JPH03183686A - Doping method in pulling up single crystal - Google Patents

Abstract

PURPOSE:To stably and continuously obtain single crystal which has uniform resistivity in the lengthwise direction by comparting the inside of a crucible into two parts with a partition having a communication and supplying a raw material of crystal and a dopant to one part of the compartments in a specified system. CONSTITUTION:The inside of a crucible is comparted into two parts with a partition having a communication part in one part. Melt consisting of a raw material of polycrystal and a dopant is formed in the comparted crucible. Seed crystal is brought into contact with the melt in one part of the compartments and pulled up to grow single crystal and also the raw material of polycrystal and the dopant are replenished to the other compartment. In this case, until the concn. CAi of the dopant in the melt contained in the other compartment is made nearly equal to the concn. Cs, i (=KoCo, i, Ko shows equibrium segregation coefficient of the dopant. Co and i show the concn. of the theoretical dopant in the melt necessary to obtain the aimed resistivity) of the theoretical dopant equivalent to the aimed resistivity rho0 of the obtained single crystal body, only the raw material of polycrystal is added. Thereafter the raw material of polycrystal and the dopant having Ko and concn. Co, i are added.

Description

[Detailed Description of the Invention] [Industrial Field of Application] The present invention relates to a doping method for pulling a single crystal, and is intended to improve the uniformity of the resistance skewer in the obtained single crystal and improve the product yield. It's about technology.

[Prior Art] As a method for manufacturing single crystals of silicon, germanium, gallium arsenide, etc., the Czochralski method (CZ method), in which a feed metal is grown and pulled from a melt in a crucible, is widely used. .

Incidentally, in recent years, when devices are manufactured using single crystals obtained by the CZ method, the range of usable resistance values of single metals tends to be limited and narrowed as the devices become more sophisticated. In particular, with the miniaturization of LSIs, strict control of the threshold voltage is required in, for example, CMOS devices, and the allowable resistance value of monoviscous crystals has been limited to a very narrow range. Therefore, in the CZ method, how to control the resistance distribution of the obtained single crystal is an extremely important issue.

Conventionally, when manufacturing silicon single crystals by the CZ method,
Typically, phosphorus, boron,
The method used is to select a dopant such as antimony, weigh the required amount, mix it with the raw material silicon, and melt it simultaneously. When a silicon single crystal is pulled up from the molten contact liquid in this way, the segregation coefficient of the added dopant is 1.
, the dopant concentration in the melt increases as the pulling progresses. Therefore, the dopant concentration contained in the silicon single crystal is low at the top side of the single crystal and high at the end, and the resistance value gradually decreases from the corner to the end, until the desired resistance value is reached. The silicon urethane used for this purpose could only be obtained from a small portion of silicon single crystals, and the yield was poor.

From this point of view, a method has been proposed in which the single crystal is continuously "raised" by replenishing the crucible with polycrystalline silicon raw materials and dopants according to the amount of pulling during the lifting operation during the single-crystal surgery. According to such a method, it is theoretically possible to keep the dopant concentration in the melt constant during the single crystal pulling operation, and the fluctuation in the resistivity value of the obtained silicon single crystal is reduced. It is conceivable that the yield of Ueno suitable for the target device will be improved.However, in reality, in conventional single crystal pulling equipment, polycrystalline silicon raw material and dopant) are placed in the crucible during the single crystal pulling operation. If it is supplied, the convection of the melt will be disturbed and the unmelted polycrystalline silicon raw material will reach the solidification zone, resulting in H' dislocation and polycrystalization of the pulled single feed.

Furthermore, recently, in such continuous pulling methods for single crystals, it is possible to reduce the f1'' of the pulled single product due to the replenishment of polycrystalline silicon raw materials and dopants, as described below, and to prevent polycrystalization. In order to achieve this, the inside of the crucible may be divided into a solidification zone where the single crystal is pulled and a melting zone where polycrystalline silicon raw materials and dopants are supplied by placing a partition wall having a through hole in the crucible. Proposal!
4 (for example, Japanese Patent Application Laid-Open No. 58-130195S;
, Special Publication No. 62-502793, Japanese Patent Publication No. 58-2609
No. 6, JP-A-62-105992, etc.).

