© B. Voraberger et al., Published by EDP Sciences, 2026

1 Introduction

The iron and steel industry, alongside cement production, ranks among the largest industrial sources of CO₂ emissions globally. A significant portion of these emissions originates from ironmaking, which remains dominated by the integrated blast furnace (BF) route, heavily reliant on coal as the primary energy source. Depending on the BF operation, typically 400–700 kg of carbon carrier is used per ton of hot metal produced. Optimizing the existing BF operation and BOF operation, e.g., by increasing the scrap rate in steelmaking, is the most immediate and effective strategy for CO₂ emission reduction [1]. However, since scrap availability is limited and high-quality steel requires lowest tramp elements, iron ore–based production will remain essential.

Direct reduction (DR), currently driven by natural gas and in the future by hydrogen, is gaining prominence as a low-carbon alternative. While the DR-EAF route is established for high-grade ores, a robust, industrially viable pathway for utilizing abundant low-grade ores is lacking. This work aims to bridge that gap by proposing and experimentally validating the Smelter-BOF route. Figure 1 compares key steelmaking routes: the conventional BF–BOF route, a scrap-based EAF route, and emerging green pathways utilizing natural gas or hydrogen-based DR followed by either an EAF or a Smelter–BOF combination. Transitioning to green steel production entails a fundamental shift in energy carriers—from carbon-based fuels to green electricity. While the BF–BOF route (a) consumes approximately 170 kWh per ton of crude steel, both hydrogen-based DR routes (c) and (d) with on-site electrolysis demand approximately 4 MWh per ton, primarily for hydrogen generation.

Consequently, CO₂ emissions in green steelmaking are highly dependent on the electrical grid's carbon intensity. In Figure 1, a grid factor of 80g/kWh is considered as an overall target announced for the year 2050. Achieving a reasonable reduction of CO₂ emissions requires access to low-carbon electricity, underscoring that the green steel transition is fundamentally an energy transition—from fossil to electrical energy. Assuming electricity energy from renewable sources only (grid factor 0 g CO₂/kWh) would result in lower CO₂ emissions, especially for the hydrogen-based DR routes (c) and (d) in Figure 1. In the DR–Smelter–BOF route (d), the total CO₂ emissions can be reduced from 723 to 393 kg/ twith 100% renewable electric energy. This corresponds to approximately an 80% reduction of CO₂ emissions compared to the current integrated route using the same low-grade iron base.

In contrast, natural gas–based DR routes require significantly less electricity, between 0.6 and 0.8 MWh per ton of crude steel, resulting in lower Scope 2 emissions. However, due to higher Scope 1 emissions from fossil fuel-based reducing gas, total CO₂ emissions remain higher under the same grid factor. Still, this pathway offers a reduction of roughly 50% compared to the conventional integrated route, which typically emits around 2 tons of CO₂ per ton of crude steel. Since carburization in the Smelter is more challenging compared to BF, and carbon additions should be only as high as necessary to allow smooth BOF operation, a hot metal carbon content of 3.5% was considered compared to 4.3% in the BF hot metal. DRI from high-grade ore, production route (c), will result in higher reducing gas consumption for the same DRI-productivity and metallization degree of 94% compared to production route (d) with lower grade iron ore due to higher iron oxide content and therefore more reduction work. Per ton of crude steel consumption figures are further influenced by yield, scrap amount, and the resulting total steel output. The energy consumption per ton of crude steel in the smelter for hydrogen-reduced DRI is higher compared to natural gas-reduced DRI (672 kWh/t vs. 650 kWh/t) since carbon in the DRI reduces the melting point and contributes to the melting and reduction process [2]. In the EAF, it is similar with additional difficulties to generate foamy slag with hydrogen-reduced DRI. The lowest CO₂ emissions can be achieved through recycling of scrap in an efficient EAF with scrap preheating, which requires the lowest electric energy consumption. With renewable energy, less than 100 kg CO₂ per ton of crude steel is possible with production route (b).

