JP5819600B2 - Semiconductor wafer evaluation method - Google Patents

Description

  The present invention relates to a semiconductor wafer evaluation apparatus and evaluation method for evaluating the impact resistance of a wafer used in a semiconductor device manufacturing process, such as a silicon wafer, and more specifically, supplies a wafer exhibiting stable impact resistance. The present invention relates to an evaluation apparatus and an evaluation method for a semiconductor wafer.

  Silicon wafers tend to have orders of magnitude smaller impact strength when hit against the wafer than static stress. There are many cases where the wafer edge is mainly hit, and it is important to evaluate the impact strength of the edge.

In the semiconductor device manufacturing process, when a silicon wafer as a material is cracked, a large loss occurs.
For this reason, there is a strong demand for semiconductor wafers that are difficult to break during device manufacture.

In semiconductor and liquid crystal manufacturing processes, especially dry etching, ion implantation, vapor deposition, etc., high temperature / rapid heating / rapid cooling has progressed, and the number of manufacturing processes performed under vacuum and dry environments has increased. Yes.
In addition, silicon wafers and glass substrates as substrates have become larger in diameter, and impact resistance has become more important.

Here, in the case of plate glass, there are various methods for measuring impact fracture strength and performing statistical processing.
For example, “the strength of the plate glass used in construction uses a thickness that can withstand the breaking strength of 1/1000 sheets” is used as a guideline.
However, there is no such standard for silicon wafers, and the development of analysis methods is desired.

Here, since the silicon wafer can be said to be a brittle material due to its crystallinity, there is a large variation in measurement values in a general material evaluation technique.
For this reason, standard equipment for evaluating and inspecting the easiness of cracking of a silicon wafer is not commercially available. For this reason, for example, evaluation methods and evaluation apparatuses such as Patent Documents 1 to 4 have been devised.

Here, the measuring apparatus and measuring method of Patent Document 1 will be briefly described. In FIG. 5, the schematic which is an example of the mechanical strength evaluation apparatus of patent document 1 is shown. (A) is a schematic side view, (b) is a schematic plan view of a part thereof.
The mechanical strength measuring apparatus 110 includes a mounting table 102 on which the semiconductor wafer 101 is mounted, a supporting unit 103 that is provided on the mounting table 102 and supports the semiconductor wafer 101 at at least two points A and B on the outer periphery of the wafer, and a load. And a load means 105 for moving the shaft 104 in parallel and pressing the tip against the outer periphery of the wafer. Then, after the semiconductor wafer 101 is placed on the mounting table 102, the supporting means 103 supports the semiconductor wafer 101 at at least two points A and B on the outer periphery of the wafer, while the load means 105 supports the tip of the load shaft 104 on the outer periphery of the wafer. And a static pressure load is applied toward the center O of the semiconductor wafer 101.
The load means 105 includes an air cylinder 105a and a pressure control valve 105b. The load shaft 104 is moved by the air cylinder 105a, and the static pressure load is controlled by the pressure control valve 105b. The pressure control valve 105b is connected to an Ar gas cylinder 106, for example.

  First, the semiconductor wafer 101 is mounted on the mounting table 102 using the measuring apparatus 110. Then, while supporting the semiconductor wafer 101 at at least two points A and B on the outer periphery of the wafer by the support means 103 provided on the mounting table 102, the load shaft 104 of the load means 105 is moved in parallel, and the tip of the load shaft 104 is moved. Is pressed against one point C on the outer periphery of the semiconductor wafer 101, a static pressure load is applied toward the center of the semiconductor wafer 101, and the static pressure load when the semiconductor wafer 101 is broken by increasing the static pressure load. taking measurement.

However, the silicon wafer breaking strength in the measuring apparatus and measuring method of Patent Document 1 is very high and varies widely. In order to evaluate the slight difference in the impact resistance of the semiconductor wafer targeted by the present invention, the sensitivity・ I am concerned that the ability of accuracy is insufficient.
Further, in the method of Patent Document 2-4, it is an evaluation method in which strength = the number of repeated hits, and considering that a silicon wafer is a brittle material, it is considered that the influence of the number of repeated hits is overestimated.

JP 2006-287139 A JP 2004-101258 A JP-A-5-288663 Japanese Patent Laid-Open No. 6-201533

Since silicon wafers are crystalline brittle materials, there is a large variation in measured values in general material evaluation techniques. In addition, there is no JIS standard for the evaluation of edge portions that are particularly problematic.
In silicon wafers several generations ago, there were cases where precipitation, strain, etc. remained in the outermost periphery, and there were cases where the fracture strength in one lot exceeded “maximum value ÷ minimum value = 10 times”.
In such a case, selection by the strength evaluation method “destructive strength = number of repeated hits” by the above-described prior art method was also possible.

