JP7218733B2 - Oxygen concentration evaluation method for silicon samples - Google Patents

Description

The present invention relates to a method for evaluating the oxygen concentration of a silicon sample.

Silicon wafers used as substrates for power devices and CIS (CMOS image sensors) are required to have a long lifetime, and low oxygen content is required to avoid shortening the lifetime. For example, in power devices, FZ silicon wafers manufactured by the floating zone (FZ) method, which is capable of lowering oxygen than the Czochralski (CZ) method, are often used. Further, as a CIS substrate, an epitaxial wafer is often used, which is obtained by depositing a low-oxygen epitaxial layer on a CZ silicon wafer grown by the CZ method. Accurately measuring the oxygen concentration of such low-oxygen wafers is very important.

Conventional oxygen concentration measurement methods include the FT-IR method (Fourier transform infrared spectroscopy) and the SIMS method (secondary ion mass spectrometry). gets worse. The detection lower limit is 0.07 (ppma-JEITA) for the FT-IR method and 0.02 (ppma) for the SIMS method.

Here, as a method for measuring the impurity concentration in a silicon wafer with high sensitivity, there is a DLTS method (transient capacitance spectroscopy). For example, in Patent Document 1, Ec is the level at the lower end of the conduction band formed on a silicon wafer by electron beam irradiation, and the level density of Ec-0.42 eV is detected by the DLTS method, and this level density is used as an index. A carbon concentration measuring method is disclosed.

JP 2018-139242 A

P. Pellegrino et al, Physical Review B, Vol 64, 195211

As described above, there are FT-IR method and SIMS method for measuring oxygen concentration in silicon wafers. The limit is 0.02 (ppma) in the method, and the detection sensitivity deteriorates at low oxygen concentrations. Therefore, methods for sensitively evaluating the oxygen concentration below the lower limit in the FT-IR method and SIMS method have been investigated. Further, Patent Document 1 discloses a carbon concentration measuring method in which an impurity level formed by electron beam irradiation is detected by the DLTS method and the level density is used as an index, but an oxygen concentration measuring method is disclosed. It has not been.

The present invention has been made to solve the above problems, and the oxygen concentration in the silicon wafer is 0.5 ppma or less, which is a low oxygen concentration, particularly below the lower limit of detection in the FT-IR method or SIMS method. An object of the present invention is to provide a method for evaluating oxygen concentration with high sensitivity.

The present invention has been made to achieve the above objects, and is a method for evaluating the oxygen concentration in a silicon sample having an oxygen concentration of 0.5 ppma or less, wherein the silicon sample is irradiated with electron beams or ions other than oxygen. Defects containing vacancies are formed in the silicon sample by irradiating a beam, the level density of the defects containing the vacancies is measured by the DLTS method, and the level density of the defects containing the vacancies is measured. A silicon sample oxygen concentration evaluation method for evaluating the oxygen concentration in the silicon sample is provided based on.

According to such a method for evaluating the oxygen concentration of a silicon sample, it is possible to provide a method for evaluating the oxygen concentration with high sensitivity when the oxygen concentration in the silicon wafer is 0.5 ppma or less, particularly 0.02 ppma or less. can.

At this time, a silicon sample with a known oxygen concentration is prepared in advance, and the silicon sample with the known oxygen concentration is irradiated with an electron beam or an ion beam other than oxygen so that the silicon sample with the known oxygen concentration contains vacancies. A defect is formed, the level density of the defect containing the vacancy is measured by the DLTS method, and a calibration curve is created between the measured level density and the oxygen concentration of a silicon sample whose oxygen concentration is known. and the oxygen concentration in the silicon sample is preferably evaluated based on the calibration curve.

By creating the calibration curve in this way, the absolute concentration of oxygen in the silicon sample can be calculated.

At this time, the oxygen concentration of the silicon sample whose oxygen concentration is known is preferably the oxygen concentration obtained by the SIMS method or the FT-IR method.

