JP2002137989A - Method of pull up single crystal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To grow a single crystal with the CZ method at a high speed without generating off specification by corresponding to vicinity of ceiling value of crystal deformation rate of the crystal. SOLUTION: The aimed temperature Tc of fused solution is set necessary to make the deforming rate coincide with the vicinity of the ceiling value, from the relation between the deformation rate of the single crystal 10 and the temperature of the fused solution of a specific point of a vicinity apart by fixed distance away from outermost peripheral part of the single crystal 10. While pulling up, the temperature of the fused solution is measured by a thermometer 13, and the aimed crystal growing speed is controlled so that the measured temperature coincides with the aimed temperature Tc.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a CZ method (Czochralski method) used for producing a single crystal such as silicon.
And a method for pulling a single crystal.

[0002]

2. Description of the Related Art The CZ method is known as a method for producing a silicon single crystal used for producing a semiconductor integrated circuit. In the production of a silicon single crystal by this method, after immersing a seed crystal in a melt of silicon formed in a crucible,
While rotating the crucible and the seed crystal in the same or opposite directions,
By pulling up the seed crystal, a silicon single crystal is grown below the seed crystal.

[0003] In such single crystal pulling, the growth rate of the single crystal is regarded as important in terms of productivity and defect distribution in a crystal plane. That is, productivity increases as the crystal growth rate increases. In addition, a region where an OSF ring is generated, which is one of the defective regions when a device is formed, moves toward the outer periphery of the crystal as the crystal growth speed increases. For these reasons, in a general operation, the target value of the crystal growth rate is often increased as much as possible.

Here, the crystal growth speed is the speed at which the crystal grows in the crystal axis direction, and has the same numerical value as the pulling speed as an operation amount viewed from the pulling device side. This crystal growth rate is usually controlled by the output of a heater that heats the crucible.

That is, the pulling speed is used for controlling the crystal diameter, and is controlled in real time in accordance with the ever-changing diameter variation. However, the pulling speed is the value of the crystal growth speed at the average pulling speed within a predetermined time. When viewed, there is a case where the target value of the crystal growth rate set in advance largely deviates.
In this case, the heater output is operated to correct the crystal growth rate to a predetermined target value. When the heater output is operated, the crystal diameter changes, but by controlling the pulling speed to suppress the change, the crystal growth rate is controlled to the target value without being affected by the diameter control. You.

However, when the crystal growth rate is increased as described above, the deformation of the crystal becomes remarkable, and as a result, the crystal cross section deviates greatly from a perfect circle, and the product diameter cannot be maintained in a part, or the quality is degraded. The crystal may separate from the liquid surface in some cases. For this reason, a technique for controlling the pulling speed based on the deformation rate represented by the following equation has been proposed in Japanese Patent Laid-Open No. Hei 6-263584. Deformation rate (%) = (maximum diameter−minimum diameter) / average diameter × 1
00

That is, according to the technique disclosed in Japanese Patent Application Laid-Open No. 6-263584, the crystal diameter is measured during the pulling to determine the deformation ratio, and the pulling is performed so that the deformation is maintained at 0.3 to 1.5%. Speed is controlled. More precisely, the target value of the crystal growth rate is corrected so that the calculated deformation rate is maintained at 0.3 to 1.5%, whereby the output of the heater and the pulling rate are changed to a predetermined value. The deformation rate is maintained at 0.3 to 1.5% while maintaining the crystal diameter.

[0008]

By the way, it is relatively easy to maintain the crystal deformation ratio in the range of 0.3 to 1.5% during pulling. However, it is difficult in an actual operation to maintain the deformation rate near the upper limit value (1.5% in this case) and perform stable pulling near the maximum growth rate corresponding to the upper limit value. is there.

The reason is that, when the deformation rate is calculated from the actually measured crystal diameter and the target value of the crystal growth rate is corrected so that it becomes the upper limit, the change in the target value of the crystal growth rate is reflected in the change in the deformation rate. By pulling length, for example, 25m
This is because a delay of about m occurs, and the response delay may cause an overshoot in which the deformation ratio exceeds the upper limit. When an overshoot occurs, the portion deviates from the product specifications and becomes a defective portion.

Therefore, in an actual operation, the target value of the deformation rate is set to a value (for example, about 1%) which is considerably smaller than the upper limit value so that overshoot does not occur.
As a result, the crystal growth speed becomes slow, and it becomes difficult to improve the productivity as expected and to exclude the region where the OSF ring occurs from the outer peripheral portion of the crystal.

