JP2783467B2 - Determination of impurity concentration in single crystal silicon rod - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a single crystal silicon rod grown by a pulling method, and more particularly to a method for accurately determining the impurity concentration in a single crystal silicon rod.

[0002]

2. Description of the Related Art In the pulling method, a single crystal seed is immersed in molten silicon in a crucible, and the single crystal seed is gradually pulled while being rotated, so that a single crystal silicon rod grows. For quality control of single-crystal silicon rods, it is necessary to understand the concentration of impurity elements in this single-crystal silicon rod.However, only a trace amount of impurity elements is contained in the molten silicon being pulled. At present, it is difficult at present to directly measure the impurity element contained in the molten silicon being pulled up.

[0003] In view of this, what is called a residual hot water analysis is to quantitatively analyze the impurity element present in the silicon remaining after solidification in the crucible after the lifting, and calculate the impurity concentration in the single crystal silicon rod from the measured value. The law had been implemented.

In a normal silicon single crystal pulling step, a single crystal silicon rod is grown from about 40 to 50 kg of molten silicon, and about 4 to 5
kg of molten silicon residue remains in the crucible. Therefore, the solidification rate in the single crystal silicon pulling step reaches about 90%. This silicon residue solidifies when the heating is stopped. However, solidification of the residue of silicon melting does not occur at the same time, but the solidification start time varies depending on the position in the crucible. This difference in the solidification start time causes segregation of impurities, and causes a significant difference in the impurity concentration distribution in the solidified silicon residue. 4 and 5 are XMA images (magnification: 40) of a silicon mass obtained from molten silicon doped with arsenic, and FIG. 4 is a photo of a slowly solidified portion of the silicon mass. ,
The fast solidified portion of the same silicon mass is shown in FIG. In the figure, the white pattern has a ratio of arsenic to silicon of 1:
The portion segregated to 1 is shown. As is apparent from a comparison between FIG. 4 and FIG. 5, even in the same silicon lump, the portion solidified late has a large white pattern ratio and a high arsenic concentration.

As is clear from the above description, the impurities uniformly contained in the molten silicon segregate during solidification, and even if a part of the solidified silicon in the crucible is quantitatively analyzed, the analysis result is the same as the sample. It changes greatly depending on the collection position. Therefore, in the method of quantitatively analyzing a sample collected from the solidified residue remaining in the crucible after the completion of pulling up single crystal silicon and estimating impurities in the single crystal silicon rod from the analysis value (hereinafter, referred to as conventional method 1), The estimated value of the impurity concentration in the single-crystal silicon rod fluctuates depending on the position where the sample to be subjected to the quantitative analysis is collected, and there is a problem that the reliability of the estimated value is low.

It has been proposed to improve the solidification rate in order to solve the problems of the conventional method for estimating impurities of a single crystal silicon rod (see the Spring 1989 Preprints, 2p-ZP-2). In this method, the solidification rate of silicon is increased to 99% or more, a small amount of silicon residue is solidified, the solidified silicon is completely dissolved with nitric hydrofluoric acid, and then quantitative analysis is performed, and the impurity concentration is estimated from the analysis result. (Hereinafter referred to as Conventional Example 2). In this method, since the segregation of impurities does not affect the analysis value of the quantitative analysis, an estimated value with less fluctuation than in Conventional Example 1 can be obtained.

[0007]

However, the conventional example 2 has a problem that the cost required for increasing the solidification rate is large and the production cost of the product is increased. In other words, although the pulling method is essentially sufficient to pull only single-crystal silicon rods that can be sold,
In order to reduce the residue for quantitative analysis, the tail portion of a single crystal that cannot be sold as a product must be continuously raised. In particular, the efficiency of the heating of the crucible deteriorated with the decrease of the molten silicon residue, and the cost required to continue the operation of the pulling machine was greatly increased.

Further, the solidified polycrystalline silicon must be dissolved with nitric hydrofluoric acid. However, since it is hardly soluble, if the silicon residue is 100 g or more, it takes a long time to dissolve. On the other hand, in order to shorten the time required for dissolution,
If the solidification rate is further increased so that the silicon residue is less than 100 g, there is a problem that the cost required for operating the above-described lifting machine further increases.

In addition, the shape of the tail portion of the single-crystal silicon rod to be pulled is difficult to control, and there is a problem in the appearance of the product.

[0010]

The uneven distribution of impurities occurs during the process of solidifying silicon, and the impurities are uniformly distributed in the molten silicon residue before solidification. Therefore, if the sample is directly collected from the molten silicon, even if the sample collection position is different, the analysis result does not change and a highly reliable estimated value can be obtained. However, if a jig for collecting a sample from high-temperature molten silicon is formed from a metal material that is not easily damaged, the molten silicon will be contaminated, so a metal jig is unsuitable. The jig is easily damaged by thermal stress when the silicon is solidified. Then, the inventor of the present application studied a method of using quartz or the like which does not cause contamination, and reached the invention described below.

The gist of the present invention is to collect a sample using a jig made of a material which does not contaminate the molten silicon remaining in the crucible after pulling up a single crystal silicon rod from the molten silicon in the crucible. Dissolving the entire amount of the sample solidified after collection to obtain a solution, quantitatively analyzing the impurity concentration in the solution, and analyzing the impurity in the single-crystal silicon rod from the quantitatively analyzed impurity concentration. Calculating the concentration.

[0012]

Embodiments of the present invention will be described below.

FIG. 1 is a schematic sectional view showing a pulling machine used in one embodiment of the present invention, in which 1 indicates a crucible. The crucible 1 receives heat from the heater 2 to keep the silicon in a molten state. While the single crystal seed held by the chuck 3 is slowly rotated by the shaft 4 connected to the rotation mechanism 5,
It is pulled up and grows a single crystal silicon rod.