As described above, the SIMS method or the FT-IR method can accurately measure the oxygen concentration of the silicon sample if the oxygen concentration is equal to or higher than the lower limit of detection, and is therefore suitable for creating a calibration curve.

At this time, the level density of the defects including the vacancies is preferably Ec-0.17 eV, where Ec is the lower end level of the conduction band.

By using such a level as an index, the oxygen concentration in the silicon sample can be evaluated with higher accuracy.

At this time, the level density of defects including vacancies is preferably Ec-0.42 eV, where Ec is the lower end level of the conduction band.

By using such a level as an index, it is possible to more accurately evaluate the oxygen concentration in a silicon sample having a particularly low oxygen concentration.

At this time, the level density of the defects including the vacancies is obtained by measuring the depth distribution of the level density of Ec-0.17 eV by the DLTS method, where Ec is the level at the lower end of the conduction band. It is preferable to evaluate the oxygen concentration depth distribution of the silicon sample from the level density depth distribution of −0.17 eV.

By doing so, it is possible to acquire the depth distribution of the oxygen concentration of several micrometers in the surface layer of the silicon sample.

At this time, the level density of a silicon sample with a known oxygen concentration is measured, a calibration curve is created between the measured level density and the oxygen concentration of the silicon sample with a known oxygen concentration, and based on the calibration curve It is preferable to evaluate the depth distribution of the oxygen concentration using

By creating a calibration curve in this way, it is possible to calculate the depth distribution of the oxygen concentration in the surface layer of several micrometers of the silicon sample.

As described above, according to the method for evaluating the oxygen concentration of a silicon sample of the present invention, it is possible to evaluate with high sensitivity an oxygen concentration as low as 0.5 ppma or less, particularly as low as the lower limit of detection in the FT-IR method or the SIMS method. can. Therefore, not only the bulk of the silicon sample but also the surface oxygen concentration can be evaluated. Also, by creating a calibration curve, the absolute concentration of oxygen in the silicon sample can be calculated.

It is the figure which showed an example of the oxygen concentration evaluation method of the silicon sample of this invention. FIG. 4 is a diagram showing DLTS spectra after irradiating electron beams to silicon samples having oxygen concentrations of 13.5 ppma and 0.042 ppma, respectively. FIG. 4 is a diagram showing the relationship between the oxygen concentration measured by the SIMS method and the Ec-0.17 eV and Ec-0.42 eV level densities of the present invention. FIG. 4 is a diagram showing a depth distribution of oxygen concentration measured by the SIMS method and a depth distribution of oxygen concentration obtained by converting the depth distribution of the Ec-0.17 eV level density of the present invention using a calibration curve; . 3 is a diagram showing the depth distribution of oxygen concentration calculated from the depth distribution of the level density of Ec-0.17 eV in Example 2 and the depth distribution of oxygen concentration measured by the SIMS method in Comparative Example 3. FIG. .

The present invention will be described in detail below, but the present invention is not limited to these.

As described above, there has been a demand for a method capable of highly sensitively evaluating low oxygen concentrations of 0.5 ppma or less, particularly below the lower limit of detection in the FT-IR method or SIMS method.

The present inventor has come up with the idea that the impurity level formed by a particle beam such as an electron beam strongly depends on impurities (for example, oxygen and carbon) existing in a silicon sample. That is, considering that there is an impurity level whose density changes depending on the oxygen concentration in the silicon sample, the authors have earnestly investigated the relationship between the impurity level density formed by the particle beam and the oxygen concentration.

As a result, the present inventors found that among the impurity levels in the bandgap of silicon measured using the DLTS method, an impurity level containing vacancies whose level density increases as the oxygen concentration increases, and We found that there are impurity levels containing vacancies whose level density decreases as the oxygen concentration increases.

Based on the above, the present inventors have extensively studied the above problems, and as a result, have found a method for evaluating the oxygen concentration in a silicon sample having an oxygen concentration of 0.5 ppma or less, wherein the silicon sample is irradiated with an electron beam or other than oxygen. to form defects containing vacancies in the silicon sample, measuring the level density of the defects containing the vacancies by the DLTS method, and measuring the level density of the defects containing the vacancies According to the silicon sample oxygen concentration evaluation method for evaluating the oxygen concentration in the silicon sample based on the potential density, the oxygen concentration in the silicon wafer is 0.5 ppma or less, especially the detection limit value or less in the FT-IR method or SIMS method. The inventors have found that a method for evaluating the oxygen concentration with high sensitivity can be provided in the case of , and completed the present invention.

Description will be made below with reference to the drawings.

FIG. 1 is a diagram showing an example of the oxygen concentration evaluation method of a silicon sample according to the present invention.

As in S1 of FIG. 1, a silicon sample whose oxygen concentration is to be evaluated is prepared. For example, as long as it is a silicon sample cut out from a silicon single crystal ingot, the shape is not limited, and a silicon chip with a size of several centimeters may be used. Specific examples include polished wafers, epitaxial wafers, and annealed wafers.

Next, as in S2, the silicon sample whose oxygen concentration is to be evaluated is irradiated with an electron beam or an ion beam other than oxygen. Thereby, defects including vacancies correlated with oxygen can be formed in the silicon sample. The electron beam irradiation dose is, for example, in the range of 5.0×10 ¹³ to 5.0×10 ¹⁵ (electrons/cm ² ), and the ion beam dose is, for example, 1.0×10 ¹¹ to 1.0×. It can be in the range of 10 ¹³ (atoms/cm ² ). By setting both the electron beam dose and the ion beam dose within this range, it is possible to more effectively prevent problems such as failure to form the desired impurity levels due to too small dose and dose. In addition, it is possible to more effectively prevent the silicon lattice from being disturbed, the crystallinity being lowered, and the target impurity level not being detected due to excessive irradiation or dose. In the case of an oxygen ion beam, the irradiated oxygen ions affect the formation of the impurity level, which affects the oxygen concentration in the silicon sample to be evaluated. Therefore, it is necessary to use an ion beam other than oxygen. .

Next, as in S3, the level density of defects including vacancies is measured using the DLTS method. Here, the details of how the present invention was discovered will be described. In particular, a method for measuring the level density of defects containing vacancies formed by irradiation with an electron beam or an ion beam and the reason for the measurement will be described in detail below.

First, the DLTS method used in the present invention will be described. In the DLTS method, a metal electrode forming a Schottky junction is normally formed on the front surface of the wafer, and a metal electrode having an ohmic junction is formed on the back surface of the wafer. When a reverse bias is applied between the two electrodes and the temperature dependence of the capacitance change in the resulting depletion layer is obtained, the capacitance change peaks at a temperature corresponding to the energy level formed by the impurities. . The impurity level density can be calculated from the capacitance change at the peak position.

Next, impurity levels containing vacancies formed by electron beam or ion beam irradiation will be described. When a silicon sample is subjected to ion implantation or electron beam irradiation, silicon at lattice positions is ejected, and vacancies and interstitial silicon are generated. Among them, it is known that an impurity level containing vacancies is formed. For example, in Non-Patent Document 1, Ec-0.17 eV and the impurity level of the vacancy oxygen complex, Ec-0.42 eV and the vacancy vacancy complex or the vacancy phosphorus complex, or the level of both There are reports on specific levels, impurities, and vacancies, such as mixed impurity levels, Ec-0.32 eV and impurity levels of vacancy oxygen-hydrogen complexes.

As an example, FIG. 2 is a diagram showing DLTS spectra after irradiating a silicon sample with an oxygen concentration of 13.5 ppma and a silicon sample with an oxygen concentration of 0.042 ppma with an electron beam. The oxygen concentration was measured by the SIMS method, and electron beam irradiation was performed at an acceleration voltage of 2 (MV) and a dose of 2.5×10 ¹⁴ (electrons/cm ² ).

