JPH09175890A - Liquid temperature control of furnace for pulling up single crystal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for controlling the liquid temperature by which the melt liquid temperature melted in a crucible can be controlled with good responsiveness in a furnace for pulling up a single crystal. SOLUTION: This method for controlling the liquid temperature is used in a furnace 1 for pulling up a single crystal equipped with a heater 3, a thermometer 5 for sensing the temperature of a melt surface, an arithmetic unit for determining a changed quantity of an applied electric power from the sensed temperature of the melt surface and the target value of the melt temperature and a device for controlling the electric power applied to the heater. The changed quantity of the applied electric power is N times that determined with the arithmetic unit (with the proviso that N is 2-5). The changed quantity of the electric power applied to the heater is made to coincide with that determined with the arithmetic unit after changing the temperature of the melt surface by a temperature corresponding to a constant ratio of a temperature difference between the sensed temperature of the melt surface and the target value of the melt temperature.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling a liquid temperature of a single crystal pulling furnace by the Czochralski method (hereinafter simply referred to as "CZ method"), and more particularly, to a method of melting a liquid in a crucible in the pulling furnace. The present invention relates to a method for controlling the liquid temperature of a melt to be controlled with good responsiveness.

[0002]

2. Description of the Related Art There are various methods for producing a single crystal. Among them, the CZ method is widely used in a method capable of industrial mass production for pulling and growing a silicon single crystal.

In this method, a single crystal is pulled and grown from a liquid surface of a melt of a crystal raw material melted in a crucible housed in a pulling furnace. Control becomes important.

For example, when a seed crystal in the process of pulling a single crystal is adapted to a melt, when it is pulled up with a small diameter in a necking process, or when the diameter is gradually increased, the shape of a shoulder (shoulder) is formed. During shoulder construction and pulling of the straight body, the responsiveness of the liquid temperature control of the melt affects the productivity and quality of the pulled single crystal.

As described above, it is important to control the liquid temperature of the melt with good responsiveness in the process of pulling the single crystal. As will be described later, the melt of the crystal raw material in the crucible in the pulling furnace is usually used. The response to the change in liquid temperature was relatively slow.
For this reason, various proposals have hitherto been made to shorten the response time for changing the liquid temperature.

Japanese Patent Application Laid-Open No. 62-123091 proposes a general method for controlling the liquid surface temperature from the relationship between the temperature of the melt and the temperature of a heater for heating the same. That is, first, the target value of the heater temperature corresponding to the target value of the melt temperature is obtained from the data showing the relationship between the melt temperature and the heater temperature, and when the heater temperature is controlled to the target value, the actual value of the melt temperature is obtained. If there is a deviation in the target value, the heater temperature is changed again according to the deviation to control the melt temperature. However, in the proposed method, although the heater temperature can be controlled with good responsiveness, the response of the change in the liquid temperature of the melt in the crucible is slow, so that a large effect cannot be achieved as the control of the liquid surface temperature.

On the other hand, Japanese Patent Application Laid-Open No. 4-325488 discloses that a liquid surface temperature is measured from a ratio of radiant energies of two wavelengths of infrared rays radiated from a molten liquid surface, and a heater is heated in accordance with a deviation from a set temperature. A method of controlling the liquid surface temperature by adjusting the power of the liquid has been proposed. In order to further shorten the response time, a transfer function between the liquid surface temperature and the electric power supplied to the heater is obtained, and the dead time included in the transfer function is determined. A method for forming a PID control system for compensating the above is disclosed. However, the disclosed method has a problem that setting conditions according to operating conditions is complicated, and it is difficult to apply the method to actual operations.

[0008]

Generally, when the temperature of a melt is controlled by changing the temperature of a heater for heating the melt in a crucible or the amount of electric power supplied to the heater, the controllability is high. It is greatly affected by operating conditions such as the charge amount of the crystal raw material in the crucible, the storage position of the crucible, or the rotation speed of the crucible. Further, even if the above-mentioned operating conditions are constant, it is affected by the deterioration with time of the heat insulating material and the heater provided in the pulling furnace. Further, in addition to such an effect, when controlling the temperature of the melt in the crucible housed in the pulling furnace, since the control is performed through the crucible having a multilayer structure, the response speed of the temperature change of the melt is reduced. Become slow. For this reason, in the conventional method including the above-described method of controlling the melt temperature, it was difficult to control the liquid temperature of the melt with good responsiveness in the process of pulling the single crystal.

