Hydrogen Marine Fuel Cells: Types, Projects, and Rules

Hydrogen marine fuel cells convert hydrogen (or a hydrogen-carrier fuel) directly to electricity through an electrochemical reaction, producing water as the only direct emission. The technology has moved from laboratory demonstration to initial commercial service between 2008 and 2026, with proton-exchange membrane (PEM) systems now operating on passenger ferries in Norway, San Francisco, and several European routes. This article covers how the major fuel cell architectures work, how onboard hydrogen is stored and bunkered, what the regulatory framework requires under the IGF Code and IMO MSC.1/Circ.1647, which vessels are operating or under development, and where the technology’s current limits sit.

The companion article Hydrogen as Marine Fuel covers the broader hydrogen propulsion picture including internal combustion engine adaptations, well-to-wake intensity, and production pathways. This article focuses on the fuel cell pathway specifically.

How a hydrogen fuel cell works

A fuel cell is an electrochemical device, not a combustion engine. Hydrogen (H2) is fed to the anode; oxygen from air is fed to the cathode. At the anode, hydrogen molecules are split into protons (H+) and electrons. The electrons flow through an external circuit, producing useful electrical current. The protons migrate through a selective electrolyte membrane to the cathode, where they combine with oxygen and the returning electrons to form water.

The net reaction is: 2H2 + O2 → 2H2O, releasing energy at an efficiency that exceeds any heat engine operating on the same fuels. No combustion occurs, so there are no nitrogen oxide (NOx), sulfur oxide (SOx), or particulate emissions. CO2 emissions at the stack are also zero. The well-to-wake CO2 intensity depends entirely on how the hydrogen was produced; green hydrogen from renewable electrolysis gives a well-to-wake intensity of approximately 5 to 10 g-CO2eq/MJ, lower than any combustion-based marine fuel.

A single fuel cell produces roughly 0.6 to 0.9 volts at load. Ship-scale power requires stacking hundreds or thousands of cells into a fuel cell stack, and multiple stacks into a fuel cell system. The electrical output is DC; an inverter converts it to AC for ship distribution or the DC feeds a bus that drives permanent-magnet propulsion motors.

Proton-exchange membrane fuel cells

PEM fuel cells use a thin solid polymer membrane, typically Nafion or a similar perfluorosulfonic acid material, as the electrolyte. The membrane conducts protons but not electrons. Operating temperature is 60 to 80 degrees Celsius for standard PEM, with high-temperature PEM variants reaching 120 to 180 degrees Celsius using phosphoric acid-doped membranes.

Marine PEM stacks operate at electrical efficiency of 50 to 60 percent (LHV basis), with efficiency declining at low and very high current densities. Start-up from cold takes minutes, not hours. Load response is fast enough to follow propulsion demand directly, though a battery buffer is standard practice to handle peak loads and protect stack membrane lifetime.

PEM cells require high-purity hydrogen, typically above 99.97 percent H2 (ISO 14687 Type D), because carbon monoxide at even a few parts per million poisons the platinum catalyst on the anode. This rules out direct use of reformer off-gas without extensive purification. Power density is high, typically 0.5 to 1.5 kW per kilogram of stack, which keeps the installation compact compared to other fuel cell types.

Ballard Power Systems (Canada) is the leading commercial marine PEM supplier, with stacks installed on MF Hydra, Sea Change, the FLAGSHIPS project vessels, and multiple harbour ferry pilots. Cummins Accelera (formerly Hydrogenics) and Toyota also supply marine-qualified PEM stacks, with Toyota’s marine FC module derived from its Mirai automotive stack. Powercell Sweden AB supplies stacks to several European OEM integrators. In September 2024, Corvus Energy’s Pelican Fuel Cell System, which uses Toyota PEM modules, received DNV type approval as the first “inherently gas-safe” marine fuel cell system, using nitrogen inerting of the fuel cell space rather than relying solely on ventilation.

Solid oxide fuel cells

Solid oxide fuel cells use a ceramic electrolyte, typically yttria-stabilized zirconia (YSZ), that conducts oxygen ions at high temperature. Operating temperature is 700 to 900 degrees Celsius for most commercial designs. At that temperature, the fuel can be reformed internally: methane, methanol, or ammonia-cracker output can all feed an SOFC without a separate pre-reformer, giving fuel flexibility that PEM cannot match.