It can be seen that the Ec-0.17 eV level density is higher in the sample with the oxygen concentration of 13.5 ppma, and conversely, the Ec-0.42 eV level density is higher in the oxygen concentration of 0.042 ppma. The Ec-0.17 eV level is related to the vacancy-oxygen complex and is an impurity level related to oxygen. Therefore, the Ec-0.17 eV level is formed more easily in the 13.5 ppma sample with a higher oxygen concentration, resulting in a higher Ec-0.17 eV level density.

The level of Ec-0.42 eV is said to be a level of a vacancy-vacancy complex, a vacancy-phosphorus complex, or a mixture of both levels. rank. When the oxygen concentration is low, the Ec-0.17 eV level is less likely to be formed, and as a result, the Ec-0.42 eV level, which is not related to oxygen, is more likely to be formed. For the above reasons, the sample with a low oxygen concentration of 0.042 ppma has a higher Ec-0.42 eV level density.

Thus, the impurity level of defects including vacancies generated by electron beam irradiation reflects the oxygen concentration in the silicon sample. Therefore, as in S4, the oxygen concentration in the silicon sample can be evaluated by using the measured impurity level as an index. However, the higher the irradiation dose, the higher the impurity level density generated by electron beam irradiation. Therefore, when evaluating the oxygen concentration, it is necessary to evaluate these impurity level densities at the same irradiation dose.

In FIG. 2, an Ec-0.32 eV level is also detected as another impurity level. Ec-0.32 eV is an impurity level related to a vacancy oxygen-hydrogen complex, and since it is an impurity level in which oxygen is involved, it can be used as an index. The level density is low compared to the level density. Furthermore, when the inventor investigated a plurality of silicon samples, the Ec-0.17 eV level had a sharper peak and a higher level density than the Ec-0.32 eV level at any level. Therefore, it is preferable to use the level of Ec-0.17 eV as an index of the impurity level related to oxygen.

Next, the range of oxygen concentration that can be evaluated in the present invention will be explained. FIG. 3 is a diagram showing the relationship between the oxygen concentration measured by the SIMS method and the Ec-0.17 eV and Ec-0.42 eV level densities according to the present invention. For the level densities of Ec-0.17 eV and Ec-0.42 eV, the oxygen concentrations after electron beam irradiation treatment at an acceleration voltage of 2 (MV) and a dose of 2.5×10 ¹⁴ (electrons/cm ² ) are different. A silicon sample was measured by the DLTS method. In the oxygen concentration range of 0.024 ppma to 0.36 ppma measured by the SIMS method, the higher the oxygen concentration, the higher the Ec-0.17 eV level density, and conversely, the Ec-0.42 eV level density is showed a tendency to decrease. On the other hand, in the range where the oxygen concentration measured by the SIMS method is 0.6 ppma or more, the level density of both Ec-0.17 eV and Ec-0.42 eV does not depend on the oxygen concentration, and the level density is almost constant Met. This is because when the oxygen concentration measured by the SIMS method is in the range of 0.6 ppma or more, the oxygen in the sample is excessive, and most of the vacancies generated by the electron beam are consumed to form the Ec-0.17 eV level. It is considered that the reason for this is that the level density of Ec-0.17 eV cannot be detected as a significant difference in the measurement by the DLTS method. From the above, it can be seen that the oxygen concentration measurement range of the present invention must be 0.5 ppma or less. On the other hand, even at 0.02 ppma or less, which was the limit of the SIMS method, it is possible to measure even lower oxygen levels by extrapolating the calibration curve of the present invention.

Next, the oxygen concentration evaluation method using the impurity level in the step S4 as an index will be described. When the level of Ec-0.17 eV is used as an index, the lower the level density, the lower the oxygen concentration can be determined. Further, when the level of Ec-0.42 eV is used as an index, the higher the level density, the lower the oxygen concentration can be determined. Therefore, even without using a calibration curve, which will be described later, it is possible to evaluate the magnitude relationship in oxygen concentration between samples at a plurality of levels.

