JP6369352B2 - Crystal growth method - Google Patents

Description

The present invention relates to the Czochralski method (hereinafter referred to as CZ).
(Hereinafter abbreviated as “method”).

  In the CZ method, a raw material is charged into a crucible and crystals are grown from a melt (silicon melt) obtained by melting the raw material. As the crystal grows, the amount of melt in the crucible decreases. As a method for compensating for this decrease, a recharge method for growing a crystal again after adding a raw material after crystal growth and a continuous charge for adding a raw material during crystal growth have been devised (for example, Non-Patent Document 1).

  Since the recharge method does not add a raw material during crystal growth, there is an advantage that there is almost no problem of dislocation of a single crystal due to the input of the raw material. However, the concentration of impurities (dopants) input for resistance adjustment increases with crystal growth due to the segregation phenomenon, and the resistivity decreases. For this reason, a dopant having a small segregation coefficient, such as P, compared to B, which has a relatively large segregation coefficient, shortens the crystal length that satisfies the resistivity standard and decreases the yield. A switching device used for power, such as an IGBT (insulated gate bipolar transistor), uses an N-type crystal with a dopant having a small segregation coefficient such as P as a dopant, resulting in high cost.

  On the other hand, the continuous charge method in which a raw material is added during crystal growth has an advantage that a high concentration of dopant due to a segregation phenomenon can be suppressed by the addition of the raw material, and a crystal having a uniform resistivity can be obtained. However, the continuous charge method has a problem of dislocation formation because the raw material must be supplied while pulling up the single crystal.

  In order to suppress this, a technique using a double crucible is disclosed (for example, Patent Document 1). A single crystal is grown inside the double crucible, and the raw material is supplied to the outside. The raw materials to be charged are granular or small pieces, and many are charged while controlling an appropriate amount from a supply device provided outside the puller (for example, Patent Document 2). However, there is a problem that the double crucible is expensive. Therefore, a method that does not use an expensive double crucible is disclosed. For example, in Patent Document 3, an isolation pipe having an inert gas exhaust pipe is provided at the tip of a pipe for supplying raw material to the melt. There is a possibility that dislocations can be suppressed without using a double crucible with such an apparatus. However, it is also a complicated structure in terms of equipment. Moreover, although the main purpose is quality improvement, a technique for controlling convection by applying a magnetic field is also disclosed (for example, Patent Documents 4 and 5). Since the convection can be suppressed by applying the magnetic field in this way, there is a possibility that there is an effect of reducing dislocation due to the additional raw material.

  Moreover, the method of supplying, after melt | dissolving a granular raw material by induction heating etc. is disclosed (for example, patent document 6, 7). However, in order to feed granular poly while controlling it, a supply device as disclosed in Patent Document 2 is necessary, and it cannot be said that the device is simple.

  Therefore, a supply method using a raw material rod instead of supply using granular poly is disclosed in Patent Document 8 and the like. With such a raw material supply, there is a possibility that a supply device such as granular poly is not necessary and the device can be simplified. However, the technique disclosed in Patent Document 8 is a method of dissolving a raw material rod and then supplying it as a melt. For this reason, in addition to the main heater for heating the crucible, a heating source for melting the raw material rod is required, and there is a problem that the apparatus becomes complicated.

JP-A 63-233092 Japanese Patent Laid-Open No. 3-60489 Japanese Utility Model Publication No. 4-84362 JP-A-6-92776 JP 2010-30860 A JP-A-5-279166 JP-A-8-183688 JP-A-2-279582

Fumio Shimura, Semiconductor Silicon Crystal Technology, P178-183, 1989

  The present invention has been made in view of the above-described problems, and an object of the present invention is to provide a crystal growth method that can easily suppress an increase in dopant concentration due to a segregation phenomenon and dislocation of crystals. .

  In order to achieve the above object, in the present invention, in the CZ method, a method for growing a crystal while supplying a raw material to a melt in a crucible during crystal growth, using the rod-shaped raw material as the raw material, the rod The solid raw material is brought into contact with the melt between the growing crystal and the crucible wall, and the rod-like raw material is inserted continuously or intermittently to grow the crystal while melting. A crystal growth method is provided.

  With such a crystal growth method, it is possible to prevent an increase in dopant concentration due to a segregation phenomenon by adding a raw material during crystal growth. In addition, by bringing the rod-shaped raw material into contact with the melt in a solid state, the crystal being grown is dislocated due to the splashing of the droplets caused by the dropping of the granular raw material of the conventional method or the vibration of the melt surface caused by the introduction of the melt. The problem is almost gone.

  Further, when the rod-shaped raw material is inserted into the pulling furnace, the rod-shaped raw material is inserted into the pulling furnace by the movable part using an insertion machine having a movable part capable of moving the rod-shaped raw material up and down. It is preferable.

  By using such an insertion machine, the rod-shaped raw material can be easily inserted into the pulling furnace.

  In this case, when the rod-shaped raw material is added, the addition is performed by using the one having an insertion portion having a gate valve that can further isolate the inside of the pulling furnace from the outside as the insertion machine. It is preferable to carry out after isolating the inside of the pulling furnace from the outside of the furnace.