Fig. 1 specific scope 1 and scope 2 CO2 emissions and electric energy consumption of current and future steelmaking routes per ton liquid crude steel (LS). For route (c) and (d) a hydrogen-based direct reduction Midrex H2 plant with an electric heater was considered for hydrogen and for natural gas-based direct reduction standard Midrex shaft with reformer. All routes considered an annual crude steel output of around 1 million tons. Route (a) with 2 different charge mix into BOF: 89/79% Hot metal and 11/21% scrap, route (b) with 100% scrap charging into Quantum EAF with scrap preheating; route (c) and (d) with hot DRI charging with a metallization of 94% and same DRI productivity of 125t/h; electric energy consumption for pelletizing and casting was not considered since it's minor. Natural gas (NG) and hydrogen (H2) values can vary based on plant and equipment configuration. System Boundaries include core process aggregates from iron ore agglomeration, BF, DR, DR+ Smelter ironmaking until EAF or BOF crude steelmaking. Dedusting, utilities and other auxiliary aggregates are not considered.

2 Iron ore availability and two-step process

For profitable operation of an EAF, high-grade DRI with low gangue content is preferred. This keeps the slag amount in the EAF process low and ensures low electric power and flux consumption, lower iron losses, and consequently a better yield. High-grade ore availability for production of high-grade DRI is limited as most of the global seaborne ores are of lower grade, well-suited for BF operation – see Figure 2. In this figure, the typical gangue content of the ores is also shown, which is close to 20% for lower-grade ores but just 3% for high-grade ores. Higher gangue content in DRI leads to increased flux requirements in the EAF to achieve the target slag basicity (CaO/SiO₂) of approximately 1.8 and ensure proper slag foaming. Consequently, lower-grade ores result in higher slag volumes, reduced EAF productivity, and increased energy demand—ultimately leading to higher operational costs. For example, the specific slag rate in EAFs using high-grade DRI is around 180 kg per ton of crude steel, whereas DRI from lower-grade ores can push this figure beyond 400 kg/t.

To overcome the limitation of processing DRI produced from low-grade ores, a two-step process is proposed [3,5].

• Step 1: Smelter – A closed, electrically heated furnace that melts and finalizes the reduction of DRI under reducing conditions, while also carburizing the metal and making slag adjustments.

• Step 2: Basic oxygen furnace (BOF) – Receives hot metal from the Smelter for refining, preserving existing infrastructure, and avoiding re-certification.

Compared to the EAF, the smelter energy input focuses less on arc heating and more on resistance heating, with a lower hearth load, enabling a longer refractory campaign. Large Soederberg electrodes are used mainly in two operation modes: immersed electrode in slag (resistance heating) and lifted electrodes with short arcing (brush-arc) between slag and electrode. The Smelter is operated with a lower slag basicity of around 1, reduced flux consumption, and overall slag generation. Additionally, the Smelter is designed to accommodate higher slag rates than EAFs, enabling more efficient processing of low-grade DRI while maintaining high iron yield and producing slag suitable for cement industry applications.

Fig. 2 Grade of iron ore traded globally; most of the ore is of low grade with higher gangue, resulting in enormous slag volumes when processed in EAF [3,4].

3 Effect of degree of metallization on the balance between DR and smelter

In an ideal case, DR and Smelter are installed at the same site, and the DRI is directly charged in hot condition into the furnace. Varying the metallization degree (MD) of the DRI allows for shifting reduction work between DR and Smelter and to optimize the process regarding productivity, energy consumption, or CO₂ emissions. An example of such variation in MD is shown in Figure 3 for a H₂-based MIDREX – Smelter combination, based on production route (d) in Figure 1, for lower-grade iron ore. The higher the MD in the DR plant, the higher is H₂ consumption and consequently the required electrical energy to produce the H₂. The remaining reduction work in the Smelter is decreasing with increasing MD of the incoming DRI, and consequently, the amount of reduction agent (carbon) and the electrical energy consumed in the Smelter. Total consumption of electrical energy is increasing with increasing MD of the DRI, while carbon consumption is decreasing, see Figure 3 (left).

CO₂ emissions of fully electrified, H₂-based DR plants with on-site electrolysis are mainly scope 2 and depend on the grid factor. In our example, 80 g of CO₂/kWh was considered, and consequently, the CO₂ emissions for the DR plant increases with MD as hydrogen and electrical energy consumption increase. On the Smelter, additional scope 1 emissions from the carbon carrier added as reducing agent need to be considered, and CO₂ emissions decrease with increasing MD as less reduction work is left for the Smelter. Specific carbon carrier additions in the Smelter drop from 93 kg to 70 kg per ton hot metal with 94% instead of 85% MD, considering a carbon fix content of 82% and 100% yield, which might be lower for industrial operation. Total CO₂ emissions are lower for high MD, where more reduction work is done in the DR using H₂, see Figure 3 (right). In an idealistic case when considering 100% electric energy from renewable sources and bio char instead of fossil carbon in the Smelter the total CO₂ emission will be lowest and not effected by the MD.

