Method for producing silicon single crystal

Abstract

To provide a method for producing a silicon single crystal by which the oxygen concentration of the silicon single crystal can be estimated with high accuracy.SOLUTION: The method includes a step of melting a silicon material to form a silicon melt M in a silica glass crucible 2, a step of measuring a surface temperature of the silicon melt in the silica glass crucible during a predetermined period from formation of the silicon melt to start of formation of a shoulder portion of a silicon single crystal C, a step of obtaining a temperature variation during the predetermined period from the measured surface temperature of the silicon melt, and a step of determining a pulling condition of a straight body portion of the silicon single crystal based on the temperature variation.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a method for producing a silicon single crystal, and more particularly to a method for producing a silicon single crystal that can keep the oxygen concentration along the crystal growth direction of the silicon single crystal grown by the Czochralski method within a specified range. [Background technology]

[0002] In growing silicon single crystals by the Czochralski method (CZ method), as shown in FIG. 7, a quartz glass crucible 51 placed in a chamber 50 is filled with polysilicon as a raw material, and the polysilicon is heated and melted by a heater 52 provided around the quartz glass crucible 51 to form a silicon melt M. Thereafter, a seed crystal P (seed) attached to a seed chuck is immersed in silicon melt M, and the seed chuck and quartz glass crucible 51 are rotated in the same direction or in opposite directions. When the seed crystal P is immersed in the silicon melt M, dislocations occur due to thermal shock, so a drawing process called necking (a process of forming a neck portion CN) is carried out to prevent the dislocations from affecting the subsequent crystal growth. Subsequently, the crystal diameter is gradually increased to grow a shoulder portion C1, and then the diameter is increased to a predetermined straight body diameter to move on to the formation of a straight body portion C2, and the silicon single crystal C is pulled and grown.

[0003] Generally, most of the hot zone inside the apparatus shown in Figure 7 is composed of graphite parts. It is known that the graphite heater 52, which reaches particularly high temperatures, and the graphite crucible 53 that holds the quartz glass crucible 51 react with oxygen that dissolves from the quartz glass crucible 51 and evaporates from the free surface of the silicon melt, causing their wall thickness to thin over time. As a result, changes in the thermal environment due to changes in the heater heat distribution and changes in the melt convection due to changes in the degree of deformation of the quartz glass crucible 51 caused by changes in the shape of the graphite crucible occur. In addition, as hot zone parts become larger, their dimensional tolerances become larger, which can lead to misalignment when assembled, resulting in variations in the thermal environment. In particular, when hot zone components become larger, the positional deviation of the hot zone components increases due to aging, which changes the thermal environment and melt convection, resulting in variations in the oxygen concentration of the silicon melt.

[0004] In order to reduce such variations in oxygen concentration, Patent Document 1 discloses a technology in which a horizontal magnetic field is applied, and then the melt surface temperature is measured with a radiation thermometer to estimate the melt convection pattern and thereby estimate the oxygen concentration.

[0005] Furthermore, Patent Document 2 discloses a technique in which the surface temperatures of silicon melt at two specific locations are measured using a radiation thermometer, the convection mode is estimated based on the magnitude of the measured temperatures, and the oxygen concentration is estimated based on the estimated convection mode to set pulling conditions. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Publication No. 2019-151499 [Patent Document 2] JP 2019-151500 A Summary of the Invention [Problem to be solved by the invention]

[0007] However, the techniques for estimating the oxygen concentration of silicon melt disclosed in Patent Documents 1 and 2 require a step of confirming that the silicon melt is forming a desired convection pattern in order to measure the surface temperature of the silicon melt, which increases the number of steps. Furthermore, if the desired convection pattern is not formed, there remains a concern that productivity may decrease due to the need to reform the convection pattern of the silicon melt to form the desired convection pattern. Furthermore, since there are usually several convection patterns in silicon melt, there is a problem in that it is difficult to make a highly accurate estimation using a method of inferring the convection pattern and then estimating the oxygen concentration from that.

