Let’s talk about our “silicon” stuff
Thread Content
Reposted from: Dr. Shi’s BLOG – Talking about Our “Silicon” Things (Part 1): Silicon on Earth. Silicon has the atomic number 14 in the periodic table, with a relative atomic mass of 28. It is a solid at room temperature, with a melting point of 1410–1414 degrees; its boiling point is much higher, at 2355 degrees, on the Celsius scale. As many people know, silicon is the second most abundant element in the Earth’s crust after oxygen. If everyone agrees that oxygen cannot be considered a mineral, then silicon is the most abundant mineral in the Earth’s crust, accounting for 27.7% of its composition! In other words, of the Earth’s crust, which is composed of over a hundred elements, silicon accounts for more than a quarter! In the Earth's crust, silicon exists primarily in various forms of oxides and silicates. The most common oxide is quartz, whose composition is silicon dioxide; crystal is a special form of quartz. Furthermore, mica, asbestos, garnet, etc., are all silicates. Silicon was first isolated as a pure element in 1823, by the Swedish chemist Berzelius, through reduction using potassium metal and silicon fluoride. Pure silicon crystals, on the other hand, are said to have been produced by the French inorganic chemist Deville in 1854. Currently, silicon is generally obtained by reducing quartz and carbon at high temperatures. However, the silicon obtained in this way is usually referred to as metallic silicon, and it is also called industrial silicon in China. Silicon was initially considered a metal because it is blue-gray in color and has a metallic luster. It was only later that I learned silicon is a semiconductor. The reason it is mistaken for metal is its electrical conductivity, which is caused by metallic impurities present in silicon; since it contains large amounts of metal elements such as iron, aluminum, and calcium, this gives metallic silicon its conductive properties. During World War II, in-depth research on silicon rectifiers began as detectors for radar waves. It was also at that time that it became clear that silicon is a semiconductor, and thus the foundation for today’s electronic and information age was laid. Silicon in semiconductors: Once it was known that silicon is a semiconductor, people began using it to manufacture various devices. Silicon devices require highly pure crystalline silicon, and the purification techniques for this material have gone through many developments, resulting in the current CZ pull-down crystal growth method and the floating zone melting method. Currently, silicon materials with a purity of 12N (99.9999999999%) are available, but generally, a purity of 10N is sufficient to meet the needs of most integrated circuits. If the purity of silicon is not high enough, semiconductor devices cannot be made. Although pure silicon is indeed a semiconductor, its carrier concentration is so low that it has little practical use. An important property of true semiconductors is that once doped with donor (N-type) or acceptor impurities (P-type), the number of carriers increases dramatically, which enables them to exhibit the characteristics necessary for practical use as semiconductors. Therefore, after producing highly pure silicon material, doping is still required; it may seem like an unnecessary step, but since the material also demands a high level of purity with regard to impurities, this process is still necessary. Silicon is a group IV element with four electrons in its outer shell. Therefore, group V elements, which have five electrons in their outer shell, are referred to as donor elements, as the extra electron can be shared with others, thereby enabling conductivity. Group III elements, which have only three electrons in their outer shell, are known as acceptor elements. Since there is one fewer electron there, it’s as if there is a hole; as a result, the electrons from the surrounding silicon atoms try to fill this hole. In this way, that hole seems to move around, just like a positron. Electrons and holes are collectively referred to as charge carriers. The element most commonly used as a donor impurity is phosphorus, mainly because it is non-toxic, readily available, and easy to dope. The element most commonly used as an acceptor impurity is boron, for reasons similar to those of phosphorus; however, it has an even more significant advantage over phosphorus in that its segregation coefficient in silicon is very close to 1. ----What does this mean? During doping, silicon and the impurity are melted together, and then a single crystal is drawn. Single crystals grow gradually from top to bottom. During the growth of a silicon crystal, all impurity elements have different concentrations at the interface between the solid silicon crystal and the liquid – the solid-liquid interface. The ratio of their concentration in the solid to that in the liquid is known as the segregation coefficient. The more the segregation coefficient is close to 1, the greater the proportion of solid and liquid is equal, and thus the impurity concentration in the drawn single crystal becomes more uniform. The closer the fractionation coefficient is to zero, the greater the difference in the ratio of solid to liquid. As a result, the tip of the single crystal drawn first contains very few impurities, while the bottom of the single crystal has a high concentration of impurities. The segregation coefficient of boron at a stationary solid-liquid interface is 0.8; it exceeds 0.9 when the solid-liquid interface is in motion. Therefore, in the drawn single crystal, the concentration of boron doping remains very uniform from start to finish. The segregation coefficient for phosphorus is 0.36; in actual crystal pulling, this value can exceed 0.5. Although it is somewhat low, it remains the highest among the Group V elements (use it as a makeshift solution!) ). The silicon in solar cells: When light shines on silicon, photons – which are another term for light – stimulate the electrons in the low-energy levels of the silicon material. These electrons, which are lazy and stationary and therefore do not conduct electricity, jump to higher energy levels, turning into free electrons that are active and move around. As a result, an electric current is generated; this effect is known as the photovoltaic effect. The first silicon solar cell was developed in 1954 at Bell Labs in the United States; its efficiency was 6% at that time, and it rose to 10% shortly thereafter. However, its cost was 1,000 times higher than that of conventional electricity, so it was used only in spacecraft. The photoelectric conversion efficiency refers to the percentage of the electrical energy generated by a photovoltaic cell compared to the energy of the light that illuminates its surface. Typically, the limit for monocrystalline silicon is 24%, while that for polycrystalline silicon is around 20%. This refers to the efficiency of a single PN junction alone, without taking into account techniques such as focusing or stacking. As in semiconductors, when donor and acceptor elements are doped into silicon, the photovoltaic effect is significantly enhanced, thereby making it possible to generate electricity using sunlight. When donor impurities are added, it is called N-type; when acceptor impurities are added, it is called P-type. Knowing the differences in the segregation coefficients among impurities helps to understand why modern solar cells prefer P-type materials as the base material for these cells. This is because boron can be doped more evenly, whereas when phosphorus is used for doping, the concentration difference between the beginning and the end of a single crystal is often large, as the impurity phosphorus accumulates from the head to the tail of the crystal. Although the photovoltaic effect is also a property of semiconductors, this effect does not require silicon to have such a high purity as in other semiconductor applications, such as integrated circuits. Research has shown that if the silicon used in solar cells can reach a purity of 6N with an appropriate level of impurities, the photovoltaic conversion efficiency can be very high. Therefore, it was not surprising in the past to use scraps and leftover materials from semiconductor silicon production for making solar cells. Before 2004, solar cells were used very little; as a result, the demand