However, recent silicon wafers have greatly improved crystal manufacturing methods and wafer processing methods. For this reason, the difference between the “maximum value ÷ minimum value” of the breaking strength in one lot is very small, and the correlation between the breaking strength and the number of repeated hits has become low.
For this reason, it is extremely difficult to evaluate the fracture strength of a silicon wafer by the conventional method, and improvement of the evaluation method has been demanded.

  The present invention has been made in view of the above problems, and an object thereof is to provide an evaluation apparatus and an evaluation method that can stably and accurately evaluate the impact strength at the edge portion of a semiconductor wafer.

In order to solve the above problems, the present invention is an apparatus for evaluating the impact strength of a semiconductor wafer,
At least a striking substance (hereinafter also referred to as a bell), dropping means for dropping the striking substance from any desired height, and a sample which is a semiconductor wafer, the wafer radial direction is a drop of the striking substance Holding means for supporting the same to the direction,
In a state where the sample is supported by the pressing means so that the radial direction of the wafer is equal to the falling direction of the hitting substance, the hitting substance is directed from the desired height to the edge portion of the sample by the dropping means. A semiconductor wafer evaluation apparatus characterized in that the apparatus can be dropped is provided.

  With such an evaluation apparatus, the impact strength applied to the edge portion of the sample made of a semiconductor wafer can be varied, and the impact strength of the semiconductor wafer can be evaluated by the so-called steer case method, so compared with the conventional case. The impact strength of a semiconductor wafer having a small variation width can be evaluated. That is, the evaluation apparatus can evaluate an impact strength value having high reliability that can be an index of the ease of cracking of a wafer in a semiconductor device manufacturing process.

Moreover, this invention is a method of evaluating the impact strength of the edge part of a semiconductor wafer, Comprising: The following process (1)-(6) can be performed at least.
(1) A sample which is a semiconductor wafer is supported by pressing means for supporting the wafer so that the radial direction of the wafer is equal to the falling direction of the striking substance, and the striking substance is separated from the sample by a predetermined weight and height. Drop on the edge.
(2) As a second experiment, a sample different from the sample in step (1) is supported by the pressing means, and if the sample is not broken in step (1), the position is higher than in step (1). If the sample is destroyed in the step (1), the position is made lower than that in the step (1), or the weight is low. Then, the hitting substance is dropped.
(3) As a third experiment, when a sample different from the samples in steps (1) and (2) is supported by the pressing means, and (2) the sample is not destroyed in step (2) The position is made higher than the process, or the impact material is dropped after being made heavy, and if the sample is destroyed in the process (2), the position is made lower than the process (2), or The impact material is dropped after the weight is reduced.
(4) The above process is repeated until the predetermined number n is reached.
(5) 50% impact from the relationship between the number of samples destroyed and the height or weight of the impacting material at that time, and the number of samples not destroyed and the height or weight of the impacting material at that time. The standard deviation (SH) of the fracture height (H ₅₀ ) and the 50% impact fracture height is calculated.
(6) The standard deviation (SE) of 50% impact fracture energy (E ₅₀ ) and 50% impact fracture energy is calculated from the H ₅₀ and the SH.

As described above, according to the present invention, the impact resistance against the dynamic stress at the edge portion of the semiconductor wafer can be evaluated by dropping the striking substance from a predetermined height. The present invention can be a test in an environment closer to an actual semiconductor device process than a general material test in which static stress is evaluated such as a bending fracture test. Moreover, since the drop test is performed a predetermined number of times, the room for accidental trouble can be reduced as compared with the single test, and the evaluation can be made with high accuracy. Furthermore, since the impact strength is changed in accordance with the immediately preceding test result, it is possible to reduce the room for measurement errors due to defects in the wafer itself, which contributes to higher accuracy.
From these results, it becomes an evaluation method capable of stably and accurately evaluating the impact strength of the semiconductor wafer.

  At this time, the sample in each step is preferably manufactured by dividing one semiconductor wafer.

  In this way, by dividing a single semiconductor wafer and preparing a sample, a plurality of impact strengths can be evaluated from a single semiconductor wafer, so that the wafer can be saved and the evaluation cost can be increased. Can be prevented. Further, although different semiconductor wafers may have slightly different impact strengths, such problems can be alleviated and a more accurate evaluation method for impact strength of semiconductor wafers can be obtained.

  At this time, it is also preferable to perform the steps (1) to (6) on a semiconductor wafer of a different type from the sample in each of the above steps, and compare the impact strength of the semiconductor wafer from the results. .

