JPH03137091A - Production of semiconductor single crystal - Google Patents

Abstract

PURPOSE:To obtain single crystal having uniform characteristics at a high yield by calculating the difference of both the measured weight of single crystal being grown and the arithmetic weight calculated from a measured diameter and judging the shape of the interface of solid and liquid from this difference and controlling the conditions of crystal growth on the basis of this judged result in the production of semiconductor single crystal by a pulling-up method. CONSTITUTION:In the production of semiconductor single crystal wherein single crystal 2 is grown by bringing seed crystal 5 into contact with the melt 4 of a raw material for crystal in a crucible 1 and pulling up the seed crystal, the constitution described hereunder is adopted. In other words, the weight of single crystal 2 which is grown while being pulled up is measured by a weight measuring means (a weight sensor 11, a buoyancy correction circuit 14, an arithmetic circuit 18). Further the diameter of single crystal 2 is measured by a diameter measuring means (a camera 12, a detecting circuit 15 for diameter of crystal, an arithmetic circuit 18). The difference of both the measured weight of single crystal 2 which is measured by the weight measuring means and the arithmetic weight of single crystal which is calculated on the basis of the diameter measured by the diameter measuring means is obtained. The shape of the solid-liquid interface of single crystal 2 and the melt 4 of the raw material for crystal is judged from this difference. The conditions of crystal growth are controlled on the basis of this judged result so that the shape of the solid-liquid interface is made to an optimum shape.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a method for manufacturing a semiconductor single crystal by a liquid confinement pulling method (LEC method) or a pulling method (CZ method), and particularly relates to a method for manufacturing a semiconductor single crystal using an optimal solid-liquid interface shape. This study concerns semiconductor single crystal growth.

[Prior art] CZ method is a method for producing crystals of Si, Ge, etc., and GaA
The LEC method, which is an improved version of the CZ method, is known as a method for producing single crystals of ■-■ group semiconductors such as s.

Semiconductor wafers manufactured by these manufacturing methods include:
It is required that there be no variation in the characteristics within the wafer surface. For this purpose, the impurity concentration must be equal at every point within the wafer surface.

When impurities are incorporated into a crystal, they have the property of segregation, so if you try to grow a crystal with a uniform impurity concentration on the wafer surface, it will solidify at the same time at every point on the wafer surface. It is necessary. In other words, it is desired that the solid-liquid interface shape is always flat.

The solid-liquid interface shape of the crystal is determined by the temperature distribution within the crystal raw material melt, and there are many parameters that can change this.

Among these, the liquid level position of the crystal raw material melt relative to the heater is an important parameter that determines the solid-liquid interface shape. That is, when the liquid level of the crystal raw material melt 4 in the crucible 1 is higher than the heater, as shown in FIG.
The solid-liquid interface shape of the single crystal 2 is concave toward the melt 4 side. Note that 3 is B, 03 of the liquid sealant. Conversely, if it is low,
As shown in FIG. 8, the solid-liquid interface shape tends to be convex toward the melt 4 side. Therefore, by optimally selecting crystal growth conditions such as the position of the melt surface relative to the heater, the solid-liquid interface shape can be made flat.

However, in crystal growth by the usual LEC method or CZ method, the solid-liquid interface shape changes with crystal growth. For this reason, the distribution of impurity atoms inside the crystal also reflects the shape of the solid-liquid interface, and the distribution is not uniform within the surface of the crystal. Therefore, in the past, the shape of the solid-liquid interface of the crystal was determined by cutting the grown crystal vertically and performing etching, or by taking an X-ray topographic image, and the shape of the solid-liquid interface was determined to be flat. The method used was to find the optimal crystal growth conditions.

[Problems to be Solved by the Invention] However, the conventional method described above does not detect the solid-liquid interface shape of the crystal during crystal growth, but detects the solid-liquid interface shape only after cutting the crystal obtained after growth. It is something to know. Therefore, in order to determine the crystal growth conditions to flatten the solid-liquid interface shape, it is necessary to first grow the crystal under certain crystal growth conditions, then cut the resulting crystal and measure the solid-liquid interface shape. Then, the flatness must be evaluated, and based on this evaluation, the flatness must be further improved (the process of changing the crystal growth conditions must be repeated many times through trial and error. In order to seek
It required a lot of development time and development cost. Furthermore, if there is a disturbance during crystal growth, the shape of the solid-liquid interface cannot be kept constant.

