Silicon, a fundamental element, underpins much of modern technology, from smartphones to solar panels. Its widespread use, surprising given its unreactive nature, stems from its semiconducting properties. Silicon is the second most abundant element in Earth’s crust, making up about 28% of its mass, primarily found as silicon dioxide, or silica, in common sand. This article details the multi-stage process of transforming sand into the high-purity silicon required for electronics.
The Raw Material
Sand, or silicon dioxide (SiO2), also known as silica, serves as the primary raw material for silicon production. Silica is a naturally abundant compound, making up a significant portion of the Earth’s crust in various forms, including quartz. Its availability makes it an economically viable starting point for large-scale manufacturing.
The stability of silica requires significant energy to break its strong bonds and liberate pure silicon. The silica used for electronic-grade silicon production must be relatively high quality, with low concentrations of metallic impurities. This initial purity helps reduce the burden on subsequent purification stages.
Basic Production: Metallurgical Grade Silicon
The first major step in transforming silica into silicon involves creating metallurgical grade silicon (MG-Si). This process uses carbothermic reduction, a chemical reaction that removes oxygen from silica with a carbon-based reducing agent. High-purity quartz sand and carbon sources like coke, charcoal, or wood chips are fed into a large electric arc furnace.
Inside the furnace, extremely high temperatures ranging from 1700°C to 2000°C are generated by electric arcs. Silicon dioxide reacts with carbon, yielding molten silicon and carbon monoxide gas. The simplified chemical reaction is SiO2 + 2C → Si + 2CO. This initial silicon product, once cooled, is about 98% to 99% pure. While suitable for industrial applications like aluminum alloys, this metallurgical grade silicon contains too many impurities for electronic uses.
Advanced Purification: Semiconductor Grade Silicon
Metallurgical grade silicon is not pure enough for electronic devices, which require impurity levels measured in parts per billion. Therefore, MG-Si undergoes extensive purification to become semiconductor grade silicon (SG-Si), also known as polycrystalline silicon. The Siemens process is the most common method for achieving this ultra-high purity.
The first stage of the Siemens process converts MG-Si into trichlorosilane (SiHCl3). This occurs by reacting finely powdered metallurgical silicon with hydrogen chloride gas at around 300°C in a reactor. Trichlorosilane is chosen for purification due to its low boiling point and high volatility.
After formation, crude trichlorosilane undergoes rigorous purification through multiple stages of fractional distillation. This separates trichlorosilane from other volatile compounds and impurities like chlorides of iron, aluminum, boron, and phosphorus, based on their differing boiling points. The result is pure trichlorosilane, with impurity levels reduced to parts per billion.
In the final step of the Siemens process, highly purified trichlorosilane gas is mixed with hydrogen and introduced into large deposition reactors. It flows over thin silicon rods, heated to 1100°C to 1150°C. The trichlorosilane decomposes on the hot surface, depositing layers of high-purity silicon. This chemical vapor deposition (CVD) process grows the rods into large, high-purity polycrystalline silicon ingots, achieving purity levels often exceeding 99.9999% (“six nines”) and sometimes reaching 99.9999999% (“nine nines”).
From Polycrystalline to Monocrystalline Silicon and its Uses
High-purity polycrystalline silicon from the Siemens process is transformed into monocrystalline silicon, a single, continuous crystal structure without grain boundaries. This transformation is crucial for electronic applications and is primarily achieved using the Czochralski method. Polycrystalline silicon is melted in a quartz crucible at temperatures exceeding silicon’s melting point of 1414°C.
A small, precisely oriented silicon seed crystal is dipped into the molten silicon. This seed crystal is pulled upwards while rotating, allowing the molten silicon to solidify around it in a continuous, ordered crystal lattice. Control over temperature gradients, pull rate, and rotation speed during growth ensures the formation of a large, cylindrical single-crystal ingot, often called a boule.
These large monocrystalline silicon ingots are sliced into thin wafers, forming the foundational substrates for microelectronic devices. This high-purity monocrystalline silicon is indispensable for manufacturing computer chips, including microprocessors and memory chips, due to its uniform electrical properties. It is also used in the production of solar cells, which convert sunlight into electricity. The multi-stage purification and crystallization processes meet the stringent purity and structural requirements for these technological applications.