JPH03252390A - Condition setting method for analyzing heat fluid in furnace - Google Patents

Abstract

PURPOSE:To simply set an analyzing condition and to improve accuracy by carrying out simulation using gas inflow temperature, speed, etc., in a furnace as boundary conditions and an analyzed mesh shape, pressure, etc., in the furnace as initial values and obtaining optimum conditions. CONSTITUTION:Gas inflow temperature and speed in a furnace under high- pressure gas, temperature of external wall, etc., are used as boundary conditions. An analyzed mesh shape, pressure in the furnace, heater power, etc., are set as initial values and simulation is carried out to obtain optimum conditions. In the operation, pressure in the furnace is changed and heater monitor temperature is fixed by the simulation to obtain correlation between heater power and gas inlet flow velocity. Simulation is done by changing the gas inlet flow velocity corresponding to the heater power and pressure in the furnace.

Description

[Detailed Description of the Invention] [Field of Industrial Application] This bolt is designed to analyze the mesh shape, This article relates to a method for setting analytical conditions for in-furnace thermo-fluid analysis, which involves setting the pressure and other physical property values in the furnace, heater power, etc. as initial values and performing a simulation to find the optimal operating conditions.

[Conventional technology]

GaAs single crystals are attracting attention as substrates for high-speed ICs and optical devices, and are being actively researched.

The mainstream method for producing single crystals is the LEC method (liquid-filled Czochralski method), which is one of the main methods for producing bulk crystals of ■-■ group compound semiconductors and is easy to produce large-diameter, circular wafers. be.

Bulk crystal growth of the ■-■ group compound semiconductor is performed under a complex thermal environment at high temperature and high pressure due to the properties of the ■-V group. In crystal growth, in order to grow crystals with lower dislocations, the thermal environment must be moderated to reduce thermal stress, and in order to further increase productivity, the temperature gradient must be increased to strengthen directional solidification. I have a request. Single crystal pulling furnaces use heat insulators and multi-stage heaters, making control extremely complex and difficult.

FIG. 10 is a diagram showing an outline of the configuration of a single crystal pulling apparatus for the LEC method, in which 31 is a water-cooled reflector, 32 is graphite, 33 is a sub-upper heater, 34 is a main heater,
35 is a sub-lower heater, 36 is a crucible, 37 is a crystal part,
38 is a melt part, 39 is B, O, and 40 is a heat insulating felt.

The inside of the single crystal pulling furnace is placed under a high pressure (10-odd atmospheres) atmosphere of non-volatile Ar (argon) gas, for example, and the inside of the crucible is heated to a melting point of GaAs of 1 by heaters (33 to 35).
It is heated to 511° (K) or higher. Then, the heater power, pulling speed, rotational speed of the crucible 36 and the crystal 37, etc. are controlled while calculating the shape of the crystal being pulled and grown from various measured values. In this case, calculation of the crystal shape during pulling and growth is performed by detecting the crystal weight using a load cell and measuring the change in the crystal weight while correcting the buoyancy due to B20339 and the surface tension of the melt.

There are many conditions for crystal growth by the LEC method, but the most important one is the thermal environment inside the furnace. In order to achieve high-quality crystal growth and high productivity, it is essential to understand the thermal environment. However, as mentioned above, the inside of the furnace is filled with a high-temperature, high-pressure atmosphere and has a complicated internal structure, so it is not possible to directly observe the entire inside of the furnace during crystal growth. Therefore, computer simulation is an essential technique to achieve a desirable thermal environment, and many analyzes have been attempted so far.

For example, temperature distribution in the furnace, temperature distribution in the crystal, thermal convection in the melt,
Analysis of the solid-liquid interface shape and intracrystalline thermal stress analysis, which achieved the crystal/melt structure, have been carried out vigorously, and results have been achieved.

[Problem to be solved by the invention 2 However, the above-mentioned various numerical analyzes and computer simulations are analyzes of each element alone, and do not involve relationships among the elements. It was not directly related to the conditions. Therefore, it is difficult to quantitatively understand how to change the t setting to optimize the thermal environment inside the reactor based on each analysis result, and it is also difficult to correlate the simulation results with the actual machine. There's a problem.

