the unbonded portion is at least three times a chamfered width in an outer peripheral portion of the semiconductor wafer before the bonding. A method for manufacturing a semiconductor junction wafer. 4. The method of manufacturing a semiconductor junction wafer according to claim 1 , wherein a silicon wafer is used as the semiconductor wafer. 5. A semiconductor junction wafer manufactured by the method for manufacturing a semiconductor junction wafer according to claim 1 .

Images