Email a friend
  • To include more than one recipient, please separate each email address with a semi-colon ';', to a maximum of 5

  • By submitting this article to a friend we reserve the right to contact them regarding Fastmarkets AMM subscriptions. Please ensure you have their consent before giving us their details.

Advancing DRI technology to make steel with hydrogen

Jan 06, 2020 | 08:00 PM |

ArcelorMittal has commissioned Midrex to design a demonstration plant to produce steel with hydrogen at the steelmaker’s Hamburg site in Germany. Richard Barrett asks Midrex expert Dr. Vincent Chevrier to explain its advantages.

In mid-September 2019, ArcelorMittal announced that it had commissioned Midrex to design a demonstration plant to produce steel with hydrogen at the steelmaker’s Hamburg site in Germany. Both companies signed a Framework Collaboration Agreement to co-operate on several projects – ranging from R&D to the implementation of new technologies. They will be governed by a number of project development agreements.

The first of these projects is to demonstrate in Hamburg the large-scale production and use of DRI (direct reduced iron) made with 100% hydrogen as the reductant. It is due to produce about 100,000 tonnes per year of DRI. Initially the hydrogen used will be made from natural gas, but it is envisaged that some – or potentially all – of the hydrogen needed will be generated by equipment powered by renewable energy sources, once it is available in sufficient quantities and at a commercially viable cost.

Carl de Maré, ArcelorMittal vice- president responsible for technology strategy, summarized that the steelmaker is working with Midrex “to learn how you can produce virgin iron for steelmaking at a large scale by only using hydrogen.”
He said that the project, combined with the steelmaker’s existing projects on the use of non-fossil carbon and on carbon capture and use, is key for the company to become carbon neutral in Europe by 2050. “Large-scale demonstrations are critical to show our ambition,” he added, while highlighting that how fast transformation will take place will depend on the political conditions for the region.

ArcelorMittal’s Hamburg works already produces steel from DRI made on the site by its existing MIDREX® Plant, which processes about 980,000 tonnes per year of iron ore pellets into DRI by reducing them to 95% metallic iron by extracting oxygen by using natural gas.

With 25 years in the iron and steel industry and having worked at Midrex for nearly ten years (in R&D and business development), Dr. Vincent Chevrier is well placed to outline the company’s already extensive experience of using hydrogen as a reductant. It is already a component of the reducing gas mixture used in its conventional DRI plant designs. He has also spent several years working on Midrex’s technology for using pure hydrogen, MIDREX H2™.

Building on experience

In a conventional MIDREX Plant, the reducing gas needed to feed the shaft furnace in which iron ore is converted to DRI is generated in a reformer. The reformer itself is usually fed with natural gas.

The reducing gas produced in the reformer typically comprises 55% hydrogen and 36% carbon monoxide. Those reducing-gas constituents are generated by the chemical reaction of the methane (CH4) in natural gas with carbon dioxide (CO2) or water (H2O), both of which are themselves generated by the reduction process in the shaft furnace and are fed in order to be consumed by the reformer (see flowsheet).

When paired with an EAF, the existing long-established Midrex technology already results in lower CO2 emissions than the BF/BOF route to steelmaking – at 1.1-1.2 kg CO2 per kg of steel, compared with 1.6-2.0 kg CO2 per kg of steel, respectively, Midrex notes. Midrex adds that its process can lower those emissions even further – to around one-third of typical BF/BOF emissions if a CO2 removal system is added.

But it is the potential of using just hydrogen as a fuel and chemical reductant in the shaft furnace (without the carbon monoxide that generates the carbon dioxide when it is also used to reduce iron ore, as now) that holds the prospect of generating much less carbon dioxide in the steelmaking industry.

Three-step transition

Midrex envisages a three-step transition for steelmakers wishing to move gradually towards hydrogen-based steelmaking, with built-in flexibility to proceed through the steps as greater volumes of affordable hydrogen generated by renewable energy sources become available.