In particular, in order to increase the length of single crystals, it is necessary to establish a technology to control the shape of the solid-liquid interface during growth using a multi-zone heater. Accurate understanding of heat transfer within the furnace is necessary. Furthermore, since crystal pulling experiments require a lot of time and effort, computer simulations are effective, but in order to improve the accuracy of these simulations, it is necessary to accurately understand the heat transfer inside the single crystal pulling furnace. It is essential.

In addition, in order to improve productivity, it is necessary to increase the size of the crucible in which a large amount of charging, ie, raw materials, is placed, and in order to meet the demand for large diameter wafers, it is also necessary to increase the size of the crucible. However, normally, the size of the entire furnace is fixed, and it is not possible to replace or transform the furnace due to various restrictions. Moreover, the heater that controls the temperature inside the furnace is
Because it is placed outside the crucible, if the crucible is made larger, the insulation part becomes thinner, which poses a safety problem for the furnace, and it becomes unclear whether crystal growth is possible. In particular, as mentioned above, the furnace is under a high-pressure (over 10 atmospheres) atmosphere of non-volatile gas, such as Ar (argon) gas, so the most important thing in modeling the heat flow inside the chamber is ,
It is a quantitative treatment with a heat balance, and L
In the EC method, handling of the unique high-pressure Ar flow is important.

A heat balance model in a furnace placed in a high-pressure atmosphere of non-volatile Ar gas will be briefly explained.

FIG. 11 is a diagram showing a heat balance model within the furnace.

The heat balance model shown in Fig. 11 is based on
Points ~■ are specified, and modeling is performed to calculate the heat balance at these points ■~ [phase]. In this model, the convection direction is gas 3 - gas 4 → upper part of the sub - upper part of the upper inside and outside of the insulation cylinder - between the outside of the insulation cylinder and the outer wall → lower part of the upper inside and outside of the insulation cylinder, and gas 1 - This indicates that the direction is from gas 2 to return to gas 3. There is heat transfer (-), thermal radiation (), and convection heat transfer () between each point 0 and 0. For example, point ■ at the bottom of the crucible (Tc,
), heat radiation (R) is generated between the main heater (TX) and the main heater (TX).
, there is heat transfer (C,) between the gas 2 (T, d) and the pedestal (T, d), and convection heat transfer (Cv) between the gas 2 (T, 2) and the gas 3 (T93), and the heat balance is maintained. It shows that there is. In this way, the amount of heat is transferred by the Ar flow, and there is a heat balance between 0 and 0 at multiple specified points (representative temperature positions) through heat transfer, heat radiation, and convection heat transfer, forming an overall heat balance model. Ru. In order to perform thermo-fluid analysis using such a heat balance model and to find the optimal operating conditions for a single crystal pulling furnace, a comparator simulation using the t1 finite element method, etc., which requires a huge amount of repeated processing, is performed.

Conventionally, for example, when performing thermo-fluid analysis in the crystal and melt parts, analysis was performed using heat transfer and convective heat transfer, and when performing thermo-fluid analysis inside the furnace, heat transfer and heat transfer were used. We are conducting an analysis using radiation. Moreover, a coupled analysis of these has not been performed, and sufficient accuracy has not been obtained. Furthermore, in conventional thermal fluid analysis using computer simulation, the temperature distribution of the heater is set as an initial value, but the temperature distribution of the heater changes depending on various conditions inside the furnace, so it is difficult to set it accurately. The accuracy of thermal fluid analysis is poor. In other words, it is not possible to quantitatively discuss the correspondence with the operating conditions of the actual machine.

In order to actually increase the size of the crucible, we first perform the analysis described above to create a furnace structure that is compatible with the increased size of the crucible, and then repeat crystal pulling experiments.
Because the accuracy was not sufficient, experiments had to be repeated many times, resulting in a large loss of time and money.

The present invention is intended to solve the above-mentioned problems, and aims to provide an analysis condition setting method for in-reactor thermal fluid analysis that can easily set analysis conditions in computer simulation and improve analysis accuracy. That is.

[A means of promise to solve problems]

To this end, the present invention uses the boundary conditions such as the gas inflow temperature, velocity, and temperature of the outer wall of the furnace heated by a heater in a high-pressure gas atmosphere, and the analysis mesh shape, pressure and other physical property values in the furnace, heater power, etc. This is an analysis condition setting method for in-furnace thermo-fluid analysis that sets initial values and performs simulation to determine optimal operating conditions.Heater power and gas inlet are used to maintain a constant heater monitor temperature by changing the pressure in the furnace through simulation experiments. The feature is that the simulation is performed by determining the relationship between the flow rate and the heater power set as an initial value, and by changing the gas inlet flow rate in accordance with the pressure in the furnace. It is characterized by determining the relationship between heater power and gas inlet flow rate from the difference in heater power when changing the pressure and the difference between gas inlet temperature and gas outlet temperature. In addition, we conduct thermal fluid analysis to determine the temperature distribution in the furnace, thermal analysis of the crystal/melt part from the temperature distribution in the furnace to determine the solid-liquid interface shape, and solid-liquid interface shape analysis to determine the shape of the solid-liquid interface and the crystal part. It is characterized by performing a simulation of a single crystal pulling furnace by separating it into a thermal stress analysis that determines the thermal stress within the crystal part from the temperature distribution.

[Effect]

In the analysis condition setting method for in-furnace thermal fluid analysis of the present invention, the relationship between the heater power and the gas inlet flow velocity is determined by changing the pressure in the furnace to keep the heater monitor temperature constant through a simulation experiment, and the heater is set as an initial value. Since the simulation is performed by changing the gas inlet flow velocity according to the power and pressure in the furnace, it is possible to solve the convection in the furnace from the gas inlet flow velocity, achieve convection and radiation, and solve the temperature distribution in the furnace from the heater power. It is possible to solve the temperature distribution with higher accuracy than the conventional method that gives the temperature distribution of the heater.

Moreover, by dividing the simulation into three parts, thermal fluid analysis, solid-liquid interface shape analysis, and thermal stress analysis, in order to find the optimal operating conditions for a single crystal pulling furnace, it is possible to make a convergence judgment in the middle, compared to the conventional method. It is possible to perform much more efficient analysis than . When designing a new furnace internal structure, it is possible to analyze the gas inlet flow velocity as a parameter and compare it with the actual furnace internal structure.

〔Example〕

Examples will be described below with reference to the drawings.

FIG. 1 is a schematic diagram illustrating one embodiment of an analysis condition setting method for in-furnace thermal fluid analysis according to the present invention.

First, consider the influence of the Ar flow on the heat balance within the chamber. The macroscopic heat balance in a vacuum can be considered as heat transfer by radiation from the outer surface of the hot zone to the surface of the glue flefter chamber. However, if the inside of the chamber is heated with a heater using Ar at high pressure, the first
As shown in the convection direction in Figure 1, the Ar gas receives buoyancy,
The heat flows upward and flows out from the upper part of the hot zone, exchanges heat at the upper part of the water-cooled reflector chamber, and then descends between the side wall of the chamber and the outer wall of the hot zone, causing heat loss. In other words, by placing the inside of the furnace in a high-pressure Ar atmosphere, loss due to convective heat transfer is added. Therefore, when the inside of the furnace is changed from a vacuum to a high-pressure Ar atmosphere, for example, in order to keep the temperature constant at the monitoring point where the main heater is located, it is necessary to increase the power of the main heater by the amount of loss due to convective heat transfer. In other words, the increase in power, the pressure inside the furnace, and convective heat transfer are closely related to maintain a thermal balance. For this reason, in computer simulation analysis, we give the gas inlet temperature and flow velocity, heater power, and outer wall temperature, solve the convection from the gas inlet flow velocity, and couple it with thermal radiation. Temperature distribution can be determined using the model.

As is well known, using the equation of convective heat transfer by a fluid, 16 atm Ar gas <p=6.12xlO-'g moving at a flow rate v (Cm/5ac) across a cross-sectional area 5 (cut)
/cIIl, C,=0. The amount of heat q (W) carried by 52 J/gK) is the temperature difference Δ between the gas inlet and outlet.
When T (K) occurs, it is expressed as q=ρC,vSΔT.

In the present invention, based on the above consideration, the boundary conditions are the flow rate and gas temperature at the gas inlet in the furnace, and the temperature at the outer wall, which are divided into meshes using the finite element method as initial values, and the mesh conditions (
The thermal fluid analysis inside the chamber is performed to determine the temperature distribution inside the furnace. to find the flow velocity V at the gas inlet, and set the boundary conditions and initial values for the thermal fluid analysis. Then, we conducted a mock experiment as shown in Figure 1.

In this simulation experiment, as will be described later, before mesh-dividing the inside of the furnace and conducting thermal fluid analysis using the finite element method, conditions such as the pressure inside the furnace and the heater monitor temperature are set accordingly. After setting these conditions, first create a vacuum inside the furnace and adjust the heater power so that the heater monitor temperature reaches the set value. Next, set the pressure in the furnace to the state in which you are going to perform thermal fluid analysis, adjust the heater power so that the heater monitor temperature reaches the set value, and check the gas inlet temperature and gas outlet temperature at that time. Measure. From these experimental data, the difference between the temperature at the gas inlet and the temperature at the gas outlet when the pressure inside the furnace is set to the same value is ΔT1. The difference Q"Qp+Qv from the power Q, 1 is determined, and the gas inlet flow velocity v (am/5ec) is determined from the above equation.

Furthermore, by changing the pressure inside the furnace and the heater monitor temperature,
Similarly, experimental data may be acquired and the flow velocity V corresponding to each set value may be determined.

When the present inventors investigated the effect of Ar flow on the heat balance in the chamber shown in Figs. 10 and 11, they found that under the condition that the main heater monitor temperature was 1285°C, the pressure was -14 psi and 233 psi. The following results were obtained.

In addition, the sub-upper heater and sub-lower heater power are 01 cross-sectional area 5 = 104crl, p = 6.12xlO-3g/cm
! , C, = 0.52 J/gK. From this, Q=
5000W, ΔT=500°C, and v=30cm/SeC are required.

Once the gas inlet flow velocity is determined in this way, the convection within the furnace can be solved based on it and the flow velocity at each point can be determined.

Therefore, since the temperature distribution is determined from the heater power of the heat generating source and the gas inlet flow rate, it is possible to determine the temperature distribution in the furnace, including the temperature distribution of the heater, which was conventionally set as a boundary condition.

Next, an analysis of optimal conditions for a single crystal pulling furnace to which the initial value setting method for in-furnace thermal fluid analysis according to the present invention is applied will be described.

This applicant has already filed an application (Patent Application No. 1-149240).
This method utilizes the analysis method (No. 1).

FIG. 2 is a diagram showing an outline of the system configuration.

The optimal condition analysis method for the single crystal pulling furnace shown in Figure 2 is as follows:
As a case study, we set the pressure inside the furnace, sub-upper heater power, sub-lower heater power, crucible position, heater shape, B2O3 thickness, etc., and also set the temperature of the hot zone outer wall, gas inlet flow rate, and temperature as boundary conditions. , physical property values such as heater power, pressure in the furnace, analysis mesh shape, etc. are set as initial values, and stain 5 ration analysis is performed.

This simulation analysis was carried out by the Thermal Fluid Analysis Department.

The feature is that the evaluation is divided into three blocks: a solid-liquid interface shape analysis section 2 and a thermal stress analysis section 3.The thermal fluid analysis section 1 solves convection from the gas inlet flow velocity and calculates thermal radiation. Finally, the solid-liquid interface shape analysis section 2 focuses on the temperature of the crystal/melt part from the temperature distribution in the furnace, and takes over the temperature distribution in the furnace as a boundary condition to analyze the temperature distribution in the furnace.・Thermal analysis of the melt part is performed to determine the solid-liquid interface shape t1 The thermal stress analysis unit 3 calculates the thermal stress in the crystal part from the solid-liquid interface shape obtained by the solid-liquid interface shape analysis part 2 and the temperature distribution of the crystal part. In the end. The data input section 4 is used to input initial value data such as case settings, boundary conditions, and heater power.

