TWI858345B - Method for manufacturing single crystal silicon and single crystal pulling device - Google Patents

Abstract

The present invention can control the resistivity along the axial direction with high precision and improve the yield rate regardless of whether the single crystal formation dislocation occurs. The present invention comprises: a step of adding a main dopant to the silicon melt when a silicon melt is formed in a crucible; a step of growing a single crystal from the silicon melt to which the main dopant is added; and a step of continuously or intermittently adding a secondary dopant having a conductivity type opposite to that of the main dopant to the silicon melt during the step of growing the single crystal; and a step of adding the secondary dopant to the silicon melt. The process is performed in the step of: transporting a flake dopant as the secondary dopant into the chamber by a push rod that can move back and forth relative to the chamber, causing the flake dopant transported into the chamber to fall onto a funnel-shaped fixture by the push rod, and injecting the flake dopant into the silicon melt through a capillary disposed on the funnel-shaped fixture.

Description

Method for manufacturing single crystal silicon and single crystal pulling device

The present invention relates to a method for manufacturing single crystal silicon and a single crystal pulling arrangement, and more particularly to a method for manufacturing single crystal silicon and a single crystal pulling arrangement capable of maintaining the resistivity along the axial direction within a specification range.

The growth of single crystal silicon by the Czochralski method (hereinafter referred to as the CZ method) is carried out in the following manner: polycrystalline silicon as a raw material is filled into a quartz crucible placed in a chamber, and the polycrystalline silicon is heated and melted by a heater placed around the quartz crucible to form a silicon melt. Thereafter, a seed crystal (i.e., a seed) mounted on a seed chuck is immersed in the silicon melt, and the seed chuck and the quartz crucible are rotated in the same direction or in the opposite direction while the seed chuck is pulled up.

In addition, many single crystal silicon produced by such CZ method is used as semiconductor material. The resistivity of the grown single crystal silicon is adjusted by the dopant added to the silicon melt. Dopants are classified into n-type and p-type. As a dopant when growing n-type crystals, P (phosphorus) is mostly used.

When adding dopants to single crystal silicon grown using the CZ method, the resistivity changes in the direction of crystal growth. This is because the dopant segregates, and as the single crystal grows, the amount of molten silicon in the crucible decreases, causing the concentration of the dopant in the remaining liquid to gradually increase, and the resistivity of the single crystal decreases continuously. The segregation coefficient of phosphorus is 0.35, but the segregation coefficient of boron, which is widely used as a dopant for p-type crystals, is lower than 0.8, and the resistivity decreases significantly from the top to the bottom compared to p-type crystals. As a result, the portion that can be used as a product becomes smaller, and there is a problem that the yield is difficult to improve.

For such a topic, for example, Patent Document 1 discloses a method of adding a main dopant and a secondary dopant (i.e., counter dope) having a polarity opposite to that of the main dopant and a smaller segregation coefficient before the crystal is pulled. By using this method, the resistivity reduction caused by the main dopant is offset by the secondary dopant, and the axial resistivity distribution of the single crystal can be improved. Here, as mentioned above, the most commonly used dopant in the manufacturing process of n-type single crystals is phosphorus (P), and the segregation coefficient of phosphorus is about 0.35. Boron (B), which is an opposite polarity element and widely used in the manufacture of devices, has a segregation coefficient of about 0.8. The segregation coefficient is larger than that of phosphorus (P), so the above technology cannot be used directly.

In view of the above-mentioned problem, Patent Document 2 discloses a method of continuously adding boron (B) as a secondary dopant to phosphorus (P) as a main dopant during the pulling process of a single crystal. If this method is used, an n-type single crystal with improved axial resistivity distribution can be manufactured by reverse doping with phosphorus (P) as a main dopant and boron (B) as a secondary dopant.

However, in the method disclosed in Patent Document 2, when dislocation occurs during the crystal growth process, the single crystal and the raw material melt already contain the secondary dopant of opposite polarity at that point in time. Therefore, when the crystal is melted again, the main dopant is insufficient and the product cannot be obtained. Therefore, there is the problem of separating and removing the dislocation portion from the raw material melt and discarding it, which greatly reduces the yield rate.

Compared to the subject of patent document 2, the invention disclosed in patent document 3 discloses the following structure: a secondary dopant thin rod and a primary dopant thin rod are provided in a chamber, and the secondary dopant thin rod is continuously or intermittently pushed into the silicon melt during the cultivation process of the straight body portion. When a dislocation is formed, the single crystal silicon is melted again, and the primary dopant thin rod is pushed into the silicon melt before pulling to add the primary dopant. [Prior technical document] [Patent document]

[Patent Document 1] Japanese Patent Publication No. 2004-307305. [Patent Document 2] Japanese Patent Publication No. 3-247585. [Patent Document 3] Japanese Patent Publication No. 2016-60667.

[The problem that the invention wants to solve]

According to the method disclosed in Patent Document 3, even if the reverse doped single crystal silicon forms dislocations during the growth process, a thin rod of the main dopant can be pushed in when it is melted again to make up for the insufficient concentration of the main dopant. However, the mechanism for pushing the thin rod of the dopant into the melt is complicated, and the concentration of the thin rod itself varies in the length direction due to segregation, so it is difficult to accurately control the concentration of the main dopant and the secondary dopant.

In addition, Patent Document 3 states that a granular dopant can be added to the melt instead of using the aforementioned structure of pushing a thin rod of the dopant into the melt. However, the specific structure when the granular dopant is added to the melt is not disclosed, and there is a problem that if the granular dopant is simply added to the melt, there is a risk of generating ripples on the melt surface and forming misalignment.