[0008] The inventors of the present invention have conducted extensive research to solve the above problems, focusing on the variation in the surface temperature of a silicon melt. They then performed frequency analysis using fast Fourier transform on the measured temperature of the silicon melt surface, and found that the variation in the surface temperature of the silicon melt (temperature fluctuation) is significantly related to the time interval obtained by dividing the crucible rotation period (seconds) by the number of divisions of the graphite crucible. It was estimated that this is because the thermal environment inside an actual single crystal pulling device is not axially symmetric, and the crucible (graphite crucible, quartz glass crucible) rotates while there is a temperature gradient within the silicon melt surface, and the divided parts of the graphite crucible, which are susceptible to thermal changes, alternately experience low-temperature and high-temperature parts, resulting in periodic temperature fluctuations. The number of divisions of the graphite crucible described above refers to the number of divisions into which the graphite crucible is divided in the vertical direction. The graphite crucible is divided into a plurality of divisions in the vertical direction, for example, 2 to 5, in order to reduce damage due to thermal expansion.

[0009] It was also found that the magnitude of the variation in the surface temperature of the silicon melt correlates with the thickness of the graphite crucible, which indicates the degree of wear. In other words, the magnitude of the variation in the surface temperature of the silicon melt serves as an indicator of the magnitude of the change in the thermal environment, the degree of wear of the graphite crucible, the degree of deformation of the quartz glass crucible, etc. The present inventors discovered that the magnitude of the variation in the surface temperature of the silicon melt correlates with the oxygen concentration in the crystal, leading to the invention of the present invention.

[0010] The present invention has been made in light of the above circumstances, and has as its object to provide a method for producing a silicon single crystal that is capable of estimating the oxygen concentration of the silicon single crystal with high accuracy. [Means for solving the problem]

[0011] The method for producing a silicon single crystal according to the present invention, which has been made to solve the above-mentioned problems, is a method for producing a silicon single crystal by pulling up a silicon single crystal by the Czochralski method, and is characterized by comprising the steps of: melting a silicon raw material to form a silicon melt in a quartz glass crucible; measuring the surface temperature of the silicon melt in the quartz glass crucible during a predetermined period from the formation of the silicon melt to the start of shoulder formation of the silicon single crystal; determining the temperature variation during the predetermined period from the temperature data related to the measured surface temperature of the silicon melt and estimating the oxygen concentration of the silicon single crystal based on the temperature variation; and determining the pulling conditions for the straight body portion of the silicon single crystal based on the temperature variation.

[0012] In addition, in the process of estimating the oxygen concentration of the silicon single crystal based on the temperature variation, it is desirable to estimate the oxygen concentration of the silicon single crystal based on the temperature variation over the predetermined period and a pre-recorded correlation between the oxygen concentration of the silicon single crystal and the temperature variation of the silicon melt. Furthermore, it is desirable that the predetermined period for determining the temperature variation be during the stabilization step, in which the silicon seed crystal is immersed in the silicon melt and a magnetic field is applied to the silicon melt.

[0013] Furthermore, it is desirable to use the variance, standard deviation, or peak intensity of the Fourier transform spectrum over a predetermined time period as the data relating to the temperature variation. Furthermore, it is desirable that the sampling interval Ts (s) for measuring temperature data relating to the surface temperature of the silicon melt is determined so that Ts≦30 / N·CR holds, where CR (rpm) is the crucible rotation speed and N is the number of divisions of the graphite crucible that holds the silica glass crucible.

[0014] Furthermore, it is desirable that the quartz glass crucible is placed inside a graphite crucible divided into multiple parts, and in the process of determining the temperature variation over the predetermined period from temperature data related to the measured surface temperature of the silicon melt and estimating the oxygen concentration of the silicon single crystal based on the temperature variation, after determining the temperature variation over the predetermined period from the temperature data related to the measured surface temperature of the silicon melt, the process further comprises a step of linking the value of the data related to the temperature variation with deterioration information of the graphite crucible to manage the graphite crucible. Furthermore, it is desirable that the specified period for determining the temperature variation be during the stabilization process, and that measurement of the surface temperature of the silicon melt in the quartz glass crucible be started during the stabilization process, and that the process of estimating the oxygen concentration and the process of determining the pulling conditions be performed, and then measurement of the surface temperature of the silicon melt in the quartz glass crucible be completed during the stabilization process.