With such a configuration, the analysis results are evaluated step by step for each block, and if the conditions at that stage are not satisfied, the process returns to the data manual section 4 to reset and repeat the simulation. Then, by feeding back to the data input section 4 so that the optimum conditions are obtained at each stage, the optimum control conditions are finally obtained from the thermal stress analysis section 3.

The system configured as described above is ultimately intended to optimally design the thermal environment within the furnace, and to grow crystals with less distortion. In other words, the question is how to derive conditions for growing high-quality crystals without destroying the furnace, without actually growing the crystals. For this reason, the temperature distribution inside the furnace and the temperature distribution near the crystal are important.

For example, it is necessary to consider whether or not the crucible shape set in a given container can be adopted, or how large the crucible can be made, and to do so, consider the shape effect of the hot zone consisting of a heater and heat insulating material, and consider multi-division. Optimize the heater placement and power ratio. In this case, if the obtained analysis result satisfies the set criteria, it means that the boundary conditions are met and there will be no temperature slowdown. For example, if the diameter of the crucible is increased by 2 inches, the thickness of the insulation material around the outer periphery will be reduced by 1 inch, so if the heater is used to heat the inside of the furnace under such a change in crucible shape, the heater power required will become abnormal. Will it never rise?
In other words, confirm the safety of the furnace and decide on the crucible position, heater length, etc.

In addition, the shape of the solid-liquid interface and the thermal stress of the crystal are used as indicators for evaluating the conditions for pulling high-quality crystals. In order to reduce defects, it is necessary to make the temperature gradient in the crystal gentler, and the temperature distribution in the crucible becomes important as such information in the vicinity of the crystal. The temperature distribution inside the crucible is determined based on the temperature distribution inside the furnace.

The analysis in each block will be detailed.

Fig. 3 is a diagram showing an example of a study model, Fig. 4 is a diagram showing an example of mesh division, Fig. 5 is a diagram showing an example of gas flow in the furnace,
Figure 6 is a diagram showing an example of the temperature distribution inside the furnace, Figure 7 is a diagram showing an example of the temperature correlation between the crucible position and the central axis of the crystal/melt part, and Figure 8 is a diagram showing an example of the temperature distribution in the furnace.
The figure shows an example of the temperature distribution of the crystal/melt part and the shape of the solid-liquid interface, and FIG. 9 shows an example of the thermal stress distribution.

In Fig. 3, 11 is an Ar atmosphere, 12 is a sub-upper heater, 13 is graphite, 14 and 22 are insulating felts, 15 is a main heater, 16 is a sub-lower heater, 17
18 is a GaAs crystal, 18 is B2O3, 19 is a GaAs melt, 20 is a crucible, and 21 is a crucible support shaft.

In the thermo-fluid analysis, by using the finite element method to divide the inside of the furnace into a mesh as shown in Fig. 4 for the study model shown in Fig. 3, it is possible to take into account the arbitrary shape of the furnace. We are making it possible for you to choose. The physical property values of various substances are defined in the mesh, and simulations are performed by changing the length and position of the heater, the crucible position, and the amount of crystal growth using the mesh. Therefore, in this simulation, we set the outer wall temperature, gas inlet flow rate, and temperature as boundary conditions, and set the mesh as initial values based on physical property values such as the crucible shape (diameter), heater length, and furnace internal structure, in addition to the boundary conditions. conditions and heater power. With these settings, the Ar gas flow in the furnace as shown in FIG. 5 and the temperature distribution in the furnace according to isothermal lines 1.2, . . . as shown in FIG. 6 are determined. Based on these results, each factor will be analyzed to evaluate whether the conditions can be changed, the possibility of crystal pulling, etc. Furthermore, as mentioned above, if the given boundary conditions meet the set criteria, the boundary conditions are met and there is no problem in terms of temperature.

The convergence of the thermo-fluid analysis is determined, for example, when the solid-liquid interface reaches the vicinity of the melting point, and when the power is changed, the gas inlet flow velocity is also changed based on the simulation experiment data, and the simulation analysis is repeated. Determination of the gas inlet flow velocity V is as follows: v=f(
T).