The inventor of this case added a secondary dopant during the pulling process of single crystal silicon, and conducted research based on the premise of controlling the axial resistivity, and then completed the present invention. The purpose of the present invention is to provide a method for manufacturing single crystal silicon, in which a secondary dopant is added during the pulling process of single crystal silicon, and reverse doping is performed to control the axial resistivity, and the secondary dopant can be supplied with a relatively simple structure. Even if the crystal is re-melted due to dislocation, the main dopant can still be supplied by the structure to make up for the insufficient concentration of the main dopant. That is, the purpose of the present invention is to provide a method for manufacturing single crystal silicon and a single crystal pulling device, which can control the resistivity along the axial direction with high precision and improve the yield rate regardless of whether the single crystal silicon has formed dislocation. [Means for solving the problem]

The method for producing single crystal silicon of the present invention completed to solve the above-mentioned problem is to form a silicon melt in a crucible by heating with a heater in a chamber, and to grow single crystal silicon by the Czochralski method. The method for producing single crystal silicon comprises: when the silicon melt is formed in the crucible, a main dopant is added to the silicon melt; a step of growing the single crystal silicon from the silicon melt to which the main dopant is added; and in the step of growing the single crystal silicon, a main dopant having a certain affinity with the main dopant is continuously or intermittently added to the silicon melt. The step of adding a secondary dopant of the opposite conductive type to the aforementioned silicon melt is performed in the step of adding the aforementioned secondary dopant to the aforementioned silicon melt: one or more flake dopants are transported into the aforementioned chamber as the aforementioned secondary dopant by a push rod that can move back and forth relative to the aforementioned chamber, the aforementioned flake dopant transported into the aforementioned chamber is dropped into a funnel-shaped fixture by the aforementioned push rod, and the aforementioned flake dopant is put into the aforementioned silicon melt through a capillary provided in the aforementioned funnel-shaped fixture.

In addition, ideally, the manufacturing method of the single crystal silicon is to perform the following steps in the step of growing the single crystal silicon: after the step of adding the secondary dopant to the silicon melt, when the single crystal silicon forms a dislocation, remelting the single crystal silicon with the dislocation; adding the flake dopant of the main dopant to the silicon melt using the push rod and the funnel-shaped fixture; growing the single crystal silicon again from the silicon melt; and in the step of growing the single crystal silicon again, continuously or intermittently adding the secondary dopant having a conductivity type opposite to that of the main dopant to the silicon melt. In addition, ideally, the manufacturing method of the above-mentioned single crystal silicon is to perform the following steps before the step of cultivating the above-mentioned single crystal silicon: applying a horizontal magnetic field to the silicon melt in the above-mentioned crucible; during the step of cultivating the above-mentioned single crystal silicon, adding a secondary dopant having a conductivity type opposite to that of the above-mentioned main dopant to the above-mentioned silicon melt, and setting the dopant drop position in the horizontal magnetic field to the 45° region to the 135° region and the 225° region to the 315° region when the magnetic field application direction is defined as 0°.

According to the above method, when a single crystal silicon is pulled from a silicon melt to which a main dopant is added for growth, a flake-shaped secondary dopant having a conductivity type opposite to that of the main dopant is continuously or intermittently added to the aforementioned silicon melt, thereby maintaining the resistivity along the axial direction of the single crystal silicon within the specification range. In particular, when the flake-shaped dopant is added to the silicon melt, the flake-shaped dopant is dropped to the correct position on the melt surface with precise positioning (pinpoint) through a funnel-shaped fixture arranged in a chamber, so that the flow of the melt surface is not affected, and the occurrence of misalignment caused by the addition of the dopant can be suppressed. In addition, since the structure is set to put the transported flake dopant into the funnel-shaped jig by the push rod, the flake dopant is supplied to the funnel-shaped jig by free fall, passes through the capillary portion in an increasing state, and falls to the molten silicon surface. Therefore, the dopant can be added to the molten silicon without clogging the capillary portion of the funnel-shaped jig. In addition, as a structure for adding dopant during the growth process of single crystal silicon, it can be realized by a relatively simple structure, and the required amount of dopant can be added in units of sheets, so precise resistivity control can be performed.

In addition, the single crystal pulling device of the present invention completed to solve the above-mentioned problem forms a silicon melt in a crucible arranged in a chamber and heated by a heater, and pulls a single crystal silicon by the Czochralski method; the single crystal pulling device is equipped with: a shielding protective member arranged to surround the above-mentioned single crystal silicon grown above the above-mentioned crucible; a funnel-shaped fixture, which is mounted on the above-mentioned shielding protective member and is composed of a conical portion and a capillary, and the above-mentioned conical portion is The funnel-shaped fixture has an upper opening, and the thin tube extends downward from the lower top of the cone portion; and the push rod is configured to be able to move forward and backward relative to the chamber, and to put one or more sheet-like dopants placed on the front end into the opening of the cone portion in the funnel-shaped fixture; the sheet-like dopants put into the opening of the cone portion in the funnel-shaped fixture fall into the silicon melt through the thin tube of the funnel-shaped fixture.

In addition, ideally, the front end of the push rod is formed into a spoon shape, and the push rod is configured to be rotatable around an axis. In addition, ideally, the single crystal pulling device is provided with: a horizontal magnetic field applying mechanism, which applies a horizontal magnetic field to the silicon melt in the crucible; and the funnel-shaped fixture is arranged in the 45° region to the 135° region and the 225° region to the 315° region when the dopant drop position in the horizontal magnetic field is set to the magnetic field application direction defined as 0°.