Electrical efficiency on pure hydrogen is 55 to 65 percent (LHV basis). In combined heat-and-power (CHP) mode, where the high exhaust temperature is recovered for heating or further power generation, total energy utilisation reaches 70 to 85 percent. The trade-off is thermal cycling: heating an SOFC from cold takes four to eight hours and rapid temperature cycling degrades the ceramic electrolyte, so SOFCs suit continuous-operation propulsion or auxiliary power rather than intermittent duty.

The marine SOFC track record begins with the FellowSHIP project: DNV, Wartsila, Eidesvik, and Norsk Hydro collaborated from 2003 to deploy a 320 kW SOFC auxiliary power unit on the offshore supply vessel Viking Lady in 2009. That system, built by Wärtsilä, ran on natural gas for several years of commercial operation and demonstrated that SOFC technology could meet class society requirements in a real sea-going environment.

Bloom Energy, Ceres Power (UK, with Samsung Heavy Industries as licensee), Mitsubishi Power, and AVL have SOFC programs with varying degrees of marine relevance. Ceres Power’s steel-cell SOFC technology, which operates around 550 to 620 degrees Celsius and tolerates faster thermal cycling than YSZ-based designs, is under development for marine auxiliary power with a Samsung Heavy Industries programme targeting bulk carriers.

Molten carbonate fuel cells

Molten carbonate fuel cells (MCFCs) use a liquid carbonate salt electrolyte retained in a ceramic matrix, operating at approximately 650 degrees Celsius. MCFCs tolerate CO2 in the fuel stream and can reform natural gas or ammonia internally after cracking. Unlike PEM cells, MCFCs do not require precious metal catalysts, reducing material cost at large scale.

FuelCell Energy (USA) is the principal commercial MCFC supplier. Their SureSource marine units have been evaluated for ship auxiliary power, particularly on cruise vessels where waste heat recovery from the MCFC exhaust can supply hotel loads. The MCFC’s large physical footprint and weight limit practical marine application to large ships where the volume is available.

One notable MCFC characteristic: the cathode consumes CO2 from the inlet air and concentrates it in the anode exhaust, meaning an MCFC system can simultaneously generate power and capture CO2 from the ship’s own engine exhaust if configured as a carbon capture device. FuelCell Energy has worked with ship operators to evaluate this dual mode, though no commercial marine carbon capture installation was in service by mid-2026.

PEM vs SOFC vs MCFC: a comparison

Hydrogen storage and fuel system architectures

The hydrogen storage challenge is as consequential as the fuel cell stack choice. Hydrogen has a mass-specific energy content of approximately 120 MJ/kg (LHV), roughly 2.7 times that of heavy fuel oil, but its volumetric energy density is extremely low without compression or liquefaction.

Compressed gaseous hydrogen

Compressed hydrogen at 350 to 700 bar is stored in high-pressure composite-wound cylinders, typically carbon-fibre-reinforced plastic (CFRP) type IV tanks. Volumetric energy density at 700 bar is approximately 4.5 MJ/litre, compared to approximately 35 MJ/litre for VLSFO. A vessel requiring 1,000 kg of hydrogen storage at 700 bar therefore needs roughly 7,800 litres of tank volume before structural volume allowance.

The compressed gas pathway suits short-range ferries with predictable refuelling at fixed berths, typically round-trip routes under 50 nautical miles. Harbor tugs, pilot boats, and river ferries in this category avoid the cryogenic complexity of liquid hydrogen. The Sea Change hydrogen ferry operating in San Francisco Bay uses compressed hydrogen at 350 bar in roof-mounted cylinders.

Liquid hydrogen

Liquid hydrogen (LH2) at minus 253 degrees Celsius (20 K) and near-atmospheric pressure has a volumetric energy density of approximately 8.5 MJ/litre, nearly double compressed gas, still only about one-quarter of VLSFO. The cryogenic storage system requires vacuum-insulated tanks, cold-box pipework, boil-off gas management, and compatible materials throughout (stainless steel, aluminium, or specific nickel alloys; carbon steel becomes brittle at cryogenic temperatures).

Boil-off is a persistent challenge. Insulation is never perfect, and hydrogen boils at minus 253 degrees Celsius, meaning any heat ingress vaporises liquid and builds tank pressure. A practical marine LH2 system must either vent boil-off safely (requiring dispersion analysis per MSC.1/Circ.1647), reliquify it onboard (energy-intensive and mechanically complex), or ensure operational patterns keep the tank pressure within acceptable bounds.

The Norled MF Hydra carries approximately 200 kilograms of liquid hydrogen in a deck-mounted cryogenic tank, bunkered by road tanker at the Hjelmeland terminal. The ferry entered regular service on the Hjelmeland-Skipavik-Nesvik route in western Norway in April 2023. MF Hydra was built by Westcon Yards and is the world’s first liquid-hydrogen-powered passenger ferry in commercial scheduled service.

Hydrogen carriers: ammonia and LOHC

Ammonia (NH3) can function as a hydrogen carrier rather than a direct fuel when paired with an onboard cracker that converts NH3 back to N2 + H2. Ammonia is liquid at minus 33 degrees Celsius at atmospheric pressure, far easier to store and handle than cryogenic hydrogen, and existing ammonia bunker infrastructure exists at many ports from the fertiliser trade. The cracked hydrogen is then fed to a PEM or MCFC stack after purification.

The cracking energy penalty is approximately 12 to 15 percent of the hydrogen’s LHV content. Several projects were evaluating ammonia-to-hydrogen-to-fuel-cell pathways as of 2025, though none had entered scheduled commercial service. DNV’s rules and IMO’s ammonia guidelines would apply to the ammonia storage portion; MSC.1/Circ.1647 applies to the fuel cell portion.

Liquid organic hydrogen carriers (LOHC) bind hydrogen chemically to an organic liquid, typically dibenzyltoluene (DBT) or toluene/methylcyclohexane, for transport and storage at ambient temperature and pressure. Hydrogen is released by catalytic dehydrogenation, consuming substantial heat. LOHC has not yet reached marine commercial deployment but is being studied by Hydrogenious LOHC Technologies and partners.

Bunkering infrastructure and port readiness

Hydrogen bunkering at scale remains one of the sharpest infrastructure gaps in the alternative-fuels portfolio. As of 2025, dedicated marine hydrogen bunkering facilities exist at a handful of locations in Norway, Japan, and California, all associated with specific ferry pilot programmes. None of these serves general commercial shipping.

Norway’s Hjelmeland terminal, operated by Norled and its hydrogen supplier, was built specifically for MF Hydra and handles liquid hydrogen deliveries by road tanker. The bunkering flow rate is slow by marine standards: a 200-kilogram fill for MF Hydra takes approximately one to two hours. For larger vessels with tonne-scale hydrogen requirements, the bunkering time under current infrastructure would extend to eight to twelve hours, comparable to LNG bunkering for medium-sized vessels but without LNG’s established safety protocols and trained workforce.

Port hydrogen supply chains face a cost structure driven by small volumes. Green hydrogen in 2024 reached the bunkering facility at EUR 8 to 15 per kilogram, well above production cost, because distribution by cryogenic road tanker adds EUR 2 to 6 per kilogram over 50 to 300-kilometre distances. Scaling to pipeline supply requires dedicated hydrogen port infrastructure with investment levels that no port authority had committed to for marine bunkering by mid-2026.

The IMO’s interim guidelines for hydrogen as fuel (from CCC 11, 2024) include requirements for bunkering operations, covering hose specifications, communications between vessel and bunker supplier, gas monitoring during bunkering, and emergency shutdown coordination. These requirements broadly follow the LNG bunkering framework in ISO 20519:2017, adapted for hydrogen’s physical properties.

Hydrogen bunkering standards and the infrastructure challenge

The liquefaction energy penalty

Liquefying hydrogen is expensive in energy terms. Compression of hydrogen to 700 bar consumes approximately 1.3 kWh per kilogram; liquefaction to minus 253 degrees Celsius consumes approximately 10 to 12 kWh per kilogram, roughly eight to nine times as much energy per kilogram as compression. At a grid electricity cost of EUR 0.10 per kWh, liquefaction alone adds EUR 1.00 to 1.20 per kilogram of hydrogen to the bunkering cost before accounting for capital, maintenance, or boil-off losses during storage and transfer.