[0015] According to the present invention, the surface temperature of the silicon melt is measured during a predetermined period of time after the formation of the silicon melt and before the formation of the straight body portion, and the temperature variation of the measured surface temperature of the silicon melt is determined. The oxygen concentration of the silicon single crystal to be pulled is then estimated based on a prerecorded correlation between the oxygen concentration of silicon single crystals pulled under the same pulling conditions and the temperature variation of the silicon melt. If it is determined that adjustment of the pulling conditions is necessary based on the estimated oxygen concentration, pulling conditions that affect the oxygen concentration, such as heater temperature, crucible rotation speed, and furnace pressure, can be adjusted to produce a silicon single crystal having the target oxygen concentration. [Effects of the Invention]

[0016] According to the present invention, it is possible to provide a method for producing a silicon single crystal that is capable of estimating the oxygen concentration of the silicon single crystal with high accuracy before growing the body portion of the silicon single crystal. [Brief explanation of the drawings]

[0017] [Figure 1]FIG. 1 is a cross-sectional view showing an example of a single crystal pulling apparatus for carrying out the method for producing a silicon single crystal according to the present invention. [Figure 2] FIG. 2 is a flow chart showing the flow of an embodiment of the method for producing a silicon single crystal according to the present invention. [Figure 3] FIG. 3 is a graph showing an example of the correlation between the oxygen concentration of a silicon single crystal and the surface temperature variation of a silicon melt. [Figure 4] FIG. 4 is a graph showing the results of an example of the present invention. [Figure 5] FIG. 5 is a graph showing an example of adjustment of the pulling conditions in an embodiment of the present invention. [Figure 6] FIG. 6 is a graph showing other results of an example of the present invention. [Figure 7] FIG. 7 is a cross-sectional view of a typical single crystal pulling apparatus. DETAILED DESCRIPTION OF THE INVENTION

[0018] Hereinafter, an embodiment of a method for producing a silicon single crystal according to the present invention will be described with reference to the drawings. However, the present embodiment will be described as an example of the present invention, and the present invention is not limited thereto.

[0019] FIG. 1 is a cross-sectional view showing an example of a single crystal pulling apparatus for carrying out the method for producing a silicon single crystal according to the present invention. This single crystal pulling apparatus 1 includes a furnace body 10 formed by stacking a pull chamber 10b on a cylindrical main chamber 10a, a graphite crucible 2 provided in this furnace body 10 so as to be rotatable about a vertical axis and movable up and down, and a silica glass crucible 3 (hereinafter simply referred to as crucible 3) held by the graphite crucible 2. This crucible 3 is rotatable about the vertical axis together with the rotation of the graphite crucible 2. The graphite crucible 2 is divided into multiple sections in the vertical direction (vertical direction).

[0020] Further, below the graphite crucible 2, there are provided a rotation drive unit 14 such as a rotation motor that rotates the graphite crucible 2 around a vertical axis, and an elevation drive unit 15 that moves the graphite crucible 2 up and down. The rotation drive unit 14 is connected to a rotation drive control unit 14a, and the elevation drive unit 15 is connected to an elevation drive control unit 15a.

[0021] The single crystal pulling apparatus 1 also includes a side heater 4 and a bottom heater 5, which are resistance heating or high-frequency induction heating types, for heating and melting the semiconductor raw material (raw material polysilicon) loaded in the crucible 3 to form a silicon melt M. As shown in the figure, the side heater 4 is disposed so as to surround the crucible 3 from the side thereof, and the bottom heater 5 is disposed below the crucible 3. The heating output of the side heater 4 is controlled by a side heater control unit 4a, and the heating output of the bottom heater 5 is controlled by a bottom heater control unit 5a.