The purpose of the present invention is to solve the problems of the prior art described above,
It is an object of the present invention to provide a novel method for manufacturing a semiconductor single crystal, which can maintain an optimal solid-liquid interface shape and easily manufacture a semiconductor single crystal with uniform characteristics at a high yield.

[Means for Solving the Problems] In order to achieve the above object, the present invention provides a method for manufacturing a semiconductor single crystal in which a single crystal is grown by bringing a seed crystal into contact with a crystal raw material melt in a crucible and pulling it up. The weight of the growing single crystal to be pulled is measured by a weight measuring means, and the diameter of the single crystal is measured by a diameter measuring means.

Next, calculate the difference between the measured weight of the single crystal measured by the weight measuring means and the calculated weight of the single crystal calculated based on the diameter of the single crystal measured by the diameter measuring means, and from this calculated weight difference. Determine the shape of the solid-liquid interface between the single crystal and the crystal raw material melt.

Based on this determination result, the crystal growth conditions of the single crystal are controlled so that the shape of the solid-liquid interface becomes optimal.

The present invention provides an L material for covering the above crystal raw material melt with a liquid sealant.
It can be applied to both the EC method and the CZ method which is not covered with a liquid sealant. Therefore, applicable materials include not only the ■-■ group compound semiconductors but also Si and Ge.

The above-mentioned measurement by the weight measuring means may be performed not only by providing a load cell or the like on the pulling shaft of the single crystal for direct detection, but also by measuring the weight of the crucible and indirectly detecting the weight of the single crystal.

The optimal shape of the solid-liquid interface is preferably flat in order to make the electrical properties of the crystal uniform.

However, depending on the material, it may be preferable to keep the solid-liquid interface shape slightly convex or concave in order to increase the single crystallization rate. In this case, the optimal shape is not flat but slightly convex or concave, and control is performed to maintain this shape.

In addition, the control parameters for the crystal growth conditions of the single crystal or the shape of the solid-liquid interface include the position of the crystal raw material melt surface with respect to the heater, crucible rotation, crystal pulling speed, crystal rotation, etc.

[Function] The actual measured weight of the single crystal measured by the weight measuring means includes the buoyancy force that the solid-liquid interface portion of the single crystal receives from the crystal raw material melt. In addition, in the case of the LEC method, B
There is also the buoyancy force received from liquid sealants such as , 03, etc.

On the other hand, the calculated weight of a single crystal, which is calculated by calculating the volume by integrating the diameters of the single crystal measured at each predetermined pulling length using a diameter measuring means, indicates that the shape of the solid-liquid interface of the single crystal is flat. It is obtained by estimating that.

Therefore, the difference between the measured weight and the calculated weight is the buoyancy force exerted on the solid-liquid interface of the single crystal. This buoyancy reflects the solid-liquid interface shape of the single crystal, and the solid-liquid interface shape of the single crystal can be determined from the buoyancy value.

This allows the shape of the solid-liquid interface to be determined in real time during crystal growth, and by immediately feeding back this result to the crystal growth conditions, the solid-liquid interface of the single crystal can always be maintained in the optimal shape.

[Example] An example of the present invention will be described below with reference to the accompanying drawings.

FIG. 1 shows an example of a crystal pulling apparatus in which the present invention is applied to the LEC method. In the figure, 1 is a crucible, and the crucible 1 contains a crystal raw material melt 4 for growing a single crystal 2 of a ■-■ group compound semiconductor. The upper surface of the crystal raw material melt 4 is covered with a liquid and B, 0, 3 as a sealant. A heater 6 is arranged around the outer periphery of the Ruppo 1, and a lower shaft is driven via a lower shaft 9 below the Ruppo 1 so that the position of the crucible 1 relative to the heater 6 can be moved up and down. A mechanism 13 is provided. Lower shaft drive mechanism 13
A lower shaft drive control circuit 17 is connected to the lower shaft drive control circuit 17 for controlling the drive speed and the like. Further, the heater 6 is connected to a heater power supply control circuit 16 for temperature control thereof.

On the other hand, above the Lupo 1, there is an upper shaft 8 that supports the seed crystal 5.
A weight sensor 11 is provided above the upper shaft 8 to measure the weight of the single crystal 2 grown from the seed crystal 5. The detection signal detected by the weight sensor 11 is corrected by a buoyancy correction circuit 14 and then input to an arithmetic circuit 18. Ruppo 1. The heater 6 and the like are housed in a high-pressure container 7 in order to prevent dissociation of the constituent elements from the melt 4.
Quartz rod 1 was placed so that the growth of single crystal 2 within it could be observed.
0 is set. A camera 12 for photographing the growing single crystal 2 is installed at one end of the quartz iron 10. The image of the camera 12 is input to a crystal diameter detection circuit 15, and a signal regarding the detected diameter of the single crystal 2 is input to an arithmetic circuit 18.