  With such a crystal growth method, the crystal growth apparatus to be used does not become large, and an additional rod-shaped raw material can be introduced easily, and the manufacturing cost can be reduced.

The cross-sectional area of the rod-shaped raw material is preferably 1/10 or less and 10 mm ² or more of the crystal for growing.

  By using such a rod-shaped raw material, it is possible to sufficiently compensate for the melt amount that has been reduced by growing crystals. Moreover, such a rod-shaped raw material can be easily dissolved.

  Moreover, the number of the rod-shaped raw materials can be plural.

  By using a plurality of rod-shaped raw materials in this way, the amount of melt that decreases during crystal growth can be easily compensated.

  Moreover, it is preferable to apply a magnetic field having a central magnetic field strength of 500 to 6000 G to the melt.

  By applying a magnetic field having such a central magnetic field strength, it is possible to prevent a relatively low temperature melt generated by melting the rod-shaped raw material from reaching a large amount in the growing crystal.

  Moreover, it is preferable to heat the said melt using the heater arrange | positioned around the said crucible, and to make the upper end of the heat generating part of the said heater above the molten metal surface which is the surface of the said melt.

  By positioning the heater heating portion in this way, radiant heat can be applied to the rod-shaped raw material, and dissolution can be performed more efficiently.

  Moreover, it is preferable to form a solidified layer in the melt.

  By forming the solidified layer in the melt as described above, the controllability of resistivity can be further improved.

  The rod-shaped raw material is a non-doped raw material or a raw material having the same conductivity type as that of the initial melt, and the dissolution ratio α of the rod-shaped raw material is (1-k) / 10 ≦ α ≦ 1 (where α is constant) It is preferable that the rod-shaped raw material is dissolved in the melt with respect to the mass crystallized from the melt within the time, and k is the segregation coefficient of the dopant introduced into the melt for resistivity control. .

  By using such a rod-shaped raw material and setting the dissolution ratio α within the above range, a decrease in resistivity can be suppressed to some extent.

The growth rate V (mm / min) of the growing crystal is 0.01 ≦ V ≦ 15000 / D ² +0.45 (where D is the diameter of the growing crystal (unit: mm)). It is preferable that

  With such a crystal growth method, solidification does not occur on the melt surface even if the rod-shaped raw material is inserted beside the crystal.

  Moreover, the insertion of the rod-shaped raw material can be started after growing for the slipback length from the start of the growth of the product part in the growing crystal.

  With such a crystal growth method, even when dislocation is generated in a portion where the length of the straight cylinder is relatively short, it can be remelted as usual.

  In addition, after growing the crystal, the raw material is recharged before the next crystal growth, and the next crystal is grown again in the next crystal growth by using the crystal growth method of the present invention. A plurality of crystals can be grown.

  With such a crystal growth method, it is possible to perform multi-pooling for growing a plurality of crystals from one quartz crucible even when the rod-shaped raw material is dissolved at a ratio smaller than the dissolution ratio of α = 1. It is.

  Further, the conductivity type of the rod-shaped raw material is the same as the conductivity type of the initial melt, the dopant concentration of the rod-shaped raw material is within ± 10% with respect to the initial melt concentration, and the dissolution ratio α of the rod-shaped raw material is 0. It is preferably within 9 ≦ α ≦ 1.1 (where α is the mass of the rod-shaped raw material dissolved in the melt with respect to the mass crystallized from the melt within a certain time).

  Thus, if α is about 1, the amount of melt is almost constant and the height of the melt surface is almost constant, that is, the height position of the melt surface decreases with crystal growth as in the normal CZ method. There is no need to raise the crucible. Further, the dopant concentration of the rod-shaped raw material is set to the initial melt concentration, that is, the same as the dopant concentration in the initial melt determined by the raw material melted before the start of crystal growth and the amount of added dopant (main dopant). The resistivity inside can be kept constant.

  Further, the rod-shaped raw material is a non-doped raw material, and the dissolution ratio α of the rod-shaped raw material is between (1-k) / 10 and 1-k (where α is relative to the mass crystallized from the melt within a certain time) , And is a mass obtained by dissolving the rod-shaped raw material in the melt, and k is a segregation coefficient of the dopant introduced into the melt for resistivity control.

  Thus, when α is between (1-k) / 10 and 1-k, the amount of melt decreases, so the crucible needs to be raised in order to keep the melt surface height to a certain degree (the thermal environment is reduced). Although it cannot be kept constant, the resistivity in the crystal length direction can be kept constant. Moreover, it is not necessary to dissolve many rod-shaped raw materials at the same time as in the case of α = 1.

  With the crystal growth method of the present invention, it is possible to easily suppress the increase in dopant concentration due to the segregation phenomenon and the dislocation of crystals accompanying the charge of the raw material. In addition, since the resistivity of the grown crystal can be accurately controlled, the length of the product portion can be increased and the yield can be improved. Thereby, a switching device used for a power device or the like and a silicon single crystal for a diode can be manufactured at low cost.

