JP2004335771A - Diffusion wafer and method for manufacturing the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a diffusion wafer wherein raw material can be reduced and the high level of dislocation density on the surface of a non-diffusion layer and the thickness of the non-diffusion layer within a standard are attained. <P>SOLUTION: The diffusion wafer has double-layer structure having a diffusion layer 1 where high-density impurity is diffused on one surface of a silicon wafer and having the non-diffusion layer where impurity is not diffused on the other surface. The diffusion wafer has dislocation density distribution where in dislocation density is gradually reduced from the surface of the non-diffusion layer to the vicinity of a prescribed surface layer N<SB>S</SB>, and gradually increased from the vicinity of the prescribed surface layer to an interface B with the diffusion layer. The inclination of the dislocation density from the surface of the non-diffusion layer to the vicinity of the prescribed surface layer is designated to be larger than that from the vicinity of the prescribed surface layer to the interface with the diffusion layer. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a diffusion wafer used in the manufacture of discrete devices (individual elements) such as transistors, diodes, and MOSFETs, and capable of improving the high-frequency characteristics by setting the dislocation density of a device formation surface to a high level, and a method of manufacturing the same.
[0002]
[Prior art]
Conventionally, as a diffusion wafer of this type, a silicon wafer has a diffusion layer in which high-concentration impurities are diffused on one surface, and a non-diffusion layer in which the surface is mirror-finished and the impurities are not diffused on the other surface. The non-diffusion layer, which has a two-layer structure, has a high dislocation density of 5,000 to 20,000 / cm ^2, and has a dislocation density from the surface of the non-diffusion layer to the vicinity of a predetermined surface layer (a predetermined depth). It has a dislocation density distribution that gradually decreases and the dislocation density from the vicinity of the predetermined surface layer to the interface with the diffusion layer gradually increases, and the dislocation density gradient from the surface to the vicinity of the predetermined surface layer and the vicinity of the predetermined surface layer Japanese Patent Application Laid-Open Publication No. HEI 11-163556 discloses a technique in which the gradient of the dislocation density up to the interface with the diffusion layer is equal.
[0003]
[Patent Document 1]
Japanese Patent No. 2888294 [0004]
This diffusion wafer is subjected to lapping evenly on both sides, and the silicon wafer loaded in the boat is deposited using, for example, POCl ₃ (phosphorous oxychloride) at a temperature of 1200 ° C. under normal pressure. At a high temperature, for example, drive-in diffusion at 1280 ° C., and after forming a high concentration diffusion layer on both sides of the silicon wafer with a non-diffusion layer in which impurities are not diffused therebetween, the diffusion layer on one side and It is manufactured by grinding and removing less than half the thickness of the non-diffusion layer and finishing the surface of the exposed non-diffusion layer to a mirror surface.
[0005]
[Problems to be solved by the invention]
However, in the conventional diffusion wafer and its manufacturing method, the dislocation density gradient from the surface of the non-diffusion layer to the vicinity of a predetermined surface layer is equal to the dislocation density gradient from the vicinity of the predetermined surface layer to the interface with the diffusion layer. In order to raise the dislocation density to a higher level, the thickness of the silicon wafer before diffusion as a material must be increased, which leads to an increase in raw materials.
There is also a problem that it is difficult to keep the thickness of the non-diffusion layer within a predetermined standard in accordance with the increase in the level of dislocation density on the surface of the non-diffusion layer.
[0006]
Therefore, the present invention provides a diffusion wafer and a method for manufacturing the same, which can reduce the amount of raw materials, increase the dislocation density on the surface of the non-diffusion layer, and standardize the thickness of the non-diffusion layer. That is the task.
[0007]
[Means for Solving the Problems]
In order to solve the above problem, the diffusion wafer of the present invention has a diffusion layer in which high-concentration impurities are diffused on one surface of a silicon wafer, and a non-diffusion layer in which impurities are not diffused on the other surface. It has a two-layer structure, and has a dislocation density distribution in which the dislocation density gradually decreases from the surface of the non-diffusion layer to the vicinity of a predetermined surface layer, and the dislocation density gradually increases from the vicinity of the predetermined surface layer to the interface with the diffusion layer. In the diffusion wafer, the gradient of the dislocation density from the surface of the non-diffusion layer to the vicinity of a predetermined surface layer is larger than the gradient of the dislocation density from the vicinity of the predetermined surface layer to the interface with the diffusion layer.
[0008]
On the other hand, a method for manufacturing a diffusion wafer is to diffuse a high-concentration impurity on both sides of a silicon wafer subjected to a lapping process, grind and remove one diffusion layer, and mirror-finish the surface of an exposed non-diffusion layer. In the method, the damage caused by the lapping of the surface to be ground and removed is made smaller than the damage caused by the lapping of the other surface.
[0009]
[Action]
In the diffusion wafer of the present invention, the dislocation density on the surface of the non-diffusion layer is much higher than that of a conventional wafer in which the thickness of the non-diffusion layer is equal.
[0010]
The gradient of the dislocation density from the surface of the non-diffusion layer to the vicinity of the predetermined surface layer, which is larger than the gradient of the dislocation density from the vicinity of the predetermined surface layer to the interface with the diffusion layer, is the dislocation density and thickness of the surface required for the non-diffusion layer. It is desirable to change it in accordance with the standard.
[0011]
On the other hand, in the manufacturing method of the diffusion wafer, the side where the damage due to the lapping process (the depth of the crushed layer having the microcrack) is reduced is larger than the side where the damage is increased due to the diffusion of impurities during the formation of the diffusion layer. The ability to absorb and relax strain stress and thermal stress due to the temperature difference between the wafer outer peripheral part and the central part becomes smaller, and the minimum value of the dislocation density distribution in the non-diffusion layer after the formation of the diffusion layer is reduced by the thickness of the non-diffusion layer. It shifts to the side where the damage was larger than the center of the sword.
[0012]
It is desirable to change the degree of damage on the side where the damage due to the lapping process is kept small in accordance with the standard of the thickness of the non-diffusion layer.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
1 and 2 are cross-sectional views showing an example of the embodiment of the diffusion wafer according to the present invention, and a dislocation density distribution diagram of a non-diffusion layer in the diffusion wafer of FIG.
[0014]
This diffusion wafer has a high concentration of impurities (typically, P (phosphorus) for n-type and B (boron) for P-type) on one (lower in FIG. 1) surface of the silicon wafer. It has a two-layer structure having a diffused layer 1 diffused and a non-diffused layer 2 on the other (upper in FIG. 1) surface having a mirror-finished surface S and no impurity diffused.
The dislocation density on the surface of the non-diffusion layer 2 serving as the device formation surface is 5,000 to 20,000 / cm ^{2 in} order to improve the gettering ability in the device process and the high-frequency characteristics when the device finally becomes an element. and a high level, and the non-diffusion layer 2, a predetermined vicinity of the surface layer from the surface S (e.g., the depth of 10~30μm from the surface) N dislocation density gradually decreases to _S, and a predetermined surface near the N has a dislocation density distribution dislocation density increases gradually to the interface B between the diffusion layer 1 from _S, the slope of the dislocation density from the surface S of the non-diffusion layer 2 to a predetermined surface near N _S is near the predetermined surface It is greater than the slope of the dislocation density to the interface B between the diffusion layer 1 from N _S.
Incidentally, the gradient of dislocation density from the surface S of the non-diffusion layer 2 to a predetermined surface near N _S is intended to be varied in correspondence with dislocation density and thickness of the standard surface S obtained in the non-diffusion layer 2 is there.
[0015]
In order to manufacture the above-mentioned diffusion wafer, first, as shown in FIG. 3 (a), a lapping process is performed on both surfaces of a silicon wafer W having a required thickness as a material using relatively coarse abrasive grains. It was a large surface D _L of the damage.
Next, as shown in FIG. 3B, one surface (the lower side in FIG. 3B) of the silicon wafer W is covered with a protective pad 3 via an adhesive tape (not shown), and then the other side. (in FIG. 3 (b) above) subjected to lapping with relatively fine abrasive grains on the surface of, and the other surface with small surface D _S of damage.
Next, the protective pad 3 is removed, and as shown in FIG. 3C, deposition and drive-in diffusion processes are sequentially performed on both surfaces of the silicon wafer W whose both surfaces have been changed to diffuse high-concentration impurities. after There has been formed a diffusion layer 1 by interposing the non-diffusion layer 2 which is not diffused into the intermediate, as shown in FIG. 3 (d), the diffusion layer 1 and the non-diffusion layer 2 of the small surface D _S side of damage was ground and removed, the surface of the exposed non-diffused layer 2 was mirror-finished to obtain a diffusion wafer as a mirror M _S.