  As described above, in the steps (1) to (6) of the present invention as described above, the impact strength of the semiconductor wafers can be accurately and quantitatively evaluated, which is very suitable for comparing the impact strengths of the semiconductor wafers. It is possible to quantitatively compare the impact strength between semiconductor wafers by first performing the evaluation process on one semiconductor wafer and then performing the evaluation process on another semiconductor wafer. become.

  As described above, it is possible to statistically evaluate the impact strength of a semiconductor wafer by varying the impact strength applied to the edge portion of a sample made of a semiconductor wafer as in the present invention. Compared with the evaluation apparatus, it is possible to improve the evaluation capability and accuracy, and to quantitatively evaluate the fragility of the edge portion of the semiconductor wafer in the semiconductor device manufacturing process.

It is the figure which showed an example of the outline of the semiconductor wafer evaluation apparatus of this invention. It is the figure which showed an example of the pressing means of the semiconductor wafer evaluation apparatus of this invention, and the sample to be used. It is the figure which expanded a part of other example of the hit | damage substance of the semiconductor wafer evaluation apparatus of this invention, and its dropping means. It is the figure which showed another example of the pressing means in the semiconductor wafer evaluation apparatus of this invention. It is the figure which showed the side surface schematic diagram of an example of the mechanical strength evaluation apparatus of patent document 1, and the one part plane schematic diagram.

Hereinafter, the present invention will be described more specifically.
As described above, it is extremely difficult to evaluate the fracture strength of a silicon wafer by the conventional method, and improvement of the evaluation method has been demanded.

  Therefore, the present inventor noticed that the impact strength when hitting the wafer is much smaller than that of the static stress in the semiconductor wafer, and that the semiconductor wafer Intensive study was conducted on the evaluation method and apparatus capable of evaluating the impact strength.

  As a result, the present inventor uses an impact resistance test method using the principle of the staircase method with a constant drop weight and an evaluation apparatus capable of performing the same, and analyzes the measurement result by the staircase method. I thought that the impact strength of the edge part of a semiconductor wafer could be statistically evaluated by this.

  The steer case method is a method of statistically analyzing the impact fracture strength from the number of samples with and without the destruction of the sample when the stress level is divided into the respective levels and the stress level value at that time (for example, Dixon, W. J. and Mood, AM, J. Amer. Stat. Assn., Vol. 43, pp. 109-126, 1948, etc.). This staircase method is often used in quality inspections, and it is known that the number of trials can be reduced as compared with the constant stimulation method, and that the accuracy is high.

  And it was discovered that if such a method and an apparatus capable of performing this were able to statistically evaluate the impact fracture strength of a semiconductor wafer such as a silicon wafer, the present invention was completed.

  Hereinafter, the present invention will be described in detail with reference to the drawings, but the present invention is not limited thereto. FIG. 1 is a diagram showing an example of an outline of a semiconductor wafer evaluation apparatus of the present invention.

As shown in FIG. 1, a semiconductor wafer evaluation apparatus 1 according to the present invention includes at least a striking substance 5, for example, a height adjusting rail 2, a slider 3, and an electromagnetic magnet 4, and the striking substance 5 has an arbitrary desired height. Is supported by a constant force so that the radial direction of the sample W, which is a semiconductor wafer, and the falling direction of the striking substance 5 are equal. Presser means 7a for the purpose.
Then, in a state where the sample W is supported by the holding means 7a so that the radial direction of the sample W and the falling direction of the hitting substance 5 are equal, the hitting substance 5 is moved from the desired height by the dropping means 15 to the edge portion of the sample W. It can be dropped vertically toward

Moreover, it is desirable to install a polycarbonate cover 6 on the holding means 7a for preventing the scattering of the fragments when the sample W is broken and the impacting material 5 after colliding with the sample W.
The striking substance 5 starts to fall from the electromagnetic magnet 4, passes through the polycarbonate cover 6, and collides with the edge portion of the sample W. Thereafter, the striking substance 5 is recovered by the polycarbonate cover 6.

FIG. 2 is a view showing an example of a holding means and a sample used in the semiconductor wafer evaluation apparatus of the present invention.
The silicon wafer W0 is a (100) wafer when expressed in a plane orientation, and a notch 9 is engraved at (011). The wafer is divided into four along the (110) cleaved surface, and a sample W which is a fan-shaped silicon piece is produced.
The sample W is placed so as to be exposed from the notch 10a, and is firmly sandwiched between the two pressing means 7a and held with a constant force.
And the pressing means 7a is installed in the apparatus base 8 so that the radial direction of the sample W may become equal to the falling direction of the hit | damage substance 5. FIG.

