JP2020007163A - Method of measuring resistivities of raw material crystal and method of manufacturing fz silicon single crystal - Google Patents

Abstract

To provide a measurement method capable of appropriately measuring resistivities and conductivity types of a raw material crystal and a method of manufacturing an FZ silicon single crystal using a raw material resistivity determined by the measurement method.SOLUTION: Provided is a measurement method of measuring resistivities of an oxygen-containing raw material crystal for use in manufacturing a silicon single crystal by an FZ method, the measurement method comprising: measuring a resistivity and a conductivity type at a plurality of circumferential locations of a sample wafer taken from a raw material crystal; determining whether the measured resistivities as appropriate or non-appropriate on the basis of deviation rates between the maximum and minimum values of the measured resistivities; determining whether the measured conductivity types as appropriate or non-appropriate on the basis of the measured conductivity types; when the measured resistivities are determined as appropriate and the measured conductivity types are determined as appropriate, adopting an average value of the measured resistivities as a raw material resistivity and the measured conductivity types as a raw material conductivity type, and when the measured resistivities are determined as non-appropriate and/or the measured conductivity types are determined as non-appropriate, failing the measurement and performing measurement again.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a resistance of a raw material crystal used for manufacturing a silicon crystal by the FZ method (floating zone method or floating zone melting method), particularly, a raw material crystal manufactured by a Czochralski (CZ) method. It relates to a method for measuring the rate.

  The FZ method is used, for example, as one of the methods for manufacturing a semiconductor single crystal such as a silicon single crystal most frequently used as a semiconductor device.

  Conventionally, n-type or p-type impurity doping is required to give a desired resistivity to a silicon single crystal. In the FZ method, a gas doping method in which a dopant gas is blown into a melting zone is known (see Non-Patent Document 1).

As the dopant gas, for example, PH ₃ or the like is used for doping P (phosphorus) as an n-type dopant, and B ₂ H ₆ or the like is used for doping B (boron) as a p-type dopant. The resistivity of a silicon single crystal varies depending on the concentration difference between these n-type dopants and p-type dopants in the crystal. However, when doping only n-type dopants or only p-type dopants in normal crystal production, The rate decreases as the dopant loading increases.

  In order to obtain a silicon single crystal having a desired resistivity, it is necessary to appropriately maintain the dopant supply amount calculated based on the resistivity of the raw material crystal and the target resistivity. By growing the single crystal by the FZ method while adjusting the concentration and the flow rate of the supplied dopant gas and maintaining the proper amount of the dopant, an FZ silicon single crystal having a target resistivity can be manufactured.

  As described above, in order to manufacture an FZ silicon single crystal having a target resistivity, it is of course necessary to supply a dopant gas having a calculated and set concentration and a flow rate. Must be appropriate. For setting the proper doping conditions, the resistivity of the raw material crystal is a very important factor. If the resistivity value and / or conductivity type of the raw material crystal is different from the true value, the resistivity of the manufactured and obtained FZ silicon single crystal will be far from the target value, and the required characteristics will not be obtained and the loss will be lost. Leads to.

  A high-purity silicon polycrystal is used as a raw material crystal for producing the FZ silicon single crystal, but is produced by a CZ method as a raw material crystal for producing the FZ silicon single crystal for the purpose of including a predetermined amount of oxygen in the FZ silicon single crystal. In some cases, a CZ silicon crystal is used (for example, Patent Documents 1 and 2). In any case, the FZ silicon single crystal having the target resistivity can be manufactured by appropriately measuring and setting the resistivity and the conductivity type of the raw material crystal and using the same for the doping calculation.

  As described above, when a CZ silicon crystal is used as a raw material crystal for producing an FZ silicon single crystal, it is desirable that the raw material crystal has as high a resistivity as possible in order to produce an FZ silicon single crystal having various resistivity bands.

  Conventionally, a high-purity silicon polycrystalline rod manufactured by the Siemens method or the like is used as a raw material crystal, and the oxygen concentration of the manufactured FZ silicon single crystal is extremely low. On the other hand, in recent years, semiconductor device manufacturing methods and required characteristics have become more diversified, and there is a demand for FZ silicon single crystals having a certain oxygen concentration. As a countermeasure, a CZ silicon crystal is used as a raw material crystal because a method using a material having a high oxygen concentration is simpler than an oxygen doping method using various methods in the production of FZ silicon single crystals.

  The resistivity and conductivity type of the CZ silicon crystal of the raw material crystal are measured in advance, and the measured value is used as the resistivity and conductivity type of the raw material crystal. Since oxygen introduced at the time of production is turned into a thermal donor, an appropriate resistivity measurement value cannot be obtained. In particular, as described above, when used in the production of FZ silicon single crystal, the tendency is remarkable because of the high resistivity as described above.

  Conventionally, after the heat treatment (oxygen donor killer heat treatment) is performed to eliminate the oxygen donor, the resistivity and the conductivity type are measured. Since the variation in resistivity in the crystal cross section of the CZ silicon crystal is relatively small, for example, The center in the sample plane was measured and used as a representative value of the resistivity.

