JPH068144A - Evaluation and control method for sand blasting condition in silicon monocrystal wafer - Google Patents

Abstract

PURPOSE:To enable a sandblasting condition to be evaluated with good repeatability by using a silicon monocrystal wafer having a <111> plane as principal plane, regarding a method for evaluating a sandblasting condition via the measurement of surface flaw density. CONSTITUTION:In a sandblasting condition evaluation method for a silicon monocrystal wafer, one side of a wafer is sandblasted and, then, subjected to oxidizing heat treatment. In addition, an oxidation induced stacking fault (OSF) is introduced onto the sandblasted surface of the wafer. Then, the flaw density of the surface is measured for evaluating a sandblasting condition. Also, a silicon wafer having <111> plane as principal plane is used. The density of OSF introduced through the oxidizing heat treatment after sandblasting any wafer having the <111> plane as principal plane, exhibits constant repeatability for a constant sandblasting condition, regardless of a manufacturing condition.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for evaluating and managing conditions in a sandblast method as a means for imparting an extrinsic gettering effect to a silicon single crystal wafer (hereinafter referred to as "wafer"). Is.

[0002]

2. Description of the Related Art In recent years, as semiconductor devices have become smaller and more highly integrated, quality requirements for wafers, which are substrate materials for semiconductor devices, have become more stringent. One of these requirements is semiconductor devices. In the above manufacturing process, there is a problem of suppressing the generation of crystal defects due to contaminants on the element formation surface of the wafer.

As a means for suppressing such crystal defects, a gettering method is known in which contaminant impurities mainly composed of heavy metals are collected in a strain field formed outside the element formation region in the element formation process. This gettering method is roughly classified into intrinsic gettering (IG) and extrinsic gettering (EG).

The method classified as extrinsic gettering mainly applies mechanical strain to the back surface of the wafer on the side opposite to the element formation surface, and a sandblast method is one of the methods. The size of the backside damage formed on the back surface of the wafer by this sandblasting method is an important element of the gettering technology, and oxidation induced stacking faults (oxidation induced stacki
It can be evaluated by the size of the density of ng fault (hereinafter abbreviated as OSF).

In the sandblasting method, a wafer 2 is placed on a conveyor belt 3 by a loader 1 as shown in a model of FIG. The wafer 2 on the conveyor belt 3 is shown in FIG.
It is conveyed in the direction of the arrow, and during the conveyance, a jet flow 5 of quartz particles or alumina particles is made to collide with the feeder 4 to perform sandblasting. The wafer 2 after the sandblast treatment is unloaded from the conveyor belt 3 by the unloader 6 and transferred to the next processing step. The main conditions that determine the effect of sandblasting are its strength and the processing time. Factors that give strength are the pressure and flow rate at which the jet flow 5 containing particles is sprayed from the feeder 4, the discharge port of the feeder and the wafer 2 to be processed.
In addition to the distance between and, the hardness of the particles used, the particle size distribution, and the concentration thereof are related. The processing time is adjusted by the belt speed of the conveyor belt 3, and its reciprocal corresponds to the processing time.

[0006]

By the way, the current wafer is cut out from a single crystal grown in the <100> orientation in order to increase the yield in manufacturing semiconductor devices.
An overwhelming number of wafers having the <100> plane as the main surface are used, and therefore, most of the wafers subjected to sandblasting also have the <100> plane as the main surface. By the way, even under the same sandblasting conditions, the OS may be different depending on the crystallographic orientation of the surface to be sandblasted.
Since it is known that the generation density of F is different, <10
The monitor during the sandblasting of the wafer having the 0> plane as the main plane is at least the same crystallographic orientation <10.
It is desirable to use a wafer whose main surface is the 0> plane.

However, in the case of a wafer having a <100> plane as a main surface, when the OSF density is measured by performing an oxidation heat treatment after sandblasting, the measured value of the OSF density is very unstable and poor in reproducibility. There was a problem that the conditions for sandblasting could not be set correctly.