The notch 10a in FIG. 2 has a small semicircular shape. Thereby, the holding region of the sample W is widened, and the deformation region is narrowed. Such a pressing means is used when evaluating “strength at the start of breakage of a minute portion” with a small deformation amount of the silicon piece.
Since the edge portion of the sample W has a fan-shaped arc shape as shown in FIG. 2, the impact force per point in the vertical direction is always applied to the striking region with the bottom of the striking material 5. In this respect, the striking material 5 is preferably cylindrical rather than spherical.

In such an evaluation apparatus 1, the strength of the breaking energy can be controlled by changing the weight (size) of the striking substance 5 and changing the height of dropping.
Specifically, when the impact strength of the semiconductor wafer is evaluated, the slider 3 is moved up and down by the height adjustment rail 2 to drop the striking substance 5 from an arbitrary height (0 to 2000 mm), and the edge of the sample W It is possible to repeatedly evaluate whether the drop reference position is lowered when the sample is damaged, and the drop reference position is raised when the sample is not damaged. It is possible to increase or decrease the weight of the striking substance instead of the height of the striking substance, but it is easier to adjust the height.
During a series of evaluations of impact strength, it is desirable to fix one of the falling height of the striking material or the weight of the striking material and change only the other.

  With such an evaluation apparatus, impacts with different strengths can be easily applied to the edge portion of a sample made of a semiconductor wafer, so that an impact applied in an actual semiconductor process can be simulated. The applied impact can be easily changed and repeated by changing the height at which the striking substance is dropped, and a statistical method such as the staircase method can be employed. Therefore, it is possible to evaluate the impact strength of a semiconductor wafer with small variations, and to obtain a highly reliable impact strength value.

  Here, when a silicon wafer is evaluated as a semiconductor wafer, since the hardness of the silicon (Mohs hardness 7) is high, it is expected that the hitting material, the pressing means, etc. will be damaged by the broken silicon pieces. In addition, as a dropping means for dropping the striking substance from any desired height, holding by an electromagnetic magnet is mechanically easy. From the viewpoint of holding these striking substances and exchanging them at the time of deterioration, the material of the striking substances is preferably chromium steel.

In the present invention, the device structure and material are not limited to the above, and a device structure capable of measuring various special conditions can be obtained.
For example, as shown in FIG. 3 which is an enlarged view of a part of another example of the impacting substance and its dropping means of the semiconductor wafer evaluation apparatus of the present invention, the dropping means 15 ′ includes a stopper 12 and a bell holding pipe 13. Thus, the falling weight holding pipe 13 corresponding to the height adjusting rail is made of polycarbonate, and a number of falling height adjusting holes 14 are opened. The bullet-shaped falling weight 5 ′ can be held at a predetermined height by the stopper 12.
If the falling weight holding pipe 13 is made of polycarbonate, there is no height restriction unlike the single axis slider system, and the height of the falling weight 5 'can be as high as 10 m or more. And it is also possible to make the silicon piece W and the falling weight 5 'collide at a specific angle by bending and holding the tip or the bevel.
The bullet-shaped drop weight 5 'can be made of a material that is harder than silicon, such as fine ceramics, but cannot be held by an electromagnetic magnet, and is evaluated by attaching a strain gauge to the drop weight. Thus, it is possible to evaluate the characteristics of more detailed fracture behavior.

Next, an example of the semiconductor wafer evaluation method of the present invention (method of evaluating the impact strength of the edge portion of the semiconductor wafer) using the semiconductor wafer evaluation apparatus of the present invention as described above is shown below. The invention is not limited to these examples.
Hereinafter, a silicon wafer is used as a semiconductor wafer, and the case where the measurement is performed by changing the height of the impact material will be described as an example, but of course the semiconductor wafer is not limited to this, and the bonded semiconductor wafer or semiconductor device is not limited to this. The impact strength of various semiconductor wafers such as mounted quartz substrates, glass substrates, and compound semiconductor substrates can be evaluated, and measurement can be performed not only by the height of the impact material but also by changing the weight of the impact material. Needless to say, you can.

First, a silicon wafer as a base of the sample, a hitting substance to be dropped, and a dropping means for dropping the hitting substance from an arbitrary predetermined height are prepared.
Here, a sample can be manufactured by dividing one semiconductor wafer. For example, five (100) wafers with a diameter of 300 mm, a conductivity type of P ^- type, a resistivity of 10 Ω · cm, an oxygen concentration of 12 ppma and a thickness of 0.78 mm are prepared, and 20 samples are prepared by dividing them into four. can do.