According to the above configuration, when a single crystal silicon is pulled from a silicon melt to which a main dopant is added for growth, a flake-shaped secondary dopant having a conductivity type opposite to that of the main dopant is continuously or intermittently added to the aforementioned silicon melt, thereby maintaining the resistivity along the axial direction of the single crystal silicon within the specification range. In particular, when the flake-shaped dopant is added to the silicon melt, the flake-shaped dopant is precisely positioned and dropped to the correct position on the surface of the melt by a funnel-shaped fixture arranged in the chamber, thereby suppressing the occurrence of dislocation caused by the flake-shaped dopant flowing along the surface of the melt and adhering to the growing crystal. In addition, since the structure is set to put the conveyed flake dopant into the funnel-shaped jig by the push rod, the flake dopant is supplied to the funnel-shaped jig by free fall, passes through the capillary portion in an increasing state, and falls to the molten surface of the silicon melt. Therefore, the dopant can be added to the silicon melt without clogging the capillary portion of the funnel-shaped jig. In addition, as a structure for adding dopant during the growth process of single crystal silicon, it can be realized by a relatively simple structure, and the required amount of dopant can be added in units of sheets, so that precise resistivity control can be performed. [Effect of the invention]

According to the present invention, a secondary dopant is added during the pulling process of single crystal silicon to perform reverse doping to control the axial resistivity, and the secondary dopant can be supplied with a relatively simple structure. Even if the crystal is melted again due to formation dislocation, the main dopant can still be supplied through the structure to make up for the insufficient concentration of the main dopant. As a result, regardless of whether the single crystal silicon is formed dislocation, the axial resistivity can be controlled with high precision and the yield can be improved.

The following is a description of the method for manufacturing single crystal silicon of the present invention using drawings. FIG1 is a cross-sectional view of a single crystal pulling device for implementing the method for manufacturing single crystal silicon of the present invention. The single crystal pulling device 1 has a furnace body 10, which is formed by stacking a pulling chamber 10b on a cylindrical main chamber 10a; the furnace body 10 has: a carbon crucible (or black lead crucible) 2, which can rotate about a lead vertical axis and is configured to be able to rise and fall; and a quartz glass crucible 3 (hereinafter also referred to as crucible 3), which is held by the carbon crucible 2. The crucible 3 can rotate around the lead vertical axis as the carbon crucible 2 rotates.

In addition, a rotation drive unit 14, such as a rotation motor, is provided below the carbon crucible 2 to rotate the carbon crucible 2 around the lead vertical axis; and a lifting drive unit 15 is provided to lift and move the carbon crucible 2. In addition, the rotation drive unit 14 is connected to the rotation drive control unit 14a, and the lifting drive unit 15 is connected to the lifting drive control unit 15a.

In addition, the single crystal pulling device 1 includes: a side heater 4 that melts the semiconductor raw material (raw material polycrystalline silicon) loaded into the crucible 3 by resistance heating to form a silicon melt M (hereinafter also referred to as the melt M); and a pulling mechanism 9 that winds the wire 6 and pulls the grown single crystal C. The wire 6 of the pulling mechanism 9 has a seed crystal P at the tip thereof.

In addition, the side heater 4 is connected to a heater control unit 4a that controls the amount of power supplied, and the pulling mechanism 9 is connected to a rotation drive control unit 9a that controls the rotation drive of the pulling mechanism 9. In addition, in the present embodiment, in the single crystal pulling device 1, for example, an electromagnetic coil 8 for applying a magnetic field (horizontal magnetic field applying mechanism) is provided on the outer side of the furnace body 10. When a predetermined current is applied to the electromagnetic coil 8 for applying a magnetic field, a horizontal magnetic field of a predetermined intensity (1000 Gauss to 4000 Gauss) is applied to the silicon melt M in the crucible 3. The electromagnetic coil 8 for applying a magnetic field is connected to an electromagnetic coil control unit 8a that controls the movement of the electromagnetic coil 8 for applying a magnetic field. In addition, in this embodiment, the MCZ method (Magnetic field applied CZ method; magnetic field applied Czochralski method) is implemented, that is, a transverse magnetic field is applied in the melt M to grow a single crystal, thereby controlling the convection of the silicon melt M to achieve the stability of the single crystal.

A radiation shield 7 is arranged above the melt M formed in the crucible 3 to surround the single crystal C. The upper and lower parts of the radiation shield 7 are opened to shield the single crystal C being grown from the side heater 4, the melt M, etc., and to rectify the gas flow in the furnace. In addition, the gap between the lower end of the radiation shield 7 and the melt surface M1 is controlled and maintained at a predetermined distance (e.g., 50 mm) according to the required characteristics of the single crystal to be grown.

In addition, the single crystal pulling device 1 is equipped with an optical measuring sensor 16, such as a CCD (Charge-Coupled Device) camera, etc., for measuring the diameter of the single crystal silicon. A small window 10a1 for observation is provided on the upper surface of the main chamber 10a, and the position change of the solid-liquid interface is detected from the outside of the small window 10a1. In addition, based on the measured single crystal diameter and crystal length, the solidification rate represented by the single crystal weight/the weight of the initial silicon raw material is calculated, and the resistivity of the crystal is estimated.

