Hot Metal
Three things worth your attention this week.
Thyssenkrupp's €3 billion Duisburg direct reduction plant will start up without hydrogen: the European Commission has approved rewriting the funding conditions, and the company's CFO called the original hydrogen requirement unrealistic for initial operations. In plain terms the shaft runs on natural gas first — and gas-based DRI arrives carburized. The day that plant moves to hydrogen, its carbon has to come from somewhere else. This week is about that carbon.
Injection carbon and charge coal specs deserve a side-by-side with the DRI certificate more often than they get one. Commercial injection carbons carry published ash in the high single digits, plus real moisture and real sulfur; carbon arriving inside the pellet carries none of that. Every kilogram of pellet carbon a shop actually burns is a kilogram of bagged carbon — ash and sulfur included — nobody had to buy.
Cleveland-Cliffs has rescoped the $500 million federal grant it won for a hydrogen-ready direct reduction plant at Middletown, Ohio, into a $1 billion blast furnace rebuild with cogeneration and AI process control, completion planned for the first quarter of 2030. Two large DRI decisions inside a week of each other say the same thing: the DRI route wins on heat cost or it doesn't get built. Carbon is part of that cost.
Prepaid fuel in the pellet
There's a habit of language in this industry worth breaking. DRI carbon gets discussed the way residuals do — a content to be "accepted," a deviation from pure iron units. Wrong frame. Carbon in DRI is fuel you've already paid to have delivered inside the raw material, in the cleanest form the shop will ever see, sitting next to the iron oxide it will help reduce. The question is never whether you want it. It's whether the furnace practice is set up to cash it.
Start with what it is. Gas-based DRI carries published carbon contents broadly between 1.5 and 4.5%, much of it combined as cementite — iron carbide — rather than free carbon dust. Cementite matters: it's bound into the metallic matrix, doesn't dust off in handling, and dissolves into the bath with the iron, releasing carbon in solution where oxygen can find it. Compare what you'd otherwise buy: injection carbons and charge coals with ash in the high single digits, sulfur you'll pay to remove later, and moisture that costs energy and — as we'll argue in three weeks' time — deserves respect for harder reasons than energy.
Now the energy math, kept honest with round numbers. One percent carbon in DRI is roughly 10 kg of carbon per tonne on a full-DRI charge. Oxidized in the bath to CO, published heats of reaction put the release around 2.5 kWh per kilogram of carbon at bath conditions — call it 25 kWh/t per percent of carbon, before post-combustion recovers anything from the CO on its way out. Move a DRI spec from 2% to 3.5% carbon and the chemical-energy swing rivals what most shops would spend serious capital to get electrically. And the same oxidation delivers the CO that feeds the foam — the financial instrument from three issues back — generated in the bath, continuously, at no injection cost.
So why isn't maximum carbon the answer? Three honest reasons, and they are the actual debate:
Carbon only pays against oxygen. The energy arrives when carbon oxidizes. Charge more carbon than the oxygen installation and blowing practice can burn inside the power-on window and the excess isn't fuel — it's refining time, a high tap carbon fighting your low-carbon grades, or FeO reduction nobody scheduled. Fuel without air is just inventory.
Carbon trades against other things in the shaft. Pushing carbon up can cost metallization or throughput, depending on the process and where it's running — and a point of metallization is a published rule-of-thumb 12-15 kWh/t back the other way, plus the yield argument from a few weeks ago. The pellet's whole package matters, not one line of the certificate.
Consistency beats content. A steady 2.5% you can plan oxygen around beats a 2-to-4% lottery every time. Variance in carbon is variance in energy input, foam behavior and tap carbon — the whole heat model wobbles with it.
Which leaves the operating rule we'd defend anywhere: order the carbon your oxygen can burn, then burn all of it. The shops that get this right treat DRI carbon as the first line of the heat's carbon balance and buy bagged carbon for the remainder. Not the other way around.
Operator's Notebook — the charge-carbon quick-calc card
Pin this at the panel. Five minutes per charge-mix change; it keeps the carbon balance honest.
Step 1 — carbon IN (kg per tonne of liquid steel):
DRI carbon: DRI kg/t charged × %C on the current certificate (not last month's) ÷ 100
Charged carbon (coal or coke in bucket or bins): kg/t × fixed carbon % ÷ 100
Injection carbon: the standard kg/t for this grade
Electrode contribution: small, roughly 1-2 kg/t — count it or note it
Scrap carbon: usually minor; include it for high-cast-iron mixes
Step 2 — carbon OUT:
Tap carbon target (aim spec, kg/t)
Slag FeO reduction duty (from your slag practice)
Balance available to burn to CO = IN − OUT
Step 3 — check against oxygen:
Oxygen needed ≈ carbon-to-burn × the stoichiometric ratio for CO (about 0.93 Nm³ per kg C), plus your standard for decarburization and lancing losses.
Required O₂ above installed blowing capacity in the power-on window → over-carboned. Cut bagged carbon first. Never the DRI.
Installed O₂ comfortably above requirement → there's headroom to discuss a higher-carbon DRI spec with supply.
Red flags: tap carbon consistently high on full-DRI heats (unburned fuel) · injection carbon at standard while certificate carbon rose (paying twice) · certificate %C changed but the heat model didn't (nobody read the certificate).
Next week: charge mix under price swings — deciding fast without breaking the furnace.
Written by active DRI-EAF operators. Anonymous by necessity, specific by design.