The inside of the furnace is under a high-pressure (for example, tens of atmospheres) atmosphere of inert gas, such as Ar (argon) gas, so when the inside of the furnace is heated by a heater, the gas convects and its flow becomes important. Furthermore, radiation at high temperatures becomes effective as the fourth power of the distance if the shape of the heater changes. Therefore, when the heater power is changed, the gas inlet flow velocity is changed accordingly to solve the convection, and the temperature distribution is solved by taking convection and radiation into account. In this way, thermal fluid analysis analyzes macroscopic temperature distribution, and for heat transfer under high temperature and high pressure, analysis using radiation and gas is important. Calculations can be performed by applying, for example, the thermally coupled Navier-3takes equation to the gas part, and applying a heat conduction equation that takes into account radiation between each surface to the structural member. At this time, the design and setting of the multi-stage heater,
If the crystal pulling position is not properly set, it will be difficult to pull the crystal. The solid-liquid interface between the crystal and the melt actually moves depending on the heat balance state, but at this stage it is assumed to be flat due to mesh division as shown.

In the solid-liquid interface shape analysis, thermal analysis of the crystal/melt part is carried out using the temperature data obtained in the thermo-fluid analysis as a boundary condition, and the solid-liquid interface shape is analyzed from the temperature distribution as shown in FIG. The solid-liquid interface shape is obtained from the temperature distribution as an isotherm of the melting point temperature, and the heater power is set to obtain the desired solid-liquid interface shape. In order to stabilize crystal growth, it is necessary that the flat solid-liquid interface shape remains constant over time, but in reality it is preferable to make the temperature gradient gentle and create a downwardly convex shape. However, if the crystal growth conditions change,
Since the way heat is transferred changes, it becomes impossible to control the raising of the crystal part in a constant manner. For example, a heat conduction equation is applied to the crystal part, and a thermo-fluid analysis that takes convection in the melt into account is applied to the melt part, thereby giving a restrictive condition that makes the solid-liquid interface the melting point. Able to solve boundary problems.

Figure 7 shows an example of the temperature correlation between the crystal and melt center axes when the crucible position is changed using the results of furnace thermal analysis. Changing the position changes the temperature gradient of the crystal part and the temperature of the melt part. This shows that it is also important to set the crucible position appropriately. In particular, although unidirectional solidification is the basis, if the temperature distribution in the melt reverses to a temperature lower than the melting point in the melt below the solid-liquid interface, crystal growth will be interrupted. Therefore, it is necessary to conduct such evaluations as well.

Furthermore, when the power of the sub-upper heater, main heater, and sub-lower heater is changed, the temperature gradient and temperature distribution will also change. FIG. 8 shows the temperature gradient and temperature distribution using equimixture lines due to a temperature difference of 1.2, . . . from the solid-liquid interface, with the solid-liquid interface being O.

In this way, by analyzing the shape of the solid-liquid interface, we can determine whether the temperature gradient of the crystal near the solid-liquid interface is suitable for high-quality crystal pulling from the temperature distribution of the crystal/melt part, how the state of the melt part is, and the temperature distribution of the solid-liquid interface. The optimal shape and homogeneous growth of crystals can be analyzed and evaluated.

In thermal stress analysis, thermal stress analysis is performed using the temperature distribution at the solid-liquid interface shape obtained by solid-liquid interface shape analysis. On the other hand, in order to obtain high quality crystals, it is desirable that the stress be low and uniform. That is, in the end,
The goal is to create crystals with less distortion, but to do so, it is necessary to perform thermal stress analysis based on the shear heat stress distribution and thermal history, as shown in Figure 9, and it is necessary to reduce thermal stress. . In addition, in this analysis, many studies have been conducted on the relationship between thermal stress and crystal defects, so this analysis has been established as one of the evaluation methods for obtaining high-quality crystals.