for silicon materials for solar applications was also low, and scrap semiconductor material was sufficient to meet those needs. But from that time on, it became clear around the world that the conventional energy sources on Earth would not last for many years, which led to an increased focus on solar energy. But solar cells are too expensive; to reduce costs, it is necessary to increase their scale of use ； But to get everyone to use it, costs must be reduced. This paradox of whether the chicken came first or the egg was resolved for the first time in 2004 by Germany. At the end of 2004, Germany **enacted the Renewable Energy Act**, declaring a subsidy of 0.57 euros per kilowatt-hour for solar power generation, which is more than ten times the price at which coal-powered electricity is sold on the grid. In this way, even if the cost of solar power generation is ten times higher than that of coal-powered generation, investing in solar power still yields higher profits than investing in coal-powered generation of the same scale. As a result, large amounts of capital flowed into photovoltaic power generation, and solar power plants as well as rooftop power systems emerged in large numbers, leading to a sharp increase in the demand for silicon materials for solar energy. Subsequently, countries such as Spain, California in the United States, Italy, and Australia also introduced various subsidy policies, leading to a surge in global demand for silicon materials. Solar cells use monocrystalline silicon or polycrystalline silicon, and monocrystalline silicon is also produced by pulling polycrystalline silicon, which has led to a severe shortage of polycrystalline silicon. This can be seen from the price: in 2004, solar-grade polysilicon cost only 20 dollars per kilogram, while by mid-2008 it had risen to 480 dollars per kilogram – a 24-fold increase over four years! In the past, after obtaining polysilicon, it was used to grow single crystals, which were then sliced to create solar cells that could be used for power generation. Now, more and more companies are also using polycrystalline silicon ingots to directly cut into wafers for battery production, with the aim of reducing costs. Although it is less efficient, the efficiency of single-crystalline silicon cells is typically around 17%, while that of polycrystalline silicon cells is around 15%. But when calculated per watt, polysilicon is still cheaper. Recently, due to the high cost of silicon materials, thin-film solar cells have also become popular. In fact, even if polysilicon is cheaper than sand, thin films still have their market. It can be made flexible, facilitating integrated production and design with buildings, which gives it irreplaceable advantages, although efficiency and degradation are issues that need to be improved. (To be continued) This post was last edited by seanjiangsu on 2009-2-13 09:38]
From ore to pure silicon: Although silicon is the most abundant element in the Earth’s crust, silicon materials used in solar energy still face such shortages due to the difficulty of purifying it. The raw material for silicon used in solar cells still comes from ores; quartz or silica ore is smelted to produce metallic silicon, which is then purified into polycrystalline silicon suitable for use in solar cells. Polycrystalline silicon is used to manufacture solar cell panels, or it can be turned into monocrystalline silicon before being used to create solar cell panels as well. The current bottleneck in the solar cell silicon material industry lies in the process of converting metallic silicon into polycrystalline silicon. Regardless of the method used later to purify polysilicon, metal silicon is usually used as the raw material; therefore, let’s first discuss the smelting of metal silicon. Smelting of metallic silicon: As mentioned earlier, metallic silicon is obtained by reducing quartz ore with carbon at high temperatures. Quartz ore, also known as silica ore, has silicon dioxide as its main component. Carbon is used as a reducing agent to reduce silicon from silica; coke, charcoal, refined coal, and petroleum coke are commonly available options. The quartz ore is crushed to an appropriate size, mixed evenly with a carbon reducing agent in a ratio of about 3:1, and placed into a electric arc furnace. An electric arc is generated to melt the silica, allowing it to undergo a reduction reaction with carbon at high temperatures. The reactions inside a electric arc furnace are actually very complex. To date, no smelting process model has been able to accurately describe the physicochemical changes that occur within a electric arc furnace. However, since this smelting process has been in use for many years, there are no major issues with it for metal silicon used in general purposes. Moreover, in practice, the experience of a skilled worker is often more useful than the models calculated by various smelting experts. The production of metallurgical silicon is an industry with overcapacity in China. It consumes a large amount of electricity – approximately 12,000 to 15,000 kWh per ton of metallurgical silicon – and in addition, the flue gases generated in the electric arc furnace cause significant pollution. Due to price competition, the profits from manufacturing metallic silicon are currently low; as a result, many factories are reluctant to invest in installing exhaust gas recovery equipment. In fact, if exhaust gas recovery and waste heat utilization are implemented properly, not only can the recovered silicon micropowder be resold, but waste heat utilization can also reduce the electricity consumption in smelting; therefore, carrying out environmental-friendly recovery in electric arc furnaces is not a loss-making endeavor. Even without considering recycling, from an environmental protection perspective, reducing emissions should be a social responsibility that companies ought to assume. The purity of metallic silicon is usually expressed as a thousandth of the percentage content of its three main impurities. These three impurities are iron, aluminum, and calcium. If the contents of these three impurities are 0.5%, 0.5%, and 0.3% respectively, it is called 553; if the contents of iron, aluminum, and calcium are 0.4%, 0.4%, and 0.1% respectively, it is called 441. If the content of one of these three impurities (usually calcium) is less than one thousandth, then a “0” is added before that digit. For example, if the contents of iron, aluminum, and calcium are 0.2%, 0.2%, and 0.02% respectively, it is denoted as 2202. Similarly, if the grade of metallic silicon is 1101, it means that its contents of iron, aluminum, and calcium are 0.1%, 0.1%, and 0.01% respectively; expressed in ppm, these values are 1000, 1000, and 100 ppm for iron, aluminum, and calcium respectively. If the total content of these three impurities is less than 1000 ppm, it is referred to as 3N metal silicon. Strictly speaking, this statement is inaccurate, because aside from these three types of impurities, there may also be considerable amounts of other impurities such as titanium, magnesium, oxygen, carbon, phosphorus, boron, etc. However, since most metal silicon factories in China currently do not have the facilities to analyze the contents of these other types of impurities, it has become a common practice to express the purity of metal silicon by subtracting the amounts of iron, aluminum, and calcium. It’s just a *common term; there’s no need to argue about its accuracy. It’s just that when others in the industry mention it, you should understand what they mean so as to avoid any misunderstandings. At present, the metal silicon plants in China that are generally of high quality are able to produce 2202 grade metal silicon stably. Recently, as more and more companies have begun developing solar cells using physical methods, higher requirements have been placed on metal silicon. Many manufacturers have started producing 1101 grade metal silicon, and they are also able to produce a certain proportion of 3N grade metal silicon. The larger the furnace capacity for metal silicon, the higher the efficiency. Currently, the furnaces in use domestically are mostly in the range of 5000KVA to 12000KVA. Internationally, furnaces with a capacity of 30,000 KVA are already in use. An investment of around 20 million RMB is sufficient to set up a metal silicon plant with a capacity of 