As described above, according to the present invention, the oxygen concentration can be evaluated with high sensitivity when the oxygen concentration in the silicon wafer is as low as 0.5 ppma or less.

Impurity level densities detected by the DLTS method cannot be compared with the oxygen concentrations of conventional SIMS and FT-IR methods, but absolute concentrations can be calculated by preparing a calibration curve. A silicon sample having a known oxygen concentration is prepared in advance in order to create a calibration curve. The oxygen concentration of a silicon sample with a known oxygen concentration is preferably measured by the SIMS method or the FT-IR method. A calibration curve is created from the oxygen concentration of the silicon sample whose oxygen concentration is known and the level densities of Ec-0.17 eV and Ec-0.42 eV in the oxygen concentration range of 0.5 ppma or less. By using this calibration curve, the level density measured by the DLTS method can be converted into the oxygen concentration obtained by the SIMS method or the FT-IR method.

FIG. 3 shows a calibration curve between the level densities of Ec-0.17 eV and Ec-0.42 eV and the oxygen concentration of 0.5 ppma or less obtained by the SIMS method. It can be seen that both levels have a high correlation coefficient ^R2 . The level of Ec-0.17 eV has a higher value of R ² than the level of Ec-0.42 eV, and since Ec-0.17 eV is an impurity level related to oxygen, the oxygen concentration of the present invention As the first index of the level density of defects including vacancies used for evaluation, the level of Ec-0.17 eV is most desirable. When the Ec-0.17 eV level is not detected, the Ec-0.42 eV level can be used as the second index.

Since Ec-0.17 eV is the level of the vacancy oxygen complex, in the range of 0.5 ppma or less, the higher the oxygen concentration, the higher the level density, and a positive correlation with the oxygen concentration is observed. . Therefore, by using this level as an index, the oxygen concentration can be evaluated. Note that the level at the bottom of the conduction band is Ec.

Ec-0.42 eV is said to be a level of a vacancy-pore complex, a vacancy-phosphorus complex, or a mixture of both, and is not an impurity level related to oxygen. However, since this level density increases as the oxygen concentration decreases, it can be used as an indicator. This is because the lower the oxygen concentration, the less likely the above-described Ec-0.17 eV level involving oxygen is formed, and the more likely the Ec-0.42 eV level not involving oxygen is formed.

By the way, the depth measured by the DLTS method depends on the width of the depletion layer and can be controlled by the reverse voltage. In the normal DLTS method described above, various impurity levels can be measured by fixing the reverse voltage and measuring the capacitance by sweeping the temperature. On the other hand, it is also possible to obtain the depth distribution of the level density in the depletion layer by fixing the level (that is, measurement temperature) and changing the reverse voltage. Therefore, by obtaining the depth distribution of the level density of Ec-0.17 eV and using it as an index, it is possible to evaluate the depth distribution of the oxygen concentration in the silicon sample.

Furthermore, by using the calibration curve, the depth distribution of the Ec-0.17 eV level density measured by the DLTS method can be compared with the depth distribution of the oxygen concentration by the conventional FT-IR method and SIMS method. , the depth distribution of the absolute concentration of oxygen in the silicon sample can be calculated. FIG. 4 shows the oxygen concentration depth distribution measured by the SIMS method for the epitaxial wafer, and the oxygen concentration depth obtained by converting the Ec-0.17 eV level density depth distribution according to the present invention using a calibration curve. distribution and shows. It can be seen that the depth distribution of the oxygen concentration converted from the depth distribution of the level density of Ec-0.17 eV substantially matches the depth distribution of the oxygen concentration obtained by the SIMS method. In this way, the DLTS method can also acquire the depth distribution of the oxygen concentration in the surface layer of several micrometers.

Incidentally, since Ec-0.42 eV is a level related to impurities that do not contain oxygen, it cannot be applied to depth profile measurement.

As described above, the level densities of defects containing vacancies generated by irradiating an electron beam or an ion beam other than oxygen are used as indicators of the level densities of Ec-0.17 eV and Ec-0.42 eV. Thus, the oxygen concentration in the silicon sample can be measured with high sensitivity.

