JPH0235455B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for manufacturing a semiconductor device in which a semiconductor element is formed on a semiconductor substrate such as a silicon wafer.

Conventionally, when gettering the generation nuclei of metal impurities and crystal defects existing in silicon wafers, it is generally done by mechanically damaging the back surface of the wafer (i.e., the surface opposite to the element formation surface) or by impurity diffusion. In this method, misfit dislocations are introduced by ion implantation, or secondary defects generated by ion implantation are utilized. In this way, the defects formed on the back surface of the wafer getter metal impurities within the wafer during the heat treatment process, thereby preventing the occurrence of micro defects on the front surface of the wafer used for device formation. do.

However, since such conventional gettering methods perform gettering processing from the back side of the wafer, when the wafer thickness increases to about 300 to 500 μm as the diameter of the wafer increases (3 to 5 inch φ), the gettering process is performed from the back side of the wafer. It becomes impossible to prevent the occurrence of defects, and the gettering effect becomes difficult to appear. If an attempt was made to strengthen the gettering effect in order to avoid this, there was a problem in that dislocations caused by the gettering treatment would reach the element forming surface (front surface) side.

The present invention has been made to correct the above-mentioned defects, and includes a step of forming a primary defect layer in a region on the element forming surface side of a semiconductor substrate (for example, a silicon wafer) by mechanical polishing or ion implantation;
A step of converting the primary defect layer into a secondary defect layer by subjecting the semiconductor substrate to a heat treatment, and a step of converting the secondary defect layer into an oxide layer (for example, a thermal oxide film using a high-pressure oxidation method); and a step of removing the oxide layer (for example, by etching), and after the oxide layer removal step, a semiconductor element is formed on the element forming surface side of the semiconductor substrate. This relates to a manufacturing method.

According to the method of the present invention, even if a semiconductor substrate having a large thickness is used, it is possible to extremely effectively prevent the occurrence of defects on the element formation side. It is possible to improve the characteristics of a semiconductor element formed in a semiconductor device.

Next, an embodiment of the present invention using a silicon wafer as a semiconductor substrate will be described with reference to the drawings.

First, the gettering method according to the present example will be explained as follows.
3 will be explained. In this case, gettering processing is performed from the wafer surface (element forming surface) side.

That is, as shown in Figure 1, the thickness is 300 to 500 μm.
A shallow (for example, within 1 μm) damaged layer 2 is created by slightly mechanically damaging the surface of a large-diameter silicon wafer 1 by polishing or by performing ion implantation (for example, boron implantation at 50 to 200 KeV). form. This damaged layer 2 is a region where primary defects exist, that is, a primary defect layer, and the density peak of primary defects exists in a region at a depth of about 2000 Å from the surface.

Next, as shown in FIG. 2, heat treatment is performed on the front side of the wafer 1 at about 1000°C to transform the damaged layer 2 into a secondary defect layer 3 containing secondary defects such as dislocation networks and dislocation loops. Convert to During this heat treatment, secondary defects act as gettering centers in the secondary defect layer 3, which getter crystal defect generation nuclei and metal impurities inside the wafer 1, so that the secondary defect layer 3 contains harmful defects and impurities at a high density. At this time, crystal defect generating nuclei and the like are naturally gettered from the silicon directly under the secondary defect layer 3, and the silicon becomes clean. Therefore, the secondary defect layer 3 functions as a gettering source, but its thickness is practically the same as that of the above-mentioned damaged layer 2, and is practically within 1 μm. Note that the above heat treatment may be performed in an inert gas.

Next, as shown in FIG. 3, the surface side of the wafer 1 is subjected to high-pressure oxidation, which will be described later, to convert the secondary defect layer 3 into an oxide layer, for example, an SiO ₂ layer. This SiO ₂ layer 4
grows to a sufficient thickness (for example, 2.2 μm or more) to cover the entire thickness of the secondary defect layer 3, thereby completely converting the secondary defect layer 3 into the SiO ₂ layer 4. It turns out. That is, if the thickness of the grown SiO ₂ layer 4 is t, the thickness of silicon consumed by the high-pressure oxidation is generally given by approximately 0.45 (constant)×t. Therefore, if t≧2.2 μm, the thickness of the consumed silicon will be approximately 1 μm or more, and the secondary defect layer 3 will be almost completely converted into the SiO ₂ layer 4.
In short, in order to remove (convert) the secondary defect layer 3, the thickness of the SiO ₂ layer 4 must be 2.2 μm or more, preferably 3 μm or more (in this case, the thickness of the silicon consumed is approximately
It is necessary to grow the film to a thickness of 1.35 μm or more.

