JP4838591B2 - Silicone coagulation mold and method for producing the same - Google Patents

Description

  The present invention relates to a silicon solidification mold used for solidifying a silicon melt and a method for producing the same, and more specifically, a mold used to obtain polycrystalline silicon for producing a substrate for solar cells, and the like. It relates to the manufacturing method.

  Silicon (Si) used for solar cells generally requires a purity of about 99.9999% (6N), various metal impurities are 0.1 mass ppm or less, and boron (B) is at least 0. .3 mass ppm or less is required. As silicon satisfying this condition, there is silicon for semiconductors manufactured using the Siemens method, that is, high-purity silicon obtained by pyrolyzing silicon chloride after distillation. However, since this Siemens method is expensive, it is unsuitable for producing silicon for solar cells, which requires a large amount of silicon and requires low cost.

  Accordingly, various techniques for producing inexpensive silicon that can be used for solar cells have been studied, and various metal impurities such as Fe, Al, and Ca except for B and P are generally removed by a unidirectional solidification method. It is done. This unidirectional solidification method utilizes the phenomenon that when a molten silicon melt is solidified, a large amount of metal impurities are distributed to the coexisting molten silicon, and only a small amount of metal impurities are taken into the solidified silicon. is there. In various techniques for manufacturing inexpensive silicon that can be used in solar cells, including this unidirectional solidification method, a process of melting and solidifying silicon is often indispensable.

  Conventionally, when a silicon melt is solidified, a quartz mold or a graphite mold, which is less likely to be mixed with impurities, is mainly used. However, when a silicon melt is poured into these molds, there is a problem that the solidified silicon lump is fixed to the mold and the silicon recovery yield is lowered. Quartz and graphite molds are usually very expensive because they require work by specialists for production and require a long production time of about 2 weeks to 1 month. There is also a problem that it is uneconomical to dispose of it without being able to use it repeatedly.

  Therefore, in order to solve the above problems, a mold made of graphite that can be assembled and disassembled, and a mold release agent is applied to the inner wall surface of the mold that comes into contact with the silicon melt to form a mold release layer. Thus, a method has been examined in which a silicon lump is taken out by decomposing the mold after the silicon is solidified. For example, a method has been proposed in which a silicon lump is taken out without damaging the mold by coating the inner wall surface of a graphite assembly mold with silicon oxide, nitride, or carbide (Patent Documents 1 and 2). . However, this (Patent Documents 1 and 2) does not describe specific means for coating the inner wall surface of a graphite mold with silicon oxide or the like.

  Further, a method has been proposed in which a mold release layer is formed by coating the inner wall surface of a mold with a slurry in which silicon nitride powder and an organic binder are dissolved in a solvent (Patent Document 3). In this method (Patent Document 3), polyvinyl alcohol, polyvinyl acetate or polyvinyl butyrate is exemplified as the organic binder (Japanese Patent Publication No. 4, column 58 to 60). These are release layers at a high temperature below the melting point of silicon. Among them, polyvinyl alcohol is preferable because it is desorbed at a low temperature (column 5, lines 9 to 11). However, in reality, when the organic binder is detached from the release layer, the remaining silicon nitride powders are very loosely sintered, so the strength of the release layer itself is weak and the release layer is damaged. There is a danger to do. If the release layer is damaged, there is a problem that silicon adheres to the mold for that portion. In order to increase the strength of the release layer itself, it may be possible to select a powder that sinters more strongly than the silicon nitride powder. However, if the sintering becomes stronger, the release layer shrinks due to the sintering. This is not practical because it may cause cracks as a result. That is, the fact is that a powder that can be easily sintered cannot be used.

  Furthermore, it is possible to prevent the silicon from sticking by improving the particle size of the powder contained in the release agent, improving the composition ratio of the release agent, or adding a dispersing agent, and enabling repeated use of the mold. Has been proposed (Patent Documents 4 to 12). However, all of these methods essentially use a powder and an organic binder, and there remains a risk that the organic binder used is detached from the release layer at a high temperature. For example, the above (Patent Document 11) is characterized in that the binder is degreased at 450 to 600 ° C., and the binder is preferably polyvinyl alcohol as described above. Further, from the viewpoint of preventing carbon uptake into silicon, it is preferable that the binder is completely removed from the release layer (paragraphs [0007] to [0011] of Patent Document 11). However, if the binder is removed, of course, the strength of the release layer itself is lowered and may be damaged. If even a part of the release layer is broken, silicon adheres to the mold from there, and the yield of silicon recovery decreases, and it becomes difficult to repeatedly use the mold targeted by these methods.

