JP5842765B2 - Method for evaluating nitrogen concentration in silicon single crystals - Google Patents

Description

  The present invention relates to a method for evaluating a nitrogen concentration in a nitrogen-doped silicon single crystal grown by the Czochralski (CZ) method.

In silicon single crystal production, nitrogen may be doped into a single crystal for reasons such as controlling crystal defects or controlling oxygen precipitates called BMD. In FZ (floating zone) crystal and the like, the nitrogen concentration may be 10 to the 14th or 15th power, but in the CZ crystal, even if the nitrogen concentration is 1 × 10 ¹⁴ / cm ³ or less, it is sufficient. It has been reported variously that there is the effect.

As a method for measuring these nitrogen concentrations, secondary ion mass spectrometry (SIMS) is effective as a local analysis, but its detection sensitivity is 10 14 14 in the middle and 1 × 10 ¹⁴ / Nitrogen concentrations below cm ³ cannot be measured.

As a simpler and more sensitive method, Fourier transform infrared spectroscopy (FT-IR) or the like is used. This nitrogen concentration measurement method is well summarized in Non-Patent Document 1.
Nitrogen in silicon is supposed to take various forms such as NN pairs, NNO, and NNOO. In general, the absorption in the infrared region due to these various modes of vibration is measured by the FT-IR method. These forms have been reported to vary with temperature. Observe all these various absorption peaks to increase sensitivity, or remove background noise due to oxygen-derived donors (hereinafter referred to as oxygen donors) by heat treatment at about 650 ° C. as in Patent Document 1. By doing so, the detection sensitivity is improved.

  Further, in Patent Document 2, it is noted that nitrogen forms a nitrogen-oxygen donor (which is in the form of ONO and may be hereinafter referred to as “NO donor”). After erasing by heat treatment, NO donors are formed by heat treatment at 500 to 800 ° C., and the nitrogen concentration is obtained from the change in resistivity generated at that time.

  However, this method requires time and effort for heat treatment twice. Also, this technique only observes changes in resistivity and does not detect nitrogen directly. In particular, when this technique is applied to a crystal having both low oxygen concentration and nitrogen concentration, the amount of NO donor to be generated is reduced, so the change in resistivity is very small. That is, a sufficient change in resistivity cannot be obtained, and this is not a reliable method for proving that nitrogen is contained in the crystal. Therefore, when the oxygen concentration is low, the highly sensitive nitrogen concentration cannot be determined.

Non-Patent Document 2 and Patent Document 3 disclose the NO donor in the low nitrogen concentration region in detail. It has been reported that when the nitrogen concentration is 1 × 10 ¹⁴ / cm ³ or less, most of the nitrogen exists as a NO donor instead of the above-described NN pair, NNO, and NNOO.
Of these, although it is not a simple method, the amount of NO donor is directly measured by far-infrared absorption at a very low temperature. When the nitrogen concentration is 1 × 10 ¹⁴ / cm ³ or less, the nitrogen concentration and the NO donor are 1: 1. Therefore, there is a possibility that the nitrogen concentration can be quantitatively measured by applying this technique.
However, this method is a method for quantifying a peak derived from an NO donor in the vicinity of 250 cm ⁻¹ , and a water vapor peak overlaps in a wide range in this wave number region. In order to remove the influence, measurement under reduced pressure is performed. Need to do. Further, the level of the NO donor is ionized at room temperature, and measurement at an extremely low temperature near the liquid He temperature is essential to neutralize this level. In this way, detection is considered possible, but it is by no means a simple method and a practical method.

JP 2003-240711 A JP 2000-332074 A JP 2004-111752 A

JEITA EM-3512 H. Ono and M.M. Horikawa Jpn. J. et al. Appl. Phys. 42 (2003) L261

As described above, a method for obtaining a nitrogen concentration using an NO donor as an index has been proposed.
However, in the case of a low oxygen concentration, the nitrogen concentration cannot be obtained appropriately, or it takes time to evaluate the nitrogen concentration, and there has been no method for simply obtaining it.

Therefore, the present invention has been made in view of the above problems, and in the case of a low oxygen concentration, the nitrogen concentration in the CZ silicon single crystal (particularly, a low concentration such as 1 × 10 ¹⁴ / cm ³ or less). It is an object of the present invention to provide a method capable of easily obtaining a).

In order to achieve the above object, the present invention is a method for evaluating the nitrogen concentration in a nitrogen-doped silicon single crystal produced by the Czochralski method, and when the silicon single crystal is produced, the oxygen concentration is 6. A silicon single crystal having a low oxygen concentration region of 0 × 10 ¹⁷ / cm ³ or less is produced, and a peak derived from an NN pair is detected by infrared absorption spectroscopy for the silicon single crystal in the low oxygen concentration region, Provided is a method for evaluating nitrogen concentration in a silicon single crystal, characterized by evaluating nitrogen concentration from a detection result.

As a result of intensive studies on the method for evaluating the nitrogen concentration in a CZ silicon single crystal, the present inventors thought that the nitrogen in the silicon single crystal is in the form of an NN pair when the oxygen concentration is low. . According to the present invention, the nitrogen concentration can be easily and directly evaluated even when the oxygen concentration in the CZ silicon single crystal is low. Moreover, it is effective even when the nitrogen concentration is low. Furthermore, since it is possible to eliminate the need to detect peaks derived from NNO or NNOO, it is simpler.
In the present application, the oxygen concentration value is based on ASTM'79.