Step 1 requires the building of an existing conventional MIDREX NG Plant or being aware that an existing one in itself has the potential to use more hydrogen. In step 2, up to 20% - 30% of natural gas– depending on the desired operating conditions – can be displaced by hydrogen as it becomes available without any modification to the existing equipment. Above 20-30% of NG displacement, some minor equipment needs to be added to compensate for the reduction in chemical reforming. Chevrier said that Midrex’s plant is flexible enough to allow changes in energy source over time to accommodate likely fluctuations in hydrogen availability.

The ultimate step requires modification of the traditional Midrex plant fed with natural gas to the company’s MIDREX H₂ flowsheet (see diagram). This step requires modifications of some of the process equipment as the process duties, such as flow rate and temperature, will change. Most notably, the all-hydrogen concept requires a reduction gas heater to rebalance the energy in the shaft furnace. For an existing plant, the Midrex reformer can easily be converted into the heater by changing its operating conditions, not the reformer itself.

“This approach offers both the ability to buy time and minimal technology risks,” noted Chevrier. Midrex is basing its calculations on the assumption of using a standard iron ore pellet feed for DRI production, so the company does not anticipate a need to change that for a plant running with just hydrogen.

Industrial-scale development While a number of existing DRI plants have already run with a high percentage of hydrogen, but not 100% hydrogen, reducing gas for years, and Midrex has conducted many relatively small-scale R&D trials with very high percentages of the gas (up to 100%), the new project at Hamburg will provide the opportunity for a large-scale industrial demonstration.

Chevrier explained that the new Hamburg plant will have a freestanding shaft furnace and no reformer of its own. It will use an offtake of spent gas from the existing DRI furnace, which will be cleaned to extract hydrogen. The balance will return to the existing plant.

“It’s the only DRI plant in Europe with the capacity to source enough hydrogen [from methane in natural gas] to make it work,” Chevrier told Metal Market Magazine. He elaborated that Hamburg provides the opportunity to generate enough hydrogen cost-effectively at the rate needed to run the demonstration plant.

“The fundamental chemistry behind it works,” said Chevrier, but he acknowledged that there are a few technical challenges to address, “so we want to make sure we address all of them in the demonstration plant,” he added.

One of those potential challenges is what the DRI product made in a shaft furnace running on pure (or nearly pure) hydrogen will look like. Factors such as its physical properties and its melting performance are two key DRI characteristics to be tested.

Low- to zero-carbon content in the product is another factor to consider, which will result in a more reactive product that is susceptible to reoxidization and one that will no longer contribute to the carbon content actually needed in the steels produced from it. In other words, where steelmakers have relied on DRI to contribute to the essential carbon content required in their steels, other means will need to be found to add carbon to the steel melt.

“We also want to see how the MIDREX Process can be optimized to handle [pure] hydrogen,” Chevrier explained. Up to now, the thermodynamics of the existing reduction process have been held in balance by the fact that iron oxide reduction by carbon monoxide is exothermic (heat-releasing), counterbalancing the endothermic (heat-absorbing) nature of reduction by hydrogen.

Consequently, while the existing reduction of iron ore in the shaft furnace in the MIDREX Process is more or less balanced in energy, the pure hydrogen process will need additional energy input (i.e. heat) to perform the reduction. “We will verify how to do this through a demonstration at a large scale,” said Chevrier. “The theory is pretty straightforward. It’s an energy rather than an ironmaking problem. If energy is available at a reasonable cost, the process is absolutely doable,” he stressed.

The fundamental challenge is that the bonds between iron and oxygen atoms in iron oxide are very strong and consequently need a lot of energy to break. Coal has traditionally been an efficient source of the energy needed and lowers the melting temperature of the iron ore reduction process. “If you are displacing the carbon by something else [to reduce carbon emissions] you need to provide a large amount of energy by other means,” he summarized.

Producing hydrogen

Two key questions are where the large volumes of hydrogen that will be needed for hydrogen-based iron ore reduction will come from, and how they will be generated without consequential carbon emissions arising from that process.

Chevrier noted that, just three years ago, 96% of global hydrogen production was from fossil fuels – nearly half of which was from natural gas, nearly a third was from oil, and just under a fifth was from coal. He added that most of this so-called ‘grey’ hydrogen is generated and consumed on the same site and so is not traded or transported.

