Ferrosilicon Production: Charge Mixing, Furnace Control, Tapping

What Ferrosilicon Is and Which Grades Are Produced

Ferrosilicon (FeSi) is a master alloy of iron and silicon, supplied as cast lumps, granules or powder, and consumed mainly as a deoxidiser and alloying addition in steelmaking and as an inoculant in the foundry industry. It is not a single product: producers ship a family of grades whose identity is fixed by silicon content and by the residual levels of aluminium, carbon, sulphur and phosphorus. The Chinese national grade table GB/T 2272 covers FeSi75, aluminium-controlled variants of FeSi75, FeSi72 and FeSi65, and these designations are widely used as the reference point in international enquiries for deoxidiser-grade material.

Grade selection is driven by three questions on the buyer side: how much oxygen must be removed, what silicon level the finished steel specification allows, and how tight the aluminium tolerance is. FeSi75 delivers the greatest deoxidising power per tonne charged and is the default for rimmed and killed carbon steels. FeSi72 and FeSi65 are used where the silicon addition must be smaller, or where a lower cost per unit of silicon is acceptable. Low-aluminium variants are specified for deep-drawing grades, electrical steels and certain foundry irons in which aluminium residuals must be held below 1.5%, 1.0% or 0.5%.

Raw Materials and Their Preparation

Three raw material streams decide the chemistry, the power consumption and the unit cost of ferrosilicon: a silicon source, a carbon reductant and an iron source.

Silicon source. Quartz or quartzite with silica content normally at 97% SiO2 or above, low alumina and low alkali content, is the standard feed. Alumina in the charge is the main route by which aluminium enters the alloy, so material with Al2O3 held near or below 0.5% is preferred for low-aluminium grades. Lump size is usually controlled within roughly 20-120 mm; an excessive fines fraction blocks gas escape and destabilises the burden.

Carbon reductant. Metallurgical coke, petroleum coke or a blend of the two supplies the carbon. Fixed carbon is typically specified at 82-85% or higher with ash kept low, because ash brings alumina, silica and iron oxide into the reaction zone outside the calculated ratio. Reductant size is generally 5-25 mm so that the carbon particles remain in contact with the quartz throughout the descent of the charge.

Iron source. Steel scrap, mill scale or iron turnings supply the iron units. The iron requirement of FeSi75 is modest, and the scrap must be clean, dry and free of non-ferrous contamination and heavy rust scale that carries uncontrolled oxygen.

Every stream is dried before charging. Moisture in the burden consumes heat, promotes violent gas evolution and upsets the surface of the charge; a routine moisture check on the day's receipts is a basic control point.

Input	Typical specification	Typical size	Function
Quartz / quartzite	SiO2 97% min, Al2O3 0.5% max preferred	20-120 mm	Silicon source
Metallurgical / petroleum coke	Fixed carbon 82-85% min, low ash	5-25 mm	Carbon reductant
Steel scrap / iron turnings	Clean, dry, free of non-ferrous metal	10-80 mm	Iron units
Electrode paste	Per the electrode supplier's grade	Self-baking blocks	Current conduction

Charge Calculation and Why Uniform Mixing Decides the Furnace Condition

The ratio of quartz to coke is not a rule of thumb; it is calculated from the reduction stoichiometry of the silicon oxides present in the actual raw materials, adjusted for the fixed carbon and ash of the reductant in use and for the operating practice of the furnace. The composition of every batch charged must therefore match the calculated batch ratio, and every batch must be weighed rather than estimated.

Uniform mixing of the charge is one of the most important operating conditions in ferrosilicon smelting. If a batch is mixed unevenly, or if material is fed selectively so that some parts of the furnace receive quartz-rich charge and others coke-rich charge, the result is a locally unbalanced reaction zone. A quartz excess leaves unreacted silica, cools the reaction zone and thickens the burden; a coke excess drives carburisation, builds silicon carbide in the hearth and raises the carbon content of the metal. Both conditions disturb the descent of the charge, change the gas permeability of the burden and cause the furnace condition to deteriorate. The consequences appear at the taphole as erratic tapping, higher fines, off-grade silicon or excessive electrode consumption.

Practical countermeasures are simple and mechanical: weigh each component, mix the batch on the charging floor or in the charging system before it reaches the furnace, feed the furnace in successive small batches rather than in large dumps, and distribute the batch across the surface instead of dumping it in one place. Small, frequent batches make it easier to hold a flat, even burden surface and to keep the composition of the charge in the furnace uniform in both the radial and the vertical direction.

Reduction Reactions Inside the Submerged Arc Furnace

Ferrosilicon is produced in a submerged arc furnace in which carbon reduces silica at high temperature. The simplified overall reaction is SiO2 + 2C = Si + 2CO. In practice the process proceeds through several overlapping steps: silica is first reduced to silicon monoxide gas and silicon carbide, as in SiO2 + C = SiO + CO and SiO2 + 3C = SiC + 2CO, and the silicon carbide then attacks further silica according to SiC + SiO2 = 2Si + CO. Silicon monoxide generated in the hot zone rises through the burden, where it is partly recaptured by carbon and partly lost with the off-gas as a fine silica fume.

The furnace therefore has to be understood as a set of zones rather than a single melt: a hot reaction zone around the electrode tips where silicon is formed and where silicon carbide accumulates if the coke balance drifts high; a cooler upper burden where preheating and the early reduction steps take place; and a hearth region where the alloy collects before tapping. Energy input for FeSi75 is in the range of roughly 8,000-9,000 kWh per tonne of alloy in typical commercial practice, with the exact figure depending on raw material grade, furnace size and the efficiency of the burden control described above.

Burden Height, Electrode Position and Crucible Control

The height of the material surface must be held within a narrow band. If the surface runs too high, the electrodes are pushed upward, the hot zone moves closer to the surface, heat is lost to the off-gas and the reaction slows. If the surface runs too low, the burden collapses around the electrodes, gas channels form, and the heat generated is again not used effectively. Either deviation degrades the furnace condition.

Correct surface control lets the electrode sit deeper in the charge. A deeper electrode position raises the temperature in the reaction zone, widens the crucible, and gives the descending charge more time in the preheating zone so that the reaction accelerates. Operators watch the burden height together with electrode current, off-gas temperature and the appearance of the surface, and adjust the charging rate rather than reacting only after the electrodes have already moved. Keeping the burden loose, avoiding hang-ups on the furnace walls and maintaining a clean taphole are all part of the same control loop: a taphole that taps slowly raises the metal level, shortens the effective reaction volume and forces the electrodes back up.

Tapping, Casting, Cooling and Crushing

Molten ferrosilicon is tapped periodically through the taphole into ladles or directly into casting beds. The alloy is cast into shallow layers so that it solidifies quickly and uniformly; slow cooling of thick sections produces coarse, hard ingots that are difficult to break and increase the generation of fines. After solidification the metal is broken, crushed and screened to the sizes required by the order.

Because ferrosilicon fractures rather than deforms, the crushing and screening circuit has a direct effect on yield. Over-crushing increases the proportion of fines below the ordered minimum size, while under-crushing leaves oversize lumps. Magnetic separation after crushing removes tramp iron from the crusher liners and broken equipment, and dust extraction controls the respirable fraction. Fines generated in handling can be briquetted or sold as a separate powder product rather than being blended back into the lump specification.

Sizing, Chemistry and Typical Commercial Grades

The table below summarises the grade and size pattern that dominates commercial supply. Values are typical for the trade and the finished composition is always tied to the agreed standard grade table for the specific order.

Grade	Silicon (%)	Aluminium max (%)	Carbon max (%)	Sulphur max (%)	Typical size (mm)
FeSi75	74.0-80.0	2.0	0.2	0.02	10-100, 10-50, 50-100
FeSi75, low-aluminium variants	74.0-80.0	1.5 / 1.0 / 0.5	0.2	0.02	10-100, 10-50
FeSi72	72.0 min	2.0	0.2	0.02	10-100, 3-10
FeSi65	65.0 min	2.0	0.2	0.02	10-100, 3-10
Granules and powder	As per grade	As per grade	As per grade	As per grade	0-1, 1-3, 3-8

Phosphorus in steelmaking grades is normally held at 0.04% max, and the trade sizes most often ordered are 0-1 mm, 1-3 mm, 3-10 mm, 10-50 mm, 10-80 mm and 50-100 mm. Oversize and undersize tolerances, expressed as the permitted percentage outside the nominal range, are agreed in the purchase specification, and the particle size determination method referenced for ferroalloys is GB/T 13247. Packaging follows GB/T 3650: lump material is commonly shipped in 1,000 kg flexible intermediate bulk containers with a liner, or in 25 kg and 50 kg bags for foundry use, and each unit is marked with the grade, the lot number, the net mass and the producer's identification.

Sampling, Chemical Analysis and Acceptance Testing

A ferrosilicon lot is accepted on chemistry, size distribution and freedom from contamination, and the credibility of the acceptance depends entirely on how the sample was taken. International practice for ferroalloy sampling is described in ISO 3713, while GB/T 4010 covers the taking and preparation of samples for chemical analysis of ferroalloys. Increments are drawn from a statistically adequate number of positions across the lot, crushed to the required nominal top size of the analytical sample, riffled and split so that the laboratory subsample represents the lot rather than a single lump.