It is a flowchart which shows an example of the procedure of the crystal growth method of this invention. It is the schematic which shows an example of the crystal growth apparatus which can be used for the crystal growth method of this invention. It is a figure which shows the result of having evaluated the axial direction resistivity distribution of the crystal | crystallization grown by Example 1, 2 and the comparative example.

  Hereinafter, the present invention will be described in more detail.

  As described above, there is a need for a crystal growth method that can easily suppress the concentration of dopants and the dislocation of crystals due to segregation.

  The inventor has intensively studied to achieve the above object. As a result, the inventors have found that a crystal growth method using a raw material supplied to the melt in the crucible during crystal growth as a rod-shaped raw material can solve the above problems, and completed the present invention.

  Hereinafter, embodiments of the present invention will be specifically described with reference to the drawings, but the present invention is not limited thereto.

  FIG. 2 is a schematic view showing an example of a crystal growth apparatus that can be used in the crystal growth method of the present invention. As shown in FIG. 2, a crystal growth apparatus (pulling furnace) 100 includes a main chamber 1 and a pulling chamber 2 connected to the ceiling (top chamber 11) of the main chamber 1 and containing the grown crystal rod 3. It has. A mechanism (not shown) for pulling up the crystal rod 3 with a wire is provided above the pulling chamber 2.

  A quartz crucible 5 for containing a melt (raw material melt) 4 and a graphite crucible 6 for supporting the quartz crucible 5 are provided in the main chamber 1, and these crucibles 5 and 6 are rotated by a drive mechanism (not shown). It is supported by a crucible shaft so that it can move up and down. A heater (main heater) 7 for melting the raw material is disposed so as to surround the crucibles 5 and 6. A heat insulating member 8 is provided outside the heater 7 so as to surround the periphery thereof.

  A gas inlet 10 is provided at the upper part of the pulling chamber 2, and an inert gas such as argon gas is introduced and discharged from the gas outlet 9 at the lower part of the main chamber 1. Further, a heat shield member 13 is provided so as to face the melt 4 so as to cut radiation from the surface of the melt 4 and to keep the surface of the melt 4 warm.

  Further, a gas purge cylinder 12 is provided above the melt 4 so that the periphery of the crystal rod 3 can be purged by the inert gas introduced from the gas inlet 10.

  The top chamber 11 is provided with an inserter 14 having a movable part, and is configured so that the rod-shaped raw material 15 can be inserted into the melt 4. This insertion machine 14 is usually provided with an insertion portion, and it is preferable to provide a gate valve capable of isolating the inside of the pulling furnace from the outside of the furnace.

  Next, the crystal growth method of the present invention will be described. The present invention is a method for growing a crystal 3 while supplying a raw material to a melt 4 in a crucible 5 during crystal growth in the CZ method using an apparatus as shown in FIG. The rod-shaped raw material 15 is brought into contact with the melt 4 between the growing crystal 3 and the inner wall of the crucible 5 while being solid, and the rod-shaped raw material 15 is inserted continuously or intermittently while being melted. A crystal growth method for growing the crystal 3.

  In this way, when a rod-shaped raw material (hereinafter referred to as a rod-shaped raw material) is melted into a melt as a solid, droplets are splashed by dropping a granular raw material in the conventional method, or vibration of the melt surface is caused by introducing a molten liquid. This eliminates the problem of dislocation of the growing crystal. Also, since a rod-shaped raw material is used, it is only necessary to control the speed at which it is inserted, a supply device for controlling the amount of raw material such as granular poly, and heating for dissolving the raw material into a molten liquid There is no need to use a device.

  In general, in the CZ method, since there is a heater for heating the melt around the crucible, the temperature distribution in the crucible is higher at the outer periphery and lower at the center. While the crystal is grown in the central portion having this low temperature, it is possible to dissolve the rod-shaped raw material between the crystal having a higher temperature and the crucible wall. Furthermore, since a portion of the rod-shaped raw material is not directly radiated from the heater by the rod-shaped raw material, there is an advantage that the crystal is easily cooled.

  FIG. 1 is a flowchart showing an example of the procedure of the crystal growth method of the present invention. Hereafter, each process of the flowchart of FIG. 1 is explained in full detail. First, as shown in FIG. 1, a rod-shaped raw material is prepared. In the present invention, the raw material supplied to the melt in the crucible during crystal growth is not particularly limited as long as it is a rod-shaped raw material. For example, it can be a prism or columnar silicon single crystal or polycrystalline silicon. In this case, the size of one rod-shaped raw material and the total amount of rod-shaped raw materials used when growing one crystal are not particularly limited, such as crucible size, diameter of crystal to be grown, conductivity type, resistivity range, weight, etc. Can be appropriately determined in consideration of the above. The rod-shaped raw material may include a dopant or a non-doped raw material.

Moreover, it is desirable that the cross-sectional area of the rod-shaped raw material is 1/10 or less and 10 mm ² or more of the crystal for growing.

Since the rod-shaped raw material is inserted beside the crystal being grown, it is not easy to melt a material having the same thickness as the crystal being grown, for example. Therefore, the cross-sectional area per rod-shaped raw material is preferably 1/10 or less of the crystal being grown. On the other hand, if the rod-shaped raw material is too thin, the effect is small no matter how fast it is melted in order to compensate for the melt amount reduced by growing crystals. Therefore, it is desirable that the cross-sectional area is 10 mm ² or more.