It is true that if the interior of the crucible is divided into a solidification zone and a melting zone by partition walls, etc., the convection of the melt will be disturbed by the supply of multi-product silicon raw materials and dopants, and the unmolten polycrystalline silicon It is thought that the risk of the raw material reaching the solidification zone will be reduced, and the risk of causing right-handed dislocation of the unit feed amount and polycrystalization will be reduced.

However, on the other hand, if a crucible with such a structure is used to perform the above-mentioned continuation, F, etching operation of a silicon single crystal, the resistivity of the resulting single crystal will be lower than that of the upper region of the single crystal. There is a great possibility that the resistivity will deviate from the desired value over a considerable length.

To elaborate on this point, in growing a single crystal, first, the target resistivity ρ of the single crystal to be obtained is grown in a crucible with a two-city structure as described above. The theoretical dopant concentration C corresponding to
A polycrystalline silicon melt containing impurities with the theoretical dopant concentration in the melt to obtain 5,1 is formed, and then a single crystal pulling operation is started in the solidification zone, and as it progresses, , the above theoretical dopant concentration Cs
, i (=, Co, i, where ko is the equilibrium segregation coefficient of the dopant.) A polycrystalline silicon raw material containing a dopant at a concentration equal to S is supplied to the melting zone in an amount commensurate with the amount of withdrawal. It is done by Theoretically, this appears to keep the dopant concentration constant in the melt in the crucible. However, in reality, there is a partition between the solidification zone and the melting zone in the crucible, so the melt in the solidification zone and the melt in the melting zone do not substantially mix, and the solidification zone Since only the reduced amount flows into from the melting zone, the dopant concentration CA of the melt at the beginning of the pulling operation, that is, in the melting zone, decreases and reaches koCo,i (as mentioned above, the concentration of phosphorus, boron, etc. The segregation coefficient ko of the dopant is smaller than 1), the solidification zone is supplied with silicon melt containing a larger amount of dopant than the amount of dopant reduced by precipitation, and as a result, the amount of dopant in the solidification zone is The actual dopant concentration CB in the melt becomes higher than the theoretical dopant concentration C8,1, and naturally the dopant concentration C5 in the precipitated single crystal also becomes higher than the theoretical concentration C3,1, resulting in a reduction in resistance. The resistivity ρ was lower than the resistivity ρ0.

[Problems to be Solved by the Invention] Therefore, an object of the present invention is to provide a novel doping method that can stably and continuously obtain a single crystal having uniform resistivity in the length direction. It is something.

[Means for Solving the Problems] The objects described in item 2 are obtained by melting a polycrystalline raw material and a dopant in a crucible that has a communicating portion in part but is essentially partitioned into two parts by a partition wall. Forming a melt and supplementing said melt in one compartment with 1'? In a method in which a single crystal is grown by contacting and pulling a single crystal, a crystal raw material and a dopant are replenished in the other compartment, and the single crystal is continuously pulled. The dopant concentration CA is the desired resistivity ρ of the resulting single crystal. The theoretical dopant concentration C1(-
, C01, where CO1 is the theoretical dopant concentration in the melt to obtain the desired resistivity, and ko is the equilibrium segregation coefficient of the dopant. ) until the polycrystalline raw material or the polycrystalline raw material and the
Adding a significantly lower concentration of dopant to C0,
ko<1), or polycrystalline raw material and C0,
A significantly higher concentration of dopants should be added to ko〉1
), then achieved by a doping method in single crystal pulling characterized by adding a polycrystalline raw material and a dopant with a concentration of l(.C0°).

The above method also involves forming a melt by melting a polycrystalline raw material and a dopant in a crucible that has a communicating portion in part but is substantially partitioned into two parts by a partition wall, and one of the crucibles is divided into two parts by a partition. In a method in which a single crystal is grown by bringing a pressed crystal into contact with the melt and pulling it, the polycrystalline raw material and dopant are replenished in the other section, and the single crystal is continuously pulled. The dopant concentration C1o in the liquid is the desired resistivity ρ of the obtained single crystal. The theoretical dopant concentration corresponding to C−,+
This can also be achieved by a doping method for single-crystal products characterized by having the following relationship for the theoretical dopant concentration Co,i in the melt to obtain (=koC0,+). be done.