In particular, a nitrogen concentration of 1 × 10 ¹⁴ / cm ³ or less can be evaluated by evaluating the nitrogen concentration in the low oxygen concentration region.
As described above, even if the nitrogen concentration is as low as 1 × 10 ¹⁴ / cm ³ or less, which is conventionally impossible to detect, the present invention can be directly evaluated.

Further, the nitrogen concentration of the entire silicon single crystal can be evaluated by calculating the segregation of nitrogen from the nitrogen concentration evaluated in the low oxygen concentration region.
Thus, it is possible to simply evaluate the entire nitrogen concentration including not only the low oxygen concentration region but also other parts of the silicon single crystal.

Further, the silicon single crystal is produced by a multiple pulling method, and the nitrogen concentration in another silicon single crystal produced by the multiple pulling method is evaluated from the nitrogen concentration evaluated for the silicon single crystal in the low oxygen concentration region. be able to.
In this way, when a plurality of silicon single crystals are produced in one batch by the multiple pulling method, the nitrogen concentration in other silicon single crystals is efficiently evaluated from the nitrogen concentration in the silicon single crystal actually evaluated. be able to.

In addition, the peak derived from the NN pair can be detected without performing a heat treatment for eliminating the oxygen donor in advance.
Thus, the evaluation method of the present invention is simpler because it is not necessary to perform heat treatment for erasing the oxygen donor.

Further, the low oxygen concentration region can be formed in one or more of a cone portion, a straight body portion, and a tail portion of the silicon single crystal.
Thus, for example, a low oxygen concentration region can be formed in any one or more of the cone part, the straight body part, and the tail part in accordance with the quality of a desired silicon single crystal to be produced.

  As described above, according to the present invention, even when the oxygen concentration in the CZ silicon single crystal is low, and even if the nitrogen concentration is low, the nitrogen concentration can be evaluated easily and directly.

It is a flowchart which shows an example of the process of the evaluation method of this invention. It is the schematic which shows an example of the preparation apparatus of a CZ silicon single crystal. It is the schematic which shows an example of the silicon single crystal which has a low oxygen concentration area | region. It is the graph showing the relationship between the nitrogen concentration calculated | required by the infrared absorption measurement in Example 1, and the nitrogen concentration calculated | required from the dopant raw material. It is the graph showing the relationship between the nitrogen concentration calculated | required by the infrared absorption measurement in the comparative example 1, and the nitrogen concentration calculated | required from the dopant raw material. It is the graph showing the relationship between the nitrogen concentration calculated | required by the infrared absorption measurement in Example 2, and a straight body length. It is the graph showing the relationship between the integrated intensity | strength of the peak derived from NO donor calculated | required by the far-infrared FT-IR method, and nitrogen concentration.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings, but the present invention is not limited thereto.
Here, the background of the completion of the present invention by the present inventors will be described in detail.
Regarding the method for evaluating the nitrogen concentration in a nitrogen-doped CZ silicon single crystal, the present inventors first characterized the NO donor. As shown in Non-Patent Document 2 described above, when the nitrogen concentration is 10 14 or less, it is not in the form of NN, NNO, or NNOO but mainly in the form of an NO donor. The present inventors investigated various crystals and determined the correlation between the amount of NO donor, oxygen concentration, and nitrogen concentration.

(NO donor amount (difference in carrier concentration))
An example of a value corresponding to the amount of NO donor is a carrier concentration difference. This carrier concentration difference Δ [n] is a difference in carrier concentration obtained from the difference between the resistivity after heat treatment for erasing the oxygen donor and the resistivity after heat treatment for erasing the NO donor.

Here, how to obtain the carrier concentration difference will be described.
The step of obtaining the carrier concentration difference mainly includes a heat treatment for erasing the oxygen donor, a subsequent measurement of resistivity, a heat treatment for erasing the nitrogen-oxygen donor, and a subsequent resistivity measurement. That is, oxygen donors and nitrogen oxygen donors are present in the crystal of the nitrogen-doped silicon single crystal grown by the CZ method, but the heat treatment for erasing the oxygen donor is relatively low temperature as will be described later. To erase the oxygen donor from the crystal and measure the resistivity. At this time, since the nitrogen-oxygen donor still remains in the crystal, the resistivity here is the resistivity in the state where there is no oxygen donor and there is a nitrogen-oxygen donor.

Next, the heat treatment for erasing the nitrogen-oxygen donor is at a relatively high temperature, and the nitrogen-oxygen donor in the crystal is erased by the heat treatment. Therefore, it is possible to measure the resistivity in the absence of oxygen donor and nitrogen oxygen donor.
Then, the carrier concentration difference caused by the nitrogen-oxygen donor can be obtained from the difference in resistivity.

