JP2000302590A - Production of crystal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a crystal production process effective in preventing pulling batch-to-batch variation from being caused. SOLUTION: This production process for producing a crystal while surely recognizing the behavior of a melt received in a crucible being used, on the basis of shape data of a standard crucible, involves: estimating the difference-in- thickness Δt between a crucible to be used and the corresponding standard crucible on the basis of the difference-in-weight ΔW between them; shifting the shape data P[n] of the standard crucible by the difference-in-thickness Δt to obtain corrected data P'[n] for the crucible to be used; and perforrming the pulling of a crystal with this crucible by using the corrected data P' [n].

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a method for producing a crystal, and more particularly, to a method for producing a crystal which is effective for preventing variation from batch to batch.

[0002]

2. Description of the Related Art Czochralski method (hereinafter referred to as "CZ method")
Is a technique in which a seed is immersed in a melt contained in a crucible, and the seed is raised to grow a crystal. In the CZ method, generally, the crystal is pulled up while controlling the liquid level of the melt at a constant level. This is for the purpose of fixing the interface between the melt and the crystal, controlling the oxygen concentration of the crystal, and controlling the crystal heat history.

[0003] According to the above-mentioned constant liquid level control, it is possible to easily and suitably control the crystal. Therefore, this constant liquid level control is currently a basic technology for producing a crystal. This constant liquid level control is achieved by ratio control that raises the crucible based on the ratio between the increase amount of the crystal growth length and the melt level decrease amount.

[0004]

However, in the conventional ratio control, it is difficult to actually keep the liquid level constant, and there has been a problem that the control result differs for each pulling batch. Such an error in the constant liquid level control affects the growth rate of the crystal, and in some cases, a desired diameter, oxygen concentration, or crystal heat history cannot be obtained, causing a reduction in yield.

In particular, in recent years, strict control of the crystal diameter, the oxygen concentration, and the heat history of crystallization has been required. It is necessary to reduce the variation of each.

Accordingly, an object of the present invention is to provide a method for producing a crystal which is effective in preventing variation in each pulling batch.

[0007]

In order to achieve the above object, according to the first aspect of the present invention, a crucible (20) is used based on shape data (P [n]) of a standard crucible (200).
2) In the method for producing the crystal (10) while grasping the behavior of the melt (12) accommodated in the crucible, the difference in wall thickness (ΔW) between the standard crucible and the crucible used is determined based on the weight difference (ΔW). Δt) is estimated, and the shape data is corrected using the estimated thickness difference. Claim 2
The described invention uses the shape data (P) of the standard crucible (200).
[N]), measuring the standard crucible weight (W _S ) indicating the weight of the standard crucible, and measuring the used crucible weight (W _C ) indicating the weight of the used crucible (202). Calculating a weight difference (ΔW) indicating a difference between the standard crucible weight (W _S ) and the used crucible weight (W _C ); and converting the weight difference (ΔW) into a thickness difference (Δt). And the shape data (P [n]) is shifted in the normal direction by the thickness difference (Δt) to obtain correction data (P ′).
[N]), and the correction data (P ′).
[N]) to determine the behavior of the melt (12). According to a third aspect of the present invention, in the second aspect of the present invention, the thickness difference (Δt) is represented by [Equation 2]. Where: Δt = wall thickness difference between standard crucible and used crucible; ΔW
= Difference in weight between the standard crucible and the crucible used; S = surface area of the standard crucible; ρ = density of the standard crucible; k = bubble rate; According to a fourth aspect of the present invention, in the second or third aspect, the shape data (P
[N]) and the correction data (P ′ [n]) are composed of coordinate data indicating an X coordinate and a Y coordinate, and the step of generating the correction data (P ′ [n]) includes the step of generating the correction data (P ′ [n]) [N]), calculating a section slope (a [n]) indicating the slope of a straight line connecting the adjacent points P [n-1] and P [n]; and calculating the calculated section slope (a [n]). ]) Calculating an average slope (a ′ [n]) indicating the average value of the adjacent data a [n] and a [n + 1] in the thickness difference (Δt)
Is decomposed into an X component and a Y component by dividing the average gradient (a ′ [n]) and the vector directed in the vertical direction into the X coordinate and the Y coordinate of the shape data (P [n]), respectively. Adding).

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (Summary of the Invention) One feature of the present invention is that a thickness difference Δt between a standard crucible and a used crucible is estimated based on a weight difference ΔW between the standard crucible and a used crucible, and the estimated thickness difference is used. To correct the shape data. The standard crucible means, for example, a crucible having a typical shape as shown in the design drawing of the crucible, and the used crucible means a crucible used for each lifting batch.

The reason why the weight difference ΔW between the standard crucible and the used crucible is determined is to capture an error component generated for each pulling batch. Then, this weight difference ΔW is converted into a thickness difference Δt.
Is converted to geometric information, and reflected in the crucible shape data. Thereby, the shape data referred to at the time of the ratio calculation becomes close to the shape of the crucible used, so that errors for each batch are substantially absorbed.

(Invention Process) Hereinafter, the process leading to the present invention will be described. First, as a cause of the variation in the constant liquid level control for each lifting batch, a difference in shape of the quartz crucible accommodating the melt can be considered. In other words, in the CZ method, the quartz crucible is replaced every time one crystal is pulled up. Therefore, if there is a difference in shape between the individual quartz crucibles, even if the same amount of melt is used, the filling depth and the melt can be reduced. A difference will occur in the diameter of the surface.

The quartz crucible is accommodated in a graphite crucible which is a susceptor, and the material is filled in this state. Since the filled material is heated by the heater, the quartz crucible softens during this heating and fits on the inner peripheral surface of the graphite crucible. Therefore, the shape of the quartz crucible is
It is decided at this stage.

Since the shape of the graphite crucible does not change much due to heating, if the inner peripheral shape of the graphite crucible is constant, the outer edge shape of the quartz crucible also becomes constant. Therefore, regarding the outer edge shape, it is considered that there is no problem in the variation of the pulling batch. On the other hand, since the inner peripheral shape of the quartz crucible is determined by the thickness of the quartz crucible, variation easily occurs. Variations in the thickness of the quartz crucible are inevitable during the manufacturing process, and it is difficult to make the thickness of all quartz crucibles constant. Such variations in the inner peripheral shape of the quartz crucible are considered to cause the following problems.

The ratio calculation, which is the basic process of the constant liquid level control, uses the inner diameter of the crucible at the portion where the surface of the melt is located (hereinafter referred to as "crucible inner diameter CI") as a calculation parameter, so that the ratio calculation is accurately performed. Therefore, it is necessary to accurately determine the diameter of the melt surface.

However, since it is difficult to measure the diameter of the melt surface by an optical method or the like, the crucible inner diameter CI generally measures the amount of change in the position of the melt surface (hereinafter referred to as liquid level MP). From the sum of the measured value and the initial depth at the time of filling the quartz crucible (hereinafter, referred to as “melt initial depth MD ₀ ”), the height at which the crucible is raised (hereinafter, “crucible height CLH”) ) And melt depth MD
Is calculated, and the melt depth MD is determined by applying the melt depth MD to the shape of the quartz crucible.

Therefore, if the inner peripheral shape of the quartz crucible is different for each pulling batch, an error occurs in the crucible inner diameter CI obtained as described above. On the other hand, it is practically difficult and complicated to measure the quartz crucible used for each batch, that is, the individual shape of the crucible used, particularly the shape after softening.

Therefore, it is necessary to grasp the error component of each pulling batch from another viewpoint. As a result of examining the difference between the individual crucibles used in this way, the present inventor has found that the weight of each used crucible has a variation of about ± 15%.

It is considered that the weight variation of about ± 15% is caused or induced by the thickness variation of each crucible used. Therefore, the inventor of the present invention has repeated the creative operation from the viewpoint of capturing the shape of each crucible to be used by utilizing the variation in weight, and has conceived a configuration effective for preventing variation in each lifting batch. Hereinafter, this characteristic new configuration will be described in detail.

(Embodiment of the Invention) The configuration of the present invention will be described below with reference to FIGS. FIG. 1 is a perspective view showing the shape of a standard crucible. As shown in the figure, the standard crucible 200 is a quartz crucible having a thickness of t _s and a surface area of S. In the following description, the surface area of the inner peripheral surface in which the melt is accommodated is described as S, but the standard crucible 200 is used.
May be S.

As described above, the standard crucible 200 means, for example, a crucible having a typical shape as shown in a design drawing. Therefore, this standard crucible 200
Thickness t _s and the surface area S of the can be determined as the data on the design from the design drawing. In addition, the standard crucible 20
0 is not limited to such a design, and one crucible actually manufactured may be used as the standard crucible 200.

FIG. 2 is a perspective view showing the shape of the crucible used. As described above, since the quartz crucible at the start of pulling is softened and fitted to the inner peripheral surface of the graphite crucible, the outer edge shape can be considered to be constant.
However, as shown in the figure, the crucible 202 used is t _c
And the standard crucible 200 has a thickness difference of Δt. The thickness difference Δt is mainly caused by a manufacturing error, and the thickness is generally non-uniform along the wall of the crucible. Therefore, it is difficult to accurately measure the thickness difference Δt. At the time of production of the crystal, the melt 12 is accommodated in the used crucible 202.

FIG. 3 is a conceptual diagram showing the concept of creating the shape data of the standard crucible 200 shown in FIG. As shown in the figure, first, the inner peripheral shape of the standard crucible 200 is approximated by a plurality of shape data P [n]. The curve x = f (y) on which the shape data is based is taken as the center section of the standard crucible 200. Each shape data P [n] is coordinate data expressed in a rectangular coordinate system, and [n] added to the end of each data point
Symbol means an identifier of each data.

