The main difference between the crucible descending method and the casting method is that the melting and crystallization of the raw materials of the former are carried out in the same crucible; the melting process of the latter is carried out in the first crucible, and the crystallization is carried out in the second crucible, as shown in Figure 1. Show. The silicon raw material is melted in the first uncoated quartz crucible, and then the molten silicon is poured into the second quartz crucible coated with silicon nitride (Si 3 N 4) film on the inner wall to crystallize the molten silicon. In addition, the crucible descending method is to pass the crucible containing molten silicon through the hot zone of the heating coil to crystallize; while the casting method is to control the temperature by adjusting the power of the heating wire. During the crystallization process, the crucible itself does not move. The liquid interface is buried under the melt, so the influence of temperature fluctuations and mechanical instability can be minimized. These two solidification methods from liquid to solid are called directional solidification (directional solidification), which is easy to cause columnar crystal growth. Therefore, chips cut from the same polysilicon crystal block will Have the same defect structure, such as grain boundaries and dislocations.
Figure 1 Using the casting method to make polysilicon
Figure 1 Polysilicon is manufactured by casting method. The silicon raw material is melted in the first quartz crucible, and then the molten silicon is poured into the second quartz crucible coated with silicon nitride (Si3N4) film on the inner wall. Compared with the crucible descending method, the casting method has shorter crystallization and cooling time.
Figure 2 The crystallization process of molten silicon in the casting method
Figure 2 is a schematic diagram of the front and cross-section of the molten silicon crystallization process in the casting method. Adjusting the power of the heating coil can change the temperature gradient of the molten silicon. The bottom gradually solidifies upwards; then the grain boundaries will grow laterally in a dendritic crystalline manner (lateral growth), forming a polysilicon layer (see Figure 2(b)). Eventually, the product grain boundaries will gradually grow from the polysilicon film to the liquid surface, forming polysilicon bulk materials, and the dendritic crystals will also become a wide range of grain boundaries, as shown in Figure 2(c).
At present, Deutsche Solar GmbH in Germany and Куocerа in Japan both use the casting method to grow polycrystalline silicon bulk materials, the mass can reach 250~400 kg, and the length, width and height are about 70 cm × 70 cm × 30 cm. German Deutsche Solar GmbH uses the casting method to grow Polycrystalline silicon bulk material, from 1997 to 2004, the mass of the bulk material has increased from 180 kg to 330 kg. Compared with the traditional method, the new type of casting method developed by Deutsche Solar GmbH can save about 30% of the growth time, and the mass of polysilicon bulk material can reach 400kg.
What are the characteristics of the crucible descending method?
The advantage of polysilicon is that it is cheap, and the bulk shape is mostly cube or cuboid, while monocrystalline silicon is mainly round or nearly square. Therefore, for general manufacturing, polycrystalline silicon has a better material usage rate than monocrystalline silicon. The disadvantage is that the conversion efficiency is slightly lower than that of monocrystalline silicon. The commonly used polycrystalline silicon bulk growth methods include: crucible descending method, casting method, heat exchange method and edge-limited flake crystal growth method. The following mainly introduces the characteristics of the crucible descending method.
The crucible descending method is also called the Bridgman-Stockbarger method. The characteristic of this method is to let the melt cool and solidify in the crucible. The solidification process starts from one end of the crucible and gradually expands to the entire melt. The crucible can be placed vertically or horizontally. The solidification process is completed through the solid-liquid interface. The interface can be moved by moving the crucible or moving the heating coil. The crucible descending method can be used to grow optical crystals (such as LiF, MgF2, CaF2, etc.), scintillation crystals (such as NaI(TI), Bi4Ge3O12, BaF2, etc.), laser crystals (such as Ni 2- :MgF2 , V2+: MgF2 etc.).
