Decarbonising the steel industry: modelling pathways in Japan

Summary

The global steel sector is one of the most emissions-intensive industries, yet there remains a lack of effective tools to support strategic planning for its decarbonisation.

To help address this gap, we developed TZ-OSeMOSYS-STEEL — an open-source energy system model designed to explore, identify, and quantify credible pathways for reducing steel industry emissions. While the underlying OSeMOSYS framework is typically used for power system modelling, this application highlights its flexibility and broader potential across hard-to-abate sectors.

To demonstrate its functionality, we applied the model to Japan’s steel sector. Analysis of model results shows that the steel sector will likely continue using BF-BOF under weak policies and low demand, but deeper decarbonisation and higher demand pathways strongly favour a transition to EAFs, especially with hydrogen-based feedstocks, by 2050.

Why are we applying system modelling to the steel industry?

The global steel sector is one of the most emissions-intensive industries, accounting for around 7–9% of global CO₂ emissions. As demand for low-carbon materials grows, the sector faces mounting pressure to define realistic and cost-effective decarbonisation pathways. To support this effort, we developed TZ-OSeMOSYS-STEEL — a new, open-source energy system model designed to explore long-term technology and investment strategies for reducing emissions in steel production. While the model is designed to be replicable across geographies, we chose Japan as a test case due to its global significance as the third-largest steel producer, its substantial export share, and the emissions intensity of its domestic steel sector. Japan’s policy targets and market pressures make it a compelling context in which to demonstrate the model’s capabilities and extract lessons for other high-emitting industrial economies.

Our focus country: Japan

Japan is the world's third-largest producer of steel, with an output of 84 million tonnes in 2023¹. Given this significant production volume and the emissions intensity of steel production, the sector is responsible for approximately 14% of the country's total CO2​ emissions² and a substantial 40-48% of its industrial CO2​ emissions³. This makes the steel industry a critical focal point for achieving national climate objectives.

The drive for decarbonisation is intensified by both national and international commitments. Japan has set a target for carbon neutrality by 2050 and aims for a 30% reduction in emissions from the steel industry by 2030, relative to a 2013 baseline⁴. However, Japan has been characterised as a "laggard" in steel decarbonisation when compared with other G7 nations, ranking poorly in policy analyses.

Moreover, the economic competitiveness of Japan's steel industry is increasingly at stake. With approximately 40% of its domestically produced steel destined for export markets, the industry faces a future where the carbon intensity of its products will significantly influence market access and competitiveness. The global demand for ‘green steel’ is on the rise, particularly from sectors like automotive manufacturing. Should Japan's steel products remain carbon-intensive while competitor nations advance their decarbonisation efforts, the industry risks losing valuable export markets or facing punitive measures such as carbon border adjustment mechanisms (CBAM).

The combination of maintaining cost-competitiveness coupled with a decarbonisation imperative creates a strong case for robust modelling to identify credible decarbonisation pathways for Japan’s steel sector. We built TZ-OSeMOSYS-STEEL⁵, a long-term systems model in order to explore, identify, and quantify these potential pathways. In this analysis, we model Japan’s steel sector from the present to 2060 to find and compare the cost-optimal mix of steel production routes in line with Japan’s decarbonisation targets, commodity price projections, and technology development. To accurately model this, it’s crucial to first understand the existing production landscape of Japan's steel sector.

Current steelmaking routes

Steel is currently produced predominantly via the Blast Furnace-Basic Oxygen Furnace (BF-BOF) route, which accounts for approximately 73-76% of Japan’s steel production². This method, which involves primary steelmaking from iron ore using coking coal as a reductant and energy source, is inherently carbon-intensive.

The rest – around 25% is produced via the Electric Arc Furnace (EAF) route. EAFs primarily use scrap steel as their main feedstock and rely on electricity as their principal energy source. Consequently, the carbon intensity of EAF steel is significantly lower than that of BF-BOF steel, particularly when the electricity consumed is generated from renewable sources.

The BF-BOF process typically emits around 2.3 tonnes of CO2​ per ton of crude steel produced (tCO2​/t-steel). In contrast, scrap-based EAF production emits considerably less, approximately 0.3 tCO2​/t-steel⁶. These figures highlight the substantial emissions reduction challenge associated with the dominant BF-BOF route in Japan.

Alternate steelmaking routes

Several alternate technological pathways are being considered – and modelled here – for the decarbonisation of Japan's steel sector. These range from incremental improvements to existing processes to technological shifts towards entirely new production methods.