Specifically, regarding the relationship between the concentration of the dopant and the resistivity, the single crystal has a dopant concentration distribution in its own longitudinal direction (the direction perpendicular to the lead when pulled). The dopant concentration distribution Cs when the solidification rate of silicon is set to g is expressed as the following formula (1). Cs=k×C ₀ ×(1-g) ^k-1 ･･･Formula (1) In formula (1), k is the equilibrium segregation coefficient, and C ₀ is the initial dopant concentration in the silicon melt. In addition, the equilibrium segregation coefficient of boron (B), which is generally used as the p-type dopant, is 0.8, and the equilibrium segregation coefficient of phosphorus (P), which is generally used as the n-type dopant, is 0.35.

To control the resistivity within the specified range, it is only necessary to adjust it in the following way: first calculate the relationship between the dopant concentration and the solidification rate in the growth of single crystal silicon, and adjust the dopant concentration based on these relationships, so that the resistivity of the single crystal is within the required range. For example, when phosphorus is used as a dopant and the resistivity of the single crystal head is set to the range of 20Ω・cm to 100Ω・cm, it is only necessary to add 0.1g to 3.5g of dopant (high concentration (10 ¹⁹ cm ^-3 level) phosphorus-containing silicon fragments with a resistivity of 1mΩ・cm to 5mΩ・cm) to a batch of about 150kg of silicon.

In addition, the single crystal pulling device 1 is equipped with: a dopant supply jig 17, which is used to supply a sheet-shaped dopant to the melt M. An opening 10a2 is provided on the upper surface of the main chamber 10a, and the opening 10a2 is connected to the lower end of the tube rod-shaped portion 18 of the dopant supply jig 17. A push rod 19 is provided in the tube rod-shaped portion 18, and the front end of the push rod 19 is spoon-shaped, and the push rod 19 can move forward and backward along the axis. A piston rod 20 is inserted into the push rod 19, and the push rod 19 is configured to move forward and backward along the piston rod 20 by driving the cylinder 21. The tube rod portion 18 is provided with a gate valve 22, and by opening the gate valve 22, the push rod 19 can move forward and enter the main chamber 10a. In addition, the cylinder 21 is connected to the cylinder drive portion 21a.

The push rod 19 is used to supply the flake dopant Dp to the silicon melt M. As shown in (a) and (b) of FIG. 2 , the front end of the push rod 19 is provided with a spoon 19a. The spoon 19a can carry 1 to 10 flake dopant Dp. The piston rod 20 and the push rod 19 are configured to be able to rotate around the axis by the rotation drive unit 25, and the flake dopant Dp loaded on the spoon 19a is freely fallen downward by the rotation of the push rod 19. In addition, the rotation drive unit 25 is connected to the rotation drive control unit 25a.

In addition, as shown in (a) and (b) of FIG. 2 , the dopant supply jig 17 has a funnel-shaped jig 23 mounted on the radiation shield 7 and made of quartz glass. As shown in the figure, the funnel-shaped jig 23 has a conical portion 23a having an opening and a capillary portion 23b extending from the front end (lower top) of the conical portion 23a in a foot shape. The opening portion 23a1 of the conical portion 23a is configured to open upward, and when the push rod 19 is inserted into the main chamber 19a to the deepest, the opening portion 23a1 of the conical portion 23a is located directly below the spoon portion 19a.

The opening diameter of the conical portion 23a is formed to be, for example, 50 mm to 100 mm so that all the flake dopants Dp falling from the spoon portion 19a can be received without overflowing, and the inner diameter of the capillary portion 23b is formed to be, for example, 10 mm to 15 mm so that the flake dopants Dp can pass through without being blocked. In addition, the distance d1 between the front end (lower end) of the capillary portion 23b and the melt surface M1 is controlled so that the state of the melt surface M1 is not adversely affected by the vertical movement of the radiation shield 7, and is controlled to be, for example, between 5 mm and 50 mm, preferably between 20 mm and 30 mm.

In addition, the tip of the capillary portion 23b is disposed at a position at least d2 (e.g., 1/10 of the crystal diameter) away from the outer surface of the single crystal C in the radial direction. This is because when the falling position of the flake dopant Dp is too close to the crystal (less than the distance d2 in the radial direction), it may adhere to the crystal before melting after falling. In addition, there is a risk that the amount of the secondary dopant introduced will increase locally because the melted dopant reaches the crystal interface before sufficient diffusion.

In addition, as shown in the top view of FIG. 3, in the horizontal magnetic field application condition of 1000 Gauss to 4000 Gauss as in the present embodiment, the front end position of the capillary portion 23b is arranged in the circumferential 45° region to the 135° region and the 225° region to the 315° region when the magnetic field application direction is defined as 0°. This is due to the following reasons. As shown by the arrow in FIG. 3, the flow direction of the melt surface changes, and in the vertical direction relative to the magnetic field application direction (0° direction), the outward flow from the crystal toward the quartz crucible is dominant. In contrast, in the magnetic field application direction, there is a portion of the flow toward the inside. In the former, the fallen dopant flakes melt along the outward flow, causing the dopant to diffuse, thus creating enough time for the added dopant to reach the solid-liquid interface.

However, in the latter, since the flow in the direction of crystal rotation is dominant and a part of the flow inward occurs, there is a risk that the fallen dopant flakes reach the crystal before melting. Furthermore, even if it melts, since the dopant easily reaches the solid-liquid interface before it fully diffuses, the secondary dopant is more likely to be introduced locally, making the resistivity distribution in the surface temporarily unstable. Therefore, the dopant drop position in the horizontal magnetic field is ideally set to: when the magnetic field application direction is defined as 0°, the 45° region to the 135° region and the 225° region to the 315° region. In addition, in the condition of no magnetic field and cusped geometry magnetic field, since there is no surface flow toward the inside in the entire circumference, it is not subject to this limitation.

