Where GPC Actually Goes: Carbon Mass Balance in Steel Ladles and Ductile Iron
Short answer
Graphitized petroleum coke is mostly a carbon correction tool: it raises carbon in a steel ladle during secondary metallurgy, or in a grey or ductile iron melt during charge preparation. Everything else - electrodes, powder injection, friction material - is real volume but secondary.
The number that should decide your grade is not fixed carbon. It is nitrogen. A cheap high-nitrogen recarburizer saves about USD 12 per tonne of iron and can cancel that saving at a scrap rate increase of under half a percent. The arithmetic for that is worked out below, and it is the reason we think most recarburizer substitutions are evaluated backwards.
Three worked examples
Supplier data sheets tell you what is in the bag. They do not tell you what it costs per tonne of metal, because that depends on your charge, your tap target and your recovery. Here are three calculations, with every assumption stated so you can substitute your own.
WE-1 · Ladle carbon trim, 100 t heat
Tap carbon 0.06%. Grade target 0.35%. Trim needed = 0.29%.
Carbon required 100,000 kg x 0.0029 = 290.0 kg C GPC required 290.0 / (0.99 x 0.92) = 318.4 kg Per tonne 318.4 / 100 = 3.18 kg/t steel Cost at USD 950/t 3.18 x 0.95 = USD 3.02 / t steel
Recovery of 92% assumed for a stirred ladle addition. If your stirring is weak and recovery is 85%, the same trim needs 3.45 kg/t and costs USD 3.28 - a 8.6% overrun that never appears as a purchasing variance.
That is the whole steel story, and it is why GPC price per tonne matters so little to a steelmaker. A move from USD 950 to USD 1,100 per tonne changes the cost of a 0.29% trim by less than half a dollar per tonne of steel. What matters far more is recovery, because recovery scales the addition directly.
WE-2 · Ductile iron, 10 t induction furnace
Charge: 70% steel scrap at 0.20% C, 30% ductile returns at 3.50% C. Tap target 3.60% C.
Charge carbon 0.70 x 0.20 + 0.30 x 3.50 = 1.19 % Deficit 3.60 - 1.19 = 2.41 % Carbon required 10,000 kg x 0.0241 = 241.0 kg C GPC required 241.0 / (0.99 x 0.90) = 270.5 kg Per tonne = 27.0 kg/t iron Cost at USD 850/t 27.0 x 0.85 = USD 23.0 / t iron
Recovery of 90% assumed for induction work with early charge addition. The same melt on a 90% scrap charge would need about 33 kg/t, which is why high-scrap foundries are the heaviest GPC consumers per tonne of metal.
Note the order-of-magnitude difference between the two applications. A steel ladle uses roughly 3 kg per tonne. A ductile iron foundry uses roughly 27. Same product, nine times the intensity - which means the same price move is nine times more painful in a foundry, and the same quality problem is nine times more consequential.
The nitrogen trap
Here is the calculation we would most like a buyer to run before switching to a cheaper recarburizer. It is also the one almost nobody runs.
Published GPC specifications typically cap nitrogen at or below 0.03 percent. Alternative carbon sources - calcined anthracite, metallurgical coke, some CPC grades - carry nitrogen in the range of several tenths of a percent. Because the foundry addition rate is large, that difference is not small when it reaches the metal.
WE-3 · Same melt, two recarburizers, addition held at 27.0 kg/t
Option A GPC, N capped at 0.03 %
27.0 kg x 0.0003 = 0.0081 kg N/t = 8 ppm in the iron
Option B alternative recarburizer, N at 0.60 % (our assumption)
27.0 kg x 0.0060 = 0.1620 kg N/t = 162 ppm in the iron
Difference 154 ppm of nitrogen, if fully absorbed
Price A at USD 850/t -> 27.0 x 0.85 = USD 23.0 / t iron
B at USD 400/t -> 27.0 x 0.40 = USD 10.8 / t iron
apparent saving USD 12.2 / t iron
Break-even one scrapped 200 kg casting at USD 500 finished value
500 / 12.2 = 41 t of iron
41 t x 5 castings/t = 205 castings
1 / 205 = 0.49 % extra scrap
Read that last line again. If substituting the cheaper material raises your scrap rate by half a percent, the saving is gone. At one percent, you are paying double for the privilege.
We want to be careful about what that calculation does and does not prove. It does not prove that GPC is always the right answer - for many grey iron and low-specification steel applications, nitrogen tolerance is generous and a cheaper recarburizer is the correct commercial choice. What it proves is that a substitution evaluated on price per tonne alone is not evaluated at all.
Where it goes, and at what size
| Application | Particle size | Typical addition | Why that size |
|---|---|---|---|
| Steel ladle carbon trim | 1–5 mm | 1–4 kg/t | Predictable dissolution under argon stirring; the standard band for secondary metallurgy |
| Induction furnace, ductile iron | 1–5 mm | 25–33 kg/t | Large addition, early in the charge; recovery depends on bath motion |
| Grey iron correction | 1–5 mm | 18–26 kg/t | Lower tap carbon target than ductile, so a smaller deficit to close |
| Powder injection and blends | 0–1 mm | process-specific | Rapid dissolution; the fines are also the cheapest grade, though only by USD 20–40/t |
| Furnace charging | 5–10 mm | process-specific | Slower dissolution, lower oxidation loss - chosen where survival matters more than speed |
| Graphite electrode manufacturing | graded | - | One carbon source among several; needle coke is typically the principal aggregate |
The counter-intuitive entry in that table is the fines row. Most buyers assume 0–1 mm material is heavily discounted because it is a by-product of screening. Published 2026 tables put 0–1 mm GPC fines at USD 700–780 against USD 720–820 for 1–5 mm standard - a gap of only USD 20 to 40 per tonne. If your process can genuinely absorb fines, that is a real saving, but it is a small one, and it is not worth re-engineering an addition system to capture.
When GPC is the wrong answer
We would rather lose a sale than sell material into an application where it cannot pay for itself. Four situations where we would talk a buyer out of it:
- The charge already carries the carbon. A cupola or a high-pig-iron charge may need no recarburizer at all, or very little. At 3 kg per tonne the graphitization premium over CPC is immaterial; there is no case to answer either way, so buy on logistics.
- Sulfur and nitrogen tolerances are both generous. If the melt can absorb what a cheaper carbon source carries, GPC's advantages - low sulfur, low nitrogen, fast dissolution - are properties you cannot monetise. Paying for them is a donation.
- The addition point is deep injection into a high-oxygen bath. Carbon that oxidises instead of dissolving is carbon you bought and did not use. Fix the oxygen first; the recarburizer grade is not your problem.
- You are buying by fixed carbon alone. Two materials at 99.0% fixed carbon can behave completely differently at the furnace if their particle size distributions differ. Recovery is a function of surface area and bath conditions, and a certificate of analysis will not tell you that.
What to actually measure
If you take one operational step from this page, log your own carbon recovery. Weigh what goes in, measure the carbon before and after, and divide. It takes a shift to set up and it changes how you buy permanently, because recovery is the multiplier that turns every price on every quote into a cost per tonne of metal.
Put alongside it a monthly independent check on fixed carbon and sulfur, and a nitrogen figure on every incoming certificate. Nitrogen is capped at or below 0.03 percent in published specifications for good material and it is the parameter most often quietly dropped from a COA. It is also, per WE-3, the one that can cost you more than the entire price difference between grades.
Frequently asked questions
What is graphitized petroleum coke used for?Principally as a recarburizer - raising carbon in steel ladles during secondary metallurgy, and in grey and ductile iron foundry melts. It also goes into graphite electrode manufacturing, powder injection and blends, and friction material. The dominant volume is carbon correction in liquid metal, and that is where the economics get decided.
How much GPC does it take to raise carbon by 0.10 percent in steel?
About 1.10 kg per tonne. A 0.10 weight percent trim needs 1.0 kg of elemental carbon per tonne; dividing by 0.99 fixed carbon and 0.92 recovery gives 1.098 kg. Scale linearly for any other trim - a 0.29 percent trim is 3.18 kg per tonne.
How much GPC does a ductile iron foundry use per tonne of iron?
Roughly 25 to 30 kg per tonne in a typical high-scrap charge. On 70 percent scrap at 0.20 percent carbon plus 30 percent returns at 3.50 percent carbon, the charge carries 1.19 percent; lifting it to a 3.60 percent target needs 2.41 percent carbon, which works out to about 27 kg per tonne at 99 percent fixed carbon and 90 percent recovery.
What carbon recovery should I assume for GPC?
Around 92 percent for a well-run stirred ladle addition and about 90 percent for induction furnace work, rising toward 95 percent when material goes in early to a clean, low-oxygen bath. Recovery is at least as much a property of your process as of the material, so your own logged figure should always beat a published one.
Why does nitrogen matter when choosing a recarburizer?
Because a cheaper high-nitrogen recarburizer moves nitrogen into the melt in proportion to the addition size, and foundry additions are large. At 27 kg per tonne, a recarburizer carrying 0.60 percent nitrogen contributes about 162 ppm against roughly 8 ppm from GPC capped at 0.03 percent - a difference that can exceed what a pinhole-sensitive casting tolerates.
What GPC particle size should I buy?
Match it to the addition point. Published routing is 0–1 mm for powder injection and rapid dissolution, 1–5 mm for standard ladle and induction furnace use, 3–8 mm for controlled pickup in steelmaking, and 5–10 mm for furnace charging where slower dissolution and lower oxidation loss matter more than speed.
Is GPC better than calcined petroleum coke as a recarburizer?
For carbon dissolution, yes - GPC has higher fixed carbon, lower electrical resistivity and better dissolution behaviour. Whether that is worth two to three times the CPC price depends on your recovery and your nitrogen and sulfur tolerance. In a low-specification melt with generous tolerances, CPC is frequently the rational buy.
Can GPC be used in an induction furnace?
Yes, with 1–5 mm as the standard size band. Induction work favours early addition into the charge rather than late correction, because there is no argon stirring to assist dissolution. Late additions into a quiet bath give lower and less repeatable recovery, which is the most common source of carbon misses in smaller foundries.
When should I not use GPC?
When carbon is already plentiful in the charge, when your sulfur and nitrogen tolerances are generous enough that a cheaper recarburizer passes, or when your addition point is deep injection into a high-oxygen bath where recovery collapses. In all three you are paying a graphitization premium for a property you cannot use - see the quotation page for what to buy instead.
Does GPC dissolve faster than other carbon additives?
Generally yes, because the graphitized structure transfers carbon into the melt more readily than amorphous carbon. The practical gain appears as higher and more repeatable recovery rather than dramatic time saving. If your constraint is tap-to-tap time the benefit is modest; if it is hitting a carbon window, it matters a great deal.
What fixed carbon grade do I need for ductile iron?
Most ductile iron work runs well on 98.5 to 99.0 percent fixed carbon with sulfur at or below 0.05 percent. Moving to 99.2 percent fixed carbon and 0.02 percent sulfur is justified where the base melt already carries sulfur near specification, or where the casting is thick-section and inclusion-sensitive. Otherwise it is money spent on a certificate.
How do I calculate GPC cost per tonne of steel?
Multiply the carbon trim in weight percent by 10 to get kilograms of elemental carbon per tonne, divide by your fixed carbon fraction and your recovery, then multiply by the GPC price per kilogram. A 0.29 percent trim at 99 percent fixed carbon and 92 percent recovery is 3.18 kg per tonne, which at USD 950 per tonne is about USD 3.02 per tonne of steel.
Technical specification caps (fixed carbon ≥99.0%, sulfur ≤0.05%, nitrogen ≤0.03%, hydrogen ≤0.01%, oxygen ≤0.30%), the graphitization threshold above 2,800 °C, particle-size routing (0–1, 1–5, 3–8, 5–10 mm), application list and the 2026 grade price bands including 0–1 mm fines at USD 700–780: ferrosilicon.co, 28 and 30 December 2025. CPC Q2 2026 China assessment at USD 500/MT: Expert Market Research, 19 August 2026. FX and tariff as cited on the price page.
What is ours, not sourced: all three worked examples. The charge compositions, tap targets, recovery assumptions of 90–92 percent, the 0.60 percent nitrogen content of the alternative recarburiser, the USD 400/t alternative price and the USD 500 finished casting value are our inputs chosen to be representative, not measurements. Rerun them with your own numbers - the method is the deliverable, not the answer. No published source we could find gives GPC consumption in kg per tonne of steel or iron, which is precisely why these calculations are here.