1. Background
The decarbonization of the steel industry under carbon neutrality goals represents a complex multi-ledger process that demands systematic trade-offs among technological pathways, resource consumption, and system benefits. Yet the reality of transformation proves far more daunting than the ideals often envisioned. According to the World Steel Association’s Sustainability Indicators 2023, global crude steel production reached 1.885 billion tonnes in 2022, while direct CO
2 emissions from the sector amounted to about 3.6 billion tonnes. This corresponds to an average emission intensity of 1.91 tonnes of CO
2 (tCO
2) per tonne of crude steel (tCO
2·(t crude steel)
−1), essentially unchanged from 1.88 tCO
2·(t crude steel)
−1 in 2020. Combined with historical output, cumulative emissions from the steel sector now exceed 150 billion tCO
2, underscoring the industry’s long-term role as one of the largest industrial sources of greenhouse gases [
1]. On average, producing 1 t of steel generates approximately 1.8-2.0 tCO
2, with the steel industry currently contributing about 24% of global industrial sector CO
2 emissions [
2], equivalent to the total exhaust emissions from global passenger vehicles [
3]. Confronted with such massive emission baselines, achieving the Paris Agreement’s 1.5 °C target and national carbon neutrality commitments requires deep structural changes in heavy-emitting industries like steel—changes that are proving significantly harder to realize than anticipated.
China’s role is pivotal in the global effort against climate change. As the world’s largest steel producer, the success of its steel industry’s green transition is crucial for meeting international climate goals. The average carbon intensity of crude steel production in China has been estimated at around 2.33 (tCO
2·(t crude steel)
−1) [
4]. Understanding both the scale and spatial distribution of these emissions is therefore crucial for evaluating feasible pathways. Recent data from the Multi-resolution Emission Inventory for China (MEIC) reveal several defining patterns [
5]. First, national CO
2 emissions rose from about 8.2 Gt in 2010 to 11.2 Gt in 2023, with the industrial and power sectors consistently dominating the total (
Fig. 1(a)). Second, emissions are geographically concentrated: In 2023, Shandong, Inner Mongolia, Jiangsu, Hebei, and Guangdong together contributed more than one-third of the national total, while the top ten regions exceeded half (
Fig. 1(b)).
Third, provincial industrial and power emissions are strongly correlated (Pearson correlation
r ≈ 0.685), forming clusters such as Hebei-Shandong-Jiangsu and Inner Mongolia-Shanxi, where steel production and electricity generation are co-located (
Fig. 1(c)). These clusters underscore both the entrenched difficulties of decarbonizing major hubs and the opportunities for integrated retrofits. Fourth, emission trajectories over the past decade highlight persistence at the top: Hebei, Shandong, and Jiangsu show consistently high industrial emissions with relatively narrow year-to-year variation, while Inner Mongolia and Liaoning display wider dispersion, reflecting shifts in industrial activity and energy supply (
Fig. 1(d)). Such contrasts indicate that some provinces are locked into enduringly high baselines, whereas others remain more volatile—differences that will shape where and how interventions can gain traction.
Taken together, these empirical insights refine how technology choices should be framed. Carbon capture, utilization, and storage (CCUS) offers scale and compatibility with existing blast furnace (BF) capacity but imposes energy penalties and water burdens. Hydrogen-based direct reduced iron (H2-DRI) promises near-zero process emissions if supported by abundant clean power and hydrogen, yet faces severe resource mismatches and regional constraints. At the same time, demand-side measures—such as greater scrap recycling and improvements in material efficiency—can reduce overall pressure on both production and infrastructure.
In sum, steel decarbonization must be understood not as an abstract technological substitution but against a concrete empirical baseline: a system in which national totals continue to rise, industrial and power emissions are tightly coupled, and a handful of provinces exert disproportionate influence. Recognizing this baseline is essential for assessing technology pathways and clarifies why policy design, resource allocation, and sequencing are as decisive as the technologies themselves.
2. Macro context: Severe transformation constraints
China, as the world’s largest steel producer accounting for approximately 56% of global output, sees its steel industry emissions represent 14%-16% of China’s total carbon emissions [
6]. Building on the MEIC-based provincial patterns highlighted above, it becomes clear that any transformation pathway must contend not only with the national aggregate but also with entrenched spatial clusters of emissions. Recent research indicates that even under optimistic scenarios with rapid national emission reductions, achieving carbon neutrality targets would require 45%-62% electrification rates, 47%-78% renewable energy share in primary energy supply, 5.2-7.9 TW of solar and wind power installations, and 64-1649 MtCO
2 of negative emission technology applications [
7]. These figures reflect the enormous difficulties of transformation—even the most optimistic scenarios require unprecedented technology deployment speeds and investment scales. Within this macro backdrop, the steel industry’s decarbonization choices are further constrained by the concentration of production in provinces such as Hebei, Shandong, and Jiangsu, which MEIC data show as persistent high emitters. This creates regional asymmetries where decarbonization measures must be tailored to localized resource conditions.
