Injection of Hydrogenous Gases into the Blast Furnace Tuyeres for Reducing CO₂ Emissions: A Review

Abstract

Transitioning to alternative fuels like natural gas (NG), coke oven gas (COG), or hydrogen presents a viable path to reduce carbon emissions in blast furnace ironmaking. A review of industrial practices and simulations reveals significant variabilities in estimating the impacts of hydrogenous gases’ tuyere injection on the blast furnace operation due to a lack of representative industrial-scale experiments and diverse modeling approaches. NG is effective yet regionally constrained, while COG, although beneficial for improving efficiency and reducing emissions, has limited availability for injection. High coke replacement efficiency ranks NG superior for emission reduction. Meanwhile, COG and cold hydrogen offer similar CO₂ reduction potentials, with carbonaceous gases offsetting COG’s higher coke replacement ratio. Preheating hydrogen can further improve its coke replacement efficacy, achieving the greatest reduction in CO₂ emissions. Hydrogen injection has yet to reach economic viability, but technological progress, scaling up, and evolving carbon legislation could alter this, prompting steel producers worldwide to advance hydrogen injection techniques in blast furnaces.

1. Introduction

The global steel industry, contributing approximately 8% to total energy system emissions or around 2.8 Gt of CO₂ annually,¹⁾ is challenging to decarbonize due to its substantial reliance on fossil fuels and significant capital investments.²⁾ Blast furnace (BF) ironmaking, the foundation of the global steel sector, accounts for the largest share of these emissions.³⁾

Nowadays, the progress in ironmaking is discussed predominantly in terms of the shift from the BF – BOF (basic oxygen furnace) steelmaking route to the hydrogen-based DRI (direct reduced iron) – EAF (electric arc furnace) route. For example, the HYBRIT project of SSAB, Sweden started pilot operations in 2020 and aims at an industrial-scale demonstration by 2026.⁴⁾ Stegra (former H₂ Green Steel), in partnership with Midrex, Paul Wurth, and SMS Group, builds a 2.5 million t per year plant in Sweden, set to launch in 2025,⁵⁾ and plans a similar facility in Spain.⁶⁾ Hydrogen-based DRI projects are taking shape in other countries worldwide.^7,8,9) However, the growth in the global share of the H₂-DRI-EAF route is limited by green hydrogen generation capacity¹⁰⁾ and the shortage of DRI-grade ores with an iron content of above 67%¹¹⁾ while the extent to which BF-grade ore converted into DRI will be accepted by the steelmakers is uncertain.¹²⁾ The International Energy Agency (IEA) anticipates that BFs will account for 83% of ironmaking in 2030 (92% in 2022), highlighting the importance of research and innovation for reducing CO₂ emissions of BF ironmaking.¹³⁾

The use of hydrogenous gases such as natural gas (NG), coke oven gas (COG), or even hydrogen to partially replace fossil fuels and reduce CO₂ emissions in the BF process is being increasingly discussed.¹⁴⁾ However, large-scale industrial hydrogen injection has not yet been implemented, so the impact on BF operation is mostly based on modeling, which has only recently been verified on an experimental BF.¹⁵⁾ Even for NG and COG, despite their long-term industrial use, estimates of optimal injection rate, coke replacement ratio (CRR), and CO₂ emissions reduction vary widely due to a lack of comparable industrial data. To address this uncertainty, this paper will review and summarize available industrial and theoretical data and compare them with the results obtained using the model proposed by the author.

The injection of hydrogen-rich gases influences multiple features of BF operation, therefore, extensive research in this domain addresses a broad spectrum of theoretical and practical aspects. These areas include the physical chemistry and thermodynamics of fuel combustion and reduction of iron oxides, the heat and mass transfer within different BF zones, optimal operational practices, and the design of tuyeres, among others. Such a breadth of analysis is too expansive for a single review. Therefore, this paper adopts a macro-level approach, concentrating on the primary indicators of the impacts of injection and the technological viability of the injection regimes.

2. Methodology Approaches for Modeling the Impacts of Injection

To predict and visualize the effects of injecting hydrogenous gases researchers apply advanced modeling tools. For example, Zhuo et al.¹⁶⁾ and Okosun et al.¹⁷⁾ applied 3D CFD (Computational Fluid Dynamics) models to study the effects of substituting hydrogen for pulverized coal injection (PCI) and NG on the processes in the raceway. Yu et al.¹⁸⁾ studied the penetration of the injected flow and shifts in the cohesive zone shape and position using a 2D model, while Zhao et al.¹⁹⁾ and Li et al.²⁰⁾ used 3D models to study hydrogen injection into the lower shaft. Mauret et al.²¹⁾ applied a 2D model developed by Paul Wurth²²⁾ to study the impacts of NG, COG, and H₂ injection on the BF process. Nogami et al.²³⁾ used a BF simulator based on multi-phase fluid dynamics, reaction kinetics, and transport phenomena fundamental framework to assess the impact of hydrogen injection on the vertical temperature and pressure patterns in the BF. Castro et al.²⁴⁾ applied a 3D five-phase numerical model, and Takatani et al.²⁵⁾ developed a 3D dynamic model considering gas-liquid-solid phase chemical reactions and phase change based on the mass, momentum, and energy conservation, successfully verified for the injection of hydrogenous gases at an experimental blast furnace.

The complexity of the process and the lack of data on the physicochemical and mechanical processes occurring in the different parts of the BF and on the quantitative aspects of vertical, radial, and circumferential patterns of the structure and granulometry of the material in the BF challenge accurate modeling. The number of assumptions often exceeds the available data confirmed in practice. Therefore, alongside the development of advanced modeling tools, simpler phenomenological analysis methods, for example, developed by Reichardt,²⁶⁾ Rist,²⁷⁾ and Kitaev,²⁸⁾ are still used to validate predictions. Many authors use relatively simple models of energy and mass balance including zonal models.^29,30,31,32) For example, Barrett et al.³⁰⁾ used a two-stage heat and mass balance model calibrated to an operating blast furnace. Yilmaz et al.,³³⁾ developed a steady-state energy and material balance model using Aspen Plus software, linked to the thermodynamic databases and equilibrium calculations in FactSage and ChemApp.

To keep the modeled operation regimes and ranges of the gases’ injection rates within technologically reasonable limits, researchers often apply constraints to various BF operation indices, for example, such as top gas temperature (TGT), raceway adiabatic flame temperature (RAFT), or pinch point temperature (a temperature level in the BF reserve zone where the difference in temperatures of gas and solid reaches minimum). Geerdes et al.³⁴⁾ determine an operation window as a condition when gas injection does not shift TGT below 110°C and RAFT below 2050°C. Many researchers share this approach with the boundary conditions widely varying for TGT (110–150°C) and RAFT (1900–2324°C), as summarized in Table 1. Some researchers apply additional conditions, e.g., Sato et al.³⁵⁾ and Barrett et al.³⁰⁾ stabilize a pinch point temperature at 1000°C. Pistorius et al.²⁹⁾ following the approach by Peacey and Davenport³⁶⁾ stabilize the wüstite-iron reduction equilibrium point at 927°C (coincides with the thermal pinch point for the BF shaft efficiency of 100%³⁷⁾). Spanlang et al.³²⁾ keep TGT constant at 150°C and allow RAFT to decrease. Several authors,^23,38,39) in addition to stable RAFT, also keep the bosh gas rate constant to stabilize a sensible heat supply to the reaction zone,²³⁾ ensure a good thermal state in the hearth, and maintain the stability of the raceway.³⁹⁾

Table 1. Industrial and theoretical data on the impacts of NG, COG, H₂ injection on CRR and CO₂ emissions (I – industrial data, P – PCI replacement, E – experimental BF, H – preheated, * - lower boundary).

