Hot Metal
Three things worth your attention this week.
Turkish import scrap topped $400 a tonne in June, gave back $40 by early July, and was most of the way back inside three weeks. Yield targets have a habit of outliving the price environment they were set in: a point of yield at $300/t metallics is one number; at $400 it's a third more. If your target predates the last big move, it isn't a target — it's a habit.
Slag processors keep marketing iron recovery — magnetic separation lines pulling metallics back out of EAF slag yards, sold as a circular-economy win. Read those brochures as an audit of the meltshop upstream: they are mining what the furnace gave away, at a second processing cost, minus a margin. Every tonne they recover at a profit is a tonne it would have been cheaper never to lose.
Steel Dynamics opened Q2 earnings season on July 20 with record steel shipments of 3.7 million tons and steel operating income up 30 percent on metal spread expansion — selling price up $105 a ton against scrap up $16, to $412 per ton melted. At $412 metallics, a point of metallic yield is worth about $4 a tonne. The spread between scrap and steel is only yours if the iron makes it to the caster.
The invoice you never see
Every shop we've worked in tracks kWh/t to the decimal and lets metallic yield float as "about 90%." That's backwards. Per dollar you can actually move, yield is the biggest line on the liquid-steel cost sheet — a few points of spread are worth as much as your electrode and refractory budgets put together — and the mechanism that drains it has a name: FeO in the slag.
The arithmetic first. Lose one point of metallic yield and you've discarded roughly 10 kg of iron units per tonne of liquid steel. With metallics in the $300-400/t band — and this summer's market living at the top of it — that point is worth $3-4 per tonne: $3-4M a year for a 1 Mt/y operation. Published benchmarks put DRI-heavy charges at roughly 87-92% metallic yield against 90-94% for clean scrap practice. The spread inside those bands, not the bands themselves, is where the money sits.
Where does the iron actually go? Mostly out the door as iron oxide. The loss is simple to write down: slag mass times FeO content times 56/72. Both factors punish a DRI shop, which is why scrap-shop yield benchmarks mislead here:
The FeO term. DRI arrives with unreduced iron oxide — every point below full metallization is oxide you must reduce in the furnace or lose to the slag. Scrap arrives metallic.
The mass term. DRI gangue drives slag volumes to a published 150-250 kg per tonne of steel, against 60-100 kg/t for a clean scrap charge. The same 25% FeO in double the slag is double the iron in the pot.
That's the FeO tax paid twice, and it's why a DRI shop that manages FeO like a scrap shop is leaving the larger share of the prize untouched.
The control lever is the carbon-oxygen balance. FeO in slag is the visible end of it: blow more oxygen than your carbon can answer, and the surplus finds iron — the equilibrium chemistry has been in Turkdogan for decades. Operationally that means matching the oxygen profile to what the charge actually carries (DRI carbon included — an argument we'll make in full in a few weeks), and using injection carbon late in the heat to claw FeO back down before tap. Shops that reconcile oxygen blown against carbon available, heat by heat, find the surplus fast.
Now the honest counter-case: do not chase minimum FeO. You need FeO in a working band — published foaming practice sits broadly in the 15-25% range — because foam needs oxygen potential, and phosphorus removal needs an oxidizing slag. Drive FeO too hard toward zero and you'll pay in injection carbon, flat foam, and a phosphorus problem that costs more than the iron you saved. The target is an optimum, not a minimum: no free oxygen you didn't need, and no slag mass you didn't need. Yield discipline is mostly the discipline of those two sentences.
One more thing worth stealing from the accountants: reconcile it. Weigh what goes in, weigh what's cast, close the loop weekly. In shops we've seen do this, the first reconciliation alone usually surfaces a yield leak nobody owned — because until that meeting, nobody was invoiced for it.
Operator's Notebook — the 8-point slag & FeO shift check
Run this once per shift. It takes ten minutes and it's the cheapest yield program there is.
Slag FeO — is there a slag analysis from this shift, and is FeO inside your working band (not just "slag looked fine")?
Slag mass sanity — flux additions plus expected gangue: does estimated slag volume match what left in the pot? A creeping pot count is a creeping tax.
Oxygen vs carbon — total O2 blown against total carbon available (charge + DRI + injected). Surplus oxygen made FeO somewhere.
Injection carbon response — did FeO trend down after late-heat injection, or are you buying carbon that isn't working?
Door losses — any slag-over-the-sill events this shift? Each one carries iron with it.
Metallization check — latest DRI metallization figure posted at the panel, not last month's average.
Slag yard feedback — magnetic content reports from slag processing, if you get them. Rising metallics in slag is the loudest alarm you'll ever ignore.
Handover line — one sentence to the next shift: FeO where it is, why, and what's in flight.
Next week: where the kWh actually go — the honest energy balance, from theoretical minimum to your meter.
Written by active DRI-EAF operators. Anonymous by necessity, specific by design.