DE112010002568B4 - Method for producing a silicon single crystal - Google Patents

Abstract

A method of producing a silicon single crystal in which a silicon single crystal is grown by a CZ method, comprising the steps of: dipping a seed crystal in a silicon melt and starting to pull the seed crystal to grow the silicon single crystal from the silicon melt allow; Forming a straight body portion of the silicon single crystal to be used as the silicon wafer, and removing an axial dislocation generated in the step of dipping and extending in the crystal pulling direction, before the step of forming a straight body portion, the step of removing the axial body Dislocation by asymmetric heating with respect to the center axis of the silicon single crystal is performed, and the fluctuation in the state of the solid-liquid interface between the silicon melt and the silicon single crystal is increased and the direction in which the axial dislocation extends, from the crystal growth direction is moved away.

Description

TECHNICAL AREA

The present invention relates to a technique suitable for using a method of producing a silicon single crystal.

STATE OF THE ART

As a method for producing a silicon single crystal which is a material of a silicon wafer, growth methods by a Czochralski method (hereinafter referred to as CZ method) are known.

Processes called CZ process are used in the production of a silicon single crystal. For example, a starting material polysilicon contained in a quartz crucible is melted using heating means, for example, a resistance heater, in a decompressing inert gas (Ar) atmosphere. A seed is dipped in the surface of a silicon melt near the melting point after melting (step of contacting seed with the melt), the melt temperature is adjusted to such a degree that the seed has affinity for the silicon melt, and when the seed crystal Affinity for it, a Impfkristallverengung in the diameter of about 5 mm is performed while the seed is pulled to remove dislocations within the seed crystal (necking step). After a necking step in which the seed crystal narrowing is performed, the crystal diameter is increased in a cone shape while the melt temperature and the pulling rate are adjusted to obtain a product diameter (shoulder forming step). When the crystal diameter reaches the product diameter, a part serving as a product is grown vertically by a constant length (step of forming a straight body), after which the crystal diameter in a conical shape (tapering step) is decreased, and the process is performed Separation of the melt, when the diameter becomes sufficiently small, finished.

In the CZ method described above, when a silicon seed crystal is brought into contact with a raw material silicon melt, dislocations (thermal shock dislocation and misfit dislocation) due to thermal shock are generated in the high-density seed crystal , The dislocation acts as a factor to hinder monocrystallization of a crystal to be grown. Therefore, the dislocation is annihilated and then it is necessary to grow a crystal. Since the cross-sectional area of a crystal to be grown generally becomes smaller, generation of dislocation is reduced. In addition, the growth of the dislocation once generated is easily stopped. For this reason, in the single crystal growth by the CZ method and the like, a method called a Dash necking method in which a neck portion having a small sectional area grows over a predetermined length is widely used is left and the dislocation is annihilated.

In the neck-necking process, more specifically, in a seed crystal necking step in which the silicon seed crystal is brought into contact with the raw material silicon melt and melted, the neck portion becomes the diameter thereof about 4 mm, shaped to be 50 to 200 mm in length. Thereafter, it is possible to carry out a process of thickening the single crystal through a shoulder portion until a predetermined diameter is reached. Thereby, it is possible to annihilate in the neck portion described above a dislocation propagated from a slip dislocation introduced into the seed crystal and to pull a dislocation-free silicon single crystal.

In addition, when a monocrystal is pulled with great weight, the dislocation introduced at the time of bringing into contact with a melt is suppressed by applying no thermal shock. Techniques that are similar to this are in JP H11-60379 A . JP 2008-87994 A such as JP H9-2898 A disclosed.

Alternatively there is, as in JP H11-12082 A discloses to produce a method for producing a large-diameter portion enlarged in diameter in front of a shoulder shape and to hold the large-diameter portion while keeping a crystal.

In addition, the disclosed JP 2001-130996 A a method in which only the neck section is allowed to grow and in the necking step an experimental proof that dislocation-free, is performed.

The US 2003/0047130 A1 describes a method for eliminating dislocations in a neck section of a large diameter silicon monocrystalline ingot, the method comprising controlling heat transfer at the melt and solid interface, thereby reducing dislocations over a reduced axial length in the neck section of a large diameter silicon monocrystalline ingot which was grown according to the Czochralski method.

In addition, the reveals US 2003/0209186 A1 a method of producing a silicon single crystal according to the Czochralski method wherein, by rapidly decreasing the pulling rate at least once while growing a neck portion of the silicon single crystal ingot, achieving zero dislocation growth in a large diameter neck portion within a comparatively short neck Cutting length so that ingots of considerable weight and large diameter can be made.

In addition, the disclosed US 2006/0144320 A1 a method and system for use in combination with a crystal growth apparatus for growing a monocrystalline ingot according to the Czochralski method, wherein a time-varying external magnetic field is applied to the melt during aiming of the ingot, and the magnetic field is selectively adjusted to provide tensile forces in the ingot To cause melt, which control the melt flow rate, while the ingot is pulled out of the melt.

From the JP H09-249 482 A are also known corresponding CZ method.

SUMMARY OF THE INVENTION

[Technical problem]

In the in JP H11-60379 A . JP 2008-87994 A such as JP H9-2898 A In the techniques disclosed, there is a problem that a completely dislocation-free state can not be realized. In particular, if the neck diameter is reduced to, for example, about 4 mm, even if growth up to a length of about 500 mm is performed, which is considered sufficient in the prior art to eliminate the dislocation, there are cases where in which the dislocation remains in places up to the straight body portion.

Besides, as occurs in JP 2001-130996 A is disclosed, when the seal within a furnace is broken to remove the neck, contamination by contaminants from the outside of the furnace or by heavy metal contaminants from the moving part of a production apparatus, and thus it is difficult to recover the starting material remaining in the quartz crucible to use. As the diameter of the wafer increases, silicon starting material having a large weight should not be discarded as waste. Therefore, the above-described method can not be applied to the actual production, and thus a method that can be modified is required.

When a moving part is placed inside a furnace, as in JP H11-12082 A Also, this movable part acts as a source of contamination, and as a result, there arises a problem that the movable part can not actually be used in pulling the silicon single crystal.

In addition, there is a problem that the state of generation and removal of the dislocation is not accurately detected.

Moreover, the mechanism of generation and removal of the dislocation is not well understood. For this reason, it was not understood how the dislocation could be completely removed. In particular, dislocation (axial dislocation) which extends in the crystal pulling direction and can not be removed by changing the crystal diameter by a dash necking method and the like exists in the dislocations generated in the neck portion. Axial dislocation (difficult to remove dislocation) exists over the entire length of the straight body segment, and it was unclear how it could be removed.

• 1. einen dislokationsfreien Zustand zu verwirklichen, indem eine Dislokation, die in einem Impfkristall-Schmelze-Kontakt-Schritt erzeugt wurde, zuverlässig zu entfernen;
• 2. den Zustand zur Erzeugung und Entfernung einer Dislokation, insbesondere einer axialen Dislokation, genau zu bestimmen;
• 3. eine Beziehung zwischen dem Verhalten einer Dislokation in einem Neck-Abschnitt und den Ziehbedingungen genau zu bestimmen;
• 4. Die Ziehbedingungen, die geeignet sind, einen dislokationsfreien Zustand einzustellen, genau zu bestimmen.

This invention has been accomplished in view of such circumstances and an object of the same is to achieve:

• 1. to realize a dislocation-free state by reliably removing a dislocation generated in a seed crystal melt contact step;
• (2) accurately determine the state of creation and removal of a dislocation, in particular axial dislocation;
• 3. accurately determine a relationship between the behavior of a dislocation in a neck section and the pulling conditions;
• 4. Determine the drawing conditions that are suitable for setting a dislocation-free state.

the solution of the problem

When growing a large-diameter silicon crystal having a diameter of 300 mm or larger, particularly about 450 mm, the diameter of the neck portion (dash-neck portion) becomes a problem for carrying a large weight. Thus, the inventors have investigated the performance of secure growth by increasing the diameter of the portion, referred to as the neck portion in the prior art, while maintaining a dislocation-free state. Here, the production of a single-crystal silicon monocrystal of several hundred kilograms to several tons is envisaged by a simple method without employing a special device serving as a source of contamination within the furnace.

In growing the large-diameter silicon single crystal, a method called "dash-necking method" is normally used. In the Dash-Necking method, the diameter at the time of growing the silicon single crystal is reduced and dislocation (thermal shock dislocation) introduced when a silicon seed crystal and a silicon melt come into contact with each other , away. In this case, a reduction in the diameter of the neck portion removes the dislocation in a simple manner. In addition, in order to grow a crystal of great weight, it is necessary to increase the diameter due to the load carried by the neck portion. The thickness of the neck portion is set to a size calculated by these relationships. In the silicon single crystal having a diameter of 300 mm, the crystal load can be supported with a diameter of about 5 mm. However, if the diameter of the silicon monocrystal is 450 mm, there may be a problem that the load can not be carried by the neck portion when the crystal is grown to a required length.

In addition, as described above, as another method, there is a method in which no thermal shock dislocation is introduced by reducing the surface temperature difference between the seed crystal and the melt (silicon melt) to thicken the neck portion. However, when the neck portion was not narrowed or the seed crystal was not set to a high temperature, a case occurred in which the growth of the silicon single crystal was achieved.

Consequently, the inventors believe that the silicon crystal can be grown with great weight by a simple process that does not require special methods or devices as long as such conditions can be clarified, and have means for accurately examining the behavior of dislocation within the Neck section associated with the dragging thereof searched.

The inventors have investigated how the dislocation behaves in the necking step. They also investigated why the axial dislocation can not be removed. They assume that, as well as the behavior of the dislocation can be clarified, it is possible to grow a silicon crystal in which a dislocation is completely removed by a simple method which does not require special methods or devices. They then searched for means of scrutinizing the behavior of the dislocation in the neck section associated with the pulling of it.

In the prior art, an analysis of the dislocation state in the neck portion was performed using X-ray topography. In this method, the cross section of the neck portion was cut to a thickness of about 1.5 mm, and then an X-ray transmission observation was performed by etching the surface with a mixed acid and the like. Therefore, only images of a certain cross section could be directly observed, the images spaced 1.5 mm apart were assembled, and a change in dislocation state accompanied with crystal growth was taken out. For this reason, only the dislocation oblique to the drawing axis, which is a central axis of the silicon monocrystal in the drawing direction, could be observed.

However, when a white X-ray beam was used as a high-energy beam of a level as disclosed in SYNCHROTRON RADIATION INSTRUMENTATION: Ninth International Conference to Synchrotron Radiation Instrumentation, AIP Conference Proceedings, Vol. 879, pp. 1545-1549 (2007), it was found that when the neck section was viewed, the dislocation was annihilated in a loop form. In addition, it was found that the dislocation existed at levels other than the normal {111} plane.

That is, the inventors clarified the existence of the above-mentioned axial dislocation for the first time by such a method. The axial dislocation was observed as a pit in the straight body portion of the silicon single crystal and in the wafer to which it was sliced. However, it was not recognized that the observed pit is the axial dislocation.

This axial dislocation can be detected in the wafer in clear differentiation from other wells by the following method.

First, in the silicon single crystal from which the wafer is cut as a wafer, the crystal habit line is appropriately formed and the dislocation-free state can be realized, and the silicon single crystal is also recognized as a monocrystal by other testing. It is possible to identify the axial dislocation in a wafer surface by performing a strain measurement using optical inspection means. As the inspection means, the deformation inspection device (SIRD; registered trademark) SirTec manufactured by JENA WAVE, which is capable of visually observing the state of deformation by internal stress, can be used.

In this case, a manifestation process is performed by heating to 900 to 1250 ° C x 1 s or more at a rate of elevating and decreasing the temperature of 10 to 300 ° C / s on the wafer to be evaluated. Thereafter, the surface properties of the wafer, including the manifested deformation, are evaluated. This manifestation process is intended to cause a large temperature difference to occur by rapid heating in the plane of the wafer and generate thermal stress caused by this temperature difference. Therefore, no heating for a long time is required, and the desired effect can be exhibited even by heating for a short period of about 1 second.

Since the axial dislocation exists over both sides of the wafer, the strain is increased by the thermal stress generated by the manifestation process in which rapid heating is performed for a very short period of time as mentioned above. On the other hand, in the investigation of the surface contamination by means of a laser surface inspection apparatus and the like, a pit similar to the axial dislocation is detected. However, in the dislocation which has no such length as to pass through both sides of the wafer, it is considered that the strain due to the thermal stress generated by the manifestation process is not increased. That is, it is believed that in the dislocation where the size thereof is less than 10% of the thickness of the wafer, the deformation by the thermal stress generated by the manifestation process does not increase.

Although the axial dislocation is shown by the above-described manifestation process, other pits and the like can not be shown by this manifestation process for the above reason. 11 FIG. 12 shows an observation example of the wafer having the manifested axial dislocation and an observation example of the wafer having no axial dislocation. FIG.

The viewing example, which is in 11 (a) is an example in which three axial dislocations are generated, and the observation example shown in FIG 11 (b) is an example in which the axial dislocation is not generated. Meanwhile, it is considered that the deformation observed at the edge of the wafer is generated by holding the wafer at the time of the heat treatment.

The axial dislocation is generated in the step of contacting the seed crystal with the melt. The axial dislocation is a dislocation that exists continuously in the growth direction of the crystal until the necking step, the step of forming a shoulder portion (increase in diameter), and the step of growing the straight body portion. At the time of wafer fabrication, axial dislocation exists only in a specific region of the wafer. In addition, the axial dislocation can be performed by methods known in the art at the time of growing the neck portion, for example, a method of adjusting the diameter increase, the diameter reduction and the length of the neck section to a certain level or higher, such as the dash necking method.

As in 12 shown comprises a silicon single crystal 60 in which the drawing is finished, a seed crystal T, a diameter reduction portion (neck portion) N0, when the silicon single crystal with a seed boundary T00 interposed therebetween is grown, a shoulder portion 60a for performing the diameter enlargement, a straight body portion 60b and an end portion 60c (Taper portion).

When the neighborhood of the seed T increases, as in 13 is shown, thermal shock dislocations Jn and misfit dislocations Jm are generated in the diameter reduction portion (neck portion) N0. The thermal shock dislocations Jn are generated at the seed crystal T-side by the thermal shock at the time of contacting the seed crystal T with the silicon melt, and are transmitted via the neck portion N0 side of the silicon single crystal to be grown 60 taken. The misfit dislocations Jm are generated when a mismatch of the lattice constant in the seed crystal T and the neck portion N0 of the silicon single crystal 60 is present.

Among the thermal shock dislocations Jn and the misfit dislocations Jm, the axial dislocation J is a dislocation growing in the growth direction (axial direction). Specifically, as stated in 12 is shown in which a 110 wafer W, consisting of the single crystal 60 is shown, the axial dislocation J is observed in the direction 10 o'clock and the direction 4 o'clock at the time of placing a cut in the direction of 12 o'clock on each of the surfaces. More specifically, with the section being a 100 direction adapted to 0 °, the axial dislocation becomes in a region in the range of 120 ° to 135 ° and a region in the range of 315 ° to 350 °, symmetrical with respect to this Region and the central axis of the silicon single crystal 60 observed. In addition, the axial dislocation only in the region of a region J1 with existence of an axial dislocation, formed of a region in which the above-mentioned regions 90 ° and 45 ° about the central axis of the silicon single crystal 60 are turned, watched. That is, the axial dislocation J extends in the growth direction (central axis direction) of the silicon single crystal 60 in the region J1. In the drawing, the region J1 shows only a region within the ranges 120 ° to 135 ° and 315 ° to 350 °.

Meanwhile, for the silicon single crystal grown by a predetermined length or size, there is a possibility that the axial dislocation is annihilated. However, at present, in the manufacturing facility, axial dislocation is first detected when the data of the wafer for which a piece of the pulled single crystal has been sliced is adjusted. The axial dislocation is normally over the entire length of a silicon single crystal and often passes through the silicon single crystal.

