Inside the race to re‑tune the blast furnace: PCI, natural gas, and CCS

Blast‑furnace ironmaking (BF) remains the industry’s workhorse, with about 70% of global steel produced via the BF/basic oxygen furnace (BOF) route (ResearchGate) (ResearchGate). That dominance comes with a carbon bill: ≈2.0–2.2 tonnes of CO₂ per tonne of hot metal (tHM, the liquid iron tapped from the furnace) (ResearchGate), feeding ~2.6 Gt CO₂ per year — around 7% of global emissions (ResearchGate).

Roughly 70% of steel’s CO₂ arises inside the BF itself (Carbon Capture Journal), so even modest efficiency gains matter. Typical BF energy use sits near ~13–14 GJ/tHM (ResearchGate). Operators are squeezing the core thermodynamics with hot‑blast stove optimization, top‑gas recovery turbines, oxygen enrichment, and fuel replacement through tuyère injection (the nozzle region where the hot blast enters).

Blast‑furnace energy and emissions baseline

The traditional BF/BOF route is emissions‑intensive relative to electric‑arc steelmaking, making fuel mix and heat recovery critical levers (ResearchGate). Two injection strategies dominate today’s optimization playbook: pulverized coal injection (PCI) and natural‑gas (NG) injection. A third — carbon capture and storage (CCS) — looms as the deep‑cut option to address the BF’s concentrated CO₂ streams.

Pulverized coal injection (PCI) performance

PCI pneumatically feeds fine coal directly through the tuyères to supplement metallurgical coke as both fuel and chemical reductant. Adoption is widespread: estimates suggest ≥65% of blast furnaces use injection, and ~75% of injection BFs use coal (ResearchGate) (ISIJ International).

The displacement effect is the big prize. Industry data put PCI’s coke replacement at ~0.85–0.95 tonnes of coke per tonne of injected coal (IIP Network). One data point: 1 tonne of PCI coal has been noted to displace about 1.4 tonnes of coking coal (indicative of ~30% coal savings) (Docslib).

Worrell et al. estimate ~3.76 GJ of energy saved per tonne of pulverized coal injected (IIP Network). Globally, lifting average PCI to ~180 kg/tHM could abate on the order of 10 Mt CO₂ per year (IIP Network) (IIP Network). In practice, many furnaces now inject 100–180 kg/tHM and trim coke by ~15–30% — for example, a 400 kg coke/tHM furnace dropping toward ~280–340 kg/tHM at heavy PCI (IIP Network). Each tonne of coke avoided (≈0.85–1.0 t coke per tPC saved) cuts ~2.0–2.5 tCO₂ (≈2.0 t/tHM typical BF emission intensity) (IIP Network) (ResearchGate). That implies ~1.7–2.4 tCO₂ reduced per tonne of PC used.

On the ground, PCI rates above 150 kg/tHM are now common in China, India, the U.S., and Europe. Theoretical limits — governed by combustion kinetics and ash behavior — run ~200–270 kg/tHM (ResearchGate). In China, co‑injection trials (natural gas plus PCI) at Chongqing Iron & Steel improved hot‑metal quality, lowered the coke rate, and lifted productivity (ISIJ International). In the U.S. and Europe, high PCI rates are also reported; POSCO cited average PCI exceeding 150 kg/tHM at its modern Kaldo furnaces.

Coal choice matters. Volatile content and grindability set raceway combustion efficiency; poor combustion risks unburned char and altered coke behavior in the hearth. The economics usually pencil: U.S. analyses put savings at $16–33 per tonne of hot metal from PCI, roughly 4–5% of HM production cost, versus capital at about US$50–55 per annual tonne of injected coal capacity for mills and injection systems (IIP Network). During pulverized coal preparation and conveyance, plants also plan for materials‑handling controls; some operators evaluate coal‑dust control options alongside PCI system design, including market offerings such as a coal‑dust suppressant.

Bottom line: under heavy PCI, typical BF coke usage can drop ~20–30%. A 2010 EPA guide summarized that ~25–30% of BF coal demand can be saved via PCI (≈0.9 t coke avoided per tPC) (IIP Network). Every kilogram of coke saved avoids about ~3.0 kg CO₂.

Natural gas injection and oxygen enrichment

Natural gas (CH₄) or other hydrogen‑rich gases (e.g., coke oven gas, syngas) can also be injected at the tuyère. In the raceway, CH₄ reforms endothermically to CO+2H₂, supplying ash‑free reductants while altering flame temperature and gas composition. NG injection rates in practice are often 20–100 m³/tHM in trials, with deployment rising where gas is inexpensive (ISIJ International).

Modeling by Li et al. (2025) shows the effect scale: adding 100 m³/tHM of NG with 6% oxygen enrichment and decreasing PCI cut the coke rate by ~21 kg/tHM and reduced CO₂ by ~60 kg/tHM (ISIJ International). In that scenario, 100 m³/t displaced about 62 kg/t of PCI and reduced coke by 21 kg/t, thus indirectly saving ~2.2 tCO₂ per tHM from the lower coke and PCI levels (ISIJ International). Other studies cited find roughly 4–6 kg of coke saved per extra 5–10 m³ of NG injected (≈0.8–1.2 kg coke per m³ NG) (ISIJ International).