C1a <C,, r (if k-<1) C, o>
C0,, (when ko>1) The present invention also provides CILI for CO,l of about 10 to 2.
About 0% low (for ko 1) or high (ko>
1) shows a doping method in single crystal pulling.

The above-mentioned aspects further include forming a melt by melting the polycrystalline raw material and the dopant in a crucible whose interior is substantially partitioned into two by a partition wall, although a communicating portion is formed in a part of the crucible, and one of the crucibles is formed by melting the polycrystalline raw material and the dopant. In a method in which a co-crystal is brought into contact with the melt in one compartment and pulled to grow a single crystal, a polycrystalline raw material and a dopant are replenished in the other compartment, and the single crystal is continuously pulled. A theoretical dopant concentration Cs in which the dopant concentration C+0 in the melt corresponds to the desired resistivity ρ0 of the obtained single crystal.
,l (=to,C,,i) for the theoretical dopant concentration cO,I in the melt, C+ +]
<Co, + (in case of koku1) C+ 0 >c
h, and the dopant concentration CAl in the melt in the other compartment.
is the desired resistivity ρ of the resulting single crystal. Until the theoretical dopant concentration C3,1 corresponding to
), or to the polycrystalline raw material and C0,+, a particularly high concentration of dopants should be added to C0,+ (if ko>1), then to the polycrystalline raw material and C,.

Single IW characterized by adding a concentration of dopants
This is achieved by a doping method in 1.

[Effect] Hereinafter, the present invention will be explained in detail based on embodiments.

FIG. 1 schematically shows the configuration of a single crystal pulling apparatus in use, which is used in the doping method for pulling a single crystal according to the present invention.

In the silicon single crystal pulling apparatus 1 shown in FIG. 1, an annular partition wall 4 having a through hole 3 in a part thereof is disposed inside a quartz crucible 2, and has a two-city structure. and,
The area inward from the annular partition wall 4 is a solidification zone 8 in which a silicon single crystal 7 is grown by contacting the melt 5 with a weighing metal 6 and pulling it up. Silicon raw material supply pipe 10 and dopant supply pipe 11
This is a melting zone 9 in which polycrystalline silicon raw materials and dopants, which are replenished from each other, are melted.

Prior to pulling a silicon single crystal, first, a multi-crystal silicon raw material and a dopant are loaded into the crucible 2, and the multi-crystal silicon raw material and dopant are passed through a cylindrical heater 1.
2 to form a melt 5. The dopant concentration C+0 in the melt 5 at the start of this pulling operation is equal in both the solidification zone 8 and the melting zone 9.

Then, the weighing wire 6 is attached to the tip of the pulling wire 13 by dipping the weighing wire 6 fixed by the metal fitting 14 into the melted itl 5 in the coagulation zone 8 and pulling it at a predetermined speed. Grow silicon single crystal 71. As the silicon single crystal 7 grown from item 6 is pulled up, the melt 5 decreases by the amount of the pulled silicon, so the weight of the pulled single crystal 7 is lifted and the wire is wound. The weight is detected by a gravimetric analyzer (not shown) attached to the taker 15, and based on the information, the polycrystalline silicon raw material and dopant are replenished from the polycrystalline silicon raw material supply pipe IO and the dopant supply pipe 11, respectively, and the crucible is The amount of melt 5 in 2 is kept constant.

Here, in a structure in which the solidification zone 8 and the melting zone 9 in the crucible 2 are partitioned by the partition wall 4, as in the single-piece product single-piece heating apparatus shown in FIG. The melt 5B in the solidification zone 8 and the melt 5B in the melting zone 9
It is considered that it does not mix with A and only the reduced amount flows into the solidification zone 8 from the melting zone 9.

