This article describes the various abnormalities encountered in the production process of polycrystalline silicon ingots, analyzes the causes of these abnormalities, and proposes some related preventive and improvement measures.

1. Silicon overflow

Polycrystalline silicon ingots include five process steps of heating, melting, crystal growth, annealing, and cooling, in which the silicon material may break due to its contained quartz ceramic crucible during melting or melting, and flows out of the crucible, often referred to as silicon. Liquid overflow. The high-temperature silicon liquid flows over the overflow wire, causing the overflow wire to melt, triggering an overflow alarm, and the system enters emergency cooling. The general overflow occurs in the melting stage and the long crystal stage, especially in the late melting and early crystal growth overflow. Overflow not only means that the furnace has no silicon ingot output, but also loses a few kilograms of silicon material, which in turn causes significant losses or even safety accidents in the hot field components. Therefore, overflow is the most serious and common production of polysilicon casting. abnormal. The possible causes of the overflow of silicon liquid are probably the following.

1) Cracked. The crucible used for holding silicon ingots is a quartz ceramic material. The production methods include injection molding and injection molding. However, regardless of the type of crucible produced, there are defects such as cracks and blowholes. These crucibles are shipped from the factory. In the past, there were usually two or more passes of the developing solution for light transmission, but there may still be flaws in missed inspections. In addition, vibrations or bumps may occur during transportation or during handling, which can cause cracks in the crucible. If these defects are not detected before loading, there is a good chance that silicon overflow will occur during the melting process. Therefore, after unpacking, it should be strictly tested before spraying. It is a more convenient and effective method to use the light source to transmit light through the five surfaces.

2) Loading extrusion. During the charging process, the positions near the corners of the crucible are especially at the four vertical angular positions. If there is a large amount of silicon material near the crucible, special attention must be paid to the space between the silicon materials. Usually, the space is more than 2 cm. Excessive loading may cause flooding. This is because the melting of the silicon material starts from the middle upper part, and the solid density of silicon is 2.33 g/cm3 and the liquid density is 2.53 g/cm3. Once the charge is too congested, the liquid silicon may solidify after it flows to the bottom of the crucible because the temperature is too cold. If there is no space for expansion, the crucible wall will be squeezed, causing the crucible to break and overflow. The closer to the corner, the more concentrated the stress is, the easier it is to overflow due to irrational charging. The overflow site actually appears near the four vertical edges, and the silicon ingot carefully observes the overflow after demoulding. The position corresponds to the inner wall of the crucible, and it is often found that the silicon material presses the silicon nitride coating and the inner wall of the crucible.

3) Unreasonable process parameters. In the process of ingot casting, during the initial period of heating and melting, the longitudinal gradient of the internal temperature of the thermal field is relatively large, and the temperature in the lower middle of the crucible is low for a long period of time. The quartz crucible ceramic material has a faster phase transformation rate above 1300°C. An excessively high heating power or heating rate can cause a large difference in the phase transformation rate of the crucible in the longitudinal direction, a large stress in the crucible wall, and a long-time tensile effect. It is prone to cracks and overflows. Therefore, many ingot processes will be held at about 1200 °C for a period of time, waiting for the temperature is relatively uniform after the upper and lower temperatures, excessive temperature gradient setting will easily lead to overflow.

2, silicon ingot oxidation

The surface of normal silicon ingots is gray steel, but some of the silicon ingots become colored after the furnace is baked, which is due to oxidation of the silicon ingot caused by oxygen in the casting process. The surface of a slight silicon oxide ingot is light yellow or colored. The heavy oxide not only has a colored surface on the silicon ingot, but also has some white particles on the surface of the silicon ingot and the graphite material. The main reason for the oxidation is air leakage, and the common location of the air leakage is the inlet and outlet valve and the growth rod. Some silicon ingots use quartz rods for measuring the long crystal rate. The quartz rods often move during the ingot process. If the seal is not good, it is easy to leak and cause oxidation. In addition, the protective gas used for the ingot is argon. If the oxygen content of the argon is too high, it will cause the oxidation and discoloration of the silicon ingot. Therefore, some companies should check the oxygen content in the argon before filling.