SUMMARY OF THE INVENTION It is an object of the present invention to overcome the above-mentioned problems and to establish a method for controlling the temperature of a melt with good responsiveness in a single crystal pulling furnace.

[0010]

Means for Solving the Problems The present inventor has provided a thermometer (for example, a radiation thermometer) for measuring the temperature of the melt surface in a non-contact manner from a window provided on the upper part of the pulling furnace. As a premise, as a result of intensive studies on a method of controlling the temperature of the melt, the following knowledge was obtained.

Normally, the power supplied to the heater for heating the melt is adjusted to control the temperature of the melt in the crucible. However, there is a certain relationship between the temperature of the melt in the thermal equilibrium state and the supplied power. Can be seen.

FIG. 1 is a diagram showing the relationship between the melt temperature and the power supplied to the heater in a thermal equilibrium state according to operating conditions and the like. A certain relationship exists between the melt temperature and the supplied power. It can be seen that the melt temperature can be controlled by adjusting the input power of the heater based on the relationship. However, the above relationship depends on operating conditions, hoisting furnace conditions, and the like (hereinafter, simply referred to as "pulling conditions"), and relationship I (shown by a solid line) and relationship II (shown by a dotted line) in FIG. ) Indicates a difference in pulling conditions. For example, the relationship II in the figure indicates the influence of the pulling furnace conditions such as the deterioration of the heat insulating material and the heater over time, even if the operating conditions are the same as the relationship I.

After the inside of the pulling furnace is in a state of thermal equilibrium and the temperature of the melt in the crucible is stabilized, the electric power supplied to the heater for heating the melt is changed stepwise based on the relation I in FIG. When the temperature is changed, the temperature of the melt in the crucible changes in a single exponential function, showing a so-called first-order lag response characteristic. Further, even if the electric power supplied to the heater is changed stepwise a plurality of times, the change in the melt temperature becomes a combination of the first-order lag response characteristics.

On the premise of the above phenomenon, the response time for a change in the liquid temperature can be shortened. That is, the relation I in FIG.
From the measured temperature (T _o ) of the melt surface and the target value (T _i ) of the melt temperature, the powers (P ₀ and P _i ) to be applied to the corresponding heaters can be uniquely obtained based on can further P ₀ and P change amount of _i from input power (delta
P) can be determined (ie, ΔP = P _i −
The P _0). Next, the determined change amount of the applied power (Δ
First, the power to be supplied to the heater is determined based on P). The initial input power is set to power (P ₀ + N · ΔP) corresponding to several times the change amount of the input power to increase the responsiveness of the liquid temperature change. As the temperature of the melt changes, the temperature of the surface of the melt becomes the temperature difference (ΔT) between the actually measured temperature (T _o ) and the target temperature (T _i ).
= T _i -T _o ), and after confirming that the temperature has changed by a certain ratio, the electric power supplied to the heater is set to an appropriate electric power (P ₀ + ΔP). By repeating such an operation, the temperature of the melt in the crucible can be changed to the target temperature with high responsiveness.

In order to show the state of change of the melt temperature with respect to the electric power supplied to the heater, the expression of "first-order lag response" or "high-order lag response" is used. Refers to a single exponential response characteristic as described above, and a high-order lag response refers to other high-order exponential response characteristics.

The present invention has been completed based on the above findings, and has as its gist a liquid temperature control method for a single crystal pulling furnace described below (see FIG. 2 described later).

That is, a heater 3 for heating the crystal melt in the crucible, a thermometer 5 for detecting the temperature of the melt surface,
An arithmetic unit that calculates an amount of power to be supplied to the heater from a temperature difference between the detected temperature of the melt surface and a target value of the melt temperature and determines a change amount of the supplied power; and a heater based on the changed amount of the supplied power. A liquid temperature control method used in the single crystal pulling furnace 1 comprising a device for controlling the power supplied to the heater, wherein a change amount of the power supplied to the heater is determined by the input device determined by the arithmetic unit. N times the change amount (however, N = 2 to 5
After the temperature of the melt surface changes by a temperature corresponding to a certain ratio of a temperature difference between the detected temperature of the melt surface and a target value of the melt temperature, and then the power supplied to the heater is changed. A liquid temperature control method for a single crystal pulling furnace, wherein the amount is made to match a change amount of the input power determined by the arithmetic unit.