Next, the weight of the standard crucible 200 (hereinafter referred to as "the weight")
"Standard crucible weight W _S ") is measured and stored. The steps up to here may be performed at least once, and need not be performed for each pulling batch.

Next, before pulling up the crystal, the weight of the crucible 202 used (hereinafter referred to as “weight crucible used W _C ”).
It was measured to calculate a standard crucible weight W _S the weight difference ΔW use crucible weight W _C. The weight difference ΔW is a parameter indicating an error component for each lifting batch. Weight difference ΔW
The calculation of [Equation 1] What is necessary is just to carry out using the above formula.

Subsequently, the calculated weight difference ΔW is converted into a thickness difference Δt. The conversion to the thickness difference Δt is performed using the standard crucible 2
00 or the density of the crucible 202 (2.2 in the case of quartz glass) and the bubble rate k. The bubble rate k is a parameter that depends on the method of manufacturing the quartz crucible, and is determined according to the manufacturing method.

The conversion of the weight difference ΔW to the wall thickness difference Δt is represented by [Equation 2] Where: Δt = wall thickness difference between standard crucible and used crucible; ΔW
= Weight difference between standard crucible and used crucible; S = surface area of standard crucible; ρ = density of standard crucible; k = bubble rate; If the bubble rate k is not known,
What is necessary is just to assume that the bubble rate k = 0. Or ρ in the above equation
(1-k) may be the bulk density of the quartz crucible.

The conversion according to the above equation is based on the assumption that the cause of the weight difference ΔW is generated as the sum of the thickness differences Δt. That is, it is an approximate value when it is assumed that the thickness difference Δt is uniformly generated along the wall surface of the crucible. This allows
Since the physical quantity called the weight difference ΔW is converted into coordinate information called the thickness difference Δt, the physical quantity can be used for correcting the shape data.

FIG. 4 shows the shape data P [n] shown in FIG.
It is a conceptual diagram which shows the shift concept of. As shown in the figure,
After the estimation of the thickness difference Δt, the shape data P [n] shown in FIG.
Is shifted in the normal direction by the thickness difference Δt, and this is used as correction data P ′ [n].

The normal direction of each data point is a direction orthogonal to the tangent of the curve x = f (y), as shown in FIG. The tangent at each data point can be determined by differentiating the curve x = f (y). Therefore, if the vector having the scalar amount Δt and oriented in the direction orthogonal to the tangent is decomposed into the X component and the Y component and added to the X coordinate and the Y coordinate of the shape data P [n], the correction data P '[N] can be obtained.

The shift direction of the shape data P [n] is given by the following equation (2).
Is positive, that is, the thickness of the standard crucible 200
Is larger than the thickness of the crucible 202 used
(T_s> T_c), The direction shown in FIG.
The result is negative, that is, the thickness of the standard crucible 200 is used.
If the thickness is smaller than the thickness of the crucible 202 for use (t_s<T
_c), The direction is opposite to the direction shown in FIG.

FIG. 5 is a conceptual diagram showing a crucible shape represented by correction data P '[n] obtained by shifting shape data P [n]. As shown in the figure, the shape of the used crucible 202 is approximately represented by a set of correction data P ′ [n], and these correction data P ′ [n]
Is approximately x = f ′ (y). If such an approximate curve is derived, the behavior of the melt 12 at the time of lifting can be easily grasped.

That is, as shown in FIG.
2 (hereinafter, referred to as “melt depth MD”) corresponds to the Y coordinate of the curve x = f ′ (y), and therefore, for example, the melt depth MD detected by the above-described method or the like. Is substituted for the variable y of this curve, the crucible inner diameter CI of the used crucible 202 is obtained.

When P [0] = (0,0), P '[0] = (0, -Δt)
And the reference point after the correction is shifted by Δt. When it is desired to set the corrected reference point P ′ [0] to (0, 0), Δt may be added to the Y coordinate of all the correction data P ′ [n].

According to the present invention configured as described above, the shape data P [n] is corrected by using the parameter of the weight difference ΔW indicating an error for each pulling batch. Variations are absorbed and controllability between batches can be improved.

[0034]

EXAMPLE (Summary) Weight difference Δ between standard crucible and used crucible
Based on W, the thickness difference Δt between the two is roughly estimated, and the crucible shape data P [n] is shifted by the thickness difference Δt,
The correction data P ′ [n] is created. Then, the crystal is pulled up using the created correction data P ′ [n] (see FIG. 4).

(Preferred Embodiment) The above-described technical idea of obtaining the thickness difference Δt using the weight difference ΔW is a very useful idea in the production of a crystal in which a quartz crucible is exchanged for each pulling batch. It is. Here, an example is shown in which this characteristic technical idea is embodied in a mode considered to be industrially preferable. Among the components described above, those that do not need to be particularly described are given the same names and the same reference numerals, and detailed description thereof will be omitted.
The embodiments described below are embodied examples of the present invention, and do not limit the present invention.