3, silicon ingot stick

A silicon nitride coating needs to be applied before the material is used as a mold release agent for silicon ingots. However, different degrees of sticking will still occur in the production process, and the silicon ingot will stick out a small part, which may cause the silicon ingots to drop corners, or even crack the entire silicon ingot, which will seriously affect the output. The cause of sticking can probably be summarized as the following situations.

1) The silicon nitride coating is too thin. Generally, the thickness of the silicon nitride coating is about 150 μm. If the coating is too thin, the silicon liquid is likely to be punctured from the pinhole or gap of the coating and contacted with the crucible in the process of contacting with the silicon liquid for several tens of hours. Once a large area of ​​silicon puncture reaction phenomenon occurs, it is easy to cause sticky pan.

2) Silicon nitride coating cracks. In the spraying process, if the silicon nitride is deposited too quickly on the crucible wall, the moisture cannot be volatilized in time, and the silicon nitride coating is likely to have a small crack during the subsequent drying process. Once the ingot is charged, the liquid silicone can easily penetrate. Cause stick pan. To avoid sticking, it is a common practice to brush off the visible part of the cracked coating and reapply it.

3) Argon flow is too high. During the melting process, the argon flow rate is too large to cause the silicon fluid to fluctuate. Especially at the end of melting, the larger argon flow rate, together with the boiling of the silicon liquid, will nitride the silicon nitride coating, especially the three-phase interface. Silicon produces a violent flushing action, which causes the coating to fall off. If a sticky pot appears at the solid-liquid interface at the upper surface of the silicon ingot, it is likely that the gas flow is large and the silicone liquid is flushed. Therefore, in many ingot process recipes, the supply of argon is appropriately reduced at the end of melting.

4) Charge the cartridge. During the charging process, if the operation is not proper, the sharp silicon material will damage the silicon nitride coating and cause sticking. In recent years, a new method has been developed to add an appropriate amount of silica sol to silicon nitride. Not only does the yttrium coating not sinter, but also the adhesion of the coating on the crucible wall is strengthened, and the adhesion and stickiness are greatly improved. .

5) Silicon nitride quality. At present, silicon nitride production brands are mainly marketed by Japan UBE and Yantai Tongli. Its silicon nitride particle size is basically within a few micrometers, and more than 90% is alpha crystal phase. If the silicon nitride powder particles are too large or too fine, there may be a greater probability of sticking problems. Î’-phase silicon nitride has a large thermal expansion coefficient, and if it has a large content, it may also cause sticky-pot problems.

4, silicon ingot crack

In production, after some silicon ingots are baked, although there are no exceptions in appearance, after infrared inspection, some cracks may be found, which are slightly more than a few centimeters in length, and occasionally appear in one of the small square ingots. The weight is Through penetration cracks, more than half of small square ingots appear scrapped, which seriously affects the yield of ingots. Cracking may be the reason for the following aspects.

1) Foreign objects fall into the ingot process. Thermal field materials can cause deterioration after a long period of use. Some graphite or C/C composite materials, such as bolts and nuts, can easily fall off into the crucible. In addition, quartz rods used to measure long crystals may be stuck inside the silicon ingot. Because of the different thermal expansion coefficients, silicon ingots that have fallen into foreign materials will crack during subsequent cooling and cooling. In the production process, in order to catch up with the output, it is common to use several furnaces per production before entering the furnace for inspection. The last silicon ingot has just been released. The furnace temperature is still a few hundred degrees Celsius, and the next furnace silicon is thrown in. After each furnace ingot is completed, it enters the inspection of the thermal field, which can greatly avoid this phenomenon.

2) The tap temperature is too high. After the general silicon ingot casting is completed, the temperature of the furnace is reduced to below 400°C, so that the ingot can be opened. If the ingot is taken too early and the temperature of the furnace is too high, the temperature difference between the ingot and the environment after the silicon ingot is baked is large, especially during the winter season. The internal thermal stress of the silicon ingot is too late to release, causing the silicon ingot to crack.