Next, the operation of this embodiment will be described.

As shown in FIG. 2, the measured weight of the growing single crystal 2 detected by the weight sensor 11 includes a buoyant force corresponding to the volume excluding B*C)s3.

Therefore, in order to correct this buoyancy, the buoyancy correction circuit 14 adds the buoyancy due to B, 033 to the measured weight from the weight sensor 11, and calculates the corrected measured weight Wmes from the arithmetic circuit 18.
is man-powered. Further, the crystal diameter detection circuit 15 uses the image input from the camera 12 to the crystal diameter detection circuit 15.
Then, as shown in FIG. 2, the diameter of the single crystal 2 for each pulling length Δh per unit time is detected, and these detected diameter values are input to the arithmetic circuit 18. The arithmetic circuit 18 calculates the diameter value D I+ D ! of these single crystals 2. ＋・
...D, the calculated weight WcaQ of the single crystal 2 on the melt 4 is calculated by the following formula.

WcaQ-Σ(πD''(n)/4)・Δh-ρ
.

n=1 Here, ρ is the density of the single crystal 2. Calculated weight Wca
i7 is the weight based on the assumption that the solid-liquid interface is flat, so if the solid-liquid interface of single crystal 2 is not flat and is convex (Fig. 2) or concave toward the melt 4 side, is the measured weight w m
es and the calculated weight Wca (do not match, and the convexity of the solid-liquid interface (
The buoyant force (ρ, -ρ, )・V that the concave) part receives from the melt 2
Only the difference appears. Note that ■ is the volume of the convex (or concave) portion of the solid-liquid interface, and ρ is the density of the melt 4.

Therefore, the difference between the calculated weight WcaQ and the measured weight Wmes is Wca12-Wmes= (9m-1) m) 1
V, and the volume of the solid skin interface part is V = (WcaQ-Wmes)/(, Om -9m
).

In this way, if the value of the volume V determined by the calculation circuit I8 is V>Q, it can be determined that the solid-liquid interface of the single crystal 2 is convex toward the melt 4 side, and the larger the value of It can be said that the degree of convexity is strong. Further, if V<Q, it can be determined that the solid-liquid interface shape is concave toward the melt 4 side. In this case, in the correction of the measurement type i1Wg+es mentioned above, the buoyancy is corrected by the weight of the melt 4 in the four parts, and V=
The value of V is determined by (WcaQ-WcaQ)/ρ.

Next, crystal growth conditions are optimized based on the result of determining the shape of the solid-liquid interface based on the value of ■. In this embodiment, the temperature of the heater 6 is controlled by the heater power control circuit I6, and
The lower shaft drive control circuit 17 drives and controls the lower shaft drive mechanism 13 to control the position of the crucible 1. By controlling these so that V=Q, the shape of the solid-liquid interface of the single crystal 2 can always be kept flat during crystal growth.

Next, a specific example in which a GaAs crystal with a diameter of 2 inches was produced using the conventional method and the pulling apparatus of this embodiment described above will be compared and described. In this example, the volume V of the convex portion of the solid-liquid interface of the single crystal 2
The position of crucible I and the temperature of heater 6 were controlled so that O≦V<7.5 cm” during crystal growth.

As a result of vertically cutting a crystal prepared by the conventional method and observing striations by AB etching, the solid-liquid interface shape was convex toward the melt side over the entire crystal, and the height of the convex part was was 30 mm at maximum. On the other hand, in the example of the present invention, the solid-liquid interface shape was almost flat, and the portion that protruded toward the melt side was about 4 mm at maximum.

In addition, a crystal produced by the conventional method was heat-treated at 9000C for 10 hours, and the electrical characteristics within the wafer plane were measured. FIG. 3 shows the measured resistivity distribution, and FIG. 5 shows the threshold voltage distribution when the FET was fabricated. The variation in threshold voltage was as large as 5QmV. On the other hand, the crystal produced in this example was also subjected to similar heat treatment, and the resistivity distribution within the wafer plane and the FET threshold voltage distribution were determined. FIG. 4 shows the resistivity distribution, and FIG. 6 shows the threshold voltage distribution.

As shown in the figure, it can be seen that in this example, the uniformity of the electrical characteristics of the wafer is significantly higher than in the conventional method. In this example, the variation in threshold voltage is 10 m.
We were able to keep it below V.

