Method for evaluating the concentration of a defect in a silicon single-crystal substrate

Abstract

Method for evaluating the concentration of a defect in a silicon single-crystal substrate, wherein the defect is formed by particle beam irradiation in the silicon single-crystal substrate, comprising the steps of: Measuring the specific resistance of the silicon single-crystal substrate, followed by irradiating the silicon single-crystal substrate with the particle beam, and remeasuring the specific resistance of the silicon single-crystal substrate after irradiation; Determine the carrier concentration in the silicon single-crystal substrate before and after irradiation based on the measured resistivity values ​​before and after particle beam irradiation to calculate a carrier concentration change ratio; and Evaluating the concentration of a W-defect based on the change ratio of the carrier concentration, wherein the VV-defect consists of a silicon atom vacancy in a Frenkel pair formed by particle beam irradiation in the silicon single-crystal substrate and another silicon atom vacancy.

Description

Technical area

[0001] The present invention relates to a method for evaluating a concentration of a crystal defect introduced into a silicon single crystal substrate by irradiation with a particle beam consisting of an electron, an ion, etc. State of the art

[0002] In the manufacturing process of semiconductor devices such as PN diodes or IGBTs, irradiation with an electron beam or various particle beams, such as a proton beam and a helium ion beam, can be performed to control their carrier lifetime. For example, Patent Document D1 discloses a PN diode in which crystal defects are formed near the PN junction interface by implanting protons. Irradiation of silicon crystals with a particle beam introduces crystal defects and correspondingly reduces the carrier lifetime. A desired lifetime can be obtained by controlling the irradiation amount.

[0003] Patent Document 2 relates to a method for measuring the lifetime value of a wafer after electron beam irradiation, and more particularly, to a measuring method for easily calculating the lifetime value after electron beam irradiation.

[0004] The following is a method for forming a defect in a silicon single crystal using particle beam irradiation. Initially, when particle beams are applied, silicon atoms forming a crystal jump from their lattice positions to form interstitial silicon atoms (I) and monovacancies (V). This pair of interstitial silicon and the monovacancy is called a Frenkel pair.

[0005] A portion of the interstitial silicon atoms is replaced by a carbon atom Cs existing at the lattice position to form an interstitial carbon Ci. This interstitial carbon is unstable and consequently bonds with an interstitial oxygen or other substitute carbon to form a composite defect, such as a CiOi or CiCs. The remaining interstitial silicon atoms are believed to form clusters by agglomerating with other interstitial silicon atoms, remaining intact in the crystals, or bonding with monovacancies to disappear again.

[0006] It is known that these defects due to interstitial carbon atoms, such as CiOi or CiCs, can be detected by photoluminescence (PL) or cathode luminescence (CL) and are affected by the oxygen concentration or carbon concentration in the crystal (Non-Patent Document 1).

[0007] On the other hand, it is assumed that the monovacancy in the Frenkel pair enters a stable state when the monovacancy is in a certain position with another monovacancy, and many of these form W defects consisting of two silicon atom vacancies. There is an example of evaluating this defect using electron spin resonance (ESR) analysis (Non-Patent Document 2).

[0008] Non-Patent Document 3 provides an overview of the characteristics of irradiation-induced defects that are important for tuning the characteristics of power devices. Citation listPatent literature Patent Document 1: Japanese Unexamined Patent Publication (Kokai) JP H08 - 102 545 A Patent document 2: JP 2012- 199 299 A Non-Patent Document 1: M. Nakamura et al., J. Electrochem. Soc., Vol. 141, No. 12, 3576 (1994) Non-Patent Document 2: S. Ishino, “shousha sonshou,” University of Tokyo Press (1979), pp. 178-183 Non-patent document 3: SIEMIENIEC, R: Irradiation-Induced Deep Levels in Silicon for Power Device Tailoring, Journal of The Electrochemical Society, J. Electrochem. Soc. 153 (2), G108-G118, 2005, DOI: 10.1149 / 1.2137649. Summary of the inventionTechnical problem

[0009] As described above, the crystal defect becomes a carrier trap due to irradiation and plays an important role in reducing lifetime. Although these quantities must be controlled in power generators, etc., to adjust the lifetime to an appropriate value, there has been a problem in that the amount of VV defects formed is difficult to quantitatively evaluate, compared to carbon-related defects, which can be quantified to a certain extent by evaluation methods such as CL.