Here, the heat treatment for oxygen donor erasure and the heat treatment for nitrogen oxygen donor erasure will be described in more detail.
Since the oxygen donor is generated in a relatively low temperature region around 450 ° C., such a low temperature thermal history is not received on the bottom side of the CZ crystal, and almost no oxygen donor is generated. On the other hand, a large amount of oxygen donors are generated on the top side of the crystal because it passes sufficiently through this thermal history region. With the recent increase in crystal length, this tendency becomes more prominent, and a large amount of oxygen donors are present on the top side and almost no oxygen donors are present on the bottom side.

  It is known that this oxygen donor is erased by a slight heat treatment at 650 ° C. for about 20 minutes, for example. Various heat treatments for erasing the oxygen donor have been proposed in addition to this, for example, there are those for high-temperature and short-time treatment using RTA (Rapid Thermal Anneal), and the temperature and time are not particularly specified here. Any heat treatment capable of erasing oxygen donors generated due to oxygen may be used.

  Further, it is described that the nitrogen-oxygen donor is extinguished by heat treatment at a relatively high temperature such as 900 ° C. in Patent Document 3, 1000 ° C. in Patent Document 2, and 1050 ° C. in International Publication No. 2009/025337. Further, the generation temperature of this nitrogen-oxygen donor is described as 500-800 ° C. in Patent Document 2 and 600-700 ° C. in Patent Document 3, and is generated at a higher temperature than that of the oxygen donor. Further, as disclosed in Patent Document 2, the amount of generation is saturated by a relatively short heat treatment. For this reason, compared with the case where oxygen donors are formed at a high density on the top side of the crystal, nitrogen oxygen donors are generated relatively uniformly. In addition, although it is not affected by the in-furnace structure or the growth rate that affects the thermal history of the grown crystal, the influence is relatively small, and the amount of nitrogen-oxygen donor greatly differs depending on the growth conditions.

From the above, as the heat treatment for oxygen donor erasure, the resistivity is measured after performing a slight heat treatment at, for example, about 650 ° C., and the calculated carrier amount is obtained. For example, if the resistivity is measured after high-temperature heat treatment at 900 ° C. or higher and the carrier amount calculated therefrom is obtained, the carrier concentration difference Δ [n] resulting from the nitrogen-oxygen donor can be obtained from the difference. Here, in order to obtain the carrier concentration from the resistivity, an Irvin curve may be used.
In addition, the measuring method of a resistivity is not specifically limited, For example, it can carry out by the four probe method etc.

(About oxygen concentration)
Next, the oxygen concentration [Oi] will be described.
The method for obtaining the oxygen concentration [Oi] can be obtained, for example, by the FT-IR method at room temperature. In [Oi], Oi is described because oxygen atoms are present at interstitial positions in the silicon crystal, and the infrared absorption at these positions is measured and expressed as oxygen concentration. Because it is. Oxygen in which oxygen precipitation heat treatment is performed and oxygen atoms form oxygen precipitates (BMD) does not cause absorption as [Oi], but the oxygen concentration referred to here is naturally not in the state of precipitation heat treatment. Is.

  When the sample has normal resistivity, the FT-IR method is used. However, when the sample is a low resistivity crystal, infrared light is absorbed and the FT-IR method cannot be used. Therefore, the oxygen concentration may be measured by a gas fusion method.

  Note that oxygen dissolved from the quartz crucible passes through the silicon melt and is almost evaporated near the surface of the melt, and only a very small part is taken into the crystal. Therefore, since the oxygen concentration in the silicon crystal changes depending on various operating conditions, it is generally measured and guaranteed by the above FT-IR method or the like.

(About nitrogen concentration)
Next, the nitrogen concentration [N] will be described.
The nitrogen concentration at this time is obtained or predicted from calculation by infrared absorption spectroscopy and segregation. Nitrogen doping in CZ silicon single crystal production is generally performed by introducing a nitrogen dopant into a crucible and dissolving it together with the silicon raw material. At this time, since nitrogen is taken into the crystal according to the segregation phenomenon, if the nitrogen concentration at a certain crystal position is obtained, the nitrogen concentration for each crystal length can be obtained by calculation. Therefore, a doping agent is introduced on the tail side of the crystal so that the nitrogen concentration is higher than the detection limit of infrared absorption spectroscopy (previously reported as 1 × 10 ¹⁴ / cm ³ ), and nitrogen at that position is added. By obtaining the concentration, a sample having a nitrogen concentration of 10 13 to 14 was obtained.

As a result of investigating and analyzing the correlation among the carrier concentration difference Δ [n], nitrogen concentration [N], and oxygen concentration [Oi] actually obtained as described above, Δ [n] is the first power of [N]. And [Oi] was found to be a linear function of approximately 2.5 to 3.5. That is,
Δ [n] = α [N] × [Oi] ^{2.5 to 3.5} + β (where α and β are constants)
It is.
From this, it can be seen that the amount of NO donor generation (that is, the difference in carrier concentration) greatly depends on the oxygen concentration.

By the way, according to Non-Patent Document 2, there is a 1: 1 correlation between the nitrogen concentration and the NO donor concentration when the nitrogen concentration is less than 1 × 10 ¹⁴ / cm ^3, but 1 when the nitrogen concentration is 1 × 10 ¹⁴ / cm ³ or more. : 1 deviates from the correlation. In combination with the above findings, this can be explained as follows. That is, when the nitrogen concentration is less than 1 × 10 ¹⁴ / cm ³ , oxygen is sufficiently abundant compared to nitrogen and most nitrogen forms NO donors. However, when the nitrogen concentration is 1 × 10 ¹⁴ / cm ³ or more, It is thought that oxygen begins to be deficient with respect to nitrogen, and other forms of nitrogen such as NN pairs increase.