Centralised liquefaction at a large industrial facility (50 to 100 tonne per day capacity) brings the liquefaction cost down to approximately USD 2.75 per kilogram of hydrogen, based on studies published in the IEA’s hydrogen cost literature. Small-scale or on-site liquefiers at individual port facilities are materially less efficient and cost three to five times more per kilogram. This means that practical LH2 bunkering for marine use depends on access to a large-scale centralized liquefaction and distribution network, which has not been built at any commercial shipping port as of mid-2026.

Compressed gaseous hydrogen avoids liquefaction but trades that energy saving for severe volumetric limitations. High-pressure tube trailers at 200 to 350 bar carry approximately 300 to 900 kilograms per trailer, meaning a vessel requiring several tonnes of hydrogen per voyage needs multiple trailer deliveries per bunkering call.

Standards for LH2 bunkering at port

As of 2024, no binding international standard exists specifically for liquefied hydrogen bunkering of ships. The Maritime Technologies Forum (MTF), a coalition of flag states and class societies, published its “Guidelines for the Development of Liquefied Hydrogen Bunkering Systems and Procedures” in August 2024. DNV, as a contributing member, stated that the MTF guidelines and their submission to IMO were critical steps in addressing the safety gap around LH2 bunkering. The document addresses pre-bunkering hazard identification, emergency shutdown system design, boil-off management during transfer, and crew competence requirements for LH2 handling.

On the standards side, CEN Workshop Agreement CWA 18157:2024, published by the European Committee for Standardization, covers hydrogen refuelling for marine vessels and sets out technical requirements including limits on filling speed based on hydrogen temperature and tank pressure. The agreement states that pressure in the ship’s hydrogen tanks must not exceed 1.25 times nominal working pressure during transfer, and specifies minimum communication protocols between the shore bunkering unit and the vessel’s fuel management system.

These guidelines are not yet equivalent to the established LNG bunkering framework. ISO 20519, which governs LNG bunkering of ships and has been in force since 2017, reflects a decade of operational experience, industry-wide incident reporting, and classification society input at a level that the LH2 standards community has not yet reached. IMO is developing mandatory LH2 bunkering requirements through the CCC sub-committee, but adopted interim guidance rather than mandatory instruments through the 2024-2025 period.

Port capital requirements and scale barriers

A port-scale LH2 bunkering facility capable of serving even a small hydrogen ferry fleet needs a dedicated cryogenic storage tank (typically 500 to 2,000 m³), pre-cooled transfer hoses, a safe-fill monitoring system, and an inert-gas purge system. Industry cost estimates for a facility at that scale start at USD 50 million to 100 million for storage and transfer infrastructure alone, before land, grid connection, or liquefier costs. No commercial port operator had committed public capital to a general-access LH2 marine bunkering facility as of mid-2026; all existing installations are purpose-built for single operators or ferry routes.

The contrast with LNG is instructive. Shell, TotalEnergies, and Gasum had commissioned 24 dedicated marine LNG bunkering vessels and 50-plus port facilities globally by 2024, backed by long-term supply contracts with shipowners who ordered LNG dual-fuel ships in volume. Hydrogen lacks both the vessel fleet (fewer than 20 hydrogen fuel cell ships in commercial service worldwide in 2025) and the multi-year supply contracts that attract infrastructure capital.

Integration with the ship power system

DC bus architecture and power conversion

A fuel cell stack produces variable-voltage DC output. At rated power, a typical marine PEM module outputs 600 to 900 V DC at the stack terminals; voltage drops as load increases and rises at light load. The standard integration approach uses a DC/DC boost converter between the stack and a stabilised DC bus, typically operating at 700 to 800 V for medium-size ferry applications. Multiple stacks connect in parallel through individual converters to the same bus.

