阅读说明：本技术 复合寿命的控制方法 (Composite life control method ) 是由 竹野博 于 2018-12-10 设计创作，主要内容包括：本发明提供一种复合寿命的控制方法,其通过进行由利用FZ法培养的掺氮硅单晶准备硅单晶基板的工序；实施热处理的热处理工序A；对硅单晶基板照射粒子束的粒子束照射工序；及对硅单晶基板进行热处理的热处理工序B,从而控制硅单晶基板的载流子的复合寿命,该方法的特征在于,在热处理工序A中,根据在准备工序中准备的硅单晶基板的氧浓度Co,使硅单晶基板中的氮向外扩散,从而调节硅单晶基板的氮浓度Cn,然后,进行粒子束照射工序。由此,能够确实地减小起因于掺氮FZ硅单晶基板的复合寿命的偏差,能够以高精度控制复合寿命。(The invention provides a method for controlling composite lifetime, which comprises preparing a silicon single crystal substrate from a nitrogen-doped silicon single crystal grown by FZ method; a heat treatment step A for performing a heat treatment; a particle beam irradiation step of irradiating a silicon single crystal substrate with a particle beam; and a heat treatment step B of heat-treating the silicon single crystal substrate to control the recombination lifetime of carriers in the silicon single crystal substrate, wherein in the heat treatment step A, nitrogen in the silicon single crystal substrate is out-diffused in accordance with the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step to adjust the nitrogen concentration Cn of the silicon single crystal substrate, and then, a particle beam irradiation step is performed. Thus, the variation in recombination lifetime due to the nitrogen-doped FZ silicon single crystal substrate can be reliably reduced, and the recombination lifetime can be controlled with high accuracy.)

1. A method for controlling recombination lifetime, which controls recombination lifetime of carriers of a silicon single crystal substrate by performing the following steps:

a preparation step of preparing a silicon single crystal substrate from a nitrogen-doped silicon single crystal grown by a floating zone melting method (FZ method);

a heat treatment step A of performing a heat treatment on the prepared silicon single crystal substrate;

a particle beam irradiation step of irradiating the silicon single crystal substrate after the heat treatment step A with a particle beam; and

a heat treatment step B of heat-treating the silicon single crystal substrate after the particle beam irradiation step,

the method is characterized in that,

in the heat treatment step a, nitrogen concentration Cn of the silicon single crystal substrate is adjusted by diffusing nitrogen in the silicon single crystal substrate out by the heat treatment based on the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step, and then the particle beam irradiation step is performed on the silicon single crystal substrate with the nitrogen concentration Cn adjusted.

2. The composite lifetime control method according to claim 1,

when the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step is less than 0.1ppma, the heat treatment step A is performed so that the oxygen concentration Co is less than 0.1ppmaThe silicon single crystal substrate after the heat treatment step A has a nitrogen concentration Cn of less than 2 × 10¹⁴atoms/cm³Wherein when the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step is 0.1ppma or more, the nitrogen concentration Cn of the silicon single crystal substrate after the heat treatment step A is made to be less than 2 × 10 in the heat treatment step A¹⁵atoms/cm³By performing said heat treatment so as to out-diffuse nitrogen,

then, the particle beam irradiation step is performed.

Technical Field

The present invention relates to a method for controlling the recombination lifetime of carriers in a silicon single crystal substrate.

Background

The following techniques have been used: in a power device (power device) such as an Insulated Gate Bipolar Transistor (IGBT) or a diode, a defect serving as a recombination center of a carrier is intentionally introduced into a silicon substrate, so that the recombination lifetime of the carrier is shortened, the switching speed is increased, and the switching loss is reduced.

As a method for controlling the recombination lifetime, there is a method of diffusing a heavy metal impurity such as gold or platinum in a substrate; and a method of irradiating a charged particle beam (which may be referred to as a particle beam) such as an electron beam, proton, or helium ion. In the method of diffusing a heavy metal impurity, since it is relatively difficult to control the concentration, the distribution in the depth direction, the in-plane uniformity, and the like, in recent years, charged particle beam irradiation has been increasingly used (see patent documents 1 to 3). Since many defects that become recombination centers are introduced near room temperature during irradiation with a charged particle beam and there are defect types that are unstable to heat, the target value of recombination lifetime can be obtained by further performing heat treatment (which may be referred to as recovery heat treatment) after irradiation with a charged particle beam to reduce the defects that are unstable to heat or adjust the defect concentration.