Current steel emission reduction technology pathways mainly concentrate in two directions: first, carbon capture and storage (CCS), which significantly reduces emissions by retrofitting existing BF-converter processes with carbon capture equipment; second, H2-DRI, which uses green hydrogen to reduce iron ore and combines with electric arc furnace steelmaking to avoid fossil carbon use from the source. These two technologies represent fundamentally different emission reduction philosophies: the former focuses on leveraging existing capacity for large-scale emission reduction, while the latter pursues near-zero emissions through process innovation. However, as the MEIC data indicate, these choices cannot be made in isolation from the provincial emission baselines and industrial-power correlations. Deeper analysis reveals that both face structural barriers that are difficult to overcome, representing not simple technology selection problems but complex decisions requiring comprehensive consideration across different stages and regions. Worryingly, even when both pathways are combined, the persistence of top-emitting provinces and the rising national baseline (2010-2023) suggest that meeting urgent emission reduction time windows may still prove impossible.
As shown in
Fig. 2(a) [
6], [
7], [
8], CCS consistently provides the largest mitigation wedge across the high, medium, and low steel-output scenarios in China's steel-sector mitigation pathway. Following Wang et al. [
6], these scenarios are defined by crude-steel output decreasing from 1053 Mt in 2020 to 746.5 Mt (high), 671.8 Mt (medium), and 578.6 Mt (low) by 2060 (−29.1%, −36.2%, and −45.1%). CCS correspondingly delivers 609.5, 536.6, and 472.4 MtCO
2·a
−1 of mitigation—equivalent to ∼31.4%-40.5% of the sector’s emissions absent measures. This persistent reliance on CCS in China’s scenarios mirrors global decarbonization pathways, underscoring a structural predicament: insufficient diversity in scalable near-term options. In line with pathways consistent with limiting global mean temperature rise to 2.0 or 1.5 °C by 2100 (e.g., Intergovernmental Panel on Climate Change (IPCC)/International Energy Agency (IEA)), scenario analyses indicate that both CCUS and hydrogen are indispensable for the sector [
9]. Globally, approximately 40% of steel emission reductions rely on CCUS achievement [
10], while hydrogen‐based ironmaking is expected to contribute over one-third of global primary steel production by 2050 [
7]. However, these projections often assume optimistic technology maturity and deployment speeds, downplaying technical bottlenecks, economic constraints, and social acceptance challenges.
3. CCS pathway: Resource burden behind scale advantages
Although CCS offers large-scale retrofitting potential, its feasibility must be assessed against the backdrop of clustered industrial-power hubs identified in MEIC, particularly in Hebei, Shandong, and Inner Mongolia. Europe’s green steel initiatives, while achieving technical breakthroughs (Sweden’s HYBRIT project produced the world’s first hydrogen-reduced fossil-free steel in 2021 [
3]), remain at demonstration stages with enormous gaps to commercial-scale applications. The United States, despite investing in both CCS and hydrogen metallurgy technology routes, announced 1.5 billion USD in funding for low-carbon ironmaking projects in March 2024, with the Advanced Research Projects Agency-Energy (ARPA-E) adding 28 million USD for more advanced research, but such investment scales are clearly insufficient relative to entire industry transformation needs [
11]. For China—with its extensive BF base—a pragmatic pathway is to deploy CCS to deliver rapid, near-term abatement while piloting hydrogen metallurgy in renewable-rich regions to build medium-to-long-term options, all while facing comparable resource and technical constraints [
3,
6]. Against this backdrop, the choice among specific CCS configurations becomes a question of integration risk, energy penalty, and site suitability.
Among CCS options for the steel sector, post-combustion amine scrubbing on BF-basic oxygen furnace (BOF) off-gases is the most mature (technology readiness level (TRL) ∼8-9), already piloted at integrated mills. Typical capture costs for steel flue gases are ∼40-70 USD·(tCO
2)
−1 under favorable heat-integration, driven mainly by solvent regeneration energy, solvent makeup, and compression [
12]. By contrast, top gas recycling (TGR)-BF with CO
2 removal benefits from higher CO
2 partial pressures and lower capture energy but remains at pilot-demo stages (TRL ∼6-8) with site-specific gas cleaning and compression needs. Oxy-fuel variants (oxy-BF/oxy-converter) simplify separation by producing CO
2-rich flue gas, yet face oxygen-supply and materials challenges and are earlier in maturity (TRL ∼5-7) [
13]. Beyond capture, transport and storage add several to a few tens of USD·(tCO
2)
−1 each, depending on distance, flow, and reservoir type; cluster-hub models generally improve economics. Taken together, amine scrubbing is deployment-ready for many integrated mills, whereas TGR-BF and oxy-fuel pathways warrant targeted demonstrations to de-risk integration and reduce energy penalties.