In the tubular rod-shaped portion 18, a freely openable and closable doping agent supply opening portion 18a is provided above the gate valve 22, and is used to supply the flake doping agent Dp to the spoon portion 19a. That is, when the gate valve 22 is closed, the spoon portion 19a is arranged at the height position of the doping agent supply opening portion 18a, and 1 to 10 flake doping agents Dp are loaded on the spoon portion 19a from the doping agent supply opening portion 18a. When additional doping agent is supplied to the melt, the doping agent supply opening 18a is closed and the gate valve 22 is opened, and the push rod 19 is driven by the cylinder 21 to enter the main chamber 10a from the opening 10a2 from the front end side. At this time, the spoon 19a is in a state where the flake doping agent Dp is placed as shown in (a) of FIG. 2 . When the spoon 19a is located directly above the opening 23a1 of the conical portion 23a, the push rod 19 is rotated around the axis by the rotation drive 25, and the flake doping agent Dp falls from the spoon 19a to the funnel-shaped jig 23 as shown in (b) of FIG. 2 .

In addition, in this embodiment, the flake dopant Dp introduced into the melt M is obtained by splitting a silicon wafer having a thickness of 500 μm or more and 1000 μm or less, which is obtained by slicing single crystal silicon containing a secondary dopant or single crystal silicon containing a primary dopant, and the flake is used as an additive. The resistivity of the single crystal silicon used as the flake dopant Dp for the dopant is measured and processed into a desired size. The dopant concentration is calculated based on the resistivity, and the amount of the added dopant can be managed by the weight of the flake.

More specifically, the flake dopant Dp needs to have a minimum weight so that when it is put into the melt surface M1, it will not be discharged out of the chamber by the inert gas directly above the melt surface M1. Therefore, ideally, the surface area of each flake is 4 ^mm2 or more. However, when the size of the flake is too large, it takes a lot of time to melt, and the risk of attachment to the single crystal during the growth process increases, so it is ideal to be less than 25 ^mm2 . Similarly, from the perspective of weight and solubility, the thickness of the flake is also ideally 500 μm or more and 1000 μm or less.

In addition, the single crystal pulling device 1 is equipped with a computer 11, which has a storage device 11a and an algorithm control device 11b, and a rotation drive control unit 14a, a lifting drive control unit 15a, an electromagnetic coil control unit 8a, a rotation drive control unit 9a, a measuring sensor 16, a cylinder drive unit 21a and a rotation drive control unit 25a are respectively connected to the algorithm control device 11b.

In the single crystal pulling device 1 constructed in this way, when growing a single crystal C with a diameter of 200 mm, for example, the pulling is performed in the following manner. That is, the raw material polycrystalline silicon (for example, 150 kg) and the silicon wafer body for adding the dopant are initially loaded into the crucible 3, and the crystal growing process is started according to the program stored in the storage device 11a of the computer 11. In addition, when manufacturing n-type single crystal silicon, a silicon wafer body containing P (phosphorus) is used as the main dopant.

Next, the furnace body 10 is set to a predetermined atmosphere (mainly an inert gas such as argon). For example, a furnace atmosphere with a furnace pressure of 60 torr to 110 torr and an argon flow rate of 40 l/min to 110 l/min is formed. Then, while the crucible 3 is rotating in a predetermined direction at a predetermined rotation speed (rpm), the raw polycrystalline silicon and the main dopant loaded into the crucible 3 are melted by the heating performed by the side heater 4 to form a melt M (step S1 in Figure 4). In addition, in step S1, the raw polycrystalline silicon can be melted while the silicon wafer for doping can be put into the crucible 3.

Next, a predetermined current is passed to the electromagnetic coil 8 for applying a magnetic field, and a horizontal magnetic field is applied to the melt M at a magnetic flux density set to a range of 1000 gauss to 4000 gauss (e.g., 3000 gauss) (step S2 in FIG. 4 ). In addition, the pulling conditions are adjusted using the power supplied to the side heater 4, the pulling speed, the strength of the magnetic field applied, etc. as parameters, so that the seed crystal P starts to rotate around the axis at a predetermined rotation speed. The direction of rotation is set to be opposite to the direction of rotation of the crucible 3. Then, the wire 6 is lowered so that the seed crystal P contacts the melt M, and after the front end of the seed crystal P is melted, it is necked to form a neck P1.

Then, the single crystal pulling process begins. That is, the crystal diameter is gradually enlarged to form the shoulder C1, and then the process of forming the straight body C2 as the product part is moved to (step S3 in FIG. 4 ). When the growth of the single crystal C begins, the computer 11 uses the measurement results of the measuring sensor 16 to calculate the solidification rate of the single crystal silicon, and estimates the resistivity of the pulled single crystal based on this (step S4 in FIG. 4 ). Here, when the resistivity drops below the critical value of the lower limit of the specification range, B (boron) is added to the melt surface M1 as a secondary dopant having a conductivity type opposite to that of the main dopant (step S8 in FIG. 4 ). That is, in order to prevent the resistivity from decreasing as the length of the single crystal silicon increases and the solidification rate increases, so that the resistivity is lower than the specification value, the doping amount of the secondary dopant is adjusted to an appropriate amount relative to the solidification rate continuously or intermittently while adding. For example, when the upper limit value of the resistance specification is set to 60Ω and the lower limit value is set to 50Ω, it is sufficient to add the secondary dopant to such an extent that the estimated resistance value becomes, for example, the lower limit value + 1% of the lower limit value (in this case, 50.5Ω). In addition, if the resistance target value is set to, for example, the upper limit value - 1% of the upper limit value (in this case, 59.4Ω), it is sufficient to add the flake dopant Dp so that the resistance value is close to this value.

