Method for manufacturing polycrystalline silicon, and polycrystalline silicon for solar cells manufactured by the method

#The present invention relates to a method for manufacturing high-purity polycrystalline silicon inexpensively and in large quantities, and in, particular, it is an object of the present invention to provide a solar cell silicon element with constancy.

#The major currents of conventional methods for manufacturing high-purity polycrystalline silicon are the Siemens method and the monosilane method. These methods are methods in which a starting-material silane gas is supplied to the interior of a high-temperature reaction furnace from a nozzle installed in the bottom part of this sealed reaction furnace, and polycrystalline silicon is manufactured by the deposition or growth of silicon effected by the pyrolysis or hydrogen reduction of the starting-material gas on the surface of a solid silicon rod (silicon seed rod or core) disposed inside the furnace.

#The starting-material silane gas that is used is a chlorosilane expressed by the formula ClnSiH4-n (n is an integer from 0 to 4) which is refined to a high purity; this may be monosilane, trichlorosilane or tetrachlorosilane used alone, or may be a mixture of these compounds. However, in the Siemens method, trichlorosilane (n=3) is the main compound used, while in the monosilane method, monosilane (n=0) is the main compound used. The silicon obtained by the pyrolysis or hydrogen reduction of these starting-material gases at a high temperature has the same composition as the silicon seed rod (hereafter referred to as the Si seed rod) set beforehand inside the furnace; accordingly, this silicon has a uniform high purity from the center part to the outer circumferential part. Since this silicon has a purity that is essential in the semiconductor industry, it is called the semiconductor grade polycrystalline silicon (SEG-Si).

#With regard to the cross-sectional dimensions of the alloy member used in the present invention instead of an Si seed rod, it is desirable that the cross section be as large as possible, both in the strength, and in order to increase the deposition area (deposition rate) of silicon in the initial stage of the reaction. Since the alloy member that is used instead of the Si seed rod is metallic, unlike ceramics or the like, this member is superior in the strength, so that slender and thin working is possible. Accordingly, if the silicon deposition rate in the initial stage of the reaction is disregarded, wire materials with a diameter of 0.5 mm can also be manufactured. The same is true of the thickness and width of plate materials. However, in the case of a wire diameter of 0.5 mm or smaller or a plate material with a thickness of 1 mm or smaller, there can be a trouble of bending or collapsing of the carrier as a result of physical forces such as electric vibrations or the like, or stresses generated in the carrier (core material) by thermal expansion during electric current heating. Furthermore, during the reaction, these surfaces may be converted to a silicide, and thus become brittle in the strength, so that reutilization becomes impossible. However, since the surface of the core material is converted into a silicide, Mo does not diffuse into the silicon, so that, even if there is contamination of the product, this contamination is extremely slight.

#Meanwhile, with regard to the maximum diameter of the wire material or maximum thickness of the plate material, the electric power required for heating increases as these values increase, so that a larger power supply unit becomes necessary. Accordingly, although the maximum diameter of a wire material or maximum thickness of a plate material is limited by the capacity of the heating unit, members inside the reaction furnace can also be heated in an auxiliary manner by using an external heating means in combination. Accordingly, if the capacity is sufficient, a diameter or thickness of 100 mm or larger is also possible. Furthermore, the shape used may be a rod, plate or the like, and the diameter may be uniform or have a tapered form. The reason for applying a tapered form is that this is convenient for facilitating withdrawal of the product from the core material.

#In order to improve the productivity, the core material may have a hollow tubular structure or duplex tubular structure in a tapered form applied so that the inside of the reaction tube can be quickly cooled, thus allowing rapid recovery of the deposited or as-grown polycrystalline silicon following the reaction. The wall thickness of the hollow tube is 0.2 mm or larger, preferably 0.5 mm or larger. In the case of a duplex tube with an outer diameter of 5 mm or larger, the diameter of the inner tube is a size that fits within the outer tube. In cases where the diameter of a hollow tube is smaller than 0.2 mm, this is undesirable for the same reasons as described above. On the other hand, in the case of the maximum diameter and maximum thickness, the upper limits for both of these are selected in accordance with the magnitude of the heating electric power capacity. In most cases, a maximum diameter of 15 mm and a maximum thickness of 3 mm are recommended. Furthermore, with regard to the thickness and material of the duplex tube set in a hollow tube, it is desirable to use the same values as those of the members used in the present invention, which have heat resistance; however, since there is no direct contact with the reducing hydrogen gas, ordinary heat-resistant materials may also be used.