For example, patent document 3 describes that, in order to shorten the reverse recovery time of the diode, heat treatment may be performed under heat treatment conditions in which crystal defects caused by irradiation with an electron beam remain, and the recombination lifetime of carriers may be set to about 0.1 to 1 (microsecond). Patent document 3 also describes that the heat treatment conditions in this case may be set to, for example, a heat treatment temperature of 350 ℃ or higher and less than 380 ℃ and a heat treatment time of about 0.5 to 2 hours. Patent document 4 describes that when the heat treatment temperature is 400 ℃ or higher, the lifetime of the wafer irradiated with protons is about 10 μ s, and this is preferable from the point of controlling the diode forward voltage, reverse recovery loss, and the like.

Switching losses in power devices are in a trade-off relationship with steady state losses, so in order to reduce overall losses, the recombination lifetime must be tightly controlled.

In addition, as for the withstand voltage characteristics of the power device, there is a problem that the resistivity of the silicon substrate varies due to the formation of an oxygen donor in the silicon single crystal substrate, and thus, in high-performance devices, FZ (Floating Zone) silicon single crystal substrates containing almost no oxygen or silicon single crystal substrates having an extremely low oxygen concentration produced by the MCZ method in which a magnetic field is applied in the CZ method are frequently used.

In this case, nitrogen is often added to the FZ silicon single crystal substrate during crystal growth in order to prevent in-furnace discharge during crystal growth, reduce crystal defects introduced during crystal growth, and improve wafer strength (patent document 5). The concentration of nitrogen introduced into the cultured single crystal can be controlled by adjusting the atmosphere gas at the time of culturing the crystal.

Disclosure of Invention

Technical problem to be solved by the invention

However, even if the conditions for particle beam irradiation and the conditions for heat treatment after particle beam irradiation are the same, there is a problem that device characteristics vary (non-patent documents 1 to 3). Since the variation in the composite lifetime directly affects the variation in the device characteristics, it is a very important technical problem to improve the variation in the composite lifetime. In particular, in recent years, as the performance of semiconductor devices has been improved, it has been required to control the recombination lifetime with high accuracy and to reduce the variation in recombination lifetime as much as possible.

As a factor of the variation in recombination lifetime, it is suspected that a certain substance contained in the silicon substrate itself is a factor, and particularly, there is a concern about the influence of impurities of carbon or oxygen.

In particular, it has been pointed out that, in a power device in which the recombination lifetime of carriers is controlled by irradiating a silicon substrate with an electron beam or a particle beam such as helium ions, an extremely small amount of carbon of 0.05ppma or less adversely affects the device characteristics (see non-patent documents 4 to 6). It is thus considered important to reduce carbon contained in the silicon single crystal substrate as much as possible.

Further, non-patent document 1 has pointed out that even when the composite lifetime control is performed in the same manner, there is a problem that the switching loss depends on the wafer, and it is considered that, among the main defects (CsI, CiCs, or CiOi) generated by the electron beam irradiation, only the CiOi activation energy depends on the wafer, and the activation tends to be higher as the oxygen concentration is higher, so that the oxygen impurity is a main factor depending on the wafer (Cs: substitutional carbon, Ci: interstitial carbon, Oi: interstitial oxygen, I: interstitial silicon, among others).

However, it has not been actually known whether these carbon or oxygen impurities are the main factors of the deviation of the recombination lifetime.

Patent document 6 describes, as a method for controlling the recombination lifetime, a method for controlling the recombination lifetime with high accuracy by adjusting the dopant concentration of a CZ silicon single crystal substrate or adjusting the dopant concentration and the oxygen concentration, thereby obtaining the recombination lifetime equivalent to that of an FZ silicon single crystal substrate.

However, patent document 6 has a problem that the main factors and control methods of the variation in recombination lifetime in the nitrogen-doped FZ silicon single crystal substrate have not been clarified yet.

Patent document 7 describes a method of screening a silicon substrate, which can reduce the variation in recombination lifetime due to the silicon substrate and control the recombination lifetime with high accuracy, and which determines that the silicon substrate is acceptable when the ratio (L T1/L T0) of the recombination lifetime L T0 after the particle beam irradiation step to the recombination lifetime L T1 after the recovery heat treatment is equal to or less than a reference value L T2.

However, the method for screening a silicon substrate described in patent document 7 has a problem that, as shown in fig. 7 of patent document 7, the probability of failing to obtain an FZ silicon single crystal substrate having an oxygen concentration of 0.4ppma or less is very high. Further, there is no effective technique for reducing the variation in the recombination lifetime of FZ silicon single crystal substrates having an oxygen concentration of 0.4ppma or less.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a method for controlling recombination lifetime, which can reliably reduce the variation in recombination lifetime caused by a nitrogen-doped FZ silicon single crystal substrate in a process for manufacturing a power device for controlling the recombination lifetime of carriers, and can control the recombination lifetime with high accuracy.