At this time, in the dopant supply jig 17, the dopant supply opening 18a is opened, and 1 to 10 flake dopant Dp containing B (boron) are placed on the spoon 19a. Then, the dopant supply opening 18a is closed, and after the gas in the tube rod 18 is replaced by the auxiliary pump (not shown), the gate valve 22 is opened, and the cylinder 21 is driven to allow the push rod 19 to enter the main chamber 10a from the opening 10a2 from the front end side. At this time, the flake dopant Dp is placed on the spoon 19a. When the spoon 19a is located directly above the opening 23a1 of the cone 23a of the funnel-shaped jig 23, the push rod 19 is rotated around the axis by the rotation drive 25, and the flake dopant Dp falls from the spoon 19a to the opening 23a1 of the cone 23a in the funnel-shaped jig 23. The flake dopant Dp accelerated by the free fall passes through the capillary 23b of the funnel-shaped jig 23 and is added to the melt. This step S8 is repeated every time the estimated resistivity of the single crystal C approaches the lower limit of the specification range.

In the above manner, the resistivity along the axis of the single crystal C is maintained within the specification range. In addition, since the flake dopant Dp is supplied to the melt surface M1 via the funnel-shaped jig 23, it can be accurately placed at a predetermined position on the melt surface M1. In addition, by placing the flake dopant Dp at a precise position, it is possible to suppress the flake dopant Dp from flowing along the melt surface and adhering to the growing crystal and causing dislocation.

In addition, when the single crystal is continuously grown and pulled to the required length without misalignment, the single crystal growth is completed (steps S6 and S7 in FIG. 4 ). That is, when the straight body C2 is formed to the required length, it moves to the final tail process, and in this tail process, the contact area between the lower end of the crystal and the melt M is gradually reduced to separate the single crystal C from the melt M and produce single crystal silicon.

On the other hand, when dislocation is formed during the growth of the single crystal (step S6 in FIG. 4), if the solidification rate of the single crystal grown so far does not reach the specified value (step S9 in FIG. 5), the crystal is melted again (step S10 in FIG. 5). After the remelting, the dopant supply jig 17 is used to add a flake dopant Dp of P (phosphorus) as the main dopant based on the solidification rate of the remelted single crystal to the melt surface M1 (step S11 in FIG. 5). This can fully compensate for the phosphorus concentration in the melt M. Thereafter, return to step S2 in FIG. 4, and pull the single crystal C again.

As described above, according to the present embodiment, when a single crystal silicon is pulled from a silicon melt to which a main dopant (phosphorus) is added for growth, a flake-shaped secondary dopant (boron) having a conductivity type opposite to that of the main dopant is continuously or intermittently added to the silicon melt, so that the resistivity along the axial direction of the single crystal C can be maintained within the specification range. In particular, when the flake dopant Dp is added to the silicon melt M, the flake dopant Dp is dropped to the correct position of the melt surface M1 with precise positioning through the funnel-shaped jig disposed in the chamber, thereby preventing the flake dopant Dp from flowing along the melt surface and adhering to the growing crystal to cause misalignment. In addition, since one or more flake dopant Dp placed on the front end of the push rod 19 is placed in the funnel-shaped jig 23, the flake dopant Dp is supplied to the funnel-shaped jig 23 by free fall, passes through the capillary portion 23b in an increasing state, and drops to the melt surface M1 of the silicon melt M. Therefore, the dopant can be added to the silicon melt M without clogging the narrow tube portion 23b of the funnel-shaped fixture 23. In addition, since the structure for adding the dopant during the growth process of the single crystal C can be realized with a relatively simple structure and the required amount of dopant can be added on a sheet basis, precise resistivity control can be performed.

In addition, in the above-mentioned embodiment, an example of manufacturing n-type single crystal silicon is described, but the present invention is not limited to this, and it can also be applied to the case of manufacturing p-type single crystal silicon. In addition, in the above-mentioned embodiment, a structure for applying a horizontal magnetic field to the silicon melt is made, but the present invention is not limited to a horizontal magnetic field, and it can also be applied to a structure with a pointed magnetic field or no magnetic field.

[Example] The method for producing single crystal silicon of the present invention is further described according to the example. [Experiment 1] In Experiment 1, 135 kg of silicon raw material was filled into a quartz crucible, and phosphorus was added as a main dopant and melted. In addition, the distance between the radiation shield and the melt surface was set to 50 mm, the furnace internal pressure was set to 65 Torr, the argon gas was flowed at a flow rate of 90 l/min, and the furnace internal environment was prepared with a transverse magnetic field strength of 3000 Gauss. Then, the crucible rotation speed was set to 0.5rpm, and the crystal rotation speed as a reference was set to 10.0rpm (in the opposite direction of the crucible rotation), and single crystals with a target crystal diameter of 200mm were grown at a pulling speed of 0.55mm/min. In addition, the resistivity specification was set to 60Ωcm to 50Ωcm, and the target resistivity when the straight body was formed was set to 59Ωcm.

[Example 1] In Example 1, single crystal growth is performed according to the above-mentioned embodiment. The timing and amount of introduction of the secondary dopant are calculated based on the solidification rate of the crystal, the phosphorus in the melt, and the boron concentration in the sheet. When approaching the lower limit of the resistivity specification, the flake dopant (secondary dopant) with the required boron concentration is placed on the spoon at the front end of the push rod. After the gate is released, the front end of the spoon is brought close to a height of 10 mm from the upper opening, and the spoon is rotated to put the flake dopant into the funnel-shaped fixture.

