What Is Metallic Silicon: Definition, Grades and Production

Definition: Silicon Metal as an Element and as a Traded Commodity

Metallic silicon, more precisely called silicon metal or industrial silicon, is the element silicon in a commercially bulk form. Silicon carries the symbol Si, sits at atomic number 14 in group 14 of the periodic table, and is a metalloid rather than a true metal: it has the grey lustre and the crystalline appearance of a metal, but it is brittle, semiconducting in its pure state and chemically closer to a non-metal in the way it bonds with oxygen and halogens.

As a commodity, metallic silicon is a refined mineral product, not a mined ore. It is made by reducing quartz or quartzite with a carbon reductant in an electric furnace, and it reaches the market as lump, granules or powder with a silicon content that is usually described in the range of about 98% to 99% for ordinary metallurgical grades. The residual elements that matter commercially are iron, aluminium and calcium, and it is their individual limits, not the headline silicon figure, that classify the material. A buyer who orders by "silicon content" alone has not written a usable specification.

The term is also occasionally applied to very high purity material, but the commodity meaning is narrower and should be kept separate. Material destined for wafers is normally called solar-grade or electronic-grade polysilicon and has a different specification system and a different price order of magnitude, so treating the two as one product in a purchasing document invites dispute.

How Industrial Silicon Is Smelted From Quartz

The industrial route is carbothermic reduction in a submerged arc furnace. The charge consists of a silica source, typically quartz or quartzite with a controlled particle size and low levels of iron and aluminium bearing minerals, together with a carbon reductant such as metallurgical coke, charcoal, petroleum coke or a blend of these. The reduction chemistry is summarised by the reaction of silica with carbon:

SiO2 + 2C = Si + 2CO

Above the theoretical balance the process is far from a single clean reaction. Silicon monoxide forms as an intermediate and can escape with the off-gas, silicon carbide can accumulate in the furnace and be consumed later, and the actual carbon demand depends on the reactivity, ash content, moisture and sizing of each reductant. The reaction zone operates at high temperature, commonly described as around 1,800 °C and above, and the furnace is a continuous submerged-arc type in which the electrodes are buried in the charge and the molten silicon is tapped at intervals and cast into beds to solidify.

Because the furnace consumes electrical energy rather than combustion heat, the cost structure of metallic silicon is dominated by the electricity price and by the quality of the charge. Industry figures quoted for specific energy consumption vary with charge purity, furnace design and operating practice, but they are commonly stated in the range of roughly 11,000 to 13,000 kWh per tonne of silicon. That single fact explains why the commercial geography of silicon metal follows cheap and stable power, and why smelters are usually operated against long-term electricity contracts. It is also why a stated "capacity" in tonnes per year means little unless it is qualified by power allocation.

Refining Routes and Purity Levels

Tap furnace silicon is cast, cooled, crushed and sized. Refining for ordinary commercial qualities concentrates on reducing the slag and iron contamination introduced by the charge and the furnace lining, and on controlling the aluminium and calcium balance. Crushing is done in stages with screening between them, and the sizing is deliberately aligned to the customer's melting or milling practice rather than to a single standard lump size.

Higher purity is reached by quite different chemistry. The dominant industrial route converts powdered silicon metal into trichlorosilane, purifies the trichlorosilane by distillation to remove boron, phosphorus and metal chlorides, and then deposits silicon in a high-temperature reactor; the Siemens-type deposition route and the fluidised-bed silane or granule route both operate on that principle. A second family of routes takes a metallurgical approach, using slag treatment, directional solidification, electron-beam or plasma melting to push impurities out of a molten silicon bath, so that upgraded metallurgical grades can be produced without chlorosilane chemistry.

Purity class	Typical silicon content	Dopant and impurity control	Principal use
Metallurgical grade (MG-Si)	About 98-99%	Fe, Al, Ca controlled by grade	Aluminium alloys, silicones, refractories
Upgraded metallurgical grade	Above 99.9%	Boron and phosphorus checked individually	Some solar cell feedstocks
Solar grade polysilicon	Very high purity, ppm to sub-ppm impurities	B, P and C limits per GB/T 25074 type specifications	Photovoltaic wafers
Electronic grade polysilicon	Highest purity, ppb level control	Specified under electronic-grade polysilicon standards such as GB/T 12963	Semiconductor devices

The commercial consequence of this table is straightforward: a supplier that quotes a single "purity percentage" without saying which class it refers to, and without stating how boron and phosphorus are controlled, has not answered the question the buyer actually asked.