Means for solving the problems

In order to solve the above-mentioned problems, the present invention provides a method for controlling a composite lifetime by performing a preparation step of preparing a silicon single crystal substrate from a nitrogen-doped silicon single crystal grown by a floating zone melting method (FZ method); a heat treatment step A of performing a heat treatment on the prepared silicon single crystal substrate; a particle beam irradiation step of irradiating the silicon single crystal substrate after the heat treatment step A with a particle beam; and a heat treatment step B of heat-treating the silicon single crystal substrate after the particle beam irradiation step, thereby controlling the recombination lifetime of carriers in the silicon single crystal substrate,

in the heat treatment step a, nitrogen concentration Cn of the silicon single crystal substrate is adjusted by diffusing nitrogen in the silicon single crystal substrate out by the heat treatment based on the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step, and then the particle beam irradiation step is performed on the silicon single crystal substrate with the nitrogen concentration Cn adjusted.

In this way, by controlling the recombination lifetime by adjusting the nitrogen concentration Cn of the silicon single crystal substrate in the heat treatment step a at a stage before the particle beam irradiation step based on the oxygen concentration Co of the silicon single crystal substrate prepared from the nitrogen-doped silicon single crystal grown by the FZ method (nitrogen-doped FZ silicon single crystal substrate), it is possible to reliably reduce the variation in recombination lifetime due to the silicon substrate.

In addition, in this case, it is preferable that the heat treatment step A is performed such that the nitrogen concentration Cn of the silicon single crystal substrate after the heat treatment step A is less than 2 × 10 when the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step is less than 0.1ppma¹⁴atoms/cm³When the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step is 0.1ppma or more, the nitrogen concentration Cn of the silicon single crystal substrate after the heat treatment step A is made to be less than 2 × 10 in the heat treatment step A¹⁵atoms/cm³The particle beam irradiation step is performed after the heat treatment is performed to diffuse nitrogen out.

In this manner, the user can easily and accurately select the desired target,when the oxygen concentration Co of the silicon single crystal substrate is less than 0.1ppma, the nitrogen concentration Cn of the silicon single crystal substrate is adjusted to be less than 2 × 10 after the heat treatment step A (before the particle beam irradiation step) in the heat treatment step A¹⁴atoms/cm³Or when the oxygen concentration Co of the silicon single crystal substrate is 0.1ppma or more, the nitrogen concentration Cn of the silicon single crystal substrate is adjusted to be less than 2 × 10 after the heat treatment step A (before the particle beam irradiation step) in the heat treatment step A¹⁵atoms/cm³Thus, the variation in recombination lifetime due to the silicon single crystal substrate can be further reliably reduced.

Effects of the invention

In the method for controlling recombination lifetime according to the present invention, the recombination lifetime can be controlled with high accuracy by adjusting the nitrogen concentration of the silicon single crystal substrate based on the oxygen concentration Co of the nitrogen-doped FZ silicon single crystal substrate before the particle beam irradiation step, and thus variation in recombination lifetime due to the silicon single crystal substrate can be reduced reliably. Further, since the nitrogen concentration of the silicon single crystal substrate before the particle beam irradiation step is adjusted by the heat treatment step after the silicon single crystal substrate is prepared, it is not necessary to lower the nitrogen concentration required for preventing the discharge in the furnace or suppressing the occurrence of crystal defects, etc. when the silicon single crystal is grown, and it is possible to avoid the deterioration of the productivity and quality of the silicon single crystal. Further, the present invention can control the recombination lifetime with high accuracy even when the nitrogen concentration is different when growing a silicon single crystal in controlling the recombination lifetime of the nitrogen-doped FZ silicon substrate, and is therefore particularly suitable for use of the nitrogen-doped FZ silicon substrate in a power device.

Drawings

FIG. 1 is a flowchart showing a method for controlling recombination lifetime of a silicon single crystal substrate according to the present invention.

Fig. 2 is a graph showing the relationship between the combined lifetime L T measured in the experimental example and the heat treatment time in the heat treatment step B.

FIG. 3 is a graph showing the relationship between the recombination lifetime L T and the bulk nitrogen concentration (bulk nitrogen concentration) after the heat treatment step A (before the particle beam irradiation step) when the oxygen concentration measured in the experimental example is 0.1ppma or more (FZ-A, FZ-B).

FIG. 4 is a graph showing the relationship between the recombination lifetime L T and the bulk nitrogen concentration after the heat treatment step A (before the particle beam irradiation step) when the oxygen concentration measured in the experimental example is less than 0.1ppma (FZ-C, FZ-D).

Detailed Description

Hereinafter, embodiments of the present invention will be described in detail, but the present invention is not limited thereto.

As described above, in the conventional technique, the recombination lifetime of carriers is controlled by adjusting the conditions of particle beam irradiation and the conditions of heat treatment after particle beam irradiation, and in this case, there is a problem that the recombination lifetime varies due to some factor caused by the silicon single crystal substrate.

As a result of diligent research, the inventors of the present invention have found that the recombination lifetime when a silicon single crystal substrate is subjected to particle beam irradiation and subsequent heat treatment varies even when the carbon concentration, which has been conventionally considered to be a main factor of variation, is substantially the same, and the recombination lifetime strongly depends on the nitrogen concentration of the silicon single crystal substrate, and have completed the present invention.

That is, it was found that the silicon single crystal produced by the MCZ method does not need to be produced by adding nitrogen, and therefore, there is no problem of variation in recombination lifetime due to nitrogen, but the FZ method requires nitrogen to be added to produce a silicon single crystal, which causes variation in recombination lifetime of an FZ silicon single crystal substrate, and that the nitrogen concentration dependency of recombination lifetime varies depending on the oxygen concentration. The present invention has been accomplished by solving the problems specific to the silicon single crystal substrate produced by the FZ method.

That is, the present invention is a method for controlling a composite lifetime, which comprises a preparation step of preparing a silicon single crystal substrate from a nitrogen-doped silicon single crystal grown by a floating zone melting method (FZ method); a heat treatment step A of performing a heat treatment on the prepared silicon single crystal substrate; a particle beam irradiation step of irradiating the silicon single crystal substrate after the heat treatment step A with a particle beam; and a heat treatment step B of heat-treating the silicon single crystal substrate after the particle beam irradiation step, thereby controlling the recombination lifetime of carriers in the silicon single crystal substrate,

in the heat treatment step a, the nitrogen concentration Cn of the silicon single crystal substrate is adjusted by diffusing nitrogen in the silicon single crystal substrate out by the heat treatment based on the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step, and then the particle beam irradiation step is performed on the silicon single crystal substrate with the nitrogen concentration Cn adjusted.

Hereinafter, the method for controlling the composite lifetime of the present invention will be described with reference to fig. 1.

First, a silicon single crystal substrate is prepared from a nitrogen-doped silicon single crystal grown by the FZ method (S1 in fig. 1). The specification (diameter, thickness, resistivity, etc.) of the silicon single crystal substrate prepared here is not particularly limited in the present invention. For example, the specification may be set according to the requirements required for the semiconductor device.

The method for preparing the silicon single crystal substrate is not particularly limited in the present invention. For example, a silicon single crystal substrate may be prepared by cutting a silicon wafer from a silicon single crystal, subjecting the silicon wafer to a chemical etching process to remove a cutting damage, and then performing chemical mechanical grinding.

Next, a heat treatment step a of heat-treating the silicon single crystal substrate is performed (S2 in fig. 1).

In the heat treatment step a, the nitrogen concentration of the silicon single crystal substrate is adjusted before the particle beam irradiation step by outdiffusing nitrogen in the silicon single crystal substrate according to the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step.

The conditions of the heat treatment step A are determined based on the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step, and when the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step is less than 0.1ppma, it is preferable that the nitrogen concentration Cn after the heat treatment step A is less than 2 × 10¹⁴atoms/cm³The conditions of the heat treatment step A are determined, and when the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step is 0.1ppma or more, it is preferable that the nitrogen concentration Cn after the heat treatment step A is less than 2 × 10¹⁵atoms/cm³Conditions of the heat treatment step A were determined.

The upper limit of the oxygen concentration Co of the silicon single crystal substrate prepared from the nitrogen-doped silicon single crystal grown by the FZ method is not particularly limited, and may be, for example, less than 1 ppma.

The lower limit of the nitrogen concentration Cn is not limited at any oxygen concentration Co. Here, the nitrogen concentration after the heat treatment step a is a nitrogen concentration in a region where the recombination lifetime is adjusted. In a manufacturing process of a semiconductor device, when a silicon single crystal substrate is thinned by grinding or the like before a particle beam irradiation process, a nitrogen concentration in a recombination lifetime region after thinning is adjusted.

The heat treatment step a is not limited to a single heat treatment step, and may be performed a plurality of times. The process may be performed before or during the process for manufacturing the semiconductor device. Further, a combination of the heat treatment step before the manufacturing step of the semiconductor device and the heat treatment step in the manufacturing step is also possible.

In the heat treatment step a, it is important to adjust the nitrogen concentration Cn of the silicon single crystal substrate before the particle beam irradiation step by diffusing nitrogen in the silicon single crystal substrate out of the silicon single crystal substrate prepared in the preparation step based on the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step, and particularly, to set such that the relationship between the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step and the nitrogen concentration Cn of the silicon single crystal substrate before the particle beam irradiation step satisfies the above-described conditions.

In this way, the variation in recombination lifetime due to the nitrogen-doped FZ silicon single crystal substrate can be reliably reduced, and the recombination lifetime can be controlled with high accuracy. On the other hand, if the heat treatment is performed under the same conditions regardless of the oxygen concentration Co of the silicon single crystal substrate without considering the adjustment of the nitrogen concentration before the particle beam irradiation step, the variation in recombination lifetime due to the nitrogen-doped FZ silicon single crystal substrate may become large.

The heat treatment step a may be performed under any conditions as long as the conditions are determined in accordance with the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step so as to allow out-diffusion of nitrogen in the silicon single crystal substrate, and particularly, it is preferable that the relationship between the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step and the nitrogen concentration Cn of the silicon single crystal substrate before the particle beam irradiation step satisfies the above-described conditions. For example, in the case of performing the primary heat treatment, the temperature is set to 1050 ℃ or higher and the time is set to 2 hours or longer, and the temperature is determined according to the oxygen concentration Co of the silicon single crystal substrate. The higher the heat treatment temperature in the heat treatment step a, the higher the diffusion rate of nitrogen, and therefore the higher the efficiency. The higher the temperature in the heat treatment step A, the longer the time, the lower the nitrogen concentration of the silicon single crystal substrate.

In the present invention, the lower limit of the nitrogen concentration of the silicon single crystal substrate after the heat treatment step a is not limited, and therefore the upper limits of the heat treatment temperature and time in the heat treatment step a are not limited. May be determined in consideration of productivity or cost, prevention of slip dislocation, and other properties.

Patent documents 6 and 7 disclose that, when an oxide passivation layer is used to suppress surface recombination in the measurement step of measuring the recombination lifetime, an oxide film can be formed on the surface of the silicon single crystal substrate before the particle beam irradiation step. The oxide film can be formed by a heat treatment in an oxidizing atmosphere, and the conditions for the heat treatment for forming the oxide film may be, for example, 900 to 1100 ℃ inclusive and 10 to 60 minutes inclusive. It is further described that instead of using an oxide film passivation layer, a chemical passivation layer without a heat treatment may be used.

Even in the oxide film forming heat treatment for the oxide passivation layer described in patent document 6 and patent document 7, nitrogen in the silicon single crystal substrate diffuses outward. However, these heat treatment conditions are not determined according to the oxygen concentration Co of the silicon single crystal substrate, and the conditions of the oxide film forming heat treatment described in patent documents 6 and 7 are not sufficient for achieving the object of the heat treatment step a in the present invention, and as a result, variation in recombination lifetime due to the silicon single crystal substrate cannot be reduced.

In order to strictly control the recombination lifetime, it is important to determine the relationship between the oxygen concentration and the nitrogen concentration of the silicon single crystal substrate at the stage of particle beam irradiation. However, since the diffusion rate of oxygen in the silicon single crystal substrate is slow, the bulk oxygen concentration after the heat treatment step a (before the particle beam irradiation step) is almost the same as the initial oxygen concentration at the stage prepared in the preparation step. Therefore, in the present invention, the oxygen concentration Co of the silicon single crystal substrate prepared in the preparation step can be regarded as the same as the oxygen concentration Co of the silicon single crystal substrate after the heat treatment step a (before the particle beam irradiation step).

Next, a particle beam irradiation step (S3 in fig. 1) of irradiating the silicon single crystal substrate after the heat treatment step a with a particle beam and a heat treatment step B (S4 in fig. 1) of performing a heat treatment after the particle beam irradiation step are performed.

The conditions of the particle beam irradiation step and the heat treatment step B to be carried out here are not particularly limited in the present invention, and may be determined so as to satisfy the requirements required for the semiconductor device, and for example, in the particle beam irradiation step, the value of 1 × 10 may be used¹⁴～3×10¹⁵/cm²The dose of (3) and an acceleration voltage of 0.5 to 2MeV are irradiated with an electron beam as a particle beam. In addition, protons or helium ions may also be irradiated.

The heat treatment step B is a heat treatment step for eliminating a part of recombination centers generated in the particle beam irradiation step and recovering the recombination lifetime, and for example, the temperature may be set to 300 to 400 ℃, the time may be set to 10 to 60 minutes, and the atmosphere may be nitrogen, oxygen, hydrogen, or the like.

In the method for controlling recombination lifetime of the present invention, the recombination lifetime can be controlled with high accuracy, and variations in recombination lifetime due to the silicon single crystal substrate can be reliably reduced.

In the present invention, based on the findings obtained by the following experiments, the variation in lifetime due to the silicon single crystal substrate is reliably reduced by using the above-described method for controlling recombination lifetime.

(Experimental example)

A silicon single crystal substrate (which may also be referred to as an FZ substrate) is prepared from a plurality of nitrogen-doped silicon single crystals grown by the FZ method. For comparison, a silicon single crystal substrate (which may also be referred to as an MCZ substrate) was prepared from a non-nitrogen-doped silicon single crystal grown by the MCZ method. No matter what kind of silicon single crystal substrate, the diameter is 200mm, the crystal plane orientation is (100), the thickness is 720-730 mu m, the conductivity type is n-type (phosphorus-doped), and the resistivity is 63-71 omega cm.

The oxygen concentration, carbon concentration and nitrogen concentration of each silicon single crystal substrate are shown in the following table. The oxygen concentration was measured by an infrared absorption method (conversion factor: JEIDA), and the carbon concentration and the nitrogen concentration were measured by Secondary Ion Mass Spectrometry (SIMS). Each concentration shown in the table is a concentration (which may be referred to as an initial concentration) at a stage (before the heat treatment step a) where the silicon single crystal substrate is prepared.

[ Table 1]

An FZ substrate (FZ-A, FZ-B) having an oxygen concentration of 0.1ppma or more is produced from a silicon single crystal grown by the FZ method using a silicon single crystal ingot grown by the CZ method as a starting material. Further, an FZ substrate (FZ-C, FZ-D) having an oxygen concentration of less than 0.1ppma was produced from a silicon single crystal grown by the FZ method using a conventional polycrystalline silicon ingot as a raw material.

Subsequently, the prepared silicon single crystal substrate is subjected to heat treatment (heat treatment step a). The heat treatment conditions (temperature/time) were set to 1000 ℃/1 hour, 1050 ℃/2 hours, 1050 ℃/4 hours, 1150 ℃/1 hour, 1150 ℃/2 hours, 1150 ℃/4 hours, and the atmosphere was set to oxygen. However, the MCZ substrate is not doped with nitrogen, and thus is subjected to heat treatment only at 1000 ℃ for 1 hour. In any of the heat treatment conditions of the heat treatment step A, the silicon single crystal substrate is inserted into a furnace maintained at 900 ℃, the temperature is raised from 900 ℃ to each temperature at a rate of 5 ℃/min, the temperature is maintained for each time, and then the temperature is lowered from each temperature to 900 ℃ at a rate of-3 ℃/min, and the silicon single crystal substrate is taken out of the furnace.

Next, the silicon single crystal substrate after the heat treatment step A was irradiated with an electron beam (particle beam irradiation step). in this case, the dose of the electron beam was set to 1 × 10¹⁵/cm²The acceleration voltage of the electron beam is set to 2 MV.

In addition, in another silicon single crystal substrate subjected to the heat treatment step a simultaneously with the silicon single crystal substrate subjected to the electron beam irradiation, the silicon single crystal substrate was polished from the front surface side and removed to about half the thickness (about 360 μm), and then the nitrogen concentration was measured by SIMS, whereby the bulk nitrogen concentration (nitrogen concentration in the vicinity of the center in the thickness direction of the substrate) after the heat treatment step a (before the particle beam irradiation step) was determined.

Subsequently, the silicon single crystal substrate irradiated with the electron beam was subjected to a recovery heat treatment (heat treatment step B), the heat treatment temperature was set to 360 ℃, the atmosphere was set to nitrogen, and the time was set to a range of 0 to 45 minutes, and then the recombination lifetime was measured (L T).

The recombination lifetime was measured by using a Microwave photoconductive decay method (Microwave Photo conductive decay method: μ -PCD method). In the μ -PCD method, first, an optical pulse having energy larger than the band gap of the silicon single crystal is irradiated, and an excessive amount of carriers are generated in the silicon single crystal substrate. Although the conductivity of the wafer increases due to the generated excess carriers, the excess carriers disappear by recombination as time passes thereafter, resulting in a decrease in conductivity. This change is detected as a temporal change in the reflected microwave power (an excessive carrier decay curve), and analyzed to determine the recombination lifetime.

The recombination lifetime is defined as the time until the concentration of excess carriers decays to 1/e (═ 0.368) by recombination (JEIDA-53-1997, "method for measuring recombination lifetime by microwave reflection photoconductive decay method of silicon wafer"). In this experiment, the time (1/e lifetime) until the reflected microwave power decayed to 1/e of the irradiation time of the light pulse was obtained as the composite lifetime. The measuring apparatus may be a commercially available measuring apparatus.

