Basic knowledge of high-purity polysilicon

Thread Content

Polysilicon material refers to a silicon material composed of two or more single-crystalline silicon grains of different sizes, and its material properties exhibit isotropy. Amorphous silicon materials are those in which silicon atoms are arranged in an orderly manner over short distances but in a disordered manner over longer distances; such materials exhibit isotropic properties. At present, the large-scale production of high-purity polycrystalline silicon is monopolized by a few developed countries such as the United States, Japan, and Germany. Since the production of polysilicon must be on a large scale (at least several thousand tons per year) to be profitable, coupled with technical complexity, proprietary nature and confidentiality, as well as the difficulties faced by new entrants in entering the market, it is quite challenging to establish an advanced and large-scale polysilicon production facility. Metallurgical-grade silicon is the raw material used to produce semiconductor polysilicon; it is obtained by reducing quartz sand (silicon dioxide) with carbon in an electric arc furnace. Although silica ores are found everywhere in nature, only a few of them can be used to produce metallurgical-grade silicon. Generally, the silica content in the ore is required to be above 97%~98%, and there are strict limits on the levels of various impurities, especially arsenic, phosphorus, and sulfur. In the metallurgical silicon used to produce high-purity polycrystalline silicon, in addition to containing over 99% silicon (Si), it also contains elements such as iron (Fe), aluminum (Al), calcium (Ca), phosphorus (P), and boron (B); their concentrations range from several tens to one thousand parts per million (in terms of molar fraction). The impurity content in semiconductor silicon should be reduced to the level of 10~9 (mole fraction), while that in solar-grade silicon should be reduced to the level of 10~6 (mole fraction). To convert metallurgical silicon into semiconductor silicon or solar silicon, it is clearly impossible to purify it while keeping it in a solid state; instead, the metallurgical silicon must be turned into a silicon-containing gas. This gas is then purified through methods such as distillation and adsorption, after which the highly pure silicon-containing gas is converted into polycrystalline silicon using chemical vapor deposition (CVD) techniques. Currently, there are three main approaches to producing high-purity polysilicon: the use of the SIMENS method (also known as the SiHCl3 method) to produce polysilicon rods ； Producing polycrystalline silicon rods using the AsiMi method (also known as the SiH4 method) ； Granular polysilicon is produced using SiH4 as the silicon source. 1. Production of polysilicon using the SIMENS method (SiHCl3 method). This method was introduced in 1954, and it quickly replaced the SiCl4 zinc reduction method that was in use at the time, becoming the method employed to this day. Its first step is to react metallurgical silicon powder with hydrogen chloride on a fluidized bed at a temperature of 250–350 degrees ； The second step is the distillation of SiHCl3, during which chlorides with different boiling points can be separated ； The third step is the deposition of silicon. Polysilicon reactors generally adopt a bell-jar design with an open end on one side. Typically, the deposition reaction for polysilicon takes 200 to 300 hours, allowing the diameter of the silicon rod deposited on the silicon bridge to reach 150 to 200 mm. 2. Production of polycrystalline silicon by the AsiMi method (SiH4 method) At the end of the 1960s, AsiMi Company proposed using SiH4 as a raw material to produce polycrystalline silicon. To produce polysilicon rods from SiH4 as a raw material, metal bell-type furnace is generally used. At high temperatures, SiH4 decomposes to produce Si and H2; the overall production cost of this method is higher than that of the SiHCl3 method. 3. Granular polysilicon manufacturing technology: This method originated from Ethyl Company’s SiH4 process. Granular polysilicon, which was commercialized in 1987, began to be produced. This technology uses a fluidized bed reactor to decompose SiH4, and the silicon formed as a result deposits on some freely flowing fine seed particles, resulting in granular polysilicon. Due to the large surface area of the crystals, the efficiency of fluidized bed reactors is higher than that of conventional Simens reactors, resulting in lower production costs for their products. The high-purity polysilicon mentioned above is the most fundamental raw material for manufacturing crystalline silicon photovoltaic cells. It is first used to produce single-crystal or polycrystalline silicon ingots, which are then cut to yield silicon wafers used in the production of crystalline silicon photovoltaic cells. 1. Single crystal silicon ingots are the raw materials used for producing single-crystal silicon photovoltaic cells. It is produced by melting high-purity polysilicon using the Czochralski method (CZ) or the floating zone method (FZ). Its diameter is about 100–300 mm, and its length can exceed 1 m. Currently, over 80% of the total output of silicon single crystals is CZ silicon, while the remaining approximately 20% is mainly FZ silicon. The FZ method does not require the use of a crucible, and it enables the production of silicon single crystals with high resistivity and purity. However, the cost of producing such silicon single crystals is high, and as the diameter of the silicon crystals increases, the production techniques face limitations. 2. Cast polycrystalline silicon (mc-Si) ingots: Photovoltaic cells made from cast polycrystalline silicon account for around 53% of the total output of photovoltaic cells, making it the most common material used in such cells. The main advantage of cast polysilicon over Czochralski-derived monocrystalline silicon is its higher material utilization rate and lower production cost ； Its disadvantages include grain boundaries, a high density of dislocations, microdefects, and a relatively high impurity concentration, which result in a significantly lower quality of the crystal compared to silicon single crystals, thereby reducing the photoelectric conversion efficiency. There are currently two main processes for producing polycrystalline silicon ingots using casting techniques: 1. The casting method involves melting high-purity polycrystalline silicon material in one crucible, and then pouring it into another preheated crucible where it cools down. By controlling the cooling rate and employing directional solidification techniques, large-grained cast polycrystalline silicon ingots can be produced. 2. Direct melting directional solidification method, abbreviated as the direct melting method, is also known as the Bridgman method. That is, high-purity polycrystalline silicon is melted directly in a crucible, and the melt is then cooled through methods such as heat exchange at the bottom of the crucible, with directional solidification technology being used to produce polycrystalline silicon ingots. The latter technology is widely used in the international industrial sector, whereas the former is currently adopted only by companies such as Solar in Germany and Kyocera in Japan. Essentially, there is no fundamental difference between these two technologies; both use the casting method to produce polysilicon. The main difference lies in whether one crucible or two crucibles are used.