Furthermore, when a release agent is applied to the inner wall surface of the mold to provide a release layer, the density of the release layer is higher on the side in contact with the silicon melt than on the mold side (Patent Document 13). ), A base layer made of a release agent obtained by adding metal silicon to silicon nitride, silicon oxide, or a mixture thereof, and a surface layer formed by a release agent made of silicon nitride, silicon oxide, or a mixture thereof. A method of providing a release layer having a two-layer structure (Patent Document 14) has been proposed. However, both of these methods require the formation of a release layer by plasma irradiation, and thus have a problem of high cost, inferior workability, and use of an expensive graphite mold.
JP 62-108515 A JP 62-260710 A US Pat. No. 5,431,869 JP-A-6-144824 Japanese Patent Laid-Open No. 7-206419 JP-A-9-175809 Japanese Patent Laid-Open No. 10-182133 JP 2001-198648 A JP 2002-239682 A JP 2002-292449 A Japanese Patent Laid-Open No. 2002-321037 JP 2003-64388 A JP 2005-46851 A JP 2005-46866 A

  Originally, when molten silicon is poured into a mold, the silicon melt collides with the inner wall surface of the mold, so when providing a mold release layer on the mold, the mold release layer itself is strong and the mold It is required to be firmly attached to the surface. However, as described above, various studies have been made on the release layer, but these are still not sufficient. Therefore, the present inventors can eliminate the possibility of impurities being mixed into silicon at the same level as quartz molds and graphite molds, even without a release layer, and more easily and quickly. As a result of intensive studies on molds that can be made, it is possible to strengthen the surface in contact with the silicon melt by forming the mold body substantially from silica glass and a phenolic resin cured product, and to apply the silicon melt. It is difficult to prevent silicon from sticking, and the impurity diffusivity to silicon is at least equivalent to that of quartz or graphite molds. Furthermore, compared to quartz molds and graphite molds The present inventors have found that a template can be obtained easily and in a short period of time.

  Accordingly, the object of the present invention is to prevent the silicon from adhering even without providing a release layer, and to eliminate the possibility of impurities being mixed into the silicon at the same level as a quartz mold or a graphite mold. In addition, another object of the present invention is to provide a silicon coagulation mold that can be manufactured easily and in a short period of time without requiring any work by a specialist.

  Another object of the present invention is to prevent silicon from sticking even without providing a release layer, and at the same level as the quartz mold and graphite mold to prevent impurities from entering silicon. Another object of the present invention is to provide a method for producing a silicon coagulation mold that can be eliminated and can be obtained easily and in a short period of time without requiring any special work.

  That is, the present invention is a mold for use in solidifying a silicon melt, wherein the mold main body is substantially made of silica glass and a phenol resin cured product.

  The present invention also relates to a method for producing a mold used for solidifying a silicon melt, wherein a mold material in which a phenol resin and silica glass are mixed is heat-cured at a temperature of 200 to 250 ° C. to obtain a predetermined shape. This is a method for producing a mold for silicon coagulation, characterized in that a mold body is obtained.

Below, while explaining the manufacturing method of the silicon coagulation mold of the present invention, the silicon coagulation mold will be described.
In the present invention, a mold body having a predetermined shape is obtained by thermosetting a mold material obtained by mixing a phenol resin and silica glass at a temperature of 200 to 250 ° C. About the silica glass which forms a casting_mold | template material, it is preferable that 95 mass% or more of silica glass with a particle size of 45-150 micrometers is contained with respect to all the silica glasses. If the particle size of the silica glass is smaller than 45 μm, the mixing with the phenol resin may not be uniform. Conversely, if the particle size is larger than 150 μm, the void on the surface of the mold body is large when the mold body is obtained, resulting in a porous shape. Prone. Therefore, when the silicon melt is poured, the silicon melt is impregnated and easily fixed. Further, when the silica glass having a particle size of 45 to 150 μm is less than 95% by mass with respect to the total silica glass, it is difficult to obtain a uniform mixture of silica glass and phenolic resin in forming the mold material, and the mold body When forming the film, the voids on the surface may become large and become porous, which is not preferable.

  Moreover, about the said silica glass, it is still more preferable that the silica glass with a particle size of less than 45 micrometers is contained less than 5 mass% with respect to the total silica glass. When the mold body obtained by the present invention actually melts the silicon melt to solidify the silicon, the mold body is usually preheated to about 1550 ° C. in advance, but when the mold body reaches about 1300 ° C., As shown in FIG. 1B, a necking phenomenon in which adjacent silica glasses having a particle diameter of 45 to 150 μm are bonded to each other (this temperature may be referred to as “necking start temperature”). At this time, if silica glass having a particle size of less than 45 μm is present, the silica glass having a particle size of less than 45 μm enters the gap between the silica glasses having a particle size of 45 to 150 μm (state shown in FIG. 1A), and assists in necking. Functions as a material. That is, a stronger necking can be obtained. When the silica glass having a particle size of less than 45 μm is contained in an amount of 5% by mass or more based on the total silica glass, the silica glass is excessively sintered and the mold body shrinks when the mold body is preheated. On the other hand, when the silica glass having a particle size of less than 45 μm is less than 3% by mass with respect to the total silica glass, it hardly acts as a necking auxiliary material. In FIG. 1, 1a represents silica glass having a particle size of 45 to 150 μm, and 1b represents silica glass having a particle size of less than 45 μm.

  Moreover, in this invention, it is preferable that casting_mold | template material contains 1-10 mass parts phenol resin with respect to 100 mass parts of silica glass. If the phenol resin content is less than 1 part by mass, there may be a problem in moldability of the mold body. Conversely, if it exceeds 10 parts by mass, the silica glass necking does not occur efficiently, and the silicon melt There is a possibility that a problem may occur in the shape retention when the silicon is poured to solidify the silicon.

  As the phenol resin used for the mold material, a commonly used known resin can be used, which may be a so-called novolac type or a resol type. This phenol resin may be mixed at room temperature as a mold material, or may be mixed as a mold material in a phenol resin solution containing a commonly used solvent.