  However, when the raw material CZ silicon crystal has a high oxygen concentration, even if the oxygen donor killer heat treatment as described above is performed, the oxygen donor cannot be completely removed, and the raw material CZ silicon crystal has a high resistivity. This will affect the resistivity and conductivity type measurements. This means that the resistivity and the conductivity type of the raw material crystal are not measured properly, and when the FZ silicon single crystal is manufactured, the dopant supply amount is determined based on these values, and even if the dopant is actually supplied, The resistivity of the FZ silicon single crystal thus obtained does not reach a target value, and the FZ silicon single crystal has a problem that it is wasted because characteristics required for a semiconductor device to be manufactured cannot be obtained.

JP 2005306653 A JP-A-2015-160800

"Floating-Zone Silicon" by WOLFGANG KELLER, ALFRED MUHLBAUER, p. 15-, MARCEL DEKKER, INC. Issue

  The present invention has been made in order to solve the above-mentioned problem, and it has been made possible to appropriately measure the resistivity and conductivity of a raw material crystal used for manufacturing a silicon single crystal by the FZ method. It is an object of the present invention to provide a method for measuring FZ, and a method for producing an FZ silicon single crystal using the raw material resistivity of the raw crystal obtained by the measuring method.

In order to achieve the above object, the present invention provides
A method for measuring the resistivity of an oxygen-containing raw material crystal used for producing a silicon single crystal by the FZ method,
(A) collecting a sample wafer from a raw material crystal,
(B) measuring the measured resistivity and the measured conductivity at a plurality of locations in the circumferential direction of the sample wafer;
(C) calculating the divergence rate between the maximum value and the minimum value among the measured resistivity values at the plurality of locations obtained in the step (b), and if the divergence rate does not exceed a predetermined threshold, the measured resistivity Appropriate, if it exceeds the threshold value, the process of determining that the measured resistivity is inappropriate,
(D) determining that the measured conductivity type is appropriate if the measured conductivity types at the plurality of locations obtained in the step (b) are all the same, and that the measured conductivity type is inappropriate if not all the same.
(E) When it is determined in the steps (c) and (d) that the measured resistivity is appropriate and that the measured conductivity type is appropriate, the measured resistivity at the plurality of locations obtained in the step (b) Is used as the raw material resistivity of the raw material crystal, the measured conductivity type is adopted as the raw material conductivity type of the raw material crystal, and the measured resistivity is not measured in the steps (c) and (d). If it is determined that the measurement and / or measurement conductivity type is inappropriate, the measurement resistivity at the plurality of locations and / or the measurement conductivity type at the plurality of locations obtained in the step (b) are rejected, and the measurement is repeated again at the plurality of locations. Measuring the measured resistivity and / or the measured conductivity of step (c) and / or starting over from step (d);
And a method for measuring the resistivity of the raw material crystal.

  According to such a measurement method of the present invention, since the resistivity and conductivity type of the raw material crystal can be properly measured, the production of the FZ silicon single crystal having the target resistivity is facilitated, and the obtained The FZ silicon single crystal can be used as a material having appropriate quality for manufacturing a target semiconductor device.

  Also, at this time, the divergence rate in the step (c) is calculated by [deviation rate] = ([maximum value of measured resistivity] − [minimum value of measured resistivity]) ÷ [minimum value of measured resistivity]. It is preferable to calculate and set the threshold value of the deviation rate to 20%.

  By calculating the deviation rate in this way, it is possible to accurately detect a case where the measurement result of the measured resistivity is inappropriate for the measured resistivity.

  Preferably, the measurement of the measured resistivity in the step (b) is performed by a four-probe method, and the measurement of the measured conductivity type in the step (b) is performed by a thermoelectromotive force method.

  If the measurement resistivity and the measurement conductivity type are measured by such a method, the measurement of the measurement resistivity with high accuracy can be performed relatively easily, and the measurement conductivity type measurement can be easily performed.

  Further, it is preferable that the raw material crystal is a crystal produced by a CZ method.

  With the raw material crystal manufactured by the CZ method, a raw material crystal having a high oxygen concentration can be obtained relatively easily, and the resistivity and conductivity type of the raw material can be accurately measured by the present invention. .

Further, it is preferable that the oxygen concentration of the raw material crystal is 6.5 × 10 ¹⁷ atoms / cm ³ or more.

According to the present invention, a resistivity close to a true value can be measured even when the oxygen concentration of the raw material crystal is 6.5 × 10 ¹⁷ atoms / cm ³ or more.

  Further, it is preferable that the resistivity of the raw material crystal is 1,000 Ωcm or more.

  According to the present invention, the resistivity of the raw material crystal can be properly measured even when the raw material crystal has a high resistivity of 1,000 Ωcm or more.

  In addition, the positions where the measured resistivity and the measured conductivity are measured at the plurality of locations in the step (b) are located at least r / 2 or more away from the center of the sample wafer, and the distance from the center is the same. Is preferred.

  As described above, when the plurality of measurement points are located at positions r / 2 or more from the center of the sample wafer and the distance from the center is the same, the measurement resistivity and the measurement conductivity type of the plurality of measurement points are relatively stable. Value.

  Further, based on the raw material resistivity of the raw material crystal obtained by the above-described measuring method and the target target resistivity of the manufactured FZ silicon single crystal, the amount of dopant to be introduced at the time of manufacturing the FZ silicon single crystal is calculated. A method for producing an FZ silicon single crystal, which comprises producing a silicon single crystal by the FZ method while adding the calculated dopant amount, is also provided.

  With such an FZ silicon single crystal manufacturing method, an FZ silicon single crystal close to the target resistivity can be easily manufactured.