6300 KVA. For chemical polysilicon, whether it is used in semiconductors or in solar applications, the purification process of converting metallic silicon into polysilicon currently relies on the Siemens method. In this method, metallic silicon is used as the raw material; concentrated hydrochloric acid is added, and a hydrochlorination reaction takes place at high temperatures to produce trichlorosilane. After being heated to vaporize, trichlorosilane is then distilled to remove impurities from the gas, resulting in highly pure trichlorosilane. This trichlorosilane is subsequently reduced using hydrogen to yield pure polysilicon. This method is called the chemical method because both the hydrochlorination and reduction of silicon are chemical processes. The method of reducing trichlorosilane with hydrogen was first adopted by Siemens, and it is also known as the Siemens process. In addition to the Siemens method, there are also methods such as zinc reduction and silane methods, all of which are chemical methods. The main problems faced by domestic chemical methods are recycling and environmental protection. The hydrochlorination of metallic silicon generates numerous harmful gases and liquids, while the reduction of trichlorosilane also results in the emission of many harmful gases due to issues with reaction efficiency. If these gases are not recycled, they not only pollute the environment but also increase the costs for businesses. However, since the recycling technology is relatively complex, no factory in China is currently capable of carrying out a full closed-loop production process for recycling. The consequences of this are, on one hand, legal risks, and on the other hand, consistently high costs. Currently, the power consumption per kilogram under the Siemens process is around 150 kWh internationally, while in China it is around 200–250 kWh per kilogram. In terms of cost, the cost per kilogram of polysilicon using the Siemens method is around 30 US dollars internationally, while Chinese companies incur a cost of nearly 70 US dollars. Companies are currently carrying out the reduction process using the silane method. Silanes can directly decompose into silicon under certain conditions, so there is no pollution issue. But the biggest problem with silanes is safety issues during production and transportation. Because silane is a highly explosive gas, the slightest carelessness can lead to an explosion. At present, there is no factory in China capable of producing more than 500 tons of silane per year. If silane gas is purchased from the United States or Japan, its price per kilogram is quite high nowadays, exceeding 2,000 RMB. The principle of zinc reduction is to use zinc as a substitute for hydrogen in the Siemens process to reduce trichlorosilane, thereby partially eliminating the environmental pollution problems associated with hydrogen reduction. But the problem with zinc reduction is that high-purity zinc is relatively difficult to obtain. Furthermore, the reduction efficiency is lower than that of hydrogen, which means there are also greater risks in terms of cost competition in the future. Polysilicon plants using the Siemens process require significant investment in addition to technical processes. If the production process starts with the conversion of metallurgical silicon into trichlorosilane and continues up to polycrystalline silicon, a factory with a capacity of 1,000 tons would require an investment of around 1.2 to 1.5 billion yuan. There is actually a lot to say about chemical polysilicon, but I won’t go into further detail in this article (after all, that’s about all I know:)). (To be continued) This post was last edited by seanjiangsu on 2009-2-13 09:39]
Overview of physical method polycrystalline silicon: The physical method, also known as the metallurgical method, is a technique for producing polycrystalline silicon by directly smelting metallic silicon; impurities are removed through high-temperature melting and directional solidification, allowing the production of metallic silicon with a purity of 6N or higher. This method was tested in the laboratories of Germany’s WACKER company as early as 1975, but since its limit was only 7N, it was completely unsuitable for use in semiconductors at that time. Moreover, solar cell applications hardly existed back then. Therefore, after achieving a purity level of 5N, WACKER stopped researching this metallurgical method once Siemens’ purification technology became commercially available. Since 2004, as the amount of silicon used in solar energy increased, by 2007 silicon used for solar energy surpassed that used in semiconductors for the first time in history; therefore, it became practical to resume research on polycrystalline silicon produced by physical methods. Many polysilicon or metallic silicon companies, including WACKER, JFE, Mitsubishi, HEMLOCK, ELKEM, and others, are once again beginning to research the production of polysilicon using physical methods. There are many companies in China that used to be involved in the production of metal silicon; since the technical barriers associated with the metallurgical method appear to be lower compared to the Siemens method, they have also begun to explore technologies for purifying polysilicon using both metallurgical and physical methods. The earliest one in China is probably Sanjing in Nan’an, Fujian. There are also many process routes in physical methods. The most common approach involves off-furnace refining of metallic silicon, during which impurities in the metal silicon are further removed through gas injection and slag formation. After that, acid washing is carried out, followed by vacuum melting; impurities are removed in the form of slag and gases by introducing reactive gases, adding reactants, and using vacuum melting. At this stage, techniques such as ion beams, electron beams, and electromagnetic fields may be employed to help remove certain impurities. Finally, through directional solidification and the effect of fractional crystallization, impurities in silicon are concentrated at the ends of the silicon ingot, resulting in silicon with relatively high purity in most of its volume. As can be seen from the above, physical polysilicon is actually produced using metallurgical methods, and metallurgy refers to high-temperature chemical reactions. Then why is it still called the physical method? This is mainly because, throughout the entire process, silicon itself, as the element to be purified, does not participate in any chemical reactions. In the chemical method, silicon undergoes a chemical change. Therefore, the physical method does not mean using only physical methods and completely avoiding chemical methods. Currently, the international industry generally refers to physical methods as metallurgical methods, and the silicon purified through such physical methods is called UMGS (Upgraded Metallurgical Grade Silicon) ), and this is because they refer to metallic silicon as metallurgical grade silicon (MGS). Silicon produced by chemical methods for use in solar cells is called solar-grade silicon (SGS, Solar Grade Silicon), while that used in semiconductors is referred to as electronic-grade silicon (EGS, Electronic Grade Silicon). I sometimes wonder that in the future, when silicon produced by metallurgical methods can be used entirely for manufacturing solar cells, and if the specifications of UMGS and SGS become exactly the same, how will these names change. For the production of high-purity metallic silicon, whether through chemical or physical methods, metallic silicon is used as the raw material. However, since the chemical method uses gas fractionation, it is not sensitive to impurities in metallic silicon; generally, metallic silicon with a purity of over 95% can be used, and even for silicone-based materials, metallic silicon up to grade 2202 is sufficient. However, the physical method relies on melting to remove impurities, so it is sensitive to the impurities contained in the metallurgical silicon used as raw material; the purer the metallurgical silicon, the better. Thus, with the rise of physical-method polysilicon, new requirements have been placed on the smelting of metallic silicon. Based on the current practices in conventional electric arc furnace processes for metallic silicon, it is generally required that the purity of metallic silicon be at least 