  When the phenol resin forming the template material is a phenol resin solution containing a solvent, the viscosity of the phenol resin solution is preferably 2 to 10 Pa · S at 25 ° C. That is, when the viscosity is less than 2 Pa · S, it means that the amount of the solvent is relatively large with respect to the phenol resin. Will take more time than necessary. On the other hand, if it exceeds 10 Pa · S, the viscosity is too high and it may not be mixed with silica glass uniformly.

  As for the phenolic resin, in the case of a solid phenolic resin, the fixed carbon in the phenolic resin is preferably 80% by mass or more. On the other hand, when the phenolic resin solution is used, The fixed carbon is preferably 40% by mass or more. As will be described later, when the mold material is heat-cured at a predetermined temperature, the phenol resin is cured to become a phenol resin cured product, which is considered to form a three-dimensional network. And, when the mold main body becomes around 1000 ° C. by preheating at the time of use, it is considered that carbonization of the phenol resin cured product starts this time. Therefore, when the proportion of fixed carbon is less than the above range, the amount of phenol resin cured product remaining when the mold body is further heated to around 1300 ° C. by preheating at the time of use (that is, the residual carbon ratio). There is a risk that the sintering of the silica glass proceeds excessively, and the mold body shrinks and deforms. The amount of fixed carbon can be determined by differential thermal / thermogravimetric analysis.

  In the present invention, when the phenol resin used for the mold material is a phenol resin that does not cure itself even when heated, the mold material contains 20 parts by mass or more of a curing agent with respect to 100 parts by mass of the phenol resin. You may make it add. If the amount of the curing agent is less than 20 parts by mass, all of the phenolic resin may not be cured. About the kind of hardening | curing agent, if used as a hardening | curing agent of a phenol resin, there will be no restriction | limiting, A hexamine can be mentioned as a specific example. In addition, since the effect will be saturated when the quantity of the hardening | curing agent to add exceeds 30 mass parts, it is good to make this value into an upper limit.

  In the present invention, the mold material is heat-cured at a temperature of 200 to 250 ° C. to obtain a mold body having a predetermined shape. When the temperature of the thermosetting treatment is lower than 200 ° C., the phenol resin in the mold material may not be sufficiently cured by heat, and the effect is saturated even when the temperature is higher than 250 ° C. In addition, what is necessary is just to select suitably about the time of a thermosetting process according to the shape etc. of the mold main body to form.

  The phenol resin that forms the mold material by the thermosetting treatment is thermoset to a phenol resin cured product (when a curing agent is used, the phenol resin and the curing agent react to form a phenol resin cured product). It is considered to form a three-dimensional network. On the other hand, regarding silica glass, it is considered that no change such as necking occurs at the temperature of the thermosetting treatment. That is, the mold main body obtained by the thermosetting treatment is substantially composed of silica glass and a phenol resin cured product. Here, substantially consisting of silica glass and phenolic resin cured product includes phenolic resin and curing agent remaining without being thermally cured to the extent that it does not affect the mold retainability of the mold body. It means that it may be. Further, it means that the third component may be included as long as the mold body does not have an influence on the effects of the present invention. Even if the phenol resin that is not thermoset remains to the extent that it does not affect, there is no particular problem because the mold body is preheated when the mold body is used.

  Although partially overlapping with the contents already described, the preferred form of the mold body of the present invention obtained by thermosetting treatment consists of 1 to 10 parts by mass of a phenol resin cured product with respect to 100 parts by mass of silica glass. The silica glass having a particle size of 45 to 150 μm is preferably contained in an amount of 95% by mass or more based on the total silica glass, and the silica glass having a particle size of less than 45 μm is contained in an amount of less than 5% by mass based on the total silica glass. Good.

In the present invention, for example, the mold material is molded by the following method and then heat-cured.
For example, a mold material is filled in a gap between a box-shaped form formed in a box shape with an open top and a core accommodated in the box-shaped form frame, and the mold material is held in a predetermined shape. In addition, if a thermosetting process is performed, a mold body having a predetermined shape can be molded. At this time, the shape of the box-shaped form and the core may be appropriately designed according to the shape of the required mold body.

  As another example, the required number of plate-shaped mold members having a predetermined shape can be produced by thermosetting a mold material for a plate that can be molded into a plate mold. If the plate-shaped mold member is assembled in a box shape so as to correspond to the bottom surface portion and the side wall portion of the mold body, a mold body having a predetermined shape can be obtained. At this time, the shape of the plate-shaped mold member and the required number of sheets may be appropriately designed according to the required shape of the mold body.

  Examples of the material of the box-shaped formwork, core, and plate formwork include commonly used molds, synthetic resin molds, etc., but from the viewpoint of dimensional accuracy and contamination stability, A synthetic resin type is preferable. Note that these molds are generally split molds, and a release agent such as tetrafluoroethylene may be spray coated to facilitate demolding. Alternatively, the mold itself may be formed from tetrafluoroethylene.

  Moreover, about the thickness of the side wall part which forms the side surface of the obtained casting_mold | template main body, it is good to set it as 10-15 mm preferably. If the thickness is less than 10 mm, the heat removal from the side surface of the mold becomes large, and it becomes difficult to obtain silicon having a columnar crystal as a product. On the other hand, about the thickness of the bottom face part used as the bottom face of a casting_mold | template main body, it is good to set it as 2-5 mm. If it is thinner than 2 mm, it will be weak in strength, and if it exceeds 5 mm, it will be difficult to remove heat.

  In the present invention, a silicon nitride solution containing silicon nitride is applied to at least the inner wall surface (inner surface of the side wall portion and the bottom surface portion) of the mold main body that is in contact with the silicon melt, and is substantially nitrided by drying. A release layer made of silicon may be formed. Since the mold body itself has poor paintability with the silicon melt, it is possible to sufficiently prevent the silicon from adhering after the silicon is solidified, but it is possible to release the silicon more reliably by providing a release layer. In addition, the release layer formed on the inner wall surface of the mold of the present invention has a good affinity and is not easily damaged. Even if the mold release layer breaks and the silicon melt is impregnated on the mold body side, there is no problem because the mold body itself can prevent the silicon from sticking. The effect which can prevent can be improved further. It should be noted that there is no problem even if the release layer contains a component other than silicon nitride as long as it does not affect the contamination of silicon with impurities.

  Examples of the solvent used as the silicon nitride solution include commonly used solvents such as water and ethanol. The silicon nitride preferably has a particle size of 50 μm or less. If the particle size is larger than 50 μm, the voids in the release layer become too large and the silicon melt is easily impregnated. Therefore, when a release layer is provided, it is preferable to use silicon nitride having the above particle size in order to sufficiently exhibit the effect.

In the mold for solidifying silicon according to the present invention, the mold body is substantially composed of silica glass and a phenolic resin cured product, and silica glass has a relatively small coefficient of thermal expansion, so that it is exposed to a high temperature state when solidifying silicon. However, the surface in contact with the silicon melt can be maintained in a strong state. Further, the surface in contact with the silicon melt has poor paintability with the silicon melt and can prevent silicon from sticking. Furthermore, impurity contamination to silicon can be eliminated in the same manner as a quartz mold or a graphite mold.
In the present invention, a mold body having a predetermined shape can be obtained by heat-curing a predetermined mold material. Therefore, the mold body can be obtained in a short time and more easily than a quartz mold or a graphite mold. Thus, the mold can be manufactured at a lower cost than a quartz mold, a graphite mold, or the like.

Hereinafter, preferred embodiments of the present invention will be specifically described with reference to the accompanying drawings.
In FIG. 2, the mold material 2 is kept in a predetermined shape by using a box-shaped mold frame 3 and a core 4 that are formed in a box shape with an upper opening, and the mold material 2 is molded by thermosetting. It is sectional explanatory drawing which shows the procedure (outline) which obtains the casting_mold | template main body 5 by this. First, a predetermined amount of silica glass having a specific particle size and a phenol resin are prepared, and these are mixed for about 15 to 20 minutes using a mixer (not shown) to obtain a mold material 2. In addition, this casting_mold | template material 2 becomes a clay-like kneaded material normally. About the phenol resin used in this case, a solid resin may be used at room temperature. For example, a phenol resin solution containing a known solvent that is usually used such as phenol or ethylene glycol may be used. Good. Moreover, when using what a phenol resin does not self-cure, you may make it add hardening agents, such as hexamine, separately.

  Next, a part of the mold material 2 obtained above is laid on the bottom part of the box-shaped mold 3 so as to have an amount corresponding to the thickness of the bottom surface part 5a of the mold body 5 [FIG. )], The core 4 is placed on the mold material 2. At this time, a predetermined gap (a gap corresponding to the thickness of the side wall portion 5b of the mold body 5) is formed between the side surface of the core 4 and the inner wall surface of the box-shaped mold 3. Next, the mold material 2 is filled in the gap between the box-shaped mold 3 and the core 4 to maintain the shape of the mold material 2 so as to be the shape of the mold body 5 [state in FIG. 2 (2)] .

  The mold material 2 prepared as described above is placed in an oven (not shown) together with the box-shaped mold 3 and the core 4, and a thermosetting process is performed in which baking is performed at a temperature of 200 to 250 ° C. in an air atmosphere. The time for the thermosetting treatment is usually 4 to 6 hours, although it depends on the shape of the mold body 5. When the thermosetting treatment is completed, the mold is cooled to room temperature, and then the core 4 and the box-shaped form 3 are removed, whereby a silicon coagulation mold composed of the box-shaped mold body 5 can be manufactured [FIG. State 3)]. If necessary, a silicon nitride solution containing silicon nitride is applied to the inner wall surface of the mold body 5 by a known method such as brushing or spraying, and dried, so that the mold body 5 is substantially made of silicon nitride. It may be a silicon coagulation mold provided with a release layer 6.

  On the other hand, FIG. 3 shows a plate-shaped mold member 12 that is prepared by filling a mold material 2 into a mold 11 for a plate that can be molded into a plate-like body, and thermosetting. FIG. 5 is a cross-sectional explanatory view showing a process of manufacturing a mold body 13 by assembling the mold mold member 12 into a box shape. First, the mold material 2 obtained in the same manner as described above is filled in the plate mold 11 [the state shown in FIG. 3 (1)], and this is put together with the plate mold 11 in an oven not shown in the figure. The thermosetting process is performed. After completion of the thermosetting treatment, the plate mold member 11 can be obtained by cooling to room temperature and removing the plate mold 11. Similar operations are performed as many times as necessary, and a predetermined number (in this case, a total of 5) of plate-like mold members 12 are produced [state shown in FIG. 3 (2)].

  Next, if the five plate-like mold members 12 obtained above are assembled in a box shape so as to correspond to the bottom surface portion 13a and the side wall portion 13b of the mold main body 13 (state of FIG. 3 (3)), the mold main body 13 It is possible to manufacture a silicon coagulation mold made of At this time, the plate-shaped mold member 12 corresponding to the side wall portion 13b of the mold body 13 can maintain the assembled state in a box shape by its own weight, but the joined portion of the plate-shaped mold member 12 is in a liquid quick-drying property. It may be fixed using an adhesive such as phenol resin, or a plate corresponding to the side wall portion 13b after being assembled in a box shape so that the plate-shaped mold member 12 corresponding to the side wall portion 13b does not fall down. The mold member 12 may be fixed by being tied with a graphite string. Furthermore, as shown in the solid line in FIG. 3 (3), the joined portion of the plate-like mold member 12 is made complementary to each other, so that the assembled state can be easily maintained. You may do it. A mold release layer may be formed on the inner wall surface of the mold body 13 assembled in a box shape, as described in the case of FIG.

  By the way, when silicon is solidified by using the mold body, it is common to preheat to about 1450 ° C. in advance by disposing a resistance heater or the like outside the outer wall surface of the mold body. At this time, carbonization of the phenolic resin cured product forming the mold body starts around 1000 ° C., but most of it is considered to be residual carbon, and from about 1300 ° C., the residual carbon ratio gradually decreases. It is considered a thing. On the other hand, for silica glass, it is considered that necking in which silica glass is bonded to each other at about 1300 ° C. Therefore, as described with reference to FIG. 3, the plate-shaped mold member 12 is assembled into a box shape and the mold body 13 is assembled. In the case of forming, the joining portion of the plate-like mold member 12 is bonded by this silica glass necking. Therefore, when the mold body 13 is formed using the plate-shaped mold member 12, there is no particular problem as long as the box-shaped state can be maintained to some extent when the plate-shaped mold member 12 is assembled.

  By the way, in FIG. 2, the description will be made using the integrated mold body 5, and in FIG. 3, the description will be made using the assembled mold body 13 in which a total of five plate-shaped mold members 12 are prepared and assembled in a box shape. However, the mold body in the present invention is not limited to those having these shapes. In other words, the shape of the mold body is not particularly limited as long as the upper part can be opened and the silicon melt can be poured, for example, the bottom surface of the mold body is rectangular, circular, elliptical, It may have a shape such as a polygon of triangles or more. Moreover, also about the shape of the side wall part of a casting_mold | template main body, arbitrary shapes can be selected and it can design suitably so that the silicon lump obtained by solidifying may become a required shape.

[Mold production]
Silica glass, a phenol resin, and a curing agent were prepared so as to have the amounts shown in Table 1, respectively, and these were kneaded with a mixer for about 20 minutes to obtain a clay-like mold material 2. For silica glass, a silica glass having a particle size of 45 to 150 μm and a silica glass having a particle size of less than 45 μm are prepared using a sieve, and the silica glass having a particle size of 45 to 150 μm is 95% by mass with respect to the total silica glass. The silica glass having a particle size of less than 45 μm was adjusted to 5% by mass. Moreover, as the phenol resin, a phenol resin solution in which the phenol resin was dissolved in phenol was used. This phenol resin solution contained 40% by mass of fixed carbon and had a viscosity at 25 ° C. of 5 Pa · S. Furthermore, hexamine (purity 99.5%) was used as a curing agent, and the amount was 0.35 parts by mass with respect to 100 parts by mass of silica glass.

  Next, the mold material 2 obtained above was held in the shape of the mold body 5 using a box-shaped form 3 and a core 4 as shown in FIG. The box-shaped form 3 is made of Teflon (registered trademark) resin and is a box-shaped form having a cubic shape with an opening at the top. The opening is 400 mm × 400 mm and the depth is 400 mm (both inner diameters). is there. The core 4 is made of Teflon (registered trademark) resin, and is a cube having a side length and height of 370 mm. First, a part of the mold material 2 was laid on the bottom part of the box-shaped form 3 so as to have a thickness of 2 mm. Next, the core 4 was placed on the mold material 2, and a gap of 15 mm was provided between the side surface of the core 4 and the inner wall surface of the box-shaped mold 3. Next, the mold material 2 is filled so as to fill all the gaps formed between the side surface of the core 4 and the inner wall surface of the box-shaped mold 3, and the box-shaped mold 3 and the core 4 are not shown in the oven. The mold material 2 was heat-cured by heating at 200 ° C. for 4 hours.

  After completion of the thermosetting treatment, the product is allowed to cool to room temperature and the core 4 and the box-shaped form 3 are removed, so that the outer diameter is 400 mm long, 400 mm wide, 400 mm high, and the inner diameter is 370 mm × 370 mm × depth. A mold body 5 having a length of 398 mm was obtained, and a silicon solidification mold was completed. In Example 1, the time required for producing the silicon coagulation mold was 6 hours.

[Performance evaluation]
The mold body 5 obtained above was placed in a chamber having an induction heating type melting furnace, and the inside of the chamber was replaced with argon gas. After the inside of the chamber is replaced with the atmospheric pressure of argon gas, 1000 kg of high-purity silicon (99.999999999 mass% or more) is melted at 1550 ° C. in the melting furnace, and the mold body 5 is increased by 5 ° C. per minute with a resistance heater. Heat to 1550 ° C. at a warm rate. Next, 100 kg of molten silicon (silicon melt) in the melting furnace was decanted into the mold body 5, and then the temperature of the mold body 5 was gradually lowered to gradually solidify the silicon from below. After the entire silicon solidified, the resistance heater and other power sources were turned off and the furnace was cooled to room temperature. After cooling to room temperature, the mold body 5 and the solidified silicon were separated. As a result, no damage was found in the mold body 5 and there was no leakage of molten silicon. Further, when the solidified silicon was taken out, the mold body 5 and the silicon were easily peeled off, and there was no silicon sticking to the mold body 5, and a silicon mass of 370 mm × 370 mm × height about 310 mm could be obtained. (Recovery 99%).

In addition, a sample for analysis was collected from the central portion of the obtained silicon lump, and qualitative analysis was performed by ICP-AES (ICP-luminescence) analysis. As a result, Fe, Al, B, and P were not detected, and it was confirmed that there was no contamination with these metal impurities. The detection limit values in the ICP-AES analysis method are Fe: 0.02 mass ppm, Al: 0.02 mass ppm, B: 0.05 mass ppm, and P: 0.05 mass ppm.
The results of the performance evaluation are summarized in Table 2.

[Mold production]
A solid resin (85% fixed carbon) was used as the phenol resin, and the outer diameter was 400 mm × width in the same manner as in Example 1 using the amount of silica glass, phenol resin, and curing agent shown in Table 1. A mold body 5 having 400 mm × height 400 mm and an inner diameter of 370 mm × 370 mm × depth 398 mm was obtained. Next, a silicon nitride solution containing 50 parts by mass of silicon nitride having a particle size of 50 μm and 50 parts by mass of ethanol is formed on the inner wall surface of the mold body 5 (the inner surfaces of the bottom surface part 5a and the side wall part 5b of the mold body 5). It was applied with a brush so that the coating amount was about 0.3 mm thick, and dried at 25 ° C. for 1 hour to form a release layer 6 made of silicon nitride, thereby completing a silicon coagulation mold. In Example 2, the time required for producing the silicon coagulation mold was 6.5 hours.

[Performance evaluation]
The performance evaluation of the silicon solidification mold obtained above was performed in the same manner as in Example 1. As a result, no damage was observed in the mold body 5 after the solidified silicon was removed, and the molten silicon leaked. There was not. Further, when the solidified silicon was taken out, the mold body 5 and the silicon were easily peeled off, and there was no silicon sticking to the mold body 5, and a silicon mass of 370 mm × 370 mm × height about 310 mm could be obtained. (Recovery 99%). Furthermore, as a result of qualitative analysis by ICP-AES analysis, there was no contamination with impurity metals.
The results are shown in Table 2.

[Mold production]
Five plate molds 11 as shown in FIG. 3 were prepared. Four of them are plate molds 11 made of Teflon (registered trademark) having an inner diameter of 400 mm × 400 mm × depth of 15 mm, and a plate-shaped mold member 12 corresponding to the side wall portion 13b of the mold body 13 is produced. Is for. The remaining one is a plate mold 11 made of Teflon (registered trademark) having an inner diameter of 400 mm × 400 mm × depth 2 mm, for producing a plate-shaped mold member 12 corresponding to the bottom surface portion 13 a of the mold body 13. belongs to. Each of these plate molds 11 was filled with the mold material 2 prepared in the same manner as in Example 1, and the plate molds 11 filled with the mold material 2 were put together in an oven not shown in the figure. The mold material 2 was heat-cured by heating at 200 ° C. for 4 hours. After the thermosetting treatment, the plate-shaped mold member 12 is taken out from each plate mold 11 and the plate-shaped mold member 12 is assembled in a box shape so as to correspond to the bottom surface portion 13a and the side wall portion 13b of the mold body 13. . In this manner, a mold body 13 having an outer diameter of 400 mm × width of 400 mm × height of 400 mm and an inner diameter of 370 mm × 370 mm × depth of 398 mm was produced, and a silicon solidification mold was completed. In Example 3, the time required for producing the silicon coagulation mold was 6.5 hours.

[Performance evaluation]
The performance evaluation of the silicon solidification mold obtained above was performed in the same manner as in Example 1. As a result, no damage was observed on the mold body 13 after the solidified silicon was removed, and the molten silicon leaked. There was not. Further, when the solidified silicon was taken out, the mold body 13 and the silicon were easily peeled off, and there was no silicon sticking to the mold body 5, and a silicon mass of 370 mm × 370 mm × height about 310 mm could be obtained. (Recovery 99%). Furthermore, as a result of qualitative analysis by ICP-AES analysis, there was no contamination with impurity metals.
The results are shown in Table 2.

[Mold production]
Five plate molds 11 that are the same as those of Example 3 are prepared, and the same mold material as that of Example 2 is filled in the plate molds 11. As in Example 3, the outer diameter is 400 mm long × horizontal. A mold body 13 of 400 mm × height 400 mm and an inner diameter of 370 mm × 370 mm × depth 398 mm was produced. Next, the outer circumferences of the four plate-shaped mold members 12 corresponding to the side wall portions 13b of the mold body 13 assembled in a box shape were fixed with a string made of graphite. A spray is applied to the inner wall surface of the mold body 13 so that a silicon nitride solution containing 50 parts by mass of silicon nitride having a particle size of 50 μm and 50 parts by mass of ethanol is applied to a thickness of about 0.3 mm. Then, it was dried at 25 ° C. for 1 hour to form a release layer made of silicon nitride, thereby completing a silicon coagulation mold. The time required for producing the silicon coagulation mold of Example 4 was 7 hours.

[Performance evaluation]
The performance evaluation of the silicon solidification mold obtained above was performed in the same manner as in Example 1. As a result, no damage was observed on the mold body 13 after the solidified silicon was removed, and the molten silicon leaked. There was not. Further, when the solidified silicon was taken out, the mold body 13 and the silicon were easily peeled off, and there was no silicon sticking to the mold body 5, and a silicon mass of 370 mm × 370 mm × height about 310 mm could be obtained. (Recovery 99%). Furthermore, as a result of qualitative analysis by ICP-AES analysis, there was no contamination with impurity metals.
The results are shown in Table 2.

[Comparative Example 1]
A graphite mold body having the same shape as the mold body obtained in Example 1 was prepared. A silicon nitride solution consisting of 100 parts by mass of silicon nitride, 5 parts by mass of polyvinyl alcohol, and 100 parts by mass of pure water was applied to the inner wall surface of the graphite mold using a brush so that the thickness was about 0.3 mm. A release layer was formed by drying at 20 ° C. for 24 hours, and the mold according to Comparative Example 1 was completed. In order to manufacture the above-mentioned graphite mold main body, a dedicated manufacturing apparatus and an engineer are required. When the release layer is formed, it takes one week to complete the mold.

The performance evaluation of the mold according to Comparative Example 1 was performed in the same manner as in Example 1. When the solidified silicon cooled to room temperature and the graphite mold were separated, the mold release at the side wall portions of the mold was poor, and silicon was confirmed to adhere to the inner surface of each side wall portion. When the portion where the silicon was fixed to the mold side wall was observed, it was confirmed that the release layer was peeled off. As for the portion to which silicon is fixed, it is expected that when the temperature of the mold is raised, polyvinyl alcohol in the release layer is decomposed and vaporized, so that the release layer does not function. For this reason, damage to the mold was confirmed, and liquid leakage of molten silicon was also observed. Further, the obtained silicon lump could not be taken out in the state of a quadrangular prism, and silicon fixed to the mold could not be recovered (recovery rate 87%). In addition, as a result of the qualitative analysis by the ICP-AES analysis method, there was no contamination with impurity metals.
The results are shown in Table 2.

[Comparative Example 2]
A quartz mold body having the same shape as the mold body obtained in Example 1 was prepared. A silicon nitride solution consisting of 100 parts by mass of silicon nitride, 5 parts by mass of polyvinyl alcohol, and 100 parts by mass of pure water was applied to the inner wall surface of the graphite mold using a spray so that the thickness was about 0.3 mm. A release layer was formed by drying at 20 ° C. for 24 hours, and a mold according to Comparative Example 2 was completed. In order to manufacture the above-mentioned graphite mold main body, a dedicated manufacturing apparatus and an engineer are required. When the release layer is formed, it takes four weeks to complete the mold.

When the performance of the mold according to Comparative Example 2 was evaluated in the same manner as in Example 1, no damage was observed in the mold after the solidified silicon was removed, and there was no leakage of molten silicon. Further, when the solidified silicon was taken out, the mold and the silicon were easily peeled off, and the silicon was not fixed to the mold, and a silicon mass of 370 mm × 370 mm × height about 310 mm could be obtained (recovery rate 99). %). Furthermore, as a result of qualitative analysis by ICP-AES analysis, there was no contamination with impurity metals.
The results are shown in Table 2.

[Comparative Example 3]
A mold body having the same shape as the mold body obtained in Example 1 was produced using a silica-magnesia castable for general fire resistance. Further, a release layer was formed on the inner wall surface of the mold body in the same manner as in Comparative Example 1, and the mold according to Comparative Example 3 was completed. It took two days to produce this mold.

When the performance of the mold according to Comparative Example 3 was evaluated in the same manner as in Example 1, no damage was observed in the mold after the solidified silicon was removed, and there was no leakage of molten silicon. Further, when the solidified silicon was taken out, the mold and the silicon were easily peeled off, and the silicon was not fixed to the mold, and a silicon mass of 370 mm × 370 mm × height about 310 mm could be obtained (recovery rate 99). %). However, as a result of the qualitative analysis by the ICP-AES analysis method, silicon contains impurity metals at respective concentrations of Al: 200 mass ppm, P: 15 mass ppm, B: 10 mass ppm, and Fe: 150 mass ppm. Contamination was confirmed.
The results are shown in Table 2.

