Steel production is one of the industries that emits the most greenhouse gases: 7.6% of global emissions. Hydrogen is one of the most promising ways to decarbonize this sector by removing the most polluting phase: the reduction of iron ore. Manufacturers are mobilizing significant resources, amounting to billions of euros, to develop these new processes as well as the production of carbon-free hydrogen to supply them.

This article is a part of our dossier on hydrogen, innovation and ecology and more precisely its practical uses.

Steel production represents one of the main sources of greenhouse gases in the world: 7.6% of global greenhouse gas emissions. It is also a booming industry: according to the world steel association (2020), global steel production increased from 1435 million tonnes in 2010 to 1875 in 2019. Decarbonizing the steel industry is one of the big challenges of the energy transition.

There is a way to largely decarbonize steel production: hydrogen.

Basically, steel and hydrogen have a rather conflicting relationship, and for good reason: hydrogen destroys steel! More precisely, it weakens the bond of the atoms and will come to associate with the carbon atoms. This causes the formation of fragile phases that endanger the structure of the object. This process involves several metals (e.g. copper, titanium) and even has a name: “hydrogen embrittlement” (or FPH). This is also one of the great challenges of transporting hydrogen: at the slightest microcrack, the gas could attack the metal.

Nevertheless, beyond this antagonism, H2 is a reducing gas (= which tends to capture oxygen, logic: H2 + 0 = H20) and it is precisely the reduction of iron ore that requires a lot of greenhouse gases. So let’s start at the beginning: how is steel produced?

How is steel produced?

The two main ways to produce steel are from iron ore or recycled steel. Contrary to what one might think, this last route represents more than a third of production in Europe. However, we are only going to focus on the first process (known as the “cast iron process”), in particular because the phase that interests us the most is the iron ore reduction phase.
The reduction of iron ore

Iron ore in its natural state is actually iron oxide (more precisely, it is called hematite). Yes, like rust! You must therefore first extract the oxygen from it: what is called “reduction” (this is the opposite of oxidation).

Example : Fe3O4 + 4CO → 3Fe + 4CO2 ; Fe2O3 + 3CO → 2Fe + 3CO2

It occurs beyond 900°C.

Melt the metal

Once it has been reduced, you still need to smelt the metal to be able to work it and remove its impurities. The melting point of iron is 1538°C.

These two operations are truly the core of steel production. Then, once you have your incandescent metal, you will work it in different ways to obtain the final product: plates or rolls of steel sheets, made up of various alloys (for example, aluminum or manganese) depending on the order. I will say little about these steps, which should not particularly change depending on the method of steel production (direct reduction or blast furnace), which we will talk about now.

Technologies of steel production

There are two processes for transforming iron ore into steel: blast furnace (Blast Furnace-Basic Oxygen Furnace, BF-BOF) and direct reduction (Direct Reduced Iron – Electric Arc Furnace, DRI-EAF). The first has largely dominated since the Middle Ages, but the second is tending to come back into fashion, being able to use only hydrogen.

Steel production in blast furnaces (BF-BOF)

Reduction and melting can take place in a blast furnace. It is a kind of huge hollow tower. In the top opening (“mouth”) are placed the coke and the ore and at the base is injected oxygen having already been heated to high temperatures (1200°C). Today we are talking about the BF-BOF (Blast Furnace-Basic Oxygen Furnace) process. The temperature varies inside the device: from 500°C at the top, we gradually reach 2100°C at the bottom. The reduction will occur in the upper part thanks to the gasified carbon of the coke and the fusion at the bottom.

Initially, charcoal was used as a reducer. Nevertheless, the increase in production made this track blocking: we lacked wood. Using coal (“mine coal”) was not a good alternative, because it contains many compounds, such as silicon or sulphur, which weaken the steel produced… The solution, found at the beginning of the 18th century, was to use a derivative of coal: coke. It is a mine coal whose volatile components have been vaporized beforehand. Coke serves as a fuel and a primary way at the same time. Indeed, its combustion (C + O2 → CO2) is very exothermic (401.67 kJ/mol). It is its carbon that will reduce iron oxide.

This process produces cast iron, whose carbon content is 2 to 6.67%. To produce steel, which has between 0.02 and 2% carbon, the metal will have to be reprocessed. It is passed through a “converter”, in which oxygen is injected, which will oxidize the carbon and leave in the form of gas (yes, it’s really not a “clean” process). This removes almost all the carbon and produces “wild steel”. In short, we can put it back, as well as other metals, afterwards (according to the composition requested by the client).