  Further discussion will be made from here, and it is expected that the amount of NO donor produced is small even if nitrogen doping is performed in the original silicon single crystal having a sufficiently low oxygen concentration. Therefore, a sample in which the oxygen concentration was actually shifted to two levels was prepared, and the amount of NO donor was measured.

As a result of this measurement, the relationship between the NO donor integrated intensity and the nitrogen concentration is shown in FIG.
In addition, the far-infrared absorption spectroscopy described in the nonpatent literature 2 was used as a quantitative determination method of NO donor amount. In far-infrared absorption spectroscopy, a peak derived from a NO donor appears at 230-255 cm ⁻¹ , and the integrated intensity is plotted on the vertical axis.
In addition, “normal oxygen concentration” in FIG. 7 refers to a sample having an oxygen concentration of 8.5 × 10 ^{17 to} 12 × 10 ¹⁷ atoms / cm ³ . The “low oxygen concentration” refers to a sample having an oxygen concentration of 6.0 × 10 ¹⁷ atoms / cm ³ or less (here, 2.6 × 10 ^{17 to} 3.8 × 10 ¹⁷ atoms / cm ³ ). .

  As is apparent from FIG. 7, the integrated intensity increases linearly with respect to the nitrogen concentration in the normal oxygen concentration sample, but the integrated intensity is weak in the low oxygen concentration sample, and hardly increases as the nitrogen concentration increases. . Therefore, as expected, it can be said that the amount of NO donor produced is small in the crystal having a low oxygen concentration.

As described above, a low oxygen concentration crystal produces a smaller amount of NO donor than a normal oxygen concentration crystal. At this time, most of the nitrogen that has not formed a NO donor forms an NN pair. This is because NNO and NNOO contain oxygen atoms, so it is considered that NN pairs are mainly formed in a crystal having a low oxygen concentration.
Therefore, in the case of a low oxygen concentration, it is considered effective to evaluate the nitrogen concentration using an NN pair as an index instead of an NO donor or the like. In particular, it was found that NN pairs that can be sufficiently detected by infrared absorption spectroscopy were formed even at a low nitrogen concentration of about 10 13.
The present inventors have found the above and completed the present invention.

Hereinafter, the nitrogen concentration evaluation method in the silicon single crystal of the present invention will be described in detail.
FIG. 1 is a flowchart showing an example of steps of the evaluation method of the present invention. As shown in FIG. 1, a nitrogen-doped silicon single crystal having a low oxygen concentration region is produced by the CZ method (step 1), and then a peak derived from the NN pair is detected by infrared absorption spectroscopy (step 2). The nitrogen concentration is evaluated from the detection result (step 3).
Hereinafter, each process is explained in full detail.

(About step 1)
A nitrogen-doped silicon single crystal is produced by the CZ method (including the MCZ method).
Here, an apparatus for producing a silicon single crystal by the CZ method will be described with reference to FIG. As shown in FIG. 2, the silicon single crystal pulling apparatus 1 is disposed around a pulling chamber 2, a crucible 3 (quartz crucible on the inside, a graphite crucible on the outside) provided in the pulling chamber 2, and the crucible 3. Heater 4, crucible holding shaft 5 for rotating crucible 3 and its rotating mechanism (not shown), seed chuck 7 for holding silicon seed crystal 6, wire 8 for pulling up seed chuck 7, wire 8 Is provided with a winding mechanism (not shown). A heat insulating material 9 is disposed around the outside of the heater 4.
The silicon single crystal 10 is pulled up by the wire 8 from the raw silicon melt 11 accommodated in the crucible 3. The silicon melt 11 is obtained by heating and melting a silicon wafer with a nitride film as a nitrogen dopant together with polycrystalline silicon as a raw material.

FIG. 3 shows an example of the silicon single crystal produced.
As shown in FIG. 3, the silicon single crystal 10 includes a cone portion 10a, a straight barrel portion 10b, and a tail portion 10c. Here, the oxygen concentration in the tail portion 10c is 6.0 × 10 ¹⁷ atoms / cm ³ or less. Although the low oxygen concentration region 12 is formed, the formation position of the low oxygen concentration region 12 is not particularly limited as will be described later. In consideration of other conditions, it may be formed in any one or more of the cone part 10a, the straight body part 10b, and the tail part 10c.

  When producing the silicon single crystal 10 using the silicon single crystal pulling apparatus 1 shown in FIG. 2, first, the polycrystalline silicon contained in the crucible 3 is heated and melted together with the nitrogen dopant as described above by the heater 4. To do. Then, the seed crystal 6 held by the seed chuck 7 is immersed in the silicon melt 11 in the crucible 3 while rotating the crucible 3. Next, the rod-shaped silicon single crystal 10 doped with nitrogen is pulled up from the silicon melt 11 while rotating and winding the wire 8.