3N, with a minimum level of 1101. The main raw material for silica is quartz ore, also known as siliceous rock. Its main component is silicon dioxide, but it usually contains metal impurities such as iron, aluminum, calcium, titanium, and magnesium, as well as non-metallic impurities such as phosphorus and boron. Carbon used as a reducing agent for silica is typically coke, petroleum coke, charcoal, and refined coal ； These carbon reducing agents often contain various types of ash, including various metal and non-metallic impurities, with particularly high levels of phosphorus. Therefore, the raw materials for metal silicon inherently contain impurities. Furthermore, during the smelting of metallurgical silicon, as the graphite electrodes used in electric arc furnaces are consumed, the impurities contained within those electrodes also end up in the metallurgical silicon product. There are also the lining and bottom materials for electric arc furnaces; in particular, some refractory materials contain large amounts of impurities such as phosphorus and boron during their production, which can also contaminate the metallurgical grade silicon produced. Therefore, to improve the purity of metal silicon, it is not only necessary to use high-quality ore, but also reductants with low impurities, electrode materials of high purity, as well as furnace bottom and lining materials with very low levels of impurities. There are no fixed rules for selecting these materials. For example, regarding silica ore, the best quartz ore currently found has a purity of 5N; the total amount of metal impurities such as iron, aluminum, and calcium is only a few PPM, while phosphorus and boron levels are generally below 1 PPM, with some even being less than 0.1 PPM. The Pro Company where I work has already identified several quartz ores with a purity higher than that of Unimin Corporation in the United States. But in fact, silica ore does not have to be that pure; it is acceptable to have some impurities, such as iron, aluminum, and calcium, even in higher amounts. Furthermore, in the raw materials, just as important as the purity of silicon is the purity of the reducing agent. Generally, carbon reductants have relatively low purity, with higher proportions of phosphorus and boron compared to ores. Among various reductants, petroleum coke has the lowest ash content, followed by charcoal, then coke, and bituminous coal having the highest ash content. The situation with electrodes is similar to that of reducing agents: high-purity graphite electrodes are superior to ordinary electrode graphite electrodes, as ordinary electrode graphite contains clay. The furnace lining can be made of materials with higher purity, such as high-alumina bricks or graphite linings. However, after several cycles of use, a layer composed of silicon carbide and silicon dioxide forms on the furnace walls, separating the lining material from the furnace charge; as a result, contamination of the silicon material by the lining is significantly reduced. In addition to material selection, the raw materials can also be processed. For example, removing the metal components from the ore through high-temperature pickling, and removing phosphorus and boron from reducing agents through high-temperature chlorination are all methods that can be employed. For example, using chlorination at a high temperature of 2500 degrees can reduce phosphorus in coke from 100 ppm to below 2 ppm. Furthermore, since it is easier to remove boron from silica than from silicon, it is also possible to add some oxidants to the smelting furnace to increase the oxidation and volatilization of non-metallic elements such as phosphorus and boron, thereby reducing their content in metallic silicon. However, regarding the potential hazards that the process may pose to the equipment as well as safety issues, protective and preparatory measures must be taken in advance. Furthermore, if high-purity metal silicon is to be produced, the smelting process, power supply system, and furnace shaking procedures must all be re-evaluated and adjusted. The focus should be on improving the furnace’s air permeability and thermal insulation, in order to maintain optimal operating conditions within the furnace. If careful attention is paid to the selection of metallic silicon and the smelting process is well controlled, it is possible to produce 3N grade metallic silicon relatively easily. Among them, metal impurities should be controlled at 70–80 ppm, while non-metallic impurities should be around 20–30 ppm. In addition to controlling the overall impurity level, there is also specific control for phosphorus and boron; it is ideal to keep phosphorus below 10 ppm and boron below 1 ppm. The production of high-purity silica and high-purity carbon: Currently, some people are also attempting to use high-purity silica in combination with high-purity carbon in a reaction, with the aim of directly producing metal silicon of 4N purity. The idea is sound in principle, but the cost may not be low, and the purity may not be high either. If the purity of silica is required to be 5N or higher, then the impurity levels in other parts of the furnace must also be very low. Conventional electric arc furnaces are not capable of meeting such purity requirements, as the level of contamination in these furnaces renders the high purity of the raw materials meaningless. It is not an easy task to create a reduction furnace environment that does not contaminate high-purity silicon material, and moreover, it is necessary for all the process parameters related to the smelting in such a reactor to be met – not only temperature, but also the atmosphere, heat field, pressure, and so on. Although the electric arc furnace is a very old smelting technique, from the perspective of mineral processing, it is currently difficult to find a furnace type that is more suitable for smelting than the electric arc furnace. Therefore, for the smelting of high-purity silica and reducing agents, a feasible approach is to refer to the structure of electric arc furnaces, use high-purity structural components, lining materials, and electrode materials, and design a new furnace type on one’s own in order to conduct metallurgical experiments on high-purity materials. Furthermore, some people think of using hydrogen to reduce silica. If one considers what will be produced when hydrogen reacts with silica and what the internal temperature will be at that time, it becomes clear just how dangerous this is. So, one must be extremely cautious. Since the metal silicon plants that supply polycrystalline silicon for physical processes are those that provide the main raw materials for solar energy, a clean energy source, it is essential to install equipment for environmental protection and waste heat recovery within these plants, so as to prevent pollution and energy waste from tarnishing the reputation of solar energy. In fact, investments in environmental recycling can typically be recouped within two to three years through the recovered silica powder or heat. (to be continued)
Extracorporeal refining for silicon purification by metallurgical methods: Although high-purity metallic silicon can be obtained through certain smelting methods and processes, achieving a purity level of around 3N is not sufficient for the physical purification of silicon intended for solar applications. Therefore, after the metal silicon smelting process, an additional step of off-furnace refining is usually added to further increase purity at a low cost. Off-furnace refining is actually not a new process. In the steel industry, almost all types of steel are currently refined. In traditional industrial silicon production, for industrial silicon with a grade of 2 or higher (221 and above), refining methods such as oxygen blowing are also commonly used. In the past, chlorine blowing was mainly used to refine metallic silicon produced in electric arc furnaces. However, since the oxygen blowing process was introduced to China, chlorine blowing was phased out of industrial use due to pollution issues, with only a few special applications remaining. In fact, oxygen blowing can basically completely replace chlorine blowing, without the hazards associated with chlorine blowing. However, the oxygen-blowing process alone is not sufficient; currently, almost all metal silicon plants are equipped with oxygen-blowing equipment, primarily to remove iron and calcium, and it also has a certain effect on removing aluminum as well. However, to provide raw material for the purification of polysilicon, a slag-forming process must be introduced during the refining process. For extracorporeal refining, a refining furnace must be installed next to the electric arc furnace; after the silicon water emerges from the electric arc furnace, it is fed into the refining furnace directly or indirectly. In the refining furnace, heating is usually carried out by induction heating to keep the silicon melt at over 1500 degrees, after which a slag-forming agent is added along with oxygen blowing. The slag former is composed of various metal oxides and salts, which can react with the impurities in metallic silicon to form a slag phase that floats on top of the silicon melt. The selection of slagging agent components is related to the impurity content and composition of metallic silicon. In the case of slagging refining, oxygen blowing has three effects: one is to utilize oxygen to react further with metal impurities such as iron, aluminum, and calcium present in silicon, thereby forming metal oxides that escape from the silicon as gases ； The second method is to use the high temperature generated by oxidation to help maintain the temperature of the silicon melt, thereby reducing power consumption ； Thirdly, it serves to stir, enabling the added slag-forming agent to react fully with the impurities in the silicon melt. After the slagging agent has fully reacted with the impurities in the silicon, the slag formed is removed, and then the silicon melt is poured into an insulating container, where a simple directional solidification process takes place. In this way, metallic silicon with a purity of 4N can be obtained. Typically, the impurity composition in 4N metallic silicon produced by extracorporeal refining is as follows: iron, aluminum, and calcium account for about 80 ppm, phosphorus is around 5–10 ppm, boron is about 1–2 ppm, and the remaining impurities make up approximately 10 ppm. Induction furnaces are commonly used for heating in refining furnaces, but other methods are also feasible. Canada’s BSI company uses natural gas mixed with air for combustion heating, and then employs the resulting flame to refine silicon melt contained in a converter; this approach takes advantage of Canada’s abundant and low-cost natural gas resources. However, natural gas has some additional advantages: the alkane gases such as methane in it are very effective at removing various impurities. Furthermore, the flame of natural gas enables thorough reaction between the slag phase and the silicon melt. Moreover, by using a converter, the silicon melt can rotate continuously, ensuring overall mixing, and the processing capacity per batch can be very high. BSI claims to be able to achieve a processing capacity of 5 to 8 tons per furnace, a figure that is difficult for other refining furnaces to match. Furthermore, it is also possible to use a resistance furnace for refining; however, at high temperatures, the oxidation problem of the resistance heating elements is difficult to address, and finding heating elements that do not oxidize at around 1500 degrees comes at a high cost. Japan’s JFE uses plasma beams to refine metallic silicon; Ar, H2O, and hydrogen are utilized to blow a plasma flame at the silicon melt, and the high temperature generated by this plasma flame (ion beam) removes boron and other impurities. However, this approach can be used as a new purification mechanism in scientific research by researchers, but it is unlikely to be applied in actual production. This is because the temperature generated by plasma is too high; the ion beam not only removes impurities but also causes significant loss of silicon itself. Those that use electron beam phosphorus removal fall under the category of vacuum melting, but they also face the same problems. Furthermore, whether it is an ion beam or an electron beam, the cost of using these energy beams is very high. Furthermore, it should be noted that safety must be prioritized during the refining process. Because slag formers have to be added at high temperatures, and oxygen must be blown in – all of these are very dangerous tasks ； When using energy beams and other refining methods, hydrogen is sometimes employed for purification, and extra caution must be taken when handling hydrogen. Also, since refining is often carried out next to electric arc furnaces, the operating environment is very important as well; the path taken by the silicon containers, as well as the tools and procedures used for adding auxiliary materials, all need to be carefully designed. It is particularly important to note that water must under no circumstances get into or leak onto the silicon melt or the hot silicon, as this could lead to catastrophic consequences. Hydrometallurgy: Many companies that use physical methods for purification do not have their own metal silicon production facilities; as a result, it is difficult for them to carry out processes of refining outside the furnace. However, if purchased metal silicon is used directly in a vacuum environment without any treatment, it leads to significant waste. Therefore, many people have adopted the pickling method, in which metallic silicon is crushed and then immersed in acid; this approach is quite effective at removing metal impurities from metallic silicon. Generally, iron can be removed by about 90%, it is more effective at removing calcium, but less effective against aluminum. Using acid alone, whether it is hydrochloric acid, sulfuric acid, or nitric acid, the effect on removing boron and phosphorus is not significant. Many experts are also researching this area; by using chemical agents such as acids, bases, complexes, and ion exchange resins, and through ion exchange processes, good results have been achieved in removing aluminum, phosphorus, and boron from silicon. However, the minimum particle size after crushing is usually between 200 and 400 mesh; if it is too large, the impurity removal effect is not significant, while if it is too small, silicon loss increases. At this scale, these room-temperature chemical methods can only remove impurities on the surface of the silicon powder; they are essentially ineffective against impurities inside the particles. Therefore, the impurity removal effect is also limited. However, since during the solidification of metal silicon, directional solidification causes most metallic impurities to accumulate at the grain boundaries, the amount of impurities on the surface of the particles is relatively higher compared to that inside the particles. Therefore, this method constitutes a valuable intermediate process for producing polycrystalline silicon via metallurgical methods. This method is also known as hydrometallurgy, and it is used by many companies that employ physical methods for the purification of polysilicon. However, each company uses different medications, as well as various processes and methods for treatment. However, the silicon material produced by wet metallurgical processes is in powder form, and when this powder is subjected to vacuum melting in a vacuum furnace, it causes significant damage to the heat field components within the furnace as well as to the vacuum system; this is a weakness of wet metallurgy. Generally, if the process is proper, hydrometallurgy can also yield 4N metallic silicon, with a purity comparable to that of silicon obtained through off-furnace refining. Due to its low power consumption and the fact that no high-temperature processes are required, many companies use hydrometallurgy as an intermediate process after metal silicon production and before vacuum melting. However, compared to off-furnace refining, hydrometallurgy is less effective at removing phosphorus and boron. In addition to processing metallic silicon, hydrometallurgy is also an effective method for recycling and reusing the end pieces and scrap material cut off after vacuum casting. Given that the various techniques for purification using physical methods are not yet mature at present, it is normal and appropriate for people to try different refining methods. Vacuum melting and directional solidification: To purify polycrystalline silicon from refined metallic silicon, vacuum melting is used as the first step in the