Ref.	CRR, kg/m3	Injection rate, m3/t-HM	CO2 emissions reduction, %/m3	O2, %	Blast temperature, °C	RAFT, °C	TGT, °C
Natural Gas Injection
21)	0.65	55	0.02	41	N.A.	2147*	90/137*
36)	0.7	100	N.A.	24.3	1127	2127*	177
35)	0.79	70/140	N.A.	34/59	1150	2225*	131/79
47)	0.7–0.8	<100	N.A.	≈25	950–1150	N.A.	N.A.
31) I	0.72–0.75/0.8	64–94/93–103	N.A.	21/26	1003–1107/1091–1124	N.A.	N.A.
29)	0.81	230	N.A.	48.5	1000	1711	152
48) I	0.68–0.87	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.
49)	0.94	127	N.A.	31.3	N.A.	N.A.	150*
42) I	0.78–1.1	62–169	N.A.	21–34.7	900–1169	N.A.	N.A.
Coke Oven Gas Injection
71) E,P	0.12	217	N.A.	28.8	1126	2202	N.A.
38) P	0.2	59–152	0.11	24–40.22	1152	N.A.	N.A.
21)	0.24	162	0.05	41	N.A.	2147*	88/137*
50,51) I	0.30–0.35	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.
70) I	0.36	35	N.A.	N.A.	N.A.	N.A.	N.A.
52)	0.4	41–203	N.A.	21.5–23.7	1122	N.A.	244
53)	0.42	240	N.A.	34	1100–1200	2000*–2200	110*–180
48)	0.45	219 (max)	N.A.	N.A.	N.A.	N.A.	N.A.
54)	0.12/0.48	173/35	N.A.	36/27	1160	2206	172
42)	0.45–0.48	50–100	N.A.	21–25	N.A.	N.A.	N.A.
42) I	0.49	80.2	N.A.	21.1	1125	N.A.	N.A.
72) E	0.49	93	N.A.	36.2	1003	N.A.	101
Hydrogen Injection
24)	0.12	225	0.03	43	NA	2472	182
30) P	0.15	211	0.04	27	1153	2050*	118*
21)	0.16	133	0.04	36	N.A.	2147*	90/137*
15) E	17%	359	0.04	39.5	1000	2097	N.A.
81)	0.21	250	0.09	37	N.A.	1890	100
55)	0.22	250	0.05	35	N.A.	1900*	130*
23)	0.22	133	N.A.	23.5	1200	2324*	140
32)	0.27	100/200	0.06	N.A.	N.A	2160/2022*	150*
53)	0.28	356	N.A.	30	1100–1200	2000*–2200	110*–180
81) H	0.33	350	0.07	37	N.A.	1901	120
33) H, P	0.34	306	0.07	21.1	1200	2150	108
39)	0.38	120	0.04	31.4	1150	2172*	200

This paper will compare the available industrial and theoretical data with the results obtained using the 1-D steady-state zonal model proposed by the author and detailed elsewhere.^40,41) It is based on Kitaev’s²⁸⁾ theory of heat exchange in shaft furnaces and Ramm’s⁴²⁾ method of calculating BF operation parameters. The change in the heat flow ratio, defined as the ratio between the heat flow capacities of the condensed material (W_m) and gas (W_g), determines three stages of heat exchange in the vertical temperature profile of the BF illustrated in Fig. 1.

Fig. 1. Schematic of heat exchange conditions and a vertical profile of W_m, W_g, t_m, t_g, and Δt in the BF.

When W_m = W_g, the cooling of the gas by 1°C in the countercurrent should be followed by an increase in the solid temperature also by 1°C.²⁸⁾ In practice, both phases’ temperatures do not change significantly along a part of the BF height called the thermal reserve zone. It accommodates a chemical reserve zone where the development of reduction processes also slows down.⁴³⁾ When W_m = W_g a minimum difference in gas (t_g) and solid (t_m) temperatures (Δt_min) is observed. The temperature in this region, referred to as the thermal pinch point, is generally considered in the range of 900–1050°C.^26,44,45) Kitaev et al.²⁸⁾ used in their studies Δt_min=20°C, whereas Ramm,⁴²⁾ based on zonal balances calculated for operational blast furnaces, argued that it varies in a wide range and may reach up to 150°C.

Various authors apply either volumetric or mass injection rates. This paper uses volumetric values in normal cubic meters abbreviated as m³. For comparison, the mass values reported by certain authors are converted to volumetric values. NG density may range from 0.55 to 0.87 kg/m³ depending on the composition.⁴⁶⁾ For conversion, a density of 0.78 kg/m³ exhibited by the NG composition listed in Table 2 under normal conditions was used. Similarly, a density of 0.457 kg/m³ was applied to COG.

Table 2. Typical chemical compositions of natural gas and coke oven gas, vol%.

Gas	CH4	C2H6	C3H8	H2	CO	CO2	N2
NG	92	3	1	0	0	2	2
COG	27	2	1	58	7	2	3

3. Industrial and Modeling Data on the Impacts of Injecting Hydrogenous Gases

3.1. Natural Gas Injection

NG injection was first deployed in Ukraine in 1957 and then used at over 30 BFs until the 1990s with a CRR of 0.7–0.8 kg/m³ observed at injection rates of up to 100 m³/t-HM (Hot Metal) for blast temperatures of 950–1150°C with the oxygen content of around 25%.⁴⁷⁾ Overall, in the former Soviet Union, 112 BFs out of a total 133 in the 1980s operated with NG injection in the range of 70–100 m³/t-HM thanks to its availability, low cost, and relatively minor capital investment requirements, providing a considerable decrease in coke rate.⁵⁶⁾ Historically, in the USA, the injection levels varied in the range of 26–128 m³/t-HM with an average of 51 m³/t-HM.⁵⁷⁾ In 1989, in North America (US and Canada) NG was injected at 23 BFs with a maximum injection rate of 75 m³/t-HM and an average of 32 m³/t-HM.⁵⁸⁾ Co-injection of NG and pulverized coal was implemented at two BFs in Ukraine in the 1990s⁵⁶⁾ and is increasingly often discussed^59,60) in the context of CO₂ emissions reduction.

Technologically, the NG injection faces constraints due to a reduction in the RAFT, potentially impeding coke gasification and ore burden smelting. Partial mitigation of this constraint is achievable through oxygen enrichment, provided it does not significantly reduce the bosh gas rate and TGT. An excessively low TGT may prolong the burden drying process, diminishing the BF’s effective height. It could also result in water condensation and wet dust particle accumulation during the initial phase of gas cleaning.³⁴⁾ Limited residence time of the injected NG and insufficient mixing with air blast in the raceway can lead to incomplete conversion of methane and other hydrocarbons into H₂ and CO. In the coke bed outside the high-oxygen area, unconverted hydrocarbons may pyrolyze into hydrogen and soot. Moderately endothermic, pyrolysis would increase the energy consumption in the BF process.⁶¹⁾ The produced soot can be entrained to the top gas, thereby reducing the CRR, and even entering the atmosphere from the scrubber water.⁶²⁾ It may also reside in the slag in the dripping zone worsening its flowability and participating in forming unstable accretion on the refractory.⁶³⁾ Therefore, achieving complete NG conversion through the optimal injection conditions, including blast temperature, oxygen enrichment, NG distribution around the tuyeres, and tuyere design, is of utmost importance.