As described above, if the results are divided into three blocks and can be analyzed sequentially while organically handing over the results to the next block, the effects of the multi-segment heater can be improved by repeating the simulation up to that step by evaluating the analysis results in the middle. etc. can be easily analyzed in various ways. For example, when looking at the temperature distribution in the crystal/melt part, it is possible to analyze how changing the heater power distribution changes the temperature distribution by repeatedly performing simulations while changing the heater power settings. can. Therefore, the point where the solid-liquid interface should be is 1500° (
K), this is the melting point of 1511° (
Finding the distribution of heater power to achieve K) can be done easily by going back and repeating the process block by block, without wasting any time. Furthermore, as a case study, if you change the pressure inside the furnace, if you change the power of the upper sub-heater or the power of the lower sub-heater, if you change the shape of the heater, or if you change the crucible position,
Changes in temperature distribution depending on changes in the thicknesses of B and 03 can also be analyzed in a relatively short time, and the optimum pressure and optimum power heater that provide the optimum operating conditions can be efficiently determined. Note that the shape of the heater, the position of the crucible, the thickness of B2O3, etc. need only be changed in the mesh shape.

The above means can also be used as is to control a single crystal pulling furnace. In this case, since the optimal design of the furnace has already been made through the above analysis, the furnace shape and boundary conditions among the furnace conditions, boundary conditions, and heater power conditions are fixed data, and the crystal pulling rate is The rotational speed of the crucible and crystal, the heater power, etc. are variables associated with crystal pulling.

Therefore, these variables may be corrected sequentially as the crystal is grown.

Therefore, the overall configuration for controlling a single crystal pulling furnace basically consists of a condition setting section that sets the conditions inside the furnace, boundary conditions, and heater power, and a condition setting section that monitors the operating status of single crystal pulling inside the furnace. The thermo-fluid analysis section 1, solid-liquid interface shape analysis section 2, It may be connected to the thermal stress analysis section 3.

Standards for evaluation and judgment at each stage are set in the thermal fluid analysis section 1, solid-liquid interface shape analysis a2, and thermal stress analysis section 3. For example, standard values are set for the temperature at a specific position in the furnace, the position of the solid-liquid interface, etc., and evaluation and judgment are performed by comparing with the analysis results in the thermo-fluid analysis section 1 and the solid-liquid interface shape analysis section 2. Correct the conditions in the setting section one by one. For example, the position of the crucible is changed in accordance with the crystal pulling speed. As mentioned earlier, the monitor section detects the weight of the crystal using a load cell, detects the temperature at each point using a thermocouple, and detects the rotational speed of the crucible and crystal using an encoder attached to the motor and motor drive current, etc. conduct. Then, the control section controls the heater power and the rotation of the crucible and crystal by feeding back the detection data from the monitor section so as to satisfy the conditions set by the condition setting section. By configuring such a control system, high-quality crystals analyzed by the thermal fluid analysis section 1, the solid-liquid interface shape analysis section 2, and the thermal stress analysis section 3 can be efficiently created. It goes without saying that the detection data of the monitor section may also be fed back to each analysis section and used as basic data for evaluation and judgment of analysis results.

Note that the present invention is not limited to the above embodiments, and various modifications are possible. For example, in the above example, the relationship between the gas inlet flow rate is determined from the difference in heater power when the furnace interior is changed from vacuum to operating pressure and the heater monitor temperature is kept constant, and the difference between the gas inlet temperature and the gas outlet temperature. However, a similar simulation experiment may be performed by changing the pressure near the operating pressure or changing the heater monitor temperature. In addition, the optimum condition analysis for a single crystal pulling furnace is characterized by using thermal fluid analysis, solid-liquid interface shape analysis, and thermal stress analysis as independent units of analysis, and connecting these in a subordinate manner. This means is shown as an example. For example, in the above example, the results of thermal fluid analysis, solid-liquid interface shape analysis, and thermal stress analysis were evaluated and the simulation conditions were changed to set the optimal conditions for the thermal environment inside the reactor. For this evaluation, parameters for evaluation and simulation condition change may be set in each analysis section, or a learning function may be added to automatically change simulation condition settings. Also,
It may also be incorporated into the control system of a single crystal pulling furnace.