purification process. Vacuum melting is a method for purifying metallic silicon under vacuum conditions. Generally, it includes two stages: vacuum slag formation and vacuum blowing. Vacuum slag formation is completely different from the slag systems used in refining. Firstly, since the raw material is already 4N polysilicon, the amount of slag former that needs to be added cannot and need not be as large as that used in extracorporeal refining. Furthermore, the term \"slagging agent\" is essentially just a synonym; it refers to something that can oxidize those impurities that are difficult to separate from silicon into compounds. These compounds either float to the surface of the silicon as slag, settle at the bottom of the crucible as precipitates, or escape as gases. In any case, the goal is to make it easier for these impurities to be separated from silicon. Vacuum blowing involves introducing gas into the molten silicon melt; the purpose is for this gas to react with the impurities present in the silicon. It also increases the contact area and time between the slag-forming agents and the silicon melt, allowing for full reaction. After the reaction, new gases are generated, which are then removed from the furnace by the vacuum system. Some experts (including so-called foreign experts) once proposed using propellers for stirring in order to increase the contact area of the slag-forming agent, but tests have shown that this method is not advisable. The main problems are the contamination and strength of the propeller material. Whether graphite material or quartz material is used, significant problems will arise. Using electromagnetic stirring is another viable option. However, if the crucible is made of graphite, the electromagnetic field in the silicon melt will be **significantly weakened, and the effect may not be as good as expected. However, the introduction of an electromagnetic field does help in the enrichment of metallic impurities in silicon materials. Silicon materials subjected to vacuum melting can typically achieve a purity of 5N. After vacuum melting, a directional solidification stage is required to further purify the non-volatile impurities in silicon. Directional solidification involves cooling the silicon melt in the crucible from the bottom, allowing it to solidify gradually as it moves upward. The segregation effect during the solidification process is utilized to concentrate impurities at the top. This process is both an ingot-making process and a purification process. However, generally, since the crystallization and melt properties of silicon materials containing impurities differ from those of high-purity materials, the heat field used in physical purification and ingot casting is different from that used in conventional ingot casting. Furthermore, it should be noted that all current thermal field analysis software has certain limitations; one cannot rely on them 100%, and practical testing is still necessary to verify their effectiveness. Practice is the sole criterion for testing truth. Directional solidification and vacuum melting can be carried out in the same furnace, or separately in two different furnaces. Since the thermal fields and vacuum requirements for melting and directional solidification are different, it is better to carry them out separately. However, after separation, two melting and solidification processes are required, which takes more time, increases power consumption, and raises the cost of the equipment. Pro New Energy currently uses a patented integrated design featuring two types of furnaces developed by itself, which has successfully addressed this issue. During ingot casting, it is essential to ensure two basic conditions: first, the temperature gradient should always decrease from bottom to top; second, the solid-liquid interface should remain as horizontal as possible. These two conditions sound simple when mentioned, but actually achieving them is not that easy. The main reason is that after solidification, silicon melt has very low thermal conductivity; as a result, it is not easy to dissipate the heat originally present inside the silicon melt and the solidified silicon ingot, as well as the latent heat released during crystallization. Even if the temperature field is designed to seem reasonable, it is not easy to keep the solid-liquid interface within the silicon ingot horizontal at all times ； Especially in the latter half of the crystallization process, as the crystal thickness increases, the complexity of the temperature field also increases. Currently, the calculation models for the heat field in ingots include the DSS type, HEM type, RPDS type, and so on ； There is a significant difference between these and the actual heat field distribution in the cast ingots. Physical polysilicon, due to its higher impurity content, is also affected by these impurities; the vapor pressures, free energies of various impurities, and even the entropy changes associated with crystallization all have an impact on the crystallization process. These factors need not be taken into consideration when casting polysilicon produced by chemical methods. The above provides a brief introduction to the principles and process flow of physical vacuum melting. Specific process conditions, such as the temperature to which heating is required, the time needed for that process, the level of vacuum required, the gases to be introduced (their composition, pressure, flow rate, and duration), the type of slag to be used, the flow rate and pressure of the gases, the structure and materials of the heat field, as well as the type of crucible to be used, all fall under the category of technical secrets held by various manufacturers. Each company has its own approaches, and it is necessary to conduct numerous experiments based on one’s own equipment conditions in order to determine the most suitable process parameters. (to be continued)
Methods of expressing silicon purity: For solar cells, the question of just how pure the polycrystalline silicon needed is remains a challenge for all polycrystalline silicon manufacturers. Some say 7N, some say 6N, and some say 5N is sufficient. Actually, all of these statements might be correct. First, we need to establish a convention for the defining terms of the material. Usually we talk about 6N or 5N; theoretically, this should refer to the number of \"9s\" obtained by subtracting from 100 the percentage concentration of all impurities present in silicon. For example, if the total concentration of all impurities is 10 ppm, which is 0.001%, then after subtracting 10, we are left with 99.999%, which is referred to as 5 nines, or 5N – where N stands for the English word \"nine\", meaning \"nine\". Therefore, 5N refers to the total number of \"9\"s before and after the decimal point in a percentage, not just the number of \"9\"s after the decimal point. Although the definition above is strict, it is by no means easy to detect all 109 elements in silicon. At present, with a precision of the order of ppm, the instruments commonly used for testing element contents are ICP-AES (Inductively Coupled Plasma Atomic Emission Spectrometer) or ICP-MS (Inductively Coupled Plasma Mass Spectrometer); GDMS (Glow Discharge Mass Spectrometer) is also used. When these instruments are used separately, they can measure 60 to 70 elements in total, but usually only 20 to 30 elements can be measured at one time. Usually, people choose the more common and abundant elements in silicon for measurement. As for the other elements, although they are not measured, they are generally below the detection limit; if present at all, their amounts are very small. Therefore, whether or not they are measured, the results remain similar. It should also be noted that when expressing the content of impurities in ppm (parts per million), there are also ppmw and ppma. ppmw refers to the concentration calculated on a weight (mass) basis; for example, if B is 2 ppmw, it means there are 2 micrograms of boron per gram of silicon ； PPMA, on the other hand, refers to the concentration calculated based on atomic density – that is, how many impurity atoms there are per million silicon atoms. For the same material, there is a certain relationship between ppmw and ppma; their ratio is identical to the ratio of the atomic weights of silicon atoms to those of impurity atoms. For example, the atomic weight of silicon is 28, while that of boron is 11; if the concentration of boron is 2 ppmw, then the corresponding value in ppma is: 2*28/11 = 5.1 ppma. Typically, ICP-AES, ICP-MS, and GDMS provide values in ppmw, whereas spectrometers such as FTIR give values in ppmma. Unless otherwise specified, the ppm values given in this article refer to ppmw. Furthermore, it has been found that certain elements in silicon, such as non-metallic elements like C, O, N, and H, will naturally decrease to a certain level—for example, below 10 ppm—once the purity of other impurities reaches a specific level. At this point, these elements no longer have an adverse effect on the performance of silicon in solar cells, such as their conversion efficiency. Some analyses indicate that levels of C and O below 20 ppm do not cause significant degradation in the solar cells nor affect their conversion efficiency; however, the author believes this finding is debatable. Therefore, as long as the concentration of these elements is not too high, there is usually no need to deduct their amounts from the total composition. Therefore, to save trouble, people now usually use 100 minus the value obtained by measuring twenty or thirty elements with conventional instruments such as ICP, to represent the purity of the silicon material being measured. This means that although the method is not entirely rigorous, it is not very different from the theoretical definition. Furthermore, this method does not take into account harmless elements such as C, N, O, and H. When the concentration of these impurities is not high enough to affect the photoelectric conversion efficiency, it is possible to disregard these irrelevant elements, allowing the impurity concentration values to correspond more accurately with the battery’s performance. Therefore, it is currently a widely accepted method of representation. For metallic silicon, since the most abundant elements in it are iron, aluminum, and calcium, the number of nines is usually indicated by subtracting the sum of the contents of these three impurities from 100. For example, if it contains 600 ppm of iron, 300 ppm of aluminum, and 100 ppm of calcium, then the total amount of these three impurities is 1000 ppm, which is equivalent to 0.1%. By subtracting this value from 100%, we get 99.9%, meaning that this metal silicon can be classified as 3N grade. Although this notation is not strict, it can still correspond to the current domestic definitions of the grade specifications for metallurgical silicon. In addition to the integer N, there are also decimal values. For example, 5.5N indicates a purity of 99.9995%, while 5.2N indicates a purity of 99.9992%. Therefore, this term allows for an intuitive and convenient way to indicate the purity of silicon material. Although it is not entirely standardized, it provides a clear method for expressing purity, and it might be worthwhile to adopt it as a common practice and promote its use. What kind of silicon can be used to make solar cells? Now, let’s talk about what purity level is required, in N units, to be able to produce solar cells. It is certain at present that 7N material can indeed be used to manufacture batteries, and impurity doping is also required during the process of growing single crystals. As for 6N, it’s not certain; it depends on the impurity content inside. If the impurities present are various metal impurities without phosphorus or boron, then such batteries can also be manufactured. And if all of that 1 ppm of impurities are boron, then because the high boron content results in too low a resistivity, phosphorus must be added to \"increase\" the resistivity; however, this in turn leads to an increase in phosphorus, so the actual amount of impurities may end up being 3–5 ppm. Although such materials can be used to produce batteries with an efficiency of 16%, they usually exhibit photodegradation. As for 5N batteries, it is usually not possible to produce them as functional batteries (or rather, the efficiency of the solar cells produced is too low, definitely less than 10%). But if the purity is increased to 5.5N, meaning the impurity content is 5 ppm, it depends on what kind of impurities are present. For example, if the boron content is greater than 1, phosphorus doping is still needed to \"increase\" the resistivity; and if all of the remaining four ppm are metal impurities, the resistivity will be even lower, requiring an even higher proportion of phosphorus doping. In this way, although solar cells can still be made, firstly their efficiency is not very high, and secondly, due to the presence of impurities, photoinduced degradation occurs. But if, among the 5 ppm of impurities, boron accounts for 0.5, phosphorus is less than 0.5, and the remaining impurities are mainly aluminum and calcium, with iron and other elements being less than 0.3, then solar cells made from such silicon materials can also achieve a very high resistivity, along with minimal attenuation. Therefore, it is still necessary to conduct many more experiments to determine what types of impurity components can be used to produce solar cells. However, a more practical conclusion is that if the total amount of impurities (the more than twenty elements commonly measured by ICP) is less than 3 ppm, with boron at 0.3–0.5 ppm, phosphorus below 0.5 ppm (corresponding to boron), iron, aluminum, and calcium each below 0.5 ppm, and the total amount of other impurities below 1 ppm, then such materials can be used to manufacture qualified solar cells. The reason why this conclusion is quite practical is that, in metallic silicon, the most abundant impurities are iron, aluminum, and calcium, while the most difficult-to-remove impurities are boron and phosphorus. If these five impurities can be reduced to the levels mentioned above, then virtually all other impurities will have been removed as well. The degradation issue of polycrystalline silicon using physical methods: Degradation in photovoltaic cells was first observed in thin-film solar cells, and it was believed at the time to be caused by lattice defects. Later, attenuation was observed in cast polysilicon cells, and it was believed that impurities present therein also had an impact on the attenuation. But exactly what these impurities are and how they exert their influence – although some experts have their own theories – none of them can fully explain the phenomena related to polycrystalline silicon produced by physical methods. It is generally believed that deep-level impurities form carrier recombination centers, thereby causing attenuation. Regarding the main substances responsible for the formation of deep-level impurities, the current authoritative view is that boron-oxygen complexes play a role. In addition, there are studies suggesting that certain heavy metals contribute to the formation of deep levels, while others attribute the degradation effect to boron-iron complexes. However, none of these claims are supported by solid experimental results, nor can they explain the attenuation phenomenon of current physical-method polysilicon. For example, the main reason for inferring that it is a boron-oxygen complex is that no significant attenuation was observed in the N-type material, which leads to the conclusion that boron is at work. But in reality, today many single-crystalline silicon factories use recycled material to produce single crystals; even when the oxygen content in the material reaches 20 ppm and boron is also present, no degradation is observed. Furthermore, the phenomenon of efficiency increasing again after annealing following attenuation cannot be explained by the theory of boron-oxygen complexes. Here, it is not a denial of some experts at home and abroad. The reason is that the impurities studied by many experts at present are those present in high-purity silicon with a purity level of 9N or higher, and their research focuses on the effects of one or two types of such impurities. Within this concentration range, the conclusions drawn by those experts may be correct, and the experimental data can also provide some support for their hypotheses. Physical-method-produced polysilicon typically has a purity level around 6N, and its impurity concentration is several hundred times higher than that of the materials mentioned above; therefore, the role played by impurities and their mechanisms may be