By preparing a sample by dividing one semiconductor wafer, a large number of samples can be manufactured from one semiconductor wafer. Therefore, waste of the wafer is hardly generated and evaluation can be performed at a low cost. And since the influence of the subtle difference of impact strength between each semiconductor wafer which generate | occur | produces when using a different semiconductor wafer can be reduced, it can be set as more highly accurate impact strength evaluation.
The dividing method may be any form or size such as wafer 4 division, wafer 12 division, wafer 24 division, and the like, and is not particularly limited.

Then, one of the previously prepared samples is selected, and the sample is supported by the pressing means.
In addition, basic conditions are determined from the material and thickness of the sample. For example, the striking substance (falling bell) has a cylindrical shape, a diameter of 8 mm, a length of 20 mm, a weight of 8.6 g, the striking substance (falling bell) starts at a height of 100 cm, and the height change level is 10 cm. be able to.

Here, as the step (1), the sample supported by the pressing means is set on the apparatus base, the striking substance held by the electromagnetic magnet is dropped, and an impact force is applied to the edge portion of the sample.
And the presence or absence of a cracking chip of a sample is checked. Regardless of whether the sample is cracked or missing, the sample is replaced with a new sample by dropping the impact material only once. The same applies to the second and subsequent samples.

  Next, as step (2), as a second experiment, a sample different from step (1) is supported by the pressing means, and when the edge portion of the sample is not broken in step (1), step (1) If the impact material is dropped from a higher position (for example, a position that is one level higher) and the edge portion of the sample is destroyed in the step (1), the lower position (for example, one level lower) than the step (1) The impact material is dropped from the position).

Then, as step (3), as a third experiment, a sample different from steps (1) and (2) is supported by the pressing means, and the edge portion of the sample is not broken in step (2) (2) The impact material is dropped from a position higher than the step (for example, a position raised by one level), and when the edge portion of the sample is destroyed in the step (2), the position lower than the step (2) ( For example, the striking substance is dropped from a position lowered by one level.
The determination of whether or not the sample is broken can be made by observing with the naked eye the presence or absence of cracks, breakage or crushing on the surface of the impact portion of the sample, but of course it is not limited to this.

And as a (4) process, the said process is repeated until it reaches predetermined times n. For the evaluation of one sample type, it is desirable that n ≧ 20, that is, the above process is repeated 20 times or more.
The experimental results at this time are shown in Table 1 below. In the table, ◯ indicates the case where the edge portion of the sample is not broken, and x indicates the case where the edge portion is broken.

Further, as the step (5), the relationship between the number of samples in which the edge portion was destroyed and the falling height of the impacting material at that time, the number of samples in which the edge portion was not destroyed and the falling height of the impacting material at that time From the above, the 50% impact fracture height (H ₅₀ ) and the 50% impact fracture height standard deviation (SH) are calculated. The 50% impact fracture height (H ₅₀ ) is estimated to be the height at which 50% of the number of tests cause fracture, and the standard deviation (SH) of the 50% impact fracture height is the standard deviation.
For the calculation of the 50% impact fracture height (H ₅₀ ) and the standard deviation (SH) of the 50% impact fracture height, it is preferable to use the calculation of the staircase method.

For example, by arranging the number of samples whose edge was destroyed and the height of the impacting material, the number of samples whose edge was not destroyed and the height of the impacting material at that time, In the presence / absence, the smaller number is selected, and the number measured for each drop height is defined as f _n .
The standard deviation (SH) of the 50% impact fracture height (H ₅₀ ) and the 50% impact fracture height is (a) H ₅₀ = H ₀ + (height displacement) × (A / C ± 1 / 2)
(B) SH = 1.62 × (displacement of height) × ((CB−A ² ) / C ² +0.029)
It can be determined by the above formula.
However, H ₀ is the height value of the impact material with respect to n = 0, and with respect to the sign of ± in the equation (a), when the number of samples whose edge portions are destroyed is fn, the edge portions are destroyed. When the number of missing samples is fn, +, A = Σnf _n , B = Σn ² f _n , and C = Σf _n .

At this time, when H ₅₀ and SH cannot satisfy the following conditions 1 and 2, it is desirable to perform the steps (1) to (5) again.
Condition 1: H ₅₀ −SH <reference height at start <H ₅₀ + SH
Condition 2: 0.5 × SH <height change level <2 × SH

In step (6), the 50% impact fracture energy (E ₅₀ ) and the standard deviation (SE) of the 50% impact fracture energy are calculated from the previously calculated H ₅₀ and SH.
50% impact fracture energy (E ₅₀ ) is, for example, E ₅₀ = mass of impact material × gravity acceleration × 50% impact fracture height (H ₅₀ )
It can be calculated from
The standard deviation (SE) of 50% impact fracture energy can be calculated from, for example, the standard deviation obtained by multiplying the height of each striking substance by the mass of the striking substance and the gravitational acceleration.