In practice, the representative values of CRR are challenging to estimate due to the variability of baseline conditions and instability of operational parameters over long periods. Ramm⁴²⁾ determined replacement ratios of 0.78–1.1 kg/m³ based on the representative data from 8 Soviet steelworks with the injection of 62–169 m³/t-HM at the blast temperatures of 900–1169°C and 21–34.7% O₂. At the Egyptian BFs, Abdel Halim³¹⁾ reports CRRs of 0.72–0.75 kg/m³ at NG rates of 64–94 m³/t-HM without O₂ enrichment and 1003–1107°C blast temperature, and 0.8 kg/m³ at 93–103 m³/t-HM with 26% O₂ and 1091–1124°C. Based on literature analysis, Babich⁶⁸⁾ indicates CRR of 0.8–0.9 kg/m³.

Modeling data also varies widely. In the CFD model by Mauret et al.,²¹⁾ NG injection of 55 m³/t-HM under a constant PCI rate resulted in a CRR of 0.65 kg/m³. In their model, to stabilize the RAFT at 2147°, an increase in oxygen enrichment from 7% to 20% was applied, although, TGT dropped to 90°C indicating a challenging operation. CO utilization ratio (η_CO=vol.%CO₂/(vol.%CO₂+vol.%CO)) increased whereas that of H₂ (η_H2=vol.%H₂O/(vol.%H₂+vol.%H₂O)) decreased causing a decrease in the η_H2/η_CO ratio from 0.99 to 0.93 attributed by the authors to water gas shift reaction (WGSR). Noteworthy, Lan et al.⁶⁴⁾ based on an extensive literature review argued that the impact of hydrogenous gases’ injection on the extent of WGSR in the BF cannot be accurately simulated and predicted. Sato et al.³⁵⁾ also modeled NG injection under the constant PCI rate and obtained an equal CRR of 0.79 kg/m³ with the injection of 70 and 140 m³/t-HM. To stabilize RAFT at 2225°C at these injection rates, oxygen enrichment of 13% and 38% was applied, followed by a TGT drop from an initial 159°C to 131°C and 79°C, respectively. Pistorius et al.²⁹⁾ modeled an injection of 230 m³/t-HM while increasing blast oxygen content from atmospheric to 48.5% and obtained a CRR of 0.81 kg/m³, although RAFT dropped from 2134°C to 1711°C. In our opinion, the conditions specified for NG injection in ref.,^21,29,35) constitute rather extreme operating regimes, particularly in terms of oxygen enrichment, and do not align with industrial practices^31,42,48) or modeling by Peacey and Davenport³⁶⁾ (Table 1).

Based on our modeling, NG injection increases W_g predominantly owing to an increase in the bosh gas rate and decreases W_m mostly thanks to a decrease in direct reduction degree. As a result, the condition W_m=W_g (corresponds to Δt_min) is reached at eventually higher temperatures (Fig. 2). Oxygen enrichment at a given NG injection rate reduces the bosh gas rate leading to a decrease in W_g and shifts the W_g=W_m region upward. Raising the blast temperature lowers the coke rate, which decreases the bosh gas rate and direct reduction degree, also shifting the W_m=W_g region upwards. The grayed areas in Fig. 2 correspond to the pinch point temperature range of 900–1050°C, representing feasible operating margins. The conditions outside them would lead to deviation from modeled operation parameters: for example, excessive NG injection may cause incomplete conversion, while the poor radial distribution of the gas can decrease the gas utilization ratio and increase the direct reduction degree. Such negative impacts, altering W_g and W_m, cannot be assessed by our model.

Fig. 2. Temperatures of gas and solid at the pinch point, RAFT, TGT vs NG injection rate for blast temperatures of 1000°C, 1200°C and 21–25% O₂. Solid, dashed, dotted lines - temperatures of solid, gas, Δt_min. Gray areas correspond to the 900–1050°C temperature range of the solid.

An increase in NG injection at a given oxygen enrichment reduces Δt_min (Fig. 2). Adjustment of the injection conditions to a reasonable pinch point temperature range ensures a relatively stable RAFT. Oxygen enrichment and raising the blast temperature lower the bosh gas rate and decrease TGT, which remains within reasonable limits. At a blast temperature of 1000°C, the modeling indicates the optimal NG injection rates of 0–15 m³/t-HM and 70–85 m³/t-HM at 21% and 27% O₂, respectively. An increase in the blast temperature to 1200°C shifts these ranges to 30–50 m³/t-HM and 92–112 m³/t-HM at 21% and 27%O₂, respectively (Fig. 3). Such injection rates represent a relatively conservative estimate, consistent with industrial data in Table 1.

Fig. 3. Optimal ranges of NG injection rate and corresponding CRR at blast temperatures of 1000°C and 1200°C.

At a blast temperature of 1000°C, the CRR decreases from 0.96 to 0.83 kg/m³ while the oxygen content in the blast increases from 21% to 27%. At 1200°C, the decrease is observed from 0.93 to 0.77 kg/m³ (Fig. 3). This is consistent with industrial practices (Table 1) and is in line with Ramm’s⁴²⁾ estimation, who argued that the theoretical value of 0.9–1.1 kg/m³ for the atmospheric blast decreases to around 0.7 kg/m³ with O₂ enrichment.

In our model, direct CO₂ emissions reductions of NG injection range from 0.07% to 0.10%/m³ and depend more on oxygen enrichment than on blast temperature. Among the sources reviewed in Table 1, this parameter was assessed only by Mauret et al.,²¹⁾ although the value of 0.02%/m³ seems to be underestimated, which aligns with the low CRR discussed above.

3.2. Coke Oven Gas Injection

COG composition (Table 2) depends on the coal type and the operational parameters of the coke oven.⁶⁵⁾ At integrated steelworks, the COG amount comprises 410–560 m³/t-coke, depending on the coal’s volatile matter. Upon purification from contaminants such as tar, benzol hydrocarbons, naphthalene, and hydrogen sulfide, COG serves as a high-calorific supplement to BF gas in hot stoves, reheating furnaces, and power generation facilities. When available in adequate volumes, COG can be injected into the BF tuyeres, a technique widely adopted since the 1960s.^42,66,67) The high hydrogen content of COG enhances its attractiveness for reducing technologies.

Several studies suggest that the CRR of COG ranges within 0.40–0.48 kg/m³.^42,48,52,68) Compared to NG, COG contains a higher proportion of hydrogen and fewer hydrocarbons, facilitating its conversion in the raceway.⁶⁸⁾ The US Environmental Protection Agency stipulates a maximum COG injection rate, governed by the thermochemical conditions within the BF, at 219 m³/t-HM.⁴⁸⁾ Ramm⁴²⁾ contends that the available COG for tuyere injection in integrated steelworks typically does not surpass 100 m³/t-HM, thus maintaining RAFT within permissible ranges without the necessity for blast oxygen enrichment.

The available industrial data on COG injection efficiency varies widely. In the mid-1980s, at the Solmer plant in France, injection of up to 35 m³/t-HM was coupled with a decrease in steam injection and an increase in blast temperature. This resulted in a CRR of 0.73 kg/m³. After stabilizing the two latter parameters, injection of over 35 m³/t-HM was followed by CRR of 0.41 kg/m³ with a beneficial effect on the BF operating conditions, especially on the burden descending.⁶⁹⁾ In Ukraine, at Zaporizhstal in the 1970s, with the injection of 80.2 m³/t-HM at a blast temperature of 1125°C, a CRR of 0.49 kg/m³ was achieved without oxygen enrichment.⁴²⁾ Recent industrial data shows some lower replacement ratios. Worldsteel informs 0.3 kg/m³ at ArcelorMittal Gijón in Spain where COG injection has been implemented since 2021.⁵⁰⁾ Paul Wurth, a key technology supplier, reports a similar value of 0.35 kg/m³ without specifying the plant.⁵¹⁾ At the ROGESA plant in Germany, COG injection has been implemented on an industrial scale since 2022 and co-injection of around 35 m³/t-HM at a constant PCI rate of around 200 kg/t-HM without increasing the oxygen enrichment (the level of enrichment is not informed) resulted in an average CRR of 0.36 kg/m³.⁷⁰⁾ For the conditions of ROGESA steelworks, Mauret et al.²¹⁾ modeled an increase in COG injection rate from 32 to 162 m³/t-HM while keeping the PCI rate constant and increasing oxygen enrichment from 7% to 20%. They obtained a CRR of 0.24 kg/m³, substantially lower than practically achieved. The modeled TGT dropped to 88°C, indicating a challenging technological regime with potentially excessive oxygen enrichment. In the modeling, an increase in COG injection rate was followed by a decrease in η_H2 and an increase in η_CO resulting in a decline in η_H2/η_CO ratio from 0.99 to 93 attributed by the authors to the impact of WGSR which contradicts the trend revealed at an experimental BF in Japan discussed below.

On an experimental scale, at the 8 m³ BF of the LKAB plant in Lulea, Sweden, Watakabe et al.⁷¹⁾ carried out hydrogenous gases injection trials in the frames of the COURSE50 project. COG injection increased the proportion of H₂ reduction and decreased the direct reduction degree, although η_CO decreased. COG partially substituted PCI, however, at the injection of 217 m³/t-HM, quite a low value of the sum of solid reductant replacement ratio (coke and coal) of 0.12 kg/m³ was achieved. The injection of RCOG (reformed COG with amplified H₂ content) into the lower shaft was also tested revealing an insufficient penetration depth of the injected gases resulting in a low H₂ utilization ratio. These studies were continued at a 12 m³ experimental BF at Nippon Steel Kimitsu in Japan, where the injection of 93 m³/t-HM at a constant PCI rate and O₂ enrichment of 15.2% (7.9% in the base case) resulted in a CRR of 0.49 kg/m³.⁷²⁾ COG injection increased the η_H2/η_CO ratio from 0.95 to 1.01 while the direct reduction degree was reduced and the utilization ratios of CO and H₂ increased, confirming the results of the study by Mousa et al.⁷³⁾ on the impact of COG injection on the reduction processes. The high oxygen enrichment resulted in a decrease in the bosh gas rate, leading to lower temperatures in the lumpy zone. This was accompanied by an upward shift in position and thinning of the cohesive zone owing to an interplay between the W_g and W_m following a decrease in direct reduction degree. TGT decreased from 135°C to 101°C, indicating that oxygen enrichment might have been excessive. The three-level horizontal probe showed that most of the hydrogen was consumed in the lower shaft, while above the mid-shaft, the hydrogen concentration in the gas phase changed insignificantly. Noteworthy, experimental data from the horizontal probing in radial and vertical dimensions representatively confirmed the results obtained using the model developed by K. Takatani et al.²⁵⁾

An increase in the H₂ utilization ratio with COG injection was also predicted in the multi-fluid model by Long et al.³⁸⁾ who estimated the PC replacement ratio of around 0.2 kg/m³ regardless of the COG injection rate in the range of 59–152 m³/t-HM. They also experimentally studied the softening-melting behavior of the mixed (sinter and pellets) burden under the gas atmospheres relevant to various COG injection rates. COG injection increased the reduction degree, resulting in a substantial increase in the softening and a slight increase in dripping temperatures, followed by a narrowing of the softening-melting temperature interval, indicating a positive impact on the permeability of the cohesive zone.

Kou et al.⁵²⁾ modeled COG injection while keeping the injected gas (air plus COG) volume constant. Injection of 41–203 m³/t-HM was coupled with an incremental increase in oxygen enrichment by 0.55% per each 3% of COG in the injected volume showing a CRR of 0.4 kg/m³ irrespectively of the injection rate.

Li et al.⁵⁴⁾ modeled injection of 35–173 m³/t-HM in three scenarios with various top burden radial distribution options. The lowest CRR of 0.13–0.22 kg/m³, attributed to an increase in wall heat losses, occurred when more coke was charged to the peripheral region, whereas the highest CRR of 0.27–0.47 kg/m³ corresponded to an elevated share of coke in the intermediary region. Higher COG injection rates were followed by lower CRR, which the authors attributed to the increased gas-solid relative velocity deteriorating the interphase heat transfer. Modeling also showed that an increase in COG injection rate has led to a higher gas-solid temperature difference, shifting the cohesive zone upwards, and shortening the thermal reserve zone.

In our modeling, similar to NG, COG injection increases the bosh gas rate and decreases the direct reduction degree causing the W_m=W_g condition to occur at higher temperatures while Δt_min becomes smaller (Fig. 4). The reviewed CRR values for COG injection (Table 1), with a few exceptions, comply with the range of 0.36–0.48 kg/m³ obtained in our model for various blast temperatures and oxygen enrichment levels (Fig. 5), being approximately two times lower than those for NG, which is consistent with the difference in heating values. An increase in oxygen enrichment decreases CRR.

Fig. 4. Temperatures of gas and solid at the pinch point, RAFT, TGT versus COG injection rate for blast temperatures of 1000°C, 1200°C and 21–25% O₂. Solid, dashed, dotted lines - temperatures of solid, gas, Δt_min. Gray areas correspond to 900–1050°C temperature range of the solid.

Fig. 5. Optimal ranges of COG injection rates and corresponding CRR at blast temperatures of 1000°C and 1200°C.

For a blast temperature of 1000°C, the optimal COG injection rates (grayed areas in Fig. 4) range within 0–30 m³/t-HM without oxygen enrichment and 120–150 m³/t-HM for 27%O₂. An increase in the blast temperature to 1200°C shifts these ranges to 55–85 m³/t-HM and 170–200 m³/t-HM, respectively (Fig. 5). The direct CO₂ emissions reductions of COG injection under the analyzed conditions are approximately two times lower than for NG and range within 0.04–0.06 %CO₂/m³ (lower values correspond to higher oxygen enrichment), complying with the value obtained by Mauret et al.,²¹⁾ but substantially lower than in a study by Long et al.³⁸⁾ (Table 1).

3.3. Hydrogen Injection

Hydrogen tuyere injection in ironmaking, being considered for CO₂ reduction, is subject to a noted mismatch by the IEA: current hydrogen project capacities fall significantly short of potential demand.¹³⁾ Unlike NG and COG, hydrogen’s economic viability lags. Yet, economies of scale in hydrogen production could shrink the cost disparity, potentially reducing the price gap to near 10% by 2030 between conventional and green hydrogen-based steel production. Climate policies might further bolster hydrogen’s appeal, contingent on resolving infrastructure and safety challenges.^3,74)

Technological concerns of hydrogen injection have been noted by researchers. Zhuo et al.¹⁶⁾ show that hydrogen substitution for PCI expands the raceway volume approximately fourfold, negatively impacting stable BF operations. Additionally, hydrogen injection alters the radial heat patterns in the hearth, leading to overheating at the periphery due to the exothermic reaction:

  

$${\mathrm{H}}_2+1/2{\mathrm{O}}_2\leftrightarrow {\mathrm{H}}_2\mathrm{O};\mathrm{\Delta }\mathrm{H}°=-241.8\mathrm{k}\mathrm{J}/\mathrm{m}\mathrm{o}\mathrm{l}$$(1)

and cooling towards the center as a result of the endothermic reaction:

  

$${\mathrm{H}}_2\mathrm{O}+\mathrm{C}\leftrightarrow {\mathrm{H}}_2+\mathrm{C}\mathrm{O};\mathrm{\Delta }\mathrm{H}°=131.3\mathrm{k}\mathrm{J}/\mathrm{m}\mathrm{o}\mathrm{l}$$(2)

which may enlarge the deadman and undermine the refractory’s stability.

Nevertheless, the interest in hydrogen injection is increasing, and several companies have already carried out industrial trials. In 2019, Thyssenkrupp Steel in Duisburg tested hydrogen injection via a single tuyere⁷⁵⁾ without subsequent scale-up. Tata Steel Jamshedpur in India conducted a 5-day hydrogen injection trial in 2023, with an expected reduction of coke consumption by 10% and carbon emissions by 7–10%,⁷⁶⁾ but the outcomes have not been reported. In 2024, Cleveland-Cliffs Inc. conducted an injection at the largest BF in North America, following the successful trial at Middletown Works in 2023, although specific results have not been disclosed.⁷⁷⁾

On an experimental scale, hydrogen injection was tested in Japan at a 12 m³ blast furnace in the frames of the COURSE50 project.¹⁵⁾ The injection of 159 m³/t-HM did not substantially change the vertical temperature profile in the BF. However, injecting 277 m³/t-HM decreased the temperature in the thermal reserve zone from 1072°C without H₂ injection to 982°C. When the injection rate was further increased to 359 m³/t-HM, the temperatures in the shaft significantly dropped, reaching 65°C below the charge level. This was likely due to the high oxygen enrichment (18.5% versus 9% for 159 m³/t-HM) excessively reducing the bosh gas rate. Co-injection of COG and H₂ did not significantly change the vertical temperature pattern compared to when only hydrogen was injected. Important aspects in the development of WGSR have been revealed. In the absence of hydrogen injection, the ratio of partial pressures (P_H2·P_CO2)/(P_H2O·P_CO) corresponded well with the WGSR equilibrium, in line with vertical probing data from an industrial blast furnace in the 1970s.⁷⁸⁾ However, introducing hydrogen lowered this ratio beneath the equilibrium value, illustrating the complexities involved in modeling hydrogenous gas injection and its effects on WGSR dynamics. At a maximum injection rate of 359 m³/t-HM carbon consumption fell by about 17% from baseline level (the CRRs have not been reported). Despite challenging technological conditions, at this injection rate, determined as the limit for stable operation, a 16% decrease in CO₂ emissions was sustained. The CO₂ emissions were reduced by 16% in agreement with the earlier developed 3D unsteady model.²⁵⁾ Subsequent trials involving the injection of preheated hydrogen yielded a 33% reduction in CO₂ emissions.⁷⁹⁾

Hydrogen injection has been extensively modeled during the past decade. CFD modeling of intensive hydrogen injection by Nogami et al.²³⁾ at constant RAFT and bosh gas flow rate showed a significant downward shift of 200–600°C isotherms along with a reduction in pressure drop. They also found that more than half of H₂O formed by the reduction regenerates to H₂ through the WGSR. However, Li et al.²⁰⁾ noted a rather high H₂ utilization ratio of 70% in their study without hydrogen injection, suggesting a potential modeling error. In addition, in our opinion, stabilizing the bosh gas rate in this and some other studies^38,39) might not necessarily ensure an optimal injection regime. A proper combination of hydrogen injection and oxygen enrichment can ensure that, while the bosh gas rate decreases, its reducing potential increases causing a decrease in the direct reduction degree which, along with the reduced coke rate, will maintain an optimal vertical pattern of heat exchange.

In the 2D model by Mauret et al.,²¹⁾ hydrogen injection of 133 m³/t-HM decreased utilization rates of H₂ and CO, while their ratio decreased from 0.99 to 0.96. A high-temperature region in the hearth appeared more localized in the periphery. Even though the RAFT was maintained within reasonable limits (2147°C versus 2197°C in the base case), TGT dropped to 90°C from the initial 124°C, indicating an excessive oxygen enrichment (15% versus 7% in the base case). Castro et al.²⁴⁾ modeled a variety of injection regimes at two levels of PCI rate (220 and 250 kg/t-HM) and in some cases (not included in Table 1) obtained quite low CRRs, e.g. 0.03 kg/m³ for 94 m³-H₂/t-HM which also might be explained by an excessive for this injection rate oxygen enrichment of 13%.

Several authors have aimed to define a theoretical maximum hydrogen injection rate. Li et al.²⁰⁾ modeled injection regimes with up to 49.5% blast hydrogen enrichment (corresponding to around 540 m³/t-HM) and argued that additional coke combustion is needed to offset the reduced heat input with the blast. Barrett et al.³⁰⁾ defined 211 m³/t-HM as a limit stipulated by optimal ranges of TGT and RAFT. Okosun et al.,¹⁷⁾ based on CFD modeling, determined the injection rate of 325 m³/t-HM as a maximum owing to a cooling effect in the raceways, disruptive for stable operation. Noteworthy, both values are lower than 359 m³/t-HM determined at an experimental BF.¹⁵⁾

Tang et al.⁸⁰⁾ suggested that injecting hydrogen could improve the CO utilization ratio if the heat share of H₂ combustion remains below 25% of the total heat generated in the raceway. CRR of 0.39 kg/m³ was defined, substantially higher than in other studies; however, in their model, CO utilization ratio increases from the initial 53% to 62% at the injection of 120 m³/t-HM while H₂ utilization drops from 41% to 30%, and η_H2/η_CO ratios appear as low as 0.77 and 0.48 without and with H₂ injection, respectively, contradicting the practice and results of other researchers.

Yilmaz et al.³³⁾ modeled the replacement of PCI with hydrogen preheated to 1200° (equal to blast temperature) and defined the injection rate of 306 m³/t-HM at oxygen content in the blast of 21.1% as an optimal regime resulting in the replacement ratio (sum of coke and PCI) of 0.34 kg/m³. This value is substantially higher than predicted in other studies without hydrogen preheating. Thanks to the heated hydrogen, RAFT was maintained at 2150°C despite the nearly atmospheric blast oxygen content.

Several studies have modeled the hydrogen injection potential for reducing CO₂ emissions. Mauret et al.²¹⁾ found that injecting 133 m³/t-HM can decrease CO₂ emissions by 6%. Tang et al.³⁹⁾ estimated a 5.2% reduction in life cycle CO₂ emissions by injecting 120 m³/t-HM. In the study by Castro et al.,²⁴⁾ CO₂ emissions reductions range from 0.01 kg/m³ at injection of 94 m³/t-HM to 0.28 kg/m³ at 9 m³/t-HM; however, the obtained values do not correlate with CRR and their model applies a rather high slag basicity of up to 1.47 (CaO/SiO₂), not feasible in BF practice. Babich⁶⁸⁾ emphasized the need for continued research to determine the potential of H₂ injection in reducing CO₂ emissions.

With the injection of preheated hydrogen, Yilmaz et al.³³⁾ and Ye et al.⁸¹⁾ predicted a CO₂ emissions reduction ratio of 0.07%/m³. In addition, Ye et al. found that injecting 350 m³/t-HM of hydrogen preheated to 1250°C reduced CO₂ emissions by 26.1%, whereas injecting less than 300 m³/t-HM led to higher CO₂ emissions from hydrogen preheating compared to the emissions reduction in the BF process. Noteworthy, in both studies, the CO₂ emissions reduction was lower than the practically achieved 33% at an experimental BF with preheated hydrogen.⁷⁹⁾

Several studies have examined the injection of hydrogen into the lower shaft.^18,19,20) However, CFD modeling by Li et al.²⁰⁾ revealed that the hydrogen flow injected into the shaft is mainly confined within the peripheral region, increasing heat loss through the BF walls. Even increased coke charging into the peripheral region did not significantly improve the penetration depth. This is consistent with the discussed earlier results of RCOG shaft injection at an experimental BF.⁷¹⁾

The impact of injecting hydrogen on the temperatures of solid and gas phases at the pinch point in our model (Fig. 6) is similar to that discussed earlier for NG and COG. CRR ranges from 0.19 to 0.29 kg/m³, depending on the injection rate, blast temperature, and oxygen enrichment (Fig. 7). This is consistent with most studies modeling the injection of cold H₂ (Table 1). A difference between the minimum temperatures of gas and solid increases in the sequence NG → COG → H₂, in line with the impacts of an increased bosh gas rate and decreased direct reduction degree on W_g and W_m. RAFT decreases and TGT increases in the same sequence. Within the optimal injection rates, RAFT varied from 2137°C to 2230°C depending upon the injection rate, injectant used, oxygen enrichment, and blast temperature.