Fig. 2 shows the relationship between the composite lifetime L T and the heat treatment time in the heat treatment step B, with respect to the conditions (temperature/time) of the heat treatment stepA, the temperature/time is 1000 ℃/1 hour in fig. 2, (B) is 1050 ℃/1 hour in fig. 2, (C) is 1050 ℃/2 hours in fig. 2, (D) is 1050 ℃/4 hours in fig. 2, (e) is 1150 ℃/1 hour in fig. 2, (f) is 1150 ℃/2 hours in fig. 2, and (g) is 1150 ℃/4 hours in fig. 2, the difference in symbols in fig. 2 indicates the difference in the silicon single crystal substrate, ○ indicates the case of FZ-A, △ indicates the case of FZ-B, □ indicates the case of FZ-C, ◇ indicates the case of FZ-D, and MCZ indicates the case of MCZ.

As can be seen from fig. 2, when the condition of the heat treatment step a was 1000 ℃/1 hour [ fig. 2 (a) ], the change in the heat treatment time of L T with respect to the heat treatment step B greatly varied depending on the difference in the silicon single crystal substrate, which indicates that L T was greatly deviated due to the silicon single crystal substrate even when the conditions of the electron beam irradiation and the subsequent recovery heat treatment were the same, furthermore, the amount of change in the heat treatment time of L T with respect to the heat treatment step B became larger as the nitrogen concentration of the FZ substrate before the heat treatment step a was higher and the oxygen concentration was lower, which indicates that L T was largely deviated due to the nitrogen concentration or the oxygen concentration even when the carbon concentration of the FZ substrate was almost the same, and indicates that the nitrogen concentration and the oxygen concentration in the FZ substrate were L T deviation, further, it was found that the change in the heat treatment time of L T with respect to the heat treatment step B was smaller as the heat treatment temperature of the heat treatment step a was higher, the deviation of L T due to the FZ substrate was very small, and the difference in the MCZ substrate was not very small as well as the MCZ difference in the MCZ substrate was L.

Next, in the case of an FZ substrate (FZ-A, FZ-B) having an oxygen concentration of 0.1ppma or more, FIG. 3 shows the relationship between the recombination lifetime L T and the bulk nitrogen concentration after the heat treatment step A (before the particle beam irradiation step). regarding the conditions (temperature/time) of the heat treatment step B, (A) is an athermal treatment, (B) is 360 ℃/15 minutes, (c) is 360 ℃/30 minutes, (d) is 360 ℃/45 minutes. the difference in symbols in FIG. 3 indicates the difference in FZ substrates, ○ indicates the case of the FZ-A, △ indicates the case of the FZ-B, and for comparison, the case of the MCZ substrate not doped with nitrogen is indicated byA dotted line, and in FIG. 3, when the bulk nitrogen concentration is the lower limit of detection (4 × 10-B)¹³atoms/cm³) Hereinafter, for convenience, the bulk nitrogen concentration is referred to as 2 × 10¹³atoms/cm³And (6) drawing.

As is clear from FIG. 3, when the oxygen concentration is 0.1ppma or more, the heat treatment time in the heat treatment step B becomes longer [ FIG. 3 (c) and FIG. 3 (d)]L T then becomes longer with bulk nitrogen concentration, but as long as the bulk nitrogen concentration is 2 × 1015atoms/cm³Hereinafter, L T is not dependent on bulk nitrogen concentration, that is, when the oxygen concentration is 0.1ppma or moreThe use of the heat treatment step A allows the bulk nitrogen concentration to be adjusted to 2 × 10 before the particle beam irradiation step¹⁵atoms/cm³Hereinafter, the deviation of L T due to the difference in FZ substrates can be minimized, and the difference of L T due to the difference in FZ substrates and MCZ substrates that are not doped with nitrogen can be minimized.

Next, in the case where the oxygen concentration is less than 0.1ppma (FZ-C, FZ-D), FIG. 4 shows the relationship between the recombination lifetime L T and the bulk nitrogen concentration after the heat treatment step A (before the particle beam irradiation step), and in the conditions (temperature/time) of the heat treatment step B, (a) is no heat treatment, (B) is 360 ℃/15 minutes, (C) is 360 ℃/30 minutes, and (D) is 360 ℃/45 minutes, the difference in symbols in FIG. 4 indicates the difference in FZ substrates, □ indicates the case of the FZ-C, ◇ indicates the case of the FZ-D, and in addition, for comparison, the case of the MCZ substrate not doped with nitrogen is indicated by a letter, and in FIG. 4, when the bulk nitrogen concentration is the lower limit of detection (4 × 10-D), the case of FIG. 4¹³atoms/cm³) Hereinafter, for convenience, this will be referred to as 2 × 10¹³atoms/cm³And (6) drawing.

