Last week’s graph (see graph#1) started to explore the relationship between crude steel production, scrap generation and the residual demand for iron extracted directly from iron ore.
This week’s graph – graph#2 – shows how greenhouse gas emissions for steelmaking vary with scrap use at individual sites. Each dot on the graph is a separate steel mill[1] – more than 200 in all. Shout outs to ResponsibleSteel and CRU for working this out. The blue dots are blast furnace – basic oxygen furnaces (BF-BOF), the orange dots are electric arc furnaces (EAF).
The horizontal axis shows the percentage of scrap being used to provide the metallic input for steel production at a given site, from 0% (so 100% iron ore) on the left, to 100% scrap on the right.
The vertical axis shows the greenhouse gas emissions per tonne of crude steel produced, including Scope 1, Scope 2 and upstream Scope 3 emissions.
What does the graph tell us?
First, greenhouse gas emissions are a lot lower towards the right of the graph than the left. No surprise there – it takes less energy to remelt scrap than it does to produce steel from iron ore in the first place.
Second, scrap usage rates vary widely between sites. As much as 30% of the metallic input for BF-BOF steelmaking can be from scrap (the average is about 12%), and EAF sites can use anything from 0% to 100% scrap with the balance made up with iron from iron ore – either pig iron from a blast furnace, or direct reduction iron (DRI). That might be surprising, but EAF is not synonymous with recycling, and BF-BOF steelmaking uses a huge quantity of scrap.
Third, there is a close relationship between the proportion of scrap used as an input, and the greenhouse gas emissions per tonne of crude steel produced, as shown by the blue dotted line. The more iron ore you use, the more energy you need. If that energy is derived from fossil fuels – either directly, or indirectly to generate electricity to power the EAF – the higher your GHG emissions. That relationship is the same whether you produce crude steel from pig iron and scrap in an EAF, or from pig iron and scrap in a basic oxygen furnace.
Which brings us to a final observation. On the left side of the graph there are a number of EAF sites – orange dots – which have much lower GHG emission intensities than the BF-BOF sites using the same proportion of scrap for input. Those low emission intesity sites are EAF sites using DRI made using natural gas, rather than blast furnace pig iron, as the source for their primary metallic input.
 Figure 2a: two ways to reduce GHG emissions for steelmaking |  Figure 2b: the future status of steelmaking |
What does the graph tell us?
Well, to really understand it we have to know how much scrap is available worldwide, and how that will change over time – Graph#1 again – and we need to go back to that graph and kick its tyres.
But even without that detail, it tells us there are two things we can do if we want to reduce GHG emissions and still produce the same amount of steel: either we have to find more scrap; or we have to reduce the GHG emissions when we make a given amount of steel from a given amount of iron ore and scrap. Those two options are shown in Figure 2a, above.
Using more scrap is the obvious choice, and Graph #1 showed that the supply of scrap will inevitably increase over time. That is a good thing – as the proportion of scrap increases the overall GHG intensity for steel production will come down, without any need for new technology, public subsidy or policy intervention.
But the supply of scrap is constrained by historical production and by the potential to recover scrap after the end of its life in use. As long as all the available scrap is used, we cannot increase the supply of scrap further or faster. The exact numbers are debated, but we can be sure that we are going to continue to produce a lot of steel from iron ore through to 2050 and beyond.
So yes, we need to maximise scrap use – but that is nowhere near sufficient. We must also reduce the GHG emissions for the production of steel from any given quantity of scrap and primary metallic input. And we need to get those emissions down to ‘near zero’.
Figure 2b shows the scale of the challenge. By 2050 or thenabouts the world has to be making steel very differently. How fast we make the change is the difference between a 1.5 degree world, and a 2.0 degree world or worse.
That means minimising Scope 2 emissions – we cannot go on using coal to generate electricity to melt scrap. It means shutting down blast furnaces that cannot capture and store their carbon emissions. And it means producing iron using hydrogen-based DRI, direct electrolysis, biocarbon, and other near zero emission technologies.
How to achieve this? What is the role of major steel users like the construction sector, and automotive sector? What are the roles for government intervention, public procurement, and trade policy? What are the implications for ‘carbon footprints’ and other performance measures?
Keep these two graphs, Graph#1 and Graph#2 to hand – and watch this space.
Footnotes
[1] Data are from the CRU Steel Cost Model and Emissions Analyis Tool, modified for alignment with the ResponsibleSteel international production standard.