[0010] The present invention was completed in view of the problems described above. It is an object of the present invention to provide a method that easily evaluates the concentration of a UV defect formed by particle beam irradiation in a silicon single-crystal substrate. Solution to the problem

[0011] To solve the problems, the present invention provides a method for evaluating a concentration of a defect in a silicon single crystal substrate, the defect being formed in the silicon single crystal substrate by particle beam irradiation, comprising the steps of: measuring the resistivity of the silicon single crystal substrate, followed by irradiating the silicon single crystal substrate with the particle beam, measuring the resistivity of the silicon single crystal substrate again after the irradiation; determining each carrier concentration in the silicon single crystal substrate before and after the irradiation based on the measured resistivity results before and after the particle beam irradiation to calculate a change ratio of the carrier concentration;and evaluating a concentration of a W defect based on the change ratio of the carrier concentration, wherein the VV defect consists of a silicon atom vacancy and is formed by the particle beam irradiation in the silicon single crystal substrate;

[0012] As described above, the present invention can evaluate the VV defect concentration based on the measurement of the resistivities before and after the particle beam irradiation and the change ratio of the carrier concentration obtained by the measurement of the resistivities, and thus provides a very simple evaluation method. Furthermore, it is possible to obtain the relationship of the VV defect concentration to an irradiation amount of a laser beam or a crystal quality (such as an oxygen concentration) from the evaluation result.

[0013] The particle beam for irradiating the silicon single crystal substrate can consist of an electron or an ion.

[0014] Such particle beam irradiation enables the formation of a Frenkel pair and a W defect due to a monovacancy V of the Frenkel pair in a silicon single crystal substrate.

[0015] A doping impurity contained in the silicon single crystal substrate may be any of phosphorus, arsenic and antimony.

[0016] Using these doping impurities, each of which has the same or larger atomic radius compared to silicon and can form a complex with a monovacancy to be stabilized, the evaluation of the VV defect concentration can be performed even more reliably using the change ratio of the carrier concentration.

[0017] It is also possible for the method according to the invention to further comprise the steps of: preparing a silicon single-crystal substrate for preliminary examination having a pre-evaluated W defect concentration separately from the silicon single-crystal substrate to be evaluated; determining a correlation based on the pre-selected W defect concentration and a change ratio of the carrier concentration before and after the particle beam irradiation with respect to the silicon single-crystal substrate for preliminary examination; and irradiating the silicon single-crystal substrate to be evaluated with the particle beam to evaluate the W defect concentration using the correlation.

[0018] Using a result from such a preliminary investigation, the relative evaluation of the VV defect concentration in the silicon single crystal substrate can be performed even more easily. Advantageous effects of the invention

[0019] As described above, the present invention makes it possible to easily evaluate the concentration of UV defects due to particle beam irradiation and obtain the ratio of the UV defect concentration to the particle beam irradiation amount or crystal quality. This makes it possible to more easily determine the required irradiation amount or crystal quality to control the lifetime of a silicon single-crystal substrate to a specific value. Brief description of the drawings Fig. 1 is a flow chart showing an example of the steps of the method for evaluating a defect concentration in a silicon single crystal substrate of the present invention; Fig. 2 is a graph showing a relationship between a change ratio of carrier concentration and an oxygen concentration; Fig.3 is a graph showing a relationship between the change ratio of a carrier concentration and an irradiation amount of an electron beam, and Fig. 4 is a flowchart showing another example of the steps of the method for evaluating a defect concentration in a silicon single crystal substrate of the present invention. Description of the embodiments

[0020] The embodiment of the present invention will be explained below, but the present invention is not limited thereto.

[0021] First, it will be described how the inventors completed the present invention.

[0022] As described above, it was previously assumed that many of the Vs (monovacancies) of Frenkel pairs formed by particle beam irradiation constitute W defects. However, the present inventors conducted careful investigations and found that the Vs mainly involve three reactions: (1) V+O→VO, (2) V+X→VX, and (3) V+V→VV. In these formulas, O represents an oxygen atom, and X represents a dopant atom.