EXAMPLES The present invention will be described in detail below with reference to examples, but these are not intended to limit the present invention.

(Comparative example 1)
Four levels of n-type silicon samples with different oxygen concentrations, which were pulled by the CZ method or the FZ method, were prepared. Samples 1 to 3 are silicon wafers grown by the FZ method, and sample 4 is a silicon wafer grown by the CZ method. The oxygen concentration of the prepared silicon sample was measured by the SIMS method. Table 1 shows the results.

As shown in Table 1, the oxygen concentration was (low) sample 1<sample 2<sample 3<sample 4 (high). The oxygen concentration in Comparative Example 1 could not be detected because Sample 1 was below the detection limit of 0.02 ppma of the SIMS method, Sample 2 was 0.025 ppma, Sample 3 was 0.21 ppma, and Sample 4 was 3.2 ppma. rice field. Next, the evaluation method of the present invention was verified using the samples 1 to 4 (Example 1, Comparative Example 2).

(Example 1)
As Example 1, silicon samples of the same level as Samples 1 to 3 of Comparative Example 1 were prepared. Next, after electron beam irradiation treatment was performed at an acceleration voltage of 2 (MV) and an irradiation dose of 2.5×10 ¹⁴ (electrons/cm ² ), the oxide film on the surface of the silicon sample was removed with hydrofluoric acid. Gold was vapor-deposited on the surface of the sample as a Schottky electrode, and gallium was imprinted on the back surface as an ohmic electrode. Then, the level densities of Ec-0.17 eV and Ec-0.42 eV of the sample were evaluated by the DLTS method. In addition, the level density was converted to the oxygen concentration using the calibration curve of FIG. 3 prepared in advance. Table 2 shows the results.

As shown in Table 2, the oxygen concentration was (low) sample 1<sample 2<sample 3 (high). The oxygen concentrations converted from both Ec-0.17 eV and Ec-0.42 eV level densities were equivalent. However, since the level of Ec-0.17 eV is more preferable as the first index as described above, the oxygen concentration in Example 1 was calculated from the level density of Ec-0.17 eV. do. In this case, the oxygen concentration of Example 1 is 0.011 ppma for sample 1, 0.024 ppma for sample 2, and 0.23 ppma for sample 3 from the level density of each sample.

In Samples 2 and 3, the oxygen concentrations of Example 1 and Comparative Example 1 were equivalent, so the validity of the present invention was shown.

From Example 1, the oxygen concentration of sample 1 was found to be 0.011 ppma in the present invention, so it is possible to evaluate a trace amount of oxygen that cannot be detected by the conventional SIMS method by using the method according to the present invention. I understand.

(Comparative example 2)
Subsequently, as Comparative Example 2, a silicon sample of the same level as Sample 4 of Comparative Example 1 was prepared, and the silicon sample was treated in the same manner as in Example 1. Then, the level densities of Ec-0.17 eV and Ec-0.42 eV of the sample were evaluated by the DLTS method. In addition, the level density was converted to the oxygen concentration using the calibration curve of FIG. 3 prepared in advance. Table 2 shows the results.

Looking at the oxygen concentration of Sample 4 of Comparative Example 2, a large divergence was observed between Comparative Examples 1 and 2.
Since the oxygen concentration measured by the SIMS method was 3.2 ppma, the oxygen concentration of sample 4 was outside the applicable range of the present invention, so such a result was obtained.

(Example 2)
In order to verify the oxygen concentration distribution of Ec-0.17 eV, a silicon epitaxial wafer was prepared as a sample and subjected to the same electron beam irradiation treatment as in Example 1. The depth distribution of the level density of Ec-0.17 eV was measured by the DLTS method, and converted into the depth distribution of the oxygen concentration using the calibration curve of FIG. The results are shown in FIG. As can be seen from FIG. 5, a distribution was obtained in which the oxygen concentration was lower on the surface side of the sample. It is considered that this reflects the out-diffusion behavior of oxygen during epitaxial growth of the sample.

