Both chemistries run on hydrogen, but they suit different jobs in an aircraft. This page sizes each one from the twin for the three steps of the hydrogen-electric ladder, and sets the results against public roadmap figures. Every model number is a development result, not validated.
Fuel-cell explorer
See the system to scale
Cells solid and to scale, balance of plant in grey X-ray. Drag to turn, pinch or scroll to zoom.
Loading the 3D view
PEM plates
SOFC tubes
Balance of plant (placeholder)
Screening estimate from archee.api.system_estimate (also fctwin system). PEM: planar cells, 370 per stack at 1.34 mm, the Toyota Mirai cell thickness [mirai], 280 cm² active area ASSUMED on a plate 2.5 times larger, the Mirai's ratio [doe-mirai]. SOFC state-of-art class (default): a linear cell fitted to one published point, 1.4 W/cm² at 0.7 V and 750 °C (Jülich anode-supported cell), drawn as 2 mm outer-diameter micro-tubes, 50 mm active, 1 mm gap (micro-tubes of 1.9 mm have reached 1.1 W/cm² at 600 °C, AIST, Science 2009); an estimate, not a twin model. SOFC lab rig cell (switch): the twin’s own 7 mm tube (0.22 mm wall, 40 mm active, 9.07 cm²) at a pitch of its diameter plus the model’s 3 mm air gap. Both in bundles of 20 × 20 tubes; details in docs/SOFC_CHECK.md. BoP kW/kW: PEM from the twin's own compressor model (2.51 bar, stoichiometry 2, 68 %, plus 0.3 kW per stack); SOFC blower ASSUMED (stoichiometry 4, pressure ratio 1.1). PEM stack kW/kg: FlyZero 7 (2026) and 16 (2050), interpolated for 2030 and 2035; system kW/kg: FlyZero mid-range; BoP mass is the difference [flyzero]. SOFC: stack only [review][nasa]. Ground-level ambient; altitude raises compressor power.
Checked against the Toyota Mirai stack[mirai][doe-mirai][anl-mirai]: cells needed at its rated point (378 against 370), compressor power (16.3 against up to 15 kW) and stack volume agree. The twin’s lab cell gave 0.44 V at the Mirai’s 1.9 A/cm² against 0.67 V, so the PEM model was recalibrated: membrane resistance refitted to that one published point (Mirai-class, the default here; switch above). It is a one-point calibration, not a validation, and it over-predicts voltage near open circuit. Stack mass still uses FlyZero targets, 3.5–8 times the Mirai’s 2.0 kW/kg. Details: docs/MIRAI_BENCHMARK.md.
Benefits and assumptions
PEM · proton-exchange membrane
Benefits
Starts in under a minute and follows load, including transients [nasa]
Mature at automotive scale: Mirai-class cells give 0.78 W/cm² at 1 A/cm² (state-of-art SOFC class: 1.20 W/cm² at 0.8 V) [twin][mirai]
Published flight roadmap: FlyZero 2.0–2.5 kW/kg stack + BoP by 2030 [flyzero]
Flown in commuter demonstrators; FlyZero TRL 6 by 2026 [flyzero][review]
Assumptions in this twin
Mirai-class calibration: membrane resistance refitted to one public Toyota Mirai point (0.67 V at 1.9 A/cm²); other coefficients from the lab anchor
80 °C, 2.53 bar, RH 0.6, cathode stoichiometry 2; sized at 1.0 A/cm² (0.784 V, 63 % LHV)
Mass from FlyZero targets, 3.5–8 times the Mirai’s demonstrated 2.0 kW/kg
The frozen lab anchor still loses to a fixed 0.7 V baseline on measured lab data (120.6 vs 34.2 mV error)
SOFC · solid oxide
Benefits
Higher efficiency: 64 % against 63 % (LHV), so 2 % less hydrogen for the same power [twin]
Less heat to reject, and at 600–1,000 °C it is high-grade: usable by a turbine or heat exchanger [twin][review]
Fuel flexible and tolerant of CO and other contaminants [nasa][review]
Steady duty suits APU, ground power and turbine-hybrid roles [nasa][review]
Assumptions in this twin
State-of-art class (default): linear cell fitted to one published point, 1.4 W/cm² at 0.7 V and 750 °C (Jülich), packaged as 2 mm micro-tubes; an estimate, not a twin model
Lab rig cell (switch): the twin’s frozen 7 mm tubular cell at 800 °C, 50 mL/min fuel, 0.24 W/cm² at 0.8 V; fuel-limited below about 0.6 V
Sized at 0.8 V (1.50 A/cm² in the selected default)
No mass model and no flight target: published figures span 0.51 kW/kg (cell) to a claimed 2.5
Start-up (tens of minutes to hours), thermal cycling and altitude not modelled; acceptance gate blocked
The use case
Scaling up step by step
Step 1 · by 2030Business + taxi, 4–9 seats350–600 kW
Step 2 · 2030sRegional, 9–100 seats600–2,000+ kW
Step 3 · 2040sLarger single-aisle4+ MW
Ladder as presented by H2FLY [h2fly]. Sizing below uses the top of each power band.
From the twin
What one cell delivers
Per square centimetre, a state-of-art SOFC cell (estimated from a published 1.4 W/cm² cell) gives 1.20 W/cm² at 0.8 V, more than the Mirai-class PEM cell's 1.41 W/cm² peak. The twin's own lab rig SOFC cell reaches only 0.35 W/cm²: it is a 7 mm research tube on a small fuel supply, not a stack cell.