At this time, the silicon single crystal 10 is pulled up so as to have the low oxygen concentration region 12.
Since the crucible 3 containing the silicon melt 11 (here, the inner quartz crucible) is composed of silicon and oxygen, oxygen atoms are eluted into the silicon melt 11. The oxygen atoms move by convection in the silicon melt and finally evaporate from the surface of the silicon melt 11. At this time, most of the oxygen evaporates, but part of the oxygen is taken into the silicon single crystal 10 and becomes interstitial oxygen (Oi).

Therefore, the oxygen elution amount is controlled by changing the rotational speed of the crucible 3, and the oxygen concentration in the axial direction (length direction) in the silicon single crystal 10 is adjusted. In addition, the flow of convection in the silicon melt 11 can be controlled by changing the number of rotations of the silicon single crystal 10 at the time of pulling, or by changing the magnetic field application conditions in the case of the MCZ method. It is also possible to adjust the oxygen concentration by adjusting the flow rate of the inert gas flowing into the pulling chamber 2 or controlling the amount of oxygen evaporated from the surface by controlling the pressure in the pulling chamber 2. .
The low oxygen concentration region 12 is formed in the silicon single crystal 10 by appropriately performing such adjustment.

As described above, the formation position of the low oxygen concentration region 12 is not particularly limited, and can be formed in any one or more of the cone portion 10a, the straight body portion 10b, and the tail portion 10c. The formation position can be determined by desired quality or the like.
That is, for example, when a silicon single crystal having an oxygen concentration of more than 6.0 × 10 ¹⁷ atoms / cm ³ is manufactured, if a low oxygen region is formed in the straight body portion in order to measure the nitrogen concentration, that portion becomes a product. In other words, it becomes a factor to lower the yield. In such a case, for example, the operation conditions are controlled so that the oxygen concentration at the tail portion is 6.0 × 10 ¹⁷ atoms / cm ³ or less, and the low oxygen concentration region 12 is formed in the tail portion 10c as shown in FIG. be able to. Since the cone portion 10a and the tail portion 10c have a larger surface area of the silicon melt 11 than the straight barrel portion 10b, oxygen is more easily evaporated and the low oxygen region 12 is easily formed. In the subsequent step, the nitrogen concentration can be evaluated using the location.
Of course, if the target oxygen concentration of the straight body portion is 6.0 × 10 ¹⁷ atoms / cm ³ or less, a sample for infrared absorption measurement is prepared from the end face when the block is produced from the straight body portion of the silicon single crystal. It can be cut out and used to directly measure the nitrogen concentration in a later step.

It is also possible to produce the silicon single crystal 12 by a multiple pulling method.
Here, the multiple pulling method means that after pulling up the silicon single crystal 10 from the silicon melt 11 accommodated in the crucible 3, the raw material polycrystalline silicon is added to the silicon melt remaining in the crucible 3. This is a method of repeatedly pulling a plurality of silicon single crystals with one crucible by repeating the steps of melting and pulling up the next silicon single crystal.
In the present invention, the multiple concentration method as described above is performed, and the nitrogen concentration is directly measured for at least one of the plurality of silicon single crystals to be pulled up. It is also possible to evaluate the nitrogen concentration in other silicon single crystals pulled by the method.

(About step 2)
A sample is cut out from the low oxygen concentration region 12 of the silicon single crystal 10 produced in step 1.
The sample shape cut out from the silicon single crystal is not particularly limited, but the sample thickness is preferably 1.5 mm or more in order to increase the signal intensity.

In detecting the peak derived from the NN pair of the sample, heat treatment for eliminating the oxygen donor may be performed in advance. However, in the present invention, detection can be performed without performing this heat treatment, and in this case, it is simpler.
As described in Patent Document 1, in the area of 950-1050Cm ^-1, very weak series of vibrational transitions absorption associated with oxygen donors (1012,1006,1000,988,975cm ^-1) appears. Since this region coincides with the wave number region where NN pair absorption occurs, the spectrum baseline tends to fluctuate for each sample to be measured. When obtaining the peak intensity of an NN pair, it is normal to obtain the absorbance spectrum by taking the difference from the spectrum of the reference that does not contain nitrogen. There is a high probability that an error will occur in the result.

Therefore, in Patent Document 1, this problem is avoided by adding a heat treatment for erasing the oxygen donor, but in the present invention, infrared absorption spectroscopy measurement is performed under a low oxygen condition of 6.0 × 10 ¹⁷ atoms / cm ³ or less. Therefore, the above oxygen donor problem can be avoided without heat treatment.

  Then, the peak derived from the NN pair is detected from the cut sample by infrared absorption spectroscopy, but this detection method itself is not particularly limited, and for example, it can be detected by the same method using the same apparatus as the conventional one. it can. It can be performed by a conventionally well-known method such as the FT-IR method.

Since the peak of the NN pair appears at 963 and 766 cm ⁻¹ , the measurement conditions are set so that this region can be measured. Since the intensity of the NN pair is very weak compared to the baseline, a reference sample is prepared, and the peak is easily analyzed by taking the spectral difference. A reference having an oxygen concentration, a resistivity, a sample thickness close to that of the measurement sample, and not doped with nitrogen can be used.