Here, it is preferable to use white X-ray topography by a white X-ray as high-energy radiation for measuring dislocation in the silicon single crystal to be grown by a CZ method. In particular, an x-ray beam having a continuous energy spectrum of 30 keV to 1 MeV, about 40 keV to 100 keV and 50 keV to 60 keV is used. Alternatively, an X-ray having a wavelength of about 0.001 nm to 0.25 nm is used. In addition, an X-ray beam was used in which the number of photons passing through a slot of 1 mm × 1 mm in a position 44 mm apart from a light source has a distribution as shown in FIG 5 is shown. An in 5 (a) The graph shown illustrates a state of an X-ray that is an example of high-energy radiation in the invention and the distribution of the number of photon energy of an X-ray. The in 5 (b) The graph shown illustrates an X-ray beam, which is an example of high-energy radiation in the invention, and the distribution of the photon number with respect to wavelength. In addition, the diameter of the X-ray beam is preferably 0.01 to 1 times the diameter of the neck portion which is the object to be measured.

Such a high energy white x-ray beam is used, whereby it is possible to three-dimensionally image the behavior of, for example, dislocation within the silicon single crystal without performing a destructive inspection, for example, slicing, and positional information about the Dislocation and the property of the dislocation first observed.

As in 4 is shown, taking into account a diffraction point (for example, diffraction point 004), a CT image is formed by a topograph image by rotating the silicon single crystal. Here, a three-dimensional position information of the dislocation line is obtained. In addition, at the same time as a white x-ray topography, as in 6 shown and carried out the properties (for example Burgers vector) of the dislocation are determined.

X-ray topography is also referred to as X-ray diffraction micrography, and is a method of viewing a spatial distribution of errors in a non-destructive manner. When an X-ray beam is caused to continuously fall on the silicon single crystal, a plurality of diffraction images, the to a diffraction point, observed. The X-ray topography analyzes the diffraction points.

Here, two-dimensional detectors which are used in consideration of the diffraction image comprising: an imaging plate in which a stimulable by light luminous phenomenon of a fluorescent plate, an X-ray film, a core plate and a lichtstimulierbarer phosphor (BaFBr: Eu ²⁺⁾ is used; an X-ray television using an image pickup tube in which a PbO film or an amorphous Se-As film sensitive to an X-ray is used as a photoconductive surface, and a CCD type X-ray detector incorporating a charge coupled CCD device uses. With respect to an X-ray detector used in the invention, it is preferable to consider the following characteristics as spatial resolution or dynamic range.
Detective Quantum Efficiency
Dynamic range
Linear strength region
Dead time and pulse loss
Light receiving area and position resolution
Non-uniformity of sensitivity
Non-linearity (or image distortion) of the position
energy resolution
time resolution
Ability to measure in real time
operational stability

In pulling the silicon single crystal, even if furnace elements such as the same carbon are used in the same device (CZ furnace) in the same manner, the generation state of the dislocation and the behavior of dislocation become dependent on the conditions at the time of drawing of the silicon single crystal. Here, the conditions at the time of drawing include, for example, the pulling rate, the height (distance) from the liquid level of the molten silicon melt to a heat shield element, the time of keeping the seed crystal in contact with the molten silicon at the liquid level of the molten silicon melt. From this, by using means for observing the behavior of dislocation within the silicon single crystal by high-energy radiation, it is possible to control the production conditions of the neck portion (the dislocation removal portion and the dislocation-free portion) at the time of using a production apparatus having a structure in which different furnaces or different carbon materials are combined to determine.

Therefore, the observation is performed by the X-ray topography, whereby it is possible to eliminate the production conditions in which the axial dislocation is generated and to specify the production conditions in which the axial dislocation is not generated. The manufacturing conditions herein include devices for pulling and producing a single crystal, means, parts, states, and methods in conjunction with all the characteristics of a single crystal and a wafer in addition to the so-called control parameters.

With a large-diameter silicon crystal having a diameter of 300 mm or larger, especially about 450 mm, the diameter of the neck portion (dash-neck portion) becomes a problem for carrying a large weight. Consequently, the present inventors have studied an increase in the diameter of the portion called a neck portion in the prior art, with the aim of producing the silicon monocrystal of large weight on the order of several hundred kilograms to several tons in the state in which there is no axial dislocation, can grow safely by a simple method without placing a special device acting as a source of contamination in the furnace.

In growing a large-diameter silicon single crystal, the neck portion is formed by reducing the diameter of the silicon single crystal at an early stage of growth, usually by using the above-mentioned dash-necking method. Thereby, the dislocation (the thermal shock dislocation and the misfit dislocation) introduced when the silicon seed and the silicon melt are brought into contact with each other in the neck section. In this case, a reduction in the diameter of the neck portion removes the dislocation in a simple manner. However, in order to grow a crystal of great weight, it is necessary to increase the diameter in view of the load borne by the neck portion. The thickness of the neck section is calculated from these relationships. In the crystal, which has a diameter of about 300 mm, the crystal load can be carried with a diameter of about 5 mm. However, when the crystal is grown to a necessary crystal length in the crystal having a diameter of 450 mm, there may arise the problem that the load can not be carried. For this reason, the inventors have studied an increase in the diameter of the neck as described above.

The inventors have found that if the diameter of the neck portion is increased in this way, the case may occur in which the above-mentioned axial dislocation J exists. In the prior art, such axial dislocation has not been confirmed.

• 1) Als Hauptprämisse, die Dislokation wandert durch die Gleitebene ({111}-Ebene im Fall von Silicium);
• 2) normalerweise wird die Dislokation in die {111}-Ebene eingeführt;
• 3) die Dislokation kann selten durch eine andere Ebene wandern, ohne dass sie in die {111}-Ebene eingeführt wird; außerdem bewegt sich selbst in diesem Fall die Dislokation normalerweise unverzüglich zu der {111}-Ebene;
• 4) die meisten der Dislokationen, die in den Silicium-Einkristall, der durch ein CZ-Verfahren produziert wird,

eingeführt werden, werden zur Zeit des Inkontaktbringens des Impfkristalls mit der Siliciumschmelze erzeugt.It has been proved that the following characteristics exist in the behavior of thermal shock dislocation Jn and misfit dislocation Jm. Here, the behavior of thermal shock dislocation Jn and misfit dislocation Jm indicates the propagation direction and the like in the process of growing the silicon single crystal in which the neck portion, the shoulder portion and the straight body portion are pulled.

• 1) As the main premise, the dislocation travels through the slip plane ({111} plane in the case of silicon);
• 2) normally the dislocation is introduced into the {111} plane;
• 3) the dislocation can rarely move through another plane without being introduced into the {111} plane; moreover, even in this case, the dislocation normally moves immediately to the {111} plane;
• 4) most of the dislocations produced in the silicon single crystal produced by a CZ method

are introduced at the time of contacting the seed crystal with the silicon melt.

Here, in Table 1, as the slip plane to which the dislocation moves and the type of the crystal to be drawn, the angle between the crystal planes is indicated. {h ₁ , k ₁ , l ₁ } {h ₂ , k ₂ , l ₂ } 100 110 111 100 90.00 ° 45.00 ° 90.000 54.74 ° 110 45.00 ° 90.00 ° 60,00 ° 90,000 35.26 ° 90,000 111 54.57 ° 35.26 ° 90.00 ° 70.53 °

When the silicon single crystal for producing a <100> wafer is produced by drawing from the silicon melt as shown in FIG 3 is shown, the dislocation generated by the thermal shock or misfit moves through the {111} plane, which is a sliding plane of the silicon single crystal, in the necking step, in which the diameter of the silicon single crystal is reduced , The dislocation moving through the slip plane reaches the surface of the silicon single crystal. It is assumed that the surface of the silicon single crystal serves as the end point of the dislocation. However, according to the above-described X-ray topography, it is known that the initial dislocation introduced by the thermal shock or by misfit can be extended in the growth direction of the crystal. This phenomenon is generated regardless of the size of the diameter of a product or the diameter of the neck portion at the time of necking. At this time, the dislocation in the direction perpendicular to the growth interface of the crystal extends between the silicon single crystal and the silicon melt without passing through the {111} plane of the slip plane, and proceeds in the growth direction of the crystal without being increased to move the sliding plane. In this way, the dislocation proceeding in the growth direction of the crystal becomes an axial dislocation.

Thereafter, the axial dislocation does not reach the surface of the silicon single crystal even if the diameter of the silicon single crystal is increased to the diameter of the product. Such a silicon single crystal having an axial dislocation has a crystal habit line similar to that of silicon. Single crystal having no axial dislocation and in the state of an ingot can not be distinguished from a non-defective product in some cases. When the wafer obtained by slicing the ingot is examined, the axial dislocation is detected as an error.

In addition, a case may occur in which the dislocation which extends in the growth direction of the silicon single crystal becomes the {111} plane during the increase in the diameter of the silicon single crystal or during the production of a part in which the diameter reaches the diameter of the product and which serves as a product, moves, and then it becomes an axial dislocation.

If the dislocation introduced by thermal shock or misfit and extending parallel to the growth direction of the crystal can be moved to the {111} plane in the necking step, the dislocation in the neck portion can be removed. That is, even if the neck portion is narrowed more than necessary or the seed crystal is not set at a temperature higher than necessary, the process may proceed to the step of shouldering, in which the diameter of the silicon single crystal in one State is increased, in which the dislocation in the neck section is removed. For this reason, the product is not affected by the dislocation generated at an early stage of crystal growth. Thereby, it is possible to suppress the number of times that the silicon single crystal is remelted, in which the axial dislocation is generated at an early stage of crystal growth, and the time of re-melting of the silicon single crystal.

Accordingly, the inventors believe that, if such conditions can be clarified, the heavy-weight silicon crystal having no axial dislocation can be grown at a low cost, and so have completed the present invention.

• 1) Der Silicium-Einkristall wird senkrecht zu der Kristallwachstumsoberfläche, die eine Grenzfläche zwischen dem Silicium-Einkristall und der Siliciumschmelze ist, wachsen gelassen. Aus diesem Grund kann der Fall eintreten, in dem die Dislokation, die zur Zeit der Bildung des Neck-Abschnitts erzeugt und eingeführt wurde, sich nicht zu der {111}-Ebene bewegt, sondern sich in Richtung senkrecht zu der Kristallwachstumsoberfläche bewegt.
• 2) Daher ist die Erfindung ein Verfahren zur Bewegung der Dislokation, die nicht in der {111}-Ebene existiert, zu der {111}-Ebene in dem Stadium, in dem der Neck-Abschnitt gebildet wird. In diesem Verfahren wird ein Mittel zur Änderung der Form der Grenzfläche zwischen dem Silicium-Einkristall und der Siliciumschmelze zur Zeit der Bildung des Neck-Abschnitts eingeführt. Die Röntgen-Topographie von Synchrotron-Strahlungseinrichtungen wird bei der Zustandsbestätigung verwendet.

The things that the inventors consider extremely important are the following.

• 1) The silicon single crystal is grown perpendicular to the crystal growth surface, which is an interface between the silicon single crystal and the silicon melt. For this reason, there may be the case where the dislocation generated and introduced at the time of formation of the neck portion does not move to the {111} plane but moves in the direction perpendicular to the crystal growth surface.
• 2) Therefore, the invention is a method of moving the dislocation that does not exist in the {111} plane to the {111} plane in the stage where the neck portion is formed. In this method, a means for changing the shape of the interface between the silicon single crystal and the silicon melt is introduced at the time of formation of the neck portion. The X-ray topography of synchrotron radiation devices is used in the condition confirmation.

A method for producing a silicon single crystal according to the invention, wherein the silicon single crystal is grown by a CZ method, comprises: dipping a seed crystal in a silicon melt and starting to pull the seed crystal to form the silicon single crystal from the silicon melt to grow; Forming a straight body portion of the silicon single crystal to be used as the silicon wafer, and removing an axial dislocation generated in the step of dipping and extending in the crystal pulling direction, before the step of forming a straight body portion, the step of removing the axial body Dislocation by asymmetric heating with respect to the center axis of the silicon single crystal is performed, and the fluctuation in the state of the solid-liquid interface between the silicon melt and the silicon single crystal is increased and the direction in which the axial dislocation extends, from the crystal growth direction is moved away.

These steps solve the above problems.

Furthermore, the invention provides the method according to claim 5.

In the step of removing an axial dislocation, the axial dislocation can be removed by expanding in the outward direction or in the manner of looping in accordance with the drawing of the crystal.

In the step of removing an axial dislocation, the shape of the solid-liquid interface between the silicon melt and the silicon single crystal may be deformed from the melt surface, thereby moving the axial dislocation extension direction away from the crystal growth direction.

In the step of removing an axial dislocation, the portion where the tangent plane of the solid-liquid interface between the silicon melt and the silicon single crystal is parallel to the surface of the melt may be made to fluctuate from the center position, thereby extending the direction of expansion the axial dislocation away from the crystal growth direction.

In the step of removing an axial dislocation, the fluctuation in the solid-liquid interface state between the silicon melt and the silicon single crystal is increased to thereby move the axial dislocation expanding direction away from the crystal growth direction.

In the step of removing an axial dislocation, the rotation of the silicon single crystal with respect to the silicon melt can be reduced to thereby increase the fluctuation in the solid-liquid interface state.

In the step of removing an axial dislocation, the magnetic field applied to the silicon melt can be reduced to thereby increase the fluctuation in the solid-liquid interface state.

In the step of removing an axial dislocation, the fluctuation in the solid-liquid interface state can be increased by hot water level vibration applied to a quartz glass crucible containing the silicon melt.

In the step of removing an axial dislocation, the silicon single crystal is heated asymmetrically with respect to the growth axis to thereby increase the fluctuation of the state of the solid-liquid interface.

In the step of removing an axial dislocation, the temperature distribution may be formed so that the solid-liquid interface between the silicon melt and the silicon single crystal is asymmetric with respect to the growth axis, thereby removing the propagation direction of the axial dislocation away from the crystal growth direction.

In the step of removing an axial dislocation, the silicon single crystal may be heated asymmetrically with respect to the growth axis to thereby asymmetrically adjust the solid-liquid interface with respect to the growth axis.

Also disclosed herein is a non-inventive method of producing a silicon monocrystal in which, in a step of removing an axial dislocation, a dislocation removal section is provided which dislocates the dislocation generated in the step of immersion by propagating outward or by annihilation in a loop shape corresponding to that shown in FIG Drawing the crystal comprises: a diameter enlarging step in which a shoulder portion is drawn to form a diameter to the necessary diameter of the straight body portion after the step of removing an axial dislocation; a dislocation behavior information acquiring step in which information on the dislocation behavior, which is a three-dimensional behavior of the dislocation, is detected in the dislocation removing section by irradiating the dislocation removing section with high-energy radiation; a step of determining that dislocation-free determining that the dislocation is removed based on the dislocation behavior information and a step of setting the pulling conditions in the dislocation-free performing conditions based on the determination that dislocation-free are performed in the step of Determination that dislocation free, be determined.

Another non-inventive method for producing a silicon single crystal disclosed herein, wherein in the step of removing an axial dislocation, a dislocation removing portion corresponding to the dislocation generated in the step of immersion by propagating outward or by annihilation in a loop shape removed from the pulling of the crystal comprises: a diameter enlarging step of pulling a shoulder portion enlarging the diameter to the necessary diameter of a straight body portion after the step of removing an axial dislocation; a dislocation behavior information acquiring step in which information of the dislocation behavior, which is a three-dimensional behavior of dislocation, is detected in the dislocation removal section by irradiating the dislocation removal section with high-energy radiation, and a dislocation-free determination step in which it is determined based on the dislocation behavior information the dislocation is removed. Of the Diameter enlargement step is started on the basis of determining that dislocation-free in the step of determining that dislocation-free.