[0023] Here, the amount of each impurity element and defect on an N-type silicon substrate will be determined, controlling the lifetime with respect to IGBT. The oxygen concentration is on the order of the 17th power to the 18th power per 1 cm 3 , for example when produced by a CZ process; and it is approximately at the upper end of the 15th power to the 16th power per 1 cm 3in an FZ process, although this depends on the oxygen concentration of the raw material. The dopant concentration is 5×10 14 atoms / cm 3 or less of the phosphorus concentration, for example, assuming that the resistivity range generally used as an IGBT will be 10 Ωcm or more. Although the amount of monovacancies formed varies greatly depending on the amount of irradiation with a particle beam and is difficult to evaluate, an example is assumed herein in such a way that the irradiation is carried out to approximately 1×10 13 up to 1×10 15 / cm 3 of Frenkel pairs in a crystal.

[0024] Under the above condition, the reaction of (1) occurs preferentially because the oxygen concentration is sufficiently high relative to the phosphorus concentration [P] or the V concentration [V]. Reactions (2) and (3) occur in the remaining V. This means that reaction (1) occurs readily, and, in contrast, the reactions of (2) and (3) hardly occur, in a crystal with a high oxygen concentration; in a crystal with a low oxygen concentration, reactions (2) and (3) readily occur instead of reaction (1), which is suppressed. In the VP and the VV, which are products of reactions (2) and (3), it is assumed that the concentrations [VP] and [VV] are determined depending on the remaining VV concentration [remaining V] ([remaining V] ∝ [VP], [remaining V] ∝ [VV]), and therefore it is possible to represent this in a simple way as [remaining V] ∝ [VP] ∝ [VV].

[0025] On the other hand, it is reported that VP at a position approximately 0.45 eV forms a level below the silicon conduction bond by DLTS (Deep Level Transient Spectroscopy) analysis, and therefore does not function as an N-type dopant. This means that the resistivity of a silicon substrate increases as the amount of VP is formed, and consequently, the [VP] can be evaluated by measuring the change ratio of the carrier concentration before and after irradiation. Therefore, the [VV] can be simply evaluated as [VV] ∝ [VP] ∝ Change Ratio of Carrier Concentration.

[0026] The present inventors have discovered the foregoing and thereby completed the present invention.

[0027] The present invention will be described in more detail with reference to the figures below, but the present invention is not limited thereto. <Erste Ausführungsform>

[0028] Fig. 1 shows an example of the steps of the method for evaluating a defect concentration in a silicon single-crystal substrate of the present invention. (Step 1: Preparation of a silicon single-crystal substrate to be evaluated) First, a silicon single-crystal substrate to be evaluated (occasionally referred to as a wafer for convenience hereinafter) is prepared. The silicon single-crystal ingot as a raw material of the wafer can be prepared by any of a floating zone method (FZ method) and a Czochralski method (CZ method). The thickness or surface treatment method of the wafer is not particularly limited.

[0029] The oxygen concentration of the wafer, although not particularly limited, can be set to approximately 5×10 15 up to 2×10 18 atoms / cm 3 which is within a range of a wafer generally produced by an FZ process or a CZ process. In such a range, the evaluation method of the present invention can be effectively used because the above defect formation mechanism discovered by the present inventor tends to be valid, although it depends on the dopant concentration and so on.

[0030] The conductivity type of the silicon can be N-type, doped with, for example, any of phosphorus, arsenic, and antimony, although it is not particularly limited. The dopant concentration in the crystal can be 5×10 14 atoms / cm 3or less. This corresponds to a resistivity of approximately 10 Ωcm or more in the case of phosphorus doping.

[0031] Regarding the reaction of a monovacancy with a dopant atom, the dopant atom forms a complex with the monovacancy to be stabilized when the atomic radius is equal to or larger than that of silicon. On the other hand, it forms a complex with interstitial silicon to be stabilized when the atomic radius is smaller than silicon. Consequently, the dopant atom can more reliably form a composite effect with the monovacancy in the case of the above elements, such as phosphorus, arsenic, and antimony, and the above mechanism is more likely to occur in the case of an element with a relatively small atomic radius, such as boron.

[0032] As described above, the reason the above examples are listed as a dopant species or a range of concentrations is that the defect formation mechanism tends to be even more valid as explained with respect to the above oxygen concentration. Phosphorus doping will be explained below as an example.