Therefore, the dopant concentration C(H) in the melt 5B in the solidification zone 8 and the dopant concentration CAi in the melt 5A in the melting zone 9 are the same concentration Ctfl at the start of pulling (t=0). However, the resistivity p of the single crystal obtained is
is the desired resistivity ρ. To make it constant, end)rn
Thirdly, the dopant concentration CB in the melt 5B in the coagulation zone 8
+ is the flu resistivity ρ for the above purpose. Let the theoretical dopant concentration of the melt Co,i be the dopant concentration CAl in the melt 5A in the melt zone 9.
(Theoretical dopant concentration C of a single crystal corresponding to resistivity ρ0
(i.e., the amount of dopant decreases as it is pulled from the solidification zone 8 as a single crystal 7, and the amount of dopant decreases as it is pulled from the solidification zone 8 as a single crystal 7.) The amount of dopant in the melt flowing from 161-
Therefore, there is no fluctuation in the dopant concentration in the melt 5B in the coagulation zone 8, and the concentrations Co and I become constant. )There is a need.

Therefore, in the first method of the present invention, the dopant concentration C in the melt 5A in the melting zone 9 is the theoretical dopant concentration C-,+ (= If a dopant with a segregation coefficient of 0 for silicon, such as phosphorus, boron, antimony, or arsenic, is smaller than 1, the silicon polycrystalline raw material or A significantly lower concentration of dopant, preferably only the silicon polycrystalline raw material, is added to the crystalline raw material and koCo,i, and then the polycrystalline silicon raw material and!
<oco, a dopant with a + concentration is added.

For example, if the dopant concentration CILI in the melt 5 (5A, 5B) at the start of pulling is the theoretical dopant concentration Co,i in the melt to obtain the target resistivity, as in the conventional method, then the obtained The dopant concentration Cs, (=, Ca,) or resistivity ρ in the heavy viscosity 1'1ir1 body is naturally the desired value Cs, + or ρ at the start of pulling (1=0). However, on the other hand, the dopant concentration CAi in the melt 5A in the melting zone 9 at that point is approximately equal to Co,i, and in the solidification zone 8 there is a dopant concentration CAi higher than the dopant concentration koCB+ reduced by the elongation. of silicon melt will be supplied from the melting zone 9, and at the next moment (1=1+), the dopant concentration in the melt 5B in the solidification zone 8 will be C,3
, becomes higher than the theoretical dopant concentration Co,i, and the dopant concentration C5t in the single crystal 7, which is naturally raised, also becomes higher than the theoretical concentration Cs,I, and the resistivity ρ becomes the target resistivity ρ. The value will be lower than . Such a concentration tendency of the dopant concentration C,3, in the melt i&5B in the solidification zone 8 is caused by the dopant concentration C3,( -koC8,), and the dopant concentration CA+ in the melt 5A in the melting zone 9 is further pulled down. The dopant concentration CAi in the melt lf1.5 A is the theoretical dopant concentration Cs,i(=to,C, ,
) until the coagulation zone 8
The dopant concentration CB in the melt 5B remains higher than the theoretical dopant concentration Co,i, and the resistivity ρ in the single crystal 7 to be pulled is the target resistivity ρ. The state of shifting to a lower value will continue. However, in the first method of the present invention, as described above, the dopant concentration CA in the melt 5A in the melting zone 9 is different from the theoretical dopant concentration Csi (=
Until C, , ”) becomes almost equal to C, ”), dopant is added to only the silicon polycrystalline raw material or to the polycrystalline raw material and the polycrystalline raw material at a significantly lower concentration than C, , at the start of pulling. (1=0) to the point in time (1=1.) when the dopant concentration CA in the melt 5A in the melting zone 9 becomes almost equal to the theoretical dopant concentration C,1,(=,Co,i) of the single crystal. The time is minimized or greatly shortened, and in the length direction of the resulting single crystal,
Resistivity ρ is the target resistivity ρ. Areas that fluctuate more greatly will decrease.

In this first method, the dopant concentration CA in the melt 5A in the melting zone 9 becomes approximately equal to the theoretical dopant concentration C, , (=, Co1) of the single crystal (1=1.). The judgment is that the dopant concentration in the lucid liquid during the pulling operation is determined in real time by a total of 4 (sipping is practically a national disaster, so it is desirable to analytically determine it in advance.

In addition, in the second method of the present invention, when a dopant with a segregation coefficient of 0 is smaller than 1 is used, the dopant concentration C+0 in the melt 5 at the beginning of pulling is set to the desired concentration of the single crystal. The dopant concentration Co,i is kept lower than the theoretical dopant concentration Co,i in the melt to obtain the jJL resistivity ρ0. As described above, when the dopant segregation coefficient l(.
+, the dopant concentration C,3, in the melt 5B in the solidification zone 8 tends to concentrate, and the resistivity ρ in the single crystal 7 gradually decreases. As shown in , the dopant concentration CIO in the melt 5 at the beginning of pulling is determined by the desired resistivity ρ of the single crystal. The theoretical dopant concentration C in the melt to obtain
If o and i are set lower than i, the resistivity ρ of the single crystal body 7 at the beginning of pulling becomes the target resistivity ρ. During the concentration period, the resistivity ρ gradually decreases from this point until reaching the target resistivity ρ. , and even less. Therefore, the resistivity ρ of the resistivity ρ in the single-connected W1 body 7 is the target resistivity ρ. There are no parts that deviate greatly from this, and the fluctuation of the skewer is reduced. In addition, the melt 5 at the beginning of pulling]
The reduction rate of the dopant concentration CILI in the melt with respect to the theoretical dopant concentration Co,i in the melt cannot be unconditionally defined because it depends on the type of dopant to be doped (segregation coefficient), etc. Approximately 10 to 2 within
Approximately 0% is appropriate.

Furthermore, in the third method of the present invention, the requirements of both the first method and the second method are applied during the single crystal pulling operation. That is, when a dopant with a segregation coefficient of 0 smaller than 1 is used, the dopant concentration C+O in the melt 5 at the beginning of pulling is
The desired resistivity ρ of the single crystal. IJ into the melt to obtain
- lower than the theoretical dopant concentration Co1,
Furthermore, from the start of the pulling operation, the melt 5 in the melting zone 9
The dopant concentration CAl at A is the desired resistivity ρ of the resulting single crystal. The theoretical dopant concentration Cs corresponding to
Until it becomes almost equal to i (=, C,, i), the silicon polycrystal raw material or the polycrystal raw material and the Co
, + is added at a significantly lower concentration, preferably only the silicon polycrystalline raw material, and then the polycrystalline silicon raw material and the dopant at the koCo,i concentration are added. Therefore, the above-mentioned effects are combined, and the resistivity ρ is the desired resistivity ρ throughout the pulled single crystal 7. It will be almost close to.

The present invention has been explained above by taking as an example the case where silicon is doped with a dopant having a segregation coefficient of less than 1 to pull down a single crystal, but the doping method for the single crystal of the present invention is The present invention can be similarly applied to other semiconductor materials produced by the same pulling method, such as germanium, gallium arsenide, etc. Also, even when doping with a dopant with a segregation coefficient larger than 1 (for example, boron for germanium), the requirements for doping with a dopant with a segregation coefficient smaller than 1 as described in It are almost the opposite, that is, the melting zone 9
The dopant concentration CAi in the melt 5A is the desired resistivity ρ of the resulting single crystal. The theoretical dopant concentration Cs, + (= koc.

) until the polycrystalline raw material and k(IC
The dopant concentration Ci0 in the melt 5 at the beginning of the initial pull-down is calculated by adding a dopant with a significantly higher concentration than 0, +, and then adding a dopant with a polycrystalline material and a dopant with a + concentration of koCo, or by simply b resistivity ρ of the concretion.

Theoretical dopant concentration C00 in the melt to obtain
, the desired resistivity ρ of the single crystal precipitated in the same way as above. This means that the fluctuations in

Furthermore, in the single crystal pulling apparatus shown in FIG. 1, the crucible has a double structure by disposing an annular partition wall 4 having a through hole 3 in a part inside the crucible 2. In the method of the present invention, the crucible in the single crystal pulling apparatus has a communication part in part, but the inside is substantially divided into two parts by the partition wall.
As long as the single crystal pulling apparatus has a structure different from that shown in FIG. 1, it can be similarly carried out.

Furthermore, the apparatus shown in FIG. 1 may have a structure in which the polycrystalline raw material and the dopant are mixed in advance and replenished into the melt 5A using two supply pipes. However, as shown in the figure, replenishment using separate supply pipes 10.11 allows the supply amount of dopant to be arbitrarily controlled relative to the supply amount of polycrystalline raw material, making it easier to control the dopant concentration. So outrageous.

Although not directly related to the present invention, a single-connected 1'TI which forms a crucible whose interior is substantially divided into two by a partition wall, although it has a communicating part in a part as in the present invention. When pulling a single crystal containing a dopant with a segregation coefficient ko smaller than 1 using a pulling device, first, a melt consisting only of polycrystalline silicon raw materials is formed in a crucible, and immediately before starting to pull the single crystal. A dopant is added only to the solidification zone side of the crucible, and the dopant concentration in the FAIi liquid in this solidification zone is set to the desired resistivity ρ of the single crystal. melt to obtain
The theoretical dopant concentration at night is Ca,l, and once the pulling operation is started, polycrystalline silicon raw material and koCo,i are added to the melting zone depending on the weight of the single crystal deposited.
Even if the single crystal pulling operation is performed by replenishing the dopant concentration, the resistivity ρ of the precipitated single crystal remains the same as the desired resistivity ρ.

Fluctuations in the temperature are kept low.

[Example] Hereinafter, the present invention will be explained in more detail with reference to Examples.

In a single crystal pulling apparatus having the configuration shown in FIG. 1, a polycrystalline silicon raw material and phosphorus (k, = 0.35) as a dopant are first melted to form a melt 5. Note that the amount of melt 5A in melting zone 9 is 11.1
kg, the amount of a liquid 5B in coagulation zone 8 is 13.9
kg. Then, a single crystal pulling operation was performed under various conditions shown below to produce a phosphorus-doped N-type silicon single crystal (6 inches in diameter).

In Example 1, the phosphorus concentration in the molten l&5 at the start of pulling was set to the desired resistivity ρ, -10 of the silicon single crystal.
Theoretical concentration C0° to obtain Ω·cm = 1. 14X
1015atoIIls/cm3, and the phosphorus concentration of the melt 5A mountain in the melting zone 9 is koC from the start of pulling.
o, + = 4 x 10” atoms/cm
3, only polycrystalline silicon raw materials are used as single crystals.
The melting zone is replenished in an amount commensurate with the weight of the single crystal deposited, and thereafter, the polycrystalline silicon raw material and phosphorus with a concentration of koCo and i are replenished into the melting zone in an amount commensurate with the weight of the single crystal deposited.

In Example 2, the phosphorus concentration in the melt 5 at the start of pulling is equal to the desired resistivity ρ of the silicon single crystal. =10
Theoretical concentration Co to obtain Ω cm, + = 1.14
About 10% lower I than X1015atoms/am3
X 1015 atoms/cm3, and from the start of pulling, multicrystalline silicon raw material and phosphorus with a concentration of koCo of 1 minute are supplied to the melting zone in an amount corresponding to the weight of single deposited metal.

Furthermore, in Example 3, the phosphorus concentration in the molten l&5 at the start of pulling corresponds to the desired claw resistivity ρ of the silicon single crystal. =
Theoretical concentration Co to obtain 10 Ω·cm, + = 1.
Approximately 10% less than 14X1015atoms/cm”
The phosphorus concentration in the melt 5A in the melting zone 9 is set to be 1 (.C,, + = 4 × 10” atoms
/cm3, only the polycrystalline + T + silicon raw material is supplied to the melting zone in an amount corresponding to the single-crystalline 1'71 pulling weight, and after that, the polycrystalline silicon raw material and the concentration are
Co, 78 phosphorus for i in an amount commensurate with the weight of the single pay raise
This is to supply the melting zone.

On the other hand, as a comparative example, based on the conventional method, the phosphorus concentration in the melt 5 at the start of pulling is equal to the desired resistivity ρ of the silicon single crystal. =Theoretical concentration co to obtain 10Ω・cm=
=1. 14X1015 atoms/cm3, multi-crystalline silicon raw material and concentration ko064 from the start of pulling.
This is to supply phosphorus to the melting zone in an amount commensurate with the weight of the single crystal pulled.

And the resistivity ρ of each single crystal obtained in this way
The ratio ρ/ρ0 between the resistivity and the target resistivity ρ0 was analyzed over the entire length of the single crystal. As is clear from the results shown in FIG. 2, in the comparative example, ρ/ρ. The variation range from =1 was 23%, but in Example 1 it was 16%.
%, in Example 2 it was 17%, and in Example 3 it was 6%, both of which had smaller fluctuation ranges than the comparative example.