While the CCS pathway scales well in principle, it also embeds significant resource demands and system-level risks. Precisely because it can retrofit existing BF-BOF assets, CCS can deliver sizable abatement quickly. Research targeting China’s steel industry [
6] shows that through integrated industry transformation pathways and source-sink matching optimization, CCS technology has potential to contribute 31.4%-40.7% of steel industry emission reduction effects (by 2060), corresponding to annual emission reductions of approximately 470-610 million tonnes of CO
2. However, such emission reduction scales come at the cost of enormous resource expenditures. Optimized deployment results indicate that nationwide CCS retrofitting has levelized emission reduction costs of approximately 187-193 CNY·(tCO
2)
−1 (about 25-30 USD·(tCO
2)
−1, which may have economic feasibility under future carbon price levels, but these cost estimates often ignore uncertainties like technical risks, operational maintenance complexity, and long-term storage safety.
More seriously, large-scale CCS deployment will impose substantial resource burdens. This does not contradict economies of scale: Scale effects reduce unit capture costs, but whole-process accounting shows that absolute energy and water use rise with the volume of CO
2 captured because post-combustion capture imposes parasitic steam/electric loads and cooling needs. Based on a recent national source-sink optimization for China’s steel sector (mentioned above) [
6], CCS retrofits would add on the order of 1.0-1.3 EJ (≈ 951-1427 PJ) of energy, equivalent to about 21.7%-22.4% higher sectoral final energy use, and imply a net water use of ∼1.60-1.69 m
3 per tCO
2 captured (enhanced water recovery partially offsets gross withdrawals). Under these whole-chain assumptions (capture + compression/transport + storage), the levelized abatement cost averages 187-193 CNY·(tCO
2)
−1 (≈ 25-26 USD·(tCO
2)
−1), with site-level heterogeneity spanning 74-542 CNY·(tCO
2)
−1 (≈ 10-73 USD·(tCO
2)
−1). The burden is geographically concentrated: Hebei alone accounts for ∼23.8%-28.6% of the added energy and ∼36.6%-40.3% of the added water, reflecting co-located steel capacity and storage access. Accordingly, suitable deployment scenarios are integrated BF-BOF clusters located within ∼250 km of storage sites (preferably oilfields) where shared pipelines/hubs lower transport and storage costs; in such clusters, carbon pricing around 239 CNY·(tCO
2)
−1 (≈ 32 USD·(tCO
2)
−1 renders most captured emissions cost-effective.
At the power system level, CCS technology does demonstrate certain synergistic value [
14]: Lowest-cost scenarios equipped with 16% fossil fuel generation with CCS can reduce network investment costs by 2.5% (about 16.8 billion USD) and decrease national electricity shortage rates by up to 9.0 percentage points by 2050. Under extreme climate events, such configurations can reduce electricity shortage risks from blizzards, sandstorms, droughts, and heat waves by 54%, 56%, 57%, and 68%, respectively. However, this system resilience improvement comes at the cost of continued fossil fuel dependence, creating inherent contradictions with deep decarbonization long-term goals. More concerning is that excessive dependence on CCS technology may delay truly clean technology development, creating “carbon lock-in” effects that lead the steel industry further down incorrect technology pathways.
4. Hydrogen metallurgy predicament: Vast gaps between ideals and reality
For H
2-DRI, the mismatch between hydrogen resource availability and steel capacity is reinforced by MEIC findings: the provinces with highest steel emissions are also those with limited renewable energy endowments. Following
Fig. 2(b), which compares deployment scales, we follow Zhang and Chen [
7] in defining moderate and aggressive scenarios as output-based rollouts by 2050: moderate ≈ 149.6 Mt·a
−1 and aggressive ≈ 408.0 Mt·a
−1 of H
2-DRI steel. In short, moderate denotes a geographically concentrated, renewables-matched rollout with near-fully green hydrogen, while aggressive denotes rapid, wide-area expansion that saturates local renewables and pulls in supplemental fuels—thereby eroding decarbonization gains. At the moderate scale, renewables can meet ∼97%-100% of hydrogen demand; at the aggressive scale, one-third to one-half must be supplemented by grid electricity and/or fossil-based hydrogen, driving up indirect emissions. Reflecting this supply constraint, life-cycle carbon intensities remain low at moderate scale—0.36, 0.15, and 0.91 tCO
2·(t steel)
−1 for solar-, wind-, and biomass-anchored H
2-DRI, respectively—but rise sharply at aggressive scale to 1.33, 0.96, and 1.12 tCO
2·(t steel)
−1, approaching natural gas-DRI levels.