[0022] Furthermore, in this single crystal pulling apparatus 1, a magnetic field applying electromagnetic coil 8 is installed outside the furnace body 10. When a predetermined current is applied to this magnetic field applying electromagnetic coil 8, a horizontal magnetic field of predetermined strength is applied to the silicon melt M in the crucible 3. An electromagnetic coil control unit 8a (magnetic field application control unit) that controls the operation of the magnetic field applying electromagnetic coil 8 is connected to the magnetic field applying electromagnetic coil 8.

[0023] That is, in this embodiment, the MCZ method (Magnetic field applied CZ method) is carried out, in which a horizontal magnetic field is applied to the silicon melt M to grow a single crystal C, thereby controlling the convection of the silicon melt M and stabilizing the single crystallization.

[0024] The single crystal pulling apparatus 1 also includes a pulling mechanism 9 that winds up a wire 6 and pulls up the grown single crystal C. A seed crystal P is attached to the tip of the wire 6 of the pulling mechanism 9. A rotation drive control unit 9a that controls the rotation drive of the pulling mechanism 9 is connected to the pulling mechanism 9.

[0025] A radiation shield 7 is disposed above the silicon melt M formed in the crucible 3, surrounding the single crystal C. The radiation shield 7 has openings at the top and bottom, and serves to shield the single crystal C during growth from excess radiant heat from the side heaters 4, the silicon melt M, etc., and also to straighten the gas flow inside the furnace.

[0026] The single crystal pulling apparatus 1 also includes a radiation thermometer 16 for measuring the surface temperature of the silicon melt M. This radiation thermometer 16 can be, for example, a two-dimensional radiation thermometer that can simultaneously measure multiple measurement points on the surface of the silicon melt M to obtain the temperature distribution on the melt surface. A small observation window 10a1 is provided on the top surface of the main chamber 10a, and the radiation thermometer 16 measures the surface temperature of the silicon melt M from the outside of this small window 10a1 in a non-contact manner.

[0027] The single crystal pulling apparatus 1 also includes a controller 11 having a storage device 11a and an arithmetic and control device 11b, and the side heater control unit 4a, the bottom heater control unit 5a, the electromagnetic coil control unit 8a, the rotation drive control unit 14a, the elevation drive control unit 15a, the rotation drive control unit 9a, and the radiation thermometer 16 are each connected to the arithmetic and control device 11b.

[0028] In the single crystal pulling apparatus 1 configured as above, when growing a single crystal C having a diameter of, for example, 200 mm, pulling is performed as follows. That is, first, silicon raw material (for example, 150 kg of raw polysilicon) is loaded into the crucible 3, and a crystal growing process is started based on a program stored in the storage device 11a of the controller 11.

[0029] First, a predetermined atmosphere (mainly an inert gas such as argon gas) is created inside the furnace body 10. For example, a furnace atmosphere with an internal furnace pressure of 65 torr and an argon gas flow rate of 90 liters / minute is created. Then, while the crucible 3 is rotated in a predetermined direction at a high rotation speed, for example, 15 rpm, the silicon raw material loaded in the crucible 3 is melted by heating with the side heater 4 and the bottom heater 5 to form a silicon melt M (step S1 in Figure 2).

[0030] When the silicon melt M is formed, a predetermined current is passed through the magnetic field applying electromagnetic coil 8, and application of a horizontal magnetic field to the silicon melt M begins in order to suppress convection (step S2 in FIG. 2).