FIG. 6 is a conceptual diagram showing an example of storing shape data P [n]. As shown in FIG.
The shape data P [n] obtained from the above shape is stored as coordinate data in an orthogonal coordinate system including an X coordinate and a Y coordinate. As shown in the figure, a data point P [n] is composed of X [n] indicating the X coordinate of the data point and Y [n] indicating the Y coordinate.

FIG. 7 is a plan view showing a method for calculating the surface area of the inner peripheral surface of the standard crucible 200. As shown in the figure, the surface area S of the standard crucible 200 is the shape data P
It is obtained as a total value of a plurality of planes divided by [n]. That is, the area S [n] of each region located between the point P [n-1] and the point P [n] is i = 1, 2 ≦ i <n, i = n + 1.
For each case, [Equation 3] Where: X [i] = X coordinate of data point P [i]; Y
[I] = Y coordinate of data point P [n]; [Equation 4] [Equation 5] It is determined by executing the above equations.

Then, the area S of each region calculated above is calculated.
Using [n], [Equation 6] The above equation is executed, and the resulting value is used as a standard crucible 200.
Surface area S.

After the execution of the above-described equation (2), the above-mentioned equation (6) is obtained.
Is performed, the surface area S obtained by performing the above is substituted into the above-described equation 3, and the thickness difference Δt is calculated.

FIG. 8 shows the shape data P [n] shown in FIG.
It is a conceptual diagram which shows the correction method of. As shown in the figure, first, adjacent points P [n−1] and P in the shape data P [n]
[N] (hereinafter referred to as “section slope a
[N]) is calculated for each data point. This section gradient a [n] is given by [Equation 7] Where: Y [n] = Y coordinate of data point P [n]; Y [n
-1] = Y coordinate of data point P [n-1]; X [n] = X coordinate of data point P [n]; X [n-1] = data point P
X coordinate of [n-1]; calculated by executing the above equation.

Next, the average value of the adjacent data a [n] and a [n + 1] in the calculated section gradient a [n] (hereinafter, referred to as “a [n + 1]”)
"Average slope a '[n]" is calculated. This average slope a ′ [n] is given by [Equation 8] It is calculated by executing the above equation.

Subsequently, a vector having the thickness difference Δt as a scalar amount and a vector oriented in the vertical direction with the average inclination a ′ [n] is decomposed into an X component and a Y component. The X component of the vector is a component indicated by ΔX in the figure, and the Y component is a component indicated by ΔY. These ΔX and ΔY exist for each data point, and the components for the n-th data are represented by ΔX [n] and ΔX [n].
Notated as Y [n].

ΔX [n] is given by [Equation 9] Calculated using the above equation, and similarly, ΔY [n] is expressed by [Equation 1
0] It is calculated using the above equation.

Therefore, the X component X '[n] of the correction data P' [n] is given by [Equation 11] The Y component Y ′ [n] of the correction data P ′ [n] calculated using the above equation is expressed by [Equation 12]. It is calculated using the above equation.

What should be noted here is the data point P [0] indicating the bottom surface of the crucible and the data point of the portion where the wall surface of the crucible is vertical. That is, since the tangent of the data point P [0] indicating the bottom of the crucible has a slope of 0 °,
The correction data P ′ [n] shifts by Δt to the Y coordinate and does not shift to the X coordinate. On the other hand, since the tangent line of the data point in the portion where the wall surface of the crucible is vertical has a slope of 90 °, the correction data P ′ [n] of the portion has Δt
Shift, not Y coordinate.

This indicates that the difference in wall thickness affects the depth of the crucible as it approaches the bottom of the crucible, and that the difference in wall thickness affects the inner diameter of the crucible as it approaches the opening of the crucible. According to the present invention, by setting the shift direction of each data point to the normal direction of each data point, the effect of the thickness difference on the crucible depth and inner diameter is substantially corrected.

FIG. 9 is a conceptual diagram showing a storage example of the correction data P '[n]. As shown in FIG.
The correction data P ′ [n] obtained by shifting [n] is
It is stored as coordinate data (X ′ [n], Y ′ [n]) in a rectangular coordinate system including the X coordinate and the Y coordinate. When pulling the crystal, coordinate data stored as shown in FIG. Hereinafter, how the data corrected in this way is used at the time of actual lifting will be described.

FIG. 10 is a partial sectional view showing the structure of a crystal manufacturing apparatus according to a preferred embodiment of the present invention. Less than,
The configuration of the crystal manufacturing apparatus will be described with reference to FIG.
In the following description, <> added after the signal name
Indicates the unit of the physical quantity.