Note that the efficiency of the process has been multiplied by technical progress. Producing a ton of cast iron required 3 tons of charcoal in the Age of Enlightenment, 1 ton of coke in 1961 and 240kg of coke in the 21st century. Despite this gain, the scale of production encourages the development of an alternative carbon-free production process. (Wikipedia)

DRI (Direct Reduced Iron) and Electric Arc Furnace (EAF)

Before blast furnaces supplanted them, that is to say from the Iron Age (Mediterranean) to the Middle Ages (in Europe), “low furnaces” were used (rq: survived until at the beginning of the 20th century certain processes, such as the Catalan forge or the Japanese tatara, but I will not detail). At the beginning, they did not heat enough to bring the iron into fusion (1538°C), then it was a choice. Indeed: it is necessary to be able to reduce the carbon content of cast iron to work it and it took a long time to learn how to do it correctly. We speak of “direct reduction” for these processes of reduction of iron below the melting temperature of the latter.

Modern variants of this process have gradually appeared since the beginning of the 20th century with the development of the electric arc furnace. It began to gain new momentum in the 1950s and several processes (HYL, SL/RN, Purofer, Midrex process) proved viable on an industrial scale. This production currently represents around 5% of the steel produced.

This time it involves reducing the ore using natural gas, coal and/or hydrogen. Coal (mine) processes work by gasification: they release a syngas (mainly CO and H2). The Midrex process uses methane to do a kind of immediate steam reforming, producing two very reductive gases: CO and H2.

The result is a pre-reduced ore (DRI, Direct Reduced Iron), which is then melted in an electric arc furnace. These processes produce “sponges” (kinds of porous bricks (obviously: the oxygen has been removed)) or metal “briquettes” (compressed at 650°C).

Direct reduced iron (DRI), also called sponge iron, is produced from the direct reduction of iron ore (in the form of lumps, pellets, or fines) into iron by a reducing gas or elemental carbon produced from natural gas or coal.

Wikipedia, Minerai de fer préréduit

Then, we find the same technique as for melting recycled steel: the electric arc furnace (Electric Arc Furnace, EAF). Then, no need to reduce the carbon content, it’s already good! Then comes the rest of the steel production process (refining, adding other components, etc.).

Hydrogen to decarbonize steel

Hydrogen is a reducing gas: it combines with oxygen to form water (H2 + O = H2O). The idea is to have it replace coke in this role so that the reduction of iron ore only emits water (H2O) and not carbon dioxide (CO2). Of course, the hydrogen will have to have been produced by low-carbon processes, such as the electrolysis of water (powered by nuclear or renewable energies) or the pyrogasification / thermolysis of biomass.

There are also carbon capture projects (eg DMX in Dunkirk, CarbHflex in Fos-sur-Mer), but I won’t go over them here.

The use of hydrogen for the direct reduction of iron ore

The DRI already seems more aligned with the energy transition than the blast furnace: on the one hand, even if the gas used contains a lot of carbon, it also contains a lot of hydrogen and on the other hand electricity is used to melt the metal (rq: the source of electricity remains to be seen in detail). Nevertheless, using only hydrogen for the reduction would be a further progress.

However, this poses some difficulties. It takes more energy to heat the reaction for reduction (reduction with hydrogen is endothermic).

Hybrit: the project of steelmakers SSAB and LKAB

HYBRIT (for “Hydrogen Breakthrough Ironmaking Technology”) Development is a Scandinavian joint venture created in 2016 bringing together SSAB (steel), LKAB (mining company) and Vattenfall (energy). Its objective is to develop a process for producing carbon-free steel using hydrogen.

They reportedly succeeded in producing the “world’s first hydrogen reduced iron sponge at pilot scale” in June 2021. These Direct Reduced Iron (DRI) sponges are made of pure iron, unlike “pellet“, which still contain oxygen. A facility to demonstrate the viability of this process would be planned at Gällivare.

The titanic investments of ArcelorMittal

ArcellorMittal announced on October 13, 2020 its ambition to reduce its CO2 emissions by 30% by 2030 and carbon neutrality by 2050 using the DRI-EAF pathway and carbon capture (they speak of the “Smart Carbon pathway”). The steelmaker has estimated the investments needed to achieve the first objective at 10 billion euros, and between 15 and 40 billion by 2050.

In France alone, the steelmaker announced in February 2022 the replacement of 3 blast furnaces on the sites of Dunkirk (Nord) and Fos-sur-Mer (Bouches-du-Rhône) by installations allowing the path of direct reduction (DRI-EAF) from hydrogen from 2027. These sites alone would represent “25% of industrial greenhouse gas emissions in France. This will represent an investment of 1.7 billion euros. (Le Monde)

The company estimated, for the transformation of its plant in Bremen and Eisenhüttenstadt, that the production of steel by this process would cost 60% more and that the conversion of the installation would represent between 1 and 1.5 billion euros. (Eurometal)

H2FUTURE: the consortium led by Siemens and Voestalpine

Voestalpine is an Austrian steel company that plans to use hydrogen to reduce ore by 2050. With this in mind, it has launched a 6 MWh PEM electrolyser project with Siemens (industry) and Verbund (energy). in Linz, financed to the tune of 18 million euros, including 12 from the European Union. The electricity needed would come from intermittent renewable energies. The initiative was validated in April 2018.