The present invention is intended to solve the above-mentioned problems, and its purpose is to improve the reproducibility of the density of OSF introduced by sandblasting conditions and the subsequent oxidative heat treatment with good reproducibility.
It is to provide a method capable of performing an accurate evaluation, and further to provide a management method capable of performing condition setting at the time of sandblasting a wafer by applying this evaluation method.

[0009]

The method for evaluating the sandblasting conditions in a silicon single crystal wafer according to the present invention comprises: sandblasting one surface of a silicon single crystal wafer, and then performing an oxidative heat treatment, the sandblasted one surface of the wafer. In the method of evaluating the condition of sandblasting by introducing oxidation induced stacking faults on the surface and measuring the surface defect density thereof, the wafer having a <111> plane as a main surface is used.

The method of managing the sandblasting conditions for controlling the oxidation-induced stacking fault density of the silicon single crystal wafer according to the present invention is characterized in that when the silicon single crystal wafer is sandblasted, the method according to claim 1
It is characterized in that the relationship between the sandblasting condition and the oxidation-induced stacking fault density is measured in advance for a silicon single crystal wafer having a 111> plane as the main surface, and the condition of the sandblasting is set based on this relationship.

Furthermore, the method of managing the sandblasting conditions of the present invention is characterized in that when the silicon single crystal wafer is sandblasted, the silicon single crystal wafer having the <111> plane as the main surface is used as a monitor.

The wafer to which the present invention is applied is <100.
After cylindrical polishing with the pulling axis of the silicon single crystal ingot grown in the> axis direction or the <111> axis direction as the center of rotation,
A disk is obtained by slicing in the direction perpendicular to the pull-up axis, chamfering the edge of the disk and lapping the main surface, and then performing an etching process to remove residual processing strain. , Which is usually called a chemical etched wafer (CW wafer). That is, the wafer to which the present invention is applied is exactly an abbreviation for such a CW wafer. The sandblast treatment is performed thereafter, and the final mirror-finished wafer is the mirror-finished wafer (PW wafer) shipped as a product.

Various methods can be considered for evaluating the degree of mechanical strain generation on the sandblasted wafer surface, but in the case of the present invention, in the context of the EG effect, it occurs when the wafer after sandblasting is subjected to thermal oxidation treatment. The method of measuring the OSF density was adopted. That is, the sandblasted wafer is heat-treated at 1100 ° C. for 1 hour in a wet oxidizing atmosphere and then cooled, and the surface oxide film is removed by hydrofluoric acid treatment. The OS produced by observing the sandblasted surface of the wafer that has been etched, washed and dried by 1 μm with an optical microscope (NIKON IC microscope)
Measure F. In the OSF measurement, the measurement magnification is 1000 times, the OSF existing in the visual field is measured using a counter at three points where the arbitrary radial direction of the wafer is divided into three parts, and the density is converted to × 10 ⁴ pieces / cm ^2. Unified by doing.

[0014]

[Operation] O introduced by performing a thermal oxidation process after sandblasting a wafer having a <111> plane as a main surface.
Regarding the density of SF, regardless of the wafer manufactured under any condition (CW wafer), a constant OSF density is reproduced under a constant sandblast condition. Therefore, if a wafer whose main surface is the <111> plane is adopted as a monitor, it is possible to set the sandblast quantitatively. Therefore, when performing the sandblast treatment on the wafer having the arbitrary main surface, at least the sandblast condition is set. Or, it can be used for control.

[0015]

EXAMPLES Next, the present invention will be described in more detail with reference to examples. Example 1 and Comparative Example 1 As a sandblasting condition for wafers etched with various etching allowances, OSF with respect to its strength was used.
The relationship of occurrence is shown by <111> main surface wafer (Example 1)
And a wafer having a <100> main surface (Comparative Example 1) were examined.