[0033] The wafer may or may not be doped with nitrogen. The carbon concentration is also not particularly limited. This is because nitrogen and carbon are assumed not to contribute to the reaction of a monovacancy V of Si. However, the amount of carbon changes the amount of defects relative to carbon due to interstitial silicon I. Consequently, an even more accurate evaluation can be obtained if this influence is taken into account when relating the inventive defect evaluation to the device properties. (Step 2: Measuring the specific resistance)

[0034] The resistivity of the above wafer is measured. However, a thermal donor or an NO donor may be generated if oxygen or nitrogen is present in the wafer. To prevent the resistivity from deviating from the intrinsic value due to the donor, it is preferable to perform a donor annihilation heat treatment beforehand. For example, in the case of a CZ wafer without nitrogen doping, the thermal donor can be sufficiently annihilated by applying a heat treatment of 650°C for approximately 20 minutes.

[0035] Regarding the resistivity measurement method, the use of a four-point probe method is simple and preferable when an electron beam is used in the subsequent particle beam irradiation, but the measurement method is not limited to this. Particularly, in the case of using an ion beam other than an electron beam, defects in a wafer due to irradiation tend to be located at a certain depth compared to the case of using an electron beam. In this case, it is preferable to use a measurement method that examines the distribution along the depth direction, such as spreading resistance (SR) measurement after angle polishing. As described above, the measurement method can be appropriately determined according to the irradiation condition and so on. (Step 3: Particle irradiation)

[0036] The wafer is then irradiated with a particle beam. The particle beam can consist of an electron or an ion. This irradiation forms crystal defects (Frenkel pairs and VV defects due to their monovacancy, etc.) in the silicon single crystal. The irradiation conditions of the particle beam are not particularly limited, and the irradiation amount can be appropriately adjusted so that an appropriate number of Frenkel pairs are formed and the above mechanism is valid. Although it is difficult to calculate a vacancy concentration from the irradiation amount, in electron beam irradiation, it is preferable to set the irradiation amount to 1×10 16 / cm 2 or less at an accelerating voltage of 2 MV. (Step 4: Measuring the specific resistance)

[0037] Then the resistivity is again measured on the wafer after irradiation with a particle beam.

[0038] In actual use as a device, annealing is usually performed after irradiation to remove unnecessary damage. However, heat treatment is unnecessary in this case, as the purpose is to evaluate the damage. However, it is of course possible to perform heat treatment if necessary. (Step 5: Calculating a change ratio of the carrier concentration)

[0039] Based on the resistivities before and after irradiation, the carrier concentrations before and after irradiation are calculated using an Irvin curve, etc., and the change ratio of the carrier concentration is further calculated. The change ratio of the carrier concentration can be determined, for example, by the following equation. Change ratio of carrier concentration = ([n]f−[n]i) / [n]i×100 Here [n]i a carrier concentration before irradiation and [n] f is a carrier concentration after irradiation. (Step 6: Evaluation of the VV defect concentration)

[0040] Using the change ratio of the carrier concentration calculated as described above, the VV defect concentration is evaluated.

[0041] Herein, a relationship between the change ratio of a carrier concentration and an oxygen concentration or an irradiation amount of the particle beam is first explained.

[0042] Fig. Figure 2 shows a correlation between oxygen concentration and the change ratio of carrier concentration. This is the result of irradiating the sample with an electron beam of 6×10 14 / cm 2of each of the phosphorus-doped N-type wafers (resistivity: approximately 60 Ωcm) with various oxygen concentrations, manufactured by an FZ process and a CZ process.

[0043] As through this Fig. As shown in Figure 2, the change ratio of the carrier concentration becomes smaller as the oxygen concentration increases. This indicates that more monovacancies Vs are consumed by oxygen as the oxygen concentration increases, and consequently, the monovacancies are difficult to form the complex PV with phosphorus according to the above considerations by the present inventors.

[0044] Fig. Figure 3 shows a correlation between the irradiation amount of an electron beam and the change ratio of the carrier concentration. All oxygen concentrations of the wafers are 6×10 15 Atoms / cm 3 .

[0045] Based on this Fig.3 it can be stated that the number of defects increases and the carrier concentration changes more when the irradiation amount of the electron beam increases.