It is an object of the present invention to provide a single crystal pulling method capable of stably maintaining a crystal growth rate corresponding to the vicinity of an upper limit of a crystal deformation rate without deviating from specifications.

[0012]

Means for Solving the Problems The heat source in the CZ furnace is a heater arranged around the crucible. The heater provides substantially uniform radiation from the surroundings to the inner crucible. At this time, the temperature of the silicon melt in the crucible rises on the outer peripheral side mainly by being heated by the heater, and generates thermal convection. In addition, forced convection is generated by rotation of the crucible. Because of the agitation by these convections, the temperature of the melt in the partial portion repeatedly fluctuates with time. However, when macroscopically and averaged over the time axis, the temperature of the melt surface becomes the highest on the side of the crucible side wall near the heater,
The temperature tends to be the lowest at the center of the crucible away from the heater.

When performing crystal growth in a CZ furnace, it is general to grow the crystal at the lowest temperature in the center. Here, the rate at which the crystal grows (solidifies) is substantially proportional to the difference between the amount of heat flowing from the solid-liquid interface to the crystal side and the amount of heat flowing from the melt to the solid-liquid interface. Therefore, when trying to increase the growth (solidification) rate of the crystal, the amount of heat removed from the solid-liquid interface to the crystal side is almost constant, so it is necessary to reduce the amount of heat flowing from the melt to the solid-liquid interface. For this reason,
By manipulating the output of the heater, the crystal growth rate is controlled.

The present inventors have conducted various studies to control and analyze deformation during crystal growth. As a result, it was confirmed that the temperature gradient in the radial direction of the melt surface near the crystal had a large effect on the deformation of the crystal. The mechanism of crystal deformation will be described based on the situation in the CZ furnace described above.

When the crystal growth rate is somewhat slow, the crystal grows in a shape close to a perfect circle. However, as the growth rate increases, that is, as the heater output decreases, the radial temperature gradient near the surface of the melt in the crucible decreases. Originally, the growth rate of a silicon crystal differs depending on the crystal plane direction, and the difference in the growth rate in the radial direction increases as the temperature gradient in the radial direction decreases. When this tendency continues for a long time, the cross-sectional shape of the crystal deviates from a perfect circle,
Observed as deformation.

That is, the cause of the crystal deformation is mainly the radial temperature gradient of the melt surface near the crystal, and as the radial temperature gradient becomes gentler, the supercooled crystal increases the deformation. Let it.

Here, when an attempt is made to measure the temperature gradient in the radial direction of the melt surface in the vicinity of the crystal, it is usually disclosed in
Although it is often considered necessary to measure the temperature between two points in the radial direction as described in JP-A-133187, the temperature at the crystal growth position is considered to be constant at 1412 ° C. under the current pulling conditions. A temperature gradient in the radial direction of the melt surface in the vicinity of the crystal can be calculated by measuring only the temperature of the point. Also, when measuring two points, the measurement position will be far from the crystal,
It is impossible to accurately reflect the temperature gradient in the radial direction of the melt surface near the crystal that actually affects the crystal deformation.

Based on this idea, the present inventors performed crystal growth while measuring the temperature near the melt surface near the crystal with a thermocouple, and determined the temperature near the melt surface when the crystal was deformed. It could be measured specifically. The contents of the experiment are as follows.

A thermocouple was placed at a position 15 mm outside the outermost periphery of the crystal, that is, near the liquid level which was 120 mm from the center when the crystal radius was 105 mm. After the crystal had a predetermined diameter, here 105 mm, the measurement was started from the point in time when the solidification rate was 30% at which the crystal growth rate was reduced. At this point, the crystal growth rate is stabilized at about 90% of the previously confirmed rate at which the deformation rate becomes 1.5%. Thereafter, the growth rate was gradually increased, and when the solidification rate reached 60%, the growth rate reached a growth rate at which the deformation rate became 1.5%, and thereafter, the growth rate was changed according to a program at a constant rate.
FIG. 1 shows changes in the crystal growth rate, deformation rate, and melt temperature at this time.