  As described above, according to the method for measuring the resistivity of a raw material crystal of the present invention, when manufacturing a silicon single crystal using the FZ method, the resistivity and conductivity of the oxygen-containing raw material crystal are properly adjusted. It can be measured, and when applied to the method of manufacturing an FZ silicon single crystal of the present invention, the manufacture of an FZ silicon single crystal having a target resistivity value becomes easy, and the obtained FZ silicon single crystal is used as a target semiconductor. It can be a material of suitable quality for device manufacture.

  This means that the loss rate of the produced crystals is reduced, and there is a great advantage in commercial production, particularly, because it leads to a stable supply of products.

It is a schematic diagram showing an example of a manufacturing process of FZ silicon single crystal of the present invention. It is a schematic diagram showing an example of a manufacturing device of FZ silicon single crystal which can be used in the present invention. It is a figure showing the comparison result of resistivity and electric conductivity of sample A and B. It is a flowchart which shows each process of the measuring method of the resistivity of the raw material crystal of this invention.

  Hereinafter, the present invention will be described in detail with reference to the drawings as an example of an embodiment, but the present invention is not limited to this.

  As described above, when the FZ silicon single crystal is manufactured using the CZ silicon crystal as the raw material crystal, the resistivity of the manufactured silicon single crystal varies between the crystals regardless of the dopant supply amount set in advance. In some cases, the situation was significantly different and was not stable. If the resistivity of the silicon single crystal manufactured in this way deviates from a target value and deviates from a specified range, the single crystal cannot be used for a target semiconductor device, resulting in a loss. is there. Therefore, even when the silicon crystal of the raw material has a high oxygen concentration, the resistance and the conductivity type have little influence on the measurement result, and the resistance and the conductivity of the raw material crystal can be properly measured. Development of a method for measuring the rate was required.

  The present inventors have conducted intensive studies on such a problem, and found that the main factor is that the resistivity of the raw material crystal deviates from the true value, and found that the peripheral factor of the sample wafer of the raw material crystal was large. If the variation of the measured resistivity measured at a plurality of points in the direction is equal to or less than a certain value, and if the measured conductivity types are all the same, the reliability of the measurement result of the measured resistivity and the measured conductivity type is high. It has been determined that the resistivity of the raw material crystal closer to the true value can be obtained, and the present invention has been completed.

That is, the present invention is a method for measuring the resistivity of a material crystal containing oxygen used when producing a silicon single crystal by the FZ method,
(A) collecting a sample wafer from a raw material crystal,
(B) measuring the measured resistivity and the measured conductivity at a plurality of locations in the circumferential direction of the sample wafer;
(C) calculating the divergence rate between the maximum value and the minimum value among the measured resistivity values at the plurality of locations obtained in the step (b), and if the divergence rate does not exceed a predetermined threshold, the measured resistivity Appropriate, if it exceeds the threshold value, the process of determining that the measured resistivity is inappropriate,
(D) determining that the measured conductivity type is appropriate if the measured conductivity types at the plurality of locations obtained in the step (b) are all the same, and that the measured conductivity type is inappropriate if not all the same.
(E) When it is determined in the steps (c) and (d) that the measured resistivity is appropriate and that the measured conductivity type is appropriate, the measured resistivity at the plurality of locations obtained in the step (b) Is used as the raw material resistivity of the raw material crystal, the measured conductivity type is adopted as the raw material conductivity type of the raw material crystal, and the measured resistivity is not measured in the steps (c) and (d). If it is determined that the measurement and / or measurement conductivity type is inappropriate, the measurement resistivity at the plurality of locations and / or the measurement conductivity type at the plurality of locations obtained in the step (b) are rejected, and the measurement is repeated again at the plurality of locations. Measuring the measured resistivity and / or the measured conductivity of step (c) and / or starting over from step (d);
And a method for measuring the resistivity of the raw material crystal.

The present invention is a method for measuring the resistivity of an oxygen-containing raw material crystal used when producing a silicon single crystal by the FZ method, and comprises the following steps (a) to (e).
Hereinafter, description will be made with reference to FIG.

[Step (a)]
The step (a) is a step of collecting a sample wafer from a raw material crystal (FIG. 4A). The sample wafer can be collected by a conventional method. For example, as a simple and practical method, a disk-shaped sample can be collected from a raw material crystal. A heat treatment for removing the influence of the oxygen donor (oxygen donor killer heat treatment) can be further performed.

  Further, the raw material crystal is preferably a crystal manufactured by the CZ method. As described above, a raw material crystal manufactured by the CZ method can relatively easily obtain a raw material crystal having a high oxygen concentration.

  Further, since it is preferable that the raw material for producing the FZ silicon single crystal is as long as possible, the sampling is preferably performed near the end of the CZ silicon crystal.

Further, it is preferable that the oxygen concentration of the source crystal be 6.5 × 10 ¹⁷ atoms / cm ³ or more. According to the present invention, a resistivity close to a true value can be measured even when the oxygen concentration of the raw material crystal is 6.5 × 10 ¹⁷ atoms / cm ³ or more.

  Further, the resistivity of the raw material crystal is preferably 1,000 Ωcm or more. According to the present invention, the resistivity of a raw material crystal can be properly measured even with a raw material crystal having a high resistivity of 1,000 Ωcm or more. Since the resistivity of a high-resistivity product varies greatly even with a slight difference in dopant concentration, it is effective to apply the present invention.

[Step (b)]
The step (b) is a step of measuring the measured resistivity and the measured conductivity at a plurality of locations in the circumferential direction of the sample wafer (FIG. 4B). The locations where the resistivity (measurement resistivity) and conductivity type (measurement conductivity type) are measured at multiple locations in the circumferential direction of the sample wafer must be within a concentric range that can be regarded as equidistant from the center of the sample disk surface It is preferable that the distance be at least a central point (r / 2) or more within a radius from the center to the outer periphery in the plane of the sample disk, but the position is not particularly limited as long as it is a plurality of points in the circumferential direction of the sample wafer. For example, it may be set to two points that are axisymmetric with respect to the center of the sample wafer, or may be set to a position that is closer to the center and that is entirely within the semicircle of the sample wafer. If a plurality of measurement points are located at a distance of r / 2 or more from the center of the sample wafer and the distance from the center is the same, the measurement resistivity and the measurement conductivity type of the plurality of measurement points are relatively stable. Become.

  The measurement of the measured resistivity can be relatively easily performed by the four probe method, and the measurement of the measurement conductivity type can be easily performed by the thermoelectromotive force method. Depending on the measurement sample, an appropriate resistivity measurement value may not be obtained because oxygen introduced at the time of manufacturing the CZ silicon crystal is converted into a thermal donor in this state. It is known that this oxygen donor is erased by, for example, a slight heat treatment at 650 ° C. for about 20 minutes. Such a heat treatment is performed as needed, and after the oxygen donor is erased, the resistivity and conductivity are erased. Can also be measured.

[Step (c)]
In the step (c), the maximum value and the minimum value of the measured resistivity measured in the step (b) are specified, the divergence rate between the maximum value and the minimum value is determined, and the divergence rate is set to a predetermined threshold value. If it is smaller, the measurement is appropriate, and if it is larger, the measured value is inappropriate.

  The deviation rate between the maximum value and the minimum value of the measured resistivity is calculated by [deviation rate] = ([maximum value of the measured resistivity] − [minimum value of the measured resistivity]) ÷ [minimum value of the measured resistivity]. be able to. The threshold value of the deviation rate can be set to 20% or less, preferably 10% or less. By calculating the deviation rate in this way, it is possible to accurately detect a case where the measurement result of the measured resistivity is inappropriate for the measured resistivity.

  In this way, it is determined whether the measured resistivity is properly measured.

[Step (d)]
In the step (d), the respective measurement conductivity types measured in the step (b) are checked. If all the measurement results are the same, the measurement is appropriate for the measurement conductivity type. This is a step of determining that the mold is inappropriate.

[Step (e)]
In the step (e), based on the results of the steps (c) and (d), if it is determined that the measured resistivity is appropriate, and if it is determined that the measured conductivity type is appropriate, the measurement is performed in the step (b). The average value of each measured resistivity is calculated, the calculated average value is defined as the raw material resistivity of the raw material crystal, and the measured conductivity type similarly measured is defined as the raw material conductivity type of the raw material crystal. If it is determined in the steps (c) and (d) that the measured resistivity is inappropriate and / or the measured conductivity is inappropriate, the resistivity and / or conductivity measured in the step (b) are determined. Is determined to be inappropriate and is not adopted as the characteristic value of the raw material crystal. The measured resistivity and / or measured conductivity type are measured again at a plurality of locations, and the process is repeated from the step (c) and / or the step (d). It is a process. At this time, the step to be repeated when one of the steps (c) and (d) is determined to be inappropriate may include a step determined to be appropriate in addition to the step determined to be inappropriate. .

  By performing the above steps (c) to (e), it is accurately determined that the measured resistivity and the measured conductivity obtained by measuring the sample wafer of the raw material crystal are appropriate. That is, a resistivity close to the true value is obtained.

  According to such a measurement method of the present invention, since the resistivity and conductivity type of the raw material crystal can be properly measured, the production of the FZ silicon single crystal having the target resistivity is facilitated, and the obtained The FZ silicon single crystal can be used as a material having appropriate quality for manufacturing a target semiconductor device.

[Method of Manufacturing FZ Silicon Single Crystal]
The present invention further provides a method for producing an FZ silicon single crystal based on the material resistivity of the material crystal obtained by the method for measuring the resistivity of the material crystal of the present invention and the target target resistivity of the FZ silicon single crystal to be produced. The present invention provides a method for producing an FZ silicon single crystal, which comprises calculating a supply amount of a dopant to be introduced sometimes, and producing a silicon single crystal by the FZ method while adding the calculated amount of the dopant. With such an FZ silicon single crystal manufacturing method, an FZ silicon single crystal close to the target resistivity can be easily manufactured.
FIG. 1 shows an example of the manufacturing process of the FZ silicon single crystal of the present invention.

  A semiconductor rod (raw material rod) 14 serving as a raw material whose resistivity is measured by the method of the present invention is placed above a high-frequency coil 16 to which a high-frequency induced current is applied, and a single crystal seed crystal 15 is placed below. The lower end of the raw material rod 14 is melted and fused to the seed crystal 15 (seeding step (a)). Further, drawing (necking) for removing dislocations generated in the crystal during this seeding is performed (( b) necking step), and then growing the crystallization-side semiconductor rod (semiconductor single crystal rod) while expanding it to a desired diameter ((c) cone part forming step). Further, while controlling the crystallization-side semiconductor rod 19 to a desired diameter and supplying a dopant gas so as to have a target resistivity, growth is performed ((d) straight body portion forming step), and the supply of the raw material is performed. Is stopped, the diameter of the crystallization-side semiconductor rod 19 is reduced, and the crystallization-side semiconductor rod is separated from the source semiconductor rod ((e) separation step). Through the above steps, a semiconductor crystal (FZ silicon single crystal) can be manufactured.

  FIG. 2 shows an example of an apparatus for producing an FZ silicon single crystal that can be used in the present invention.

  An upper shaft 12 and a lower shaft 13 are provided in a chamber 11 of the FZ silicon single crystal manufacturing apparatus 1. A semiconductor rod having a predetermined diameter is attached to the upper shaft 12 as a raw material semiconductor rod 14, and a seed crystal 15 is attached to the lower shaft 13. Further, a high-frequency coil 16 for melting the raw material semiconductor rod 14 is provided, and a silicon single crystal (crystallization-side semiconductor rod) 19 can be grown while moving the melting zone 18 relatively to the raw material semiconductor rod 14. During the growth, the dopant gas can be supplied from the dopant gas doping nozzle (dopant gas supply means) 20. The dopant gas is determined based on the raw material resistivity of the raw material crystal 14 obtained by the measuring method of the present invention and the target target resistivity of the manufactured FZ silicon single crystal. Calculated and supplied based on this result. The downward arrow in the drawing indicates the direction of the crystal movement.

  First, for example, a silicon polycrystal rod having a predetermined diameter is attached to the upper shaft 12 as the raw material semiconductor rod 14 whose resistivity is measured by the resistivity measuring method of the present invention, and a seed crystal 15 is attached to the lower shaft 13. After the raw semiconductor rod 14 is melted by the high-frequency coil 16, it is fused to the seed crystal 15. The crystallization-side semiconductor rod 19 grown from the seed crystal 15 is dislocation-free by the aperture 17, and is lowered relative to the high-frequency coil 16 while rotating both shafts. The crystallization-side semiconductor rod 19 is grown while moving upward.

  After forming the aperture 17, the crystallization side semiconductor rod 19 grown from the seed crystal 15 is grown while expanding to a desired diameter to form a cone portion, and the raw material semiconductor rod 14 and the crystallization side semiconductor rod 19 are formed. The crystallization side semiconductor rod 19 is grown while controlling to a desired diameter to form a straight body portion.

  Then, the melting zone 18 is moved to the upper end of the material semiconductor rod 14 to complete the growth of the silicon single crystal 19, the diameter of the crystallization-side semiconductor rod 19 is reduced, and the crystallization-side semiconductor rod 19 is removed from the material semiconductor rod 19. Then, a semiconductor crystal is manufactured by separating the semiconductor crystal from the semiconductor substrate 14.

  Since the resistivity required for FZ silicon single crystal ranges from less than 1 Ωcm to several thousands to 10,000 Ωcm or more, the method of adjusting the resistivity of FZ single crystal by adding a dopant during the production of the single crystal is used. The raw materials used must have versatility. For this reason, it is desirable that the resistivity of the raw material crystal be as high as possible, and it is more preferable that the amount of dopant contained in the raw material crystal be as small as possible.

  In order to meet the demand for FZ silicon single crystal having a certain oxygen concentration, it is better to adopt a method using raw materials having a high oxygen concentration from the beginning than to adopt an oxygen doping method in the production of FZ silicon single crystal. It is more convenient. In this case, it is preferable to use a CZ silicon crystal as a raw material crystal.

As for the CZ silicon crystal, the desirable characteristics as a raw material for producing an FZ silicon single crystal are the same as those described above. Therefore, it is general to produce a CZ silicon crystal having a high resistivity without adding an additional dopant and use it as a raw material crystal. is there. Here, one of the reasons for using the CZ silicon crystal as a raw material crystal is to obtain an FZ silicon single crystal having an oxygen concentration within a predetermined range. In this case, it is desirable that the oxygen concentration of the CZ silicon crystal of the raw material be 50 times or more the desired oxygen concentration of the FZ silicon single crystal (for example, Patent Document 2). The oxygen concentration can be high, for example, 6.5 × 10 ¹⁷ atoms / cm ³ or more.

  As described above, the resistivity of the FZ silicon single crystal is adjusted by adding a dopant with a dopant gas during single crystal growth. The necessary amount of the dopant to be added is calculated from factors such as the target resistivity and conductivity type of the FZ silicon single crystal, the resistivity and conductivity type of the source crystal, and the crystal growth conditions. Determine the concentration and supply amount.

  As described above, in order to manufacture an FZ silicon single crystal having a target resistivity, it is important to reliably supply a dopant gas having a calculated and set concentration and flow rate.

  However, even though the supply amount of the dopant gas is confirmed to be a predetermined value, the resistivity of the manufactured single crystal may not be as set. One of the reasons is that the calculation of the dopant addition amount itself is inappropriate. Among the calculation factors of the amount of dopant addition, the resistivity and conductivity type of the FZ silicon single crystal are as predetermined, and if the factors such as the crystal growth conditions are different from the actual condition, it should be detected. If the resistivity is not as set, the cause can be specified. On the other hand, since the resistivity and conductivity type of the raw material crystal cannot be confirmed after the production of the FZ silicon single crystal even if the resistivity and conductivity type are different from the true values, it is determined that the resistivity and the conductivity type are not properly determined. It has been found that the resistivity of the crystal is largely different between the crystals, which is a cause of instability.