The main control unit 30 makes full use of the seed control unit 32, the crucible control unit 48, and the heater control unit 34,
The binary control of the crystal growth diameter GD and the seed rising speed SL is executed. The main controller 30 determines the seed rising speed SL, the crucible rising speed, and the heater temperature in order to achieve the binary control, and uses the determined values as the seed controller 3.
2, the crucible control unit 48, and the heater control unit 34. Further, the main controller 30 controls the melt 1
In order to keep the liquid level of No. 2 constant, liquid level constant control is performed in which the crucible 14 is raised at a predetermined ratio as the crystal 10 grows. With this constant liquid level control, it is possible to treat the rising height of the seed 18 and the crystal growth length GL as equal.

The seed controller 32 has a control mechanism for raising and lowering and rotating the seed 18 and a weight sensor 26 for measuring the crystal growth weight GW (see FIG. 11).
The seed 18 is raised at the seed raising speed SL determined by.

The crucible control unit 48 has a control mechanism for raising and lowering and rotating the crucible 14 (see FIG. 11), and raises the crucible 14 at the speed determined by the main control unit 30.

The heater control unit 34 outputs HCNT <W based on the output HPWR <volt> signal of the main control unit 30.
/ H> signal, and outputs the generated signal to the heater 16. As a result, heater 16 has HCNT <W /
h>, heat is supplied to the crucible 14.

The liquid level sensor 28 is disposed above the melt 12 and optically detects the liquid level MP. Then, the detected value is output to the main control unit 30 as an MP <volt> signal.

The heat retaining cylinder 40 is disposed on the outer periphery of the heater 16, and retains the heat emitted from the heater 16 inside thereof, thereby improving the efficiency of supplying heat to the crucible 14.

The temperature sensor 42 is disposed inside the heat retaining cylinder 40 and detects the temperature around the heat retaining cylinder 40. The detected temperature is used as a TMP <volt> signal as the main control unit 3.
Output to 0. Note that, instead of the temperature sensor 42, a radiation thermometer may be provided around the heat retaining cylinder 40 to measure the temperature of the shield material forming the inside of the heat retaining cylinder 40.

The chamber 38 hermetically accommodates the crystal 10 and hot zone components such as the crucible 14 and the heater 16 therein. Argon gas is supplied into the chamber 38.

The crucible shaft 46 is connected to the crucible support 44.
And is moved up and down and rotated by the power supplied from the crucible control unit 48. Crucible support 44
Puts the crucible 14 on the upper surface and sets the crucible shaft 4
6 to follow the vertical movement and the rotation. as a result,
The crucible 14 moves up and down and rotates.

FIG. 11 shows the configuration of the seed control unit 3 shown in FIG.
FIG. 2 is a block diagram illustrating a configuration of a crucible control unit 2;
Hereinafter, the configurations of the seed control unit 32 and the crucible control unit 48 will be described with reference to FIG.

The first motor amplifier 54-1 receives the output SL <volt> signal of the main control unit 30 as a setting signal, and refers to the rotation speed of the first gear 52-1 to generate the motor drive power SCNT <volt>. Generate. Then, the generated signal is output to the first motor 50-1.

The first motor 50-1 is connected to the first gear 52-1 according to the output SCNT of the first motor amplifier 54-1.
To rotate. As a result, the wire drum 24 rotates, the wire 22 is wound, and the seed 18 is raised. When lowering the seed 18, the first motor 50-1 is rotated in the reverse direction.

The first rotary encoder 56-1 converts the rotation speed of the first gear 52-1 into a pulse signal and outputs the pulse signal to the first pulse counter 58-1. The first pulse counter 58-1 counts the pulse signals received from the first rotary encoder 56-1 and outputs the counted result to S
The signal is output to the main control unit 30 as an LH signal (seed rising height). When the seed 18 is moving down, the count value of the first pulse counter 58-1 is decremented.

A structure for rotating the seed 18 is provided in the seed control unit 32 in addition to the structure shown in FIG. This configuration is similar to the configuration for raising the seed 18 described above, and the description is omitted here.

The second motor amplifier 54-2 receives the output CL <volt> signal of the main control unit 30 as a setting signal, and refers to the rotation speed of the second gear 52-2 to generate the motor drive power CCNT <volt>. Generate. Then, the generated signal is output to the second motor 50-2.

The second motor 50-2 is connected to the second gear 52-2 according to the output CCNT of the second motor amplifier 54-2.
To rotate. As a result, the crucible shaft 46 moves upward, and the crucible 14 rises. When lowering the crucible 14, the second motor 50-2 is rotated in the reverse direction.

The second rotary encoder 56-2 converts the rotation speed of the second gear 52-2 into a pulse signal and outputs the pulse signal to the second pulse counter 58-2. The second pulse counter 58-2 counts the pulse signals received from the second rotary encoder 56-2, and outputs the counted result to C
It is output to the main control unit 30 as an LH signal (crucible rise height). When the crucible 14 is descending, the count value of the second pulse counter 58-2 is decremented.

In the crucible control unit 48, in addition to the structure shown in FIG. This configuration is similar to the configuration for raising the crucible 14 described above, and the description is omitted here.