[0031] After the application of the horizontal magnetic field begins, the radiation thermometer 16 starts measuring the surface temperature of the silicon melt M, and each time the measurement data is updated, it is sent to the controller 11 (step S3 in FIG. 2). The sampling interval Ts (s) of the measurement by the radiation thermometer 16 is set so that Ts≦30 / N CR holds, where CR (rpm) is the crucible rotation speed and N is the number of divisions of the graphite crucible 2. This is because the variation in the surface temperature of the silicon melt M is closely related to the frequency obtained by dividing the crucible rotation speed (rpm) by the number of divisions of the graphite crucible, and so it is necessary to set a sampling frequency that can capture this fluctuation period. According to the sampling theorem, the required sampling frequency fs must be at least twice the frequency obtained by dividing the crucible rotation speed by the number of divisions of the crucible, so the following formula (1) is established: fs≧2 CR N=CR N / 30 Formula (1) Since fs=1 / Ts, the sampling interval Ts is given by the following equation (2). Ts≦30 / N CR...Formula (2)

[0032] Furthermore, the pulling conditions are adjusted using parameters such as the initial power supply to the side heater 4 and the bottom heater 5, and the pulling speed, and the seed crystal P starts to rotate around its axis at a predetermined rotation speed. The rotation direction is set to the opposite direction to the rotation direction of the crucible 3. This starts the step of stabilizing the temperature of the silicon melt M, i.e., the stabilization step (step S4 in FIG. 2).

[0033] Next, the wire 6 is lowered to bring the seed crystal P into contact with the silicon melt M, and after the tip of the seed crystal P is melted, necking occurs to form a neck portion CN (step S5 in FIG. 2). Once the neck portion CN is formed, the formation of a shoulder portion C1, which gradually expands the crystal diameter, begins (step S6 in FIG. 2).

[0034] Meanwhile, the controller 11 calculates the temperature variation (e.g., standard deviation) of the surface temperature of the silicon melt M measured by the radiation thermometer 16 during a predetermined period from the formation of the silicon melt M in step S1 to the start of shoulder formation of the silicon single crystal C, for example, during the predetermined period from the stabilization step in step S4 to the shoulder formation step in step S6. The predetermined period is preferably during the stabilization step, which is a temperature stabilization step. From the viewpoint of reliability, the time is preferably set to a range of 10 to 180 minutes. Furthermore, it is desirable to start temperature measurement using this radiation thermometer 16 after the seed crystal has been immersed in the melt, since immersing the seed crystal in the melt during temperature measurement will affect temperature variations. Furthermore, it is desirable to start temperature measurement after a magnetic field has been applied, since accurate temperature measurement is difficult in the absence of a magnetic field due to the vibration of the melt surface. Then, the oxygen concentration is estimated based on the obtained temperature variation (step S7 in FIG. 2). To explain the method for estimating the oxygen concentration, first, the silicon single crystal pulling conditions (pulling conditions in this embodiment) are specified, the silicon single crystal C is pulled, and the oxygen concentration of the produced silicon single crystal C is obtained in advance.

[0035] Furthermore, when pulling up the silicon single crystal C, the temperature variation of the silicon melt M during a predetermined period from the start of the step of forming the silicon melt M to the step of forming the shoulder of the silicon single crystal C is obtained in advance. Then, the correlation between the oxygen concentration of the silicon single crystal C and the temperature variation of the silicon melt M under those pulling conditions is recorded. To keep this record, it is specifically desirable to manufacture at least five silicon single crystals C. An example of the correlation between the oxygen concentration of silicon single crystal C thus obtained and the temperature variation of silicon melt M under the pulling conditions is shown in the graph of Fig. 3. In Fig. 3, the vertical axis represents the oxygen concentration at the same straight body length, and the horizontal axis represents the temperature variation (standard deviation) of the surface temperature of silicon melt M.

[0036] The oxygen concentration is estimated in step S7 of Fig. 2 by applying the obtained temperature variation (standard deviation) of the surface temperature of the silicon melt M to the correlation between the oxygen concentration of the silicon single crystal C and the temperature variation of the silicon melt M under the pulling conditions, for example, as shown in Fig. 3. At this time, if the estimated oxygen concentration differs from the desired oxygen concentration, the oxygen concentration needs to be adjusted.