  When a CZ silicon crystal is used as a raw material crystal in the production of an FZ silicon single crystal, the resistivity and the conductivity type of the CZ silicon crystal must be properly grasped. As a method for this purpose, a disk-shaped sample is taken from a CZ silicon crystal as a raw material crystal, subjected to a heat treatment for removing the influence of an oxygen donor, and then, for example, a resistivity is measured by, for example, a four-point probe method. There is a method in which the conductivity type is measured by an electromotive force method, and the resistivity and conductivity type of the CZ silicon crystal are determined. Conventionally, at the measurement position, the CZ silicon crystal has a relatively small variation in resistivity in the cross section of the crystal, and thus, for example, is set at one point in the center of the sample plane.

  The inventor has found that the cause of the deviation of the resistivity of the manufactured silicon single crystal from the target value and deviating from the specified range is mainly that the resistivity of the raw material crystal deviates from the true value. For the purpose of confirming that the resistivity sample of the CZ silicon crystal as the raw material was properly measured, the resistivity and conductivity type in the sample surface were measured, and the dopant concentration of the sample was measured. The resistivity and the conductivity type were calculated from the dopant concentration measurement results, and the measured values and the calculated values were compared and examined. Hereinafter, this examination will be described.

The resistivity is measured in the sample diameter direction by the four-point probe method, and the conductivity type is measured by the thermoelectromotive force method at the center of the sample radius range, r / 2 portion (a portion on the outer peripheral side by 中心 of the sample radius r from the center). , And the outer peripheral portion (at a position 10 mm from the outer periphery). The dopant concentration was measured at the center, r / 2 part, and outer periphery of the sample radius range by photoluminescence measurement.
FIG. 3 shows the results of comparison of the resistivity of the two examples of samples A and B, and Table 1 shows the results of comparison of the conductivity type.

  FIG. 3A shows the result of the resistivity comparison of Sample A. According to the measured values, the resistivity distribution has a shape with a lower peripheral portion and a higher central portion. In the conductivity type measurement results of Sample A shown in Table 1, the entire surface was determined to be p-type. On the other hand, in the calculated values, although the resistivity distribution slightly fluctuates, the values are substantially the same at three points in the plane. In the conductivity type measurement results shown in Table 1, the entire surface was determined to be p-type, similarly to the measured values.

  FIG. 3B shows the result of the resistivity comparison of Sample B. According to the measured values, the resistivity distribution has a shape in which the outer peripheral portion is low, has a local maximum value near the r / 2 part, and becomes lower toward the central portion. In the conductivity type measurement results of Sample B shown in Table 1, the outer peripheral portion was determined to be p-type and the central portion was determined to be n-type, and a clear determination could not be made in the vicinity of r / 2. This is because, when a straight line from the outer periphery to the center is traced, the resistivity starts from a low resistivity of the p-type and increases steadily, and changes from the p-type to the n-type near the r / 2 portion, and the resistivity further increases. Is changing from high to low. On the other hand, in the calculated values, the resistivity distribution still has a slight variation, but has substantially the same value at three points in the plane. The conductivity type measurement results shown in Table 1 were different from the measured values, and the entire surface was determined to be p-type.

  In both cases of the samples A and B, the measured value and the calculated value of the resistivity substantially coincide with each other in the vicinity of the outer peripheral portion of the sample, and the measured value and the calculated value of the conductivity type also coincide. When this moves to the r / 2 part, the central part, and the position inside the sample, the resistivity shows a difference between the measured value and the calculated value, and the conductivity type shows a case where the measured value differs from the calculated value.

  This can be rephrased that, at the inside position of the sample, the measured value tends to show the n-type compared to the calculated value. Here, the calculated value is a numerical value based on the difference between the dopant concentration measurement result, that is, the difference between the n-type impurity concentration such as phosphorus or arsenic and the p-type impurity concentration such as boron or aluminum, and conforms to the original resistivity value. It is considered something. The fact that the n-type tendency is stronger than the calculated value means that the donor amount is increased due to an effect other than the dopant, and it is considered that the influence of the thermal donor, in this case, the oxygen donor is apparent.

  From this viewpoint, even if the resistivity and conductivity are measured at the center of the sample as in the prior art, an appropriate value cannot be obtained. Can be measured, it can be said that the measurement position is preferably on the outer peripheral side of the in-plane r / 2.

  Here, as described above, the raw material crystal used for manufacturing the FZ silicon single crystal has a very high resistivity, and the frequency of the measured resistivity value changes more frequently than in the case where a low resistivity crystal is measured. For example, in the measurement by the four probe method, the degree of contact of the probe changes due to a slight difference in the state of the sample surface, and the measured resistivity value greatly changes. That is, there are cases where the measurement is not appropriate. Alternatively, high resistivity means that it is close to the p-type and n-type change points, and the conductivity type may also change.

  Therefore, it is conceivable that an inappropriate measured value may be adopted as the resistivity or conductivity type of the raw material crystal, and the manufactured FZ silicon single crystal may be a factor deviating from the target resistivity. . However, as in the prior art, for example, in the measurement of only one location in the center of the sample, it is not possible to detect whether or not the measured value is incorrect even if the measured value is incorrect.

  Here, from the principle of the method of manufacturing the raw material crystal by the CZ method, the resistivity distribution in the cross section of the raw material crystal has a symmetric shape with the center as an axis. That is, since the resistivity is the same within a concentric range that can be regarded as equidistant from the center of the source crystal cross section, if the resistivity measurement is performed at a plurality of positions within this range, the obtained resistivity values will be the same. Should be. Similarly, the conductivity type should be essentially the same within this range. Utilizing this phenomenon, if the measured resistivity is not equal or the measured conductivity type is not the same, it can be determined that the measurement is inappropriate.

  Therefore, the present inventor collected sample wafers for 1,000 raw material crystals having a size of 1,000 Ωcm or more, and performed resistivity measurement at two locations at a position 10 mm from the outer periphery of each sample. At this time, when the divergence rate between the maximum value and the minimum value of the measured resistivity was confirmed, it was found that the average value was 3.8%, and the average value + 3σ was 21.5% (σ is variation in resistivity). . When the resistivity was re-measured for 22 samples having a divergence rate of more than 20% among the samples, all the divergence rates were less than 10%. Further, the resistivity was re-measured for nine samples having a deviation rate of more than 15% and less than 20%, and all of them also had a deviation rate of less than 10%. On the other hand, 20 samples having a divergence rate of less than 10% were selected and re-measured, and the divergence rate hardly changed.

  From this, as described above, the resistivity measurement at a plurality of positions within the concentric range that can be regarded as equidistant from the center of the source crystal cross section, if the deviation rate is determined by a threshold, whether the measurement is appropriate or not The threshold value of the divergence rate used for determining whether the resistivity measurement is appropriate or inappropriate at this time is at least 20%, preferably 10%.

  In order to obtain the resistivity of the CZ silicon crystal, which is a raw material for producing FZ silicon single crystals, it is more appropriate to measure the dopant concentration by photoluminescence measurement and calculate the resistivity from the results. In order to perform the above method on all the raw material rods, it is not realistic in terms of lead time, cost, and the like, and a simpler method is required.

  Therefore, as in the present invention, a method of measuring a plurality of positions at which the same resistivity and the same conductivity type are obtained, determining whether the measurement is appropriate or inappropriate, and re-measuring if the measurement is determined to be inappropriate However, it is extremely effective for setting a practical and appropriate resistivity value and conductivity type. Even if an incorrect measurement value is included, the number of measurement points is extremely increased and an average value is calculated, thereby effectively invalidating the incorrect resistivity measurement value without performing re-measurement. Although a method of obtaining an appropriate resistivity value or adopting a large number of conductive types is also conceivable, this takes a great deal of time. In the present invention, since the measurement position can be set to two points, it is more advantageous in terms of lead time, cost, and the like.

  Hereinafter, the present invention will be described more specifically with reference to Examples, but the present invention is not limited thereto. The oxygen concentration of the CZ silicon raw material crystals used in the following Examples and Comparative Examples was almost the same.

(Example 1)
A sample wafer is sampled from a CZ silicon raw material crystal (oxygen concentration is 6.5 × 10 ¹⁷ atoms / cm ³ or more) used as a raw material for producing FZ silicon single crystal, and is located 10 mm from the outer periphery of the sample wafer. For two different points, the resistivity (measured resistivity) was measured by the four probe method, and the conductivity type (measured conductivity type) was measured by the thermoelectromotive force method. Next, the threshold value of the divergence rate is set to 20%, the divergence rate of the obtained two measured resistivity values is obtained. When the deviation rate exceeded 20%, the same resistivity measurement was performed again. At this time, the divergence rate was calculated as [deviation rate] = ([maximum value of measured resistivity] − [minimum value of measured resistivity]) ÷ [minimum value of measured resistivity]. Similarly, when the two measured conductivity types match, the conductivity type of the raw material crystal was adopted, and when the two conductivity types did not match, the same conductivity type measurement was performed again.

  The required number of CZ silicon source crystals whose resistivity and conductivity type were determined in this manner were prepared. The determined raw material resistivity was 1,000 Ωcm or more in each case.

  In order to produce an n-type 50 Ωcm (target resistivity) FZ silicon single crystal, the amount of dopant added was calculated using the above-mentioned raw material resistivity, and a single crystal was produced while supplying the dopant as it was. An FZ silicon single crystal was obtained. The variation (σ) in the resistivity of each of the obtained FZ silicon single crystals with respect to the target resistivity was within 1.4%.

(Example 2)
In order to manufacture a p-type FZ silicon single crystal of 3,500 Ωcm (target resistivity), 20 FZ silicon single crystals were formed using the same procedure as in Example 1 except that the threshold value of the deviation rate was set to 10%. I got it. The variation (σ) in the resistivity of each of the obtained FZ silicon single crystals with respect to the target resistivity was 8.4%.

(Comparative Example 1)
The resistivity of the in-plane central portion of the sample wafer sampled from the CZ silicon crystal used as a raw material for manufacturing the FZ silicon single crystal was measured by the four-point probe method, and the conductivity type was measured by the thermoelectromotive force method. The other prepared CZ silicon crystals were measured in the same manner, and the raw material resistivity was determined.

  In order to produce an n-type 50 Ωcm (target resistivity) FZ silicon single crystal, the amount of dopant added is calculated using the raw material resistivity measured above, and the single crystal is produced while supplying the dopant as it is. And 30 FZ silicon single crystals were obtained. At this time, the resistivity variation (σ) between the single crystals was 4.0%.