Fig. 3 Effect of the metallization degree for hydrogen reduced DRI on the electric energy consumption, smelter carbon additions and total CO2 emissions (scope 1 and 2). Basis of calculation same as in 1 production route (d) with grid factor of 80g CO2/kWh: Pelletizing + Midrex H2 plant with electric heater + Smelter + BOF for around 1 million ton of crude steel production per year, which is relevant for the heat losses.

4 Process principle and validation, furnace design and upscaling

A schematic of the Smelter process is shown in Figure 4 (left). The Smelter can be charged with various types of raw materials, including hot and cold DRI in the form of pellets, fines, or compacted briquettes. By default, hot DRI (HDRI) or hot compact iron (HCI) is preferred to maximize the use of sensible heat. Additional iron-bearing byproducts, such as slags, dust, and shredded scrap, can be used in proportions of up to 10%.

A carbon carrier, serving both as a reductant and carburizing agent, is mixed with the DRI and fluxes in a distribution bin located above the furnace. This mixture is then fed into the furnace through multiple feed legs. For rapid carburization, carbon injection lances or post-tapping carbon additions are also foreseen.

The Smelter is a closed furnace equipped with Söderberg electrodes, with electric heating provided primarily through resistance heating and brush arcing. Its closed design minimizes air ingress, creating a reducing atmosphere that enables efficient final reduction of the feed material. This results in low FeO content in the slag and high iron yield. To ensure optimal refining in the BOF, the hot metal should be carburized to a carbon content above 3.5%, providing sufficient chemical energy and blowing time for effective phosphorus and nitrogen removal [6]. Hence scrap rate in BOF after a Smelter will be lower than after the BF without additional measures for scrap rate increase as described in [1]. Slag and metal tapping of the Smelter is performed via tap holes, which are opened and closed using drilling and gunning machines.

As illustrated in Figure 4 (right), the carburization mechanism is essential for the Smelter. Due to the low density of the carbon carrier, it will mainly float on top of the slag and direct contact with the molten metal bath cannot be ensured. Therefore, carbon particles must come into contact with iron carriers, allowing carbon to dissolve into liquid iron droplets in the slag before they descend into the bath.

Fig. 4 Left: Smelter process flowchart with input and output; right: Schematics of the carburization mechanism inside the smelter.

5 Materials and methods

Smelter operation has been tested in a 600 kg pilot furnace, as detailed in [7]. Across three campaigns and eight test runs, various iron carriers—including DRI pellets, fines, and hot briquetted iron (HBI)—were processed with differing carbon contents and metallization degrees. Both coal and biochar were tested as carbon carriers for the reduction process to reduce residual iron oxides in the DRI and to carburize the molten metal bath. With good mixing of DRI and carbon carriers, a carbon content of more than 4% in the hot metal was achieved, meeting the requirements for optimal BOF operation.

The furnace has a hearth diameter of 926 mm in newly relined condition, three 150 mm electrodes, and a maximum. operation power of 400 kW. The first two campaigns were conducted using a sintered MgO-refractory, while the last one utilized an MgO-C and Al₂O₃-SiC-C-based material. Slag samples were analyzed by X-ray fluorescence (XRF) analysis and titration methods. Optical emission spectroscopy (OES) and LECO were used to determine the hot metal composition. Melt temperature was measured by standard Heraeus disposable thermocouples.

Table 1 summarizes the input materials of all eight heats, performed in three campaigns. Starting with DRI from DR-grade iron ores, also direct reduced product (DRP) from lower-grade ore in fine condition without agglomeration, hot briquetted iron (HBI), cold briquetted iron from fines (CBI), also mixtures with BOF slag and unreduced BF-grade iron ore pellets were tested. Anthracite and biochar were tested as carbon carriers. BF slag was added to simulate the higher slag rate from lower grade ores.

Table 2 and Table 3 summarize the major properties of the above-listed input materials.

6 Results

Figure 5 shows the slag samples in the ternary CaO-SiO-Al₂O₃ system. The values have been normalized to these three components. In principle, basicity (CaO)/(SiO₂) remained relatively stable. The normalized alumina content, mainly defined by the iron ore basis, remained between 10 and 20%.

As can be seen in Figure 6 center, MgO was high in campaigns 1 and 2. This was a result of refractory dissolution. By replacing it with Al₂O₃-SiC-C, and MgO-C-based ones for campaign 3, the problem was solved, as highlighted by the significantly lower MgO content. The left diagram shows a box diagram with [C] and B2. While the latter shows a stable behavior, the former fluctuated around an average of ∼3%. The right one shows the temperatures in each heat.

Fig. 5 Slag samples in the ternary Al2O3-CaO-SiO2 system, calculated with Factsage™ 7.3 using FToxid database.

Fig. 6 Left:– carbon content, (FeO), and basicity; center – (MgO) in campaigns 1 and 2 vs. campaign 3; right:– temperatures in form of box diagrams showing the 25 and 75% quartiles.

7 Upscaling and development roadmap

In a further step, Primetals Technologies, in collaboration with voestalpine and Rio Tinto, is developing an industrial demonstration plant under the Hy4Smelt project, integrating HYFOR with a Smelter capable of producing up to 3t/h of hot metal. The smelter in the Hy4smelt project is a round furnace with 3 electrodes and a maximum active power of around 3 MW, which allows feeding of fines as well as compacted DRI. The next phase involves a 37t/h rectangular furnace with 6 Soederberg electrodes in the HYREX project with POSCO, using sinter feed and fluidized bed technology. Engineering for a full-scale 1.25 Mtpa Smelter is complete, with commercialization targeted for 2028, pending final decisions based on prototype learnings [8]. Figure 7 illustrates the Primetals Smelter development timeline.

Fig. 7 Roadmap for development and stepwise upscaling from process validation tests, Hy4smelt round smelter pilot plant, HYREX rectangular smelter pilot plant and full size rectangular industrial smelting furnace.

8 The smelter as a recycling aggregate

In conventional iron and steelmaking, significant volumes of slag are generated—particularly in the blast furnace (BF), where slag rates can exceed 280 kg per ton of hot metal. As the industry transitions toward green steel production, BF and BOF slag volumes are expected to decline, while EAF slag volumes will increase substantially. Unlike BF slag, which is widely utilized as a secondary cementitious material (SCM), EAF slag currently has limited applications and is often landfilled.

Smelter slag, due to its similarity to BF slag in terms of basicity and composition, is well-suited for use as an SCM. Figure 8 illustrates the slag phase diagram and compares the composition and basicity of various slags (BF, Smelter, BOF, EAF) with cement clinker. The orange arrow indicates the potential for reduction with carbon, while the blue arrows show simultaneous reduction with e.g., carbon carrier, FeSi, and modification by fluxes e.g. quartzite.

To qualify as a cement clinker substitute, slag must meet specific chemical criteria to allow the formation of alite (C₃S) and belite (C₂S), which are essential for hydraulic activity. For use as SCM, latent hydraulic behavior is sufficient. BF slag, with a basicity B = C/S between 0.8 and 1.2, is typically granulated to achieve high glass content. Smelter slag, with similar properties, is expected to meet these requirements as well.

BOF slag, with a basicity around 3 and high iron oxide content, is less suitable for direct use in cement applications. However, tests have shown that reduction treatment followed by controlled cooling and grinding can modify BOF slag to make it viable as a clinker substitute [9], see the orange arrow in Figure 8, or SCM, see the blue arrow in Figure 6. EAF slag, with specific rates of 100–180 kg/t and high contents of FeO (up to 35%), MgO, and Cr₂O₃, presents additional challenges. Its current use is limited to low-value applications such as road construction, and in some regions, landfilling is required due to strict chromium limits.

The Smelter offers a promising solution for valorizing EAF slag by enabling both metal recovery and mineral modification through a thermal reduction process, see the blue arrow in Figure 8. By transferring most metallic oxides (e.g., Fe, Cr, V) into the metal phase and adjusting slag basicity during the liquid stage, the resulting slag can be granulated and used as an SCM or as a decarbonized raw material (e.g., limestone substitute). The composition and hydraulic properties of the modified slag will determine its economic value. The recovered metal fraction can be reused as scrap substitute or directly as hot metal, depending on its composition and the plant specific options for liquid charging.

Figure 9 presents a process flowchart for EAF slag valorization in a closed melting and reducing furnace. Reduction occurs at the slag–metal interface and within slag-entrained metal droplets. However, compared to DRI processing, slag treatment poses unique technical and economic challenges:

• Steelmaking slags contain significantly higher iron oxide levels (up to 35%), requiring more reduction work and higher energy and reductant consumption.

• Cr and V oxides are reduced only after FeO levels are sufficiently low, necessitating high reduction degrees and stirring to promote droplet formation and transfer of reduced metals to the bath.

• High share of scrap in the charge mix will increase the Cr and V oxides in the EAF slag which will after the reduction process end up in the recovered metal and slag fraction and lower its value potentially making the business case less attractive.

• High reduction degrees also transfer phosphorus to the metal bath, potentially lowering its value or showing the need for further dephosphoration treatment.

• The energy consumption for treatment of cold slag will be close to 1000 kWh per ton input slag – which is double compared to the required energy for liquid slag of 490 kWh per ton input slag material for industrial furnaces.

• As electric energy consumption and costs are the main part of the operational expenses for this pyrometallurgical treatment maximized energy efficiency and to lowering electric energy consumption steelmaking slag is key to make this business case viable.

• EAF Slag should therefore be preferably charged in liquid/hot form to the Smelter. For EAF slag, this is challenging due to continuous slag overflow and resulting temperature and viscosity drop while tapping. Batchwise charging of hot slag could be feasible.

• Cement and secondary cementous material usage is strictly standardized and regulated and must follow country specific guideline. One of the current challenging limits on the composition is the total Cr content of below 600 ppm in many European countries for usage as SCM.

Furnace size of such plant will depend on operation philosophy (batch vs. continuous operation), reductants used and conditions of charged EAF slag. The choice of reductants affects energy consumption, offgas volumes, flux demand, and operating costs. Reductants like FeSi, aluminum, calcium carbide instead of carbon-based material (coal, coke, biochar) reduce scope 1 CO₂ emissions. While the reduction process of iron oxides with carbon is endotherm, FeSi and Aluminum will reduce electric energy consumption due exothermic reaction and therefore lower equipment size and CAPEX mainly due to lower off-gas flow rates and lower required energy input. However, these alternative reduction materials might contribute to higher OPEX and higher Scope 3 CO₂ emissions.

In Table 4, a comparison of different technologies for handling and treating steelmaking slags is presented. On the far right, the conventional and most widely applied method is shown: cooling BOF/EAF slag in a slag pit, followed by cold processing through crushing, screening, and magnetic separation. This approach allows recovery of metallic iron only, while the oxidic metals, representing the major part, are lost. Furthermore, the mineral fraction remains unmodified, and due to slow cooling, the glass content is low—limiting the slag’s use to low-value applications such as road construction. In regions with strict chromium regulations, even these applications may be prohibited, resulting in landfilling. The main advantage of this established method is its simplicity and low operational cost, with minimal capital investment required.

In contrast, the far-left side of Table 4 illustrates reducing treatment in a Smelter. This method enables full recovery of metallic content and modification of the slag’s mineral composition, making it suitable for use in the cement industry as a secondary cementitious material (SCM). The resulting high-value output products can generate additional revenue, helping to offset the higher capital expenditure (CAPEX) for furnace installation and auxiliary systems, as well as the operating expenses (OPEX), which are primarily driven by electricity consumption.

Intermediate solutions—such as basicity adjustment for free lime stabilization followed by granulation—may achieve sufficient glass content for SCM use but do not recover oxidic iron or other valuable metals [10]. As a result, current limitations on iron and chromium content for SCM applications remain unresolved, restricting their use in the cement industry. However, stabilized slag may be used as filler material.

Direct dry, air or water granulation without basicity modification fails to recover oxidic metals or generate sufficient glass content, rendering the slag unsuitable for SCM applications. In contrast, reducing treatment in a Smelter offers a comprehensive solution: full metal recovery, slag modification, and production of high-value SCM [11]. This supports circular economy principles by eliminating landfilling and contributes to CO₂ emission reductions in both the steel and cement sectors.

Fig. 8 Slag phase diagram and typical composition, basicity, and specific amounts of iron- and steelmaking slags (BF, Smelter, BOF, EAF) and cement clinker. The orange arrow indicates the modification due to reducing treatment, the blue arrows indicate simultaneous reduction and modification towards BF/Smelter slags; Phase diagram calculated by Factsage™7.3 using FToxid database.

Fig. 9 Process flowchart for EAF slag valorization in a closed melting and reducing furnace followed by slag granulation and grinding.

9 Discussion

In the Smelter process validation tests all material mixes were molten successfully. Considering an industrial Smelter operation, the following major targets must be achieved:

• Melting of DRI with additives and final reduction of residual FeO.

• Achieving a hot metal suitable for BOF-steelmaking.

• Modifying the slag for cement production.

Looking at the tests in principle and the measured temperatures, Target a) was achieved. (FeO) consistently remained very low with 1.1% in average. This low (FeO) levels confirms the higher iron yield of the smelter compared to the EAF with (FeO) > 25% and allows usage in cement industry since its composition is very similar to BF slag with typical (FeO)∼0,5%, details see Figure 8. The temperature fluctuated between ∼1420 and 1650 °C. That can be explained by the small furnace size and will be more stable in large-scale operation. Hot metal carbon, relevant for Target b), was on the lower side with ∼3% on average but > 4% maximum. By premixing the carbon carriers and the iron carriers, carburization was generally improved. After relining the furnace with MgO-C and Al₂O₃-SiC-C-based refractories for the third campaign, (MgO) could be kept stable. Additionally, B2 and (Al₂O₃) were in acceptable ranges, which should allow the smelter slag to be used in cement industry as secondary cementitious material such as BF Slag. Therefore, all three targets were achieved in the pilot-scale tests. Clear message regarding Electrical energy consumption of the smelter for the same hot DRI (temperature, composition) will be higher in the Smelter compared to the EAF due to additional energy required for carbothermic final reduction and since in the EAF also chemical energy through side wall burners is used for melting.

The tests have shown that the smelter can efficiently process lower grade DRI with different metallization degree, chemical composition and size and is therefore increasing the operational flexibility of DRI plant by enabling variations in iron oxide and DRI quality. The optimum MD depends on several factors such as availability and cost of iron oxide materials, natural gas, hydrogen and electric energy as well as specific amount, composition and temperature of reducing gas, and needs to be evaluated for each target setting case. The achievable MD is also influenced by the direct reduction (DR) technology applied. The typical range for the MD for shaft furnace-based DR technology is from 92–96% [12]. For the EAF, but also for the Smelter higher MD of DRI is preferred to reduce energy consumption and increase productivity. Due to the reducing condition, the higher carbon additions and carbon yield the Smelter can process DRI with lower MD more efficient than EAF.

In the Development Roadmap the ongoing scale up from the smelter process validation tests to full industrial plant is described. The main furnace equipment principles of round and rectangular Smelters including Soederberg electrodes are known from other industries or applications. Hence focus will be to validate Smelter operations including consumptions figures, maintenance issues and dedicated equipment features in full industrial size. Primetals Technologies has chosen a step wise scale up with a scale up factor of below 10 and dedicated targets and risk mitigation for each scale up. Some of the main challenges, like Carburization, carbon yield and refractory lifetime will be investigated in detail in each step.

Utilization of smelter for steelmaking slag valorization in reducing treatment followed by granulation would be currently the most comprehensive and circular solution with benefits for both steel and cement industry. Higher CAPEX and higher electric energy demand compared to current slag handling in slag pit and mechanical processing or other technologies such as simple modification and granulation present roadblocks. Potential upsides of slag valorization with reducing treatment could be restrictions for landfilling and cement industry which is searching for alternative to declining amount of BF slag. Downsides of this business could be legislative restriction in usage of modified slag as SCM and additional CAPEX for treatment of metal product for usage in steel industry.

10 Conclusion

This work demonstrates that the Smelter–BOF route is a viable and efficient solution for closing the raw material gap in green steel production. Process validation tests on pilot scale level confirmed the Smelter’s ability to process various forms of low-grade DRI and carbon carriers, achieving FeO levels below 1.5% in slag and hot metal carbon contents exceeding 3.5%, which are essential for BOF operation. The resulting slag exhibited properties comparable to blast furnace slag, indicating its suitability as a secondary cementitious material. Process modeling further showed that, when powered by renewable electricity, the Smelter–BOF route can reduce CO₂ emissions by up to 78% compared to the conventional BF–BOF route, while also enabling slag valorization and metal recovery.