(Comparative Example 3)
Subsequently, in order to verify the results of Example 2, an epitaxial wafer of the same level as in Example 2 was used as a sample and evaluated by the SIMS method. The results are shown in FIG. According to FIG. 5, it can be seen that the oxygen concentration tends to be high on the bulk side of the sample and low on the surface side. In particular, the oxygen concentration in the region from the surface to 5 μm was below the detection limit of the SIMS method, and the oxygen concentration could not be measured accurately.

Here, when comparing the two, it seems that there is a divergence in the oxygen concentration, but this is because the background derived from the SIMS method apparatus is added. This background is dominant up to 5 μm from the surface, and the oxygen concentration is almost constant at 0.02 ppma. In order to eliminate this effect, a background component of 0.02 ppma was subtracted from the original profile.

In the SIMS method, the oxygen concentration is below the detection limit in the region 2 to 4 μm from the surface, so the depth distribution of oxygen concentration cannot be accurately measured. The behavior that the oxygen concentration decreases can be captured. As described above, since the DLTS method can measure the depth distribution of the level density in the surface layer of several μm, it is possible to obtain the oxygen concentration depth distribution in the surface layer of several μm by using the calibration curve.

As described above, according to the oxygen concentration evaluation method of the present invention, it was shown that a very small amount of oxygen concentration, which cannot be detected by the conventional SIMS method, can be evaluated.

It should be noted that the present invention is not limited to the above embodiments. The above-described embodiment is an example, and any device having substantially the same configuration as the technical idea described in the claims of the present invention and exhibiting the same effect is the present invention. included in the technical scope of

Claims (7)

A method for evaluating the oxygen concentration in a silicon sample having an oxygen concentration of 0.5 ppma or less,
irradiating the silicon sample with an electron beam or an ion beam other than oxygen to form defects containing vacancies in the silicon sample, and measuring the level density of the defects containing the vacancies by a DLTS method; A method for evaluating oxygen concentration in a silicon sample, wherein the oxygen concentration in the silicon sample is evaluated based on the level density of defects including the vacancies. A silicon sample with a known oxygen concentration is prepared in advance, and the silicon sample with the known oxygen concentration is irradiated with an electron beam or an ion beam other than oxygen to remove defects including vacancies in the silicon sample with the known oxygen concentration. Formed, the level density of defects containing the vacancies is measured by the DLTS method, a calibration curve is created between the measured level density and the oxygen concentration of a silicon sample with a known oxygen concentration, and the 2. The method for evaluating oxygen concentration in a silicon sample according to claim 1, wherein the oxygen concentration in the silicon sample is evaluated based on a calibration curve. 3. The method for evaluating oxygen concentration of a silicon sample according to claim 2, wherein the oxygen concentration of the silicon sample whose oxygen concentration is known is obtained by SIMS method or FT-IR method. 4. The level density of the defects including the vacancies is Ec-0.17 eV, where Ec is the level at the bottom of the conduction band. The method for evaluating the oxygen concentration of a silicon sample according to 1. 4. The level density of the defects including the vacancies is a density of Ec-0.42 eV, where Ec is the lower end level of the conduction band. The method for evaluating the oxygen concentration of a silicon sample according to 1. The level density of the defects including the vacancies is obtained by measuring the depth distribution of the level density of Ec-0.17 eV by the DLTS method, where Ec is the level at the lower end of the conduction band. 5. The method for evaluating oxygen concentration of a silicon sample according to claim 1, wherein the depth distribution of oxygen concentration of the silicon sample is evaluated from the depth distribution of the level density of 17 eV. measuring the level density of a silicon sample with a known oxygen concentration; creating a calibration curve between the measured level density and the oxygen concentration of the silicon sample with the known oxygen concentration; 7. The method for evaluating the oxygen concentration of a silicon sample according to claim 6, wherein a depth distribution of the concentration is evaluated.

Images