Since 2004, companies have begun to use the crucible descending method to grow polycrystalline silicon bulk materials, with a growth rate of up to 10 kg/h. The growth method is as follows: the silicon raw material is placed in a quartz crucible, and the inner wall of the crucible is coated with a layer of nitride For the silicon (Si3N4) film, the melting and crystallization of the raw materials are carried out in the crucible. The main purpose of plating silicon nitride film is to prevent polysilicon from sticking to the quartz crucible during the crystallization process, causing the crucible or silicon crystal to crack. Increase the temperature to melt the silicon raw material, and gradually move the crucible down so that the bottom of the crucible passes through the area of higher temperature gradient. The whole crystallization process will begin to crystallize from the bottom of the crucible and gradually extend upward. The solid-liquid interface will gradually move upward as the crucible position drops to complete crystallization. Graphite resistance heating is generally used, and the insulation system is graphite tube and molybdenum tube.
Figure 1.1 is a schematic diagram of the crucible descending method
Figure 1.1 The crucible descending method, the melting and crystallization process simultaneously exist in a quartz crucible coated with silicon nitride (Si3N4) film on the inner wall. The crystallization process is to slowly move the molten silicon and crucible below the heating coil to leave the coil. The crystallization process is The report is complete.
Figure 1.2 shows the relationship between temperature and crucible moving position.
Figure 1.2 The relationship between the temperature of the crucible descending method and the crucible moving position, the crystal crystallization process occurs in the region of the temperature gradient in the figure
The silicon crystal grown by the crucible descending method needs to have an appropriate thermal field, including the resistance heater, insulation material, the size of the crucible and the relative position of the heater, adjust these factors to make the temperature gradient of the melting zone smaller, and the crystallization zone The temperature gradient becomes larger. The shape and position of the solid-liquid interface of the crucible descending method are closely related to the integrity and defects of the crystal. Generally speaking, the solid-liquid interface can be divided into three shapes: convex, flat, and concave. From the perspective of reducing dislocations and other defects and avoiding internal stress, the flat interface is the most ideal situation, but in the actual crystal growth process Middle; The concave interface will cause crystals to grow from the edge of the crucible to the center, easy to form polycrystalline, and impurities and bubbles will form inclusions. Therefore, a convex solid-liquid interface is usually maintained, but the convex interface is likely to cause uneven radial temperature distribution and generate internal stress. During the crystal growth process, as the crucible gradually drops, the part in the high temperature zone decreases, and the part in the low temperature zone gradually increases, which will cause the solid-liquid interface to move to the high temperature zone, and the temperature gradient of the interface will become smaller. At this time, it is easy to appear that the crystallization speed is greater than the crucible falling speed, causing bubbles or inclusions inside the crystal. Generally, it can be solved by increasing the control temperature or reducing the crucible falling speed.
Overall, the crucible descending method has these characteristics.
Advantages of the crucible descending method :
(1) The shape of the crystal can be controlled by the shape of the crucible.
(2) The growth direction of the crystal can be determined by the seed crystal, if there is no seed crystal. The crystal will grow along the optimal direction (preference).
(3) Because the growth environment is closed or semi-closed, the volatilization of melt and dopants can be avoided.
(4) The molten silicon starts to crystallize from the crucible wall, which can prevent the molten silicon from being further contaminated by the quartz crucible.
(5) The operation is simple, and large-size crystals can be grown. A single growth furnace can grow several crystals at the same time, which is suitable for industrial mass production.
Disadvantages of the crucible descending method:
(l) The process of crystal growth takes place inside the crucible, so it cannot be observed.
(2) The high growth rate easily makes the temperature gradient of the crystal too large, causing the crystal to break.
(3) The thermal stress of silicon crystal rods is relatively large, which leads to high dislocation density and uneven grain boundary distribution.
(4) The inner wall of the crucible must be specially coated to prevent the crystal from sticking to the crucible, causing the crucible or crystal to break.
(5) During the crystallization process, internal stress is easily introduced into the crystal from the crucible, so the thermal expansion coefficient of the crucible material should be smaller than that of the crystal, and the inner wall of the crucible must be very smooth to prevent stress.
(6) During the growth process, the crystal does not rotate, so the uniformity of the crystal, especially the doped silicon crystal, is worse than that of the Czochralski method.
