JPH034516B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a method for accurately detecting crystal diameter in single crystal growth in which a crucible is pushed up. Conventionally, an optical method has been used to detect the diameter of a crystal being grown by a pulling method. It is a method that detects the reflection of laser beam irradiation near the crystal during growth, the radiation position of a high-temperature heat generating part (fusion ring) generated on the liquid surface around the crystal, or the transmission image of X-rays or the radiation emitted from the crystal. This method uses a TV to monitor the infrared image generated by the crystal, and extracts an electrical signal proportional to the crystal diameter from the video signal. However, the drawback of such an optical method is that it is susceptible to eccentricity of crystal rotation, fluctuation of reflected light, fogging of a viewing window, etc. Furthermore, in the case of X-rays, there are fatal drawbacks such as the complexity of the equipment and the danger that X-rays may pose to the human body. Furthermore, this optical detection method cannot be used depending on the crystal material. For example, conditions specific to compound semiconductors, such as the intense thermal convection of the high-pressure gas in the pressure chamber and the need to observe through a liquid encapsulant layer covering the melt, make optical detection methods impossible. be. On the other hand, a method of detecting the crystal diameter by detecting the crystal weight and precisely measuring the crystal length has been attempted. This method overcomes the drawbacks of optical crystal extraction methods. However, this method can only be applied in cases where it is sufficient to take into account the natural drop in the liquid level due to a decrease in the melt. Generally, keeping the growth interface (liquid level) in a fixed position stabilizes temperature conditions and is advantageous for growth. Therefore, crucible elevation control may be performed to maintain the interface in a fixed position. Therefore, the above-mentioned method has applicability limitations, and there is a need for further development. The present invention adopts the weight detection method and precision length measurement method, takes into account the drop in the liquid level due to melt reduction and the crucible elevation control to compensate for the drop in liquid level, and accurately adjusts the crystal diameter in real time during growth. The purpose is to be measurable. In particular, the earlier application for the present invention (patent application filed in 1983)
-58646 (Special Publication No. 59-358765) is characterized by solving the problem of correcting the rise of the crucible to compensate for the drop in liquid level. Embodiment 1 FIG. 1 shows a first embodiment of the present method, and shows a block diagram of the apparatus. Reference numeral 1 denotes a weight sensor connected to the pulling shaft 3 to detect the weight of the crystal 11. If it is connected to the crucible push-up shaft 5, the crystal weight will be detected from the amount of decrease in the melt 14. 2 is a precision length measuring device that measures the pulled length of the crystal in conjunction with the pulling axis. 4 is also a precision length measuring device, which measures the lifting distance of the crucible in conjunction with the crucible push-up shaft 5. 6 is a crucible push-up control mechanism, which is controlled by a computer 8. 7 is an interface for inputting the outputs of the weight sensors 1, 2, and 4 and the length measuring device to the computer 8. 8
is a computer, the output of which is sent to the push-up mechanism 6 through an interface 9. Note that 10 is a pulling speed controller, 12 is a growth interface, 13 is a liquid capsule, 14 is a melt, and 15 is a crucible. The principle of crystal diameter detection using this method and crucible elevation control will be explained. The cross section of the crystal during growth becomes approximately circular. Therefore, we consider a crystal to be an aggregate of disks with different diameters and the same thickness. The smaller the thickness of the disk, the closer the approximation to the actual crystal will be. Here, the minute length ΔL of the crystal grown in the pulling direction is defined as the thickness of the disk. Since ΔL is a minute length, it changes slowly. Other factors can be ignored, and the weight change ΔW measured by the weight sensor can be regarded as the weight change of the crystal. Therefore, if the crystal diameter (that is, the diameter of the disk) is D and the crystal density is ρ _s , then the weight change ΔW is ΔW = ρ _s・π・(D/2) ²・ΔL (1) Figure 2 shows the crystal growing. FIG. 2 is a diagram illustrating the lowering of the liquid level and the pushing up of the crucible. (The above-mentioned disk is shown as a hatched area with diameter D and thickness ΔL.) 16 is a seed crystal, 17 is a crystal, 18 is the liquid level position before the crystal grows ΔL, and 19 is when the liquid level is 18. This indicates that the position of the growth interface has shifted upward by Δl. Reference numeral 20 indicates the current liquid level position, 21 indicates the melt, and 22 indicates a cylindrical flat-bottomed crucible. 18 is the position of the liquid level at an arbitrary time,
The figure shows that the crystal was pulled up by Δl by the pulling mechanism, and the crucible was pushed up by Δx during that time, and the crystal grew by ΔL. The relationship between Δl, ΔL, and Δx can be expressed by the following formula. ΔL=Δl/1−ρ _S /ρ _C (D/D _C ) ² −Δx (2) (ρ _s , ρ _L are the densities of the crystal and melt, D is the crystal diameter, and D _c is the inner diameter of the crucible) Therefore, the crystal diameter D can be found from equations (1) and (2) using the following equation. D=D ₀ {-(1-α+β)+〓(1-α+β)
² +4αβ／2αβ} ^1/2 (3) However