[Effects of the Invention] As described above, according to the present invention, a polycrystalline raw material and a dopant are melted in a crucible that has a communication portion in a part but is essentially partitioned into two parts by a gap. A seed crystal is brought into contact with the melt in one compartment and pulled up to grow a single crystal, while a polycrystalline raw material and a dopant are replenished in the other compartment to continuously grow a single crystal. In the method of pulling a body, the dopant concentration C1 in the melt at the beginning of pulling is determined by the desired resistivity ρ of the single crystal. The dopant concentration in the melt in the other compartment of the crucible is set to be slightly closer to the koCo,i value than the theoretical dopant concentration in the melt to obtain C6,1, and as the pulling operation progresses, the dopant concentration in the melt in the other compartment of the crucible is By adding polycrystalline raw materials and dopants so that the I(.Co,
Moreover, the desired value can be obtained, and the product yield can be improved.

[Brief explanation of drawings]

Figure 1 schematically shows the configuration of the single crystal pulling equipment used in the method of the present invention in use, and Figure 2 shows the selection in the longitudinal direction of the single crystal pulled under the conditions of a platform scale. This shows the variation in resistivity. 1... Single crystal pulling device, 2... Quartz crucible, 3.
... Through hole, 4... Annular partition wall, 5.5A, 5B... Melt, 6... Pressed crystal, 7...
Single crystal, 8... solidification zone, 9... melting zone,
10... Polycrystalline silicon raw material supply tube, 11... Dopant supply tube, 12... Cylindrical heater, 13...
Pulling wire, 14...Metal fittings for mounting, 15... Pulling wire winding machine. Patent originator Nippon Steel Corporation

Claims (1)

[Scope of Claims] (1) In a crucible that has a communicating portion in part but is substantially partitioned into two parts by a partition wall, a polycrystalline raw material and a dopant are melted to form a melt; In the method of continuously pulling a single crystal by bringing a seed crystal into contact with the melt in one compartment to grow a single crystal, and replenishing a polycrystalline raw material and a dopant in the other compartment, Dopant concentration C_A_ in the melt of the other compartment
i is the theoretical dopant concentration C_s_i (=k_oC_o_,
_i, where C_o_, _i is the theoretical dopant concentration in the melt to obtain the desired resistivity, and k_
o is the equilibrium segregation coefficient of the dopant. ), a significantly lower concentration of dopant is added to the polycrystalline raw material or to the polycrystalline raw material and k_oC_o_,_i (if k_o<1), or a significantly lower concentration of dopant is added to the polycrystalline raw material and k_oC_o_,_i until the A doping method for pulling a single crystal, characterized in that a high concentration of dopant is added to (if k_o>1), and then a polycrystalline raw material and a dopant of k_oC_o_,_i concentration are added. (2) In a crucible that has a communicating portion in part but is essentially partitioned into two parts by a partition wall, a melt formed by melting a polycrystalline raw material and a dopant is formed, and in one of the parts, the melt is In the method of continuously pulling a single crystal by bringing a seed crystal into contact with a liquid and pulling it to grow a single crystal, the polycrystal raw material and dopant are replenished in the other compartment, and the single crystal is continuously pulled. The dopant concentration C_i_0 in is the theoretical dopant concentration C_s_,_ corresponding to the target resistivity ρ_0 of the obtained single crystal body.
i (=k_oC_o_,_i, where k_o is the equilibrium segregation coefficient of the dopant). Doping method in crystal pulling. C_i_0<C_o_, i (if k_o<1) C_i
_0>C_o_, _i (if k_o>1) (3) C_
3. The doping method for pulling a single crystal according to claim 2, wherein C_i_0 is about 10 to 20% lower (when k_o<1) or higher (when k_o>1) with respect to o_,_i. (4) The dopant concentration C_i_0 in the melt at the beginning of pulling is as described in claim 2 or 3, and the replenishment of the polycrystalline raw material and dopant impurities during the pulling growth process is as described in claim 1. Doping method in single crystal pulling.