3) The process settings are irrational. After the directional growth is completed, because the temperature difference between the bottom and the top of the silicon ingot is large, the steel cage needs to be closed, and the furnace temperature is maintained at about 1300° C. for annealing. If the annealing time is too short, the large thermal stress inside the silicon ingot cannot be effectively released. Internal cracks may occur during cooling. In addition, for large-scale silicon ingots, too fast cooling process settings can easily lead to cracking.

4) High-temperature silicon ingots come in contact with metal. After the silicon ingot is baked, the surface temperature is still several hundred degrees Celsius. Generally, the temperature is cooled down to about 100 DEG C. to start removing the crucible, so that the silicon ingot is released from the mold, and then the silicon ingot is transferred to the next sandblasting process. In this process, it is impossible to use fixtures to make contact with the silicon ingots. If the temperature of the silicon ingot is still high, the metal with high thermal conductivity is in contact with the silicon ingots, which may induce cracking of the silicon ingot. Therefore, after the silicon ingot is released, try to avoid "high temperature operation", especially in the cold winter season, which can effectively reduce the occurrence of cracking of the silicon ingot, which is also beneficial for the reduction of the rate of subsequent slicing of silicon wafers.

5) Sticking cracks. Viscose tanning is the most common cause of cracks in silicon ingots. Even though some silicon ingots appear to be slightly sticky and the appearance of a few centimeters or even smaller tantalum pieces adheres to the silicon ingots, the silicon ingots are still The occurrence of cracks, especially when the location of stickies occurs at the bottom and side of the silicon ingot, has the highest probability of occurrence. In addition, when large-grain silicon ingots (single-crystal silicon ingots) are grown, the problem of cracking due to sticking defects is more likely to occur, and the silicon ingot is often cracked.

5, infrared detection of shadows and hard inclusions

After the silicon ingot has been squared into small silicon blocks, the defects of the silicon ingot must be detected by an infrared flaw detector. The principle of infrared flaw detection is that the infrared light emitted by a specific light source can penetrate a silicon block of 200 mm depth and then be captured and imaged by an infrared detector. Pure polysilicon crystals barely absorb wavelengths in this band, but if there are defects such as microcrystals, conglomerates, hard inclusions, and cracks in the silicon block, these defects will absorb infrared light and will appear dark in the imaging system. Some of the punctate dark areas that present stripes, clusters, or masses are often referred to as shadows. The formation of shadows is probably caused by the following factors.

1) The growth of crystallites is caused by excessive growth of crystallites. After the directional solidification starts, if the temperature is too low or the longitudinal temperature gradient is too large to form a large number of nucleation centers, the silicon ingot rapidly grows, and then crystallites are generated, and the infrared imaging shows a large area of ​​band-like shadows. The most common shadow in production tends to occur in the silicon block in the center of the silicon ingot. The longitudinal position is most common in the middle and lower parts of the silicon side, precisely because this position is the fastest average crystal growth rate.

According to our production experience, the typical trend of the growth rate of the long crystal is that the steel cage has just opened at the beginning of one or two, the growth rate of the crystal is often below 1cm/h, and the following hours are the fastest, reaching 1.7cm/h-1.9. Cm/h, even more than 2cm/h, gradually stabilizes to 1.1 cm/h-1.5 cm/h after the middle period of growth. The average speed of the entire growth process is about 1.2cm/h-1.3cm/h. If the growth rate exceeds 2 cm/h for a long time, it is easy to form microcrystalline shadows in this area. In the early period of growth, the solid-liquid interface often has a transition process from micro-convex to micro-convex. In this process, the period of rapid growth of crystal growth is relatively easy to generate shadows, especially the high content of impurities. Under the circumstances, impurities do not effectively segregate to produce numerous nucleation centers, thereby forming crystallites. Therefore, setting a reasonable formula process and controlling a reasonable growth rate of crystals is necessary to reduce the proportion of shadows.