[Test conditions and method] <111 in crystal orientation
> And <100> main surface of P-type single crystal (resistivity about 1
For a wafer whose main surface is the <111> plane, the etching amount is 25 μm and 3 μm, respectively.
Etching to 8 μm and 55 μm, <1
With respect to the wafer having 00> as the main surface, three wafers (CW wafers) each having an etching amount of 25 μm, 38 μm, and 48 μm were prepared. For each of these wafers, the setting conditions of the sandblasting strength (OSF generation amount after sandblasting, evaluated as the number / piece / cm ² ) using a P-type wafer with a <111> main surface as a monitor were 300,000, 500,000 and 700,000. Processing was performed at three levels.

[Test Results] Each 3 after sandblasting
OS for 3 observation points (total 9 points)
The F density was inspected, and the relationship between the average amount of the generated OSF, the set intensity of sandblast, and the etching amount of the wafer was shown in FIG. 1 (Example 1) and FIG. 2 (Comparative Example 1).

What is clear from FIGS. 1 and 2 is <11.
Both the 1> main surface wafer and the <100> main surface wafer correspond to the strength of the sandblast strength, and the same tendency is seen in that the generation of OSF increases or decreases. In the case of Example 1 (FIG. 1), the same sandblasting strength setting conditions were not affected by the wafer etching amount in any case, and the values were close to the setting conditions. On the other hand, Comparative Example 1
In the case of (FIG. 2), the generation of OSF corresponds to the etching amount of the wafer and shows an exponential increase. This fact has almost no effect on the etching amount of the <111> major surface wafer, so that no matter what etching amount the CW wafer is used, if the setting condition of the sand blast strength is constant, a constant OSF is maintained. In the case of <100> main surface wafers, the OSF generation value varies greatly depending on the etching amount, while the reproducibility of occurrence is obtained. Therefore, control any quality condition as a CW wafer including the etching amount. Unless it is done, reproducibility of the OSF generation value cannot be obtained. In other words, this fact is the reason why the <100> main surface wafer cannot be used as a monitor in the past. Although the data is omitted, it has been confirmed that the same result is obtained for the N-type wafer as for the P-type wafer.

Example 2 and Comparative Example 2 As a sandblasting condition for wafers etched with various etching allowances, the relationship between the amount (specified by the time of sandblasting) and the OSF generation is shown in the table below. Example 2) and a <100> main surface wafer (Comparative Example 2) were examined. [Test Conditions and Methods] Wafers manufactured from the same single crystals as those used in Example 1 and Comparative Example 1, that is, P-type single crystals having crystal orientations <111> and <100> as principal planes (resistivity of about 1Ωcm ordinary resistance product), the etching amount is 2 for each of the wafers whose main surface is the <111> plane.
A wafer (CW wafer) and a mirror surface wafer (CW wafer) that were etched to have a thickness of 5 μm, 38 μm, 55 μm, and had a <100> plane as a main surface were etched to have an etching amount of 25 μm, 38 μm, 48 μm PW wafers) were prepared in triplicate.
Each of the wafers was used as a setting condition for the sandblasting strength of the P-type monitor wafer on the <111> main surface by 5
The amount of sandblasting is fixed at 0,000 pieces / cm ² as shown in FIG.
The conveyor belt speed (m / min) in the sand blasting apparatus, which is shown in a simulated manner at 1.8, 1.
The sandblasting treatment was carried out by setting to 2, 0.8 and 0.6, and the reciprocal thereof (min / m; corresponding to treatment time) was taken as the sandblasting amount.

[Test Results] The OSF density at three observation points (total of 9 points) was inspected for each of the three wafers after sandblasting, and the average value of the OSF generation amount and the set amount of sandblasting (time ) And the etching amount of the wafer are shown in FIG. 3 (Example 2) and FIG. 4 (Comparative Example 2). 3 and 4 are common to each other, there is a clear proportional relationship between the increase in the amount of sandblast (processing time) and the amount of OSF generated.
In the case of (FIG. 3), regardless of the wafer etching amount, that is, regardless of the manufacturing condition of the CW wafer, the sandblasting time and the OSF generation amount show a constant relationship, while the comparative example Comparative example 1 in 2 (FIG. 4)
Similar to the case, there is a significant difference between the sandblasting time and the OSF generation amount for wafers with various etching amounts. Therefore, unless elucidation of this cause and a standard for management are established, a wafer having a <100> plane as a principal surface cannot be used as a monitor for evaluating or managing sandblast conditions.

Example 3 A wafer having a <111> major surface was used as a monitor for sandblasting for 20 days, and its suitability as a monitor was examined.

[Test Conditions and Method] A wafer having the same conductivity type and resistivity as those used in Examples 1 and 2 having a <111> plane as a main surface (etching amount: about 37 μm).
m) 20 sheets were prepared. In the test, each morning, each of the monitor wafers having the <111> plane as the main surface was sandblasted simultaneously with the other wafers, and then the monitor wafers were inspected for OSF generation by a predetermined method. As a sandblast condition, the set strength is 500,000 pieces / cm when using a <111> monitor.
² and the belt speed was 1.2 m / min under the common conditions.

[Test Results] The test results are shown in FIG. It can be seen that the wafers used for the monitor have a reproducible value with the level value corresponding to the sandblast strength set to the above-mentioned 500,000 pieces / cm ² .

[0024]

According to the sandblasting condition evaluation method of the present invention, the sandblasting condition can be evaluated with good reproducibility. Further, according to the sandblasting condition management method of the present invention, the sandblasting condition can be controlled so that the OSF density of the silicon single crystal wafer becomes a predetermined value.

[Brief description of drawings]

1] Etching amount and OS of <111> P-type wafer
It is a graph which shows the relationship of F density.

2] Etching amount and OS of <100> P-type wafer
It is a graph which shows the relationship of F density.

FIG. 3 is a graph showing the relationship between the OSF density of a <111> P-type wafer, the reciprocal of the belt speed (sandblast amount), and the wafer etching amount.

FIG. 4 is a graph showing the relationship between the OSF density of a <100> P-type wafer, the reciprocal of the belt speed (sandblast amount), and the wafer etching amount.

FIG. 5 is a graph showing changes in OSF density in working days in Example 3.

FIG. 6 is an explanatory diagram simulating the sandblast method.

[Explanation of symbols]

 1 Loader 2 Wafer 3 Conveyor Belt 4 Feeder 5 Jet 6 Unloader

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Shoji Nakamura 1393 Yashiro Odaira, Sarahaku-shi, Nagano Nagano Electronics Co., Ltd. 72) Inventor Moriaki Torigoe 596-2, Jono Koshi Nitta, Nakagibiki-gun, Nakakubiki-gun, Niigata Naoetsu Electronic Industry Co., Ltd. Electronic Industry Co., Ltd. (72) Inventor Michito Sato No.150 Ohira, Odakura, Saigo Village, Nishishirakawa-gun, Fukushima Prefecture Shin-Etsu Semiconductor Co., Ltd. Shirakawa Factory

Claims (3)

[Claims] 1. A sandblasting method is carried out by sandblasting one side surface of a silicon single crystal wafer, followed by oxidation heat treatment, introducing oxidation-induced stacking faults on the sandblasted one side surface of the wafer, and measuring the surface defect density. In the method for evaluating the conditions of (1), a silicon single crystal wafer having a <111> plane as a main surface is used as the wafer, and an evaluation method of a sandblasting condition in the silicon single crystal wafer. 2. When the silicon single crystal wafer is sandblasted, the <11
A silicon single crystal wafer having a 1> plane as a main surface is preliminarily measured for a relationship between a sandblasting condition and the oxidation-induced stacking fault density, and the sandblasting condition is set based on this relationship. A method for controlling sandblasting conditions for controlling the oxidation-induced stacking fault density of a crystal wafer. 3. The silicon single crystal according to claim 1 or 2, wherein a silicon single crystal wafer having a <111> plane as a main surface is used as a monitor when the silicon single crystal wafer is sandblasted. Management method for sandblasting conditions on wafers.