FIG. 12 is a block diagram showing a configuration of heater control unit 34 shown in FIG. As shown in the figure, the heater control unit 34 is configured by a feedback control system using a thyristor and a power sensor. Since such a configuration is a well-known technique, detailed description will be omitted.

FIG. 13 is a block diagram showing the configuration of the first block of main control unit 30 shown in FIG. Hereinafter, the configuration of the first block will be described with reference to FIG. In the following description, the parameters included in the transfer function are unified as follows.

K_V= Speed conversion constant, K_T= Temperature conversion constant
Number, T_DV= Speed control system differential time, T _DT= Temperature control system
Differentiation time, T_IV= Speed control system integration time, T_IT= Temperature
Control system integration time, α_V= Differential coefficient of speed control system, α_T= Warm
Degree control coefficient, P_V= Proportional gain of speed control system, P_T
= Temperature control system proportional gain.

The first amplifier 66-1 converts the digital input signal SLH into SLH <mm>, and converts this SLH <mm>.
Is the crystal growth length GL <mm>, and the first operation execution unit 6
8-1, a target diameter determining unit 78, and a target speed determining unit 80 shown in FIG. The first amplifier 66-1
The subsequent stage is configured by software.

The target diameter determining unit 78 determines the crystal growth length GL.
Is stored in advance as a program pattern, and GL <mm> is applied to the program pattern to determine a corresponding target diameter. Then, the determined value is set as GD (GL) <mm> and the first operation execution unit 68-
1 to the fourth operation execution unit 68-4 shown in FIG.

The first operation execution unit 68-1 calculates [Equation 13] Where: Dcrystal = specific gravity of crystal 10; π = pi; GL = crystal growth length; GD (GL) = target diameter;
The above calculation is executed to predict a target weight corresponding to the target diameter. Then, the predicted weight GPW <g> is output to the second subtractor 70-2.

The third amplifier 66-3 converts the analog input signal GW <volt> to GW <g>, and the GW <g>
To the second subtractor 70-2. This third amplifier 66
The latter part of -3 is constituted by software.

The second subtractor 70-2 calculates GPW <g> and G
The difference between W <g> is calculated to generate a weight deviation GWD <g>, and the generated value is referred to as a D-type speed control amplifier 72 and a PID
It is output to the mold temperature control amplifier 74.

The D-type speed control amplifier 72 is given by [Equation 14] GWD <g> is processed by the transfer function to generate a seed rising speed manipulated variable SLC <mm / min>. And
The generated value is output to the fourth subtractor 70-4 shown in FIG.

The PID type temperature control amplifier 74 is represented by [Equation 1]
5] GWD <g> is processed by the above transfer function, and the temperature manipulated variable T
Produces C <° C>. Then, the generated value is output to the fifth subtractor 70-5 shown in FIG.

FIG. 14 is a block diagram showing a configuration of the second block of main control unit 30 shown in FIG. Hereinafter, the configuration of the second block will be described with reference to FIG.

The second amplifier 66-2 converts the digital input signal CLH into CLH <mm>, and this CLH <mm>
To the third subtractor 70-3. The second stage of the second amplifier 66-2 is configured by software.

The seventh amplifier 66-7 outputs the analog output MP <volt> of the liquid level sensor 28 shown in FIG.
mm> and converts the MP <mm> to a third subtractor 70-3.
Output to The subsequent stage of the seventh amplifier 66-7 is configured by software.

The third subtractor 70-3 calculates MD ₀ and MP <m
CLH <mm> is subtracted from the sum of m> to generate a melt depth MD <mm>. Then, the generated value is shown in FIG.
5 is output to the second operation execution unit 68-2.

FIG. 15 is a block diagram showing a configuration of a third block of main control unit 30 shown in FIG. Hereinafter, the configuration of the third block will be described with reference to FIG.

The target speed determining unit 80 determines the crystal growth length GL
Is stored in advance as a program pattern, and GL <mm> is applied to the program pattern to determine a corresponding target speed. Then, the determined value is output to the fourth subtractor 70-4 as SL (GL) <mm>. This SL (GL) <mm / min> is the target value of the crystal growth rate GR.

The fourth subtractor 70-4 outputs SL (GL) <m
m / min> and SLC <mm / min>, and the result is taken as the fourth amplifier 66-4 and the fourth operation execution unit 6
8-4.

The fourth amplifier 66-4 is connected to the fourth subtractor 7
The output of 0-4 is converted to an analog signal SL <volt> and output to the first motor amplifier 54-1 shown in FIG. The subsequent stage of the fourth amplifier 66-4 is configured by hardware.

The second calculation execution unit 68-2 calculates the crucible inner diameter CI corresponding to the melt depth MD input in the following procedure with reference to the crucible shape table 82 storing the values shown in FIG. I do.