2) There are too many impurities in the silicon melt, or can not be fully eliminated, resulting in impurity-type shadows and hard inclusions. If there are too many impurities in the raw material, for example, a large amount of head and tail hides are used to feed the material, etc. After the ingot is opened, the proportion of shadows detected increases significantly. This type of shadow is most common in clusters. In addition, if a high-resolution infrared flaw detection instrument is used, some dispersed dot shadows can be detected in the center of the small square ingot, and the color is lighter. The general diameter is one to several millimeters in size. After polishing of the small square ingots, infrared flaw detection is performed. The shadows are more clearly visible, and a few hundred micrometers or even more fine dot shadows can be found.

After dissolving these cluster-shaped shadows with strong acid, it is easy to get some inconveniences. These contaminants are either black blocks or rod-shaped yellow and transparent. The two often co-exist, and these are often called Hard inclusions. Studies have shown that these black blocks are β-SiC for the inclusion phase and the yellow transparent rod-like inclusion phase is β-Si3N4.

The clustered shadows are removed as defectives in subsequent processing. However, those with small punctiform particles tend to be undetectable or have been intentionally or unintentionally ignored. The Mohs hardness of silicon is 6.5, while the Mohs hardness of the two inclusion phases of β-SiC and β-Si3N4 is 9.2 and 9.0, respectively, which is significantly higher than that of silicon. These two types of inclusions are a serious hazard to subsequent slicing, especially SiC inclusions, because the abrasive used for cutting is also SiC. If the particle size of inclusions is larger than the diameter of the cutting wire, it is very easy to cause breakage in the slicing process. Even if the wire is continuously broken, there may be visible line marks on the silicon wafer, which will seriously affect the output of excellent products. Those finer, hard inclusions, even though the slicing process is normal, will cause severe leakage of these hard inclusions and lower the photoelectric conversion efficiency after the silicon wafer is made into a battery.

The main source of carbon in silicon ingots is the reaction of C or CO vapors with silicon liquid during the high temperature process, and the source of nitrogen is mainly caused by silicon nitride release agent inside the crucible wall or falling into the silicon liquid, or dissolved into the silicon liquid by dissolution. The silicon solution then undergoes phase change and crystal growth during several tens of hours of high temperature. In the top silicon material of some purified silicon ingots, it is often easy to find many visible acicular or rod-shaped β-Si3N4. Therefore, controlling the source of carbon and nitrogen is an effective method for effectively reducing shadows or hard inclusions. For example, adding a composite material cover at the top of the crucible and rationally designing the gas flow path to expel the CO vapor as quickly as possible can reduce the reaction with silicon liquid. Effectively suppress the carbon content of the entire silicon ingot. Adding a certain proportion of silica sol high-temperature adhesive inside the silicon nitride paste can enhance the adhesion of the silicon nitride coating, effectively reducing the loss of the coating and the nitrogen content entering the silicon liquid.

In addition, after the ingot is completed, most of the hard inclusions are within 10 mm of the top of the silicon ingot or in the edge leather. However, after this silicon material is removed and then processed by processes such as sandblasting, it is recycled. With continuous circulation, the inclusions in these recycled scraps are increasing, resulting in an increase in the concentration of inclusions in the silicon liquid after the feedstocks are formed. During the growth of silicon ingots, some inclusions inevitably flow into the silicon ingot due to convection or sedimentation, resulting in hard formation. Inclusions. Therefore, when the proportion of ingredients in the ingredients, such as controlling the ratio of undercuts, can effectively reduce the occurrence of hard inclusions.

6. Concluding remarks

Polycrystalline silicon ingot casting is an important part of photovoltaic solar energy manufacturing. This article describes some common abnormalities or defects encountered in the production process of polycrystalline silicon ingots, analyzes the causes of these abnormalities, and proposes some related preventive and improvement measures. For practical production has a certain guiding significance.

Embedded Adjustable Angle Spotlight

Embedded Spotlight,Led Downlight,Embedded Adjustable Light,Ceiling Light Embedded

FOSHAN CITY KURUI LIGHTING CO.,LTD , https://www.kuruilighting.com

Posted on