However, recent research reveals an unexpected phenomenon [
7]: The emission reduction benefits of hydrogen-based steelmaking do not correlate linearly with deployment scale, and beyond certain thresholds, “diminishing returns or even reversals in emission reduction gains” may occur. At aggressive deployment levels, carbon intensities increase markedly to 1.33, 0.96, and 1.12 tCO
2 per t steel for the solar-, wind-, and biomass-anchored pathways, respectively, representing increases of 3.7-fold (solar), 6.4-fold (wind), and 1.2-fold (biomass). According to the related research [
7], this counterintuitive outcome does not result from higher on-site emissions of H
2-DRI/electric arc furnace (EAF) itself, but rather from upstream energy constraints. Once the availability of nearby renewables is saturated, additional demand for electrolysis and EAF electricity, as well as hydrogen production, must be met by supplemental sources—primarily grid electricity and, in the biomass case, coal-derived hydrogen. The life cycle assessment (LCA) encompasses upstream renewable construction, hydrogen production, EAF power consumption, and these supplemental inputs [
7]. Because the grid and fossil-based hydrogen carry much higher emission factors, their contribution dominates the carbon footprint and offsets the advantages of green hydrogen, leading to significantly higher life cycle intensities in aggressive scenarios. This finding challenges the conventional assumption that larger deployment scales inherently deliver greater decarbonization benefits, and instead highlights structural limitations that hydrogen metallurgy may face in large-scale applications.
This scale-effect reversal originates in structural constraints on green-hydrogen supply that are unlikely to be resolved in the near term. Northwest and southwest regions have abundant solar and wind resources but limited steel capacity, while major steel-producing regions like north and east China have relatively limited renewable resource endowments, approximately 52.6%-55.8% of operating steel units nationwide lack conditions for locally obtaining sufficient green hydrogen. Specifically, Hebei, Jiangsu, Liaoning, Shandong, and Shanxi Provinces collectively account for 56.7% of national hydrogen demand, but their solar, wind, and biomass energy technology development potentials represent only 1.9%, 10.4%, and 21.7% of national totals, respectively. This regional mismatch implies that even under moderate hydrogen-metallurgy roll-out, local renewables can only just meet ∼97%-100% of hydrogen needs; once scaling aggressively, fossil-based hydrogen would inevitably fill about one-third to one-half of the shortfall—eroding the original decarbonization advantage of H2-based pathways.
Against this constraint landscape, near-term CCS deployment in steel should prioritize post-combustion retrofits at integrated hubs, given their higher TRL and faster abatement per unit of new infrastructure [
12]. In parallel, scaling demonstrations of TGR-BF with integrated CO
2 removal and advancing oxy-fuel variants where oxygen supply chains are available will broaden the portfolio and de-risk future options. Shared CO
2 pipelines and storage hubs are critical to lower unit transport/storage costs and improve bankability [
15]. To ensure apples-to-apples comparisons—both across CCS configurations and vis-à-vis H
2-DRI—future assessments should report transparent cost boundaries distinguishing capture-only from capture-plus-transport-and-storage (T&S).
Set against CCS, near-term H
2-DRI deployment presents additional concerns.
Fig. 2(c) clearly demonstrates significant differences in annual emission reduction potential between CCS and H
2-DRI technologies through nested comparisons. CCS technology’s annual emission reductions of 472.4-609.2 Mt [
6] significantly exceed H
2-DRI technology’s moderate deployment of 219.91-333.65 Mt [
7]. This gap reflects not only technology scale characteristics but a practical dilemma under tight decarbonization timelines: We may need to prioritize CCS—despite its higher resource intensity—because it delivers larger, faster abatement, while H
2-DRI remains constrained by renewables and hydrogen supply. Moreover, evidence that excessive roll-out can drive six-fold increases in hydrogen-metallurgy carbon intensity further challenges its suitability as the primary decarbonization pathway in the near term.
Water resource constraints further deteriorate hydrogen metallurgy prospects. H
2-DRI technology has water intensities of 1.47-7.30 m
3 per tonne steel, with water consumption surging under aggressive deployment scenarios [
7]. For comparison with CCS (Section 3), given that steel production emits approximately 2 tCO
2 per tonne steel, this translates to 0.74-3.65 m
3 per tonne CO
2 avoided—overlapping with but potentially exceeding CCS water intensity (1.60-1.69 m
3 per tCO
2 captured) at the higher end. At the sectoral scale, solar hydrogen production schemes consume up to 3.009 billion m
3 annually, while wind schemes also increase to 1.766 billion m
3. Converting to water requirements per tonne CO
2 emission reduction, solar and wind schemes increase 2.5-fold and 4.5-fold, respectively from moderate to aggressive stages. More concerning, under aggressive hydrogen metallurgy schemes, Hebei Province bears the heaviest water consumption due to undertaking substantial hydrogen production tasks, with annual water consumption reaching 622 million m
3 (solar scheme) or 539 million m
3 (wind scheme), causing already drought-stressed Hebei’s industrial water use to surge dramatically, potentially triggering serious water resource crises.
Fig. 2(d) shows the relationship between unit product costs and emission reduction costs, further highlighting the economic dilemmas in technology selection [
6]. The background density contours (kernel density estimation (KDE)) indicate where technology deployment scenarios cluster in cost-benefit space, with darker orange regions representing more common or economically feasible deployment conditions under the modeled assumptions. The economic viability of H
2-DRI is highly sensitive to deployment scale and the type of energy input, creating substantial uncertainty. Under moderate deployment scenarios, solar- and wind-based H
2-DRI pathways can achieve net economic benefits, with cost reductions of 12.12% and 13.55%, respectively, but only at a carbon price of around 363 CNY (≈ 50 USD) per tonne CO
2—a level far above current market conditions. When scaled aggressively, however, costs increase sharply and surpass those of traditional processes. Biomass-based H
2-DRI remains prohibitively expensive in all cases, with costs exceeding conventional routes by 9.03% under moderate deployment and 1.75% even under aggressive rollout [
7]. Importantly, these cost estimates for H
2-DRI assume near-site renewable hydrogen supply and therefore exclude long-distance hydrogen storage and transport (such as pipelines, liquefaction, or ammonia carriers). By contrast, CCS estimates incorporate the full chain of capture, transport, and storage. This boundary difference means that the apparent competitiveness of H
2-DRI is likely conservative, and actual system-wide costs could be even higher once distribution infrastructure is included. Without robust policy incentives and stable carbon pricing, firms are unlikely to adopt these new technologies at the scale required.
5. Systemic challenges: Contradictions between time windows and technology maturity
When combined with the empirical MEIC trends, the broader systemic challenge becomes evident: Rising national totals, persistent provincial baselines, and regional resource mismatches all interact with technology maturity issues.
Fig. 2(e) showing key low-carbon technology capability heatmaps provides more macro but equally concerning technology development perspectives [
8]. From 2020 to 2050, various low-carbon technology development trajectories show obvious staged characteristics, but technology maturation and large-scale deployment speeds may not match climate target urgency [
16]. Early stages are dominated by coal-based technologies (including coal with CCS), mid-term stages see solar emerging as the highest capability technology alongside rapid hydrogen development, and later stages position carbon capture and negative emission technologies as critical. However, this technology evolution pathway assumes various technologies can develop and deploy smoothly as planned, ignoring technology development uncertainties, path dependencies, and lock-in effects. CCS technology’s importance gradually emerges in mid-to-late periods, while H
2-DRI technology’s development window mainly concentrates in mid-term stages when hydrogen technology matures, but if hydrogen technology development speeds fall short of expectations, entire steel decarbonization processes will face serious delays.
The realistic predicament is that the steel industry faces not merely technology selection problems, but multiple challenges involving resource constraints, economic feasibility, time windows, and system complexity. The respective strengths and weaknesses of CCS and H2-DRI represent scale versus intensity debates in emission reduction strategies, yet neither can address the decarbonization challenge. While CCS pathways can at the macro scale achieve the largest emission reduction pie, their per-unit emission reduction resource costs are enormous and may delay truly clean technology development; hydrogen metallurgy pathways, while achieving near zero-carbon emission intensities under ideal conditions, face severe renewable energy supply constraints, with large-scale deployment potentially leading to significant emission reduction effect deterioration. More severely, both technology pathways face enormous water resource pressures, which may become decisive factors limiting technology deployment given increasingly scarce water resources.
Future steel industry emission reduction must pursue challenging multi-technology integration, phased advancement routes, yet success probabilities remain concerning. Short-term, while using CCS for retrofitting key steel enterprises and industrial clusters represents realistic choices for rapidly reducing absolute emissions, such choices themselves reflect our lack of better alternatives. Meanwhile, developing H2-DRI demonstration projects in renewable resource-rich, suitable regions serves more to maintain technology development possibilities rather than holding excessive expectations for large-scale applications. Current factors limiting hydrogen promotion include lack of large-scale, long-distance hydrogen transport pipeline networks and storage infrastructure, requiring massive investments and long-term planning with enormous risks under increasing economic uncertainties.
Medium-term, while regions can theoretically gradually form complementary emission reduction patterns based on resource endowments, realistic interest coordination, technology standard unification, and cross-regional cooperation face enormous challenges. Building cross-regional power transmission channels and hydrogen transport pipeline networks requires unprecedented investment scales and policy coordination, with extremely difficult realization under intensifying local protectionism and resource competition. Long-term, visions of power systems approaching 100% decarbonization with dramatically reduced hydrogen production costs may be overly optimistic, with technology development uncertainties and economic-social system complexities potentially preventing timely goal achievement.
Policy-level challenges are equally severe. While establishing reasonable, long-term predictable carbon pricing signals represents key means for guiding industry investment and technology selection, the feasibility of substantially raising carbon prices remains questionable under economic growth pressures and intensifying international competition. Resource guarantees face multiple constraints: Cross-regional water resource allocation involves complex stakeholder interest coordination [
17]; the energy consumption and costs of seawater desalination technology remain high [
18]; and optimizing power dispatch to prioritize clean electricity for emission reduction projects may conflict with power system economics and stability. Infrastructure construction forward-looking planning requires massive government investments and long-term commitments, yet such commitment sustainability remains concerning under increasing fiscal pressures and uncertain investment returns [
19].
6. Breakthrough directions: Seeking solutions within constraints
The preceding analysis has outlined a sobering reality: China’s steel sector faces not merely technology selection problems but multiple interlocking challenges involving resource constraints, economic feasibility, time windows, and system complexity. Yet beneath these macro-level dilemmas lies a more fundamental tension—the gap between systems-level strategic assessment and the granular engineering realities that ultimately determine whether theoretical pathways can translate into operational success.
Consider the H
2-DRI-EAF pathway, which our analysis presents as achieving near-zero process emissions under ideal conditions. At the process level, however, a critical challenge emerges: Unlike BF that remove impurities through slag formation, H
2-DRI retains nearly all gangue from iron ore embedded in the reduced product [
20]. Standard electric arc furnaces are not designed to handle this burden, requiring additional melting and slag removal steps—such as dedicated electric smelting furnaces (ESFs) or H
2-DRI + ESF + BOF configurations [
21]—that fundamentally alter energy consumption and cost structures [
20]. These process modifications can increase electrical energy requirements and lengthen tap-to-tap times, though operational adjustments including continuous DRI feeding, maintaining hot heel liquid baths, inert gas stirring, and controlled carbon addition have shown promise in improving efficiency, with industrial trials demonstrating that optimized 50% DRI charges can require less energy per tonne than 100% scrap charges [
20]. Similarly, the BF-BOF-CCUS pathway faces practical constraints in capturing CO
2 from multiple emission points, where achieving greater than 90% overall capture proves extraordinarily difficult despite theoretical potential, with practical abatement likely limited to approximately 60% for retrofitted systems due to infrastructural complexities of capturing emissions from all individual point sources [
22,
23]. These process-level complexities underscore that macro-level pathway comparisons, while essential for identifying structural constraints, remain incomplete without detailed techno-economic analysis and LCA performed on an equivalent productivity basis with transparent cost boundaries.
This recognition illuminates the inherent limitations of the present study. Our grounding in empirical MEIC emission data enables us to identify structural constraints—provincial clustering, resource mismatches, rising national emissions from 8.2 Gt in 2010 to 11.2 Gt in 2023—that any technology must navigate. Yet this systems-level focus constrains our ability to perform bottom-up engineering cost analysis or to differentiate among critical operational adjustments that affect energy efficiency and economics at the plant level. The boundary asymmetry in cost estimates further complicates comparison: BF-BOF-CCUS projections typically incorporate full capture-transport-storage chains, with estimates ranging from 500 to 600 USD per tonne in recent assessments [
24], whereas H
2-DRI-EAF economics often assume near-site renewable hydrogen supply and exclude long-distance transport infrastructure costs, with cost projections highly sensitive to hydrogen price (1.5-3 USD·kg
−1) and carbon pricing (15-50 USD·(tCO
2)
−1), where modest carbon prices of 15 USD·t
−1 can enable green H
2-DRI-EAF at 1.5 USD·(kg H
2)
−1 to undercut conventional BF-BOF costs [
25]. Similarly, while we quantify resource burdens—CCUS imposing approximately 1.0-1.3 EJ of additional energy (equivalent to 21.7%-22.4% higher sectoral final energy use) and net water use of approximately 1.60-1.69 m
3 per tCO
2 captured, with Hebei bearing ∼36.6%-40.3% of added water consumption [
6], and H
2-DRI exhibiting water intensities of 1.47-7.30 m
3 per tonne steel that surge dramatically under aggressive deployment scenarios [
10], with H
2-DRI-EAF energy requirements of approximately 3.5-4.5 MW·h of electricity per tonne for hydrogen production and EAF melting, potentially requiring 2-3 GW of dedicated renewable capacity for a 1 million tonnes per annum (MTPA) facility [
24]—these figures do not constitute a full multi-criteria LCA tracking upstream impacts, land transformation, critical material depletion, and slag management differences across consistent boundaries [
26].
These limitations reflect a broader methodological challenge: Effective decarbonization analysis must bridge systems-level constraints with plant-level engineering realities, yet few frameworks successfully integrate both scales. Moving forward, viable pathways demand research that combines detailed process modeling of engineering variants—quantifying energy penalties, capital costs, and emissions on a per-tonne basis [
24]—with transparent economic comparisons distinguishing what costs are and are not included, comprehensive LCA tracking multiple environmental dimensions with regional specificity [
26], and dynamic scenario analyses simulating technology learning, infrastructure buildout, and policy evolution over time rather than static 2050 snapshots. Critically, such analyses must integrate demand-side measures—material efficiency, longer product lifespans, and increased scrap recycling as building stock matures—not as separate options but as essential components reducing pressure on both CCUS-intensive retrofits and resource-intensive H
2-DRI primary production. Scrap-based EAF steel demonstrates significantly lower carbon footprints (approximately 0.7 tCO
2·(t steel)
−1) compared with conventional BF-BOF (approximately 2.3 tCO
2·(t steel)
−1) [
24], though scrap availability currently limits this route’s expansion potential.
Within these constraints, several breakthrough directions remain viable. Regionalized deployment strategies aligning CCS retrofits with steel-intensive northern provinces possessing geological storage access while piloting H
2-DRI in renewable-rich western regions can leverage local comparative advantages. Targeted innovation in long-distance hydrogen transport, electrolytic ironmaking, and multi-point capture efficiency could ease structural barriers. LCA indicates that while BF-BOF-CCUS faces practical limitations with capture rates likely around 60% or lower in retrofitted systems [
22,
23] and achieves at best approximately 49% emission reduction in optimized configurations [
26], H
2-DRI-EAF with green hydrogen can deliver up to 90% or greater direct emission cuts, achieving substantially lower carbon intensities compared to conventional routes [
24,
26,
27], though with shifted burdens to renewable infrastructure land use and critical mineral requirements [
26]. Policy frameworks establishing stable carbon price signals, coordinating cross-regional infrastructure investment, and prioritizing water allocation for strategic projects will prove as decisive as technological progress itself.
The steel sector’s trajectory thus becomes a test case for whether large-scale industrial decarbonization can succeed under real-world constraints. Our empirical analysis establishes the structural realities any pathway must confront—rising emissions, coupled industrial-power systems, concentrated provincial influence—yet translating these insights into actionable strategies requires the granular, multi-scale analysis that bridges systems constraints with engineering realities. The decarbonization paradox we identify remains valid: Neither CCS nor hydrogen metallurgy alone can deliver deep reductions within limited time windows [
22,
25]. Only through integrated transformation encompassing demand reduction, regionalized deployment, infrastructure development, and disruptive innovation—orchestrated within governance systems reconciling carbon neutrality with resource security and regional disparities—can the sector navigate its challenge. In this sense, China’s pathway will not only determine domestic climate outcomes but illuminate whether industrial transformation at the scale and speed climate targets require is achievable under the messy, constrained conditions characterizing real-world change.
7. Conclusions
This perspective has examined China’s steel decarbonization challenge through the integration of empirical emission data and process-level techno-economic analysis, revealing systematic gaps between systems-level projections and operational realities. Drawing on MEIC 2010-2023, we established an empirical baseline where national CO2 emissions rose from 8.2 to 11.2 Gt, industrial and power sectors remained tightly coupled, and five regions—Hebei, Shandong, Jiangsu, Inner Mongolia, and Guangdong—exerted disproportionate influence on sectoral trajectories. Against this backdrop, we assessed the two most prominent technological pathways and found that neither can independently deliver the deep, rapid reductions climate targets demand.
Our analysis contributes three key findings that challenge conventional technology assessments. First, full-chain resource accounting reveals that CCS—despite offering the largest near-term abatement potential, contributing over 40% of industry reductions by 2060—imposes substantial burdens often invisible in aggregate scenario modeling. CCS retrofits would require approximately 1.0-1.3 EJ of additional energy, equivalent to 21.7%-22.4% higher sectoral consumption, and 1.60-1.69 m3 of water per tonne CO2 captured. These burdens concentrate heavily in already water-stressed industrial hubs, with Hebei alone bearing approximately 36.6%-40.3% of added water consumption.
Second, H2-DRI exhibits counterintuitive scale effects that fundamentally challenge assumptions of linear decarbonization benefits from technology scale-up. Under moderate deployment scenarios aligned with local renewable supply, life-cycle carbon intensities remain low—from 0.15 to 0.91 tCO2 per tonne steel depending on energy source. At aggressive deployment levels, however, emission intensities increase multi-fold as renewable supply saturates and fossil-derived hydrogen fills the gap, with some pathways showing increases of 1.2- to 6.4-fold compared to moderate deployment. Third, boundary asymmetries in economic comparisons systematically distort technology rankings. CCS cost estimates typically incorporate full capture-transport-storage chains (approximately 500 to 600 USD·(tCO2)−1), whereas H2-DRI projections often exclude long-distance hydrogen infrastructure—pipelines, liquefaction facilities, or ammonia carriers—making direct comparisons misleading without transparent boundary definitions.
Process-level engineering complexities compound these systemic constraints in ways rarely captured by aggregate modeling. H2-DRI retains gangue from iron ore embedded in reduced products, requiring additional melting steps or dedicated ESF that fundamentally alter energy consumption and cost structures. Industrial trials show that optimized operations can partially mitigate these burdens, yet the fundamental challenge remains. Similarly, BF-BOF-CCUS faces practical multi-point capture limitations—capturing emissions from BF, BOF, sinter plants, and auxiliary processes simultaneously—that reduce achievable abatement to approximately 60% in retrofitted systems despite theoretical potential exceeding 90%. These granular engineering realities, typically absent from macro-level scenario analyses, ultimately determine whether ambitious pathways can translate from theoretical potential to operational deployment.
The steel decarbonization challenge thus extends beyond technology selection to encompass resource geography, infrastructure inertia, and economic path dependencies that decades of investment have locked into place. These structural tensions are not unique to steel or to China. Cement, chemicals, aluminum, and other energy-intensive industries face fundamentally similar constraints: Production concentrated where legacy infrastructure exists rather than where low-carbon resources are abundant; transformation speeds constrained by capital stock turnover and supply chain reorganization; deployment scales limited by regional resource carrying capacity rather than global theoretical potential. The paradox we identify in steel—that aggressive technology deployment can paradoxically increase rather than decrease emissions when resource constraints bind—may foreshadow similar challenges in other hard-to-abate sectors.
This recognition raises a critical methodological question: How wide is the gap between what systems-level scenario analyses project as feasible and what process-level engineering reveals as operationally achievable? Our findings suggest this gap may be substantial. When modeling assumes smooth technology substitution and elastic resource supply, yet empirical analysis reveals gangue retention requirements, multi-point capture inefficiencies, and multi-fold emission intensity increases under aggressive deployment, the divergence between theoretical pathways and practical implementation becomes apparent. Current integrated assessment models and scenario analyses, while essential for long-term planning, may systematically underestimate the friction inherent in large-scale industrial transformation.
Closing this gap requires new analytical frameworks that integrate detailed process engineering with regional resource constraints, infrastructure requirements, and dynamic techno-economic assessment across consistent boundaries. Such frameworks must distinguish between pathways that are theoretically elegant—achieving deep emission cuts on paper—and those achievable under the time, resource, and economic constraints characterizing real-world industrial systems. They must account for non-linear responses to deployment scale, regional resource mismatches, and the engineering modifications required when ideal conditions do not hold. Critically, they must integrate demand-side measures not as optional add-ons but as essential components of any viable transformation strategy.
In this sense, the steel sector’s trajectory serves not merely as a decarbonization case study but as an empirical test of whether the fundamental restructuring modern economies require is achievable at the scale and speed climate stabilization demands. Our analysis establishes that neither CCS nor H2-DRI alone can deliver the necessary reductions within available time windows. Only through integrated strategies encompassing demand reduction, regionalized deployment aligned with local resource endowments, and targeted innovation in hydrogen transport, electrolytic ironmaking, and capture efficiency, can the sector navigate its challenge. Whether China—and by extension, the global industrial system—can orchestrate such integrated transformation under real-world constraints of resource scarcity, economic feasibility, and governance capacity will determine not only domestic climate outcomes but the broader feasibility of industrial decarbonization at the pace the climate crisis requires.
PDF
(1450KB)