As is clear from fig. 4, when the oxygen concentration is less than 0.1ppma, the longer the heat treatment time in the heat treatment step B, the stronger the dependency of the bulk nitrogen concentration of L T becomes, but when the bulk nitrogen concentration is 2 × 10¹⁴atoms/cm³That is, when the oxygen concentration is less than 0.1ppma, the bulk nitrogen concentration is adjusted to 2 × 10 before the particle beam irradiation step by using the heat treatment step A¹⁴atoms/cm³Hereinafter, the deviation of L T due to the difference in FZ substrates can be minimized, and the difference of L T due to the difference between the FZ substrate and the MCZ substrate not doped with nitrogen can be minimized.

Further, as is clear from the results shown in fig. 3 and 4, when the oxygen concentration is 0.3ppma or less, the bulk nitrogen concentration of the FZ substrate may be adjusted to 2 × 10 in the heat treatment step a before the particle beam irradiation step¹⁴atoms/cm³Hereinafter, the recombination lifetime can be controlled with high accuracy regardless of the difference in the raw material or the difference in the nitrogen concentration when the FZ silicon single crystal is grown, and the difference between the FZ method and the MCZ method without nitrogen doping.

As described above, by adjusting the nitrogen concentration of the silicon single crystal substrate before the particle beam irradiation step based on the oxygen concentration of the silicon single crystal substrate whose recombination lifetime is to be controlled, the recombination lifetime can be controlled with high accuracy, and thus variations in recombination lifetime due to the silicon single crystal substrate can be reliably reduced. Further, since the nitrogen concentration of the silicon single crystal substrate before the particle beam irradiation step is adjusted in the heat treatment step a after the silicon single crystal substrate is prepared, it is not necessary to lower the nitrogen concentration required for preventing the discharge in the furnace or suppressing the occurrence of crystal defects, etc. when the silicon single crystal is grown, and it is possible to avoid the deterioration of the productivity and quality of the silicon single crystal.

The reason why the combined lifetime of the electron beam irradiation and the subsequent heat treatment depends on the nitrogen concentration is considered as follows.

When a silicon substrate is irradiated with a high-energy particle beam, silicon atoms at lattice positions are ejected, and interstitial silicon (hereinafter, also referred to as I) and vacancies (hereinafter, also referred to as V) as shells thereof are generated. The excessively generated I and V are unstable when they are individually separated, and therefore they are recombined (V + I → 0), or I and V are aggregated with each other, or they react with light element impurities contained in the silicon substrate to form a complex. Further, the cluster of I or V or the complex of I or V and the light element impurity forms a deep level in the band gap of silicon, and functions as a recombination center of carriers, thereby reducing the recombination lifetime.

As a defect associated with vacancy V, it is known that V reacts with substitutional phosphorus Ps to form VP (V + Ps → VP). Further, V may react with interstitial oxygen Oi to form VO (V + Oi → VO), and V may further react with VO to form V₂O(V+VO→V₂O). Furthermore, V reacts with each other to form VV (V + V → VV). In the presence of nitrogen, V also reacts with N to form VN (V + N → VN). Since V competes with P, O or N, respectively, VN is likely to form at higher nitrogen concentrations, other complexes associated with V may not be readily formed.

On the other hand, as defects associated with interstitial silicon I, it is known that I reacts with substitutional boron Bs to form interstitial boron Bi (I + Bs → Bi), and further, Bi reacts with Oi to form BiOi (Bi + Oi → BiOi). In the presence of carbon, I reacts with substitutional carbon Cs to form interstitial carbon Ci (I + Cs → Ci), and Ci reacts with Oi and Ci and Cs to form CiOi and CiCs (Ci + Oi → CiOi, Ci + Cs → CiCs). In addition, I also reacts with each other to form I clusters (I + … → In). In the presence of nitrogen, V reacts with N, so that the recombination of V with I is inhibited, and thus a complex related to I may be easily formed.

Since the reaction of I or V with light element impurities depends on the respective absolute and equilibrium concentrations, it is very complicated to guess which complex is dominant. When the heat treatment is further performed, the complex disappears or changes in form, which further complicates the process.

As shown in the above experimental examples, it is considered that when the nitrogen concentration of the silicon single crystal substrate becomes high, a thermally unstable compound is easily formed by the high energy particle beam irradiation, and therefore the compound is easily disappeared by the subsequent heat treatment, and the recombination lifetime after the heat treatment is increased.