In the above-described method for producing a silicon single crystal, high energy radiation is in the energy range of 40 keV to 70 keV, and the dislocation removal portion can be in a rotating state with high energy radiation when the number of revolutions is 0.1 to 30 rpm to be irradiated.

A silicon single crystal is produced by one of the manufacturing methods described above.

A silicon wafer is made of the silicon single crystal described above.

The method for producing a silicon single crystal according to the invention, in which a silicon single crystal is grown by a CZ method, includes the step of immersion, the step of forming a straight body, and the step of removing an axial dislocation such as mentioned above. This makes it possible for the first time to remove the axial dislocation for which no removal method is known in the prior art.

In the step of removing an axial dislocation, it is possible to remove the axial dislocation before the diameter enlarging step in which the shoulder portion is pulled by moving the propagation direction of the axial dislocation away from the crystal growth direction. For this reason, it is possible to pull the silicon single crystal in which there is no dislocation extending to the straight body portion as an axial dislocation. Therefore, it is possible to produce a wafer in which no axial dislocation exists from the straight body portion of the silicon single crystal having no axial dislocation.

In the step of removing an axial dislocation, the dislocation may be removed by propagating outward or by annihilation in a loop shape in accordance with the pulling of the crystal. That is, the dislocation is moved to the sliding plane whose direction is different from the crystal growth direction. The dislocation may be caused to migrate to the outside or be annihilated with another dislocation. This makes it possible to prevent the axial dislocation from being generated. Therefore, a single crystal in which there is no dislocation extending as an axial dislocation to the straight body portion can be drawn from the silicon melt. Thus, it is possible to exclude a non-perfect crystal, which is referred to in the art as monocrystalline, but in practice has an axial dislocation, and to pull a silicon single crystal which has no axial dislocation.

In the step of removing an axial dislocation, the shape of the solid-liquid interface between the silicon melt and the silicon single crystal may be deformed from the surface of the melt, thereby moving the propagation direction of the axial dislocation away from the crystal growth direction.

The propagation of the dislocation does not move to the {111}, which is a slip plane, but moves to the direction perpendicular to the crystal growth surface, so that the crystal first grows in the direction perpendicular to the crystal growth surface. That is, as in 16 (a) is shown in the solid-liquid interface K between the silicon melt 3 and the grown single crystal 6 the dislocation extends in the normal direction of the tangent plane in which the dislocation j is used as a contact. When the solid-liquid interface K is parallel to a liquid level 30 that is, the solid-liquid interface K is in a planar state as shown in FIG 16 (a) is shown, the dislocation j extends in the direction indicated by arrows in the melt 3 , Therefore, the dislocation j of the axial direction (crystal growth direction) shown in the vertical direction in the drawing extends in the direction of crystal growth. When the dislocation j extends in the direction of crystal growth as it is, the position of the dislocation j in the diameter direction in the diameter increasing portion of the silicon monocrystal changes, but is not eliminated. Thus, the dislocation j extending in the crystal growth direction may become an axial dislocation provided to extend in the crystal growth direction even in the straight body portion.

The process of deforming the shape of the solid-liquid interface between the silicon melt and the silicon single crystal from the surface of the melt includes a process of forming a solid-liquid interface K1 between the silicon melt 3 and the grown silicon single crystal 6 in a downward convex shape, as in 16 (b) is shown. Furthermore it includes, as in 16 (c) is shown a process of forming a solid-liquid interface K2 between the silicon melt 3 and the grown silicon single crystal 6 in an upward convex shape.

As a result, the dislocation j extends as in 16 (b) is shown in the normal direction of the tangent plane in which the dislocation j is used as a contact in the solid-liquid interface K1 corresponding to the pulling of the crystal. As in 16 (c) 2, the dislocation j additionally extends in the normal direction of the tangent plane in which the dislocation j is used as the contact in the solid-liquid interface K2 in accordance with the pulling of the crystal.

Therefore, each of the dislocations j extends as shown in FIG 16 (b) is shown, in the outward direction of the silicon single crystal according to the pulling of the crystal. Alternatively, a variety of dislocations j, as in 16 (c) is shown, collected in the direction of the central axis of the silicon single crystal. As a result, the dislocations can be removed by escape to the outside of the silicon single crystal or by annihilation with another dislocation. Thus, it is possible to prevent the dislocation j from becoming an axial dislocation, that is, a difficult-to-dislodge dislocation.

• (a) Wenn die Grenzflächenform flach ist, erstreckt sich die Dislokation in vertikaler Richtung (der Mittelachsenrichtung und der vertikalen Richtung des Silicium-Einkristalls).
• (b) Wenn die Grenzflächenform nach unten gerichtet konvex ist, erstreckt sich die Dislokation in Richtung des äußerenen Umfangs des Silicium-Einkristalls.
• (c) Wenn die Grenzflächenform nach oben gerichtet konvex ist, erstreckt sich die Dislokation in Richtung der Innenseite des Silicium-Einkristalls.

The relationship between the shape of the solid-liquid interface and propagation direction of the dislocation is summarized as follows.

• (a) When the interface shape is flat, the dislocation extends in the vertical direction (the central axis direction and the vertical direction of the silicon single crystal).
• (b) When the interface shape is convex downward, the dislocation extends toward the outer periphery of the silicon single crystal.
• (c) When the interface shape is convex upward, the dislocation extends toward the inside of the silicon single crystal.

These forms of solid-liquid interface can be regulated by the pull rate. That is, as the pulling speed is increased, the shape of the interface becomes convex upward. As the pull rate is reduced, the shape of the interface becomes convex downward. In addition, more accurate control can be performed by changing the intensity of the applied magnetic field, the rotation speed of the crucible, the rotation speed of the crystal, and the like.

In addition, the diameter of the crystal may fluctuate with a change in the pulling rate. When the pulling speed is decreased to prevent the axial dislocation from being generated by removing the dislocation, it is possible to remove the dislocation by escaping toward the outside of the crystal and prevent the axial dislocation from being generated. In this case, there is no problem that the diameter is reduced to such an extent that it is capable of holding the crystal to be drawn. In addition, when the pulling speed is increased, the dislocation can be removed by annihilation toward the central axis, or can be removed by escape to the outside as it is, and thus the axial dislocation can be prevented from being generated. In this case, there is no problem that the diameter is increased to such an extent that raw material is generated as waste. For this reason, the range of fluctuation in the diameter of the crystal is not particularly limited as long as the pulling of the straight body portion of the silicon single crystal is not affected.

In addition, deviation of the propagation direction of the dislocation from the pulling direction (axial direction) of the silicon single crystal by changing the pulling speed is also advantageous for removing the dislocation.

In the step of removing an axial dislocation, the portion where the tangent plane of the solid-liquid interface between the silicon melt and the silicon single crystal is parallel to the surface of the melt can be made to fluctuate from the center position to the propagation direction of the axial direction Dislocation away from the crystal growth direction. Both the solid-liquid interface K1, which is formed in a downward convex shape, in 16 (b) and the solid-liquid interface K2, which is formed in upwardly directed convex shape, shown in FIG 16 (c) have shapes that are rotationally symmetric when using the central axis of the crystal as the axis of rotation. Thus, they have shapes which are axisymmetric with respect to the axis line of the central axis of the silicon single crystal. The dislocation may proceed in accordance with the center axis line of the silicon single crystal. That is, the dislocation may be in the crystal center. In In this case, it is considered, as mentioned above, that even if the level of the convex shape of the solid-liquid interface mold is changed to some extent, the propagation direction of dislocation of the crystal center was not changed from the central axis direction.

The process of causing the portion where the tangent plane of the solid-liquid interface between the silicon melt and the silicon single crystal is parallel to the surface of the melt to fluctuate from the center position includes: a process of forming the solid-liquid Interface K3 between the silicon melt 3 and the grown single crystal 6 in a downward convex shape, as in 17 (a) is shown; a process in which the shape of the solid-liquid interface K3 is caused to fluctuate so that it is not axisymmetric, so that the portion Kd in which the direction of the normal line of the tangent plane becomes a vertical direction, that is, the central crystal axis direction is away from Kc, which is the crystal center position; a process in which the solid-liquid interface K4 between the silicon melt 3 and the grown single crystal is formed in an upwardly directed convex shape as shown in FIG 17 (b) and a process of causing the shape of the solid-liquid interface K to fluctuate such that it is not axisymmetric, so that the portion Kd in which the direction of the normal line of the tangent plane becomes a vertical direction, that is, the crystal center axis direction, away from Kc, which is the crystal center position.

The control of the deviation of the shape of the solid-liquid interface K3, which in 17 (a) or the solid-liquid interface K4 shown in FIG 17 (b) of the axial symmetry is shown: a process of adjusting the rotational ratio of the silicon melt 3 to the grown crystal 6 to zero, that is, stopping the crystal rotation with respect to the melt; a process of removing a magnetic field applied during the pulling; a process of distorting the temperature distribution of the grown crystal so as not to be symmetrical with the center axis of the crystal, and a process of changing the state of the silicon melt from a state suitable for the growth of the crystal symmetrical to the central axis of the crystal.

The fluctuation is generated from the state in which the convex distribution of the silicon melt is axisymmetric to the central axis of the crystal by reducing or stopping the rotation of the crystal with respect to the melt. The symmetry of the convex distribution is disturbed and thus crystal growth does not become axisymmetric. As a result, the crystal is grown although the silicon single crystal does not have a uniform diameter. That is, in the solid-liquid interface, the temperature distribution which is axisymmetric to the central axis of the silicon single crystal is disturbed. The portion in which the direction of the normal line of the tangent plane of the solid-liquid interface becomes a vertical direction, that is, the center axis direction of the crystal can be moved to a position away from the center of the crystal.

The fluctuation is followed from the state where the convexion in the silicon melt is regulated by the magnetic field by stopping the application of the magnetic field during the pulling. The growth of the crystal does not become axisymmetric to the central axis of the crystal. As a result, the crystal is grown, though the crystal does not have a uniform diameter. That is, in the solid-liquid interface, the temperature distribution which is axisymmetric to the central axis of the silicon single crystal is disturbed. The portion where the direction of the normal line of the solid-liquid interface tangency plane becomes a vertical direction, that is, the central axis direction of the crystal can be moved to a position away from the center of the crystal.

In producing the temperature distribution which is not axisymmetric to the crystal, only an unbalanced portion (a portion of the diameter direction) of the grown crystal is heated and cooled, or heated or cooled. Specifically, only a specific portion in the vicinity of the solid-liquid interface in the crystal that is rotated is irradiated with a laser beam, and heating is performed in this portion. In addition, in the furnace for pulling the silicon single crystal, a gas flow is conducted in the vicinity of the surface of the silicon melt in the diameter direction of the crystal, and only one side of the crystal is cooled. A temperature regulating agent that heats or cools a part of the crystal may develop the following configurations: a configuration in which the temperature controlling agent is rotated following the rotation of the crystal; a configuration in which on-off switching can be performed in accordance with the rotation of the crystal and a configuration in which a position is irradiated with a laser beam or a position for ejecting a cooling gas by rotation can be regulated. Such temperature regulating means may be mounted in a fixed position.

The temperature regulating agent does not thermally shield the silicon melt at a part of the crystal, and non-axisymmetric temperature distribution is built up in the crystal. For example, a notch rotated in synchronism with the crystal may be disposed near the lower end of a heat shield disposed inside the furnace for pulling the silicon single crystal which reduces the heat radiated from the silicon melt onto the crystal becomes. More specifically, a pulling device in which an opening to be opened and closed can be disposed at the lower end of the heat shield, and a rotating means for rotating the heat shield is arranged in synchronism with the rotation of the crystal.

In order to change the state of the silicon melt from the state that is suitable for crystal growth axially symmetric to the central axis of the silicon single crystal, it is considered to bring about a state where heating of the quartz glass crucible for receiving the silicon melt deviates from the axial symmetry , For example, a position that is heated may be rotated in accordance with the crucible being rotated.

In addition, the state of a part of the quartz glass crucible, for example, about 1/4 of the inner wall in the circumferential direction, is caused to be different from the other parts, and thus the occurrence of pulsation, which is considered to be a cause of vibration of the Melting levels (liquid levels) of the silicon melt is made higher. Specifically, in the inner wall, in the range of 10 cm from the upper end of the crucible and in the area of one quarter in the circumferential direction, the bubble content in the depth (thickness position) of 0.5 mm to 1 mm from the surface can be increased by about 30% (FIG. 25 to 35%) higher than the bubble content of the inner wall surface in the other areas.

In addition to the above, it is possible to adopt one of the methods for changing the shape of solid-liquid interface from axis symmetry by changing parameters such as pulling rate, applied magnetic field strength, crucible rotation speed, crystal rotation speed, Heating state of the crystal and the silicon melt to adjust.

In the step of removing an axial dislocation, the fluctuation of the state of the solid-liquid interface between the silicon melt and the silicon single crystal may be increased to thereby move the direction of expansion of the axial dislocation away from the crystal growth direction. As mentioned above, the fluctuation of the solid-liquid interface works so that the growth direction of the dislocation is directed away from the growth direction of the crystal by various means. The dislocations are removed or annihilated with each other by escaping to the outside of the crystal, whereby it is possible to prevent the axial dislocation from being formed.

In the step of removing an axial dislocation, the rotation of the silicon single crystal with respect to the silicon melt can be reduced to thereby increase the fluctuation in the solid-liquid interface state.

In the step of removing an axial dislocation, the strength of a magnetic field applied to the silicon melt can be reduced to thereby increase the fluctuation in the solid-liquid interface state.

In the step of removing an axial dislocation, the fluctuation in the solid-liquid interface state can be increased by melt level vibration generated on the quartz glass crucible for accommodating the silicon melt.

In the step of removing an axial dislocation, the silicon single crystal may be asymmetrically heated with respect to the growth axis to thereby increase the fluctuation in the solid-liquid interface state.

In the step of removing an axial dislocation, the temperature distribution will be such that the solid-liquid interface between the silicon melt and the silicon single crystal is asymmetric with respect to the growth axis to move the propagation direction of the axial dislocation away from the crystal growth direction.

In the step of removing an axial dislocation, the silicon single crystal is asymmetrically heated with respect to the growth axis to thereby asymmetrically adjust the solid-liquid interface with respect to the growth axis.

Also disclosed herein is a non-inventive method of producing a silicon single crystal in which, in the step of removing an axial dislocation, a dislocation removal portion which generates the dislocation generated in the step of immersion by propagating outward or by annihilation in loop form removed, formed according to the pulling of the crystal, the method comprising: a diameter enlarging step of pulling a shoulder portion increasing in diameter to the necessary diameter of the straight body portion after the step of removing an axial dislocation; a dislocation behavior information acquiring step in which information on dislocation behavior, which is a three-dimensional behavior of the dislocation, is detected within the dislocation removal section by irradiating the dislocation removal section with high-energy radiation; a step of determining that dislocation-free determining that the dislocation is removed based on the dislocation behavior information and a step of setting the pulling conditions where the conditions for dislocation-free performing the drag based on the determination are dislocation-free , in the step of determining that is dislocation free, to be determined.

This makes it possible to clarify the behavior of the dislocation for which the behavior in the prior art was not clarified. In addition, it is possible to annihilate the dislocation by escaping to the outside with the growth of the dislocation removal portion (dislocation-free portion). Alternatively, the dislocation is annihilated in a loop shape and made a dislocation-free state, and then the drawing conditions for increasing the diameter of the silicon single crystal can be easily developed.