[0046] As in Fig. 2 and Fig. As shown in Figure 3, the irradiation amount of the particle beam changes depending on the oxygen concentration and the irradiation amount of the particle beam. Herein, it is assumed that [VP] ∝ [VV] as described above, and thus, it can be indirectly understood that more W defects are formed in a wafer with a larger change ratio in the carrier concentration. As described above, it is possible to easily determine the VV defect concentration using the change ratio of the carrier concentration. Although the specific value for the VV defect concentration is not determined, a relative VV defect concentration between all samples can be evaluated.

[0047] By obtaining data in advance to show the relationship between the oxygen concentration or the irradiation amount of a particle beam and the change ratio of the carrier concentration, which is related to the VV defect concentration, it is possible to determine the relationship between the oxygen concentration and the VV defect concentration. This makes it possible to evaluate how much irradiation amount is required to make the irradiation damage equivalent when the oxygen concentration of a wafer is changed, for example.

[0048] It is believed that certain types of impurities may affect the formation of an irradiation defect if they are included in the wafer in addition to the above example of carbon. Therefore, it is possible to perform an even more precise evaluation by including a step of considering the influence of these impurities, if necessary, when evaluating a wafer containing a high concentration of impurities not considered in the above embodiment or a wafer with a high impurity concentration that differs greatly among samples. This consideration can be appropriately performed according to the evaluation precision, cost, time, and effort. <Zweite Ausführungsform>

[0049] As another embodiment of the present invention, it is possible to conduct a preliminary examination before the evaluation of a silicon single crystal substrate to be actually evaluated. Fig. 4 shows an example of the steps of the second embodiment. [Preliminary investigation](Step 1: Preparation of a silicon single-crystal substrate for preliminary investigation)

[0050] First, a silicon single-crystal substrate for preliminary inspection (sometimes referred to as a preliminary inspection wafer for convenience hereinafter) is prepared. For this preliminary inspection wafer, for example, the same type of wafer as the wafer to be inspected that will be prepared later can be used, except that the VV defect concentration has already been inspected. Details such as oxygen concentration can be determined as in the first embodiment, for example. Herein, this is explained by illustrating an example of preparing a plurality of preliminary inspection wafers with the same oxygen concentration, dopant species, resistivity, etc., with one wafer to be evaluated, but the number is not particularly limited. (Step 1': Evaluation of the VV defect concentration using the same silicon single crystal substrate as a preliminary investigation)

[0051] It is difficult to directly determine the amount of VV defect concentration. Therefore, it is possible to evaluate the VV defect concentration in the pre-test wafer beforehand and indirectly based on the recovery characteristic of a diode fabricated in the particle beam irradiation manufacturing step under the same conditions as the particle beam irradiation step of Step 3 in a pre-test as described later, using the same type of wafer as a pre-test wafer. Since a VV defect in a wafer serves as a carrier recombination center, more VV defect concentration defects result in a shorter annihilation time and improved recovery characteristic. Therefore, the recovery characteristic is assumed to represent the amount of VV defects formed.

[0052] Herein, diodes are fabricated by changing the irradiation amount for a number of pre-test wafers, and in each of these diodes, a recovery characteristic is previously determined as an indication of the VV defect concentration in a pre-test wafer.

[0053] It is of course possible to use parameters other than the recovery property, which correspond to the VV defect concentration. (Step 2 to Step 5: Specific resistance measurement to calculate the change ratio of the carrier concentration)

[0054] A resistivity measurement, a particle beam irradiation, and a resistivity re-measurement are performed on each pre-test wafer to determine the carrier concentrations in each pre-test wafer before and after irradiation, and each change ratio of the carrier concentration is calculated.

[0055] In step 3 (particle beam irradiation), the particle beam is irradiated under the same conditions as the diode manufacturing step in the previous step 1'. In this case, the irradiation amount differs for each pre-test wafer. The specific processes in these steps can be performed, for example, in the same manner as in the first embodiment. (Step 6: Determining a correlation)

[0056] A correlation is determined between the VV defect concentration (herein, the quality of a diode's recovery characteristic) and the change ratio of the carrier concentration in each pre-test wafer, calculated as described above. This allows the VV defect concentration (quality of the diode's recovery characteristic) to be obtained, which corresponds to each change ratio of the carrier concentration when the oxygen concentration, etc., are under the same conditions. [General inspection]

[0057] (Steps 7 to 11: Preparation of a silicon single-crystal substrate to be evaluated to calculate the carrier concentration change ratio) Then, the main test is performed. This means that an evaluation is carried out on an actual wafer to be evaluated. The wafer to be evaluated is prepared and subjected to resistivity measurement, particle beam irradiation with a predetermined irradiation amount, and further resistivity measurement. The carrier concentrations in the wafer to be evaluated are determined before and after irradiation to calculate the carrier concentration change ratio. (Step 12: Evaluation of the VV defect concentration using the correlation)

[0058] The VV defect concentration of the wafer to be evaluated (herein, the quality of a diode's recovery characteristic) is evaluated using the change ratio of the carrier concentration calculated as described above, and the correlation is determined in the preliminary test. This means that the calculated change ratio of the carrier concentration is applied to the above correlation to determine the corresponding VV defect concentration (the quality of the recovery characteristic). In this way, it is possible to easily perform a relative evaluation of the VV defect concentration using the preliminary test.

[0059] Furthermore, it is not necessary to perform the preliminary test every time. It is possible to use previously collected measurements to demonstrate a correlation between the change ratio of carrier concentration and VV defect concentration. In this case, the steps required for the preliminary test can be omitted, and the evaluation can be carried out even more simply. Example

[0060] The present invention will be described more specifically below with reference to Examples and Comparative Examples, but the present invention is not limited thereto. (Example 1)

[0061] A phosphorus-doped silicon single-crystal ingot with a diameter of 200 mm was produced using the CZ and FZ processes, respectively. The oxygen concentrations in the crystals, measured by FT-IR, were 3.5×10 17 Atoms / cm 3in the CZ crystal and 0.3×10 17 Atoms / cm 3 in the FZ crystal. Each of the resistivities was approximately 60 Ωcm. Furthermore, the carbon concentrations were approximately 5×10 14 atoms / cm 3 to reduce the occurrence of carbon-related defects that affect device properties.

[0062] First, polished wafers (PW) were prepared from the FZ crystal. Subsequently, PN diodes were fabricated from these FZ wafers. These diodes were subjected to electron beam irradiation at five levels with an accelerating voltage of 2 MV during the fabrication step.

[0063] The amount of VV defect formed was evaluated indirectly based on a recovered property of the fabricated diode, as it is difficult to directly measure the amount of VV defect formed, as described above. It is assumed that the quality of the recovery property represents the extent of the VV defect concentration. The evaluation results are shown in Table 1.

[0064] On the other hand, electron beam irradiation at 5 levels was performed under the same conditions during diode fabrication using the same PW method as the previous PW, and the resistivities before and after irradiation were determined using a four-point probe method. Based on the measurements, the carrier concentrations before and after irradiation were determined. The resulting change ratio of the carrier concentration is also shown in Table 1. [Table 1] Manufacturing process of a crystal Irradiation level Irradiation amount [ / cm 2 ] Recreational property Change ratio of carrier concentration (%) FZ 1 2E+14 failed 17,7 FZ 2 5E+14 failed 23 FZ 3 1E+15 Success 28,8 FZ 4 2E+15 Success 33,2 FZ 5 4E+15 Success 41,3

[0065] Table 1 shows the correlation between the VV defect concentration (quality of the recovery property) and the change ratio of the carrier concentration, determined as described above. It was also shown that the recovery property resulted in success when the change ratio of the carrier concentration was approximately 29% or more.

[0066] The evaluation of the VV defect concentration (evaluation of the recovery property) and the calculation of the carrier concentration change ratio were performed on CZ wafers, which are polished wafers (PBs) made of CZ crystal, in the same manner as above. The results are shown in Table 2. [Table 2] Crystal manufacturing process Irradiation level Irradiation amount [ / cm 2 ] Recreational property Change ratio of carrier concentration (%) CZ 1 2E+14 failed 8,6 CZ 2 5E+14 failed 11,2 CZ 3 1E+15 failed 17,6 CZ 4 2E+15 failed 22,9 CZ 5 4E+15 Success 29,3

[0067] As can be seen from Table 2, the recovery property results in success, as in the case of the FZ wafers, when the change ratio of the carrier concentration was approximately 29% or more.