As described above, the impact resistance against the dynamic stress of the semiconductor wafer can be evaluated by dropping the impact material from a predetermined height as in the step (1). The present invention can be a test in an environment closer to an actual semiconductor device process than a general material test such as a bending fracture test.
Also, as shown in step (4), by repeating steps (2)-(3) until a predetermined number of times n, the room for accidental trouble can be reduced compared to a single test. It can be evaluated with high accuracy. Furthermore, since the impact strength is changed in accordance with the immediately preceding test result, it is possible to reduce the room for measurement errors due to defects in the wafer itself, which contributes to higher accuracy.
Then, as in steps (5) to (6), H ₅₀ and SH are calculated, and E ₅₀ and SE are calculated to quantitatively evaluate the impact strength of the semiconductor wafer to the extent that it can be evaluated with other semiconductor wafers. be able to.
For these reasons, the impact strength of the semiconductor wafer can be evaluated stably and accurately.

  Of course, the whole wafer may be used without dividing the wafer in the stage of preparing the sample. An example is shown in FIG.

The holding means 7b in FIG. 4 has a holding jig structure capable of measuring various special conditions.
Silicon is a cubic structure with a diamond structure, and there is anisotropy in physical properties such as Young's modulus. In the pressing means 7b in FIG. 4, it is possible to evaluate the impact strength at a specific crystal plane using a whole wafer.
Further, the notch 10b in FIG. 4 has a large semi-elliptical shape. As a result, the holding area of the silicon wafer W0 is reduced and the deformation area is widened.

At this time, it is necessary to shift the notch 9 from the falling position of the impacting material 5 by θ in the figure so that the impacting material 5 does not collide with the notch 9 of the silicon wafer W0. The value of θ at this time can be arbitrarily selected in the range of 0 to 360 ° counterclockwise from the position of the notch 9.
For example, in the case where a crystal orientation (100) wafer has a notch (011), the value of θ is preferably set to 45 °, which is a standard direction. However, anisotropic crystals such as silicon have different measured values of impact fracture strength depending on the value of θ even on the same wafer, so it is necessary to arbitrarily select the optimum angle θ for measurement.

A silicon wafer having a diameter of 300 mm is as thin as 0.78 mm compared to the area. It may be bent greatly by its own weight. The pressing means 7b and the notch 10b in FIG. 4 are used when evaluating the macro fracture strength of a silicon wafer when such a phenomenon such as elastic deformation or buckling is observed.
As described above, according to the present invention, it is possible to evaluate various subtle fracture behaviors and characteristics by changing the shape of the notch and the hitting substance in the sample holding method.

In addition, the steps (1) to (6) can be performed on other types of semiconductor wafers to compare the impact strengths of the semiconductor wafers.
If the impact strength evaluation method described above is used, the impact strength is evaluated by a statistical method, and therefore the impact strength can be evaluated accurately and quantitatively.
Therefore, it is very suitable for comparing the impact strength between semiconductor wafers, and is suitable for comparing the impact fracture strength of various silicon wafers.

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
Experiments were performed using a semiconductor wafer evaluation apparatus as shown in FIG. As target wafers, 10 silicon wafers having a diameter of 300 mm, a conductivity type of P ^- type, a resistivity of 10 Ω · cm, an oxygen concentration of 12 ppma, a thickness of 0.78 mm, and a crystal orientation (100) were prepared. Each of the two sheets was heat treated as it was. That is, a comparative evaluation of “20 pieces of silicon not subjected to heat treatment and 20 pieces of silicon subjected to heat treatment” was performed.
At this time, the heat treatment was performed under the conditions of (1000 ° C., 16 hr) after (800 ° C., 2 hr).
In addition, the impact material at this time is made of chromium steel, heat-treated product, the shape is a cylinder, the diameter is 8 mm, the length is 20 mm, the weight is 8.6 g, the height at the start of the impact material is 100 cm, the height The change level was 10 cm.

At this time, BMD defects could not be detected with a 90-degree scattering LST (MO441) BMD defect measuring machine for 20 silicon pieces that were not heat-treated.
The 50% impact fracture energy and its standard deviation were as follows.
50% impact fracture energy (E50) = 0.082J
Standard deviation of 50% impact fracture energy (SE) = 0.014J

In addition, for 20 silicon pieces subjected to heat treatment, BMD defects of 1.0 × 10 ⁹ / cm 3 were detected by 90-degree scattering LST (MO441) measurement.
At this time, the 50% impact fracture energy and its standard deviation were as follows.
50% impact fracture energy (E50) = 0.101J
Standard deviation of 50% impact fracture energy (SE) = 0.009J

When a population average test (significance level 0.05) was performed from the distribution of fracture strength of the silicon pieces not subjected to heat treatment and the silicon pieces subjected to heat treatment, a significant difference was found in the population average.
Further, it was found that the silicon wafer used in the examples had many BMD defects due to heat treatment, and the 50% impact fracture energy (E50) was increased by about 23%.
As described above, the evaluation apparatus of the present invention makes it possible to statistically evaluate the impact strength of the edge portion of the semiconductor wafer. As a result, the effect of improving the evaluation ability was obtained as compared with the conventional inspection machine.

(Comparative Example 1)
The impact strength of the silicon wafer under the same conditions as in Example 1 was evaluated using the mechanical strength measuring apparatus described in Patent Document 1 as shown in FIG. In addition, 20 silicon wafers to be heat-treated and 20 silicon wafers to be heat-treated were prepared as samples.

At this time, with respect to 20 silicon wafers that were not heat-treated, the 90-degree scattering LST (MO441) BMD defect measuring machine could not detect BMD defects.
In addition, the mechanical strength of the edge part of a silicon wafer, its average value, and its standard deviation were as follows.
Mechanical strength: 45Kg-85Kg
Average mechanical strength = 55.1 kg
Standard deviation of mechanical strength = 9.1Kg

Further, about 20 silicon wafers subjected to heat treatment, BMD defects of 1.0 × 10 ⁹ / cm 3 were detected by 90-degree scattering LST (MO441) measurement.
At this time, the mechanical strength, average value, and standard deviation of the edge portion of the silicon wafer were as follows.
Mechanical strength: 46Kg ~ 83Kg
Average mechanical strength = 54.8 Kg
Standard deviation of mechanical strength = 10.1 Kg

In the measurement results of Comparative Example 1, no difference in fracture strength due to the presence or absence of heat treatment could be confirmed. The silicon wafer breaking strength in this apparatus is very high, and it is difficult to consider that such a high load is applied in the semiconductor device manufacturing process. In addition, the broken wafer had a very finely crushed shape, and a form different from a crack at the time of device manufacture was seen.
These causes are presumed to be that when a static load is applied to the wafer, a large internal stress accumulates due to the deformation of the wafer such as buckling, and the wafer is shattered at the moment of occurrence of the breakdown.

  The measurement results of Comparative Example 1 and the results of Example 1 of the present invention cannot be simply compared because the units are different. However, in Example 1 of the present invention, it is considered that the detection accuracy is higher than that of the measurement apparatus of Comparative Example 1 in that the difference in fracture strength due to the presence or absence of heat treatment can be evaluated.

(Comparative Example 2)
The evaluation apparatuses described in Patent Documents 2 to 4 cannot be simply compared with the results of the present invention because the wafer holding method and the evaluation means are different.
However, in any of the above evaluation apparatuses, it is difficult to vary the impact strength given to the wafer at one time. For this reason, the breaking strength = the number of repeated hits is a criterion for evaluation.
Therefore, by using the semiconductor wafer evaluation apparatus of the present invention, whichever of the “evaluation of fracture strength with variable impact strength (Example 1)” and “evaluate the fracture strength based on the number of repeated impacts (Comparative Example 2)” It was compared whether it was excellent as an analysis method.

Specifically, first, a sample which is a wafer piece is prepared by the same method as in the first embodiment, and is held on the apparatus base. The heat treatment for the target wafer and the wafer piece is under the same conditions as in the first embodiment.
Thereafter, a bell is prepared under the same conditions as in Example 1 and dropped toward the edge of the sample. Whether or not the sample has been destroyed is determined by the naked eye with respect to the presence or absence of cracks, breaks or fractures on the surface of the impact site of the sample.

At this time, if the edge portion of the sample is not broken, the bell is dropped again from the same height (100 cm) as before. And it keeps hitting repeatedly many times until destruction of a sample occurs.
Then, the number of repeated hits until the edge portion of the sample is broken is recorded, and the impact test is repeatedly performed on each of the 20 samples that were not heat-treated and the heat-treated sample.
From the number of repeated hits obtained as a result of the above experiment, the number of hits Σ (total number of hits × total number of 20 shots) and the average number of hits until breaking are calculated. The results at this time are shown in Table 2 below.

At this time, about 20 pieces of silicon that were not heat-treated, 11 pieces were destroyed at the first hit, three pieces were destroyed at the second hit, and the last one was destroyed at the tenth hit.
The results were 54 Σ hits and an average hit of 2.7 hits / break.