Physical and Chemical Properties That Matter to Buyers

Silicon melts at about 1,414 °C and boils at about 3,265 °C, which places it far above the melting range of aluminium and well above common steelmaking temperatures in its own right, so it is added to melts as a solid that dissolves rather than as a pre-melted liquid. Its density in the solid state is around 2.3 g/cm3, noticeably lower than that of molten aluminium and steel, which affects how lump dissolves and where it tends to float in a ladle or furnace bath.

The material is brittle and behaves elastically only over a very small strain; it fractures rather than deforms. In practice this means that the size distribution of a shipment changes during transit, because corners break off during handling, and it is one reason why screened sizing with a stated fines tolerance is written into contracts rather than assumed.

Thermally, silicon has a low coefficient of expansion and good stability, which is the basis of its use in refractory and ceramic bodies and in glass. Electrically, it is a semiconductor whose resistivity falls as temperature rises, and it forms a stable native oxide that protects the surface and, in the melting shop, forms a thin silica skin on stored lumps. Chemically it is inert in ordinary conditions but reacts at temperature with oxygen to give silica, with carbon to give silicon carbide, with nitrogen to give silicon nitride, and with halogens and hydrogen chloride, which is the reaction that opens the silicones industry.

For the buyer, the practical properties are these: melting behaviour in the bath, dissolution rate as a function of size, surface oxidation accumulated during storage, moisture pick-up in fines, and the trace chemistry that determines reaction selectivity in a chemical reactor. Mechanical strength of the lump is not usually specified, but friability is indirectly captured through the fines allowance.

Metallic Silicon Compared With Ferrosilicon, Polysilicon and Silicon Carbide

Four silicon-bearing materials are sold into overlapping markets and are frequently mixed up in purchasing documents. The differences are in iron content, in physical form and in the standard that governs them.

Material	Composition	Governing specification	Main function
Metallic silicon	Si about 98-99%, low iron by grade	GB/T 2881; analysis per GB/T 14849 series	Aluminium alloying, chemical feed, refractories
Ferrosilicon	Iron-silicon alloy, FeSi75, FeSi72, FeSi65	GB/T 2272	Steel deoxidation and alloying, inoculants
Polysilicon	High purity silicon, ppm to ppb impurities	Solar-grade per GB/T 25074; electronic grades per GB/T 12963	Solar cells and semiconductor devices
Silicon carbide	SiC, a compound rather than the element	GB/T 2480	Abrasives, refractories, deoxidation

The functional difference is worth stating plainly. Ferrosilicon carries deliberate iron and is optimised for steelmaking, where iron is not an objection and the price per unit of silicon is what matters. Silicon metal is bought where the iron is an objection or where the silicon must enter a chemical reaction rather than a steel bath. Polysilicon is bought for its electrical behaviour. Silicon carbide is a compound, and its specification work concerns grain size and SiC content rather than silicon content.

Substitution between them is only occasionally practical. In steel deoxidation, silicon metal can replace part of the ferrosilicon addition, but at a different cost per unit of silicon and with a different effect on the iron balance and the slag. In aluminium alloying, ferrosilicon is not a substitute, because the iron it carries is usually a contaminant for the alloy concerned. In the chemical industry, silicon metal is the only suitable feed of the four, and silicon carbide is not interchangeable with it.

What a Correct Specification Sheet Contains

Because metallic silicon is a specification product, the purchasing document should be able to stand as the arbitration reference if a dispute arises. A workable specification names the grade or the individual element limits, the silicon content basis, the particle-size range with an explicit oversize and fines allowance, the packing format and unit weight, the sampling and analysis methods, and the documents that must accompany delivery.