[Formula] α=Δx/Δl, β=ρ _s /ρ _L (D ₀ /D _c ) ² (The range of α is 0≦d≦β/1−β) In other words, ΔW, Δl, and Δx are precisely The crystal diameter D can be found by measuring. ΔW, Δl, and Δx are detected by the measuring instruments 1, 2, and 4 shown in FIG. 1, respectively. Also, from equation (2), the condition for ΔL=Δl is expressed by the following equation. α=β/1−β (4) If this relationship is maintained and the crucible is raised, D=D ₀
That is, the liquid level is controlled to a fixed position. Having described the principle above, the actual operation of the apparatus shown in FIG. 1 will now be explained. The pulling length Δl of the crystal was set to 0.2≦Δl≦0.5 (mm).
(See Figure 3) The accuracy of the precision profilometer is ±1μ, allowing for sufficiently accurate measurements. On the other hand, the weight sensor is susceptible to noise, so the noise is processed using a filter or the like. Here, the following operation is repeated during the growth, and the crystal diameter can be continuously determined.
First, while pulling Δl, the crystal weight Wi was measured n times,
As Wi=(ΣWi)/n, as the previous Δl, as the previous Δl
The difference from -1 measured during pulling is ΔW=
--1. Furthermore, detect the crucible push-up distance Δx while raising Δl, and from the values of Δl, ΔW, and Δx (3)
Determine the crystal diameter using the formula. All these calculations are performed by the computer 8. Next, the next crucible push-up distance Δx is determined by equation (4) and outputted as an input to the crucible push-up presser 6. Returning to the initial operation By performing the above operations, we can accurately measure the crystal diameter in real time and control the liquid level in a fixed position to grow crystals under stable conditions.
I was able to do this for crystals. Embodiment 2 Instead of using the precision length measuring devices 2 and 4 in FIG. 1 of the first embodiment, a clock and a shaft movement speed measuring device (for example, a tachometer generator) are used. Time t ₁ , t ₂ (t ₂
〓t ₁ ) is measured and set as Δt = t ₂ − t ₁ , and the crystal pulling rate is
If v _l and crucible push-up speed v _x are given, Δl and Δx in the above equation (2) can be found by the following equation. Δl=v _l Δt, Δx=v _x Δt (5) Since the operating principle is the same as that of the first embodiment, the explanation will be omitted below. As explained above, the present invention has the following advantages. (a) Accurate numerical measurements of growing crystallometers can be performed in real time. (b) When pushing up the crucible, the crystal diameter can be determined accurately even if the crucible pushing speed changes. (c) Even if the crystal diameter changes and the degree of descent of the melt changes, it is possible to maintain the liquid level at almost the same position by adjusting the crucible push-up speed, allowing crystal growth under stable temperature conditions. (d) Since the crystal diameter can be detected, automatic crystal growth is possible by determining the crystal diameter setting location and feeding back the deviation to the growth conditions.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of the apparatus of the present invention, FIG. 2 is a diagram explaining the operating principle of the method, and FIG. 3 is a diagram showing the relationship between crystal pulling length and crystal diameter error. 1... Weight sensor, 2, 4... Precision length measuring device,
3... Crystal pulling axis, 5... Crucible pushing axis, 6
... Crucible push-up control mechanism, 7, 9 ... Interface, 8 ... Computer, 10 ... Pulling speed controller, 11 ... Crystal, 12 ... Growth interface,
13...liquid capsule, 14...melt liquid, 15...
Crucible, 16...Seed crystal, 17...Crystal, 18...
...Position of the liquid level before the crystal grows by ΔL, 19...Position of the growth interface when the liquid level is 18, 20...Position of the current liquid level, 21...Melt, 22...Cylindrical flat-bottomed crucible.

Claims (1)

[Claims] 1. In single crystal growth by the pulling method, the crystal pulling length is measured using a first precision length measuring device, and the unit crystal pulling length Δl is set to 0.2 mm or more and 0.5 mm or less,
For each unit crystal pulling length, measure the push-up length Δx of the crucible using a second precision length measuring device and detect the increase in crystal weight ΔW, and calculate the crystal diameter D using the following formula D = D ₀ {-(1-α+β) +√(1-α+β) ² +4αβ/
2αβ} ^1/2 [Formula] α=Δx/Δl, β=ρ _s /ρ _L (D ₀ /D _c ) ² (However, the inner diameter of the crucible is D _c , and ρ _s and ρ _L are the crystal and melt, respectively. A single crystal precision crystal diameter detection method characterized by: