A Leading Ferro Alloy Manufacturer

Analysis of Metal Silicon Production from Slicing Silicon Sludge Using Medium-Frequency Furnace (Part 1)

Home » News » Analysis of Metal Silicon Production from Slicing Silicon Sludge Using Medium-Frequency Furnace (Part 1)

Share

Over the years, I have been involved in the utilization of slicing silicon sludge from the photovoltaic industry to produce metal silicon. Here, I will share an analysis of using intermediate frequency (IF) furnaces for this purpose, based on my experiences and observations. We hope to encourage and support my colleagues in the field.

Silicon metal alloy project
Silicon metal alloy case

Introduction to Silicon Sludge

The photovoltaic industry chain involves processes from polysilicon and monocrystalline silicon to solar photovoltaic panels. The transition from monocrystalline silicon ingots to photovoltaic panels entails a series of processing steps that produce various silicon sludges. The main silicon sludge production processes include the crystal pulling process (producing crystal pulling silicon sludge) and the silicon ingot slicing process (producing slicing silicon sludge). In the silicon ingot wire-cutting equipment, silicon wafers are cut in a water bath, and the silicon powder generated during the process enters the water. The wastewater containing silicon powder is precipitated and filtered to form silicon sludge cakes with a water content of typically 45-60%. These two types of silicon sludge have different silicon contents: crystal pulling silicon sludge has a lower silicon content (60-80% total silicon on a dry basis, 35-65% elemental silicon) while slicing silicon sludge has a higher silicon content (85-97% total silicon, 70-95% elemental silicon). The photovoltaic industry primarily produces slicing silicon sludge, accounting for 85-90% of total silicon sludge. For example, 1 GW of silicon wafers (182mm size) annually produces approximately 1800 tons of slicing silicon sludge with 55-60% water content. In 2021, China’s silicon wafer production was about 227 GW, which translates to an annual silicon sludge production of approximately 400,000 tons.

Methods of Utilizing Slicing Silicon Sludge

Currently, domestic companies utilize slicing silicon sludge primarily through intermediate frequency furnace (IF furnace) smelting. The silicon sludge is dried and then melted in an IF furnace to produce metallurgical-grade silicon. This method borrows from the process of smelting silicon slag in IF furnaces, with key technical characteristics as follows:

Smelting Process in Intermediate Frequency Furnace

(1) Heating Non-magnetic Elemental Silicon to Its Melting Point

Silicon is a semiconductor and is almost non-magnetic at room temperature, with a high melting point of 1436°C. How can non-magnetic, high-melting-point silicon powder be melted in an IF furnace? The reason lies in the fact that liquid silicon, once formed, exhibits metallic magnetic properties, allowing it to be heated by induction. However, silicon powder from room temperature to its melting point of 1436°C cannot be heated by induction due to its lack of magnetic properties. Therefore, a method called “red introduction” is used. A small amount of coke or waste electrode blocks is placed at the bottom of the IF furnace, and a graphite electrode is inserted. Upon electrification, a short circuit forms between the electrode and the coke, creating an arc. Once the arc is established, small amounts of silicon powder are added around the arc. The high temperature of approximately 2000°C around the arc melts the silicon powder, forming silicon melt. Once a certain depth of silicon melt (usually 15-20%) is achieved in the furnace, the electrode is removed, and the presence of the magnetic silicon melt allows the IF furnace to begin induction heating.

(2) Melting Large Quantities of Silicon Powder

Once a certain depth of silicon melt is formed in the IF furnace, induction heating begins to transfer heat to the silicon melt, which then transfers heat to the silicon powder. The silicon powder gradually melts, similar to how flour dissolves in boiling water. When added to the boiling silicon melt, silicon powder forms a semi-molten paste before fully melting. This melting method, relying on heat conduction from the silicon melt, necessitates multiple, small, and uniform additions of silicon powder. If too much silicon powder is added at once, it forms thick layers that cannot be heated by the induction coil, leading to increased oxidation. The oxidation of silicon powder produces silicon dioxide (SiO₂), which has a melting point above 1800°C. This high melting point slag crust complicates further melting. Hence, multiple small additions of silicon powder are required, prolonging smelting time and increasing energy consumption. For example, an 8-ton IF furnace (with a nominal capacity for melting steel) takes 2-2.5 hours to melt 1.4-1.5 tons of dry silicon powder, consuming 3000-3500 kWh per ton of silicon melt. Depending on the elemental silicon content, the yield of silicon melt is usually 600-800 kg, with a high-level operation achieving a silicon recovery rate of 60-70%.

(3) Temperature Control in Smelting

Suppose furnaces are typically used in the steel industry, designed primarily with steelmaking in mind, where the temperature is controlled at 1550-1600°C. At these temperatures, elemental silicon melts, but the slag, mainly composed of silicon dioxide, does not. To melt the high-melting-point slag, higher temperatures are required, generally controlled at 1700-1750°C. This ensures the slag melts and maintains fluidity, increasing energy consumption. IF furnaces use low-refractory lining materials, and the high temperatures accelerate the erosion of the furnace lining, increasing slag volume. Additionally, non-magnetic slag accumulates on the silicon melt, further complicating heating and smelting. The accumulation of slag on the furnace mouth after multiple smelting cycles necessitates regular cleaning, requiring multiple IF furnaces to operate in rotation, increasing investment and energy consumption.

(4) Handling Silicon Melt and Large Volumes of Slag

Suppose furnace smelting of silicon powder produces a significant amount of slag, sometimes equal to or greater than the silicon yield. Factors influencing slag volume include the elemental silicon content of the silicon sludge, oxidation loss during drying and the operational level of the IF furnace. The main sources of slag are the inherent silicon dioxide in the silicon powder and the added slag-forming materials like lime to reduce the melting point of silicon dioxide. Due to the large slag volume, manual slag removal is necessary before extracting the silicon melt, increasing labor intensity and extending the smelting cycle.

(5) Production Intensity and Environment

As described, IF furnace smelting of silicon sludge requires small, frequent additions, monitored by operators. The process involves handling fine silicon powder (2000-2500 mesh, 0.5-2.5 µm), with operators adding powder manually under high-temperature conditions (over 1600°C). The process generates smoke and fumes from decomposing organic matter and moisture, creating harsh working conditions with significant health risks from silicon dust and combustion by-products. Additionally, manual slag removal from the furnace adds to the labor intensity.

Despite these challenging and hazardous conditions, how effective and beneficial is the production of metallurgical silicon from silicon sludge using IF furnaces? Stay tuned for the next installment! To be continued!