[0018]

FIG. 2 is a diagram illustrating a schematic configuration of a single crystal pulling furnace for carrying out the liquid temperature control method of the present invention. As shown in FIG. 1, a heater 3 for heating a crystal melt 4 in a crucible 2 in accordance with input electric power is provided in a target pulling furnace 1.
And a thermometer 5 (for example, a radiation thermometer) for measuring the temperature of the melt surface in a non-contact manner through a window provided at the upper part of the pulling furnace.
An arithmetic unit for determining power to be supplied to the heater from a temperature difference between a detected temperature of the melt surface and a target value of the melt temperature to determine a change amount (ΔP) of the supplied power; A device for controlling the electric power supplied to the heater based on the amount.

Normally, when the electric power supplied to the heater for heating the melt in the single crystal pulling furnace is changed stepwise to change the liquid temperature, a crucible having a multilayer structure composed of a graphite crucible and a quartz crucible is formed. Since the temperature of the melt is controlled via the above, it is assumed that the change in the melt temperature in the crucible exhibits a high-order delay response characteristic. However, recently, both graphite crucibles and quartz crucibles housed in the furnace have been manufactured to be thin and airtight, so that the response characteristics of the melt temperature have been improved, and the melt itself has been improved. The convection heat transfer revealed that the response time of the temperature change was significantly shortened.

FIG. 4 to be described later is a diagram showing an example of the result of actually measuring the response state of the melt temperature when the input power is changed in a stepwise manner in the pulling furnace of the actual machine. From this, it can be confirmed that the change in the melt temperature in the crucible in the pulling furnace actually used for the production of a single crystal does not need to consider the response of the higher-order lag, and the response is almost the first-order lag.

Therefore, in order to examine in detail the relationship between the input power and the response of the melt temperature in the pulling furnace, modeling was performed under the condition that the response of the melt temperature was a first-order lag response, and the heat transfer was performed. Created a simulator. Using this simulator, the relationship between the input power and the response of the melt temperature was examined variously, and the following points were focused on.

When the input power is changed in a stepwise manner by the change amount (ΔP) of the input power determined from the actually measured value of the melt temperature and the target value, the melt temperature is within a certain temperature difference from the target temperature. It takes a long time to reach and stabilize.

On the other hand, the step-like input power is set as a change amount larger than the change amount (ΔP) of the input power, and thereafter the input power is made to coincide with the change amount (ΔP) of the input power at an appropriate timing. If this is the case, the response time for the melt temperature to reach the target temperature can be greatly reduced. In this case, it is difficult to determine the timing of switching the input power irrespective of operating conditions and hoisting conditions such as hoisting furnace conditions. This timing is determined by the temperature between the actually measured temperature of the melt and the target temperature. It is solved by determining from the difference ratio.

The present inventor theoretically supports the above-mentioned point of view on the assumption that the response of the melt temperature in the pulling furnace actually used for the production of a single crystal is almost a first-order response. And

The response [h (t)] when the input power is changed stepwise by the change amount (ΔP) of the input power is expressed by the following equation.

H (t) = 1−e ^{−t / T} where t: elapsed time T: time constant Here, the response [f (t) when the input power is changed to N times ΔP )] Is f (t) = N × (1−e ^{−t / T} )... Next, the response [g] when the input power is changed to {− (N−1)} times ΔP is applied. (T)] is represented by the following equation.

G (t) = − (N−1) × (1-e− ^{t / T} ) When a time when the response [f (t)] expressed by the equation matches the target temperature is t _0. , F (t ₀ ) = N × (1−e ^{−to / T} ) =
1, and the following relationship is obtained.

[0028] ^{e -t0 / T = 1-1 / N} ··· On the other hand, the synthesis function in the case of adding the response of formula in t ₀ hours after the addition of [f (t)] [g (t)] The derivative of is
The following equation is obtained.

F ′ (t + t ₀ ) + g ′ (t) = N / T × e− ^{(t + t0) / T−} (N−1) / T × e− ^{t / T} (3) By substituting, this value becomes 0 (zero).

That is, after the elapse of the time t ₀ , the melt temperature is maintained at a constant value. Therefore, after adding an electric power corresponding to N times the change amount (ΔP) of the input power to the power input to the heater, In this case, by adding {-(N-1)} times the amount of change (ΔP) to the electric power supplied to the heater, the response time of the melt temperature can be greatly reduced. Here, switching the amount to be added to the electric power supplied to the heater from the input power N times ΔP to the input power {− (N−1)} times ΔP means that N times ΔP from the following equation. This means switching from the electric energy to the electric energy of ΔP.

[ΔP · N + ΔP · {− (N−1)}] = ΔP A specific configuration of the arithmetic unit that determines the change amount of the input power in the present invention will be described with reference to FIG.

FIG. 3 is a flowchart showing a calculation configuration for determining the change amount of the input power in the method of the present invention. The configuration in the order of steps is as follows.

(Step 1) In response to a command to change the melt temperature, the measured temperature (T _o ) of the melt surface detected by the thermometer is input.

(Step 2) Target value of melt temperature (T _i )
And a temperature difference (ΔT) between the actual measured temperature of the melt (T _o ) and the target value of the melt temperature (T _i ) is obtained.

(Step 3) Measured temperature (T _o ) of the melt
From the target value of the melt temperature (T _i ) and the temperature difference (ΔT),
Electric power to be supplied to the heater, that is, the measured temperature of the melt (T
input power for maintaining _o) (P ₀₎ and melt the target temperature (T _i) to order the input power (P _i) is calculated, further amount of change from the difference between the input power input power ([Delta] P =
P _i -P ₀ ) is determined.

(Step 4) Based on the determined input power change amount (ΔP), the power input to the heater is first reduced by an amount (N · Δ) corresponding to N times the input power change amount.
P) The input power setting [I] is performed so as to increase (set power: P ₀ + N · ΔP).

(Step 5) Next, the temperature of the melt surface is determined by the temperature difference (Δ) between the measured temperature (T _o ) and the target temperature (T _i ).
The timing when the applied power is switched when the temperature changes by a temperature corresponding to a certain ratio of T) is set.

(Step 6) After confirming the above switching timing, the electric power supplied to the heater is changed to an appropriate electric power (set electric power: P ₀ + ΔP) to set the input electric power [II], and the crucible is set. The melt temperature in the inside can be changed to the target temperature with good responsiveness.

In the present invention, when N is increased, the control accuracy is deteriorated and overshoot or the like is expected to occur. Therefore, N is set in the range of 2-5. More preferably, N is desirably in the range of 2-3.

The input power is switched at a point in time when the temperature of the melt surface changes by a temperature corresponding to a certain ratio of the temperature difference between the actually measured temperature and the target temperature, and this ratio is in the range of 50 to 70%. It is desirable to do. This ratio is examined based on the simulation result using the heat transfer simulator.

According to the simulation results on the condition that the time for the melt temperature to reach the target temperature (the time for the difference from the target temperature to be within 2 ° C.) can be shortened and the liquid temperature does not exceed the target temperature. When the power supplied to the heater is set to be twice as large as the change in the input power (when N is set to 2), the above ratio is set to 60 to 70%, and the change in the input power is set to 3 to 5 times. (When N is 3 to 5)
It became clear that it was desirable to set it to 50 to 60%.

As a device for controlling the electric power supplied to the heater in the pulling furnace of the present invention, a control device using a thyristor which is generally used may be used.

As described above, when the operating conditions such as the charge amount of the crystal raw material, the storage position of the crucible, or the rotation speed of the crucible are changed, or the operating conditions are the same, the heat insulating material and the heater deteriorate with time. The method of the present invention can cope with a case where the pulling furnace condition changes such as deterioration of the hot zone characteristic. That is, in the method of the present invention, the relationship between the temperature of the melt in the thermal equilibrium state and the power supplied to the heater may be corrected according to the pulling conditions of the single crystal. In particular,
The relationship between the melt temperature and the power supplied to the heater when the temperature of the melt is stable (for example, the amount of change in the liquid temperature within 45 minutes is within ± 1.5 ° C) each time the lifting conditions are changed or periodically. Is calculated, and based on the relationship, the power (P ₀ and P _i ) to be supplied to the heater and the change amount (ΔP = P _i −P ₀ ) of the supplied power are determined in FIG. 3 (step 3). Modify the configuration. By making such a correction, the temperature of the melt in the crucible can always be controlled with good responsiveness.

[0044]

EXAMPLES The effects of the liquid temperature control method for a single crystal pulling furnace according to the present invention will be specifically described based on examples.

Comparative Example 120 kg of polycrystalline silicon was charged into a 22 ″ crucible housed in a pulling furnace and completely melted to bring the inside of the pulling furnace into a thermal equilibrium state. From liquid temperature (T ₀ ) of 1425 ° C to target temperature (T _i )
The input power was changed in steps so that the temperature became 1435 ° C., the liquid temperature of the melt was increased, and the responsiveness of the melt temperature was confirmed.

From the relationship between the melt temperature and the electric power supplied to the heater under the operating conditions of the embodiment, the electric power (P ₀ ) for maintaining T ₀ is 53.0 KW, and the temperature difference (Δ between T ₀ and T _i (Δ
The change amount (ΔP) of the input power determined from T) is 0.6 KW
Therefore, the step-like input power (P ₀ + ΔP) is 5
3.6KW.

FIG. 4 is a diagram showing the results of actual measurement of changes in the applied power and the melt temperature in the comparative example. The symbols of the applied power and the melt temperature in the figure are those defined above.

As a result of the actual measurement, the temperature of the melt is equal to the temperature difference (ΔT).
It takes a long time of about 110 minutes to reach 63%, and it takes about 270 for the difference from the target temperature to be within 2 ° C.
Needed a minute. The temperature of the melt was measured at a position 100 mm from the center of the crucible with a radiation thermometer, and a smoothing treatment was performed to remove the temperature fluctuation due to the influence of convection.

(Example of the Present Invention) Responsiveness of the melt temperature was confirmed under the same conditions as the comparative example.

FIG. 5 is a view showing the results of actual measurement of changes in the applied power and the melt temperature in the example of the present invention. Also in the example of the present invention, the change amount (ΔP) of the input power determined from the temperature difference (ΔT) between T ₀ and T _i is 0.6 KW. 54.2KW (P ₀ + 2 · ΔP) corresponding to double power,
When the melt temperature changes by a temperature corresponding to 60% of the temperature difference (ΔT), the power supplied to the heater is 53.6 KW (P ₀ +
ΔP). At this time, it takes 60 minutes to reach 63% of the temperature difference (ΔT), and it takes about 90 minutes for the difference from the target temperature to be within 2 ° C.
Needed a minute. As is clear from comparison with the comparative example, according to the example of the present invention, the response time can be reduced to about 1/3 of the comparative example, and the liquid temperature can be controlled with good responsiveness.

[0051]

According to the liquid temperature control method of the present invention, the temperature of the melt in the crucible housed in the single crystal pulling furnace can be accurately controlled with good responsiveness. In addition, since the liquid temperature can be controlled with excellent responsiveness even when the operating conditions and the hot zone characteristics change, the thermal history of the grown single crystal is stabilized, and the quality can be improved. .

[Brief description of the drawings]

FIG. 1 is a diagram showing a relationship between a melt temperature in a thermal equilibrium state obtained according to a pulling condition and a power supplied to a heater.

FIG. 2 is a diagram illustrating a schematic configuration of a single crystal pulling furnace for implementing the liquid temperature control method of the present invention.

FIG. 3 is a flowchart showing a calculation procedure for determining a change amount of input power in the method of the present invention.

FIG. 4 is a view showing the results of actual measurement of the response state of the melt temperature when the electric power supplied to the pulling furnace is changed stepwise as a comparative example.

FIG. 5 is a diagram showing the results of actual measurement of changes in applied power and melt temperature according to the method of the present invention.

[Explanation of symbols]

1 ... pulling furnace, 2 ... crucible, 3 ... heater, 4 ... melt,
5. Thermometer

Claims (1)

[Claims] 1. A heater for heating a crystal melt in a crucible, a thermometer for detecting a temperature of the melt surface, and a heater based on a temperature difference between a detected temperature of the melt surface and a target value of the melt temperature. Used in a single crystal pulling furnace comprising an arithmetic device for calculating the amount of power to be applied to the battery and determining the amount of change in the amount of power applied, and a device for controlling the amount of power applied to the heater based on the amount of change in the amount of power applied. Liquid temperature control method, wherein the amount of change in the power supplied to the heater is N times the amount of change in the supplied power determined by the arithmetic unit, and the temperature of the melt surface is the same as the detected temperature of the melt surface. After changing by a temperature corresponding to a certain ratio of the temperature difference from the target value of the melt temperature, the change amount of the power supplied to the heater is made to match the change amount of the supplied power determined by the arithmetic unit. A liquid temperature control method for a single crystal pulling furnace. However, N is in the range of 2 to 5.