The crucible 1 holds about 40 kilograms of molten silicon 6, and the pulling machine stops the growth of the single crystal silicon rod when the single crystal silicon rod reaches about 35 kilograms (solidification rate 90%), The single crystal silicon rod is moved to the cooling zone 7.

Next, the shutter 8 is closed, and a sample is collected using a jig for collecting the sample. The jig 9 has a detachable tip portion 9a made of high-purity quartz.
As shown in FIG. 2, it is moved in the direction of arrow A, and a required amount (5 to 30 g) of the molten silicon residue is scooped off at the tip 9a made of high-purity quartz. The jig 9 from which the molten silicon has been scooped moves in the direction opposite to the direction of the arrow A, and reaches the tray 10 made of high-purity quartz. After reaching the receiving tray 10, the tip 9a is cooled together with the molten silicon, so that thermal stress is generated due to a difference in thermal expansion coefficient and the tip 9a is broken. However,
Since the damaged tip portion 9a and the sample only scatter on the receiving tray 10, they can be completely collected.

The tray 10 moves in the direction of arrow B, and the fragments at the tip 9a and the sample are collected in a Teflon beaker. Since a considerable portion of the sample is attached to the debris, hydrofluoric acid is placed in a Teflon beaker to completely dissolve and remove the quartz. After washing the remaining sample with water, weigh the sample,
Dissolve with a predetermined amount of nitric hydrofluoric acid. As mentioned above, samples
Since it is 30 g, dissolution is completed in a short time.

Next, the concentration of trace metal impurities in the solution was determined by ICP-MS (Inductively Coupled).
ed Plasma Mass Spectromet
ry) or quantitative analysis by an analytical method such as flameless atomic absorption to determine the impurity concentration. When the impurity concentration is thus determined, the impurity concentration Cos in the single crystal silicon top portion is quantified by (Equation 1). Note that Cl represents the impurity concentration in the molten silicon remaining in the crucible, g represents the solidification rate when Cl was obtained, and k represents the segregation coefficient of the impurities. (Equation 1) assumes that impurities are contained in the initial state.

[0018]

## EQU1 ## Cos = kCl / (1-g) ^k-1

Assuming that the impurities in the molten silicon remaining in the crucible were inherited from the molten silicon before the single crystal silicon rod was pulled up, the impurities contained in the top portion (solidification rate 0) of the single crystal silicon rod can be calculated from the above equation. A quantitative value Cos of the concentration is obtained. FIG. 3 shows quantitative values Cos (indicated by small circles) of Fe, Cr, and Ni measured by the method according to the present embodiment, and quantitative values (indicated by small squares) measured from a sample having a solidification rate of 99% or more in Conventional Example 2. ).
As for the quantitative value Cos measured in this example, a sample obtained from a residue having a solidification rate of 90% was analyzed by an ICP-MS method. The horizontal axis shows Fe, Cr, and Ni, respectively, and the vertical axis shows quantitative values. The deflection coefficients of Fe, Cr and Ni are each 8
The calculation was performed as × 10 ⁻⁶ , 3 × 10 ⁻⁵ , and 3 × 10 ⁻⁵ . As is evident from FIG. 3, the two agree well, and it can be understood that accurate quantitative values can be obtained in this example. In this embodiment, the sample is collected in a molten state and slightly (5-10
In contrast to g), a sample solution for analysis can be prepared, whereas in Conventional Example 2, several hours are required to improve the solidification rate, and further, the entire amount of solidified sample (100 to 200 g) is dissolved.
It took several more hours. Therefore, in the present embodiment, the time required for the analysis can be reduced by about 10 hours as compared with the conventional example.

[0020]

As described above, according to the present invention, a sample is directly collected from the molten silicon residue, and the impurity concentration in the single-crystal silicon rod is determined based on the analysis result. This has the effect that it can be obtained in a short time with a slight increase in cost.

Further, since the pulling of the single-crystal silicon rod is completed immediately after the growth of the product portion,
The appearance of the tail is not impaired.

[Brief description of the drawings]

FIG. 1 is a schematic sectional view of a lifting machine used in one embodiment of the present invention.

FIG. 2 is a sectional view showing a sample collecting jig according to an embodiment of the present invention.

FIG. 3 is a graph showing quantitative values of impurity concentrations in a sample.

FIG. 4 is a photograph showing a part of the crystal structure of silicon solidified in a crucible.

FIG. 5 is a photograph showing a crystal structure of another portion of silicon solidified in a crucible.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Crucible 2 Heater 3 Seed chuck 6 Molten silicon 7 Cooling zone 8 Shutter 9 Sample collection jig 9a Tip part 10 Receiving tray

──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Hisashi Furuya 1-297 Kitabukuro-cho, Omiya-shi, Saitama Prefecture Central Research Laboratory, Mitsubishi Materials Corporation (72) Inventor Junichi Matsubara 1-297 Kitabukuro-cho, Omiya-shi, Saitama Hiroyuki Kameda, Examiner, Central Research Laboratory, Mitsubishi Materials Corporation (58) Field surveyed (Int.Cl. ⁶ , DB name) G01N 33/00 G01N 21/31 JICST file (JOIS)

Claims (1)

(57) [Claims] 1. A step of collecting a sample of molten silicon remaining in a crucible after pulling up a rod of single-crystal silicon from the molten silicon in the crucible using a jig made of a material that does not contaminate the molten silicon; Dissolving the entire amount of the post-solidified sample to obtain a solution, quantitatively analyzing the impurity concentration in the solution, and calculating the impurity concentration in the single-crystal silicon rod from the quantitatively analyzed impurity concentration. And a method for quantifying the impurity concentration of a single-crystal silicon rod.