(1) Third subtractor 70-3 shown in FIG.
Using the melt depth MD output from the crucible shape table 82 as a search key, the data point n-1 satisfying the following condition is obtained.
And n.

Y ′ [n−1] ≦ MD <Y [n] where: Y [n−1] = crucible depth at data point n−1; MD = melt output from third subtractor 70-3 Depth; Y [n] = crucible depth at data point n (2) Using the values of data points n−1 and n obtained as a result of the above search, [Equation 16] The above equation is executed, and the obtained CI <mm> is output to the fourth operation execution unit 68-4.

The third operation execution unit 68-3 calculates [Equation 17] Where: W ₀ = initial charge weight; MD ₀ = initial melt depth; CI (y) = crucible inner diameter corresponding to crucible depth y; MD ₀ <mm> obtained by executing the above equation 14 to a third subtractor 70-3. This melt initial depth MD ₀ is calculated and stored at the start of the lifting, and is stored.
Thereafter, the stored value is used.

The fourth operation execution unit 68-4 calculates [Equation 18] Where: Dcrystal = specific gravity of crystal 10; GD
(GL) = target diameter; Dmelt = specific gravity of melt 12;
CI (MD) = the diameter of the crucible 14 at the portion where the liquid surface of the melt 12 is in contact; SL (GL) = the crystal growth speed output by the target speed determining unit 80; SLC = the manipulated variable based on the weight deviation; And execute the result.
Output to 6-5.

The fifth amplifier 66-5 has a relation of CL <mm / mi.
n> into an analog signal CL <volt>, and
To the second motor amplifier 54-2. The subsequent stage of the fifth amplifier 66-5 is configured by hardware.

FIG. 16 is a block diagram showing the configuration of the fourth block of main control unit 30 shown in FIG. Hereinafter, the configuration of the fourth block will be described with reference to FIG.

The fifth subtractor 70-5 generates a heater temperature HT <° C.> by taking the difference between the set temperature Tset <° C.> of the heater 16 and TC <° C.>. Then, the generated value is output to the sixth amplifier 66-6.

The sixth amplifier 66-6 converts HT <° C.> into an analog signal HT <volt>, and outputs a sixth subtractor 70-
6 is output. The subsequent stage of the sixth amplifier 66-6 is configured by hardware.

The sixth subtractor 70-6 outputs HT <volt>
The difference between the signal and the output TMP <volt> of the temperature sensor 42 is calculated to generate a temperature deviation HTD <volt>. Then, the generated signal is output to the PID type temperature control amplifier 84.

The PID type temperature control amplifier 84 is expressed by [Equation 1]
9] The HTD <volt> is processed by the transfer function to generate a power signal HPWR <volt>. Then, the generated value is output to the heater control unit 34 shown in FIG.

With the above configuration, the liquid level constant control is stabilized, and the production of crystals is favorably performed. Although the above-described embodiment is an example in the case of performing the liquid level constant control, the present invention can be applied to the case where the liquid level is moved, and it is needless to say that the same effect can be expected. In addition, the present invention can be expected to produce a more suitable crystal by combining it with a technology for converting the actual charge weight into data or a technology for correcting wire elongation.

[0097]

As described above, according to the present invention,
It is possible to provide a method for producing a crystal that is effective for preventing variation in each pulling batch.

Further, according to the present invention, the shape data P [n] is corrected by using the parameter of the weight difference ΔW indicating the error for each pulling batch, so that the shape variation of the quartz crucible is absorbed, and An improvement in controllability between batches can be expected.

[Brief description of the drawings]

FIG. 1 is a perspective view showing the shape of a standard crucible.

FIG. 2 is a perspective view showing the shape of a crucible used.

FIG. 3 is a conceptual diagram showing a concept of creating shape data of the standard crucible 200 shown in FIG.

FIG. 4 is a conceptual diagram showing a concept of shifting shape data P [n] shown in FIG.

FIG. 5 is a conceptual diagram showing a crucible shape represented by correction data P ′ [n] obtained by shifting shape data P [n].

FIG. 6 is a conceptual diagram illustrating a storage example of shape data P [n].

FIG. 7 is a plan view showing a method for calculating the surface area of the inner peripheral surface of the standard crucible 200.

8 is a conceptual diagram showing a method of correcting the shape data P [n] shown in FIG.

FIG. 9 is a conceptual diagram illustrating a storage example of correction data P ′ [n].

FIG. 10 is a partial cross-sectional view illustrating a configuration of a crystal manufacturing apparatus according to a preferred embodiment of the present invention.

11 is a block diagram showing a configuration of a seed control unit 32 and a crucible control unit 48 shown in FIG.

FIG. 12 is a block diagram showing a configuration of a heater control unit 34 shown in FIG.

FIG. 13 is a block diagram showing a configuration of a first block of a main control unit 30 shown in FIG.

14 is a block diagram showing a configuration of a second block of the main control unit 30 shown in FIG.

15 is a block diagram showing a configuration of a third block of the main control unit 30 shown in FIG.