Fig. 6. Temperatures of gas and solid at the pinch point, RAFT, TGT versus H₂ injection rate for blast temperatures of 1000°C, 1200°C and 21–25% O₂. Solid, dashed, dotted lines - temperatures of solid, gas, Δt_min. Gray areas correspond to 900–1050°C temperature range of the solid.

Fig. 7. Optimal ranges of H₂ injection rates and corresponding CRR at blast temperatures of 1000°C and 1200°C.

Our model predicts CO₂ emission reductions for hydrogen injection of around 0.06%/m³, aligning with the data in Table 1. NG injection shows the highest CRR and CO₂ emissions reduction. Hydrogen has the lowest CRR while its CO₂ emissions reduction potential is greater than that of COG because the higher CRR of the latter is offset by a significant amount of carbonaceous gases. At a blast temperature of 1000°C, the optimal hydrogen injection rates are 0–40 m³/t-HM without oxygen enrichment and 180–235 m³/t-HM at 27% O₂. An increase in the blast temperature to 1200°C shifts these ranges to 82–145 m³/t-HM and 260–320 m³/t-HM, respectively (Fig. 7). The maximum values generally align with the data in Table 1, although, according to our model, the optimal operation conditions with elevated hydrogen injection rate can be achieved while the oxygen content is kept within 30%. Oxygen enrichment increases the optimal range of the injection rates by eventually smaller increments in all three cases (NG, COG, and H₂ injection).

Overall, the review indicates a significant divergence in the outcomes of computational models and industrial or experimental trials, complicating the scale-up from theoretical predictions and pilot trials to industrial operations. To bridge this gap, focused research should further advance the validity of modeling and comprehensively evaluate hydrogenous gases’ impact on BF operations in order to ensure accurate predictions of CRR and CO₂ reduction, key factors for techno-economic efficiency.

4. Conclusions

(1) Tuyere injection of hydrogen and hydrogenous gases, such as NG and COG, provides a viable way to reduce the carbon footprint of blast furnace ironmaking. Despite notable advancements in industrial applications and modeling, there is a considerable disparity in estimating the impacts of injection on the BF operation and determining the most effective injection regimes due to the lack of representative industrial trials and variations in modeling approaches and assumptions.

(2) While NG injection is a mature technology, its use is limited to a few regions due to availability and cost. Optimized injection conditions, including blast temperature, oxygen enrichment, tuyere design, etc. are crucial for achieving a high coke replacement.

(3) Nowadays COG injection sees renewed attention with several companies adopting this technology in the last few years even though the available resource for the BF injection is limited considering the role of COG in the energy mix of integrated steelworks. Practical experience and modeling have shown a positive impact of COG injection on the BF process.

(4) NG has the highest potential for reducing CO₂ emissions due to its superior coke replacement ratio (around 0.85 kg/m³). Although COG has a higher coke replacement ratio (around 0.4 kg/m³) compared to cold hydrogen (around 0.25 kg/m³) their CO₂ emissions reduction potentials are comparable (up to around 0.06%/m³) owing to the presence of carbonaceous gases in COG. Hydrogen preheating can enhance coke replacement and ensure the highest CO₂ emission reduction levels.

(5) The cost efficiency and positive technological impacts of injecting NG and COG have been proven, whereas hydrogen injection is not yet economically profitable and not deployed on a large scale. With technological advancements, scale-up, and carbon policies, this situation may change in the next decade, motivating steelmaking companies worldwide to work towards implementing hydrogen injection.

(6) Further research is essential for reliable predictions of the impacts of the hydrogenous gases’ injection on the blast furnace operation, coke replacement ratios, and CO₂ reduction potential, key factors for the techno-economic efficiency.

References

• 1)  IEA, Emissions Measurement and Data Collection for a Net Zero Steel Industry, (April 2023), https://www.iea.org/reports/emissions-measurement-and-data-collection-for-a-net-zero-steel-industry, (accessed 2024-09-25).
• 2)   S.  Durga,  S.  Speizer and  J.  Edmonds: Energy and Climate Change, 5 (2024), 100152. https://doi.org/10.1016/j.egycc.2024.100152
• 3)  European Parliamentary Research Service, Carbon-free steel production, (April, 2021), https://www.europarl.europa.eu/RegData/etudes/STUD/2021/690008/EPRS_STU(2021)690008_EN.pdf, (accessed 2024-09-25).
• 4)  SSAB, HYBRIT: A new revolutionary steelmaking technology, https://www.ssab.com/en/fossil-free-steel/insights/hybrit-a-new-revolutionary-steelmaking-technology, (accessed 2024-09-25).
• 5)   SMS  Group, H₂ Green Steel, https://www.sms-group.com/plants/projects/h2-green-steel, (accessed 2024-09-25).
• 6)   Stegra,  H₂ Green Steel and Iberdrola announce €2.3 billion Green hydrogen venture, (December 2, 2021), https://stegra.com/news-and-stories/h2-green-steel-and-iberdrola-announce-23-billion-green-hydrogen-venture, (accessed 2024-09-25).
• 7)  Salcos, µDRAL, https://salcos.salzgitter-ag.com/en/mydral.html, (accessed 2024-09-25).
• 8)  FCW, Chile’s CAP Develops Green Hydrogen Pilot Plant for Decarbonized Steel Production, (March 29, 2023), https://fuelcellsworks.com/news/chiles-cap-develops-green-hydrogen-pilot-plant-for-decarbonized-steel-production, (accessed 2024-09-25).
• 9)  AIST, LIBERTY Group: Exploring New Hydrogen-Based DRI Plant in France, (February 22, 2021), https://www.aist.org/liberty-group-exploring-new-hydrogen-based-dri-plant-in-france, (accessed 2024-09-25).
• 10)  IEEFA, Hydrogen unleashed: Opportunities and challenges in the evolving H2-DRI-EAF pathway beyond 2024, (February 27, 2024), https://ieefa.org/resources/hydrogen-unleashed-opportunities-and-challenges-evolving-h2-dri-eaf-pathway-beyond-2024, (accessed 2024-09-25).
• 11)  IEEFA, Solving Iron Ore Quality Issues for Low-Carbon Steel, (August 2022), https://ieefa.org/sites/default/files/2022-08/Solving%20Iron%20Ore%20Quality%20Issues%20for%20Low-Carbon%20Steel_0.pdf, (accessed 2024-09-25).
• 12)  Midrex, Outlook for DRI Production and Use, (March 2024), https://www.midrex.com/tech-article/outlook-for-dri-production-and-use/, (accessed 2024-09-25).
• 13)  IEA, Steel, https://www.iea.org/energy-system/industry/steel, (accessed 2024-09-25).
• 14)   L.  Holappa: Metals, 10 (2020), 1117. https://doi.org/10.3390/met10091117
• 15)   C.  Kamijo,  Y.  Matsukura,  H.  Yokoyama,  K.  Sunahara,  K.  Kakiuchi,  H.  Sakai,  K.  Nakano,  Y.  Ujisawa and  K.  Nishioka: ISIJ Int., 62 (2022), 2433. https://doi.org/10.2355/isijinternational.ISIJINT-2022-099
• 16)   Y. T.  Zhuo,  Z. J.  Hu and  Y. S.  Shen: Fuel, 302 (2021), 120804. https://doi.org/10.1016/j.fuel.2021.120804
• 17)   T.  Okosun,  S.  Nielson and  C.  Zhou: JOM, 74 (2022), 1521. https://doi.org/10.1007/s11837-022-05177-4
• 18)   X.  Yu,  Z.  Hu and  Y.  Shen: Fuel, 302 (2021), 121092, https://doi.org/10.1016/j.fuel.2021.121092
• 19)   Z.  Zhao,  X.  Yu,  Y.  Li,  J.  Zhu and  Y.  Shen: Fuel, 348 (2023), 128641. https://doi.org/10.1016/j.fuel.2023.128641
• 20)   J.  Li,  S.  Kuang,  L.  Jiao,  L.  Liu,  R.  Zou and  A.  Yu: Fuel, 323 (2022), 124368. https://doi.org/10.1016/j.fuel.2022.124368
• 21)   F.  Mauret,  M.  Baniasadi,  H.  Saxén,  A.  Feiterna and  S.  Hojda: Metall. Mater. Trans. B, 54 (2023), 2137. https://doi.org/10.1007/s11663-023-02822-4
• 22)   M.  Baniasadi,  F.  Mauret,  K.-P.  Kinzel,  P.  Bermes,  M.  Baniasadi,  C.  Castagnola,  H.  Saxén,  S.  Hojda and  R.  Lin: Proceedings of the 8th ECIC & 9th ICSTI Conferences, Bremen, Germany, (2022).
• 23)   H.  Nogami,  Y.  Kashiwaya and  D.  Yamada: ISIJ Int., 52 (2012), 1523. https://doi.org/10.2355/isijinternational.52.1523
• 24)   J. A.  de Castro,  C.  Takano and  J. I.  Yagi: J. Mater. Res. Technol., 6 (2017), 258. https://doi.org/10.1016/j.jmrt.2017.05.007
• 25)   K.  Takatani,  T.  Inada and  Y.  Ujisawa: ISIJ Int., 39 (1999), 15. https://doi.org/10.2355/isijinternational.39.15
• 26)   P.  Reichardt: Archiv für das Eisenhüttenwesen, 1 (1927), 77. https://doi.org/10.1002/srin.192700187
• 27)   A.  Rist and  G.  Bonnivard: Rev. Met. Paris, 60 (1963), 23. https://doi.org/10.1051/metal/196360010023
• 28)   B. I.  Kitaev,  Y. G.  Yaroshenko and  V. D.  Suchkov: Heat Exchange in Shaft Furnaces, Pergamon Press, Oxford, UK, (1967), 301.
• 29)   P. C.  Pistorius,  J.  Gibson and  M.  Jampani: Applications of Process Engineering Principles in Materials Processing, Energy and Environmental Technologies, ed. by S. Wang et al., The Minerals, Metals & Materials Series, Springer International Publishing, Cham, Switzerland, (2017), 283. https://doi.org/10.1007/978-3-319-51091-0_26
• 30)   N.  Barrett,  S.  Mitra,  H.  Doostmohammadi,  D.  O’dea,  P.  Zulli,  S.  Chew and  T.  Honeyands: ISIJ Int., 62 (2022), 1168. https://doi.org/10.2355/isijinternational.ISIJINT-2021-574
• 31)   K. S.  Abdel Halim: Materials Letters, 61 (2007), 3281. https://doi.org/10.1016/j.matlet.2006.11.053
• 32)   A.  Spanlang,  W.  Wukovits and  B.  Weiss: Berg Huettenmaenn Monatsh, 165 (2020), 243. https://doi.org/10.1007/s00501-020-00963-6
• 33)   C.  Yilmaz,  J.  Wendelstorf and  T.  Turek: J. Cleaner Prod., 488 (2017), 154. https://doi.org/10.1016/j.jclepro.2017.03.162
• 34)   M.  Geerdes,  R.  Chaigneau,  I.  Kurunov,  O.  Lingiardi and  J.  Ricketts: Modern Blast Furnace Ironmaking: an Introduction, IOS Press BV, Amsterdam, The Netherlands, (2015), 218.
• 35)   M.  Sato,  K.  Takahashi,  T.  Nouchi and  T.  Ariyama: ISIJ Int., 55 (2015), 2105. https://doi.org/10.2355/isijinternational.ISIJINT-2015-264
• 36)   J. G.  Peacey and  W. G.  Davenport: The Iron Blast Furnace: Theory and Practice, Pergamon Press, Oxford & New York, (1979), 251.
• 37)   M.  Naito,  A.  Okamoto,  K.  Yamaguchi,  T.  Yamaguchi and  Y.  Inoue: Nippon Steel Technical Report, 94 (2006), 103. https://www.nipponsteel.com/en/tech/report/nsc/pdf/n9417.pdf
• 38)   H. M.  Long,  H. T.  Wang,  W.  Zhao,  J. X.  Li,  Z. G.  Liu and  P.  Wang: Ironmaking & Steelmaking, 43 (2016), 450. https://doi.org/10.1080/03019233.2015.1108480
• 39)   J.  Tang,  M.  Chu,  F.  Li,  Z.  Zhang,  Y.  Tang,  Z.  Liu and  J.  Yagi: Journal of Cleaner Production, 278 (2012), 123191. https://doi.org/10.1016/j.jclepro.2020.123191
• 40)   V.  Shatokha: Int. J. Miner. Metall. Mater., 29 (2022), 1851. https://doi.org/10.1007/s12613-022-2474-8
• 41)   V.  Shatokha: Steel Research Int., (2023), 2300397. https://doi.org/10.1002/srin.202300397
• 42)   A. N.  Ramm: Modern Blast Furnace Process, Metallurgiya, Moscow, USSR, (1980), 303 (in Russian).
• 43)   Y.  Hashimoto,  Y.  Sawa,  Y.  Kitamura,  T.  Nishino and  M.  Kano: ISIJ Int., 58 (2018), 2210. https://doi.org/10.2355/isijinternational.ISIJINT-2018-177
• 44)   D. M.  Kundrat,  T.  Miwa and  A.  Rist: Metall. Trans. B, 22 (1991), 363. https://doi.org/10.1007/BF02651235
• 45)   E. T.  Turkdogan: Metall. Trans. B, 9 (1978), 163. https://doi.org/10.1007/BF02653682
• 46)   H.  Dembicki, Jr.: Practical Petroleum Geochemistry for Exploration and Production, Elsevier, Amsterdam, The Netherlands, (2017), 135. https://doi.org/10.1016/B978-0-12-803350-0.00004-0
• 47)   N. A.  Gladkov: Fundamental and Applied Problems in Ferrous Metallurgy, ISI NASU, Dnipro, Ukraine, (2009), 37 (in Russian).
• 48)  US EPA: Available and Rmerging Technologies for Reducing Greenhouse Gas Emissions from The Iron and Steel Industry, (2012), https://www.epa.gov/sites/default/files/2015-12/documents/ironsteel.pdf, (accessed 2024-09-25).
• 49)   M.  Jampani,  J.  Gibson and  P. C.  Pistorius: Metall. Mater. Trans. B, 50 (2019), 1290. https://doi.org/10.1007/s11663-019-01538-8
• 50)  Worldsteel, ArcelorMittal: CO₂ reduction by means of coke oven gas co-injection in blast furnace, (2022), https://worldsteel.org/case-studies/environment/arcelormittal-co2-reduction-by-means-of-coke-oven-gas-co-injection-in-bclast-furnace/, (accessed 2024-09-25).
• 51)   Paul  Wurth: Coke oven gas injection into blast furnace, (January 2020), https://www.paulwurth.com/wp-content/uploads/2020/08/Flyer-Coke-Oven-Gas-Injection-en.pdf, (accessed 2024-09-25).
• 52)   M.  Kou,  H.  Zhou,  L. P.  Wang,  Z.  Hong,  S.  Yao,  H.  Xu and  S.  Wu: Processes, 7 (2019), 528. https://doi.org/10.3390/pr7080528
• 53)   X.  Gao,  R.  Zhang,  Z.  You,  W.  Yu,  J.  Dang and  C.  Bai: Materials, 15 (2022), 6078. https://doi.org/10.3390/ma15176078
• 54)   Z.  Li,  Z.  Qi,  Z.  Wang,  L.  Zhang,  D.  Liang and  Q.  Dong: International Journal of Hydrogen Energy, 47 (2022), 31109. https://doi.org/10.1016/j.ijhydene.2022.07.036
• 55)   S.  Ye,  Z.  Xie,  Y.  Wang,  W.  Wu,  R.  Wang,  Y.  Zhang and  X.  Lu: Journal of Cleaner Production, 425 (2023), 139042. https://doi.org/10.1016/j.jclepro.2023.139042
• 56)   A.  Babich,  S.  Yaroshevskii,  A.  Formoso,  A.  Cores,  L.  Garcia and  V.  Nozdrachev: ISIJ Int., 39 (1999), 229. https://doi.org/10.2355/isijinternational.39.229
• 57)   J.  Stubbles: Energy Use in the U.S. Steel Industry: An Historical Perspective and Future Opportunities. U.S. Department of Energy (2000), https://www.energy.gov/eere/amo/articles/itp-steel-energy-use-us-steel-industry-historical-perspective-and-future, (accessed 2024-09-25).
• 58)   W.-K.  Lu and  M. G.  Ranade: ISIJ Int., 31 (1991), 395. https://doi.org/10.2355/isijinternational.31.395
• 59)   K.  Wang,  J.  Zhang,  H.  Zhou,  J.  Ni,  Y.  Hu,  J.  Huang,  S.  Li,  X.  Ma,  S.  Wu and  M.  Kou: Applied Thermal Engineering, 230 (2023), 120817. https://doi.org/10.1016/j.applthermaleng.2023.120817
• 60)   A.  Majeski,  A.  Runstedtler,  J.  D’alessio and  N.  Macfadyen: ISIJ Int., 55 (2015), 1377. https://doi.org/10.2355/isijinternational.55.1377
• 61)   S.  Nielson,  T.  Okosun,  B.  Damstedt,  M.  Jampani and  C. Q.  Zhou: Processes, 9 (2021), 1447. https://doi.org/10.3390/pr9081447
• 62)  US EPA: Pollution Effects of Abnormal Operations in Iron and Steel Making - Volume 111. Blast Furnace lronmaking, Manual of Practice, (1978), 93.
• 63)   A.  Dmitriev,  M.  Zolotykh,  Y.  Chesnokov,  O.  Ivanov,  K.  Chen and  G.  Vitkina: Proceedings of METEC & 2nd ESTAD 2015, European Steel Technology and Application Days Düsseldorf, Germany, Düsseldorf, (2015), 1.
• 64)   C.  Lan,  Y.  Hao,  J.  Shao,  S.  Zhang,  R.  Liu and  Q.  Lyu: Metals, 12 (2022), 1864. https://doi.org/10.3390/met12111864
• 65)   C.  Banerjee,  S.  Agarwal,  P. S.  Dash and  A.  Roy: Thermal Science and Engineering Progress, 25 (2021), 101025. https://doi.org/10.1016/j.tsep.2021.101025
• 66)   N.  Takamatsu,  K.  Kurihara,  A.  Hatanaka,  G.  Saitoh and  H.  Kaku: Nippon Steel Technical Report, 101 (2012), 79. https://www.nipponsteel.com/en/tech/report/nsc/pdf/NSTR101-12_tech_review-2-1.pdf
• 67)   W.  Liu,  H.  Zuo,  J.  Wang,  Q.  Xue,  B.  Ren and  F.  Yang: International Journal of Hydrogen Energy, 46 (2021), 10548. https://doi.org/10.1016/j.ijhydene.2020.12.123
• 68)   A.  Babich: Ironmak. Steelmak., 48 (2021), 728. https://doi.org/10.1080/03019233.2021.1900037
• 69)   N.  Jusseau,  A.  Dufour,  B.  Metz and  P.  Duperray: Rev. Met. Paris, 83 (1986), 367. https://doi.org/10.1051/metal/198683050367
• 70)   R.  Lin,  A.  Feiterna,  D.  Senk,  A.  Babich and  J.-P.  Simoes: Proceedings of 8th ECIC & 9th ICSTI Conferences, Bremen, Germany, (2022).
• 71)   S.  Watakabe,  K.  Miyagawa,  S.  Matsuzaki,  T.  Inada,  Y.  Tomita,  K.  Saito,  M.  Osame,  P.  Sikström,  L. S.  Ökvist and  J.-O.  Wikstrom: ISIJ Int., 53 (2013), 2065. https://doi.org/10.2355/isijinternational.53.2065
• 72)   K.  Nakano,  H.  Sakai,  Y.  Ujisawa,  K.  Kakiuchi,  K.  Nishioka,  K.  Sunahara,  Y.  Matsukura and  H.  Yokoyama: ISIJ Int., 62 (2022), 2424. https://doi.org/10.2355/isijinternational.ISIJINT-2022-117
• 73)   E. A.  Mousa,  A.  Babich and  D.  Senk: ISIJ Int., 53 (2013), 1372. https://doi.org/10.2355/isijinternational.53.1372
• 74)  European Parliament, The potential of hydrogen for decarbonising steel production, (December 14, 2020), https://www.europarl.europa.eu/RegData/etudes/BRIE/2020/641552/EPRS_BRI(2020)641552_EN.pdf, (accessed 2024-09-25).
• 75)  Thyssenkrupp, Injection of hydrogen into blast furnace, (February 3, 2021), https://www.thyssenkrupp-steel.com/en/newsroom/press-releases/thyssenkrupp-steel-concludes-first-test-phase-successfully.html, (accessed 2024-09-25).
• 76)  Reuters, India’s Tata Steel begins hydrogen gas injection trial in blast furnace, (April 24, 2023), https://www.reuters.com/business/sustainable-business/indias-tata-steel-begins-hydrogen-gas-injection-trial-blast-furnace-2023-04-24/, (accessed 2024-09-25).
• 77)  Cliffs, Cleveland-Cliffs Completes Successful Blast Furnace Hydrogen Injection Trial at Indiana Harbor #7 Blast Furnace, (January 26, 2024), https://www.clevelandcliffs.com/news/news-releases/detail/620/cleveland-cliffs-completes-successful-blast-furnace, (accessed 2024-09-25).
• 78)   H.  Itaya,  T.  Fukutake,  K.  Okabe and  T.  Nagai: Tetsu-to-Hagané, 62 (1976), 472 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.62.5_472
• 79)   Nippon  Steel, The world’s highest reduction in blast furnace CO₂ emissions of 33% was confirmed by injecting heated hydrogen into the Super COURSE50 test furnace for blast furnace hydrogen reduction technology, (February 6, 2024) https://www.nipponsteel.com/news/20240206_100.html (in Japanese), (accessed 2024-09-25).
• 80)   Z.  Tang,  Z.  Zheng,  H.  Chen and  K.  He: Energy Technology, ed. by A. A. Baba et al., The Minerals, Metals & Materials Series, Springer International Publishing, Cham, (2021), 149. https://doi.org/10.1007/978-3-030-65257-9_14
• 81)   S.  Ye,  Z.  Xie,  Y.  Wang,  W.  Wu,  R.  Wang,  Y.  Zhang and  X.  Lu: Journal of Cleaner Production, 425 (2023), 139042. https://doi.org/10.1016/j.jclepro.2023.139042