(About step 3)
And it evaluates by converting into the nitrogen concentration from the obtained peak intensity. The peak intensity (absorbance) is converted into an absorption coefficient α according to the Lambert-Beer rule. For conversion to the nitrogen concentration, for example, a JEITA conversion coefficient of 1.83 × 10 ¹⁷ can be used.
However, as described above, even with low oxygen concentration crystals with an oxygen concentration of 6.0 × 10 ¹⁷ atoms / cm ³ or less, the generation amount of NO donors should be different if the oxygen concentration is greatly different. The amount will also change. Therefore, it is desirable to obtain a conversion factor for each oxygen concentration level.
As described above, the nitrogen concentration in the silicon single crystal in the low oxygen concentration region can be evaluated.

Here, in the CZ method, impurities are incorporated into the crystal according to segregation. Therefore, after obtaining the nitrogen concentration in a partial region of the CZ silicon single crystal, the nitrogen concentration in other parts of the silicon single crystal is calculated from the segregation calculation. Can be sought. In particular, the nitrogen concentration of the entire single crystal can be obtained.
That is, in the case of a doping method in which a nitrogen dopant is introduced when the raw material is melted before pulling the silicon single crystal, the nitrogen concentration in the silicon single crystal is determined according to the segregation coefficient and solidification rate of nitrogen. Therefore, when the nitrogen concentration in a part of the pulled silicon single crystal (low oxygen concentration region) is obtained by the method described above, the nitrogen concentration in other parts of the silicon single crystal can be easily calculated by calculating the segregation of nitrogen. Can be sought.

Further, if an evaluation method using the calculation of the segregation of nitrogen is developed, it can be applied to a plurality of silicon single crystals pulled by the above-described multiple pulling method.
That is, a nitrogen dopant is introduced at the time of melting the raw material before the first pulling of the silicon single crystal, a silicon raw material is poured into the residual melt after the pulling of the first silicon single crystal, and a melt is formed again. In the case of the multiple pulling method in which the second one is pulled up from the melt and the steps are repeated, if the nitrogen concentration in the low oxygen concentration region is determined for one of the plurality of pulled up crystals by the method described above, the silicon The nitrogen concentration of the entire single crystal can be determined, and further the nitrogen concentration in other silicon single crystals in the same batch can be determined.

For example, if there is one silicon single crystal with a known nitrogen concentration, the nitrogen concentration in the residual melt after pulling up the silicon single crystal can be determined, and the nitrogen concentration contained in the silicon single crystal to be pulled up next is determined from there. Can be sought. If a silicon single crystal with a known nitrogen concentration is produced in the middle of a batch, the nitrogen concentration of the silicon single crystal pulled before that can be similarly determined.
Thus, the nitrogen concentration in the silicon single crystal in the low oxygen concentration region can be evaluated, and then the nitrogen concentration in another silicon single crystal can be evaluated efficiently.

  In addition, it is considered that an error is likely to occur when the nitrogen concentration of a plurality of silicon single crystals is obtained from the result of one silicon single crystal, but nitrogen has a segregation coefficient as small as about 0.0007, Since it hardly evaporates, it can be said that there is little difference between the amount of nitrogen remaining in the melt between the calculation and the actual measurement.

As described above, there has conventionally not been a simple nitrogen concentration evaluation method for low oxygen concentration crystals. However, if the evaluation method of the present invention is used, the detection of peaks derived from NN pairs can detect NNO or NNOO-derived peaks. There is no need to detect a peak, and the nitrogen concentration in the low oxygen concentration region can be evaluated easily and directly. Moreover, evaluation is possible even at a low nitrogen concentration of 1 × 10 ¹⁴ / cm ³ or less, which was conventionally set as the lower limit of detection.
In addition, by utilizing the calculation of nitrogen segregation, it is possible to easily evaluate the nitrogen concentration in the entire silicon single crystal and also in other silicon single crystals in the multiple pulling method.

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)
The evaluation method of the nitrogen concentration in the silicon single crystal of the present invention was carried out. The silicon single crystal pulling apparatus shown in FIG. 2 was used for the production of the silicon single crystal.
A silicon wafer with a nitride film was introduced as a dopant into the raw material, and a nitrogen-doped silicon single crystal having a low oxygen concentration region was pulled up by the CZ method. Samples a and b were cut out from the low oxygen concentration region of the silicon single crystal. Reference R1 was prepared from another nitrogen non-doped silicon single crystal.
In samples a and b and reference R1, the respective oxygen concentrations were 3.0 × 10 ¹⁷ , 2.6 × 10 ¹⁷ , 3.8 × 10 ¹⁷ atoms / cm ³ , and the conductivity type was all N-type and resistivity Are 56, 43, and 39 Ωcm, respectively, and the thicknesses are all 1.9 mm.

And each sample was measured using the infrared absorption measurement, and the difference spectrum with a reference was acquired.
The peak intensity (absorbance) derived from the NN pair appearing at 963 cm ⁻¹ is 3.4 × 10 ⁻⁵ and 7.3 × 10 ⁻⁵ for samples a and b, respectively, and the absorption coefficient α determined according to the Lambert-Beer rule is The results were 4.1 × 10 ⁻⁴ and 8.7 × 10 ⁻⁴ cm ⁻¹ , respectively. When the conversion factor of 1.83 × 10 ¹⁷ was used, the respective nitrogen concentrations could be obtained as 7.6 × 10 ¹³ and 1.6 × 10 ¹⁴ / cm ³ .