As can be seen from FIG. 1, the deformation rate reflects the crystal growth rate, but changes later than the change in the growth rate. This is because the influence on the crystal deformation rate after a change in the radial temperature gradient on the melt surface near the crystal appears only after the crystal has grown to some extent (25 mm in this case). For this reason, when the target value of the crystal growth rate is changed based on the crystal deformation rate, it is necessary to predict and control the time lag at which the deformation rate changes with delay.

On the other hand, the melt temperature reflects the crystal growth rate, and responds quickly to the change in the crystal growth rate. Therefore, a change in the deformation rate can be predicted in advance from a change in the melt temperature,
As a control target of the deformation rate, the melt temperature is excellent in response. Further, the melt temperature (critical temperature) corresponding to the upper limit of the deformation rate (here, 1.5%) can be obtained.

Based on the above results, a target value of the melt temperature measured by a thermocouple is set, and the target value of the crystal growth rate is maintained so that the target temperature is maintained at the target value. Was corrected. The target value of the crystal growth rate is originally determined by a table, and if the crystal growth rate is controlled according to the table, the melt temperature is determined uniquely by a function of the solidification rate and the target value, This cannot be controlled to be constant. Therefore, here, the target value of the crystal growth rate was corrected so that the melt temperature was maintained at the target value Tc. The target value Tc of the melt temperature is the upper limit value of the deformation rate (here, 1.
5%) near the melt temperature (critical temperature).

FIG. 2 shows changes in the deformation rate and the melt temperature at this time. By keeping the temperature near the melt surface at a specific point near the crystal near the critical temperature, the crystal deformation rate can be kept constant near the upper limit (here, 1.5%), and the deformation rate is within the allowable range. It can be seen that the crystal growth rate can be increased while maintaining the temperature within the range.

The single crystal pulling method of the present invention has been completed on the basis of this finding. In pulling a single crystal by the CZ method, the deformation rate of the single crystal during pulling and the melting in the radial direction near the crystal are increased. From the relationship with the liquid temperature gradient, a target value of the melt temperature gradient necessary to match the deformation rate to the target value is set, and the melt temperature gradient is measured during pulling of the single crystal. The target value of the crystal growth rate is corrected so that the measured melt temperature gradient is maintained at the target value.

It is preferable that the target value of the crystal deformation rate is close to the upper limit of the allowable range of the deformation rate because the crystal growth rate can be increased.

The melt temperature gradient is calculated from the measured temperature of the surface of the melt at a specific point near the crystal at a predetermined distance from the outermost periphery of the crystal, with the melt temperature at the crystal growth position as the crystal melting point. preferable.

That is, in the actual operation, it is not necessary to particularly consider the melt temperature gradient. From the relationship, a target value of the melt temperature required to match the deformation rate to the target value is set,
The melt temperature may be measured during the pulling of the single crystal, and the target crystal growth rate may be corrected so that the measured melt temperature is maintained at the target value.

The thermometer for measuring the melt temperature may be a thermocouple immersed in the melt, but a non-contact thermometer such as a two-color thermometer is more preferable. A thermocouple measures the temperature near the surface of the melt, while a two-color thermometer measures approximately the surface temperature of the melt. If the difference is corrected, a two-color thermometer or another non-contact thermometer can be used for controlling the growth rate using the crystal deformation rate as an index. Since the thermocouple is brought into direct contact with the melt, there are problems in terms of contamination and durability, but the non-contact type thermometer does not have these problems. When a two-color thermometer is used, it is effective to remove stray light due to irregular reflection from the melt surface, set an observation angle, and adjust for a change in crystal diameter, etc. for improving accuracy.

The target value of the melt temperature is, specifically,
A melt temperature corresponding to 80 to 95% of the upper limit of the deformation rate is preferred. The distance from the outermost periphery of the crystal to the measurement point is 110 to 115% of the crystal diameter (230 mm for an 8-inch diameter).
To 240 mm). Incidentally, the upper limit of the deformation rate is 1.0 to 1.5%.

[0030]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 3 is a configuration diagram of a pulling apparatus suitable for performing the single crystal pulling method of the present invention.

The pulling apparatus shown in FIG. 3 has a main chamber 1 and a pull chamber 2 as a furnace body of a CZ furnace. Crucible 3 in the center of main chamber 1
Are arranged, and the heater 4 is arranged on the outer peripheral side. The crucible 3 includes a quartz crucible and a graphite crucible outside the quartz crucible, and holds the silicon melt 10. The crucible 3 is placed on a support shaft 5 called a pedestal, and is driven vertically and circumferentially by the elevation and rotation of the support shaft 5.