At this time, in the step of determining that dislocation-free, the dislocation state is determined from the acquired information. For this reason, it is possible to accurately determine the behavior of the dislocation within the dislocation removal section (dislocation-free section). Thereby, it is possible to effectively use the dislocation removal by creating a loop, which is unknown in the prior art. For this reason, it is necessary in the prior art to reduce the diameter of the dislocation-free portion, which is called a neck portion, more than necessary. Even if the weight of the crystal pulled out of the silicon melt increases, the load can thereby be carried by the dislocation-free section. Accordingly, it is possible to draw out of the silicon melt a silicon single crystal having a larger diameter than ever before. In addition, in the prior art, a long period of time as a production period and an increase in the number of operations have been caused by performing unnecessary process steps with respect to the step of immersion and the like. However, according to the above-described method, it is possible to pull a large weight silicon single crystal having a diameter of about 450 mm, which has no axial dislocation and is in a dislocation-free state without a long time as a production time and a long time Increase in the number of steps to be caused.

In the present invention, moreover, there is disclosed a method of manufacturing a silicon single crystal not according to the present invention, wherein in the step of removing an axial dislocation, a dislocation removing portion which removes the dislocation generated in the step of immersion by expanding in the outward direction or by annihilation in a loop shape, is formed in accordance with the pulling of the crystal, the method comprising: a diameter enlarging step of pulling a shoulder portion enlarging the diameter to the necessary diameter of the straight body portion after the step of removing an axial dislocation; a dislocation behavior information acquisition step in which information on the dislocation behavior, which is a three-dimensional behavior of dislocation, is detected in the dislocation removal section by irradiating the dislocation removal section with high-energy radiation, and a determination of dislocation-free determination based on the dislocation behavior information that the dislocation is removed. The diameter increasing step is started on the basis of the determination that is dislocation-free in the step of determining that dislocation-free.

Thereby, it is possible to observe the behavior of the dislocation generated at the time of pulling in real time by a nondestructive inspection, which is impossible to perform in the prior art. For this reason, after confirming that the dislocation in the dislocation-removing portion is removed (dislocation-free state), it is possible to pull the shoulder portion (diameter-enlarging portion) and the straight body portion. Therefore, the dislocation is removed, whereby it is possible to safely and easily pull the dislocation-free crystal having no axial dislocation.

In the above-described method for producing a silicon single crystal, the above-mentioned high-energy radiation is in the energy range of 40 keV to 70 kEV.

It is possible to sufficiently detect state information about the dislocation necessary for growth of the dislocation-free crystal by high-energy radiation having such energy.

In addition, the dislocation removing portion can be irradiated with the above-mentioned high-energy radiation in the rotational state at a rotational speed of 0.1 to 30 rpm.

This makes it possible to obtain three-dimensional information of the dislocation. It is also possible to obtain information in a non-destructive way.

Radiation of high-energy radiation in the rotational state includes: a process in which the dislocation-removal portion rotates centered in the crystal growth direction on the rotation center axis, so that this dislocation-free portion is irradiated with high-energy radiation; alternatively, an operation of irradiating high-energy radiation to acquire dislocation information in a relative movement state, wherein a high-energy radiation source and the dislocation-free portion are in a state comparable to the above-described rotation ,

At an early stage of pulling the silicon single crystal by a CZ method, the step of dipping, the necking step and the step of forming a shoulder portion are performed. In the step of immersion, the seed crystal is rotated and lowered, and the apical portion of the seed crystal is dipped in the surface of the silicon melt. After the apical portion of the seed crystal is immersed in the silicon melt, the lowering of the seed crystal is stopped. The silicon melt and the seed crystal are caused to have sufficient affinity for each other.

When the seed crystal is brought into contact with the silicon melt, a meniscus is normally formed at the interface therebetween due to the surface tension of the silicon melt. The melt fluctuates immediately after the crystal raw material has melted, but strongly at the local temperature. In addition, the entire melt undergoes a noticeably large temperature fluctuation and becomes unstable.

For this reason, the step of dipping is performed after the crystal source material has melted and after a certain time has elapsed. In this case, when the temperature of the liquid level at the time of contacting the seed crystal with the silicon melt is excessively high, the apical portion of the seed crystal is melted and separated from melt. In contrast, when the temperature of the liquid level of the silicon melt is excessively low, the crystal grows around the apical portion of the seed crystal. Then, the crystal in the surface of the melt is suspended around the seed crystal. In such a state, if the process moves from the step of dipping to the necking step, a dislocation is generated in the neck section.

For this reason, it is necessary to stabilize the temperature of the melt as the process moves from the step of dipping to the necking step. After confirming that the apical portion of the seed crystal is immersed in the silicon melt, causing the seed crystal and the silicon melt to have sufficient affinity for each other and stabilizing the temperature of the silicon melt, the process proceeds from the dipping step to the necking step above.

That is, the process of causing the seed crystal and the silicon melt to have affinity for each other includes: a process in which the temperature of the surface (liquid level) of the silicon melt is determined by the shape of the interface at the time of contacting the process of controlling the electric power of the heater on the basis of the determined temperature of the liquid level to adjust the heat input to the silicon melt.

In other words, the process of causing the seed crystal and the silicon melt to have affinity for each other includes: a process in which the power of the heater is adjusted to adjust and stabilize the temperature of the melt surface Meniscus having a predetermined shape is formed around the apical portion of the seed crystal in the state in which the growth rate of the seed crystal is zero. In addition, the meniscus is not always formed only when the seed crystal and the silicon melt are caused to have affinity with each other. The meniscus is formed at the interface between the crystal and the melt itself at the time of crystal growth in the necking step and the like, following the step of causing the seed crystal and silicon melt to have affinity with each other.

• (1) Wenn das CZ-Verfahren eingesetzt wird, wird zuerst das Kristallausgangsmaterial in dem Tiegel geschmolzen und dann wird der Impfkrisatall in die Schmelze, die in dem Tiegel enthalten ist, eingetaucht und es wird bewirkt, dass der Impfkristall Affinität für die Schmelze hat. Danach wird der Impfkristall gezogen und der Necking-Schritt, bei dem der Neck-Abschnitt gebildet wird, wird durchgeführt. Als Nächstes werden der Schulter-Abschnitt und Körperabschnitt des Einkristalls gebildet. In einem solchen Verfahren zum Herstellen eines Silicium-Einkristalls wird, nachdem der Neck-Abschnitt mit einer vorbestimmten Länge in dem oben genannten Necking-Schritt gebildet wurde, bewirkt, dass der Neck-Abschnitt Affinität für die Schmelze hat, und dann kann der kontinuierliche Neck-Abschnitt gebildet werden.
• (2) Wenn das MCZ-Verfahren verwendet wird, wird zuerst das Kristallausgangsmaterial in dem Tiegel geschmolzen und dann wird der Impfkristall in die Schmelze, die in dem Tiegel enthalten ist, eingetaucht und es wird bewirkt, dass der Impfkristall Affinität dafür hat. Danach wird der Impfkristall (hoch)gezogen und es wird der Necking-Schritt zum Bilden des Neck-Abschnitts durchgeführt. Als Nächstes werden der Schulterabschnitt und der Körperabschnitt des Einkristalls gebildet. In einem solchen Verfahren zum Herstellen eines Einkristalls wird, nachdem der Neck-Abschnitt mit einer vorbestimmten Länge in dem oben beschriebenen Necking-Schritt gebildet ist, bewirkt, dass der Neck-Abschnitt Affinität für die Schmelze hat, und dann kann der kontinuierliche Neck-Abschnitt gebildet werden.

After causing the seed crystal to have affinity for the melt in the dipping step, regardless of a CZ method or an MCZ method, and forming the neck portion having a predetermined length in the necking step, it is caused the formed neck section again has affinity for the melt and the continuous neck section is formed. This makes it possible to shorten the length of the neck portion necessary for the dislocation-free state to remove the dislocation from the neck portion and to improve the dislocation-free rate.

• (1) When the CZ method is employed, first, the crystal source material in the crucible is melted, and then the seed is immersed in the melt contained in the crucible and the seed crystal is caused to have affinity for the melt. Thereafter, the seed crystal is pulled and the necking step in which the neck portion is formed is performed. Next, the shoulder portion and body portion of the single crystal are formed. In such a method of producing a silicon single crystal, after the neck portion having a predetermined length is formed in the above-mentioned necking step, the neck portion is made to have affinity for the melt, and then the continuous neck Section are formed.
• (2) When the MCZ method is used, first, the crystal source material in the crucible is melted, and then the seed crystal is immersed in the melt contained in the crucible, and the seed crystal is caused to have affinity for the seed crystal. Thereafter, the seed crystal is pulled (high) and the necking step for forming the neck portion is performed. Next, the shoulder portion and the body portion of the single crystal are formed. In such a method of manufacturing a single crystal, after the neck portion having a predetermined length is formed in the above-described necking step, the neck portion is made to have affinity for the melt, and then the continuous neck portion be formed.

• (3) Bei der Herstellung des Silicium-Einkristalls gemäß (1) oder (2) kann die Länge des Neck-Abschnitts, der zuerst im Necking-Schritt gebildet wird, auf 20 mm oder größer eingestellt werden. In dem oben beschriebenen Necking-Schritt kann, nachdem der Neck-Abschnitt auf eine Temperatur, die höher als die Temperatur der Schmelze zur Zeit der Bildung des Neck-Abschnitts ist, erhöht wurde, der Neck-Abschnitt dazu gebracht werden, dass er Affinität für die Schmelze hat. Außerdem wird die Temperatur der Heizvorrichtung zum Schmelzen des Kristallausgangsmaterials in dem Tiegel gemessen und die Temperatur der Heizvorrichtung kann auf der Basis des Resultats der Temperaturmessung reguliert werden, um die Temperatur der Schmelze einzustellen.

When the MCZ method is used, the transverse magnetic field applied to the melt may be in the range of 0.2 T to 0.4 T (2000 G to 4000 G).

• (3) In the production of the silicon single crystal according to (1) or (2), the length of the neck portion, which is first formed in the necking step, can be set to 20 mm or larger. In the necking step described above, after the neck portion is raised to a temperature higher than the temperature of the melt at the time of formation of the neck portion, the neck portion can be made to have affinity for the melt has. In addition, the temperature of the heater for melting the crystal raw material in the crucible is measured, and the temperature of the heater can be regulated on the basis of the result of the temperature measurement to adjust the temperature of the melt.

The processes described above in which the seed crystal is caused to have affinity for the melt (silicon melt) and causes the neck portion to have affinity for the melt include a process in which the meniscus shape of the contact interface at the time of the melt Contacting the crystal with the melt, for example, the overhang of the crystal habit line. That is, it includes an operation in which the temperature of the surface of the melt is determined, the power of the heater based thereon is regulated, the heat input to the melt is adjusted, and the temperature of the surface of the melt is stabilized.

By causing the seed crystal and the neck portion to have affinity for the melt, the dislocation is easily removed in this way. However, if it is caused that the seed crystal and the neck portion have affinity for the melt, it is even possible to produce the straight body portion having no axial dislocation in any of the production processes of the condition described above in the invention.

A silicon single crystal is prepared by any of the manufacturing methods described above.

A silicon wafer is made of the silicon single crystal described above.

In the present invention, moreover, there is disclosed a non-inventive method for producing a silicon single crystal in which the silicon single crystal is grown by a CZ method, the method comprising: a step of contacting a seed crystal with a silicon melt and starting to draw the silicon single crystal; a dislocation removing step in which a dislocation removing portion is formed which removes a dislocation generated in the step of immersion by propagating outward or by annihilation in a loop shape in accordance with the pulling of the crystal; a diameter enlarging step in which a shoulder portion is drawn by increasing the diameter to the necessary diameter; a step of forming a straight body in which a straight body portion is pulled; a dislocation behavior information acquiring step in which information on the dislocation behavior, which is a three-dimensional behavior of the dislocation, is detected in the dislocation removing section by irradiating the dislocation removing section with high-energy radiation; a step of determining that dislocation-free based on the dislocation behavior information is determined to be dislocation-free, and a step of setting the pulling conditions in which the dislocation-free performing conditions are determined on the basis of the determination that dislocation-free in the step of determining that dislocation free, be determined. This solves the problems described above.

In the present invention, moreover, there is disclosed a non-inventive method for producing a silicon single crystal in which the silicon single crystal is grown by a CZ method, the method comprising: a dipping step in which a seed crystal is brought into contact with a silicon melt and pulling of the silicon single crystal is started; a dislocation removing step in which a dislocation removing portion is formed which removes a dislocation generated in the step of immersion by propagating outward or by annihilation in a loop shape in accordance with the pulling of the crystal; a diameter enlarging step in which a shoulder portion is drawn by increasing the diameter to the necessary diameter; a step of forming a straight body in which a straight body portion is pulled; a dislocation behavior information acquiring step in which information about the dislocation behavior, which is a three-dimensional behavior of the dislocation, is detected in the dislocation removal section by irradiating the dislocation removal section with high energy radiation, and a dislocation free determination step is determined based on the dislocation behavior information the dislocation is removed.

The diameter enlargement step may be started based on the determination that dislocation-free in the step of determining that dislocation-free.

In the above-described method for producing a silicon single crystal, the high-energy radiation is in the energy range of 40 keV to 70 keV, and the dislocation-removing portion can be irradiated with high-energy radiation in the rotating state at a rotation number of 0.1 to 30 rpm.

In the present invention, there is also disclosed a non-inventive method for producing a silicon single crystal in which the silicon single crystal is grown by a CZ method, and which comprises: a dipping step in which a seed crystal is brought into contact with a silicon melt; a pulling of the silicon single crystal is started; a dislocation removing step in which a dislocation removing portion that removes a dislocation generated in the step of immersion by propagating outward or by annihilation in a loop shape is formed in accordance with the pulling of the crystal; a diameter enlarging step in which a shoulder portion is drawn by increasing the diameter to the necessary diameter; a step of forming a straight body in which a straight body portion is pulled; a dislocation behavior information acquiring step in which information on the dislocation behavior, which is a three-dimensional behavior of dislocation, is detected in the dislocation removal section by irradiating the dislocation removal section with high energy radiation, a step of determining that dislocation free based on the dislocation behavior information is determined the dislocation is removed, and a step of setting the pulling conditions based on the determination that dislocation-free, in the step of determining that dislocation-free, for dislocation-free performance.

This makes it possible to clarify the behavior of the dislocation for which the behavior in the prior art is not clarified. In addition, it is possible to annihilate the dislocation by escaping to the outside with the growth of the dislocation-removing portion (dislocation-free portion) or to annihilate the dislocation in a loop shape. This makes it possible to bring the dislocation-free section to a dislocation-free state which has no axial dislocation. The dislocation-free section is brought into a dislocation-free state and then the drawing conditions at the time of increasing the diameter of the silicon single crystal can be easily developed.

In the step of determining that dislocation-free, the dislocation state is determined from the information obtained in the previous step. For this reason, it is possible to accurately determine the behavior of the dislocation in the dislocation removal section (dislocation-free section). Thus, it is possible to effectively use the dislocation removal by the formation of a loop, which is unknown in the prior art. Since the diameter of the silicon single crystal is reduced, it becomes possible to increase the diameter of the dislocation-free portion called the neck portion, which has never been known before. Therefore, even if the weight of the crystal pulled out of the silicon melt increases, it is possible for the silicon single crystal to be carried by the dislocation-free portion, and it is possible to draw a large-diameter silicon single crystal. Therefore, it is possible to omit unnecessary operations concerning the step of immersion and the like, which cause a long production period and an increase in the number of operations. It is possible to pull a heavy-weight silicon single crystal having a diameter of about 450 mm in a dislocation-free state having no axial dislocation.

In the present invention, there is also disclosed a non-inventive method for producing a silicon single crystal in which the silicon single crystal is grown by a CZ method, and which comprises: a dipping step in which a seed crystal is brought into contact with a silicon melt; the pulling of the silicon single crystal is started; a dislocation removing step in which a dislocation removing portion that removes a dislocation generated in the step of immersion by propagating outward or by annihilation in a loop shape is formed in accordance with the pulling of the crystal; a diameter enlarging step in which a shoulder portion is drawn by increasing the diameter to the necessary diameter; a step of forming a straight body in which a straight body portion is pulled; a dislocation behavior information acquiring step in which information on the dislocation behavior, which is a three-dimensional behavior of dislocation, is detected in the dislocation removal section by irradiating the dislocation removal section with high-energy radiation, and a dislocation-free determination step is determined based on the dislocation behavior information; that the dislocation is removed. The diameter enlargement step is started on the basis of the determination that is dislocation-free in the step of determining that dislocation-free.

