From the Feher Cycle to the Cluster Mesh: A History of Supercritical CO₂ Power

What Is the Supercritical (Feher) Cycle?

The user's question refers to the "Feyer cycle" — this is almost certainly the Feher cycle, named for engineer Ernest G. Feher, who first proposed using carbon dioxide above its critical point (31.1 °C / 73.8 bar) as the working fluid in a closed Brayton-type power cycle.

Above its critical point, CO₂ is neither a conventional gas nor a liquid — it is a dense, low-viscosity fluid that behaves like a liquid for pumping purposes and like a gas for expansion through a turbine. That combination lets a supercritical CO₂ (sCO₂) cycle extract more work per unit of turbomachinery size than a steam Rankine cycle or an air-based Brayton cycle, because the fluid is far denser at the compressor inlet, which sharply cuts compression work and shrinks every component in the loop.

Simplified closed-loop supercritical CO₂ Brayton cycle, the basic architecture proposed by Feher in 1967 and still used in today's sCO₂ turbine systems, including Infinity Turbine's Cluster Mesh cells.

Because the cycle is closed, the CO₂ never leaves the loop — the heat source (gas flame, waste heat, solar receiver, or reactor) is external to the working fluid, which is one reason the cycle has been studied for such a wide range of applications over the past six decades.

A Brief History of Supercritical CO₂ Power Production

The four archival documents reviewed for this article trace the sCO₂ cycle from a 1966 patent through a miniaturized 1968 Air Force turbine, a 1971 Navy submarine power system, and a 1976 NASA/GE study of full utility-scale plants — decades before the current wave of commercial interest.

1948

Early groundwork: partial-condensation CO₂ cycles

Sulzer Brothers explored a partial-condensation CO₂ Brayton concept, an early precursor to closed-loop CO₂ power cycles, though it did not operate in the supercritical regime that later defined the Feher cycle.

1966

Ernest G. Feher patents the "Supercritical Cycle Heat Engine"

U.S. Patent 3,237,403, issued March 1, 1966, describes a modified Rankine cycle using CO₂ above its critical point as the recirculating working fluid. This patent is cited directly as prior art in the 1971 Navy patent reviewed for this article (US 3,736,745), which credits Feher with providing "a nearly complete background for a thorough understanding" of supercritical thermal power systems.

1967

The cycle gets its name

Feher presents "The Supercritical Thermodynamic Power Cycle" (Douglas Paper No. 4348) to the Intersociety Energy Conversion Engineering Conference in Miami Beach, Florida — the paper that established what the industry now calls the Feher cycle: high thermal efficiency, compact components, no erosive phase-change through the turbine, and no cavitation risk at the pump because the fluid never leaves the supercritical/dense-liquid region.

1968

Air Force funds the first hardware — and it is tiny

Astropower Laboratory (McDonnell Douglas), under Feher himself, delivers AFAPL-TR-68-100, "Investigation of Supercritical (Feher) Cycle," for the Air Force Aero Propulsion Laboratory. The design: a Combined Rotating Unit pairing a single-stage axial turbine, a single-stage radial pump, and a solid-rotor alternator on self-acting (working-fluid-lubricated) bearings — 10 kWe net output at a design speed of 800,000 RPM. Even at that scale, the report projected a 0.70 turbine efficiency and 0.67 pump efficiency, evidence that the Feher cycle scales down to genuinely small, high-speed hardware without collapsing in efficiency.

1971–73

U.S. Navy adapts the cycle for undersea power

Horace E. Karig patents US 3,736,745, "Supercritical Thermal Power System Using Combustion Gases for Working Fluid," assigned to the Secretary of the Navy. It uses recirculated combustion gas (mostly supercritical CO₂) as the sole working fluid, explicitly citing Feher's 1966 patent as background, and claims roughly 42% overall thermal efficiency in a compact package suited to the high ambient pressures found at ocean depth — a use case where the small size and closed-cycle nature of supercritical CO₂ systems mattered as much as raw efficiency.

1976

NASA and GE test the cycle at utility scale

The Energy Conversion Alternatives Study (ECAS), a NASA/ERDA/NSF-sponsored General Electric report (NASA CR-134948), evaluates a supercritical CO₂ Brayton cycle side-by-side with advanced steam, closed-cycle gas turbine, liquid-metal-topping, MHD, and fuel-cell systems for large, coal-fired central power stations. The sCO₂ cycle lands in the 40–45% efficiency band with several other advanced cycles, and its estimated operating & maintenance cost (1.2 mills/kWh) is the lowest of the closed-cycle options studied. This is the moment the cycle was seriously considered for the "one huge turbine" model of a central power plant — and, for reasons of cost, materials, and the maturity of competing technologies at the time, it was not the concept the utility industry ultimately built.