Innovating traditional routes: COURSE50, SuperCOURSE50, and CCS

Introduced in 2008 and 2018 respectively, COURSE50 and SuperCOURSE50 are Japanese steel industry initiatives aimed at reducing CO2 emissions from conventional blast furnace–basic oxygen furnace (BF-BOF) steelmaking.

The COURSE50 project aims to achieve an approximate 30% reduction in CO2​ emissions by 2030 (compared to 2013 levels) through measures including the injection of hydrogen into blast furnaces and the application of Carbon Capture and Storage (CCS). Building on this, the SuperCOURSE50 project targets a ~50% CO2​ reduction by using heated external hydrogen, with commercialisation anticipated by 2040. Recent tests in experimental blast furnaces— which are several hundred times smaller than their commercial-scale counterparts — have demonstrated reductions in emissions intensity of up to 43%. CCS remains an essential component for both COURSE50 and SuperCOURSE50 to meet their stated emission reduction goals.4 However, the deployment of CCS in Japan faces considerable hurdles, including the availability of suitable geological storage sites and the high costs associated with capture, transport, and long-term storage.

These pathways are widely regarded as transitional measures as they offer partial emission reductions while retaining reliance on fossil-based processes and CCS. The modelling will assess their net CO2​ reduction potential, factoring in emissions associated with non-“green” hydrogen production. Other critical outputs are , the cost-effectiveness of these modifications and the risk of these assets becoming stranded if deeper decarbonisation is required post-2030 or 2040. These pathways with CCS all present a dilemma: whether they are a pragmatic bridge to a lower-carbon future or a costly diversion of capital that might only achieve partial decarbonisation. Given the substantial investment required for BF modifications and CCS infrastructure, and the significant uncertainties surrounding large-scale, cost-effective CCS deployment in Japan due to limited storage capacity, modelling can help compare these options against more transformative, albeit challenging, alternatives.

Expanding electric routes: the role of scrap, EAF efficiency, and renewable power

Increasing steel production via EAFs is a prominent decarbonisation strategy. Japan generates approximately 37 million tonnes of steel scrap annually, with around 85% being recycled back into steelmaking, predominantly through EAFs. Maximising the use of domestic scrap is a key element of many decarbonisation scenarios. Technological advancements are also broadening the capabilities of EAFs, enabling them to produce a wider range of steel grades, including some that were traditionally the exclusive domain of the BF-BOF route. This expands the potential market share for EAF-based steel.

The cornerstone of decarbonising the scrap-EAF route is the provision of renewable electricity. A significant increase in renewable energy generation would be needed to power an expanded EAF fleet and to produce green hydrogen for other steelmaking routes. Estimates suggest an additional 3 TWh of renewable electricity will be needed by 2030 to support a shift towards EAFs. To illustrate the scale, a new EAF with a 2 million tonne per annum (Mtpa) capacity would require approximately 0.76 TWh of electricity annually, which could require the construction of a dedicated solar power installation of around 530 MW.

When modelling this pathway, critical inputs include scenarios for scrap availability (considering domestic supply versus imports, and constraints related to scrap quality), the pace and cost of renewable energy deployment, and the capital costs associated with new EAF installations. The impact of government policies, particularly those that might favour or disadvantage EAFs compared to BF-based routes, is also a crucial parameter. While Japan possesses significant domestic scrap resources, the quality of obsolete scrap can make it challenging to produce the full spectrum of steel grades demanded by the market. Models should therefore explore scenarios that incorporate investments in advanced scrap sorting and cleaning technologies, or the strategic blending of scrap with virgin iron units like Direct Reduced Iron (DRI) or Hot Briquetted Iron (HBI), to overcome these quality limitations and increase the decarbonisation potential of the EAF route.

Given the high energy costs associated with domestic production of HBI, Japanese steelmakers are investing in overseas HBI projects, particularly in regions with abundant and cost-effective natural gas or future potential for green hydrogen, such as the Middle East and Australia. This import reliance ensures a stable supply of high-quality metallics, positioning HBI as a key enabler of Japan's transition to a more sustainable steelmaking model.

The hydrogen frontier: H2-DRI and its potential

One key long-term decarbonisation pathway involves the use of hydrogen, ideally "green" hydrogen – produced via electrolysis powered by renewable energy – to directly reduce iron ore. This Direct Reduced Iron (DRI) is then typically melted in an EAF to produce steel (H2-DRI-EAF). This process has the potential to achieve greater than 95% emissions reduction compared to the conventional BF-BOF route.

Multiple projects around the globe are pursuing such H2-DRI technology. In Japan, major steel manufacturers are engaged in its development, including JFE Steel, and Nippon Steel, who are aiming for industrial-scale H2-DRI production by around 2040. To support this, the Japanese government's Green Innovation Fund provides financial support for the development of hydrogen-related technologies, including those applicable to steelmaking.