Thereby, it is possible to observe the behavior of the dislocation generated at the time of drawing in real time by a non-destructive inspection, which was impossible to accomplish in the prior art. Therefore, after confirming that the dislocation of the dislocation-removing portion is removed and brought into a dislocation-free state, the silicon single crystal can be pulled, whereby the shoulder portion (diameter enlargement portion) and the straight body portion can be formed. For this reason, it is possible to safely and easily pull the dislocation-free crystal which has no axial dislocation in the shoulder portion and the straight body portion.

In the above-described method for producing a silicon single crystal, the high-energy radiation may be in the energy range of 40 keV to 70 keV.

As a result, the high energy radiation has sufficient energy to provide information about the dislocation state necessary for the growth of the dislocation-free crystal.

The dislocation removal section may also be irradiated with the above-mentioned high-energy radiation in a rotation state at a revolution number of 0.1 to 30 rpm.

This makes it possible to obtain three-dimensional information of the dislocation. Further, it is possible to obtain three-dimensional information about dislocation in a non-destructive manner.

Here, irradiation of high-energy radiation in the rotational state includes: a process in which the dislocation-removal portion rotates centered in the crystal growth direction on the rotation center axis, so that this dislocation-free portion is irradiated with high-energy radiation; alternatively, an operation of irradiating high-energy radiation to acquire dislocation information in a relative motion state, wherein a high-energy radiation source and the dislocation-free section are in a state coincident with the rotation described above; Rotation is comparable.

[Advantageous Effects of Invention]

It is possible according to the invention to clarify the behavior of the axial dislocation, whose behavior in the prior art is not clarified. In addition, it is possible to annihilate the dislocation by escaping to the outside with the growth of the dislocation removing portion. Alternatively, it is possible to annihilate the dislocation in a loop shape.

In this way, the dislocation is annihilated, whereby it is possible to increase the diameter by bringing the dislocation-free portion into a dislocation-free state and thereafter pulling the silicon single crystal and growing the straight body portion. Alternatively, after it has been confirmed that the dislocation-free portion becomes a dislocation-free state, it becomes possible to grow the silicon single crystal and grow the shoulder portion (diameter expanding portion) and the straight body portion. Thereby, it becomes possible to maintain the diameter necessary for supporting the weight of the silicon single crystal. It is thus possible to safely and easily pull the dislocation-free crystal of large weight having no axial dislocation out of the silicon melt.

According to the method for producing a silicon single crystal of the invention, it is possible to realize a dislocation-free state by removing the dislocation generated in the step of contacting the seed crystal with the silicon melt. Furthermore, it is possible to accurately determine the situation of generation and removal of the dislocation. In addition, it is possible to accurately determine the relationship between the state of dislocation in the neck portion and the pulling conditions for the silicon single crystal. It is also possible to accurately determine the pulling conditions for the silicon single crystal capable of setting a dislocation-free state.

BRIEF DESCRIPTION OF THE DRAWINGS

1 FIG. 10 is a schematic cross-sectional view illustrating a manufacturing apparatus used in a method of manufacturing a silicon single crystal according to a first embodiment. FIG.

2 FIG. 10 is a flow chart illustrating the method of manufacturing a silicon single crystal according to the first embodiment in connection with other non-process steps of the present invention. FIG.

3 (a) to (e) are explanatory diagrams of steps S02 and S03 according to the first embodiment.

4 FIG. 10 is an explanatory diagram of step S12 according to the first embodiment. FIG.

5 (a) and 5 (b) are graphs illustrating the characteristics of high-energy radiation in step S12 according to the first embodiment.

6 FIG. 10 is an explanatory diagram of step S12 according to the first embodiment. FIG.

7 are image data of a dislocation-free portion obtained in step S12 according to the first embodiment.

8th are image data of the dislocation-free portion obtained in step S12 according to the first embodiment; 8 (a) shows a state in which dislocations are not removed, 8 (b) shows a state in which dislocations are annihilated in a loop shape, and 8 (c) shows a state in which dislocations are removed.

9 FIG. 12 is a cross-sectional view schematically illustrating the manufacturing apparatus used in the method of manufacturing a silicon single crystal according to a second embodiment. FIG.

10 FIG. 10 is a flowchart illustrating the method of manufacturing a silicon single crystal according to the second embodiment in connection with other non-process steps of the present invention.

11 are image data representing a viewing result of a wafer; 11 (a) FIG. 12 is an example of viewing of the wafer having axial dislocation, and FIG 11 (b) is an observation example of the wafer having no axial dislocation.

12 Fig. 12 is a schematic diagram illustrating a silicon single crystal produced by a prior art Dash necking method.

13 Fig. 10 is an enlarged view illustrating the silicon single crystal produced by the prior art Dash necking method.

14 FIG. 12 is a cross-sectional view schematically illustrating the manufacturing apparatus used in the production of a silicon single crystal according to a third embodiment. FIG.

15 FIG. 12 is a cross-sectional view schematically illustrating the manufacturing apparatus used in the method of manufacturing a silicon single crystal according to the third embodiment. FIG.

16 (a) to 16 (c) are explanatory diagrams for explaining step S03 according to the third embodiment.

17 (a) and 17 (b) are explanatory diagrams for explaining step S03 according to the third embodiment.

18 are image data of the dislocation-free portion obtained in step S12 according to the third embodiment; 18 (a) shows a state in which dislocations are not removed, and 18 (b) and 18 (c) are states where dislocations are removed.

DESCRIPTION OF EMBODIMENTS

Hereinafter, a method for producing a silicon single crystal according to a first embodiment of the invention will be described with reference to the drawings.

1 FIG. 10 is a schematic elevational view illustrating a silicon single crystal manufacturing apparatus according to the embodiment. FIG.

As in 1 is shown, a CZ furnace, which is an apparatus for producing a silicon single crystal using a Czochralski method (hereinafter referred to as CZ method), comprises a crucible 1 , a heater 2 , a drawing axis 4 , a seed crystal chuck 5 , a heat shield element (heat shield) 7 and a magnetic field generating element (magnetic field generating device) 9 ,

The crucible 1 has a double structure, is arranged in the middle part in a chamber, has a quartz crucible 1a and a graphite crucible 1b , The quartz crucible 1a takes a silicon melt 3 on. The graphite crucible 1b is located on the outside of the quartz crucible 1a and holds the quartz crucible 1a , The crucible is made using a pedestal 1c turned and moved up.

The heater 2 is arranged on the outside of the crucible. As a heater 2 For example, a resistance heater may be used. The heater 2 heats the silicon melt 3 in the crucible 1 is at a temperature of the melting point or higher and is not limited to a resistance heating device as long as the temperature can be maintained.

The drawing axis 4 rotates about the axis parallel to the vertical direction at any rotation speed. In addition, the drawing axis moves 4 in the axial direction parallel to the vertical direction at any speed.

The seed crystal chuck 5 is at the bottom of the drawing axis 4 arranged and holds a seed crystal T of silicon.

If a silicon single crystal 6 in the CZ oven, which is in 1 is bred, causes the drawing axis 4 lowers and the seed crystal T attached to the seed crystal chuck 5 is attached, in the silicon melt 3 is immersed. After that, while the crucible and the drawing axis turn, the pull axis 4 so as to raise the seed crystal T and the silicon single crystal 6 to pull (high).

The heat shield element 7 is above the crucible 1 arranged such that it is the lateral side of the growing silicon single crystal 6 surrounds and the upper side of a part of the liquid level of the silicon melt 3 surrounds. The heat shield element 7 comprises a graphitic outer shell of cylindrical shape and in the interior of the outer shell is filled graphite felt. The heat shield element 7 is between the heater 2 and the surface of the silicon melt 3 and the lateral side of the silicon single crystal 6 arranged and shielded against radiant heat. The inner surface of the heat shield element 7 is oblique, so that the inner diameter of the heat shield element 7 gradually decreases from the upper end to the lower end. The outer surface of the upper part of the heat shield element 7 is similar to the inner surface of the same obliquely. The outer surface of the lower part of the heat shield element 7 is nearly parallel to the vertical direction. This reduces the thickness of the lower part of the heat shield element 7 to the lower side near the silicon melt 3 to.

The width (thickness) W of the lower part of the heat shield element 7 in the radial direction is for example about 50 mm. The inclination θ of the inner surface of the heat shield element 7 which is a surface of an inverted truncated cone with respect to the vertical direction is, for example, about 21 °. The height H1 of the liquid level of the silicon melt 3 to the lower end of the heat shield element 7 is in the range of, for example, about 10 mm to 250 mm. The height H1 can be set, for example, to 50 mm or 100 mm. In addition, the heights H1 can be appropriately set in each step described later.

A magnetic field generating element 9 is on the outside of the heater 2 arranged. The magnetic field, the magnetic field generating element 9 is generated, may be a horizontal magnetic field or a cusp magnetic field. The strength of the horizontal magnetic field can be 2000 G to 5000 G (0.2 T to 0.5 T). The strength of the horizontal magnetic field can be 3000 G to 4000 G (0.3 T to 0.4 T). The strength of the horizontal magnetic field is preferably 3000 G to 3500 G (0.30 T to 0.35 T). The height of the magnetic field center may be in the range of 150 mm below the liquid level of the silicon melt 3 up to 100 mm above the liquid level of the silicon melt 3 be. The height of the magnetic field center is more preferably 75 mm below the liquid level up to 50 mm above the liquid level. Optionally, the magnetic field can not by the magnetic field generating element 9 be generated.

Next, a description will be given of a method for producing a silicon single crystal in which the in 1 shown CZ furnace is used.

When the silicon single crystal 6 is prepared, the state of dislocation to be generated and the behavior of the dislocation may change due to a change in the drawing conditions. Thus, in the method of producing a silicon single crystal according to the embodiment, the pulling conditions of the silicon single crystal capable of removing the dislocation are adjusted.

As in 2 14, the manufacturing method according to the embodiment includes a step of setting the drawing conditions S00, a preparation step of drawing S01, a step of immersion S02, a step of removing a dislocation S03, a diameter increasing step S04, a step of forming a straight body S05, a step of forming the tail (tapering step) S06, a measurement preparing step S11, a dislocation behavior information acquiring step S12, and a step of determining that dislocation-free, S13.

First, in step S00, the drawing conditions that allow the silicon single crystal to become 6 is in a dislocation-free state. Here, the drawing conditions include: the holding time of the seed crystal (seed) T before contact with the silicon melt 3 ; the contact time with the silicon melt 3 the silicon seed T and the temperature of the seed T based on the contact time; the distance (height H1) between the heat shield element 7 and the silicon melt 3 ; the width W of the lower end of the heat shield element 7 in the diameter direction; the conditions of an atmosphere inside the furnace; the number of revolutions and the pulling speed of the drawing axis 4 at the time of pulling the silicon single crystal; the number of revolutions of the crucible 1 ; the length of the dislocation-free portion (neck portion, dislocation removal portion) N, which will be described later, in the drawing direction; the state of the applied magnetic field and the heating state, that is, the temperature distribution in the grown silicon single crystal. The drawing conditions may be set to any values in an initial state before the production of the silicon single crystal 6 be set.

The pulling conditions of the silicon single crystal 6 In step S00, moreover, data obtained by pulling the silicon single crystal can be used 6 was done several times under different conditions before. In this case, the conditions are based on the specification of the producing silicon single crystal 6 selected.

Next, in step S01, the high-purity silicon polycrystal is placed in the crucible 1 injected. The weight of the silicon polycrystal to be injected is approximately in the range of 100 kg to 400 kg. At this time, a dopant for adjusting a product to a predetermined resistivity on the basis of whether the product is a p-type or an n-type is injected so as to have a predetermined concentration in the silicon melt. In addition, the concentrations of carbon, nitrogen and the like are adjusted to adjust or set the gettering ability or resistivity of the product. The dopant may optionally be injected.

In the embodiment, the silicon polycrystal of 300 kg is placed in the crucible 1 injected. The concentration of the dopant is adjusted so that the resistance of a straight body portion 6b of the silicon monocrystal to be produced 6 for example, 12 Ωcm will.

The atmosphere in the CZ furnace is set to an inert gas atmosphere, and the pressure of the inert gas is set in the range of 1.3 kPa to 13.3 kPa (10 Torr to 100 Torr). In the embodiment, the atmosphere in the CZ furnace is set, for example, as an Ar gas atmosphere at a pressure of 50 Torr (6.666 kPa). In addition, the atmosphere in the CZ furnace may contain hydrogen gas.

In addition, a horizontal magnetic field of, for example, 3000 G (0.3 T) is generated by the magnetic field generating element 9 is generated and then the silicon polycrystal with the heater 2 heated and melted. At this time, the height of the center of the magnetic field becomes within a range of 75 mm below the liquid level of the silicon melt 3 to 50 mm above the liquid level of the silicon melt 3 established. When using a CZ (MCZ) method in which the magnetic field is applied, the transverse magnetic field applied to the silicon melt 3 is set to a range of 0.2 T to 0.4 T (2000 G to 4000 G).

Next, in step S02, as shown in FIG 3 (a) the seed crystal T attached to the seed crystal chuck 5 is attached, close to the silicon melt 3 come and is kept for a predetermined time and heated. Thereafter, the seed crystal T is rotated, and lowered, and then comes into contact with the silicon melt 3 brought as it is in 3 (b) is. The lower end of the seed crystal T becomes the silicon melt 3 immersed (step of immersion). At this time, the dislocation t in the lower end of the seed crystal T is generated by a thermal shock (see arrow A1). At this time, each of the temperatures of the seed crystal T and the silicon melt is 3 set so that the temperature conditions are in a predetermined range. Specifically, the heating state of the heater 2 is set to form a meniscus having a predetermined shape around the apical portion of the seed crystal T and the temperature of the silicon melt 3 is set.

Next, in step S03, as shown in FIG 3 (c) is shown, the silicon single crystal 6 pulled while the crucible 1 and the drawing axis 4 turn around. At this time, the shape of the interface between the silicon single crystal 6 and the silicon melt is shaped in a concave shape (see arrow A2). The pulling conditions of the silicon single crystal 6 use the conditions set in step S00. The dislocations generated in step S02 at this time become propagated toward the outside of the silicon single crystal 6 annihilated in the diametric direction or are annihilated in a loop shape, depending on the drawing conditions. This will, as in the 3 (d) and 3 (e) is shown, the dislocations t removed and the dislocation-free neck portion, that is, the dislocation-free portion N, is formed (see arrow A3).

The dislocation-free portion N in which the dislocations t are removed is formed as shown in FIG 3 (c) is shown, the shape of the interface (meniscus) between the silicon single crystal grown in the lower end of the seed crystal T, and the silicon melt 3 becomes a predetermined shape. The shape of the meniscus formed in the lower end of the seed T, or the silicon single crystal 6 changes depending on the heating conditions of the heater 2 and the drawing conditions, for example, the speed of raising the drawing axis 4 , The temperatures of the seed T and the silicon melt 3 For example, they are respectively set to the temperatures within a predetermined range, thereby making it possible to control the shape of the meniscus formed in the lower end of the seed crystal T or the silicon single crystal 6 adjust.

If the dislocation-free section N, which has no formed dislocation t, as shown in 3 (d) is formed, the dislocations t among the dislocations generated in step S02 are formed so as to be outside of the silicon single crystal 6 extend in the diameter direction. In addition, the crystal habit line becomes the outer periphery of the silicon single crystal 6 formed (see arrow A4). The portion of the crystal habitus line on the inside of the diameter direction is in a dislocation-free state which has no dislocation. When the silicon single crystal 6 has grown, as is in 3 (e) Further, the crystal habit line continues at the lower side (see arrow A5) and the dislocation-free portion is formed continuously (see arrow A6). On the lower side of the dislocation-free section, a dislocation-free initial cone is also formed. In this way, the dislocations t become by spreading toward the outside of the silicon single crystal 6 annihilated in the diametric direction or are annihilated in a loop shape. As a result, the dislocation-free portion N is formed.