The novel contribution of this work lies in experimentally validating a two-step process that integrates a Smelter with BOF refining, offering an industrially scalable pathway for utilizing abundant low-grade ores and supporting circular economy principles through slag valorization.

Future work will focus on industrial-scale demonstration and optimization. Key priorities include:

• Operational validation in the Hy4Smelt and HYREX projects, including continuous feeding of hydrogen reduced fines and compacted DRI.

• Long-term refractory performance and maintenance strategies under reducing conditions.

• Detailed techno-economic analysis of the Smelter–BOF route, including energy consumption, CAPEX/OPEX, and slag valorization business cases.

• Further research on slag treatment for compliance with cement industry standards and environmental regulations.

These steps will be critical to confirm the technical and economic feasibility of full-scale deployment and to accelerate the transition toward low-carbon steelmaking.

Funding

Part of this research was funded by Austrian FFG Basisprogramm: Project Number 56152711.

Conflicts of interest

The authors declares no conflict of interest in regards to this article.

Data availability statement

Data associated with this article cannot be disclosed due to legal/ethical/other reason.

Author contribution statement

Conceptualization, B.V. and G.W.; Methodology, A.P.; Validation, A.P., K.P. and R.M.; Formal Analysis, R.M; Data Curation, A.P.; Writing – Original Draft Preparation, B.V.; Writing – Review & Editing, G.W.; K.P.; Visualization, B.V.; Supervision, G.W.; Funding Acquisition, B.V.

References

All Tables

All Figures

Fig. 1 specific scope 1 and scope 2 CO2 emissions and electric energy consumption of current and future steelmaking routes per ton liquid crude steel (LS). For route (c) and (d) a hydrogen-based direct reduction Midrex H2 plant with an electric heater was considered for hydrogen and for natural gas-based direct reduction standard Midrex shaft with reformer. All routes considered an annual crude steel output of around 1 million tons. Route (a) with 2 different charge mix into BOF: 89/79% Hot metal and 11/21% scrap, route (b) with 100% scrap charging into Quantum EAF with scrap preheating; route (c) and (d) with hot DRI charging with a metallization of 94% and same DRI productivity of 125t/h; electric energy consumption for pelletizing and casting was not considered since it's minor. Natural gas (NG) and hydrogen (H2) values can vary based on plant and equipment configuration. System Boundaries include core process aggregates from iron ore agglomeration, BF, DR, DR+ Smelter ironmaking until EAF or BOF crude steelmaking. Dedusting, utilities and other auxiliary aggregates are not considered.

In the text

	Fig. 2 Grade of iron ore traded globally; most of the ore is of low grade with higher gangue, resulting in enormous slag volumes when processed in EAF [3,4].
In the text

Fig. 3 Effect of the metallization degree for hydrogen reduced DRI on the electric energy consumption, smelter carbon additions and total CO2 emissions (scope 1 and 2). Basis of calculation same as in 1 production route (d) with grid factor of 80g CO2/kWh: Pelletizing + Midrex H2 plant with electric heater + Smelter + BOF for around 1 million ton of crude steel production per year, which is relevant for the heat losses.

In the text

	Fig. 4 Left: Smelter process flowchart with input and output; right: Schematics of the carburization mechanism inside the smelter.
In the text

	Fig. 5 Slag samples in the ternary Al2O3-CaO-SiO2 system, calculated with Factsage™ 7.3 using FToxid database.
In the text

	Fig. 6 Left:– carbon content, (FeO), and basicity; center – (MgO) in campaigns 1 and 2 vs. campaign 3; right:– temperatures in form of box diagrams showing the 25 and 75% quartiles.
In the text

	Fig. 7 Roadmap for development and stepwise upscaling from process validation tests, Hy4smelt round smelter pilot plant, HYREX rectangular smelter pilot plant and full size rectangular industrial smelting furnace.
In the text

	Fig. 8 Slag phase diagram and typical composition, basicity, and specific amounts of iron- and steelmaking slags (BF, Smelter, BOF, EAF) and cement clinker. The orange arrow indicates the modification due to reducing treatment, the blue arrows indicate simultaneous reduction and modification towards BF/Smelter slags; Phase diagram calculated by Factsage™7.3 using FToxid database.
In the text

	Fig. 9 Process flowchart for EAF slag valorization in a closed melting and reducing furnace followed by slag granulation and grinding.
In the text