Next, an embodiment of the pond of the present invention will be explained using a crystal pulling apparatus shown in FIG. This embodiment is characterized in that X-rays are used as a means to detect the diameter of the single crystal 2.
Other points are the same as in the above embodiment. That is, in this embodiment, an X
A ray generating device 19 and an X-ray image receiving device 20 are provided in opposition.

Then, X-rays generated by the X-ray generator 19 are irradiated onto the single crystal 2 and its surroundings, and the transmitted X-rays are received by the X-ray receiver 20, and an electrical signal of the obtained external image of the single crystal 2 is generated. The configuration is such that the data is input to a crystal diameter detection circuit 15.

When a GaAs single crystal with a diameter of 2 inches was produced using this apparatus while controlling the crystal interface, a single crystal with excellent uniformity was obtained as in the previous example.

[Effect of the invention] According to the present invention, there are the following effects.

(The shape of the solid-liquid interface is determined from the buoyancy of the solid-liquid interface of the single crystal, which is the difference between the actual measured weight of the single crystal and the calculated weight calculated from the diameter.) By being able to detect the shape of the solid-liquid interface in real time and feeding it back to the crystal growth conditions, it has become possible to control the shape of the solid-liquid interface during crystal growth, which was previously impossible. Excellent semiconductor single crystals can be manufactured with good reproducibility. Furthermore, it is possible to effectively cope with disturbances that disturb the crystal growth state.

(2) In the past, in order to find the optimal crystal growth conditions, it was necessary to prepare a number of crystals and cut the crystals for evaluation, but in the present invention, such crystal growth conditions This eliminates the need for development time and development costs.

(3) Since the shape of the solid-liquid interface can be kept almost flat, it is also an effective way to avoid situations where the solid-liquid interface shape becomes too convex and the crystal comes into contact with the bottom of the crucible during growth. be.

[Brief explanation of the drawing]

Fig. 1 is a configuration diagram showing a crystal pulling device used in an embodiment of the present invention, Fig. 2 is an enlarged vertical cross-sectional view showing the single crystal growth situation in the crucible of the same device, and Fig. 3 is a conventional method. Figure 4 is a diagram of the in-plane resistivity distribution of a crystal wafer produced by the present invention, and Figure 5 is a diagram of the resistivity distribution in the plane of a crystal wafer produced by the conventional method. Figure 6 is a threshold voltage distribution diagram of an FET manufactured using a crystal wafer manufactured according to the present invention, and Figures 7 and 8 are diagrams of threshold voltage distribution diagrams of an FET manufactured using a crystal wafer manufactured according to the present invention. FIG. 9 is a longitudinal cross-sectional view showing the corresponding solid-liquid interface shape, and FIG. 9 is a configuration diagram of a crystal pulling apparatus used in another embodiment of the present invention. 1 is a crucible, 2 is a single crystal, 3 is B, 03.4 is a crystal raw material melt, 5 is a seed crystal, 6 is a heater, 7 is a high pressure container, 8 is an upper shaft, 9 is a lower shaft, 10 is a quartz rod , 11 is a weight sensor,
12 is a camera, 13 is a lower shaft drive mechanism, 14 is a buoyancy correction circuit, 15 is a crystal diameter detection circuit, 16 is a heater power supply control circuit, 17 is a lower shaft control circuit, 18 is a calculation circuit, 19 is an X-ray generator, 20 is an X-ray image receiving device. Fig. 2 Structure of crystal pulling device Fig. 1 Fig. 3 Fig. 4 Threshold voltage distribution of conventional FET Fig. 5 Threshold voltage distribution of PET of this embodiment Fig. 6 Fig. 7 Fig. 8 Figure 9: Structure of crystal pulling device

Claims (1)

[Claims] In a method for producing a semiconductor single crystal in which a single crystal is grown by bringing a seed crystal into contact with a crystal raw material melt in a crucible and pulling it, the weight of the growing single crystal to be pulled is measured by a weight measuring means. At the same time, the diameter of the single crystal is measured by a diameter measuring means, and the difference between the measured weight of the single crystal measured by the weight measuring means and the calculated weight of the single crystal calculated based on the diameter measured by the diameter measuring means. The shape of the solid-liquid interface between the single crystal and the crystal raw material melt is determined from this difference, and the crystal growth conditions of the single crystal are controlled based on this determination result so that the solid-liquid interface shape is optimal. A method for manufacturing a semiconductor single crystal, characterized in that: