What makes sCO₂ turbines different — and why efficiency matters now

Supercritical CO₂ turbines outperform conventional steam Rankine systems because CO₂ operating beyond its critical point (31.1°C, 7.39 MPa) simultaneously exhibits liquid-like density and gas-like viscosity — a combination that dramatically reduces compression work and allows turbomachinery to be built far more compactly. In a closed-loop Brayton cycle, this translates directly into higher thermal efficiency from a smaller physical footprint, making sCO₂ an attractive working fluid across nuclear, solar, fossil, and waste-heat recovery applications.

The efficiency range within the retrieved dataset spans from approximately 40% for simple regenerative configurations to as high as 49% for recompression layouts at turbine inlet temperatures of 700°C. This compares favourably with steam cycles at equivalent heat source temperatures, and is the principal driver behind sustained R&D investment across multiple sectors. According to WIPO, clean energy power generation technologies have been among the fastest-growing patent categories globally since 2015 — a macro trend that provides the commercial backdrop for sCO₂’s rising filing activity.

The technology encompasses several sub-domains: turbine aerodynamic design (radial inflow, radial outflow, axial, and multi-stage configurations); full-system cycle optimisation (recompression, partial cooling, intercooling, and cascaded layouts); working fluid modification via CO₂-based binary mixtures; thermal energy storage integration; and direct-fired combustion architecture for Allam-cycle variants. The dataset retrieved for this analysis spans 2009 to 2025, with a clear acceleration in publication and filing density from 2018 onward.

From feasibility to commercial filings: the sCO₂ innovation timeline

The sCO₂ innovation record divides cleanly into four phases, each marked by a distinct shift in research focus — from cycle-level feasibility, through turbomachinery characterisation, to the commercial patent filings that now signal protectable product architectures.

The foundational period (2009–2014) was defined by cycle-level feasibility work. The 2014 NREL comparison of sCO₂ cycle configurations established the recompression and partial cooling cycles as leading candidates for concentrated solar power (CSP). A 2013 French study on coal-fired power plant integration demonstrated potential overall efficiencies of 41.5–44.5% using sCO₂ Brayton cycles combined with carbon capture — an early signal of the technology’s decarbonisation credentials.

The cycle characterisation period (2015–2018) was anchored by KAIST, whose 2015 review characterised sCO₂ cycle development status across nuclear, solar, and fossil applications. By 2018, detailed turbomachinery design codes had emerged, including the Institute of Engineering Thermophysics (Chinese Academy of Sciences) compressor design methodology — a sign that component-level engineering was maturing alongside system-level analysis.

The turbine design intensification period (2019–2022) produced the highest concentration of turbine-specific results in the dataset. Cranfield University designed a 510 mm diameter, 100 MWth-scale radial turbine running at 21,409 rpm. Xi’an Jiaotong University applied deconvolutional neural networks to reconstruct turbine pressure and temperature fields at 60,000 rpm. Khalifa University explored 600 geometric variants of an 8 MW radial turbine using deep neural network optimisation — a methodological step-change that compressed years of CFD iteration into a single optimisation run.

“Recompression sCO₂ Brayton cycles reach as high as 49% efficiency at 700°C inlet temperatures — a performance level that positions sCO₂ as a credible replacement for steam Rankine cycles in next-generation CSP and nuclear plants.”

The commercialisation period (2022–2025) is defined by active patent filings. Hanwha Power Systems filed an EP patent on a supercritical carbon dioxide power generation system in December 2022. Toshiba Energy Systems filed an EP patent on a combustor structure for sCO₂ gas turbines in September 2025. The University of the Basque Country filed an EP patent on a multi-recuperator regenerative Brayton cycle with auxiliary compressors as recently as December 2025. These filings signal a transition from academic research toward protectable commercial architectures.

Four technical clusters shaping sCO₂ turbine design

The retrieved patent and literature record organises into four distinct technical clusters, each representing a coherent research programme with its own leading institutions, design methodologies, and commercial implications.

Cluster 1: Radial inflow turbine design and optimisation

The radial inflow turbine is the most widely studied configuration in the dataset, suited for power outputs from 0.1 MW to approximately 100 MW. Research has converged on computational design-optimisation frameworks coupling one-dimensional mean-line codes with three-dimensional CFD validation and real-gas property databases. The Institute of Engineering Thermophysics (Chinese Academy of Sciences) achieved a total-to-static efficiency of 85.77% for a 2.1 MW turbine in 2022. Khalifa University’s 2021 deep neural network study explored 600 geometric variants of an 8 MW radial turbine. Xi’an Jiaotong University applied deconvolutional neural networks to reconstruct turbine pressure and temperature fields at 60,000 rpm, demonstrating that AI-augmented design is becoming a standard methodology in this field.

Cluster 2: Radial outflow and axial multi-stage turbines

For larger-scale or higher-pressure-ratio applications, radial outflow and multi-stage axial turbines offer structural advantages including distributed enthalpy drop and improved bearing life. A 2022 study from the School of Energy Systems investigated radial outflow designs at 600°C, 200 bar inlet conditions across four mass flow rates. Korea Maritime and Ocean University developed a multi-stage preliminary design algorithm for a two-stage radial outflow turbine, validated via CFD. Xi’an Jiaotong University completed thermodynamic and aerodynamic design of a 10 MW three-stage axial sCO₂ turbine in 2019. This sub-field is comparatively less mature within the dataset but is attracting increasing attention as power outputs scale toward utility-class requirements.

Cluster 3: Full-system cycle architecture and recompression configurations

A significant portion of the dataset addresses system-level cycle layout optimisation rather than turbine component design in isolation. Recompression, partial cooling, and multi-recuperator configurations are identified as efficiency-maximising approaches. The University of the Basque Country’s December 2025 EP patent claims improved heat recovery via N≥3 recuperators in series with N or N-1 auxiliary compressors — a novel architectural claim that distinguishes from prior dual-recuperator recompression art. Hanwha Power Systems’ EP patent covers dual compression sections, dual regeneration sections, a power transmission stage, and parallel turbine operation to reduce RPM and distribute torque. Harbin Engineering University developed an in-house code coupling turbomachinery geometric design with cycle performance prediction for Generation IV nuclear applications.

Cluster 4: CO₂ binary mixture working fluids

A growing research thread addresses replacing pure CO₂ with binary mixtures to shift the critical temperature, improve cycle efficiency, and adapt cycles to dry-cooling environments — particularly relevant for CSP in arid climates. Universidad Politécnica de Madrid evaluated eleven CO₂ binary mixtures in 2019, reporting thermodynamic efficiency improvements of 3–4%. A 2021 follow-up assessed CO₂/COS, CO₂/H₂S, CO₂/NH₃, and CO₂/SO₂ mixtures in both simple and complex recompression configurations. The University of London investigated CO₂/TiCl₄, CO₂/NOD, and CO₂/C₆F₆ mixtures with specific focus on turbine design sensitivity to dopant fraction. Despite substantial academic investigation, no patents specifically claiming these fluid compositions in turbine applications appear in the retrieved dataset — a potential whitespace discussed further in the strategic implications section.

Where sCO₂ turbines are being deployed: application domains

Concentrated solar power (CSP) is the most heavily represented application domain in the dataset, followed closely by Generation IV nuclear reactors, fossil fuel and waste heat recovery, and direct-fired Allam-cycle architectures. Each domain presents a distinct set of temperature requirements, scale constraints, and competitive dynamics.

Concentrated Solar Power (CSP) leads the dataset. sCO₂ cycles are regarded as leading candidates to replace steam Rankine cycles in central receiver tower systems, driven by efficiency advantages at 400–800°C heat source temperatures and potential for reduced levelised cost of electricity. A 2020 University of Seville study specifically addressed the potential of sCO₂ power cycles to reduce LCOE for contemporary CSP plants. Dynamic system modelling for CSP — including a 2023 year-long solar simulation from the National Technical University of Athens — signals a shift toward bankable project designs, as noted by analysts at the IEA tracking renewable energy cost trajectories.

Generation IV nuclear reactors represent the second major domain, with KAIST consistently positioning sCO₂ cycles as the preferred conversion technology for microreactor concepts. A 2023 KAIST comparative evaluation of gas Brayton cycles for micro-nuclear reactors reinforced this positioning. The Nuclear Power Institute of China contributed a 2021 study on temperature feedback effects in sCO₂-cooled reactors, indicating parallel national programmes tracking the technology.

Fossil fuel and waste heat recovery applications include sCO₂ Brayton cycles as bottoming cycles for gas turbine exhaust and as integrated cycles in coal-fired plants. A Cranfield University techno-economic analysis published in 2021 examined sCO₂ integration with coal-fired power plants. Performance improvements over organic Rankine cycles are demonstrated at higher-temperature waste heat sources above 350°C — a threshold relevant to industrial process heat recovery. Standards bodies including ISO are developing frameworks for advanced power cycle efficiency measurement that will affect how sCO₂ performance claims are verified in commercial projects.

Direct-fired sCO₂ and Allam-cycle architectures represent the most commercially advanced frontier in the dataset. Toshiba’s 2025 EP patent claims a combustor structure penetrating the turbine casing perpendicular to the rotor axis, with integrated fuel and oxidant supply parts for direct sCO₂ combustion — an architecture that enables inherent carbon capture by operating with an oxy-fuel oxidant. This is one of very few patents in the dataset claiming specific mechanical architecture for Allam-cycle-class sCO₂ turbines.

Geographic and assignee landscape: who holds the patents

Innovation in sCO₂ turbine technology is distributed across a wide set of institutions rather than concentrated in a small number of dominant assignees — a pattern that reflects the technology’s academic origins and the relatively early stage of commercial consolidation.

China is the most prolific country in the dataset by publication count. Xi’an Jiaotong University, the Institute of Engineering Thermophysics (Chinese Academy of Sciences), Harbin Engineering University, Central South University, and Lanzhou Jiaotong University all appear across turbomachinery component design, cycle optimisation, deep-learning approaches, and nuclear reactor coupling. However, Chinese institutions are underrepresented in the active patent records retrieved — a gap that may reflect publication-first commercialisation strategies, domestic-only patent filing strategies not captured in this dataset, or early-stage technology status. IP strategists should conduct dedicated Chinese patent office searches to fill this gap.

South Korea is prominent through KAIST (nuclear and microreactor applications) and Korea Maritime and Ocean University (radial outflow turbine design). Hanwha Power Systems holds active EP and JP patents on sCO₂ power generation systems, demonstrating Korean industrial engagement at the commercial patent level. The JP filing previously associated with Doosan Heavy Industries & Construction Co., Ltd. reinforces the depth of Korean OEM investment in this space.

Japan holds the highest-profile industrial patent in the dataset through Toshiba Energy Systems & Solutions Corporation’s September 2025 EP filing on combustor architecture for sCO₂ gas turbines. Spain is represented by Universidad Politécnica de Madrid (working fluid mixtures sub-domain) and the University of the Basque Country, which holds the most recent EP patent filing in the dataset (December 2025). Russia‘s Moscow Power Engineering Institute produced multiple publications on Allam-cycle CO₂ turbine flow path design and kinematic efficiency analysis, representing a distinct national programme on oxy-fuel sCO₂ turbines. UK contributions are anchored at Cranfield University across radial turbine design, combined cycle integration, and aero bottoming cycle studies.

Active patent jurisdictions in the dataset are EP (3 active patents) and JP (1 active patent from Hanwha, 1 from Doosan). No US patents appear in the retrieved results, though US institutions including NREL and Arizona State University are active in the academic literature. The EPO‘s growing role as the preferred jurisdiction for sCO₂ commercial filings by Asian OEMs is consistent with broader trends in clean energy patent strategy documented in EPO’s annual patent index.

Strategic whitespace and IP implications for 2026

The sCO₂ patent landscape contains several identifiable whitespace areas and competitive pressure points that should inform R&D prioritisation and IP strategy for organisations active in this field.

Patent whitespace in combustor and turbine integration for direct-fired cycles. Toshiba’s 2025 EP combustor patent is one of very few in this dataset claiming specific mechanical architecture for Allam-cycle-class sCO₂ turbines. R&D teams targeting the direct-fired sCO₂ market should map freedom-to-operate carefully around this filing and its family members before committing to combustor geometry choices. The USPTO record for equivalent US filings should be checked separately, as no US counterparts appear in the retrieved dataset.

Korean and Japanese OEM engagement signals near-term commercial hardware. Hanwha Power Systems and Toshiba Energy Systems holding active EP patents — not merely academic publications — indicates that sCO₂ turbine systems are advancing toward commercial product development in Asia-Pacific. Western OEMs should assess competitive positioning against these filings, particularly in the EP jurisdiction where both companies have chosen to protect their system-level and combustor architectures.

China’s academic dominance has not yet translated to visible patent filings in this dataset. Chinese institutions account for the largest share of literature results but are underrepresented in the active patent records retrieved. This may reflect publication-first commercialisation strategies, domestic-only patent filing strategies not captured here, or early-stage technology status. IP strategists should conduct dedicated Chinese patent office (CNIPA) searches to fill this gap before drawing conclusions about Chinese competitive positioning.

AI-driven turbomachinery design creates a capability gap. Teams that have deployed neural-network-based design optimisation — as demonstrated by Xi’an Jiaotong University’s field reconstruction at 60,000 rpm and Khalifa University’s 600-variant optimisation study — can iterate through thousands of design variants at marginal computational cost. Organisations relying solely on traditional CFD workflows face a meaningful design-cycle disadvantage in this field.

CO₂ binary mixture working fluids represent an underpatented opportunity. Despite substantial academic investigation of CO₂/TiCl₄, CO₂/SO₂, CO₂/H₂S, and CO₂/C₆F₆ dopant systems — with reported efficiency improvements of 3–4% — no patents specifically claiming these fluid compositions in turbine applications appear in the retrieved dataset. This represents a potential whitespace for composition-of-matter or method-of-use patent filings by organisations active in CSP or geothermal sCO₂ applications, particularly in arid climates where dry-cooling compatibility is commercially critical.

Dynamic system modelling is the next capability frontier for CSP deployment. Two 2023 publications — from the University of Chinese Academy of Sciences on CSP system control strategy and from the National Technical University of Athens on year-long solar dynamic simulation — signal a shift from static design-point analysis toward fully dynamic system modelling. This capability will underpin bankable project designs for utility-scale sCO₂ CSP plants and is likely to become a differentiating competency for engineering firms pursuing project finance.