  Furthermore, the number of rod-shaped raw materials can be plural.

  As described above, since the rod-shaped raw material is inserted into the melt next to the growing crystal, it is difficult to melt the thick material. Therefore, the raw material is thinned, but if it is thinned, it is difficult to compensate for the amount of melt that decreases. Therefore, it is preferable to compensate for the decreasing amount of melt by melting a plurality of rod-shaped raw materials.

  Further, as described above, the shadowed portion of the rod-shaped raw material cools by blocking direct radiation from the heater, so that, for example, a plurality of rod-shaped raw materials are arranged concentrically around the crystal to form a rod-shaped material. It is also possible to increase the crystal growth rate by simultaneously promoting melting of the raw material and cooling of the grown crystal. An example of a method of arranging a plurality of rod-shaped raw materials concentrically around the crystal is a method of arranging a plurality of inserters 14 shown in FIG. 2 concentrically around the crystal. Note that a plurality of rod-shaped raw materials may be melted by sequentially adding rod-shaped raw materials to one insertion machine. Further, a plurality of rod-shaped raw materials may be simultaneously loaded in one insertion machine.

  Next, as shown in FIG. 1, crystals are grown while the rod-shaped raw material is dissolved. In the present invention, the rod-shaped raw material prepared as described above is brought into contact with the melt between the growing crystal and the crucible wall as a solid, and the rod-shaped raw material is inserted or melted continuously or intermittently. , Grow crystals.

  Specifically, before the crystal growth, after charging the raw material (the raw material here is not limited to the rod-shaped raw material) to the crucible, heating the raw material with a heater disposed around the crucible, and forming a melt When the crystal is grown, the rod-shaped raw material is inserted and melted. Here, examples of the crystal that can be grown in the present invention include a silicon single crystal and a compound semiconductor single crystal.

  In the present invention, a crystal can be grown by the MCZ method (magnetic field application CZ method) by providing a magnetic field application device (not shown) on the outer periphery of the main chamber 1 shown in FIG. At this time, it is preferable to apply a magnetic field having a central magnetic field strength of 500 to 6000 G to the melt.

  Patent Documents 4 and 5 disclose techniques for applying a magnetic field to the melt in the continuous charge method. Although these main purposes are to make the quality of the crystal to be grown uniform, convection is suppressed by applying a magnetic field. In the case of the present invention, the rod-shaped raw material is dissolved beside the crystal being grown. If the relatively low-temperature melt produced here reaches a large amount in the growing crystal, solidification may occur and dislocation may occur. In order to prevent this, it is effective to apply a magnetic field to the melt.

  The magnetic field strength is preferably 500 G or more and 6000 G or less as the central magnetic field strength. If the central magnetic field intensity is 500 G or more, convection can be sufficiently suppressed. If the intensity of the central magnetic field is 6000 G or less, the apparatus for generating the magnetic field is not large, and the problem of the leakage magnetic field to the surroundings hardly occurs. Further, since it is impossible to deny the possibility that the quality of the crystal to be grown deteriorates under a very strong magnetic field strength, it is preferable to set the central magnetic field strength to 6000 G or less.

  Further, as described above, in the present invention, it is preferable to heat using a heater in which the melt is arranged around the crucible. In this case, it is preferable that the upper end of the heat generating portion of the heater be above the molten metal surface that is the surface of the melt.

  In the present invention, it is not necessary to newly prepare a heating source in order to dissolve the rod-shaped raw material. Basically, a rod-shaped raw material is inserted into a melt heated by a main heater arranged around a crucible, and dissolved. At this time, the rod-shaped material is more easily dissolved by applying radiant heat from the heater to the rod-shaped material. Since silicon transmits infrared rays at low temperatures, heating by radiation is inefficient, but the rod-shaped raw material in the present invention is at a high temperature while directly above the melt, and can sufficiently absorb direct radiation from the heater. is there. Therefore, by setting the upper end of the heat generating portion of the heater (main heater) above the melt surface, radiant heat can be applied to the rod-shaped raw material, and dissolution can be performed more efficiently.

  Here, the heat generating portion of the heater is, for example, a portion having a slit in a graphite heater. In general, a graphite heater has a structure in which slits are alternately cut from the top and bottom, one or two circuits, and a current flows from the terminal to the terminal. In a general heater, a slit that is inserted from above and a slit that is inserted from below are adjacent to each other and overlap each other in the height direction, which is the thinnest part and has a high electrical resistance. This high resistance portion is the portion that generates the strongest heat when electricity is applied, and is called a heat generating portion. In the case of a general graphite heater, the heat generating portion referred to here is a portion where the upper and lower slits are adjacent to each other and overlap in the height direction. Further, when the structure is not a general structure, the resistivity is higher than that of the surrounding area, which means that the portion is likely to generate heat.

  Moreover, it is preferable to form a solidified layer in the melt.

  In the melt, a method of forming a solidified layer mainly on the bottom of the quartz crucible is called a DLCZ (Double Layer CZ) method and is known as a method for controlling resistivity. In this method, a single crystal is grown on the surface of the melt, while a solidified layer formed in the melt is dissolved to make the dopant concentration in the melt uniform. If the solidified layer can be controlled well, the resistivity can be made uniform, and this technology can be applied to control the oxygen concentration.

  In order to make the resistivity uniform, it is necessary to control that the solidified layer is enlarged in the early stage of crystal growth and the solidified layer is reduced in the final stage. However, in general, the control for gradually dissolving the solidified layer is not easy, and there are problems that it takes time to form the initial solidified layer and that the solidified layer contacts the grown crystal.

  Therefore, the present invention can be used as a technique for compensating for the difficulty in controlling the solidified layer in the DLCZ method. For example, it is possible to supplement the resistivity control of the DLCZ method by inserting a larger amount of rod-shaped raw material in a scene where the solidified layer is difficult to dissolve and not inserting a rod-shaped raw material in a scene where the solidified layer is easily dissolved.

  By doing so, it is possible to obtain a crystal with high uniformity of resistivity even with a rod-shaped raw material that is small relative to the weight of the crystal to be grown.

  Furthermore, it is possible to set the oxygen concentration to a desired value while forming a solidified layer in the melt and to control the resistivity mainly by using the present invention.

  When inserting the rod-shaped raw material into the pulling furnace, it is preferable to insert the rod-shaped raw material into the pulling furnace with the movable part using an insertion machine having a movable part capable of moving the rod-shaped raw material up and down.

  In this way, the rod-shaped raw material can be inserted by an apparatus composed of only a movable part that can move up and down.

  However, since the diameter of the rod-shaped raw material is usually smaller than the crystal diameter, the length of the rod-shaped raw material tends to be longer. If an attempt is made to provide a corresponding range of motion, the apparatus becomes large.

  Therefore, it is desirable that the insertion portion in the insertion machine has a structure that can isolate the inside of the pulling furnace from the outside of the furnace by a gate valve or the like. After isolating the inside of the pulling furnace from the outside of the furnace, a rod-shaped raw material is added, and if this is inserted into the pulling furnace after the previous rod-shaped raw material, the insertion portion can be prevented from becoming long.

  Also, if the rod-shaped raw material is too close to the growing crystal, it will cause dislocations, so that the distance between the outer periphery of the crystal and the outside of the rod-shaped raw material may be inserted at a position at least 2 cm away. preferable. On the other hand, if it is too close to the quartz crucible, there is a possibility of contact due to deformation of the quartz crucible. Therefore, it is preferable that the distance between the inner wall of the quartz crucible and the outside of the rod-shaped raw material is at least 2 cm.

  As described above, the rod-shaped raw material may include a dopant or a non-doped raw material. Moreover, the melting mass of the rod-shaped raw material within a certain time is not particularly limited. For example, it is preferable that the rod-shaped raw material is a non-doped raw material or a raw material having the same conductivity type as that of the initial melt, and the dissolution ratio α of the rod-shaped raw material is (1−k) / 10 ≦ α ≦ 1. Here, α is the mass of the rod-shaped raw material dissolved in the melt with respect to the mass crystallized from the melt within a certain time. k is the segregation coefficient of the dopant introduced into the melt for resistivity control.

  In this case, the conductivity type of the rod-shaped raw material is the same as the conductivity type of the initial melt, the dopant concentration of the rod-shaped raw material is within ± 10% with respect to the initial melt concentration, and the dissolution ratio α of the rod-shaped raw material is 0.9 ≦ More preferably, α ≦ 1.1 or less. Furthermore, it is particularly preferable that the dopant concentration of the rod-shaped raw material is equal to the initial melt concentration, and the dissolution ratio α of the rod-shaped raw material is 1.

  On the other hand, the rod-shaped raw material may be a non-doped raw material, and the dissolution ratio α of the rod-shaped raw material may be between (1-k) / 10 and 1-k.

  Since there may be not only one rod-shaped raw material but also a plurality of rod-shaped raw materials, the mass of the rod-shaped raw material dissolved in the melt at the dissolution ratio α is the sum of the dissolved masses of all rod-shaped raw materials. In addition, the fixed time at the dissolution ratio α may be very short when the rod-shaped raw material is continuously added, and is basically the moment. On the other hand, when inserting intermittently, it is a time longer than at least the insertion interval. Momentarily, α may be 0 or may exceed 1. In particular, when the number of times of charging is small, it may be necessary to take from the start to the end of crystal growth.

  Here, the reason why the dissolution ratio is (1-k) / 10 ≦ α ≦ 1 will be described. In the following description, when the conductivity type of the rod-shaped raw material is the same as the conductivity type of the initial melt and the dopant concentration of the rod-shaped raw material is about the same as the initial melt concentration, the dissolution ratio α of the rod-shaped raw material is about 1. The reason why it is preferable and when the rod-shaped raw material is a non-doped raw material will be described together with the reason why the dissolution ratio α of the rod-shaped raw material is preferably between (1-k) / 10 and 1-k.