Next, although not shown, the SiO ₂ layer 4 is removed by etching to expose the cleaned and defect-free crystal plane of the wafer 1. Then, semiconductor elements such as diodes and transistors are formed on this crystal plane by a conventionally known method. As is clear from the above, the crystal plane of the wafer for device formation has been completely cleaned by removing the SiO ₂ layer 4 which has completely converted the secondary defect layer 3. The characteristics of the semiconductor element formed in this manner are extremely excellent.

An important point in the method described above is that a high-pressure oxidation method is used in the step shown in FIG. 3 (conversion of secondary defect layer 3 into SiO ₂ layer 4). According to this high-pressure oxidation method, an oxide film of 2.2 μm or more can be easily formed at a relatively low temperature, and the amount of silicon consumed is as described above. This high-pressure oxidation must be carried out at a pressure higher than atmospheric pressure and at a temperature at which defects do not grow significantly, but it is carried out at a temperature of 1000°C or lower (preferably 950°C or lower), for example under a pressure of 6.6 kg/ ^cm2. is desirable. In other words, a state in which the growth of stacking faults that can grow due to high-pressure oxidation is suppressed (almost zero).
If you try to form a SiO ₂ layer with a thickness of about 2.2 μm or more,
It has been experimentally confirmed that the temperature must be below 1000°C (preferably below 950°C). Conversely, when the temperature exceeds 1000°C, the defects tend to extend, and the size of the stacking fault becomes larger than the thickness of the SiO ₂ layer 4. Therefore, unless the oxidation treatment is performed under conditions that prevent the defects from extending, It becomes difficult to eliminate defects in crystal planes.

As an example, when high-pressure oxidation is performed at 950°C for 60 minutes to form an oxide film with a thickness of 2 μm,
When the pressure is 25Kg/ ^cm2 and the process is carried out at 1000℃ for 60 minutes to form a 2μm thick oxide film, the pressure should be 25Kg/cm2.
15Kg/ ^cm2 . In any of these cases, since the temperature is low and the pressure is considerably higher than 6.6 Kg/cm ² , stacking faults can be prevented from extending. However, the oxidation time is also relevant, and generally the pressure can be lowered by lengthening the oxidation time, but it may be arbitrarily determined in consideration of workability and the thickness of the oxide film.

As described above, according to the above-described high-pressure oxidation method, a thick oxide film can be formed at a relatively low temperature, and the growth of a secondary defect layer can be substantially suppressed. Therefore, a larger amount of the silicon surface layer (secondary defect layer) can be consumed or removed than with conventional oxidation methods. In addition, in conventional methods, it is common to sequentially form device active regions on the element formation surface. Removing an area and then forming another active area there
This is something that cannot be considered from conventional concepts. However, in the above example, by actively applying such a process as a gettering process,
This has resulted in extremely excellent effects.

In addition, in the above-mentioned embodiment, it is also possible to form the damaged layer 2 by impurity diffusion. However, in this case, the impurities enter into the interior of the wafer 1 during the subsequent heat treatment process, so the formation of the damaged layer 2 in the present invention must be performed by mechanical polishing or ion implantation.

Next, another embodiment in which the present invention is applied to a method of manufacturing a MOS-FET will be described with reference to FIG.

First, as shown in FIG. 4A, an N-type silicon substrate 1 is surface oxidized to form a SiO ₂ film 5, and then, as shown in FIG. 4B, a P-type source is formed using the SiO ₂ film 5 as a mask. Region 6 and drain region 7 are formed by diffusion. Next, as shown in FIG. 4C, ions 8 such as Si ⁺ , Ar ⁺ , Cl ⁺ are irradiated from the surface side to implant these ions into the surface region of the substrate.
As a result, a shallow ion-implanted layer, ie, a damaged layer 9, is uniformly formed.

Next, by annealing in an inert gas, the damaged layer 9 is converted into a secondary defect layer 10, as shown in FIG. 4D, and defect generation nuclei and metal impurities in the substrate 1 are gettered. Then, as shown in FIG. 4E, after removing the SiO ₂ film 5 by etching, _the surface is oxidized under high pressure, and as shown in FIG. Convert. Next, as shown in FIG. 4G, SiO ₂
After removing the film 11 by etching, a gate oxide film 12 is newly formed as shown in FIG. 4H. Thereafter, by a conventionally known method, a source metal electrode and a drain metal electrode are deposited to correspond to the source region 6 and drain region 7, respectively, and a gate metal electrode is deposited on the gate oxide film 12. Complete MOS-FET.

The present invention has been described above based on examples, but
This embodiment can be further modified based on the technical idea of the present invention. For example, depending on the thickness of the secondary defect layers 3 and 10, the thickness of the oxide film formed by high-pressure oxidation may be controlled based on the above-mentioned formula of approximately 0.45×t=thickness of the secondary defect layer. In addition, the growth of stacking faults during high-pressure oxidation can be prevented by the above-mentioned low-temperature pressure oxidation, but the reaction (unfaulting reaction) due to the formation of dislocation networks (secondary defects) after ion implantation
It is thought that this can also be prevented by converting it into a dislocation loop. In any case, stacking faults can be effectively prevented from penetrating the cleaned silicon surface.

Furthermore, the surface treatment of the semiconductor substrate performed in the present invention is not only for the purpose of gettering as described above.
All or part of an insulating layer such as a polysilicon layer or nitride layer formed on the wafer surface is converted into an oxide by oxidation, and after this oxide is removed, a new layer is formed in the removed area. It can also be applied to repairs.
Furthermore, if the growth of defects occurs only to such an extent that it does not substantially pose a problem, normal atmospheric pressure oxidation can be performed for a short time instead of the above-mentioned high pressure oxidation.
Furthermore, the semiconductor substrate may be made of a material other than silicon.

Since the present invention has the above-described configuration, it is possible to form and oxidize the primary and secondary defect layers and remove the oxide layer before forming the semiconductor element on the oxide-removed portion of the semiconductor substrate. can, therefore,
Even if a thick semiconductor substrate is used, it is possible to extremely effectively prevent the occurrence of defects on the element formation side. It can be made excellent. In particular, in the present invention, since the primary defect layer is formed by mechanical polishing or ion implantation, sufficient damage is caused only to a shallow region (for example, within 1 μm) of the semiconductor substrate, compared to the case where it is formed by impurity diffusion. can be given. Therefore, there is no fear that impurities will enter the interior of the semiconductor substrate in the subsequent heat treatment step.

[Brief explanation of drawings]

1 to 3 show one embodiment of the present invention, in which FIG. 1 is a cross-sectional view of a state in which a damaged layer is formed on the wafer surface, and FIG. Figure 3 is a cross-sectional view of the secondary defect layer converted to an oxide layer, and Figures 4A to 4H are MOS-FETs of the present invention.
FIG. 4 is a cross-sectional view showing another embodiment applied to the manufacturing method in order of steps. In addition, in the symbols used in the drawings, 1... silicon wafer, 2, 9... damaged layer, 3, 10...
... Secondary defect layer, 4, 11 ... SiO ₂ film, 8 ... Impurity ion.

Claims (1)

[Scope of Claims] 1. A step of forming a primary defect layer in a region of the semiconductor substrate on the element forming surface side by mechanical polishing or ion implantation, and heat-treating the semiconductor substrate to transform the primary defect layer into a secondary defect layer. a step of converting the secondary defect layer into an oxide layer; and a step of removing the oxide layer, and after the oxide layer removal step, the device of the semiconductor substrate is removed. A method for manufacturing a semiconductor device in which a semiconductor element is formed on a forming surface side.