Next, in step S04, the diameter of the silicon single crystal becomes 6 increased linearly by the pulling rate of the silicon monocrystal 6 , the number of rotations of the drawing axis 4 , the number of rotations of the crucible 1 and the heating conditions of the heater 2 be regulated. Thereby, the diameter at the lower side of the dislocation-free portion N in which the silicon single crystal becomes 6 who in 3 (e) is shown in a dislocation-free state, enlarged. As in 1 is shown at the lower side of the dislocation-free portion N of the silicon single crystal 6 a conical shoulder section 6a educated.

Next, in step S05, the pull axis becomes 4 by a predetermined length in the state where the diameter of the silicon single crystal 6 is held at a diameter of, for example, 300 mm or 450 mm, which is the diameter of the silicon wafer serving as a product, is pulled up. This will, as in 1 is shown, the cylindrical straight body portion 6b which has a constant diameter in which silicon single crystal 6 formed (step of making a straight body).

Next, in step S06, the diameter of the silicon single crystal becomes 6 decreased linearly by the pulling rate of the silicon single crystal 6 , the number of revolutions of the drawing axis 4 , the number of revolutions of the crucible 1 and the heating conditions of the heater 2 be regulated. Thereby, a conical end portion or taper portion (not shown), whose direction is opposite to that of the shoulder portion 6a is in the silicon single crystal 6 educated. Thereafter, the silicon single crystal 6 from the silicon melt 3 separated by the end portion and the silicon melt 3 be separated from each other.

Next, in step S11, preparation for measuring the state of dislocation of the dislocation-free portion N is performed. Specifically, first, the dislocation-free portion N is made of the silicon single crystal 6 separated. As in 4 2, the separated dislocation-free portion N is supported by a support portion R1 so as to be coaxial with the pulling axis by using the central axis N4 of the silicon single crystal 4 rotated as a rotation axis. The support portion R1 is disposed so as to be capable of changing the direction of the center axis N4 which is a rotation axis.

Next, in step S12, as shown in FIG 4 2, the dislocation-free portion N is irradiated with white X-rays as high energy B radiation from an irradiation element R, and diffracted X-rays of the white X-rays irradiating the dislocation-free portion N are detected by a detection element d including a CCD (dislocation behavior information detecting step ). The irradiation element R comprises an absorber, a shutter and a diffractometer and has a shutter absorber used in regulating the time resolution and the measurement against a heat load. The detection element d has a function as a diffractometer. The detection element d has sufficient performance to detect white X-ray topography.

In the prior art, the analysis of the dislocation to be detected by X-ray topography was carried out in the following manner. First, the cross section to be analyzed is cut as a disk to a thickness of about 1.5 mm. Thereafter, the cut surface to be analyzed is etched using a mixed acid and the transmission diagram is observed by X-ray irradiation. Thus, only the transmission diagram in the cross section of a predetermined interval is obtained, and images, for example, at a pitch of 1.5 mm, are synthesized to analyze a change in the state of dislocation associated with crystal growth. For this reason, only the dislocation propagating obliquely to the pulling axis of the silicon single crystal can be observed.

On the other hand, when the dislocation-free portion N is observed using the white X-rays as high-energy radiation as disclosed in SYNCHROTRON RADIATION INSTRUMENTATION: Ninth International Conference to Synchrotron Radiation Instrumentation, AIP Conference Proceedings, Vol. 879, pp. 1545-1549 (2007) , the dislocations annihilated in a loop shape can be measured. This confirms that the dislocations are in levels other than the normal {111} plane. By using such high-energy white X-ray, it is possible to three-dimensionally image the behavior, for example, the dislocation within the silicon single crystal, without performing the destructive inspection, for example, by slicing. As a result, positional information of the dislocation and dislocation characteristics can be three-dimensionally observed. As such observation device, the large synchrotron radiation system (Spring-8) called BL28B Japan Synchrotron Radiation Research Institute can be used.

The white X-rays have a continuous spectrum and can be in the energy range from 30 keV to 1 MeV. The energy of the white X-rays may be 40 kEV to 100 kEV or 50 kEV to 60 kEV. The wavelength of the white X-rays may be 0.001 nm to 0.25 nm. Thereby, it is possible to sufficiently consider the state of dislocation in the dislocation-free section N.

5 (a) is a graph in which the vertical axis indicates the number of photons and the horizontal axis indicates the energy, showing the distribution of the number of photons passing through a slot of 1 mm × 1 mm in a position at a distance of 44 m the X-ray source of the irradiation element R is obtained.

5 (b) FIG. 12 is a graph in which the vertical axis indicates the number of photons and the horizontal axis indicates the wavelength, showing the distribution of the number of photons passing through a slot of 1 mm × 1 mm in a position at a distance of 44 m from FIG X-ray source of the irradiation element R is obtained.

Here, the X-ray diameter may preferably be in the range of 0.01 to 1 times the diameter of the dislocation-free portion N which is an object to be measured.

In the embodiment, the dislocation-free portion N is irradiated with white X-rays having an energy of 40 keV to 70 keV, which is high-energy radiation B, while being rotated at a rotation number of 0.1 rpm to 30 rpm. Specifically, while the dislocation-free portion N is rotated using the center axis N4 in the crystal growth direction as a rotation axis, the dislocation-free portion N from the irradiation element R is irradiated with high-energy radiation. Alternatively, while the irradiation element R is rotated centered on the central axis N4, the dislocation-free portion N from the irradiation element R is irradiated with high-energy radiation. Both the irradiation element R and the dislocation-free portion N can be rotated by using the axis N4 as a rotation axis. It is possible to detect three-dimensional information about the dislocation of the dislocation-free portion N in a non-destructive manner by, for example, performing a Fourier transform on the detection result.

In addition, in step S12, as shown in FIG 4 is shown, the irradiation angle ω of the high energy B radiation with respect to the axis N4 of the dislocation-free portion N of the silicon single crystal 6 set. With a focus on a diffraction spot (for example, 004 diffraction spot), a CT image is formed by a topograph image by rotating the dislocation-free section N. Thereby, three-dimensional position information about the dislocation line is obtained. The three-dimensional behavior of the dislocation in the dislocation-free portion N obtained in this way is recorded as the first information.

In addition, properties of dislocation, for example, Burgers vectors, are determined by white X-ray topography, which in 6 shown is measured. In 6 The detection positions d1 to d7 correspond to the diffraction direction of the white x-ray beam. Each of the detection directions corresponds to a position for detecting the Laue topograph, for example in the crystal orientations [-1 1 1], [-1 1 3], [0 0 4], [1-13] and [1 -11].

Here, -1 in the above crystal orientations means 1 with a superline.

Next, in step S13, it is determined whether the dislocation is removed based on the first information including the three-dimensional behavior of the dislocation detected in step S12. Specifically, from the image data of the dislocation-free portion N shown in FIG 7 , it is determined whether the dislocation is removed. That is, as in 8 (a) is shown, when the dislocation extends along the central axis N4, it is determined whether the dislocation is removed. As in 8 (b) is shown, if the dislocations exist but are annihilated in a loop shape, it is determined that the dislocations are removed. As in 8 (c) is shown, if the dislocations can not be confirmed, it is determined that the dislocations may be removed.

Next, the determination in step S13 that is dislocation-free is returned to step S00.

That is, when the dislocation of the dislocation-free portion N is removed by the drawing conditions first set in step S00, the drawing conditions are recorded in a database, and the drawing conditions are used in step S00. When the dislocation of the dislocation-free portion N is removed by the drawing conditions first set in step S00, the drawing conditions in the database are recorded as not being used in step S00, and new drawing conditions are set as appropriate in step S00 in order to bring the dislocation-free section N into a dislocation-free state.

As described above, according to the method of manufacturing a silicon single crystal of the embodiment, the dislocation-free portion N is formed, thereby enabling the dislocation involved in the step of contacting the seed crystal T with the silicon melt 3 was generated in the shoulder portion 6a and the straight body portion 6b of the silicon single crystal 6 Will get removed. This makes it possible, the dislocation-free state of Schulterabsbhnitts 6a and the straight body section 6b of the silicon single crystal 6 to realize.

In addition, it is possible to accurately determine the situation of the generation and removal of the dislocation by step S12 and step S13.

Further, it is also possible to accurately determine the pulling conditions of the silicon single crystal capable of producing a dislocation-free state by repeating step S00 to step S06 and step S11 to step S13.

Even if the diameter of the neck portion of the silicon single crystal 6 , that is, the dislocation-free portion N, is not reduced more than necessary, it is also possible to remove the dislocation in the dislocation-free portion N. Accordingly, it is possible to use the silicon single crystal 6 with heavy weight having a diameter of about 450 mm, in which the dislocation is removed to grow safely by a simple method, no special methods or devices are required.

Hereinafter, a method for producing a silicon single crystal according to a second embodiment of the invention will be described with reference to the drawings.

In the embodiment, the components corresponding to the first embodiment mentioned above are given the same reference numerals and the description thereof is omitted.

As in 9 is shown, the CZ furnace of the embodiment comprises the crucible 1 , the heater 2 , the drawing axis 4 , the seed crystal chuck 5 , the heat shield element 7 , the magnetic field generating element 9 , the irradiation element R and the detection element d.

The irradiation element R irradiates the dislocation-free portion N while being drawn while it is located in a chamber with high-energy radiation B. The detection element d detects diffracted light of the high-energy radiation B to which the dislocation-free portion N is irradiated. The irradiation element R and the detection element d are arranged so as to be capable of expressing the angle ω of the high energy B with respect to the pulling axis 4 , which is close to the axis of the axis N4 to adjust.

The method for producing a silicon single crystal according to the embodiment includes step S00 to step S06 according to the first embodiment described above, as shown in FIG 10 is shown. However, the method does not include step S11 to step S13 after step S00 to step S06, and differs from the first embodiment described above in that it between step S03, and step S04 includes a dislocation behavior information acquisition step S22 and a determination that dislocation-free S23.

In step S22, similarly to the above-described step S12, the dislocation-free portion N from the irradiation element R is irradiated with high energy B radiation, and the diffracted light is measured by the detection element d. In the embodiment, the irradiation of the dislocation-free portion N with the high energy B radiation and the detection of the diffracted light within the CZ furnace are performed while pulling the silicon single crystal.

In step S23, according to the above-described step S13, it is determined whether the dislocation of the dislocation-free portion N is removed. When the dislocation of the dislocation-free portion N is removed, step S04 is started. Further, if the dislocation of the dislocation-free portion N is not removed, the process is returned to step S03, and the dislocation-free portion N is further grown, or the dislocation-free portion N is re-grown by remelting. In addition, the drawing conditions are returned to step S00, and drawing conditions different from the withdrawn drawing conditions are set.

According to the embodiment, in the 3 (a) to 3 (e) shown states measured in real time. For this reason, it is possible to start with step S04 after confirming that the dislocations t in the dislocation-free portion N are removed. Accordingly, there is no case in which the silicon single crystal 6 is pulled into a state in which the dislocations t in the shoulder section 6a and the straight body portion 6b not removed. Thus, it is possible to shorten the time required for the silicon single crystal 6 manufacture. In addition, it is possible to obtain the yield ratio when pulling the silicon single crystal 6 to improve.

Next, a third embodiment of the invention will be described with reference to the drawings.

In the embodiment, a description is given of a method of manufacturing a silicon single crystal capable of removing the axial dislocation extending in the pulling direction of the silicon single crystal over the entire length of the silicon single crystal.

The axial dislocation is a dislocation among the dislocations generated in the neck portion, which is not removable in the prior art, even if the diameter of the silicon single crystal is reduced by a Dash necking method. When a wafer is produced by slicing the straight body portion of the silicon single crystal into thin slices, the axial dislocation has been detected as a pit in the art.

The axial dislocation can be distinguished from the well and detected by the following method.

First, deformation is measured in the surface of the wafer of the silicon single crystal, using, for example, the deformation inspecting device (SIRD: Registered Trade Mark) SirTec manufactured by JENA WAVE, which is capable of distributing a stress state observed by optical means of investigation.

Specifically, on the wafer, a manifestation method of heating is performed for 2 seconds or more at a rate of temperature increase in the range of 10 ° C / s to 300 ° C / s and at a temperature in the range of 900 ° C to 1250 ° C , The surface properties of the wafer, including the manifested distortion, are evaluated. In this manifestation method, a large temperature difference occurs due to the rapid heating in the plane of the wafer, and thereby the thermal deformation or the thermal stress is caused. Thus, no heating for a long time is required, but heating may be performed for a short period of about 1 second.

The axial dislocation exists through the surface and back of the wafer. Therefore, in the axial dislocation, the strain increases due to the thermal stress generated by the above-described manifestation method. On the other hand, the axial dislocation is detected as a pit in the surface contamination test, however, in the dislocation having no such length as to pass through the surface and back of the wafer, the strain due to the thermal stress generated by the manifestation method increases. Not. Specifically, in the dislocation in which the length thereof is less than 10% of the thickness of the wafer, the strain due to the thermal stress exhibited by the manifestation method does not increase.

Although the axial dislocation is proved by the above-described manifestation method, for the above reason, other pits and the like are not proved by this manifestation method. 11 (a) FIG. 11 shows a viewing example of the wafer having the proven axial dislocation. FIG. 11 (b) Fig. 11 shows an observation example of the wafer having no axial dislocation. It is assumed that the deformation observed at the edge of the wafer is generated by holding the wafer at the time of the manifestation process.

The axial dislocation is generated in the step of contacting the seed crystal with the silicon melt. The axial dislocation exists in the direction of the silicon monocrystal which is grown, that is, continuously with the neck portion, the shoulder portion, and the straight body portion. The axial dislocation occurs in a specific region when the silicon single crystal is processed into a wafer. The axial dislocation can not be removed by a dash necking technique of the art.

As in 12 shown comprises a silicon single crystal 60 produced by the Dash Necking method in the relevant art, a seed crystal T0, a neck portion N0, a shoulder portion 60a , a straight body section 60b and an end portion (rejuvenation portion) 60c ,

When the neighborhood of the seed T0 is increased, as in 13 is shown, thermal shock dislocation Jn generated in the seed crystal T0 and continuing to the neck portion N0 is generated by the thermal shock when the seed crystal T0 is brought into contact with the silicon melt. If there is no match in the lattice constant between the seed T0 and the neck portion N0, a misfit dislocation Jm is generated in the neck portion N0.

Neither the thermal shock dislocation Jn nor the misfit dislocation Jm grown in the growth direction of the silicon single crystal has the possibility of becoming a dislocation J. In the art, the axial dislocation J is observed as follows. First, as in 12 shown is the straight body section 60b of the silicon single crystal 60 thinly cut and a plurality of wafers W of the {110} cross-section are obtained. If a notch (notch) on each wafer W towards 12 Clock is attached, the axial dislocation J in the direction 10 Clock and the direction 4 Watch observed. More specifically, with the position of the notch attached in the {100} direction set at 0 °, the axial dislocation J in a region J1 becomes symmetrical with respect to the central axis of the silicon single crystal in the range of 120 ° observed to 135 ° and in the range of 315 ° to 350 °. In addition, the axial dislocation J is observed in a region obtained by rotating the region symmetrically with respect to the center axis of the silicon single crystal by 90 ° and 45 ° centered at the center axis. That is, in the region J1, the axial dislocation J extends in the growth direction (center axis direction) of the silicon single crystal. It shows 12 only a region J1 in the range of 120 ° to 135 ° and in the range of 315 ° to 350 °, wherein the position of the notch is set to 0 °.