The flake doping agent passes through the inner part of the thin tube of the funnel-shaped fixture and falls to the surface of the melt to melt. After falling, no wave front or solidification on the surface of the melt is generated, and good single crystal growth can be performed. This operation was repeated five times while estimating the resistivity, and no misalignment occurred, and the operation was completed without misalignment. Figure 6 shows the change of resistivity in the direction of crystal growth. As shown in Figure 6, each time reverse doping is repeated, the resistivity that has dropped to the lower limit of the specification can be raised to the upper limit of the specification. It was confirmed that this can greatly improve the yield of resistivity even for products with narrow resistivity.

[Comparative Example 1] In Comparative Example 1, no secondary dopant was added during the growth of single crystal silicon. The other conditions were the same as those of Example 1. The results of this Comparative Example 1 are shown in FIG6. As shown in FIG6, it was confirmed that when no secondary dopant was added as in Comparative Example 1, as the curing rate increased, the resistivity along the axial direction decreased and did not meet the specification.

[Experiment 2] In Experiment 2, the pulling was started under the same pulling conditions as in Example 1, and at the time point when reverse doping (addition of secondary dopant) was performed twice, the growth of the straight body was interrupted and the crystal was melted again. Then, after the remelting, the phosphorus atom deficiency caused by the boron added in the reverse doping was added by the flake silicon of the main dopant. Thereafter, the pulling of the single crystal silicon was started, and the resistivity at the beginning of the formation of the straight body was estimated. Examples 2 to 4 were implemented under the same conditions, and the resistivity at the beginning of the formation of the straight body before the pulling was interrupted (before remelting) and the resistivity at the beginning of the formation of the straight body after remelting were estimated and verified respectively. The results of Examples 2 to 4 are shown in Table 1. As shown in Table 1, the resistivity before and after re-melting can be within the specification range, and the resistivity at the beginning of the straight body formation can be stabilized.

[Table 1] Embodiment 2 Embodiment 3 Embodiment 4 average Standard Deviation Before melting again 60.30Ωcm 58.74Ωcm 58.87Ωcm 59.30Ωcm 0.87 After melting again 59.11Ωcm 60.17Ωcm 58.67Ωcm 59.31Ωcm 0.75

[Experiment 3] In Experiment 3, single crystal pulling was performed under the same conditions as in Example 1 in a horizontal magnetic field 3D (three-dimensional) computer simulation, and the flow direction of the melt surface during the pulling process was verified. In addition, as pulling conditions, the flow of single crystals with diameters of 200 mm, 300 mm, and 450 mm on the melt surface was observed. As a result, in any of the above crystal diameters, the flow direction of the melt surface changed, and in the direction perpendicular to the magnetic field application direction (0° direction), the outward flow from the crystal toward the quartz crucible was dominant. In contrast, there was a portion of inward flow in the magnetic field application direction.

It can be seen from this that if the dopant flakes are dropped into an area where the outward flow from the crystal toward the quartz is dominant, the dropped dopant flakes will diffuse while melting along the outward flow, creating sufficient time for the added dopant to reach the solid-liquid interface. On the other hand, if the dopant flakes are dropped into an area where a portion of the inward flow exists, there is a risk that the dropped dopant flakes will reach the crystal before melting. In addition, it is determined that even if there is a molten dopant, it is easy to reach the solid-liquid interface before it is fully diffused, so the secondary dopant is introduced locally more often, making the resistivity distribution in the surface temporarily unstable. Therefore, based on the results of Experiment 3, it was confirmed that the ideal drop position of the dopant in the horizontal magnetic field is: the 45° area to the 135° area when the magnetic field application direction is defined as 0°, and the 225° area to the 315° area.

1: Single crystal pulling device 2: Carbon crucible 3: Crucible (quartz glass crucible) 4: Side heater 4a: Heater control unit 6: Wire 7: Radiation shield 8: Electromagnetic coil for magnetic field application 8a: Electromagnetic coil control unit 9: Pulling mechanism 9a: Rotation drive control unit 10: Furnace body 10a: Main chamber 10a1: Small window 10a2: Opening 10b: Pulling chamber 11: Computer 11a: Storage device 11b: Calculation control device 14: Rotation drive unit 14a: Rotation drive control unit 15: Lifting drive unit 15a: Lifting drive control unit 16: Measuring sensor 17: Dopant supply jig 18: Tube rod 18a: Dopant supply opening 19: Push rod 19a: Spoon 20: Piston rod 21: Cylinder 21a: Cylinder drive 22: Gate valve 23: Funnel jig 23a: Cone 23b: Capillary 23a1: Opening 25: Rotation drive 25a: Rotation drive control C: Single crystal C1: Shoulder C2: Straight body d1: Distance d2: Distance Dp: Flake dopant M: Silicon melt (melt) M1: Melt surface P: Seeding S1 to S11: Steps

[Fig. 1] is a cross-sectional view of a single crystal pulling device for implementing the single crystal silicon manufacturing method of the present invention. [Fig. 2] (a) and (b) are cross-sectional views showing a portion of the single crystal pulling device in Fig. 1 in an enlarged manner. [Fig. 3] is a top view showing the flow direction of the melt surface due to the influence of the horizontal magnetic field. [Fig. 4] is a flow chart of the single crystal silicon manufacturing method of the present invention. [Fig. 5] is a flow chart following the flow chart of Fig. 4. [Fig. 6] is a graph showing the results of Experiment 1 of the embodiment.