completely different, and many new effects may occur. Pro Company is currently working with several research institutions on studies regarding the degradation of polycrystalline silicon produced by physical methods. Preliminary research results indicate that the degradation of such polycrystalline silicon is related to impurities present in the material; these impurities are not necessarily boron-based, but are more likely to be metallic impurities. However, as the primary acceptor impurity, if boron combines with any other element, it will lead to a reduction in carriers, thereby also contributing to a decline in the photoelectric conversion efficiency. Currently, new theoretical models are being developed for CP polysilicon materials, with breakthroughs expected in the near future. Progress on Pro’s CP method: Some online visitors have said that Pro Company itself does not have any products available. Although this might be an attempt to provoke a reaction, it still ends up working against them. I will now update everyone on Pro’s progress. Shanghai Pro New Energy Co., Ltd. has been using physical methods for the production of UMGS, developing its own unique patent-based technology and manufacturing processes that are entirely owned by the company. To distinguish itself from other competitors, Pro named its method the CP method (Chemical Physics Method). As of June 30, 2008, Pro Company had used its own CP method for physical purification to produce polycrystalline silicon ingots with a grade of 5.5N, a cross-sectional size of 700x700, and a weight of 250 kilograms. Among them, the boron content is 1.3–1.5 ppmw, phosphorus is 2–4 ppmw, and the content of other metal impurities is 1.5–2 ppm. The polycrystalline silicon wafers obtained by directly cutting cubic slices from polycrystalline silicon ingots have a resistivity of 1.6–2.7 ohm-cm, and a minority carrier lifetime of 2–3 microseconds; the parameters seem to be quite good ； However, when using polycrystalline silicon wafers alone to manufacture cells, the efficiency is only 8–10%. By using the aforementioned material preparation methods but employing this type of silicon material for single-crystal pulling and slicing, the resistivity of the silicon wafers ranges from 0.5 to 3 ohm-cm, and the minority carrier lifetime is greater than 2 microseconds; these values are now on par with those of single-crystalline silicon wafers produced by chemical methods. However, the solar cells produced from these silicon wafers have a conversion efficiency of only 16%, which is still somewhat lower compared to the 17% achieved by chemical methods. Moreover, there is also the issue of attenuation. Based on the results of attenuation tests conducted at Sun Yat-sen University and Xiamen University, the conversion efficiency can remain stable between 13% and 15% after exposure to strong light, with a minimum level of no less than 13%. This result is slightly lower than that published by CSI, but it still places it in a leading position among domestic silicon materials. Figure: Resistance rate and minority carrier lifetime distribution graphs of single-crystalline silicon wafers produced by Pro UMGS. (a) Resistance rate distribution graph of single-crystalline silicon wafers made by Pro UMGS; (b) Minority carrier lifetime distribution graph of such wafers. Caption: At the Shanghai International Photovoltaic Conference on June 8, 2008, a domestic company that produces polycrystalline silicon using metallurgical methods displayed graphs showing the minority carrier lifetime distribution of its single-crystalline silicon. There was a significant difference between the central area and the surrounding areas in those graphs; the company jokingly referred to this pattern as a “monkey butt” and claimed that all silicon materials produced by metallurgical methods would exhibit such a pattern. However, as can be seen from the figure above, Pro Company’s metallurgical polysilicon does not exhibit this phenomenon, which indicates that the \"monkey butt\" effect is not inevitable or universal. According to the analysis by experts from the Pro Renewable Energy Research Institute and those from collaborating research institutions, the formation of the \"monkey butt\" phenomenon is still attributed to an excessive amount of impurities; these impurities accumulate from the edges toward the center of the crystal rod during crystal pulling, thereby reducing the lifetime of the minority carriers in the central part. It should be noted that the silicon materials mentioned above were used by Pro Company in June 2008 to produce monocrystalline silicon wafers; these are not the current results, let alone the final results. After June 30, 2008, PRO conducted another round of purification tests, resulting in further reduction of impurity levels. Currently, it is working together with relevant downstream manufacturers to analyze the produced monocrystalline silicon, polycrystalline silicon wafers, and solar cells; the results will be published at a later date. Currently, PRO is still conducting further tests to continue improving purity and reducing impurities, in order to lower attenuation and increase the conversion efficiency once it has been stabilized. It is hoped that in a short period of time, our UMGS will be able to reach a level equal to or similar to that of SGS’s international Siemens method. Additionally, I’d like to provide some information on energy consumption. Pro Company currently consumes 20,000 kWh of electricity per ton of polysilicon produced (this figure could be reduced to 15,000 kWh in the future). Converted to per solar cell, this amounts to 0.2 kWh. Assuming each cell generates 2 watts of power, it would take only 100 hours of power generation to recover all the electricity used in producing the polysilicon. At 5 hours of generation per day, this means it would take around 20 days. Of course, the subsequent processes of pulling single-crystal silicon and manufacturing solar cells also require electricity, so it takes longer to actually recover the electrical energy. However, the electricity consumption in these later stages is much lower than that involved in producing polysilicon. However, considering that the energy consumption for producing polysilicon using the Siemens method is between 150,000 and 250,000 kWh per ton, which is 7 to 12 times higher than that of the physical method for polysilicon purification, the advantages of the physical method cannot be ignored just based on this factor alone. Furthermore, if we consider the entire production process starting from silica ore, including the smelting of metallic silicon, extracorporeal refining, vacuum melting, polycrystalline silicon ingot casting, single-crystalline silicon drawing, slicing, cell production, and finally solar cell modules, the total energy consumption using Pro’s CP method does not exceed 0.5 kWh per cell; therefore, the recycling time is 250 hours. Even in areas with category 2 lighting conditions, it only takes 3 months to recover all the energy consumed in production. Some experts in the industry claim that over the entire lifespan of solar cells (20 years), the electricity generated from solar energy is not sufficient to cover the amount of energy required to produce those cells; this fallacy must be put to an end. Regarding the issue of pollution emissions that concerns many people, Pro will ensure that no harmful pollutants are emitted from its factories in Shanghai (in fact, vacuum melting does not generate any pollution at all). In order to completely change the perception of high energy consumption and pollution associated with polysilicon production, Pro will begin building a 1,000-ton CP-process polysilicon production demonstration plant in Nanhu District, Shanghai, by the end of 2008. This plant will gradually adopt wind and solar power for electricity generation, with full utilization of such sources achieved within two years; it will thus be the world’s first polysilicon plant that relies entirely on wind and solar power. Here, those interested in using UMG materials to make batteries are also welcome to communicate with Pro. There is never an end to talking about the physical methods for purifying polysilicon; I will share some of my insights with you in the future. For now, I will put a stop here, but we will continue to discuss related topics going forward. (End of text)