Thus, in sample a, 10 to the 13th power level of low-concentration nitrogen, which was thought to be undetectable, could be detected.
Further, the nitrogen concentrations of a and b were calculated from the number of nitrogen atoms contained in the starting nitride film, and were found to be 3.6 × 10 ¹³ and 7.2 × 10 ¹³ / cm ³ , respectively.

A plot of these results is shown in FIG. The horizontal axis represents the nitrogen concentration calculated from the dopant material, and the vertical axis represents the nitrogen concentration in the infrared absorption measurement by the method of the present invention.
Although there was a difference in absolute value between the calculated value and the infrared absorption measurement value, the tendency of the calculated value and the measurement value was the same. The two points show a substantially proportional relationship, and it can be said that the quantitativeness by the evaluation method of the present invention is sufficient.
Note that the difference in absolute value can be adjusted by examining the relationship between the measured value and the calculated value according to the present invention for various samples as necessary, and performing appropriate calibration or conversion.

(Comparative Example 1)
In the same manner as in Example 1, a silicon wafer with a nitride film was introduced into the raw material as a dopant, and the nitrogen-doped silicon single crystal was pulled up by the CZ method. Samples c and d were cut out from the silicon single crystal, and a reference R2 was prepared from another nitrogen non-doped silicon single crystal.
In the samples c and d and the reference R2, the respective oxygen concentrations are 11.7 × 10 ¹⁷ , 11.5 × 10 ¹⁷ , 12.1 × 10 ¹⁷ atoms / cm ³ , the conductivity type is all P-type, and the resistivity Are 24, 20 and 24 Ωcm, respectively, and the thicknesses are all 1.9 mm.

And each sample was measured using the infrared absorption measurement, and the difference spectrum with a reference was acquired.
When the peak intensity (absorbance) derived from the NN pair appearing at 963 cm ⁻¹ was obtained, the peak could not be detected in sample c, and it was 4.9 × 10 ^{−5 in} sample d. The nitrogen concentration of sample d obtained using a conversion factor of 1.83 × 10 ¹⁷ was 1.1 × 10 ¹⁴ / cm ³ .
Moreover, the nitrogen concentration of the samples c and d obtained by calculation from the dopant material was 3.0 × 10 ¹³ and 6.3 × 10 ¹³ / cm ³ , respectively.

Thus, the nitrogen concentration calculated from the raw material is similar in the sample a of Example 1 and the sample c of Comparative Example 1 (3.6 × 10 ¹³ , 3.0 × 10 ¹³ / cm ³ ), but in the sample c, Since the oxygen concentration exceeded 6.0 × 10 ¹⁷ atoms / cm ³ , it was not possible to detect the low concentration of nitrogen in the 10 13th range.

FIG. 5 shows the results of samples c and d. For comparison, FIG. 5 also shows the results of samples a and b of Example 1.
Since the peak of sample c could not be detected, the measured value was 0 / cm ³ .
Compared to the two points of samples a and b of Example 1, sample d has a nitrogen concentration obtained from infrared absorption lower than the calculated value. Similarly to FIG. 7, this is data supporting that a large amount of NO donor is generated at the normal oxygen concentration as in the sample d and the amount of NN pairs is smaller than that of the low oxygen concentration crystal.

(Comparative Example 2)
Sample e was prepared and the nitrogen concentration was investigated in the same manner as in Comparative Example 1 except that the oxygen concentration was 7 × 10 ¹⁷ atoms / cm ^3. Sample A in Example 1 was examined in the same manner as Sample d. The nitrogen concentration obtained from the infrared absorption was lower than the calculated value, compared with the two points b and b. This is also because the amount of NN pairs is small.

In the method disclosed in Patent Document 2 described above, it is considered necessary to increase the oxygen concentration in order to accurately quantify the low nitrogen concentration. On the other hand, in the above comparative example, when the oxygen concentration is high, the low nitrogen concentration cannot be quantified, which is seemingly contradictory.
This is because the form of nitrogen detected by each method is different. In Patent Document 2, NO donors are indirectly quantified from the change in resistivity before and after the heat treatment, and the higher the oxygen concentration in the silicon single crystal, the greater the amount of NO donors generated, which is considered to increase the sensitivity.

On the other hand, in the method of the present invention, quantification is performed by detecting NN pairs. As mentioned above, nitrogen takes several types of forms in a silicon single crystal, but there are NN pairs that do not contain oxygen and other forms that contain oxygen, and the presence of nitrogen depends on the oxygen concentration in the silicon single crystal. The ratio will change. Therefore, in the quantification method using the NN pair, the detection sensitivity of the nitrogen concentration is low when the oxygen concentration is high, but the detection sensitivity of the nitrogen concentration is increased under the low oxygen concentration condition, and the oxygen concentration is reduced to 6.0 × 10 ¹⁷ atoms / If it is less than cm ³ , highly sensitive determination of nitrogen concentration becomes possible.