Further, about 20 silicon pieces subjected to the heat treatment, 8 pieces were destroyed at the first hit and the last one was destroyed at the seventh hit.
The result was an average number of hits of 2.6 hits / sheet until the number of Σ hits was 52.
From the above results, the analysis method of “Evaluation of Fracture Strength by Number of Repeated Blows (Comparative Example 2)” could not find a difference in fracture strength between varieties depending on the presence or absence of heat treatment.

In the case of Comparative Example 2, one sample specimen that was not easily broken greatly influenced the measurement result. When a sample specimen that was not easily cracked was observed closely, a micro dent of a size that could not be confirmed with the naked eye was generated by the blow, and the puncture was expanded, leading to destruction.
In this method, it is necessary to repeatedly hit the same point many times with the right strength for damage. Adjustments are difficult because the impact strength of the right strength varies depending on the variety.
Further, in order to reduce the influence of the presence or absence of a sample specimen that is not easily destroyed, it is necessary to increase the number of evaluation sheets. There is also a problem that the evaluation time is very long. Moreover, since the occurrence of micro dents is likely to vary, it is difficult to perform reproducible measurement.
Therefore, it is considered that the evaluation method of Comparative Example 2 has a risk of overestimating the influence of repeated hits, considering that the silicon wafer is a brittle material.

On the other hand, the “destructive strength evaluation method with variable impact strength (Example 1)” of the present invention is replaced with a new sample regardless of whether or not the sample is destroyed in a single impact. This eliminates the effect of micro dents on fracture strength.
In Example 1 of the present invention, a slight difference in impact resistance due to the presence or absence of heat treatment can be statistically evaluated. Therefore, it is expected that detection sensitivity / accuracy is higher than that of a conventional strength measuring apparatus.

  The present invention is not limited to the above embodiment. The above-described embodiment is an exemplification, and the present invention has any configuration that has substantially the same configuration as the technical idea described in the claims of the present invention and that exhibits the same effects. Are included in the technical scope.

DESCRIPTION OF SYMBOLS 1 ... Semiconductor wafer evaluation apparatus, 2 ... Height adjustment rail, 3 ... Slider,
4 ... electromagnetic magnet 5, 5 '... striking substance (bell), 6 ... polycarbonate cover,
7a, 7b ... holding means, 8 ... device base, 9 ... notch,
10a, 10b ... notches, 12 ... stoppers,
13 ... Drop weight holding pipe, 14 ... Drop height adjustment hole, 15, 15 '... Drop means,
DESCRIPTION OF SYMBOLS 101 ... Semiconductor wafer, 102 ... Mounting stand, 103 ... Support means,
104 ... Load shaft, 105a ... Air cylinder, 105b ... Pressure control valve,
106 ... cylinder, A, B ... the point where the wafer is supported,
C: Point where the tip of the load shaft is pressed, O: Wafer center,
W ... sample, W0 ... silicon wafer.

Claims (3)

A method for evaluating the impact strength of a semiconductor wafer, comprising performing at least the following steps (1) to (6).
(1) A sample which is a semiconductor wafer is supported by pressing means for supporting the wafer so that the radial direction of the wafer is equal to the falling direction of the striking substance, and the striking substance is separated from the sample by a predetermined weight and height. It is dropped on the edge and an impact is applied to a specific crystal plane.
(2) As a second experiment, a sample different from the sample in step (1) is supported by the pressing means, and if the sample is not broken in step (1), the position is higher than in step (1). the dropped the striking material from high comb, (1) if the sample has been destroyed in the process, (1) the specific crystal plane location to drop the said striking material from low comb than step Shock.
(3) As a third experiment, when a sample different from the samples in steps (1) and (2) is supported by the pressing means, and (2) the sample is not destroyed in step (2) position to drop the blow material from high comb than step (2) if the sample has been destroyed in the process, the by dropping the striking material from low comb position than (2) step A shock is applied to a specific crystal plane.
(4) The above process is repeated until the predetermined number n is reached.
(5) the broken number and height of the striking material at the time of the sample, the number of broken not sample from the height relationship of the striking material at that time, 50% impact fracture height ( H ₅₀ ), the standard deviation (SH) of the 50% impact fracture height is calculated.
(6) The standard deviation (SE) of 50% impact fracture energy (E ₅₀ ) and 50% impact fracture energy is calculated from the H ₅₀ and the SH.   The method for evaluating a semiconductor wafer according to claim 1, wherein the sample in each step is prepared by dividing one semiconductor wafer. The steps (1) to (6) are also performed on a semiconductor wafer of a type different from the sample in each step, and the impact strength of the semiconductor wafer is compared based on the results. The evaluation method of the semiconductor wafer of Claim 1 or Claim 2.

Images