[0068] Based on Table 1 and Table 2, it can be said that the damage amount evaluated based on the change ratio of the carrier concentration is an effective indicator. That is, it was found that evaluating the VV defect concentration based on the change ratio of the carrier concentration is an effective method in the present invention. Furthermore, it is possible to grasp the relationship between the irradiation amount and the VV defect concentration, the oxygen concentration, and so on, in each of the manufactured CZ wafers and FZ wafers, as shown in Table 1 and Table 2. (Example 2)

[0069] For each of the FZ wafers and the CZ wafer, which were manufactured separately from the previous ones (whose oxygen concentrations, etc., were the same as those manufactured to produce Table 1 and Table 2), an electron beam was irradiated at an irradiation level different from the irradiation levels described in Table 1 and Table 2. Their carrier concentration change ratios were calculated using Table 1 and Table 2 as the preliminary test results. These two change ratios were 29% or more. Therefore, any recovery characteristic was considered a success when manufacturing a PN diode by irradiating the same type of wafer as the wafers with an electron beam at the same irradiation level, considering the carrier concentration calculation results in Table 1 and Table 2 of the preliminary test.

[0070] When the PN diodes were actually fabricated, the evaluated recovery characteristics were a success as pre-evaluated. (Comparison example)

[0071] FZ wafers and CZ wafers of the same type as in Example 1 were fabricated. First, the PN diodes were fabricated by simply irradiating the FZ wafers with an electron beam at an irradiation amount as described in Table 1 (irradiation levels 1 to 5). Unlike in Example 1, no resistivity measurements were performed and no calculations were made of the carrier concentration change ratio. The relationship between the irradiation amount and the quality of the recovery characteristics was calculated. This relationship was similar to that shown in Table 1.

[0072] Consequently, in the case of manufacturing the PN diodes from the CZ wafers in the same manner as the FZ wafers, each quality of the recovery characteristic of the PN diodes based on the CZ wafers was evaluated to be the same as in the FZ wafers with the same irradiation amount.

[0073] In fact, however, the relationship between the irradiation amount and the quality of the recovery property fell to the results shown in Table 2. This means, as demonstrated, for example, by the results of irradiation levels 3 and 4, that the results in CZ wafers were, contrary to expectations, a failure, although the recovery properties were successful in FZ wafers. As with the previous discussion, the device properties in wafers with different oxygen concentrations may differ from each other due to the difference in each irradiation damage, even if the irradiation conditions are the same.

[0074] As can be seen from the evaluation results of this comparative example which considers only the irradiation condition, compared with that of Examples 1 and 2 carrying out the present invention, the present invention which considers the change ratio of the carrier concentration in evaluating the VV defect concentration as described above is effective.

Claims

[1] A method for evaluating a concentration of a defect in a silicon single crystal substrate, wherein the defect is formed in the silicon single crystal substrate by particle beam irradiation, comprising the steps of: Measuring a resistivity of the silicon single crystal substrate, followed by irradiating the silicon single crystal substrate with the particle beam, measuring the resistivity of the silicon single crystal substrate again after the irradiation; Determining each carrier concentration in the silicon single crystal substrate before and after irradiation based on the measured resistivity results before and after particle beam irradiation to calculate a change ratio of the carrier concentration; and Evaluating a concentration of a W defect based on the change ratio of the carrier concentration, wherein the VV defect consists of a silicon atom vacancy in a Frenkel pair formed by the particle beam irradiation in the silicon single crystal substrate and another silicon atom vacancy. [2] The method for evaluating a concentration of a defect in a silicon single crystal substrate according to claim 1, wherein the particle beam for irradiating the silicon single crystal substrate is an electron beam or an ion beam. [3] The method for evaluating a concentration of a defect in a silicon single crystal substrate according to claim 1 or 2, wherein a doping impurity contained in the silicon single crystal substrate is any one of phosphorus, arsenic and antimony. [4] The method for evaluating a concentration of a defect in a silicon single crystal substrate according to any one of claims 1 to 3, further comprising the steps of: Preparing a silicon single crystal substrate for preliminary testing with a pre-evaluated concentration of the W defect, separately from the silicon single crystal substrate to be evaluated; Determining a correlation based on the pre-evaluated concentration of the W defect and a change ratio of the carrier concentration before and after the particle beam irradiation with respect to the silicon single crystal substrate for preliminary examination; and Irradiation of the silicon single crystal substrate to be evaluated with the particle beam to evaluate the concentration of the W defect using the correlation.

Images