However, in the silicon single crystal grown to a normal length or greater, the axial dislocation J has the possibility of being annihilated. However, when the wafer of a silicon single crystal having a normal length is considered, the axial dislocation exists over the entire length of the silicon single crystal, and the axial dislocation J often passes through the silicon single crystal in the central axis direction.

In the method of growing the silicon single crystal, at the time of pulling the neck portion, the shoulder portion, and the straight body portion, there are the following characteristics in the direction of propagation of the thermal shock dislocation Jn and the misfit dislocation Jm.

First, the thermal shock dislocation Jn and the misfit dislocation Jm normally move through the {111} plane, which is a slip plane of silicon.

Second, the thermal shock dislocation Jn and the misfit dislocation Jm are normally introduced into the {111} plane.

Third, the thermal shock dislocation Jn and the misfit dislocation Jm immediately move to the {111} plane, although they can rarely move through another plane without entering the {111} plane.

Fourth, most of the thermal shock dislocations Jn and the misfit dislocations Jm are generated at the time of contacting the seed crystal with the silicon melt.

When the silicon monocrystal is pulled to make a <100> wafer, it is considered that the thermal shock dislocation Jn and the misfit dislocation Jm are affected by the {111} plane, which is a sliding plane of the silicon monocrystal of forming a neck portion and reach the surface of the silicon single crystal. However, according to the observation result by the above-described X-ray topography, the initial thermal shock dislocation Jn and the misfist dislocation Jm do not extend to the {111} plane of the slip plane but in the direction perpendicular to the interface between the silicon single crystal and the Si Silicon melt, regardless of the diameter of the neck section. Therefore, the thermal shock dislocation Jn and the misfit dislocation Jm do not move to the slip plane, but often extend in the growth direction of the silicon single crystal.

Here, there may be a case that the thermal shock dislocation Jn and the misfact dislocation Jm do not reach the surface of the silicon monocrystal even if the diameter of the silicon monocrystal increases to the diameter of the product, and the axial dislocation within the Straight body section exists. In this way, there may occur a case that the silicon single crystal having the axial dislocation within the straight body portion has a crystal habit line similar to the silicon single crystal having no axial dislocation in the straight body portion, and this is not of a non-defective one Product in the state of an ingot can be distinguished. The axial dislocation is confirmed as an error at the time of manufacturing the wafer from the ingot.

In addition, a case may occur in which the thermal shock dislocation Jn and the misfit dislocation Jm extending in the growth direction of the silicon single crystal move to the {111} plane during the formation of the shoulder portion or during the formation of the straight body portion ,

When the thermal shock dislocation Jn and the misfist dislocation Jm can be moved to the {111} plane in the step of forming a neck portion, it is possible to remove the axial dislocation even if the diameter of the neck portion does not decrease or the seed crystal is not brought to a high temperature. Therefore, as long as the thermal shock dislocation Jn and the misfit dislocation Jm can be moved to the {111} plane in the neck portion, the product is not affected by the axial dislocation even if the shoulder portion and the straight body portion are formed. In addition, it is possible to suppress the number of times of remelting and remelting time in an early growth stage of the silicon single crystal. That is, as long as the thermal shock dislocation Jn and the misfit dislocation Jm can be moved to the {111} plane in the neck portion, it is possible to obtain the silicon single crystal of large weight in which the axial dislocation in the shoulder portion and in the straight body section is not present, to breed at low cost.

As described above, the thermal shock dislocation Jn and the misfit dislocation Jm do not proceed to the {111} plane at the time of formation of the neck portion, but in the direction perpendicular to the interface between the silicon single crystal and the silicon melt. Therefore, in the embodiment, a description will be made of a method of moving the thermal shock dislocation Jn and the misfist dislocation Jm that are not in the {111} plane to the {111} plane in the step of forming a neck portion , Specifically, the shape of the interface between the silicon single crystal and the silicon melt at the time of formation of the neck portion is regulated. To confirm the conditions, the X-ray topography is used.

Hereinafter, a method for producing a silicon single crystal according to the embodiment will be described in detail with reference to the drawings.

The method for producing a silicon single crystal according to the embodiment comprises step S01 to step S06 and step S11 to step S13 according to the method for producing a silicon single crystal according to the first embodiment, which is disclosed in U.S.P. 2 is shown. Alternatively, the method includes step S01 to step S06, step S22, and step S23 according to the method of manufacturing a silicon single crystal according to the second embodiment, which is described in US Pat 10 is shown. In the embodiment The method differs from those of the first embodiment and the second embodiment described above in that the above-mentioned axial dislocation is removed in step S03.

In addition, the CZ furnace of the embodiment differs from the CZ furnace used in the first embodiment or the second embodiment described above in that it does not include a temperature-regulating mechanism for the dislocation-free portion N. All other configurations are the same as those of the CZ furnace according to the first embodiment or the second embodiment.

As in 14 is shown, the CZ furnace of the embodiment comprises a laser irradiation element La and a gas supply element G as a temperature regulating mechanism of the dislocation-free portion N of the silicon single crystal. The laser irradiation element La and the gas supply element G are arranged so as to be centered on the pulling axis 4 synchronous with the rotation of the drawing axis 4 rotate. That is, the laser irradiation element La and the gas supply element G heat and cool the silicon single crystal coextensive with the pulling axis 4 turns, from a predetermined direction.

In addition, the laser irradiation element La and the gas supply member G need not be arranged to rotate centered on the pulling axis. In this case, the laser irradiation element La and the gas supply element G are fixed, and a laser beam and cooling gas can be applied and supplied intermittently at a predetermined timing from the laser irradiation element La and the gas supply element G.

Alternatively, as in 15 is shown in the CZ oven of the embodiment, an openable and closable opening 7a at the lower end of the heat shield element 7 is formed as a temperature regulating mechanism of the dislocation-free portion N of the silicon single crystal. In this case is a cylindrical element 7a in the upper part of the heat shield element 7 arranged and a lid part 7b which has an opening which the passage of the drawing axis 4 allowed, at the top of the cylindrical element 7a arranged. A rule element 41 that is the raising and lowering of the drawing axis 4 is regulated, is on the lid part 7b arranged. The heat shield element 7 , the cylindrical element 7a , the lid part 7b , the rule element 41 and the drawing axis 4 rotate integrally around the axis, with the drawing axis 4 consistent with the rotation of the draw axis 4 , That is, the heat shield element 7 only shields in a predetermined direction of the silicon single crystal, which is aligned with the drawing axis 4 turns, not from the heat of the heater 2 , In other words, the heater 2 heats the silicon single crystal, which is aligned with the drawing axis 4 turns, through the opening 7a from a predetermined direction.

In the embodiment, first, the seed crystal T with the silicon melt is formed in step S02 3 brought into contact, as it is in the 3 (a) and 3 (b) is shown.

Thereafter, in step S03, the pulling axis is caused 4 goes up, and the silicon single crystal is grown under the seed crystal T, whereby the dislocation-free portion N is formed.

In step S02, the seed crystal T is rotated and lowered, and then the lower end of the seed crystal T becomes the silicon melt 3 immersed. After the lower end of the seed T into the silicon melt 3 is submerged, the lowering of the seed crystal T is stopped, and it is sufficiently effected that the seed crystal T and the silicon melt 3 Have affinity to each other. Part of the lower end of the seed crystal T becomes the silicon melt 3 integrated.

When the silicon melt 3 is in a stable state and the seed crystal T with the silicon melt 3 is brought into contact in step S02, the surface tension of the silicon melt acts 3 at the interface between the seed crystal T and the silicon melt 3 , and so a meniscus is formed.

However, immediately after the silicon polycrystal used as the starting material has melted, the silicon melt fluctuates 3 strong in terms of local temperature and gets an unstable state. For this reason, after the silicon polycrystal serving as the raw material is melted, step S02 is performed after a lapse of a predetermined time.

When the temperature in the liquid level of the silicon melt 3 is excessively high, the lower end of the seed crystal T is melted at the time of pulling the silicon single crystal and thus separated therefrom.

When the temperature in the liquid level of the silicon melt 3 is excessively low, a crystal grows on the lateral side of the lower end of the seed crystal T, resulting in an overhanging state. In such a state, when the process goes from step S02 to step S03, a new dislocation is generated in the dislocation-free section N.

When the process goes from step S02 to step S03, it is necessary that the temperature of the silicon melt 3 is in stable condition. For this reason, after the lower end of the seed crystal T is caused to flow into the silicon melt 3 is submerged, the lower end of the seed crystal T has a sufficient affinity for the silicon melt 3 has, and after the temperature of the silicon melt 3 is stabilized, the process proceeds from step S02 to step S03.

When causing the lower end of the seed crystal T affinity for the silicon melt 3 has, the seed crystal T in the silicon melt 3 is immersed and then the shape of the interface between the seed crystal T or the silicon single crystal and the silicon melt 3 observed. In addition, the overhang of the crystal habit line is observed by observing the interface. The temperature of the liquid level of the silicon melt 3 becomes the shape of the interface between the seed crystal T or the silicon single crystal and the silicon melt 3 and the overhang of the crystal habitus line. Electrical energy, that of the heater 2 is regulated on the basis of the estimated temperature and the heat input to the silicon melt 3 is set. That is, when causing the lower end of the seed crystal T affinity for the silicon melt 3 has, the heating conditions are through the heater 2 is set so that a meniscus having a predetermined shape is formed at the lower end of the seed crystal T in the state where the growth rate of the seed crystal T is zero. The temperature of the liquid level of the silicon melt 3 is set and so does the silicon melt 3 stabilized.

Thus, after it has been effected in step S02, the seed crystal T affinity for the silicon melt 3 has pulled the silicon single crystal in step S03 to a predetermined length, thereby forming the dislocation-free portion N having a predetermined length. The length of the dislocation-free portion N, which is formed after causing the seed crystal T to have affinity for the silicon melt 3 is preferably set to 20 mm or larger. Thereafter, the lower end of the formed dislocation-free portion N having a predetermined length is made to have affinity for the silicon melt similar to the seed crystal T 3 and then the dislocation-free portion N having a predetermined length is formed.

In addition, in the embodiment, in step S03, the lower end of the seed crystal T is caused to have affinity for the silicon melt 3 in a state in which the temperature of the silicon melt 3 is regulated as described above, and then the dislocation-free portion N is formed with a predetermined length. The lower end of the formed neck portion, which has a predetermined length, is caused to have affinity for the silicon melt 3 in the state where the temperature is set in a similar manner as for the lower end of the seed crystal T, and then the dislocation-free portion N having a predetermined length is further formed. Thereby, it is possible to shorten the length of the dislocation-free portion N necessary for the dislocation-free state and to reliably remove the axial dislocation and dislocation of the dislocation-free portion N.

In addition, in step S03, after the temperature of the dislocation-free portion N becomes a temperature above the temperature of the liquid level of the silicon melt 3 was increased, causing the dislocation-free portion N affinity for the silicon melt 3 Has. In addition, the temperature of the heater 2 at the time of melting the silicon polycrystal, which serves as a starting material, measured at the temperature of the heater 2 on the basis of the measurement result, and then the temperature of the silicon melt can be adjusted.

The silicon single crystal becomes almost perpendicular to the interface K between the silicon single crystal and the silicon melt 3 , which is a surface in which the crystal is grown, is grown. For this reason, the dislocations j, as in 16 (a) is shown when the interface K between the silicon single crystal and the silicon melt 3 Similarly, it does not progress to the {111} plane, which is a sliding plane of the silicon single crystal, but advances in the direction (arrow direction in the drawing) nearly perpendicular to the interface K. Therefore, when the interface K is a nearly horizontal plane surface, there may occur a case where the dislocations j become the axial dislocations extending in the growth direction of the silicon single crystal from the interface K to the shoulder portion and the straight body portion of the silicon body. Single crystals extend.

The shape of the interface between the silicon single crystal and the silicon melt 3 varies depending on the pulling rate of the silicon single crystal, that is, the speed of rising of the pulling axis 4 , Consequently, it is considered that the speed of raising the drawing axis 4 is changed and thus the shape of the planar interface K, which is almost parallel to the horizontal surface, as in 16 (a) is shown changed to a curved shape.

When the pulling rate of the silicon monocrystal is less than a predetermined speed, as shown in FIG 16 (b) is shown; The interface K1 between the silicon single crystal and the silicon melt changes 3 to a downward convex shape. When the pulling rate of the silicon monocrystal is higher than a predetermined speed as shown in FIG 16 (c) is shown, the interface K2 between the silicon single crystal and the silicon melt changes 3 to an upward convex shape.

That is, it is possible to form the interfaces K1 and K2 between the silicon single crystal and the silicon melt 3 to regulate by regulating the pulling rate of the silicon single crystal.

The silicon single crystal is grown in the direction perpendicular to the interfaces K1 and K2. In addition, the dislocations j proceed in the growth direction of the silicon single crystal. That is, the dislocations j propagate in the normal direction of the virtual tangent plane at the interfaces K1 and K2 between the silicon single crystal and the silicon melt 3 continued.

For this reason, most of the dislocations j, as in 16 (b) that is, when the interface K1 is a curved surface that is downwardly directed convex, toward the outside of the dislocation-free portion N of the silicon single crystal in the diameter direction. As in 16 (c) 2, when the interface K2 is a curved surface that is convex upward, the dislocations j propagate toward the inside of the dislocation-free portion N of the silicon single crystal in the diameter direction.

In this way, the shape of the interface K is regulated and deformed into a curved shape, whereby it is possible to move the propagation direction of the dislocations j in the direction from the direction of the central axis which is the growth direction of the silicon single crystal. is different. That is, at a point where the virtual tangent plane is not horizontal at arbitrary points on the boundaries K1 and K2, the dislocations j proceed in the normal direction of the virtual tangent plane. Therefore, at a point where the virtual tangent plane is not horizontal at arbitrary points at the interfaces K1 and K2, the traveling direction of the dislocations j moves in the direction different from the direction of the central axis which is the growth direction of the silicon single crystal differentiates. Thereby, the thermal shock dislocation Jn and misfit dislocation Jm that do not exist in the {111} plane can be moved to the {111} plane to remove them.

That is, when the interface K1 is a curved surface that is downwardly convex, the dislocations j propagate toward the outside of the dislocation-free portion N of the silicon single crystal in the diameter direction to reach the outer peripheral surface of the silicon single crystal , and are removed in the dislocation-free section N in this way. In addition, when the interface K is a curved surface that is convex upward, the dislocations j propagate toward the inside of the dislocation-free portion N of the silicon single crystal in the diameter direction and are annihilated, and are thus removed in the dislocation-free portion N. ,

Therefore, it is possible to prevent the axial dislocations J from being generated by removing the dislocations j propagating from arbitrary points at the interfaces K1 and K2 in which the virtual tangent plane is not horizontal.

When causing the dislocations j to go toward the central axis of the silicon single crystal and be annihilated or caused to proceed to the outer peripheral surface of the silicon single crystal and thus removed, there may occur a case where the pulling rate of the silicon monocrystal fluctuates, the diameter of the silicon monocrystal to be formed fluctuates. At this time, even if the diameter of the silicon single crystal decreases to a certain extent, no problem occurs as long as it is a diameter capable of holding the silicon single crystal finally to be formed. In addition, even if the diameter of the silicon single crystal increases to a certain extent, there are no problems. as long as it has a diameter that rejects any starting material as waste. As long as the diameter is in the range in which the pulling of the straight body portion of the silicon Single crystal is not adversely affected, the diameter of the silicon single crystal may vary to a degree.

It may be the case that the direction of progress of the dislocation deviates from the pulling direction of the silicon monocrystal due to a change in the pulling speed. This phenomenon is advantageous for the purpose of dislocation removal.