Inject syngas (CO+H2) into the blast furnace

These solutions would require a complete change in steel production. However, these gigantic installations are extremely expensive, many running into billions of euros (!). This dimension alone runs the risk of radically limiting their development.

An intermediate method consists in injecting hydrogen into the blast furnace, probably in the form of synthesis gas (syngas), composed essentially of hydrogen and carbon monoxide. This is all the more practical since the transformation of coal into coke releases a syngas (“coke oven gas” or “coal gas”) rich in hydrogen (62.7% H2 on average). It seems to me that this gas must be treated beforehand to remove the sulfur (in particular?). There are ThyssenKrupp and ArcelorMittal which have embarked on this path.


On November 11, 2019, ThyssenKrupp began testing the injection of hydrogen (delivered by Air Liquide) into a blast furnace in Duisburg, Germany, which is expected to decrease CO2 emissions from the corresponding steel production by 20 %. . The test, costing 2.7 million euros (including 40% federal subsidies) was launched in November 2019 and was to last 14 months. The company nevertheless plans to ultimately take the same paths as ArcelorMittal: a an way by DRI-EAF and a way by carbon capture.

ArcelorMittal: the Igar project

ArcelorMittal is also exploring this axis in Dunkirk, with the Igar project:

At ArcelorMittal Dunkerque, the group is developing a hybrid blast furnace process, which involves the use of DRI gas injection technology in the blast furnace vessel as well as gas injection into the blast furnace nozzles. furnace, using plasma technology to create a reducing gas. This is the first large-scale implementation of what is in essence a hybrid blast-furnace/DRI technology. Eventually, it will make it possible to inject green hydrogen into the blast furnace as soon as it becomes available.

ArcelorMittal, 13 février 2020

(I’m not sure that this project is still relevant (see the complete transformation of the blast furnaces announced above, in 2022). To be checked.)

The SALCOS and WindH2 projects by Salzgitter ag

Salzgitter plans to replace its last three blast furnaces with direct reduction units and electric arc furnaces. This is its “SALCOS” (SAlzgitter Low CO2 Steelmaking) strategy. In line with the logic of Salcos, the WindH2 project consists, with Avacon (energy) and Linde (hydrogen specialist) of producing hydrogen which would be intended to fuel the transition of the steelmaker. The latter put 50 million euros in the project. More information here.

Out of category: electrolysis of iron by Boston Metal

This is a somewhat special track, since it is neither in the blast furnace route (BF-BOF), nor in the direct reduction route (DRI-EAF), but in a specific process: electrolysis of iron ore (Melted oxide electrolysis, MOE). It does not even use hydrogen and therefore does not in principle fall within our field of study, however this process is too close to the subject (it is a question of electrifying the production of steel) and too interesting for not be approached.

The reaction would be as follows: Fe2O3 + e- => Fe + O2. For this to work, the metal must be molten, so the operation must be done at more than 1538°C. (Wiencke et al. 2018)

This process is operated by Boston Metal, a company in which Bill Gates and BMW have notably invested. It’s still a startup: it raised $20 million in 2019 and another 60

The facilities would also be much more modular, increasing the necessary investments from the order of a billion for traditional steelmaking to the scale of a million dollars. This is an extremely interesting aspect for creating ecosystems bringing together electricity production, steel production and transformation of the latter. The process would even be 35% cheaper than conventional steel production, requiring less materials and infrastructure.

The Siderwin project, which notably involves ArcelorMittal, is also trying to exploit the electrolysis of ore, but it seems less advanced.

Which decarbonization model will prevail? Only the future will tell…

  • World steel association (2020), Steel Statistical Yearbook 2020 concise version, A cross-section of steel industry statistics 2010 – 2019
  • Wiencke J., Lavelaine H., Panteix P-J., Petitjean C. et Rapin C., Electrolysis of iron in a molten oxide electrolyte, Journal of Applied Electrochemistry, 2018,
  • Les articles Wikipedia sur ces sujets sont très complets : ; ;
  • C’est pas sorcier « Le dire c’est bien, le fer c’est mieux » (absolument fantastique pour comprendre rapidement la production du fer)
  • Sénat (2019), Sur les enjeux de la filière sidérurgique dans la France du XXIe siècle : opportunité de croissance et de développement, Rapport d’information n° 649 (2018-2019) de Mme Valérie LÉTARD, rapporteure, fait au nom de la MI enjeux de la filière sidérurgique, déposé le 9 juillet 2019

Pour aller plus loin :