Modelling the H2-DRI pathway requires careful consideration of several critical factors. These include the projected cost of green hydrogen in Japan, which is anticipated to be high compared to other regions due to land constraints for renewable energy deployment and potential reliance on imports. Secondly, the availability of high-grade iron ore, which is the preferred feedstock for current DRI processes, also presents a challenge as such ores are not abundantly available globally, and processing lower-grade ores to improve their grade would increase costs, which is a more significant obstacle than the resource availability. Furthermore, models must incorporate realistic estimates for the capital and operational costs of H2-DRI plants and their associated EAFs, as well as the extensive infrastructure requirements for green hydrogen production, storage, and transportation.

It is also clear that the success of H2-DRI in Japan is not merely a function of the steelmaking technology itself but is deeply intertwined with the broader national strategy for establishing a large-scale, cost-effective green hydrogen economy. This, in turn, hinges on a revolutionary expansion of renewable energy capacity. Therefore, any steel sector decarbonisation models focusing on H2-DRI must be linked with, or informed by, comprehensive national energy system models that project renewable energy capacity, hydrogen production costs, and the potential scale and cost of hydrogen imports under various policy and technological development scenarios.

Open

Methane: a critical challenge

It is important to note that while steelmakers focus on reducing CO2 from their production facilities, significant and persistent methane leaks from both traditional BF-BOF and newer DRI routes remain a critical emissions challenge. For the BF-BOF process, the issue lies upstream with fugitive methane released during the mining of metallurgical coal, a key raw material. These substantial emissions occur long before the coal reaches the steel plant and are therefore unaffected by on-site CO2 capture technologies. In the case of DRI, which predominantly uses natural gas as a reductant, methane—the primary component of natural gas—leaks from extraction wells, pipelines, and processing facilities along the supply chain. Consequently, even as steel plants implement technologies to cut their direct CO2 emissions, these indirect, fugitive methane releases from both coal and natural gas supply chains will persist, undermining the overall climate benefits of the emission reduction efforts.

Carbon capture and storage (​​CCS)

The widespread adoption of Carbon Capture and Storage (CCS) within Japan's steel sector is fraught with challenges, casting doubt on its near-term viability as a primary decarbonization pathway. Integrated steel mills, with their multiple and varied emission sources such as blast furnaces, coke ovens, and power plants, present a complex engineering problem for comprehensive carbon capture. This complexity contributes to a low national adoption rate and a history of lower-than-promised capture rates in demonstration projects, such as the partial underperformance at the Tomakomai facility⁷.

Globally, CCS projects have a high failure rate, a trend that exacerbates the uncertainty for Japanese steelmakers already grappling with unclear cost structures and significant energy requirements for capture, transport, and storage⁸. The technical and economic hurdles of transporting CO2, whether by pipeline or ship, to suitable storage sites are substantial. Furthermore, Japan's geological landscape, characterised by numerous fault lines, raises significant concerns about the long-term security and liability of CO2 storage.

This underscores the immense challenge and scepticism surrounding the large-scale deployment of CCS as a silver bullet for the Japanese steel industry's carbon emissions and its readiness to deploy CCS at scale⁹.

Scenario setup

We set up a matrix of scenarios to explore decarbonisation pathways for Japan’s steel sector out to 2050. The underlying model identifies the cost-optimal mix that meets the user-defined policy targets and constraints. The scenarios considered are a combination of 3 demand scenarios (low, medium, and high) and 3 decarbonisation constraints (unconstrained, net zero, and net-zero-scope3). Together, these combinations result in 9 scenarios. Comparing the scenarios across the demand and decarbonisation axes reveals how each influences the overall steel production mix by 2050.

The three demand projections are provided by Nippon Steel Research Institute (NSRI) and described in the methodology document.

Model results

A changing production mix

Analysis of our model results shows that BF-BOF remains a significant share of steel production in the low demand condition, regardless of the emission reduction goal, representing almost 100% of steel production in the unconstrained scenario, 60% in the NZ scenario and 50% in the NZ-Scope3 scenario.

EAF begins to emerge from 2035 – representing over 15% of steel production from 2035 onwards across all decarbonisation pathways in the medium demand scenario. The share stays the same in the unconstrained scenario, increases to 80% for NZ, and reaches 100% for NZ-Scope3. In the first two cases, the EAF is supplied entirely with HBI-NG while in NZ-Scope3 results in a combination of supply from HBI-NG (40%) and HBI-H2 (60%).