Chemical analysis of ferrosilicon is carried out according to the GB/T 4333 series, which provides methods for the determination of silicon, aluminium, carbon, sulphur and phosphorus. In practice the receiving inspection covers: silicon content within the grade range; aluminium against the agreed maximum for low-aluminium grades; carbon, sulphur and phosphorus against the limits above; moisture on the as-received mass; particle size distribution; and a visual check for slag, refractory, wood, plastic and other foreign matter. A certificate of analysis accompanies each lot and a retained sample is kept for a defined period so that a dispute can be resolved against the original material rather than against a re-sampled portion.

Cost Drivers Buyers Should Understand

Ferrosilicon is an energy-intensive commodity and its price structure follows the production physics. Electricity is the largest single cost element, so smelting capacity located in regions with low seasonal power tariffs and stable supply has a structural advantage, and tariff windows move offers through the year. Quartz quality and reductant price set the second layer of cost: a clean, low-alumina quartzite costs more per tonne but reduces the consumption of reductant and the risk of off-grade metal. Grade and specification add the third layer, because controlling aluminium to 1.0% or 0.5% requires a tighter selection of raw materials and more careful furnace operation than producing standard FeSi75, and is priced accordingly. Finally, size fraction, packaging type and inland transport behave as separate lines in a quotation: granulated material produced by an extra crushing and screening step carries a different cost from lump, and a lined one-tonne bag costs more than a 50 kg woven bag.

For comparison of offers, the useful basis is the delivered cost per tonne of contained silicon for a defined grade and size, with the delivery term stating clearly whether the price is ex-works or freight-inclusive and whether the quoted value includes duties. Two prices that differ by a few percent per tonne of alloy can reverse when converted to cost per unit of contained silicon and when size distribution and packaging are brought to the same basis.

FAQ

Q: What raw materials are used to produce ferrosilicon?
The three essential inputs are a silicon source such as quartz or quartzite with about 97% SiO2, a carbon reductant such as metallurgical or petroleum coke with a high fixed-carbon content, and an iron source such as clean steel scrap or iron turnings. Electrode paste is also consumed as a process material.

Q: Why must the furnace charge be mixed uniformly?
Because the quartz-to-coke ratio is calculated for the batch as a whole. An unevenly mixed or selectively fed charge creates local excesses of silica or of carbon, which disturb the reaction zone, thicken or carburise the burden and cause the furnace condition to deteriorate, with off-grade silicon, erratic tapping and higher fines as the visible results.

Q: What is the typical silicon content of FeSi75 ferrosilicon?
FeSi75 is specified in the range of 74.0-80.0% silicon, with aluminium controlled to 2.0% max in the standard grade and to lower maxima such as 1.5%, 1.0% or 0.5% in the low-aluminium variants. Carbon is normally 0.2% max and sulphur 0.02% max.

Q: Which chemical reactions produce silicon in the furnace?
The overall reduction is SiO2 + 2C = Si + 2CO. In the furnace this proceeds through intermediate steps including SiO2 + C = SiO + CO and SiO2 + 3C = SiC + 2CO, followed by the attack of silicon carbide on silica, SiC + SiO2 = 2Si + CO. Silicon monoxide gas formed in the hot zone rises through the burden and is partly recaptured by carbon.

Q: What particle sizes of ferrosilicon are commonly supplied?
Standard trade sizes include 0-1 mm, 1-3 mm, 3-10 mm, 10-50 mm, 10-80 mm and 50-100 mm, with lump material normally supplied as 10-100 mm. The permitted oversize and undersize fractions are agreed in the specification and determined by the particle size test method for ferroalloys, GB/T 13247.

Q: How is a ferrosilicon lot sampled and accepted?
Sampling follows ISO 3713 and the sample for chemical analysis is taken and prepared under GB/T 4010, with increments drawn from many positions across the lot and reduced by riffling and splitting. Chemistry is determined by the GB/T 4333 series. Acceptance is based on the certificate of analysis for the lot, the measured particle size distribution, moisture and freedom from foreign matter, with a retained sample kept for dispute resolution.

Q: How much energy does ferrosilicon smelting consume?
In typical commercial practice, producing one tonne of FeSi75 consumes on the order of 8,000-9,000 kWh of electrical energy, together with approximately 1.8-2.0 tonnes of quartzite and 0.9-1.0 tonnes of coke as a reductant. The figures vary with raw material quality, furnace size and the quality of burden control.