[0037] 8, if oxygen concentration adjustment is not required, the temperature measurement by the radiation thermometer 16 is terminated. On the other hand, if oxygen concentration adjustment is required, the pulling conditions that affect the oxygen concentration of the silicon single crystal C are adjusted, and then the temperature measurement by the radiation thermometer 16 is terminated (step S9 in FIG. 2).

[0038] Specifically, crystal growth conditions such as heater heating temperature, crucible rotation speed, and furnace pressure are adjusted. For example, when growing silicon single crystal C, the temperature variation of the measured surface temperature of silicon melt M is expected to vary depending on changes in the thermal environment. In this case, if data is previously obtained from measurements of the oxygen concentration of silicon single crystal C previously pulled using the same apparatus, showing that the oxygen concentration of silicon single crystal C tends to decrease as the temperature variation increases, as shown in the graph of Figure 3, for example, at least one of the following pulling conditions is adjusted: setting the heater temperature higher, increasing the crucible rotation speed, or decreasing the furnace pressure. Conversely, if data is previously obtained showing that the oxygen concentration of silicon single crystal C tends to increase as the temperature variation increases, at least one of the following pulling conditions is adjusted: setting the heater temperature lower, decreasing the crucible rotation speed, or increasing the furnace pressure.

[0039] The pulling conditions are adjusted as necessary, and once the shoulder portion C1 is formed, the controller 11 controls the lifting drive control unit 15 using the lifting drive control unit 15a to keep the pulling speed constant at a predetermined speed, and proceeds to the process of forming the straight body portion C2, which will become the product part (step S10 in Figure 2).

[0040] When the body portion C2 is formed to a predetermined length, the process proceeds to the final tail portion process (step S11 in FIG. 2). In this tail portion process, the contact area between the bottom end of the crystal and the silicon melt M gradually decreases, and the single crystal C and the silicon melt M are separated, thereby producing a silicon single crystal C.

[0041] As described above, according to the embodiment of the present invention, the surface temperature of the silicon melt M is measured during a predetermined period of time after the formation of the silicon melt M and before the formation of the straight body portion, and the temperature variation is determined from the temperature data related to the measured surface temperature of the silicon melt M. Then, the oxygen concentration of the silicon single crystal C to be pulled is estimated based on the pre-recorded correlation between the oxygen concentration of silicon single crystal C pulled under the same pulling conditions and the temperature variation of the silicon melt M. Then, if it is determined that adjustment of the pulling conditions is necessary based on the estimated oxygen concentration or temperature variation, pulling conditions that affect the oxygen concentration, such as the heater temperature, crucible rotation speed, and furnace pressure, can be adjusted to produce a silicon single crystal C having a target oxygen concentration. As described above, according to the method for producing a silicon single crystal of this embodiment, by using the correlation between the oxygen concentration of the silicon single crystal C and the variation in the silicon melt M under the pulling conditions, which is recorded in advance, the oxygen concentration of the silicon single crystal C to be pulled can be estimated with high accuracy before the straight body portion of the silicon single crystal is grown, and the pulling conditions for the straight body portion can be adjusted based on the estimation results.

[0042] In the above embodiment, the standard deviation is used as an index of temperature variation, but the variance over a predetermined time may also be used as an index of temperature variation. Furthermore, the peak intensity of the Fourier transform spectrum is, in principle, related to the magnitude of the variation in the surface temperature of the silicon melt M, and is correlated with the variance and standard deviation. Therefore, the peak intensity of the period accompanying the rotation of the crucible 3, and the power spectrum that can be calculated therefrom, may also be used as an index of temperature variation.