[0016]
Here, in order to measure the dislocation density distribution of the non-diffusion layer, first, lapping with # 800 abrasive grains was performed on both sides of a 125 mm-diameter n-type silicon wafer by the CZ method, and 15 μm on each side, 30 μm in total. Then, a protective pad made of a fluororesin having a thickness of 1.5 mm and a diameter of 125 mm is pasted on one surface through an adhesive tape to protect one surface, and then the other surface is coated with # 1200 abrasive grains. An n-type silicon wafer was obtained by changing the lapping process by 10 μm to change the damage on both sides.
Next, an n-type silicon wafer having both surfaces changed is charged into an electric furnace, the temperature in the furnace is kept at 1200 ° C., and oxygen, nitrogen and POCl ₃ gas are introduced into the furnace and heat-treated for 180 minutes. After forming a diffusion layer on both sides of the n-type silicon wafer, the n-type silicon wafer is heat-treated at a temperature of 1290 ° C. for 300 hours in an argon gas atmosphere to form a diffusion layer in which impurities are further diffused. Thereafter, the diffusion layer on the side with less damage was removed by grinding to expose the non-diffusion layer.
Next, the ground surface was polished with a margin of about 20 μm, and the damage caused by the grinding was completely removed to obtain a diffusion wafer. Then, the surface of each of these diffusion wafers is polished every 20 μm in the depth direction to prepare a sample, and a sample is prepared by angle wrap to show the tendency of the change in the dislocation density from the diffusion wafer surface to the diffusion layer interface. The dislocation density was evaluated. The results are shown in FIG.
[0017]
As can be seen from FIG. 4, the dislocation density distribution of the non-diffusion layer, the dislocation density gradually decreases from the interface B ₁ of the small surface D _S side of the damage to a position beyond the center C of the thickness of the non-diffusion layer, there minima after a, toward the interface B ₂ larger surface D _L side damage by increasing dislocation density, and the gradient of the dislocation density in the boundary of the minimum value smaller surface D _S side of the damage, a large surface damage It is greater than the slope of the dislocation density of D _L side.
It should be noted that the above-mentioned tendencies vary depending on the electrical resistance, oxygen concentration, type, surface damage, diffusion conditions, diffusion layer thickness, etc. of the silicon wafer before diffusion. Is desirable.
[0018]
【The invention's effect】
As described above, according to the diffusion wafer and the method of manufacturing the same of the present invention, the dislocation density on the surface of the non-diffusion layer is significantly higher than that of the conventional one in which the thickness of the non-diffusion layer is equal. In addition, raw materials can be reduced.
Further, by changing the gradient of the dislocation density from the surface of the non-diffusion layer to the vicinity of a predetermined surface layer, both the specification of the dislocation density and the thickness of the surface required for the non-diffusion layer can be satisfied.
[Brief description of the drawings]
FIG. 1 is a sectional view showing an example of an embodiment of a diffusion wafer according to the present invention.
FIG. 2 is a diagram showing a dislocation density distribution of a diffusion layer in the diffusion wafer of FIG.
3 (a) to 3 (d) show an example of an embodiment of a method for manufacturing a diffusion wafer according to the present invention. FIG. 3 (a) is an explanatory view of a first step, and FIG. (C) is an explanatory view of a third step, and (d) is an explanatory view of a final step.
FIG. 4 is a diagram showing a dislocation density distribution of a non-diffusion layer in a specific example of the diffusion wafer of FIG. 1;
[Explanation of symbols]
First diffusion layer 2 non-diffusion layer S surface _{N S} predetermined surface near the B interface W Silicon wafer _{D L} larger surface _{D S} less surface _{M S} mirror of damage damage

Claims (2)

The silicon wafer has a two-layer structure having a diffusion layer in which a high-concentration impurity is diffused on one surface and a non-diffusion layer on which the impurity is not diffused on the other surface. The dislocation density gradually decreases to near the surface layer of the diffusion wafer, and has a dislocation density distribution in which the dislocation density gradually increases from the vicinity of the predetermined surface layer to the interface with the diffusion layer, the predetermined surface layer from the surface of the non-diffusion layer A diffusion wafer characterized in that the gradient of the dislocation density to the vicinity is larger than the gradient of the dislocation density from the vicinity of a predetermined surface layer to the interface with the diffusion layer. In a method for manufacturing a diffusion wafer, a high-concentration impurity is diffused on both surfaces of a silicon wafer subjected to a lapping process, one diffusion layer is ground and removed, and the surface of an exposed non-diffusion layer is mirror-finished. A method of manufacturing a diffusion wafer, characterized in that damage caused by lapping of a surface on the other side is made smaller than damage caused by lapping of another surface.

Images