On methods, the normal structure is a chemical grade or element-limit clause referencing GB/T 2881 for industrial silicon, an analysis clause referencing the GB/T 14849 series for the determination of iron, aluminium and calcium, a sampling clause referencing GB/T 4010 for ferroalloys or ISO 3713 for international contracts, a sizing clause referencing GB/T 13247, and a packing clause referencing the general ferroalloy packing rules of GB/T 3650. Where the material is intended for a chemical reactor, the clause set is extended with powder particle-size distribution, surface oxidation and heavy-metal limits, because those variables influence silane yield and can trigger additional product-safety documentation.

On documents, a lot-level certificate of analysis with the actual test results, the lot or heat identification, the packing list and the analysis method quoted per element is the baseline. A declaration of no radioactivity is commonly requested for ferroalloy and silicon shipments. Where an export destination applies specific import controls, the certificate of origin and the tariff classification become part of the same document set, and inspection is normally carried out on a quantity-proportional basis rather than on a single first bag.

Two practical comments complete the picture. First, size tolerances should be agreed as percentages of the consignment, with a stated sampling procedure, because a visual check of one bag proves nothing about the lot. Second, moisture in fines and powders is worth a separate clause: it changes the effective silicon content per tonne delivered, it affects handling, and it is the most common cause of a weight-versus-analysis argument at the port.

Common Misconceptions About Metallic Silicon

Metallic silicon is a metal. It is a metalloid. It looks metallic and conducts electricity, and in commercial language it is called metal silicon, but it cannot be forged or rolled and it fails in a brittle manner. Treating it as a structural metal leads to unrealistic expectations about mechanical properties.

A higher silicon percentage is always better. For a given specification, silicon is normally reported by difference, so a higher reported silicon figure may simply reflect a looser iron, aluminium or calcium result. Grade comparison should be made on the individual element limits, not on the derived silicon number.

Silicon metal and ferrosilicon are interchangeable. They carry different iron levels, different designations and different standard references, and they are consumed by different processes. Using one as a substitute for the other changes the melt chemistry and the deoxidation balance.

All silicon metal is the same feed for chemicals. Reactor performance is sensitive to aluminium, to the iron-to-calcium ratio, to powder sizing and to surface condition. Two shipments meeting the same grade designation can behave differently in a reactor if those secondary variables are not controlled.

Stored silicon metal keeps indefinitely without care. It does not decompose, but lumps slowly oxidise and fines absorb moisture. Grade segregation, cover and moisture-barrier packing for fines are the minimum precautions, and retained lot samples are what make any later comparison possible.

FAQ

Q: Is metallic silicon the same as silicon metal?
Yes. Metallic silicon, silicon metal, industrial silicon and metallurgical-grade silicon are commercial names for the same bulk product, typically about 98-99% Si and graded by iron, aluminium and calcium limits under GB/T 2881.

Q: How is metallic silicon produced?
Quartz or quartzite is reduced with a carbon reductant in a submerged arc furnace, following the overall reaction SiO2 + 2C = Si + 2CO. The molten silicon is tapped, cast, cooled, crushed and screened to the required size.

Q: Why does the electricity price matter so much in this product?
Because the furnace uses electrical energy rather than combustion heat, and industry practice for specific energy consumption is commonly quoted in the range of roughly 11,000 to 13,000 kWh per tonne of silicon, the power tariff is a dominant part of the cost base.

Q: What is the difference between metallurgical grade and polysilicon?
Metallurgical-grade silicon is about 98-99% pure and is used in aluminium alloys, silicones and refractories. Polysilicon is far purer and is produced by chlorosilane purification or comparable routes; solar-grade material is addressed by GB/T 25074 type specifications and electronic grades by standards such as GB/T 12963.

Q: Which tests are used to verify a shipment?
Chemical analysis is carried out under the GB/T 14849 series for iron, aluminium and calcium; sampling follows GB/T 4010 or ISO 3713; particle size is tested under GB/T 13247; and packing terms follow the general ferroalloy packing rules of GB/T 3650.

Q: Can silicon metal be used instead of ferrosilicon in steelmaking?
Only partially. Silicon metal can replace a share of the ferrosilicon addition where a low-iron, low-carbon silicon source is required, but the cost per unit of silicon, the iron balance and the effect on the slag are different, so the substitution has to be assessed per heat rather than assumed.