[0043] The crucible 3 is placed inside a graphite crucible 2 that is divided into multiple sections. The thermal environment inside an actual single crystal pulling apparatus is not axially symmetrical; the crucible 3 rotates while a temperature gradient exists within the silicon melt surface, and the divided sections, which are susceptible to thermal changes, alternate between low-temperature and high-temperature sections, resulting in periodic temperature fluctuations. Furthermore, the magnitude of this temperature fluctuation, i.e., the magnitude of the variation in the surface temperature of the silicon melt M, correlates with the thickness and other factors that indicate the degree of wear of the graphite crucible 2. In other words, the magnitude of the temperature variation is an indicator of the degree of deviation in the thermal environment, the degree of wear of the graphite crucible 2, the degree of deformation of the crucible 3, etc. Therefore, the step of estimating the oxygen concentration of the silicon single crystal based on the temperature variation may include a step of determining the temperature variation over the predetermined period from temperature data relating to the measured surface temperature of the silicon melt M, and then managing the graphite crucible 2 by linking the value of the data relating to the temperature variation with deterioration information of the graphite crucible 2. This allows the graphite crucible 2 to be replaced at an appropriate timing. [Example]

[0044] The method for producing a silicon single crystal according to the present invention will be further described based on examples. In Examples 1 and 2 and Comparative Example 1, silicon single crystals were produced by pulling them using the well-known HMCZ method according to the flow chart shown in Figure 2. The various production conditions were, for example, a pulling rate of 0.3 to 2.0 mm / min, a body length of 500 to 2000 mm, a magnetic field strength of 2000 to 4000 Gauss, a crucible rotation speed of 0.1 to 15.0 rpm, and an Oi range of 3E17 to 1.7E18. Ten crystals were pulled for each of Examples 1 and 2 and Comparative Example 1 under the same conditions except for the change in the crucible rotation speed, and the average values ​​were used as the values ​​for Examples 1, 2, and Comparative Example 1. 3 to 5, both the vertical and horizontal axes are normalized values.

[0045] (Experiment 1) In Experiment 1, using the single crystal pulling apparatus shown in Figure 1, the temperature of the silicon melt surface was measured at sampling intervals of 5 seconds one hour before the start of neck formation in the stabilization process (temperature stabilization process), and the standard deviation was calculated and used as an index. In Example 2, the standard deviation in the graph of Figure 3 is about 1.0, and pulling was performed without adjusting the pulling conditions. Here, a standard deviation of 1.0 means that when the desired oxygen concentration is set to 1.0, the standard deviation of the temperature of the silicon melt surface that roughly corresponds to this is set to 1.0. Therefore, when the standard deviation is 1.0, it is expected that a silicon single crystal with the desired oxygen concentration can be pulled without changing the pulling conditions for the straight body portion. In Comparative Example 1, the standard deviation was about 0.7 in the graph of Fig. 3, and pulling was performed without adjusting the pulling conditions. As shown in Fig. 3, when the standard deviation is 0.7, it is expected that the oxygen concentration will be about 1.08 higher than the desired concentration.

[0046] The results of Example 2 and Comparative Example 1 are shown in the graph of Fig. 4. The horizontal axis of the graph in Fig. 4 represents the straight body length, and the vertical axis represents the oxygen concentration. The results of Example 2 showed that an oxygen concentration close to the desired one was obtained for the length of the straight body portion, but Comparative Example 1 produced a single crystal with an oxygen concentration higher than the desired one. Therefore, in Example 1, a silicon single crystal was produced by the following method. The standard deviation was approximately 0.7 in the graph of FIG. 3, which could result in a high-oxygen crystal, as in Comparative Example 1. Therefore, the amount of oxygen change when the standard deviation in FIG. 3 changed from 1.0 to 0.7 was calculated, and the estimated amount of oxygen change was corrected. Specifically, the crucible rotation speed was changed as shown in the graph of FIG. 5. FIG. 5 is a graph showing the change in crucible rotation speed in Experiment 1, with the vertical axis representing the crucible rotation speed and the horizontal axis representing the body length. That is, in Example 1, the pulling conditions were adjusted by lowering the crucible rotation speed from that at the start of pulling to reduce the oxygen concentration. More specifically, in Example 1, after the shoulder was formed and pulling of the body began, the crucible rotation speed was gradually reduced from 2.0 rpm to 1.7 rpm.

