If you read the literature of the European Recycling Industries Confederation (EuRIC) you might get the impression that recycling steel is only possible using electric arc furnaces¹: ‘… two […] steelmaking routes, the Blast Oxygen Furnace (BOF) [sic], which relies on iron ore, and the Electric Arc Furnace (EAF), which uses recycled scrap steel’.²

This is, of course, nonsense. What are the facts?

In reality, both the Blast Furnace – Basic Oxygen Furnace (BF-BOF) and EAF steelmaking pathways use both primary (iron ore) and secondary (scrap) sources of iron, as shown in the main illustration for this post, from the Institution of Structural Engineers (IStructE) paper, ‘The Role of Scrap in Steel Decarbonisation’³.

In the electric arc furnace electricity is fired through metal, heating it until it melts.  That electricity may be generated from fossil fuels, renewable sources or nuclear power.  The input materials for an EAF can in principle be anything between 100% scrap and 100% primary iron.  The actual proportion will depend on factors including the relative costs of energy, the cost, quality and availability of scrap and primary inputs, and on the quality specifications of the steel being produced. 

The blast furnace in contrast produces (primary) pig iron from iron ore, typically using coal as its energy source.  The carbon in the coal is burnt to generate heat and also reacts with the iron ore, splitting it into metallic iron and carbon dioxide.  This chemical separation of iron and from oxygen requires more energy than simply re-melting scrap.  Oxygen is injected in the oxygen furnace to burn off excess carbon, generating more heat.  Scrap metal is added to control the process – a bit like adding ice cubes to reduce the temperature of boiling water – albeit at 1500^o centigrade.

So both steelmaking pathways use both primary steel and scrap as inputs. But how does scrap consumption in the two pathways compare quantitatively?

The global production of crude steel in 2024 was 1,890 million tonnes (Mt). If we assume an optimistic yield loss of just 2.5% for conversion, we would need 1,928 Mt of iron from iron ore and scrap to produce that quantity of crude steel. worldsteel and the Bureau of International Recycling (BIR) estimate that in 2024 steelmakers used 630 Mt of ferrous scrap in total, and the balance of ferrous input must be made up from primary sources.

At least 6% of the metallic input to the blast furnace is likely to be scrap, but proportions of 15% to 20% are common, and higher percentages are possible⁴. The global average scrap content of steel produced through the BF-BOF route is estimated to be around 12.5%, and around two thirds of steel (71%) is produced via the BF-BOF route, slightly less than one third (29%) by EAF.

Based on these data⁵, in 2024 BF-BOF steelmaking would have consumed around 172 Mt of scrap (Figure 1b), compared to 458 Mt consumed for EAF production (Figure 1c). So, 27% of the steel sector’s total scrap use was for BF-BOF steelmaking, compared to 73% used for EAF production, as shown in Figure 2c.

On the other side of the equation 1,203 Mt (93% of the total) of primary iron would have been used for BF-BOF steelmaking, and 95 Mt (7% of the total) for the production of steel by the EAF route. The average scrap content of EAF produced steel was 82.5%, and the average primary metallic content was 17%.

So let’s be honest.  Most scrap consumption (73%) goes for EAF steelmaking – but a highly significant minority (27%) – is used for so-called ‘primary’ production. EAF production is mostly (83%) scrap-based – but it also uses a significant proportion (17%) of primary metallic material. Much of that will be direct reduced iron (DRI), produced using natural gas rather than coal, but some will be pig iron, together with other primary metals.

The coal-based production of iron and steel needs to be phased out as soon as possible, as does the coal- and gas-based production of electricity.  Those are challenges that all steelmakers – whether their inputs are mostly of scrap or mostly of primary metallic iron need to address. As ever, it is not either/or, it is both/and. Effective policies need to be designed accordingly.

Maximising scrap use is essential – but the recycling industry and EAF steelmakers are well aware that the utilisation of pre- and post-consumer steel scrap is already at close to maximum levels, and that the main effect of policy instruments that increase demand for scrap steel will be to increase profitability, rather than to increase the overall supply or mitigate climate change.

Investment and focussed support for recycling is needed, for example to improve the separation of steel from other metals in the materials recovery supply chain.  Users of copper and other ‘tramp’ elements should recognise and reward the end-of-life recovery of those metals, which are far more valuable, per tonne, than scrap steel, and which are otherwise lost from productive use. Carmakers should design cars for recyclability. Blunt policies to favour high recycled content steel do not achieve any of this.

The recycling industry ought to be more honest in its lobbying. 

---

1. A previous version of this article referenced a quote from a 2022 paper by the Brussels-based NGO Sandbag which claimed that “recycling steel is only possible using electric arc furnaces” (Sandbag (2022) European Scrap Steel Floats Away under Carbon Market Incentives). The paper has now been corrected, and this post amended accordingly. ↩︎
2. EuRIC (2025) Why the Sliding Scale is a Slippery Slope for Defining Green Steel ↩︎
3. Arnold et al (2025) The Role of Scrap in Steel Decarbonisation ↩︎
4. worldsteel (2021) Scrap use in the steel industry ↩︎
5. I have used conservative assumptions to estimate the use of scrap for BF-BOF production.  If the yield loss of metallic inputs in conversion to crude steel is 5% (rather than 2.5%), and if the average global consumption of scrap in BF-BOF production is 13.5% (rather than 12.5%), then: 30% of total scrap use for steelmaking would be via the BF-BOF pathway (rather than 27%), and 70% by EAFs (rather than 73%);  90% of primary iron would be used by BF-BOFs, and 10% by EAFs; and the average scrap content of EAF-produced steel would be 77% rather than 82.5%. ↩︎
6. Figures are based on BIR and worldsteel data for 2024, assuming 12.5% scrap input for the BF-BOF route, and 2.5% metallic yield loss.  Area is proportional to mass of metallic input. In both figures 1 and 2 the total metallic input is greater than crude steel production, due to the conversion loss. ↩︎