The pull chamber 2 includes the main chamber 1
On top of the center. A pulling mechanism 6 is provided at the top of the pull chamber 2. At the center in the pull chamber 2, a wire 7 as a pulling shaft is hung from a pulling mechanism 6. A seed chuck 8 for holding a seed crystal is connected to a lower end of the wire 7. The pulling mechanism 6 pulls up while rotating the wire 7 at a predetermined speed.

In operation, a predetermined amount of the melt 10 is formed in the crucible 3 while maintaining a predetermined atmosphere in the furnace.
While the seed crystal held by the seed chuck 8 is immersed in the melt 10, the wire 7 is pulled up while rotating at a predetermined speed. Thus, a silicon single crystal 11 is grown below the seed crystal.

For controlling the diameter and the growth rate during the growth, a single crystal 11 is placed above the outer peripheral portion of the main chamber 1.
A camera 12 is installed to measure the diameter of the sample, and a non-contact thermometer 13 such as a two-color thermometer is installed to measure the surface temperature of the melt 10.

The camera 12 photographs a fusion ring formed on the outer periphery of the lower end of the single crystal 11 from a window provided on the outer periphery of the ceiling of the main chamber 1, and measures the diameter of the single crystal 11 from the ring diameter. The non-contact thermometer 13
From the window provided on the outer periphery of the ceiling of the main chamber 1,
The surface temperature of the melt 10, more specifically, the surface temperature at a specific point at a predetermined distance from the outer periphery of the single crystal 11 is measured in a non-contact manner. These measurement data are input to a crystal diameter and growth rate control system described later.

In the control system, the diameter measurement value input from the camera 12 is compared with a preset diameter target value, and the growth rate required to reduce the deviation to 0 is determined by PID control using the deviation. The diameter measurement value is controlled to a target diameter value by operating the pulling speed by the pulling mechanism 6 based on the output value of the growth rate.

The output value of the growth rate is also compared with a target value of the growth rate, and a PID control using the deviation outputs a set value of the heater temperature required to reduce the deviation to zero. The output of the heater 4 is operated to obtain a value. When the output of the heater 4 is increased, the crystal diameter becomes smaller, and the pulling speed is reduced to prevent this change in diameter. Conversely, when the output of the heater 4 is reduced, the crystal diameter increases, and the pulling speed is increased to prevent this change in diameter. Thus, the crystal growth rate is controlled to the target value while the crystal diameter is controlled to the target value.

Here, the target value of the crystal growth rate is corrected based on the melt surface temperature measured by the non-contact type thermometer 13. More specifically, the relationship between the deformation rate of the single crystal 11 and the surface temperature of the melt at the temperature measurement point is determined in advance, and the melt surface temperature corresponding to the vicinity of the upper limit of the deformation rate determined from this relationship is set to the target value. As the Tc, the target value of the crystal growth rate is corrected so that the measured value of the melt surface temperature by the non-contact type thermometer 13 is maintained at the target value Tc.

That is, the crystal growth rate is limited by the crystal deformation rate. However, when the crystal deformation rate is calculated and the target value of the crystal growth rate is corrected, a control delay occurs, and the upper limit of the crystal deformation rate is obtained. When the crystal growth rate corresponding to the vicinity is set to the target value, the specification is deviated due to the overshoot.
For this reason, it is necessary to lower the target value of the crystal growth speed to a speed corresponding to a value considerably smaller than the upper limit of the deformation rate.

However, when the target value of the crystal growth rate is corrected based on the melt surface temperature measured by the non-contact type thermometer 13, the target value of the crystal growth rate is adjusted to correspond to the vicinity of the upper limit of the crystal deformation rate. Even if it goes up to the level that does, the specification does not deviate due to overshoot.

This is because the deformation rate of the crystal corresponds to the radial temperature distribution of the melt in the vicinity of the crystal without any time delay (see FIG. 2). The measured value of the surface temperature of the melt by the non-contact type thermometer 13 is the surface temperature of the melt at a specific point at a certain distance from the outermost periphery of the single crystal 11, and the surface temperature of the melt at the outermost periphery of the single crystal 11 is constant. Since this is the crystal melting point (1412 ° C.), it corresponds to the radial temperature distribution of the melt near the crystal. Accordingly, the relationship between the deformation rate of the single crystal 11 and the melt surface temperature at a specific point is determined in advance, and the melting temperature slightly lower than the melt surface temperature (critical temperature) corresponding to the upper limit of the deformation rate determined from this relationship. If the target value of the crystal growth rate is corrected so that the liquid surface temperature is set to the target value Tc and the measured value of the melt surface temperature by the non-contact type thermometer 13 is maintained at the target value Tc, the deformation rate While maintaining a high crystal growth rate substantially corresponding to the upper limit of the above, the deformation rate is managed within an allowable range.