  One of the merits of the continuous charge method in which raw materials are continuously added during crystal growth is the above-described uniform resistance, and the other is the uniform thermal environment. When the raw material of the same amount as the crystallized mass is dissolved, that is, α = 1, the melt amount is constant and the height of the melt surface is constant, that is, the melt accompanies crystal growth as in the normal CZ method. The height position of the surface does not decrease. Therefore, it is not necessary to keep the melt surface at a substantially constant height by raising the crucible. For this reason, it is possible to keep the crucible position and the heater heat generation state in the same state with respect to the crystal length direction. Therefore, it becomes easy to obtain crystals of uniform quality. In this case, it is most preferable that α = 1.

  At this time, by dissolving the rod-shaped raw material having the same conductivity type and the same concentration as the dopant concentration in the initial melt determined from the raw material melted before the start of crystal growth and the amount of added dopant (main dopant), The resistivity can be kept constant. Of course, dissolving the rod-shaped raw material with the same concentration may mean simultaneously dissolving a plurality of raw materials such as a low-concentration raw material and a high-concentration raw material so that the average value is equal to the initial melt concentration.

  Although the above is the most preferable production method, since the present invention dissolves the rod-shaped raw material beside the growing crystal, in order to dissolve at a dissolution ratio α = 1, many rod-shaped raw materials are dissolved simultaneously. It can be difficult. Therefore, it is preferable to add a non-doped rod-shaped raw material containing no dopant at a dissolution ratio α = (1−k). Thereby, the resistivity in the crystal length direction, which is the most important purpose, can be kept constant.

  However, in this case, α is not 1, and the amount of melt decreases. Therefore, it is necessary to raise the crucible in order to keep the melt surface height constant. Therefore, although the thermal environment cannot be kept constant, it is in a direction in which it is made uniform as compared with the case of a normal CZ method in which no heat is input. Of course, if it is not a rod-shaped raw material containing no dopant at all, but a dopant having the same conductivity type as that added to the initial melt, α is adjusted to be larger than (1-k) to 1 By doing so, the resistivity in the length direction can be kept constant. At this time, it is preferable that the ascending speed of the crucible is controlled according to the value of α so that the melt surface height is maintained at a substantially constant height.

  Furthermore, even if the resistivity is not constant in the length direction, the length of the product portion can be increased and the yield can be greatly improved by suppressing the decrease in resistivity to some extent. In particular, in N-type crystals where the segregation coefficient of the dopant is small and the resistivity is drastically reduced, a diameter-enlarged portion (cone portion) produced before reaching the product portion and a diameter-reduced portion (rounded portion) formed after the product portion. ) And other non-product parts are high, so even if the product part is made several tens of percent longer, the effect is great. In order to enjoy such an effect, it is preferable that α ≧ (1-k) / 10.

  From the above viewpoint, it is preferable that the dissolution ratio α of the rod-shaped raw material is between (1−k) / 10 and 1. Here, the same conductivity type as the main dopant initially introduced with the dopant in the rod-shaped raw material was assumed, from non-doping to a concentration equivalent to the initial melt concentration. However, the present invention is not limited to this, and the present invention can be applied even when the dopant concentration in the rod-shaped raw material is higher than this, or when sub-dopants having different conductivity types are included. Furthermore, it can be applied to the case where a solidified layer is formed in the melt. Of course, in this case, it may be considered that the optimum range protrudes from the dissolution ratio (1-k) / 10 ≦ α ≦ 1. Therefore, more accurately, it is important to determine the optimum value of α according to the average concentration of the rod-shaped raw material to be charged and the state of the solidified layer formed in the melt, but it is approximately (1-k) / 10. It is sufficient if ≦ α ≦ 1.

In the present invention, the growth rate of the crystal being grown is not particularly limited, but the growth rate V (mm / min) of the crystal being grown is preferably 0.01 ≦ V ≦ 15000 / D ² +0.45. Here, D is the crystal diameter (unit: mm) during growth.

  The reason why the continuous charging of the rod-shaped raw material disclosed this time was not performed 15-30 years ago, in which many patent documents related to continuous charging can be cited, is considered to be one of the causes. At that time, device technology and quality evaluation technology were not as advanced as they are now, so the quality of silicon single crystals was not as important as it is now, and productivity and yield were emphasized. For this reason, the crystal growth rate was generally high. In order to increase the growth rate, it is necessary to cool the crystal, and the crystal is grown with a lower heater power. For this reason, the melt temperature in the crucible becomes lower, and if a rod-shaped raw material as disclosed herein is inserted beside the crystal, solidification occurs and it may be difficult to make a single crystal.

  However, along with the advancement of devices and evaluation techniques, reduction of Grown-in defects is required, and accordingly, the crystal growth rate has been reduced. This increased the heater power and the melt temperature in the crucible. As a result, even when the rod-shaped raw material is inserted beside the crystal, solidification does not occur, and a single crystal can be obtained.

As described above, the present invention is preferably applied to an operation in which a Grown-in defect is reduced. The range of the growth rate depends on the diameter of the crystal, and is approximately 15000 / D ² +0.45 or less. On the other hand, if the growth rate is 0.01 mm / min or more, productivity can be maintained. Therefore, the range to which the present invention is applied (the range of the growth rate of the growing crystal) is preferably 0.01 ≦ V ≦ 15000 / D ² +0.45.

  Further, the rod-shaped raw material may be inserted during crystal growth, and the insertion timing is not particularly limited. For example, the insertion of the rod-shaped raw material may be started after growing for the slipback length from the start of the growth of the product portion in the crystal being grown.

  Usually, in the production of a single crystal in the CZ method or the MCZ method, single crystallization is inhibited for some reason, and dislocations often occur. When dislocations are generated, dislocations generated at the dislocations are slip-backed to a portion that has already been single-crystallized but is still in a high temperature state. For this reason, a single crystal can be obtained only by the length obtained by subtracting the slip-back length from the length at which dislocations have occurred. In addition, the top part of the crystal has a quality other than resistivity, such as oxygen concentration and grown-in defect characteristics, which do not fall within the standard values, so the part excluding a certain length from the top part is set for the product. Has been. Therefore, when the crystal is dislocated with a relatively short length, the product cannot be obtained at all due to the portion not suitable for the top product and the slipback. For this reason, dislocation crystals are submerged in the melt and remelted, and single crystal production is performed again.

  For example, when the melting of the rod-shaped raw material having a dopant concentration lower than the initial melt concentration is started together with the start of the growth of the straight body portion, dislocation occurs in a portion where the length of the straight body is relatively short, and then remelted, it has been dissolved until then. The dopant concentration in the melt is lowered by the amount of the rod-shaped raw material. As a result, the resistivity becomes higher than intended. Therefore, the insertion of the rod-shaped raw material may be started after the growth of the slip-back length from the start of the growth of the product portion in the growing crystal. Furthermore, it is more preferable to start the insertion within the range where the resistivity satisfies the standard with only the initial melt (that is, the insertion of the rod-shaped raw material before the resistivity becomes lower than the standard). By doing in this way, even when dislocation is formed in a portion where the length of the straight body is relatively short, it can be remelted as usual.

  After the crystal is grown as described above, the raw material (the raw material here is not limited to the rod-shaped raw material) is recharged before the next crystal growth, and the crystal growth method of the present invention is used again in the next crystal growth. Then, the next crystal is grown, that is, the rod-shaped raw material is brought into contact with the melt between the growing crystal and the crucible wall while being solid, and the rod-shaped raw material is continuously or intermittently inserted and dissolved. A plurality of crystals may be grown from one quartz crucible by growing the next crystal.

  As described above, when the dissolution ratio α = 1, the amount of melt does not decrease. Therefore, after growing one crystal, it is possible to start growing the next crystal without recharging the raw material. However, the present invention includes dissolving the rod-shaped raw material at a ratio smaller than the dissolution ratio of α = 1. In this case, after the growth of one crystal, the melt amount is reduced compared to the initial stage. In this case, the raw material and the necessary dopant can be added and melted by recharging to prepare for the next crystal pulling. Thereby, it can be grown while inserting the rod-shaped raw material in the next crystal growth. Of course, the dissolution ratio α and the target resistivity can be changed for each crystal. By repeating such a process, it is possible to perform multi-pooling for growing a plurality of crystals from one quartz crucible.

  EXAMPLES Hereinafter, although an Example and a comparative example are shown and this invention is demonstrated more concretely, this invention is not limited to this Example.

Example 1
An N-type crystal having a resistivity of 50 Ωcm ± 7% and a diameter of 206 mm was grown using the crystal growth apparatus whose schematic diagram is shown in FIG. The initial charge amount of the raw material was 200 kg, and a crystal having a straight body length of 100 cm was grown by applying a horizontal magnetic field having a central magnetic field strength of 4000 G. The average single crystal growth rate was approximately 0.7 mm / min. Since the crystal top portion may not have quality other than resistivity, P (phosphorus) was initially doped with the aim of a straight body of 10 cm and an upper limit of 53.5 Ωcm.

  Further, four 4 cm prismatic rod-shaped raw materials were mounted on an insertion machine and inserted during straight body growth. The rod-shaped raw material was produced by cutting out from a single crystal having a high resistivity of 3000 Ωcm or more, which is hardly doped with a dopant. Of course, the crystal from which the rod-shaped raw material is cut does not have to be a single crystal, and polycrystalline silicon can also be used.

  The insertion speed of the rod-shaped raw material was controlled in synchronization with the single crystal growth speed, and was 1.5 times the single crystal growth speed. The dissolution ratio α (= raw material dissolution mass / crystallization mass) is about 0.288. Insertion of the rod-shaped raw material started at the same time as the growth of the straight body. In addition, the gate valve attached to the insertion machine was closed on the way, and the crystal was grown while adding the rod-shaped raw material. The resistivity at the center of the crystal thus obtained is shown in FIG.

(Example 2)
After growing the crystal in Example 1, the raw material was added by recharging, the raw material in the crucible was 200 kg, and the second crystal was grown by applying a horizontal magnetic field with a central magnetic field strength of 4000 G. The target resistivity is the same as in Example 1, N-type, 50 Ωcm ± 7%. This time, after the growth of the slip portion from the start of the growth of the product part in the crystal being grown, the rod-shaped raw material is added from the product start length (10 cm) + slip back length (about 20 cm) = 30 cm. It was decided to. Therefore, the dopant concentration in the initial melt was additionally doped at the time of recharging so that the concentration was 53.5 Ωcm at the product starting length of 10 cm, that is, the same concentration as when no raw material was added.