(Comparative Example 2)
In order to produce a p-type FZ silicon single crystal of 3,500 Ωcm (target resistivity), 20 FZ silicon single crystals were obtained by the same procedure as in Comparative Example 1. At this time, the resistivity variation (σ) between the single crystals was 18.9%.

As is clear from the above results, according to the present invention, even when a raw material crystal containing oxygen is used in the production of a silicon single crystal using the FZ method, the resistivity and the conductivity type are reduced. When the FZ silicon single crystal can be properly measured and is applied to the method for manufacturing an FZ silicon single crystal of the present invention, the production of the FZ silicon single crystal having a target resistivity value becomes easy, and the obtained FZ silicon single crystal is obtained. It can be a material having an appropriate quality for manufacturing a target semiconductor device. Then, even for a crystal in which the oxygen concentration of the raw material crystal is 6.5 × 10 ¹⁷ atoms / cm ³ or more, the resistivity close to the true value can be measured.

  Further, based on the raw material resistivity of the raw crystal and the target resistivity of the FZ silicon single crystal to be manufactured, the amount of dopant to be introduced at the time of manufacturing the FZ silicon single crystal is calculated, and the FZ silicon single crystal is manufactured based on the calculated amount. Therefore, it is possible to reduce the variation in the resistivity between the obtained FZ silicon single crystals regardless of the resistivity and the conductivity type of the FZ silicon single crystal to be manufactured.

  Note that the present invention is not limited to the above embodiment. The above embodiment is an exemplification, and the present invention has substantially the same configuration as the technical idea described in the claims of the present invention, and has the same effect. Within the technical scope of

DESCRIPTION OF SYMBOLS 1 ... Single crystal manufacturing apparatus, 11 ... Chamber, 12 ... Upper axis, 13 ... Lower axis,
14: raw material semiconductor rod, 15: seed crystal,
16 ... high frequency coil, 17 ... throttle, 18 ... melting zone,
19: silicon single crystal (crystallized semiconductor rod),
20 ... Dopant gas doping nozzle (dopant gas supply means).

Claims (8)

A method for measuring the resistivity of an oxygen-containing raw material crystal used for producing a silicon single crystal by the FZ method,
(A) collecting a sample wafer from a raw material crystal,
(B) measuring the measured resistivity and the measured conductivity at a plurality of locations in the circumferential direction of the sample wafer;
(C) calculating the divergence rate between the maximum value and the minimum value among the measured resistivity values at the plurality of locations obtained in the step (b), and if the divergence rate does not exceed a predetermined threshold, the measured resistivity Appropriate, if it exceeds the threshold value, the process of determining that the measured resistivity is inappropriate,
(D) determining that the measured conductivity type is appropriate if the measured conductivity types at the plurality of locations obtained in the step (b) are all the same, and that the measured conductivity type is inappropriate if not all the same.
(E) When it is determined in the steps (c) and (d) that the measured resistivity is appropriate and that the measured conductivity type is appropriate, the measured resistivity at the plurality of locations obtained in the step (b) Is used as the raw material resistivity of the raw material crystal, the measured conductivity type is adopted as the raw material conductivity type of the raw material crystal, and the measured resistivity is not measured in the steps (c) and (d). If it is determined that the measurement and / or measurement conductivity type is inappropriate, the measurement resistivity at the plurality of locations and / or the measurement conductivity type at the plurality of locations obtained in the step (b) are rejected, and the measurement is repeated again at the plurality of locations. Measuring the measured resistivity and / or the measured conductivity of step (c) and / or starting over from step (d);
And a method for measuring the resistivity of a raw material crystal.   The divergence rate in the step (c) is calculated as [deviation rate] = ([maximum value of measured resistivity] − [minimum value of measured resistivity]) ÷ [minimum value of measured resistivity]. The method for measuring the resistivity of a raw material crystal according to claim 1, wherein a threshold value of the rate is set to 20%.   The measurement of the measured resistivity in the step (b) is performed by a four-probe method, and the measurement of the measured conductivity type in the step (b) is performed by a thermoelectromotive force method. 3. The method for measuring the resistivity of a raw material crystal according to item 2.   The method for measuring the resistivity of a raw material crystal according to any one of claims 1 to 3, wherein the raw material crystal is a crystal produced by a CZ method. The method for measuring the resistivity of a raw material crystal according to any one of claims 1 to 4, wherein the oxygen concentration of the raw material crystal is 6.5 × 10 ¹⁷ atoms / cm ³ or more.   The method for measuring the resistivity of a raw material crystal according to any one of claims 1 to 5, wherein the resistivity of the raw material crystal is 1,000 Ωcm or more.   In the step (b), the positions where the measured resistivity and the measured conductivity are measured at the plurality of locations are located at least r / 2 or more away from the center of the sample wafer, and the distance from the center is the same. The method for measuring the resistivity of a raw material crystal according to any one of claims 1 to 6.   A method for measuring the resistivity of the raw material crystal according to any one of claims 1 to 7, based on the raw material resistivity of the raw crystal and a target target resistivity of the manufactured FZ silicon single crystal. A method of calculating the amount of dopant to be introduced at the time of manufacturing an FZ silicon single crystal, and manufacturing the silicon single crystal by the FZ method while adding the calculated amount of dopant.

Images