Since the temperature of the heater and the pressure inside the furnace are constantly monitored, this data and the data detected by the load cell described earlier on the crystal weight are manually input to this system and configured to feed back to the temperature control of the heater. Good too. In this case, changes in the solid-liquid interface shape affect the diameter of the grown crystal.

That is, when the solid-liquid interface shape is convex downward, the buoyancy becomes large, so the weight of the crystal detected by the load cell appears to be lighter at the top. Therefore, the controllability of the diameter may be improved by determining the solid-liquid interface shape from the change in crystal weight, calculating the crystal diameter, and feeding it back to the control of the heater temperature.

〔Effect of the invention〕

As is clear from the above explanation, according to the present invention, in thermo-fluid analysis, the heater power is set and the gas inlet flow velocity is determined from the relationship obtained from the simulated experiment data corresponding to the heater power. , since convection within the furnace is solved, it is possible to perform highly accurate thermal fluid analysis by taking convective heat transfer and thermal radiation into consideration. Moreover, the temperature distribution inside the single crystal pulling furnace due to the multi-segment heater can also be estimated with high accuracy. Furthermore, by analyzing the optimal conditions for a single crystal pulling furnace using divided simulation analysis, we are able to efficiently and precisely create the desired thermal environment even in single crystal pulling equipment that utilizes multi-segmented heaters, which was difficult in the past. It can also be estimated by

[Brief explanation of drawings]

Fig. 1 is a schematic diagram for explaining an example of an analysis condition setting method for in-furnace thermal fluid analysis according to the present invention, and Fig. 2 is a diagram for explaining an example of an optimum condition analysis method for a single crystal pulling furnace. diagram,
Fig. 3 is a diagram showing an example of a study model, Fig. 4 is a diagram showing an example of mesh division, Fig. 5 is a diagram showing an example of gas flow in the furnace,
Figure 6 is a diagram showing an example of the temperature distribution inside the furnace, Figure 7 is a diagram showing an example of the temperature correlation between the crucible position and the central axis of the crystal/melt part, and Figure 8 is a diagram showing an example of the temperature distribution in the furnace.
The figure shows an example of the temperature distribution in the crystal/melt part and the shape of the solid-liquid interface, Figure 9 shows an example of the shear stress distribution in the crystal cross section, and Figure 10 shows a single crystal bow lifting device using the LEC method. FIG. 11 is a diagram showing a heat balance model in the furnace. 1...Thermo-fluid analysis section, 2...Solid-liquid interface shape analysis section,
3...Thermal stress analysis section, 4...Data input section. Figure 3 Figure 7

Claims (4)

[Claims] (1) The boundary conditions are the gas inflow temperature, speed, outer wall temperature, etc. in the furnace heated by a heater in a high-pressure gas atmosphere,
This is an analysis condition setting method for in-furnace thermo-fluid analysis in which the analysis mesh shape, pressure and other physical property values in the furnace, heater power, etc. are set as initial values and the optimum operating conditions are determined by simulation. Calculate the relationship between heater power and gas inlet flow rate to keep the heater monitor temperature constant by changing the pressure of An analysis condition setting method for in-furnace thermal fluid analysis. (2) The relationship between heater power and gas inlet flow rate is determined from the difference in heater power when the pressure inside the furnace is changed from vacuum to operating pressure, and the difference between gas inlet temperature and gas outlet temperature. An analysis condition setting method for in-furnace thermal fluid analysis according to claim 1. (3) Thermofluid analysis to determine the temperature distribution in the furnace; solid-liquid interface shape analysis to determine the solid-liquid interface shape by thermally analyzing the crystal/melt part from the temperature distribution in the furnace; 2. The method for setting analysis conditions for in-furnace thermal fluid analysis according to claim 1, wherein the simulation of the single crystal pulling furnace is performed separately in a thermal stress analysis for determining thermal stress in the crystal part from temperature distribution. (4) Analysis conditions for in-furnace thermofluid analysis according to claim 3, characterized in that the simulation is repeated by changing the heater power for each case of the in-furnace pressure, crucible position, heater shape, etc. to determine the optimal operating conditions. Setting method.