  The crystal diameter, crystal length, growth rate, rod-shaped raw material shape / number and material were the same as those in Example 1. However, since the addition of the rod-shaped raw material is started in the middle of the product portion, the product portion length that satisfies the resistivity standard is shorter than that in the first embodiment. Therefore, the insertion speed of the rod-shaped raw material was slightly higher than that of Example 1 and was 1.6 times the single crystal growth speed. The dissolution ratio α (= raw material dissolution mass / crystallization mass) is about 0.307. The resistivity at the center of the crystal thus obtained is shown in FIG.

(Comparative example)
Crystal pulling was performed under the same conditions as in Example 2 except that the rod-shaped raw material was not inserted. The resistivity at the center of the crystal thus obtained is shown in FIG.

  FIG. 3 is a diagram showing the results of evaluating the axial resistivity distribution of the crystals grown in Examples 1 and 2 and the comparative example. In FIG. 3, the vertical axis represents resistivity (Ωcm), and the horizontal axis represents crystal length (cm). As shown in FIG. 3, the resistivity at the center of the crystal obtained in Example 1 was within the target specification from a straight cylinder of 10 cm to a straight cylinder of about 97 cm. Further, in Example 2, the resistivity at the center of the obtained crystal is as shown in FIG. 3, and it was within the target specification from the straight cylinder of 10 cm to the straight cylinder of about 80 cm. On the other hand, as shown in FIG. 3, in the comparative example in which the rod-shaped raw material was not inserted, the straight body from 10 cm to about 50 cm was within the target standard. However, this length was significantly shorter than in Examples 1 and 2.

  The present invention is not limited to the above embodiment. The above-described 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 any device that exhibits the same function and effect is the present invention. It is included in the technical scope of the invention.

1 ... main chamber, 2 ... pulling chamber, 3 ... crystal rod,
4 ... Melt (raw material melt), 5 ... Quartz crucible, 6 ... Graphite crucible,
7 ... Heating heater (main heater), 8 ... Heat insulation member, 9 ... Gas outlet,
10 ... Gas inlet, 11 ... Top chamber, 12 ... Gas purge cylinder,
13 ... Heat shield member, 14 ... Inserting machine, 15 ... Rod-shaped raw material,
100: Crystal growth apparatus (pulling furnace).

Claims (10)

In the CZ method, a crystal is grown while supplying a raw material to a melt in a crucible during crystal growth, wherein a rod-shaped raw material is used as the raw material, and the rod-shaped raw material is grown as a solid and a crucible. The rod-shaped raw material is continuously or intermittently inserted into and melted between the walls, and the crystal is grown while being melted, and the melt is heated using a heater arranged around the crucible. And, the upper end of the heat generating part of the heater is above the molten metal surface that is the surface of the melt to give radiant heat from the heater to the rod-shaped raw material ,
The rod-shaped raw material is a non-doped raw material, and the dissolution ratio α of the rod-shaped raw material is between (1-k) / 10 and 1-k (where α is a rod relative to the mass crystallized from the melt within a certain time And a k is a segregation coefficient of a dopant introduced into the melt for resistivity control.) .   When inserting the rod-shaped raw material into the pulling furnace, using an insertion machine having a movable part capable of moving the rod-shaped raw material up and down, and inserting the rod-shaped raw material into the pulling furnace by the movable part. The crystal growth method according to claim 1, wherein the crystal growth method is characterized.   When the rod-shaped raw material is added, the addition is performed by using the one having an insertion portion having a gate valve capable of isolating the inside of the pulling furnace from the outside of the pulling furnace as the inserting machine. The crystal growth method according to claim 2, which is performed after isolating the inside of the furnace from the outside of the furnace. 4. The crystal growth method according to claim 1, wherein a cross-sectional area of the rod-shaped raw material is 1/10 or less and 10 mm ² or more of a crystal to be grown. 5.   5. The crystal growth method according to claim 1, wherein the number of the rod-shaped raw materials is plural.   The crystal growth method according to any one of claims 1 to 5, wherein a magnetic field having a central magnetic field strength of 500 to 6000 G is applied to the melt.   The crystal growing method according to claim 1, wherein a solidified layer is formed in the melt. The growth rate V (mm / min) of the growing crystal is set to 0.01 ≦ V ≦ 15000 / D ² +0.45 (where D is the diameter of the growing crystal (unit: mm)). The crystal growth method according to any one of claims 1 to 7 , wherein the crystal growth method is performed. The insertion of the rod-shaped raw material is started after growing for the slipback length from the start of the growth of the product part in the crystal being grown, The method according to any one of claims 1 to 8 , Crystal growth method. After the crystal is grown, the raw material is recharged before the next crystal growth, and the next crystal is grown again by using the crystal growth method according to any one of claims 1 to 9 in the next crystal growth. A crystal growing method characterized by growing a plurality of crystals from one quartz crucible by growing.

Images