Electrolysis of water is another option, which accounted for the 4% balance of global hydrogen production but is rising rapidly. While the basis of electrolysis technology is well established, unless it is powered by renewable rather than fossil- fuel sources of energy, its use pushes carbon dioxide emissions further upstream. Chevrier also pointed out that the cost of hydrogen produced that way is currently too high for many applications at prevailing US electricity prices – about double the cost of hydrogen from steam reforming.

He also noted, however, that many projects are under way to advance the hydrogen economy, given its potential value in many energy sectors. While its production via renewable energy sources is one clear low-carbon route, a pragmatic approach is that when power produced by any means exceeds the immediate demand for it, excess power can be used to produce hydrogen as a form of energy storage. That hydrogen could also be sold as a product for industrial processes like steelmaking or for hydrogen-powered vehicles.

Current modern technologies for hydrogen production, such as proton exchange membrane (PEM), generate about 200 Nm³ per hour of hydrogen per MW of power. Consequently a 20 MW plant – at the upper end of the size range at present – could generate 4,000 Nm³ per hour of hydrogen. By contrast, Chevrier indicated that 60,000 Nm³ per hour of hydrogen can be substituted for about 20,000 Nm³/h of natural gas in a 2.0 million tonne per year DRI plant, which represents just 30% of the total natural gas consumption.

To summarize, the amount of power needed by electrolysis to produce the volumes of hydrogen needed for DRI production is very big. For example, a large DRI MIDREX Plant producing about 1.8 million tpy of DRI could require the power generated by more than 200 large offshore wind turbines, Chevrier estimated. He also noted that, with its location in Northern Germany, near the coast, Hamburg is well placed to access off-shore wind power.

Addressing challenges

Chevrier emphasized that cost-effective production of ‘green’ hydrogen will be essential for the three-step process that Midrex envisages will be used for steelmakers to progress to full hydrogen-based DRI production. At 6-12 cents per kWh, energy prices in Europe are about 4-10 times higher than where they need to be for cost-effective hydrogen production as of now, he noted, confirming that the cost of electricity drives the cost of hydrogen. “Both capital and operating costs need to be three to four times lower than they are now,” he added.

“Some people say that will never happen,” he acknowledged, but “wind power has achieved that in the past 10-15 years,” he added. He also stressed that the technology for hydrogen production is improving tremendously, and they can produce larger and larger volumes of hydrogen.

While it has already been established that the MIDREX Process can run with a wide range of hydrogen content, in making the transition from today’s typical reductant gas composition to the future’s, the question arises as to how swiftly a DRI plant could react to rapidly changing levels of hydrogen availability.

Chevrier acknowledged that, by its nature, renewable energy supply like solar or wind power, for example, can fluctuate in short timeframes according to weather conditions. He also noted that while Midrex has indeed already demonstrated that its technology can work with anything from 0-100% hydrogen – adding that the process can also be turned down from full to nearly half capacity within the period of a day if desired – the company is also looking at its responsiveness to fluctuating hydrogen availability.

He said that a relatively small change of hydrogen content of up to 5% could be made in a matter of hours, and that the process is more than capable of handling seasonal variations of hydrogen generation, or even on a shorter term weekly basis.

The answer to more rapid changes in hydrogen supply is a buffer stock. “Storage will be key to the hydrogen economy,” said Chevrier, pointing out that hydrogen is cheaper to store than electricity and that its costs of transportation are similar to the costs of electricity transmission. As a gas that has to be contained for safety, it also has very low yield loss during transmission and transportation.
He added that hydrogen can be added to an existing natural gas pipeline and that, in Germany, the country’s extensive pipeline network has potential to store and distribute hydrogen in the transition period.


Chevrier observed that the construction of 20 MW electrolysers are in progress and that there are already announcements of building 100 MW electrolysers.

“Commercial hydrogen-based DRI production may be just 5-10 years away where power is already cheap,” said Chevrier. “Ultimately, it depends on how you view global energy supplies,” he added. He believes that across all forms of energy production, there should be sufficient to provide enough hydrogen.

By Richard Barrett

To view the entire issue, please click here.


Latest Pricing Trends Year Over Year