  The silicon solidification mold according to the present invention has a strong surface that contacts or collides with the silicon melt, and avoids impurity contamination to the same extent as a quartz mold or a graphite mold when silicon is solidified. In addition, since the surface of the mold main body has poor paintability with the silicon melt, it is possible to prevent silicon from being fixed as much as possible. Therefore, it is suitable as a mold for producing polycrystalline silicon used for a solar cell substrate or the like using a unidirectional solidification method or the like. Further, according to the present invention, since the above-mentioned silicon solidification mold can be manufactured easily and in a short period of time compared to a quartz mold or a graphite mold, it is also industrial in terms of cost and workability. Is extremely advantageous.

FIG. 1 (a) is a schematic diagram showing the state of silica glass forming the mold body of the present invention. FIG. 1 (b) is a diagram in which silica glass is poured into the mold body to solidify the silicon during use. It is a schematic diagram which shows a mode that it couple | bonds. FIG. 2 is a cross-sectional explanatory view showing a process (outline) of manufacturing a mold body by thermosetting a mold material using a mold and a core. FIG. 3 is a cross-sectional explanatory view showing a process (outline) of manufacturing a mold body by producing a plurality of plate-shaped mold members using a plate mold and assembling them into a box shape.

Explanation of symbols

1: Silica glass, 1a: Silica glass (particle size: 45 to 150 μm), 1b: Silica glass (particle size: 45 μm or less), 2: Mold material, 3: Box-shaped frame, 4: Core, 5, 13: Mold Main body, 5a, 13a: bottom portion, 5b, 13b: side wall portion, 6: release layer, 11: plate formwork, 12: plate mold member.

Claims (19)

  A mold for solidifying silicon melt, wherein the mold body is substantially composed of silica glass and a cured phenol resin.   The mold for silicon coagulation according to claim 1, wherein the mold body is composed of 1 to 10 parts by mass of a phenol resin cured product with respect to 100 parts by mass of silica glass.   The mold for solidifying silicon according to claim 1 or 2, wherein the side wall forming the side surface of the mold body has a thickness of 10 to 15 mm.   The silicon coagulation mold according to any one of claims 1 to 3, wherein the silica glass having a particle size of 45 to 150 µm is contained in an amount of 95 mass% or more based on the total silica glass.   The silicon coagulation mold according to claim 4, wherein the silica glass having a particle size of less than 45 µm is contained in an amount of less than 5 mass% with respect to the total silica glass.   The mold for silicon solidification according to any one of claims 1 to 5, wherein a release layer substantially made of silicon nitride is provided on an inner wall surface of the mold body.   The mold for solidifying silicon according to claim 6, wherein the silicon nitride has a particle size of 50 μm or less.   A method for producing a mold used for solidifying a silicon melt, wherein a mold material in which a phenol resin and silica glass are mixed is heat-cured at a temperature of 200 to 250 ° C. to obtain a mold body having a predetermined shape. A method for producing a mold for silicon coagulation characterized by the above.   The mold material is filled in the gap between the box-shaped mold formed in a box shape with the upper opening and the core housed in the box-shaped mold, and the mold body is thermoset treated. The method for producing a silicon coagulation mold according to claim 8.   A mold material for a plate capable of molding a plate-shaped body is filled with a mold material, and a predetermined number of plate-shaped mold members are produced by thermosetting the mold material, and these plate-shaped mold members are used as mold bodies. 9. The method for producing a silicon coagulation mold according to claim 8, wherein a mold body is obtained by assembling in a box shape so as to correspond to the bottom surface portion and the side wall portion.   The method for producing a mold for silicon solidification according to any one of claims 8 to 10, wherein the mold material contains 1 to 10 parts by mass of a phenol resin with respect to 100 parts by mass of silica glass.   The method for producing a silicon coagulation mold according to any one of claims 8 to 11, wherein the silica glass having a particle size of 45 to 150 µm is contained in an amount of 95 mass% or more based on the total silica glass.   The method for producing a silicon coagulation mold according to claim 12, wherein the silica glass having a particle size of less than 45 µm is contained in an amount of less than 5 mass% with respect to the total silica glass.   The method for producing a silicon coagulation mold according to any one of claims 8 to 13, wherein when the phenol resin is a phenol resin solution containing a solvent, the viscosity of the phenol resin solution at 25 ° C is 2 to 10 Pa · S.   The method for producing a silicon coagulation mold according to any one of claims 8 to 14, wherein when the phenol resin is a phenol resin solution containing a solvent, the fixed carbon in the phenol resin solution is 40% by mass or more.   The method for producing a silicon coagulation mold according to any one of claims 8 to 13, wherein when the phenol resin is a solid phenol resin, the fixed carbon in the phenol resin is 80 mass% or more.   The method for producing a silicon coagulation mold according to any one of claims 8 to 16, wherein the mold material contains 20 parts by mass or more of a curing agent with respect to 100 parts by mass of the phenol resin.   The silicon solidification mold according to any one of claims 8 to 17, wherein a silicon nitride solution containing silicon nitride is applied to an inner wall surface of the mold body and dried to form a release layer substantially made of silicon nitride. Production method.   19. The method for producing a silicon solidification mold according to claim 18, wherein the silicon nitride has a particle size of 50 [mu] m or less.

Images