Because NG cools the raceway (endothermic reforming), operators boost blast oxygen or preheat the NG. Co‑injection with PCI widens the operating window. Oxygen enrichment alone (~2% more O₂) can cut the coke rate by ~3% (≈25 kg/tO₂ per 100 kg coke) (ISIJ International), enabling higher NG shares. Case experience (e.g., Chongqing Iron & Steel) reported improved hot‑metal quality and lower coke rates with modest NG injection (ISIJ International).

The carbon math underscores the trade‑offs. On a per‑carbon basis, 1 kg of CH₄ roughly replaces 1 kg of carbon from coal and, if fully combusted, yields about 2.75× less CO₂. Thus, 100 m³ (≈80–90 kg) of NG replaces ~80–90 kg of C, saving ∼220–250 kg CO₂ versus coal. In the cited modeling, the co‑injection/oxygen‑enriched setup delivered ~60 kg CO₂/tHM saved for 100 m³/t (ISIJ International), reflecting the partial substitution and reduced PCI. In practice, high‑NG systems (≥50–100 m³/tHM) typically lower coke use by ~5–10% when coupled with oxygen enrichment.

Carbon capture and storage (CCS) feasibility

Even with PCI and NG, BFs emit ≈2+ tCO₂/tHM. Post‑combustion CCS is the main pathway to deep cuts, since capturing BF CO₂ targets roughly 70% of a steel plant’s emissions (Carbon Capture Journal). At an integrated site (coke ovens, sinter, BF), those BF‑related processes can run to about 80% of site CO₂. BF top gas (typically 20–25% CO₂) and hot‑blast stove flue gases are attractive capture targets; chemically scrubbing the BF gas at ~90% capture would cut integrated plant CO₂ by ~63% (0.9×0.7) (Carbon Capture Journal).

The technology is nascent in steel. As of 2024, only two carbon‑capture projects operate in the sector (≈1% of capacity), and just one BF‑BOF CCS scheme is planned, versus dozens of hydrogen‑based projects (Transition Asia). Costs are a hurdle: retrofitting a BF can run ~$50–100+ per tonne CO₂ captured (IEAGHG studies vary). Yet pilots show promise. An ArcelorMittal/Sekisui trial processing real BF gas in Spain achieved ~90% CO₂ purity in capture and 90% conversion efficiency, indicating technical feasibility (Carbon Capture Journal).

Energy and site integration matter. Full BF CCS can theoretically slash plant CO₂ by ~60–80%. Practically, capture would target BF runner gas and, potentially, converter gas stacks; with appropriate stove configurations (e.g., oxygen‑fired stoves), >50% site reduction is plausible. Capturing 90% of the BF’s ~70% emission share alone yields ~63% site reduction (Carbon Capture Journal). Residual emissions from coke ovens and reheating remain, making CCS essential for deep decarbonization. Chemical scrubbing units involve solvent management; operators often provision accurate chemical metering and corrosion control in these loops, where utilities such as a dosing pump and a corrosion inhibitor may be specified in the balance of plant.

Policy is catching up. CCS imposes significant low‑grade heat demand (often met by steam extraction) and high capex, so deployment to date is limited to models and small pilots. Few jurisdictions mandate CCS in steel. Indonesia is one to watch: Presidential Reg. 14/2024 and MEMR Reg. 16/2024 (effective December 2024) establish permits for carbon storage areas (WIPKs), laying legal groundwork for future CCUS across sectors, potentially including steel (Ashurst). For capture trains that employ chemical scrubbing, some projects also plan upstream gas‑cleanup or solvent handling consumables; vendors market CO₂/H₂S removal solvents such as amine systems as part of generic gas‑treatment toolkits.

Economics, policy signals, and next‑decade outlook

Today’s gains are incremental but material. PCI remains a “low‑hanging fruit”: ~20–30% coke cuts under heavy injection; ~3.76 GJ/t PC injected in energy savings; and potential global abatement on the order of 10 Mt CO₂/yr if average PCI reaches ~180 kg/tHM (IIP Network). NG co‑injection (often 20–100 m³/tHM) brings 5–10% coke savings at high injection levels when tuned with oxygen enrichment, with modeling examples showing ~60 kg CO₂/tHM reductions for 100 m³/t plus 6% O₂ (ISIJ International).

CCS, meanwhile, is the lever for deep cuts — potentially >50% site reductions with BF gas capture — but sits early on the cost and policy curve (Transition Asia) (Carbon Capture Journal). As regulatory frameworks such as Indonesia’s new carbon storage rules mature (Ashurst), capture projects will have clearer permitting paths. Until then, blast‑stove efficiency, top‑gas recovery, oxygen enrichment, and tuned injection mixes will carry most plants’ energy‑efficiency and emissions‑intensity goals. For operators upgrading utility blocks around these projects, standard industrial housekeeping applies, from conveyance dust control to routine water and chemical handling across the site’s auxiliaries.

Sources: Efficiency and injection figures: (IIP Network) (ISIJ International) (ResearchGate) (ResearchGate). Emission shares: (ResearchGate) (Carbon Capture Journal). Economics and regulation: (IIP Network) (Ashurst).