PEM Mirai-class 80 °CPEM lab cell 75 °C (frozen anchor)SOFC state-of-art class 750 °C (estimated)SOFC rig cell 800 °CSOFC rig cell 700 °C Twin model outputs, not validated. PEM Mirai-class: membrane resistance refitted to a public Toyota Mirai point, 2.53 bar; lab cell: frozen anchor, 2.51 bar; both RH 0.6, stoichiometry 2. SOFC: frozen v1 tubular cell, 3 % steam, 50 mL/min. The SOFC curves bend down near 0.6 A/cm² because the rig fuel flow runs out (90 % utilisation at 0.5 V); a stack fed more fuel would go further.
Sized for each step
Same power, different machines
Each stack is sized for the net electrical power of the step: PEM at 1.0 A/cm², SOFC at 0.8 V. Which cell you assume changes the answer. At 2,000 kW: Mirai-class PEM vs state-of-art SOFC: hydrogen 96 against 94 kg/h, SOFC area 0.7× PEM; Mirai-class PEM vs the rig SOFC cell: hydrogen 96 against 94 kg/h, SOFC area 3.2× PEM; lab PEM cell vs state-of-art SOFC: hydrogen 115 against 94 kg/h, SOFC area 0.5× PEM; both lab cells: hydrogen 115 against 94 kg/h, SOFC area 2.7× PEM.
Step 1 · by 2030Business + taxi, 4–9 seats · sized at 600 kW
Step 3 · 2040sLarger single-aisle · sized at 4,000 kW
PEM
SOFC
PEM
SOFC
PEM
SOFC
Operating point
0.656 V · 1.00 A/cm²
0.800 V · 0.30 A/cm²
0.656 V · 1.00 A/cm²
0.800 V · 0.30 A/cm²
0.656 V · 1.00 A/cm²
0.800 V · 0.30 A/cm²
Active area
91.4 m²
247 m²
305 m²
825 m²
610 m²
1,649 m²
Hydrogen at full power
34.4 kg/h
28.2 kg/h
115 kg/h
94 kg/h
229 kg/h
188 kg/h
Heat to reject
546 kW at ~75 °C
340 kW at ~800 °C
1,820 kW at ~75 °C
1,133 kW at ~800 °C
3,639 kW at ~75 °C
2,266 kW at ~800 °C
Stack + BoP mass
240 – 300 kg (FlyZero 2030 target)
No published target
667 – 800 kg (FlyZero 2035 target)
No published target
1,143 – 1,333 kg (FlyZero 2050 target)
No published target
Screening estimate from archee.api.size_stack: stack only, 100 % hydrogen utilisation, LHV heat basis, no compressor, converter or balance-of-plant losses. PEM mass uses FlyZero stack + BoP targets [flyzero]; there is no comparable SOFC target.
Try it
Size your own case
Computed in your browser with the same code as the web API (/api/v1/size); SOFC current density is interpolated from the twin's 800 °C curve.
Beyond the curves
How they differ in an aircraft
PEM
SOFC
Source
Operating temperature
65–95 °C today; FlyZero target 80 °C (2026) → 100 °C (2035)
600–1,000 °C
[nasa][flyzero][review]
Start-up
Under 1 minute
Tens of minutes to hours (ceramic stack must be heated)
[nasa]
Load following
Base load and transients
Base load only: the ceramic electrolyte limits transient response
Cell level 0.51 kW/kg as tabulated in a 2025 review; NASA claims up to 2.5 kW/kg for an advanced design; no FlyZero target
[flyzero][h2gear][zeroavia][review]
Heat
Low-grade, near 80 °C, with stack efficiency typically 40–50 %: large radiators or vapour-compression cooling; HT-PEM (PBI, 30–90 °C hotter) eases this
High-grade, hundreds of °C: usable by a turbine or heat exchanger, but complex thermal management
[flyzero][h2gear][review]
Fuel and contaminants
Needs clean hydrogen; CO poisons platinum (HT-PEM is more tolerant)
Many fuels (CO is a fuel); tolerant of contaminants
FlyZero: TRL 6 by 2026; flown in commuter demonstrators
Low TRL for flight; no flight role identified in the 2025 review
[flyzero][review]
Natural aviation role
Main propulsion (Steps 1–3)
APU, ground power and turbine-hybrid concepts; long, steady duty
[nasa][review]
The crossover question
What a ground deployment proves
Ground markets can retire some risks before flight. They cannot prove power per kilogram, altitude, vibration or certification, which must be funded as flight-specific work.
Risk
PEM ground route
SOFC ground route
Carries over to flight
Stack durability under real hours
Backup and ground power, buses
Data-centre and distributed power (steady duty suits SOFC)
Partial: flight duty cycle differs
Manufacturing cost, yield and UK supply chain
Volume from ground markets
Volume from data-centre orders
High
Controls, health monitoring and AI
Fleet data to train and validate models
Fleet data to train and validate models
High
Thermal management
Low-grade heat rejection at scale
High-grade heat recovery and integration
Partial: ram air and altitude differ
kW/kg, altitude, vibration, certification
Not proved on the ground
Not proved on the ground
None: fund separately
Same aircraft, third energy store
Where the battery fits
Neither cell is flown alone. A fuel-cell regional aircraft also carries a battery buffer for start-up and transients, which the battery twin can size.
Battery role in a fuel-cell aircraft
Published requirement
Fuel-cell start-up and transients (regional)
50 kWh, 200 Wh/kg, peak above 5C [ati-battery]
Full electrification, regional / single-aisle
Above 1,000 / 1,500 Wh/kg [ati-battery]
Scalability of fuel-cell propulsion with battery hybridisation
Still unclear; would need much larger batteries [ati-battery]
Both models are frozen development candidates. The PEM model loses to a fixed 0.7 V baseline on measured data; the SOFC acceptance gate is blocked. See the evidence board.
No mass model: specific power comes from published targets and claims, not from the twin.
SOFC thermal cycling, start-up energy and altitude effects are not modelled.
Published figures are quoted from public sources only.