Boundaries K1 and K2, which are in 16 (a) and 16 (b) are curved surfaces which are parallel to the central axis, which is parallel to the growth direction of the silicon single crystal. That is, when the central axis of the silicon single crystal becomes an axis of rotation, the interfaces K1 and K2 are curved surfaces that are rotationally symmetric to each other. For this reason, the virtual tangent plane at the intersection of the interfaces K1 and K2 with the center axis of the silicon single crystal may be horizontal. In this case, the dislocations j, which are generated at the central axis of the silicon monocrystal, proceed at the central axis of the silicon monocrystal in the growth direction of the silicon monocrystal. These dislocations j may be the axial dislocation extending from the interfaces K1 and K2 to the shoulder portion and the straight body portion of the silicon single crystal.

Consequently, in step S03 of the embodiment as shown in FIG 17 (a) and 17 (b) is shown, the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 formed into a curved surface and deformed in a shape that is asymmetric to the central axis of the silicon single crystal. Specifically, the temperature distribution in the direction perpendicular to the center axis of the silicon single crystal, that is, in the diameter direction, is generated in the interfaces K3 and K4.

When the temperature distribution in the diameter direction, that is, the temperature gradient in the diameter direction, is not generated in the silicon single crystal as shown in FIGS 16 (b) and 16 (c) is shown, the interfaces K1 and K2 between the dislocation-free portion N and the silicon melt 3 a curved surface symmetrical with the central axis, which is parallel to the growth direction of the silicon single crystal.

Here, the dislocation-free portion N of the silicon single crystal is heated or cooled from one direction, with the in 14 or 15 shown CZ furnace is used, whereby the temperature distribution is generated perpendicular to the central axis direction of the silicon single crystal in the interfaces K1 and K2.

If the in 14 is used, the neighborhood of the interfaces K1 and K2 of the dislocation-free portion N is heated from one direction by the laser beam irradiation from the laser irradiation element La.

In addition, from the side opposite to the laser irradiation element La, cooling gas is supplied through the gas supply element G. Thereby, the neighborhood of the interfaces K1 and K2 of the dislocation-free portion N from the direction opposite to the direction in which the dislocation-free portion N is heated is cooled.

If the in 15 shown CZ furnace is caused to radiant heat of the heater 2 and the silicon melt 3 the neighborhood of the interfaces K1 and K2 of the dislocation-free portion N from the opening 7A of the heat shield element 7 is reached, and the dislocation-free portion N is heated from one direction.

In this way, a temperature gradient becomes perpendicular to the direction of rising of the drawing axis 4 generated in the interfaces K1 and K2.

Thereafter, the interfaces K3 and K4 between the dislocation-free portion N of the silicon single crystal and the silicon melt 3 as it is in the 17 (a) and 17 (b) is shown, a curved surface which is symmetrical to the central axis, which is parallel to the growth direction of the silicon single crystal. Thereby, a horizontal portion Kd of the interfaces K3 and K4 moves from a portion Kc on the center axis of the silicon single crystal to a position away from the outside in the diameter direction.

In addition, the shapes of the interfaces K3 and K4 change between the silicon single crystal and the silicon melt 3 also depending on the strength of a magnetic field, the magnetic field generating element 9 is generated, the rotational speed of the crucible 1 and the rotational speed of the Silicon single crystal, that is, the rotational speed or rotational speed of the drawing axis 4 , By being regulated, the shapes of the interfaces K3 and K4 become between the silicon single crystal and the silicon melt 3 More precisely, the position of the horizontal section Kd is also regulated more precisely.

The silicon melt 3 and the silicon single crystal can be brought into a relatively quiet state, for example, by the rotational speed of the crucible 1 and the rotational speed of the drawing axis 4 be regulated. When the silicon melt 3 and the silicon single crystal are relatively stopped, thus, the fluctuation of convection of the silicon melt becomes 3 generated. This is the convection of the silicon melt 3 not symmetrical to the center axis of the silicon single crystal and the growth of the silicon single crystal is not symmetrical to the center axis of the silicon single crystal. That is, the temperature distribution in the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 is not symmetrical to the central axis of the silicon single crystal. Thereby, the horizontal portions Kd of the interfaces K3 and K4 move to positions away from the portions Kc at the center axis of the silicon single crystal toward the outside in the diameter direction.

In addition, the strength of a magnetic field can be set to zero by the strength of a magnetic field generated by the magnetic field generating element 9 is generated is reduced. The convection of the silicon melt 3 is adjusted correctly by the magnetic field. Therefore, when the magnetic field is annihilated during the pulling of the silicon single crystal, the convection of the silicon melt becomes 3 generates a fluctuation. This is the convection of the silicon melt 3 not symmetrical with the center axis of the silicon single crystal and the growth of the silicon single crystal is not symmetrical to the center axis of the silicon single crystal. That is, the temperature distribution in the interfaces K3 and K4 between the silicon single crystal and the silicon melt is not symmetrical with the center axis of the silicon single crystal. Thereby, the horizontal portions Kd of the interfaces K3 and K4 move to a position away from portions Kc at the center axis of the silicon single crystal toward the outside in the diameter direction.

In addition, in the state where the heater 2 asymmetric to the drawing axis 4 is arranged, the heating device in synchronism with the rotation of the drawing axis 4 to be turned around. This is the convection of the silicon melt 3 not symmetrical with the center axis of the silicon single crystal, and the temperature distribution of the silicon melt 3 is not symmetrical to the central axis of the silicon single crystal. As a result, the growth of the silicon single crystal is not symmetrical to the center axis of the silicon single crystal. That is, the temperature distribution in the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 is not symmetrical to the central axis of the silicon single crystal. Thereby, the horizontal portions Kd of the interfaces K3 and K4 move to positions away from the portions Kc at the center axis of the silicon single crystal toward the outside in the diameter direction.

In addition, properties of a portion of the inner peripheral surface of the quartz crucible may be caused to be caused 1a may differ from the properties of the other sections. Specifically, the portion having about 1/4 of the entire inner circumferential surface may be a changed portion different from the other portions in a range of about 10 cm from the upper end of the quartz crucible 1a be educated. In this case, in the changed portion, the bubble content in a position where the depth from the surface is in the range of 0.5 mm to 1 mm is higher than the bubble content in the other portions, in a range of 25%. up to 35%. For example, the blister content of the changed section may be 30% higher than the bladder content of the other sections.

In this way, the properties of a portion of the inner peripheral surface of the quartz crucible become 1a made different from the properties of the other sections, whereby the convection of the silicon melt 3 is not symmetrical to the central axis of the silicon single crystal and the growth of the silicon single crystal is not symmetrical to the central axis of the silicon single crystal. That is, the temperature distribution in the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 is symmetrical to the central axis of the silicon single crystal. Thereby, the horizontal portions Kd of the interfaces K3 and K4 move away from portions Kc on the center axis of the silicon single crystal to the outside in the diameter direction.

In addition to the method described above, similarly, methods can be developed in which the positions of the horizontal portions Kd of the interfaces K3 and K4 are moved by changing parameters, for example, the heating state of the heater 2 , the strength of a magnetic field generated by the magnetic field generating element 9 is generated, the rotational speed of the crucible and the rotational speed of the drawing axis 4 ,

In this way, the horizontal portions Kd of the interfaces K3 and K4 move from the center axis of the silicon single crystal to the outside in the diameter direction, and thus the propagation direction of the dislocation j generated at the central axis of the silicon single crystal is from the direction of the Center axis, which is the growth direction of the silicon single crystal, has been changed. In addition, the horizontal portions Kd of the interfaces K3 and K4 can be appropriately moved. Thereby, the thermal shock dislocation Jn and the misfist dislocation Jm which are not in the {111} plane can be reliably removed into the dislocation-free portion N by moving to the {111} plane. Thereby, it is possible to reliably remove the dislocation j extending from any points on the interfaces K3 and K4 and to reliably remove the axial dislocation J.

According to the embodiment, therefore, it is possible to produce a silicon single crystal having no axial dislocation extending to the straight body portion which can not be removed by the prior art method, in addition to the effect in the method of production of a silicon single crystal described in the above-described first embodiment and second embodiment. It is possible to manufacture a wafer having no axial dislocation from the straight body portion of the silicon single crystal having no axial dislocation.

[Example 1]

In this example, a 300 kg silicon polycrystal is injected into the crucible. The concentration of a dopant is adjusted so that the resistance value of the straight body portion of the silicon single crystal to be produced becomes 12 Ωcm. In addition, the atmosphere in the CZ furnace is set to an Ar gas atmosphere with a pressure of 50 Torr (6.666 kPa).

The diameter of the seed crystal is set at 300 mm to grow the dislocation-free section.

Here, step S00 to step S06 and step S11 to step S13 are performed using five different conditions from sample 1 to sample 5 shown in Table 2 as the drawing conditions in step S00 shown in FIG 2 , carried out. Specifically, as shown in Table 2, five different combinations of the pulling rate of the drawing axis, the distance (height) of the heat shield spacing from the liquid level of the silicon melt, and the retention time for which the seed crystal was immersed in the silicon melt under holding were used. [Table 2] Pulling speed (mm / mm) Distance from the melt (mm) Retention time (min) dislocation Sample 1 1.0 50 5 there Sample 2 1.0 50 100 there Sample 3 1.0 100 100 there Sample 4 1.5 100 100 loop Sample 5 0.5 150 100 dislokationsfrei

In step S13, which is in 2 is shown, the condition of each sample was considered. As in 8 (a) is shown in sample 1, the dislocation extends in the axial direction, and the dislocation is removed. As in 8 (b) is shown, in sample 4, although the dislocation exists, the dislocation is annihilated in a loop form, and the dislocation is removed. As in 8 (c) is shown in sample 5, the dislocation is removed.

Accordingly, as for the drawing conditions in step S00, the pulling speed of the pulling axis is set to a range of 0.5 mm / min to 1.5 mm / min, the distance (height) of the silicon melt is set to a range of 100 mm 150 mm and the retention time is set to 100 min. Thereby, it was possible to remove the dislocation of the dislocation-free portion and to remove the dislocation of the shoulder portion and the straight body portion of the silicon single crystal.

[Example 2]

When the silicon single crystal is prepared using the prior art method as disclosed in U.S. Pat 18 (a) As shown, the axial dislocation extending in the central axis direction of the silicon single crystal exists in the neck portion of the silicon single crystal.

Next, the removal of the axial dislocation was carried out in a similar manner to Example 1.

As a result, in step S13, which is in 2 is shown, determines that the axial dislocation in the dislocation free section is removed, as shown in FIG 18 (b) is shown. Alternatively it moves, as is in 18 (c) is shown, the propagation direction of the dislocation to the {111} plane, which is different from the central axis direction of the silicon single crystal, and in the dislocation-free section, the dislocation reaches the outer peripheral surface of the silicon single crystal and is removed.

INDUSTRIAL APPLICABILITY

A method for producing a silicon single crystal according to the invention comprises: a step of contacting a seed crystal with a silicon melt, and then pulling the seed crystal (high) to grow a silicon single crystal; a step of removing a dislocation generated in the silicon single crystal by causing the dislocation to migrate toward the outside of the silicon single crystal in the diameter direction, or by annihilating the dislocation in a loop shape to form a dislocation-free portion in the silicon Single crystal, and a step in which the silicon single crystal is pulled by forming the dislocation-free portion and enlarging the silicon single crystal to a predetermined diameter to form a straight body portion.

A method for producing a silicon single crystal according to the invention solves at least one of the following problems: realization of a dislocation-free state by removing the dislocation generated in the step of contacting the seed crystal with the silicon melt; to accurately determine the situation of generation and removal of the dislocation; for accurately determining a relationship between the state of dislocation in the neck portion and the pulling conditions of the silicon single crystal and for accurately determining the pulling conditions of the silicon single crystal capable of setting a dislocation-free state.

LIST OF REFERENCE NUMBERS

1a

quartz crucible

3

silicon melt

4

Drawing axis (middle axis)

6

Silicon single crystal

6a

shoulder portion

6b

straight body section

B

High energy radiation

N

dislocation-free section (dislocation removal section, axial dislocation removal section)

T

seed

K1, K2, K3, K4

Interface (solid-liquid interface)

Claims (11)

A method of producing a silicon single crystal in which a silicon single crystal is grown by a CZ method, comprising the steps of: dipping a seed crystal in a silicon melt and starting to pull the seed crystal to grow the silicon single crystal from the silicon melt allow; Forming a straight body portion of the silicon single crystal to be used as the silicon wafer, and removing an axial dislocation generated in the step of dipping and extending in the crystal pulling direction, before the step of forming a straight body portion, the step of removing the axial body Dislocation by asymmetric heating with respect to the center axis of the silicon single crystal is carried out, and the fluctuation in the state of solid-liquid The interface between the silicon melt and the silicon single crystal is increased and the direction in which the axial dislocation extends is moved away from the crystal growth direction. A method of producing a silicon single crystal according to claim 1, wherein, in the step of removing an axial dislocation, rotation of the silicon single crystal with respect to the silicon melt is reduced and fluctuation in the solid-liquid interface state is increased. The method for producing a silicon single crystal according to claim 1, wherein in the step of removing an axial dislocation, the magnetic field intensity applied to the silicon melt is reduced, and the fluctuation in the solid-liquid interface state is increased. The method for producing a silicon single crystal according to claim 1, wherein the asymmetric heating is performed by heating the silicon single crystal from one direction through an opening of a heat shield portion, wherein the heat shield shielding against heat is disposed above a crucible so as to be the heat shield lateral side of the growing silicon single crystal and surrounding the upper side of a part of the liquid level of the silicon melt is performed, and an openable and closable opening at the lower end of the heat shield element is formed, and the asymmetric heating is carried out, wherein the heat shield element and a The pulling axis of the silicon single crystal is integrally rotated about the axis which is consistent with the pulling axis in synchronism with the rotation of the pulling axis. A method of producing a silicon single crystal in which a silicon single crystal is grown by a CZ method comprising the following steps: Immersing a seed crystal in a silicon melt and starting to draw the seed crystal to grow the silicon single crystal from the silicon melt; Forming a straight body portion of the silicon single crystal to be used as the silicon wafer; and Removing an axial dislocation generated in the step of immersion and extending in the crystal pulling direction, before the step of forming a straight body portion, wherein the step of removing the axial dislocation is performed by forming a temperature distribution that is asymmetric with respect to the center axis of the silicon single crystal at a liquid-solid interface between the silicon melt and the silicon single crystal, and the fluctuation in the state of solid state Liquid interface is increased and the direction in which the axial dislocation extends, is moved away from the crystal growth direction. The method for producing a silicon single crystal according to claim 5, wherein the asymmetric temperature distribution is generated by heating the liquid-solid interface from one direction using laser beam irradiation from a laser irradiation element, while the liquid-solid interface from the side facing the laser irradiation element is cooled by using cooling gas supplied through a gas supply member, and the laser irradiation member and the gas supply member are rotated centered along a pulling axis of the silicon single crystal in synchronization with the pulling axis of the silicon single crystal. A method of producing a silicon single crystal according to claim 5, wherein the asymmetrical temperature distribution is formed by setting the strength of a magnetic field from a predetermined value to zero. The method of manufacturing a silicon single crystal according to claim 5, wherein the asymmetric temperature distribution is formed by heating a crucible by arranging a heater on the outside of the crucible asymmetrically with respect to a pulling axis of the silicon single crystal, and heating the heater synchronously with the crucible Rotation of the drawing axis is rotated. The method for producing a silicon single crystal according to claim 5, wherein the asymmetric temperature distribution by using a quartz crucible in which a property of a portion of the inner peripheral surface of the quartz crucible is different from that of the other portions of the inner peripheral surface of the quartz crucible, so that occurrence of pulsation in a part of the silicon melt, which receives the quartz glass crucible, made higher and causes a vibration of the melt level. A method for producing a silicon single crystal according to claim 5, wherein, in the step of removing an axial dislocation, rotation of the silicon single crystal with respect to the silicon melt is reduced. A method of producing a silicon single crystal according to claim 5, wherein in the step of removing an axial dislocation, the magnetic field intensity applied to the silicon melt is reduced.

Images