The result of growing a silicon single crystal by the CZ method using such a method will be described.

The growth conditions are as follows: quartz crucible diameter 560 mm, crystal diameter 8 inches, crystal rotation speed 12 rpm, crucible rotation speed 6 rpm, charge amount 110 kg, Ar flow rate 30 L, furnace pressure 25 Torr. A two-color thermometer was used as the non-contact thermometer, and the distance from the outermost periphery of the crystal to the temperature measurement point was 15 mm. The upper limit of the crystal deformation rate is 1.5%, and the corresponding melt surface temperature (critical temperature) is 1420 ° C. This critical temperature is different from the actual temperature because it is obtained by multiplying the output of the thermometer by a coefficient.

The target value Tc of the melt surface temperature is changed to a deformation ratio of 1.4.
%, The target value of the crystal growth rate was corrected so that the measured value of the melt surface temperature was maintained at the target value Tc. The process time was 16 hours, and the occurrence rate of the quality out part due to the overshoot of the deformation rate was 1%.

On the other hand, when correcting the target value of the crystal growth rate by calculating the deformation rate, the target value of the deformation rate must be increased from the vicinity of the upper limit in order to suppress the occurrence rate of the quality-out portion to about 1%. It had to be drastically reduced, which limited the crystal growth rate and increased the process time to 19 hours. On the other hand, when the target value of the deformation rate is increased to near the upper limit, the process time becomes 16 hours, but the occurrence rate of the quality out portion due to the overshoot of the deformation rate becomes 15
% Has been reached.

[0046]

As described above, the single crystal pulling method of the present invention can stably maintain the crystal growth rate corresponding to the vicinity of the upper limit of the crystal deformation rate. Therefore, the productivity can be improved without causing the specification deviation due to the overshoot of the deformation rate. OS
The generation region of the F-ring can be stably excluded from the outer periphery of the crystal.

[Brief description of the drawings]

FIG. 1 is a graph showing changes in the crystal deformation rate and the melt temperature when the crystal growth rate is increased.

FIG. 2 is a graph showing changes in the melt temperature and the deformation rate when the crystal growth rate is controlled so that the melt temperature matches a target value.

FIG. 3 is a configuration diagram of a single crystal pulling apparatus suitable for carrying out the single crystal pulling method of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Main chamber 2 Pull chamber 3 Crucible 4 Heater 5 Support shaft 6 Pulling mechanism 7 Wire 8 Seed chuck 10 Melt 11 Single crystal 12 Camera 13 Non-contact type thermometer

Claims (4)

[Claims] Claims: 1. In pulling a single crystal by a CZ method,
From the relationship between the deformation rate of the single crystal being pulled and the melt temperature gradient in the radial direction near the crystal, a target value of the melt temperature gradient required to make the deformation rate equal to the target value is set. In addition, it is characterized in that the melt temperature gradient is measured during the pulling of the single crystal, and the target value of the crystal growth rate is corrected so that the measured melt temperature gradient is maintained at the target value. Single crystal pulling method. 2. The melt temperature gradient is calculated from the surface temperature of the melt at or near the crystal at a predetermined distance from the outermost periphery of the crystal, with the melt temperature at the crystal growth position as the crystal melting point. 2. The method for pulling a single crystal according to 1. 3. In the pulling of a single crystal by the CZ method,
From the relationship between the deformation rate of the single crystal during pulling and the melt temperature at a specific point in the vicinity of the crystal at a predetermined distance from the outermost periphery of the crystal, the melt temperature required to match the deformation rate to its target value Is set, and the melt temperature is measured during the pulling of the single crystal, and the target value of the crystal growth rate is corrected so that the measured melt temperature is maintained at the target value. A single crystal pulling method. 4. The single crystal pulling method according to claim 1, wherein the target value of the deformation rate is near an upper limit of an allowable range of the deformation rate.