16 is a block diagram showing a configuration of a fourth block of the main control unit 30 shown in FIG.

[Explanation of symbols]

 10: crystal, 12: melt, 14: crucible, 16: hi
, 18 ... seed, 20 ... seed chuck, 22 ...
Wire, 24: wire drum, 26: weight sensor,
28: liquid level sensor, 30: main control unit, 32: seed control
Section, 34: heater control section, 38: chamber, 40 ...
Insulating cylinder, 42: temperature sensor, 44: crucible support, 46
... crucible shaft, 48 ... crucible control unit, 50-1 ... No.
1 motor, 50-2 ... second motor, 52-1 ... first
Gear, 52-2: second gear, 54-1: first motor
54-2 ... second motor amplifier, 56-1 ... first
Rotary encoder, 56-2 ... second rotary encoder
Coder, 58-1 ... first pulse counter, 58-2 ... second
2 pulse counter, 60 thyristor controller, 6
2: AC / DC converter, 64: Power sensor, 66-1: No.
1 amplifier, 66-2: 2nd amplifier, 66-3: 3rd amplifier
, 66-4 ... fourth amplifier, 66-5 ... fifth amplifier, 6
6-6: sixth amplifier, 66-7: seventh amplifier, 68-1
... first operation execution unit, 68-2 ... second operation execution unit, 68-
3 ... third calculation execution unit, 68-4 ... fourth calculation execution unit, 70
-2: second subtractor, 70-3 ... third subtractor, 70-4 ...
Fourth subtractor, 70-5 ... fifth subtractor, 70-6 ... sixth subtraction
A calculator, 72: D-type speed control amplifier, 74: PID-type temperature
Operation amplifier, 78: Target diameter determination unit, 80: Target speed determination
Fixed part, 82: crucible shape table, 84: PID type temperature
Control amplifier, 86 ... PD type speed operation amplifier, 88 ... I type
Temperature operating amplifier, 100: Crystal growth rate determining unit, 102
... Ratio calculation unit, 104 ... Liquid level moving speed addition unit, 106 ...
Liquid level fluctuation detector, 108: speed converter, 200: standard
Crucible, 202: crucible used, a [n]: section inclination,
a '[n]: average inclination, CI: crucible inner diameter, CL: crucible
Boiling speed, CLH: Crucible rising height, GD: Crystal growth
Diameter, GL: crystal growth length, GR: crystal growth rate, GW
... crystal growth weight, GWD ... weight deviation, MD ... melt depth
Well, MD ₀… Melt initial depth, MP… Liquid level, P [n]…
Shape data, P '[n]: correction data, SL: on seed
Lifting speed, SLH: Seed rising height, Δt: Wall thickness difference, ΔW
... weight difference, W_C... weight of crucible used, W_S… Standard crucible weight
amount,

Claims (4)

[Claims] 1. The shape data (P) of a standard crucible (200)
[N]), while grasping the behavior of the melt (12) contained in the used crucible (202),
In the method for manufacturing 0), the thickness difference (Δt) between the standard crucible and the used crucible is estimated based on the weight difference (ΔW), and the shape data is corrected using the estimated thickness difference. A method for producing a crystal, comprising: 2. The shape data (P) of a standard crucible (200)
[N]); measuring a standard crucible weight (W _S ) indicating the weight of the standard crucible; and using a crucible weight (W) indicating the weight of the used crucible (202).
Measuring a _C), said standard crucible weight _{(W S)} and using the crucible weight _(W C)
Calculating a weight difference (ΔW) indicating the difference between the two, and converting the weight difference (ΔW) into a thickness difference (Δt); and converting the shape data (P [n]) into the normal direction. Creating a correction data (P ′ [n]) by shifting by the thickness difference (Δt); and melting (1) based on the correction data (P ′ [n]).
2) a method for producing a crystal, comprising a step of grasping the behavior. 3. The thickness difference (Δt) is expressed by [Equation 2]. Where: Δt = wall thickness difference between standard crucible and used crucible; ΔW
3. The weight difference between the standard crucible and the crucible used; S = surface area of the standard crucible; ρ = density of the standard crucible; k = bubble rate; 4. The shape data (P [n]) and the correction data (P ′ [n]) are composed of coordinate data indicating an X coordinate and a Y coordinate. The creating step includes a section gradient (a) indicating a gradient of a straight line connecting the adjacent points P [n-1] and P [n] in the shape data (P [n]).
[N]), and adjacent data a in the calculated section slope (a [n]).
Average slope (a ′) indicating the average value of [n] and a [n + 1]
[N]), and a vector which has the thickness difference (Δt) as a scalar quantity, and which has the average inclination (a ′ [n]) and the vector oriented in the vertical direction is represented by an X component and a Y component.
4. The method for producing a crystal according to claim 2, further comprising a step of decomposing the crystal into components and adding the X and Y coordinates of the shape data (P [n]).