(Example 2)
The evaluation method of the present invention was carried out. The nitrogen concentration of the entire silicon single crystal was determined from the nitrogen concentration in the low oxygen concentration region. In addition, here, it calculated | required about six places on behalf of the following.
First, a silicon wafer with a nitride film was introduced into the raw material as a dopant, and the nitrogen-doped silicon single crystal was pulled up by the CZ method. At that time, the silicon single crystal was fabricated so that the oxygen concentration was 6.0 × 10 ¹⁷ atoms / cm ³ or less in the entire region of the straight body of the silicon single crystal. And when the sample was cut out at six locations of the straight cylinder positions 20, 45, 70, 95, 120, and 150 cm from the cone side of the produced ingot and the oxygen concentration at each position was measured, all were 3.0 × 10 ¹⁷ to It was within the range of 3.8 × 10 ¹⁷ atoms / cm ³ .

Next, when the nitrogen concentration in the portion of the straight cylinder of 20 cm was determined by infrared absorption measurement (about NN pair), it was determined to be 6.2 × 10 ¹³ / cm ³ . From this value, a theoretical curve of segregation calculated with a segregation coefficient of nitrogen of 0.0007 was obtained.
FIG. 6 shows the nitrogen concentration and theoretical curve of the part of the straight cylinder 20 cm with respect to the position of the straight cylinder length, and also the measured values for the nitrogen concentration of other parts (45, 70, 95, 120, 150 cm). Showed.

  As shown in FIG. 6, it can be seen that the nitrogen concentration at other sites is close to the theoretical value of segregation. Therefore, it was confirmed that the nitrogen concentration of the entire silicon single crystal can be sufficiently estimated from the segregation calculation.

(Example 3)
The evaluation method of the present invention was carried out.
Three silicon single crystals having a diameter of 200 mm were pulled by a multiple pulling method under the conditions of an initial melt weight of 200 kg, a pulling crystal weight of 150 kg, and a recharge raw material weight of 150 kg.
At this time, the nitrogen dopant was introduced only when the raw material was melted before the first pulling, and no additional nitrogen dopant was introduced during recharging. The operating conditions during pulling of the silicon single crystal are the same as in Example 2. When a sample was cut out from a portion of the straight body 100, 150 cm of each ingot and the oxygen concentration at each point of the ingot was measured, all were in the range of 2.9 to 3.6 × 10 ¹⁷ atoms / cm ³ .

First, when the infrared absorption measurement (about NN pair) of the first straight body of 150 cm was performed, the nitrogen concentration was determined to be 1.5 × 10 ¹⁴ / cm ³ .
From this quantitative value, the nitrogen concentration at the 100 cm portion of the same first silicon single crystal and the sample cutting positions (100, 150 cm) of the other second and third silicon single crystals was obtained by calculation. Subsequently, the nitrogen concentration at each position was actually determined by performing infrared absorption measurement (for NN pairs). The results are shown in Table 1.

  As shown in Table 1, it was confirmed that the measured value and the calculated value were almost the same value. Therefore, in the multiple pulling method, it was confirmed that the nitrogen concentration in another silicon single crystal can be sufficiently estimated from the nitrogen concentration in the low oxygen concentration region of one silicon single crystal.

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

1 ... silicon single crystal pulling device, 2 ... pulling chamber, 3 ... crucible,
4 ... heater, 5 ... crucible holding shaft, 6 ... seed crystal, 7 ... seed chuck,
8 ... wire, 9 ... insulating material, 10 ... silicon single crystal,
10a ... cone portion, 10b ... straight barrel portion, 10c ... tail portion,
11 ... silicon melt, 12 ... low oxygen concentration region.

Claims (5)

A method for evaluating the nitrogen concentration in a nitrogen-doped silicon single crystal produced by the Czochralski method,
When producing the silicon single crystal, a silicon single crystal having a low oxygen concentration region having an oxygen concentration of 6.0 × 10 ¹⁷ atoms / cm ³ or less is produced.
The silicon single crystal in the low oxygen concentration region is not subjected to a heat treatment for eliminating the oxygen donor in advance, and the peak derived from the NN pair among the peaks derived from the NN pair, NNO, and NNOO by infrared absorption spectroscopy. A method for evaluating a nitrogen concentration in a silicon single crystal, wherein only the nitrogen is detected and the nitrogen concentration is evaluated from the detection result. 2. The method for evaluating a nitrogen concentration in a silicon single crystal according to claim 1, wherein a nitrogen concentration of 1 × 10 ¹⁴ / cm ³ or less is evaluated by evaluating a nitrogen concentration in the low oxygen concentration region.   3. The nitrogen in the silicon single crystal according to claim 1, wherein the nitrogen concentration of the entire silicon single crystal is evaluated by calculation of nitrogen segregation from the nitrogen concentration evaluated in the low oxygen concentration region. 4. Concentration evaluation method. The silicon single crystal is produced by a multiple pulling method,
4. The nitrogen concentration in another silicon single crystal produced by a multiple pulling method is evaluated from the nitrogen concentration evaluated for the silicon single crystal in the low oxygen concentration region. The method for evaluating a nitrogen concentration in a silicon single crystal according to claim 1. The low oxygen concentration region, the cone portion of the silicon single crystal, the straight body portion, the silicon according to any one of claims 4 to claims 1, characterized in that formed on any one or more of the tail portion Method for evaluating nitrogen concentration in a single crystal.

Images