1: Single crystal pulling device

2: Carbon crucible

3: Crucible (quartz glass crucible)

4: Side heater

4a: Heater control unit

6: Wires

7: Radiation protection

8: Electromagnetic coil for applying magnetic field

8a: Electromagnetic coil control unit

9: Lifting mechanism

9a: Rotation drive control unit

10: Furnace body

10a: Main chamber

10a1: Small window

10a2: Opening

10b: Pulling room

11: Computer

11a: Storage device

11b: Calculation control device

14: Rotary drive unit

14a: Rotation drive control unit

15: Lifting drive unit

15a: Lifting drive control unit

16: Measurement sensor

17: Doping agent supply jig

18: Cylinder tube rod-shaped part

18a: Opening for supplying doping agent

19: Push rod

19a: Spoon

20: Piston rod

21: Cylinder

21a: Cylinder drive unit

22: Gate valve

23: Funnel-shaped fixture

25: Rotary drive unit

25a: Rotation drive control unit

C: Single crystal

C1: Shoulder

C2: Straight body

M: Silicon melt (melt)

M1: Melt surface

P: Seed crystal

Claims (4)

A method for producing single crystal silicon comprises: heating a silicon melt in a crucible by a heater in a chamber, and growing the single crystal silicon by the Czochralski method; the method for producing single crystal silicon comprises: adding a main dopant to the silicon melt when the silicon melt is formed in the crucible; The method comprises the steps of: growing the single crystal silicon in a silicon melt; and continuously or intermittently adding a secondary dopant having a conductivity type opposite to that of the primary dopant to the silicon melt during the growing of the single crystal silicon; and in the step of adding the secondary dopant to the silicon melt, performing: adding one or more flake dopants as a precursor The secondary dopant is transported into the aforementioned chamber; the aforementioned flake dopant is dropped onto the funnel-shaped fixture, and the aforementioned flake dopant is added into the aforementioned silicon melt via a capillary disposed on the aforementioned funnel-shaped fixture; before the step of cultivating the aforementioned single crystal silicon, a step of applying a horizontal magnetic field to the aforementioned silicon melt in the aforementioned crucible is performed; during the step of cultivating the aforementioned single crystal silicon, in the step of adding a secondary dopant having a conductivity type opposite to that of the aforementioned main dopant into the aforementioned silicon melt, the dopant drop position in the horizontal magnetic field is set to the 45° region to the 135° region and the 225° region to the 315° region when the magnetic field application direction is defined as 0°. A method for manufacturing single crystal silicon as described in claim 1, wherein in the step of growing the single crystal silicon, after the step of adding the secondary dopant to the silicon melt, when the single crystal silicon forms a dislocation, the following steps are performed: remelting the single crystal silicon with the dislocation formed; adding a flake dopant of the main dopant to the silicon melt; regrowing single crystal silicon from the silicon melt; and in the step of regrowing the single crystal silicon, continuously or intermittently adding the secondary dopant having a conductivity type opposite to that of the main dopant to the silicon melt. A single crystal pulling device is provided, wherein a silicon melt is formed in a crucible which is arranged in a chamber and heated by a heater, and a single crystal silicon is pulled by the Czochralski method; the single crystal pulling device is provided with: a shielding protective member which is arranged to surround the single crystal silicon grown above the crucible; a funnel-shaped fixture which is mounted on the shielding protective member and is composed of a cone portion and a thin tube, wherein the cone portion is open at the top and the thin tube extends downward from the lower top of the cone portion; and a push rod which is used to put one or more flake-shaped dopants into the funnel. The opening of the cone portion in the funnel-shaped fixture; the aforementioned flake dopant introduced into the opening of the cone portion in the funnel-shaped fixture is dropped into the aforementioned silicon melt through the aforementioned capillary of the funnel-shaped fixture; the aforementioned single crystal pulling device is further equipped with: a horizontal magnetic field applying mechanism, which applies a horizontal magnetic field to the aforementioned silicon melt in the aforementioned crucible; the aforementioned funnel-shaped fixture is arranged in the 45° area to the 135° area and the 225° area to the 315° area when the dopant drop position in the horizontal magnetic field is set to the 0° defined magnetic field application direction. A single crystal pulling device forms a silicon melt in a crucible that is arranged in a chamber and heated by a heater, and pulls a single crystal silicon by the Czochralski method; the single crystal pulling device is provided with: A shielding protective member that is arranged to surround the cultivated single crystal silicon above the crucible; a funnel-shaped fixture that is mounted on the shielding protective member and is composed of a cone and a thin tube, the cone is open at the top, and the thin tube extends downward from the bottom of the cone; and a push rod that is configured to be able to move forward and backward relative to the chamber and to put one or more sheet-shaped dopants placed on the front end into the front end of the funnel-shaped fixture. The opening of the cone portion; the flake dopant dropped into the silicon melt through the capillary of the funnel-shaped fixture before being put into the opening of the cone portion; the single crystal pulling device is further equipped with: a horizontal magnetic field applying mechanism, which applies a horizontal magnetic field to the silicon melt in the crucible; the funnel-shaped fixture is arranged in the 45° area to the 135° area and the 225° area to the 315° area when the dopant drop position in the horizontal magnetic field is set to the 0° magnetic field application direction; wherein the front end of the push rod is formed into a spoon shape, and the push rod is configured to be rotatable around an axis.