[0047] The results of Example 1 are shown in the graph of Figure 4. As shown in Figure 4, the oxygen concentration was within the target value ([Oi] = 8E-17 (atoms / cm 3 )) from the temperature variation index, the oxygen concentration ([Oi] = 8.7E-17(atoms / cm 3 )) and adjust the pulling conditions to obtain the desired oxygen concentration ([Oi]=8E-17(atoms / cm 3 )) crystals were obtained. 6 shows the probability that the pulled silicon single crystal contains blocks that do not meet the oxygen concentration specifications. As shown in the graph of FIG. 6, the silicon single crystal pulled in Example 1 was able to reduce the probability of containing blocks that do not meet the specifications by about 40% compared to Comparative Example 1. [Explanation of symbols]

[0048] 1. Single crystal pulling device 3. Quartz glass crucible 4 Side heater 5 Bottom heater 6 wire 7 Radiation Shield 8a Electromagnetic coil control unit 14 Rotation drive unit 14a Rotation drive control section C Silicon single crystal M Silicon melt C1 Shoulder C2 Straight body part CN neck

Claims

1. A method for producing a silicon single crystal by pulling a silicon single crystal by the Czochralski method, A step of melting a silicon raw material to form a silicon melt in a quartz glass crucible; measuring the surface temperature of the silicon melt in the quartz glass crucible during a predetermined period from the formation of the silicon melt to the start of shoulder formation of the silicon single crystal; a step of determining a temperature variation during the predetermined period from temperature data relating to the measured surface temperature of the silicon melt, and estimating an oxygen concentration of the silicon single crystal based on the temperature variation; determining pulling conditions for the body portion of the silicon single crystal based on the temperature variation; 1. A method for producing a silicon single crystal, comprising:

2. In the step of estimating the oxygen concentration of the silicon single crystal based on the temperature variation, 2. The method for producing a silicon single crystal according to claim 1, wherein the oxygen concentration of the silicon single crystal is estimated based on the temperature variation over the predetermined period and a prerecorded correlation between the oxygen concentration of the silicon single crystal and the temperature variation of the silicon melt.

3. 2. The method for producing a silicon single crystal according to claim 1, wherein the predetermined period for determining the temperature variation is during a stabilization process, during which the silicon seed crystal is immersed in the silicon melt and a magnetic field is applied to the silicon melt.

4. 2. The method for producing a silicon single crystal according to claim 1, wherein the data relating to the temperature variation is a variance, a standard deviation, or a peak intensity of a Fourier transform spectrum over a predetermined time period.

5. The sampling interval Ts (s) for measuring the temperature data relating to the surface temperature of the silicon melt is The method for producing a silicon single crystal according to claim 1, characterized in that the crucible rotation speed is determined so that Ts≦30 / N·CR holds, where CR (rpm) is the number of divisions of the graphite crucible that holds the quartz glass crucible and N is the number of divisions of the graphite crucible that holds the quartz glass crucible.

6. The quartz glass crucible is placed in a graphite crucible divided into multiple parts, a step of determining a temperature variation for the predetermined period from temperature data relating to the measured surface temperature of the silicon melt, and estimating an oxygen concentration of the silicon single crystal based on the temperature variation, After determining the temperature variation for the predetermined period from the temperature data relating to the measured surface temperature of the silicon melt, 2. The method for producing a silicon single crystal according to claim 1, further comprising a step of managing the graphite crucible by linking the data value relating to the temperature variation with deterioration information of the graphite crucible.

7. The predetermined period for obtaining the temperature variation is during a stabilization process, 2. A method for producing a silicon single crystal according to claim 1, characterized in that measurement of the surface temperature of the silicon melt in the quartz glass crucible is started during the stabilization process, and the process of estimating the oxygen concentration and the process of determining the pulling conditions are performed, and then measurement of the surface temperature of the silicon melt in the quartz glass crucible is finished during the stabilization process.

Images