DRI-Gas plays a small role as a transitional route in the high-demand scenarios from 2030 to 2045, representing around 2.5% of steel production — a level that could be met by a single DRI plant¹⁰. BF-BOF is completely phased out by 2045 under NZ-Scope3. All steel produced by EAF in 2050, with two-thirds using HBI-H2 and the remaining using HBI-NG.

The cost of decarbonizing

Comparing the costs and emissions of each scenario gives the CO2 abatement cost for each. These are calculated as the difference in system costs between the decarbonisation scenarios and the unconstrained scenario, divided by the difference in emissions between the two scenarios. The results show a relatively low cost range of 7-9 $/tCO2 across the net-zero scenarios. However, these increase significantly - between 3 to 9x - across the Net-zero (Scope 3) scenarios. This is driven by the fact that the EAF route with HBI-NG as a feedstock is already the cheapest route on an LCOS basis, and its emissions intensity is 0.7 tCO2/tSteel compared with 2 tCO2/tSteel for the currently dominant BF-BOF (Coke) route. However, when Scope 3 emissions are included for the HBI-NG feedstock, steel production shifts to a combination of HBI-NG and HBI-H2 fed EAF routes. The higher cost of HBI-H2 is then reflected in the higher abatement cost.

Conclusions

BF-BOF will prove difficult to retire, due to a considerable degree of invested capital, deeply entrenched supply chains for raw materials like coking coal and iron ore, and a workforce with specialised skills in this traditional technology. The transition towards alternative, lower-carbon routes, or the significant modification of existing BF-BOF assets, involves not only the technical feasibility and cost of new technologies but also the challenge of managing the phase-down or replacement of these massive, long-lived industrial facilities. There is also significant potential for stranded assets, where existing infrastructure becomes economically unviable before the end of its operational life due to changing regulatory or market conditions.

While EAFs currently represent about a quarter of Japan's steel production, their full decarbonisation potential hinges on two main factors: the availability of high-quality scrap and the greening of the electricity grid. The emissions from EAFs are predominantly Scope 2, arising from electricity consumption. Therefore, as Japan pursues its national goals to increase the share of renewable energy in its power generation mix, the carbon footprint of EAF steel will correspondingly decrease. Power system modelling is required to explore options that involve quickly increasing renewable energy supply, specifically for industrial applications. These options would reinforce the findings here that expanding EAF production, coupled with a rapidly decarbonising grid, could offer a stronger and quicker way to cut emissions in the short to medium term – possibly more than what current national plans, which focus more on BF-based producers⁹ – suggests.

The model results show that Japan’s steel sector follows different decarbonisation pathways depending on demand levels and emissions targets. In low-demand scenarios, the BF-BOF route remains a major production method. For medium demand, EAF production, – particularly with HBI-NG and HBI-H2 – becomes dominant from 2035 in net-zero scenarios. In high-demand scenarios, DRI-Gas plays a minor transitional role, with BF-BOF being phased out by 2045 under Net-zero (Scope 3), leading to all steel being produced by EAF, predominantly using HBI-H2.

The CO2 abatement costs are relatively low for Net-zero scenarios (7-9 $/tCO2) but rise sharply for Net-zero (Scope 3) scenarios, mainly due to the higher cost of HBI-H2 when considering the full lifecycle emissions.

What's next?

TZ-OSeMOSYS-STEEL is designed to be replicable in other countries and regions; scalable to different spatial resolutions from asset- to national-level. It can be used to increase transparency and accessibility of system modelling for the global steel sector, a significant source of emissions.

TZ-OSeMOSYS-STEEL- is available under an AGPL open-source license. Access it on Github here.

References:

1. Production Data for 2024, Japan Iron and Steel Federation
2. Steel Policy Scorecard: Country Profile – Japan, E3G (2023)
3. Japan’s Green Steel Transformation: Policy Recommendations, Renewable Energy Institute (2023)
4. Net Zero with Green Steel, JapanGov (2024)
5. TZ-OSeMOSYS-STEEL is a novel implementation of TZ-OSeMOSYS, an open-source energy system modelling package developed by TransitionZero. It builds on the widely used OSeMOSYS modelling framework.
6. Using scrap — from steel recycling mills — in EAFs significantly cuts carbon emissions compared to traditional virgin steel production, as it bypasses the energy-intensive process of reducing iron ore with coal.
7. Report of Tomakomai CCS Demonstration Project, METI, NEDO, and JCCS (2020)
8. Carbon Capture Has a Long History. Of Failure., Institute for Energy Economics and Financial Analysis (2022)
9. CCS Readiness Index, Global CCS Institute
10. Building Direct Reduction Plants Brochure, Midrex Technologies