1976–2000s

A long, quiet middle period

With no utility-scale demonstration funded, sCO₂ power research largely went dormant for roughly three decades. Turbomachinery materials, high-pressure seals, and compact heat-exchanger manufacturing had not yet caught up to what the 1968 and 1976 studies had already shown on paper.

2008

Sandia National Laboratories restarts the field

Sandia begins building a supercritical CO₂ Brayton test loop in stages: a simple compressor loop, then a heated, un-recuperated Brayton loop, then a full recompression loop — the modern research program that revived Feher's cycle for grid-scale use.

2022

First sCO₂ electricity delivered to a real grid

On April 12, 2022, Sandia engineers heated their closed-loop sCO₂ system to 600°F (about 315°C) and fed power to the Sandia–Kirtland Air Force Base grid for nearly an hour, peaking at 10 kW — a small number in absolute terms, but the first time a Feher-type cycle had ever been synchronized to a live electrical grid. The work was funded through the Department of Energy's Supercritical Transformational Electric Power (STEP) program.

2024–2026

Pilot-scale demonstration and commercial re-entry

Under the DOE STEP program, Southwest Research Institute designed and built a roughly 10 MWe sCO₂ pilot plant, with full-scale testing beginning in 2024. In parallel, rising natural-gas-fired sCO₂ interest — driven largely by AI data-center power demand and multi-year backlogs for large-frame gas turbines — has brought the cycle back as a commercial proposition, including modular natural-gas-fired systems such as Infinity Turbine's Cluster Mesh sCO₂ turbine generators.

Why this matters: every hardware data point in this timeline — the 1968 10 kWe turbine at 800,000 RPM, the 2022 Sandia 10 kW grid test — is a small machine. The Feher cycle's efficiency advantage over steam has never depended on building bigger; it comes from the density of the fluid itself. That is the technical premise behind numbering-up small sCO₂ cells instead of scaling up one large turbine.

The Conventional Baseline: One Huge Turbine

To put the small-cluster approach in context, it helps to see what a conventional large central power plant actually looks like. A 2019 engineering analysis of a typical 600 MW coal-fired supercritical steam plant (a different use of the word "supercritical" — here referring to steam pressure above water's critical point, not a CO₂ cycle) offers a useful real-world reference point for the "one huge turbine" model still dominant in utility power generation.

Typical 600 MW coal-fired supercritical steam unit

Main steam conditions: ~538°C at 16.7 MPa (a single, massive HP/IP/LP turbine train)
Boiler efficiency: ~88–91% depending on coal grade
Turbine cycle heat rate: ~8,000–9,300 kJ/kWh across the load range
Overall plant efficiency: ~34–41%, with the best efficiency only at near-100% load
Efficiency drops sharply below 40% load, and the plant cannot be built, moved, or resized incrementally

What "one huge turbine" implies

Single point of failure for hundreds of megawatts
Years of civil works, permitting, and construction before first power
Capacity is added in enormous, indivisible increments
Efficiency is optimized for one design point, not for partial loads
Represents the same "scale-up" logic later applied to the largest gas-fired and sCO₂ turbine concepts of the 1970s

Coal supercritical steam (600 MW unit)

~39%

Simple-cycle gas turbine

35–42%

Combined-cycle gas turbine

up to ~60%

sCO₂ Brayton cycle (theoretical ceiling)

~50%

Infinity Turbine IT1000 (design target, 500°C+)

40%+*

Approximate thermal efficiency ranges by technology. Coal-plant figures from the reviewed 2019 engineering study; combined-cycle and simple-cycle ranges are typical industry figures; sCO₂ Brayton ceiling per Sandia National Laboratories. *IT1000 figure is Infinity Turbine's stated design target, not an independently measured result.

Infinity Turbine's Cluster Mesh: Natural-Gas-Fired sCO₂ at Small Scale

Infinity Turbine's Cluster Mesh Supercritical CO₂ Turbine Generator system applies the Feher cycle's compactness advantage directly: instead of one large turbine, it deploys many small, modular sCO₂ turbine-generator cells that operate together — the company describes this as "number-up" instead of "scale-up." The flagship unit, the IT1000, is a 1 MW natural-gas-fired sCO₂ power block housed in a standard 20/40 ft hi-cube shipping container.

IT1000 — 1 MW modular sCO₂ power block

Closed-loop supercritical CO₂ Brayton cycle core
Turbine inlet temperature: 500–700°C
Fuel: natural gas, with optional solar-thermal or waste-heat input
Output: 1,000 kW nominal at 480 V AC, 3-phase
Design target: ~20% efficiency at 250°C rising to 40%+ at 500°C and above
Packaging: containerized or trailer-mounted for rapid, mobile deployment

How the cluster mesh scales

Add IT1000 (or IT250 / IT10MW) cells incrementally as demand grows
Multiple units paralleled for 5, 10, or larger MW-class sites
Individual cells can be serviced or replaced without taking the whole system offline
Because the working fluid is CO₂ running a transcritical cycle, the same hardware can also provide direct cooling — relevant for AI data centers that need both power and heat rejection

Scale-up vs. number-up: one large turbine concentrates capacity, lead time, and risk in a single unit; a cluster mesh spreads the same eventual capacity across many small, independently deployable and serviceable sCO₂ cells.

Small Cluster vs. One Huge Turbine: What Actually Changes

The case for numbering-up rather than scaling-up rests on more than nameplate efficiency. Below is a comparison of the two approaches for gas-fired prime power, drawing on Infinity Turbine's published Cluster Mesh positioning alongside the historical efficiency and cost figures reviewed above. Figures for the Cluster Mesh system are the manufacturer's stated design targets rather than independently audited performance data.

Attribute	One large gas turbine	Cluster Mesh sCO₂ (small cells)
Typical lead time to first power	5–6 years for large-frame (100–400 MW) machines	Weeks to a few months per cell
Deployment	Fixed site, extensive civil works, foundations, exhaust stacks	Containerized or trailer-mounted; minimal site prep
Scalability	Added in large, indivisible increments	Cell-by-cell, matched to actual demand growth
Redundancy	Single point of failure; N+1 means buying a whole extra unit	Inherent — individual cell outages are absorbed by the mesh
Noise	Often exceeds 100 dB near the machine; drives siting and permitting constraints	Closed-loop sCO₂ and compact turbomachinery are markedly quieter
Working fluid	Air (open cycle); low density, more compression work	Supercritical CO₂ (closed loop); dense fluid, less compression work
Cooling integration	Separate system and capital budget	Transcritical CO₂ cycle can double as a cooling system for co-located loads (e.g., data centers)
Efficiency at partial load	Optimized for one design point; degrades off that point	Multiple cells can be dispatched to match load, keeping each cell near its design point
Capital exposure to over-provisioning	Must size for future peak demand today	Capacity added incrementally as demand materializes

Large gas turbine (100–400 MW class)

5–6 yr

Small modular reactor

2030s+

Utility solar + battery

1–3 yr

Cluster Mesh sCO₂ cell

weeks–mo.

Approximate time from order to first power, by prime-power technology. Cluster Mesh and large-turbine figures reflect Infinity Turbine's published positioning and general industry procurement reporting for large-frame turbines circa 2024–2026.

The historical documents reviewed above make the underlying physics case: the Feher cycle's efficiency has never come from bulk. The 1968 Air Force turbine hit useful efficiency at 10 kWe and 800,000 RPM; the 1976 NASA/GE study found the same cycle competitive at utility scale. What has changed since is manufacturing — machining, high-speed bearings, compact heat exchangers, and power electronics have matured to the point where many small sCO₂ cells can be built as standardized, factory-produced units rather than one-off, site-built machines. That shift is what makes "number-up" a practical strategy today in a way it was not in 1976.

Outlook

Supercritical CO₂ power has now been studied, in some form, for nearly 60 years — from Feher's original patent and the Air Force's miniature 1968 turboalternator, through the Navy's undersea power system and NASA/GE's utility-scale evaluation, to Sandia's 2022 grid-connected demonstration and the Department of Energy's ongoing STEP pilot plant. What is new in 2026 is the demand signal: AI data centers need dispatchable, on-site power faster than the traditional large-turbine supply chain can deliver it. That is the opening modular, natural-gas-fired sCO₂ systems like Infinity Turbine's Cluster Mesh are built for — not by building a bigger version of the 1976 utility turbine, but by turning the same dense-fluid cycle into a small, standardized building block that can be numbered up as fast as the load grows.