From the DC bus, one or more DC/AC inverters drive propulsion motors and, through a separate transformer, supply the vessel’s 400 V or 690 V AC auxiliary distribution network. This fully electric architecture is already familiar from battery-electric ferry designs; the fuel cell replaces or supplements the shore-charged battery as the primary energy source.

The SF-BREEZE concept ferry developed for the US Department of Transportation (a 150-passenger San Francisco Bay vessel study) specified 41 PEM modules arranged in four 30 kW stacks each, feeding a 3,200 V DC main bus through individual DC/DC converters, driving two 2,000 kW AC motor waterjet propulsors via three-phase inverters. That level of modular parallelism was chosen to allow individual module bypass without loss of propulsion capability.

Battery hybridisation and load following

Fuel cells respond to load changes more slowly than diesel engines or batteries. A PEM stack changes output at roughly 10 to 20 percent of rated power per second when unconstrained, but driving the stack to follow rapid load transients directly accelerates membrane degradation by cycling the membrane humidity and current density. Marine operational profiles include fast transients: high current demand during berthing, acceleration, and anchoring; low current during slow steaming; zero current at layover.

Battery energy storage, typically lithium iron phosphate (LFP) or NMC chemistry in the 200 to 500 kWh range for a ferry application, buffers these transients. The energy management system (EMS) holds the fuel cell at a steady mid-range operating point while the battery absorbs peaks and troughs. DC bus voltage is kept within plus or minus 5 percent of nominal by the battery’s bidirectional converter acting as bus former. MF Hydra’s 330 kWh battery buffer relative to its 200 kW nominal fuel cell output is sized to handle harbour manoeuvring with fuel cells at steady state.

The dual-source architecture also provides a natural redundancy layer. If a fuel cell module trips offline, the battery sustains propulsion long enough for an EMS response or operator intervention. The 200 kW Ballard FCwave installation on MF Hydra is split across two modules; each can operate independently, so a single module failure leaves the vessel with 100 kW, enough to complete a ferry crossing at reduced speed.

SOLAS safe-return-to-port implications

SOLAS Chapter II-1, Regulation 55 (equivalent design and arrangement) and the safe-return-to-port provisions introduced in amendments adopted 2006 (in force 1 July 2010) apply to new passenger ships of 120 metres length or more with three or more main vertical zones. The regulation requires that the essential systems, including propulsion, remain operational after a fire or flooding casualty within any single compartment and that the vessel can proceed to a port of refuge under its own power.

MSC.1/Circ.1647 addresses the interaction between fuel cell systems and these safe-return requirements. For a passenger ship with a hydrogen fuel cell propulsion system, the guidelines require that the fuel cell installation is separated into at least two independent zones, each capable of sustaining minimum safe-transit speed after a casualty in the other zone. This parallels the redundant main-engine or propulsion-generator requirements in a conventional diesel-electric ferry. Where a hydrogen fuel cell system relies on a single fuel cell room, the flag state may require demonstration by equivalent analysis that the casualty scenario still leaves the vessel mobile.

Draft SOLAS amendments discussed at MSC in 2023 clarified that single-fuel installations, which includes hydrogen fuel cell ships that cannot switch to an alternative fuel, must demonstrate propulsion redundancy by other means: typically by segregating fuel cell modules across fire zones and providing an independently powered emergency propulsion unit. These amendments were targeted for entry into force 1 January 2026. Flag state surveyors applying the Red Ensign Group’s safe-return-to-port survey guidance note that fuel-cell-propelled passenger ships require explicit fuel supply isolation valve logic integrated into the damage-control plan.

Insulation monitoring and DC electrical safety

A large DC bus carries fault energy orders of magnitude higher than a standard AC distribution panel. MSC.1/Circ.1647 requires continuous insulation monitoring (IMD) of the fuel cell DC output and the main DC bus. An earth fault on an ungrounded IT DC bus doesn’t trip the system immediately, but a second earth fault in a different polarity creates a direct short. The circular requires audible and visual alarms on first fault detection, automatic reduction of fuel cell output to safe standby levels on second fault confirmation, and interlocking with the hydrogen supply shutoff valve.

Equipment in the fuel cell room and any adjacent gas-risk zone must be rated IEC 60079 for Zone 1 (gas explosion protection), explosion group IIC (hydrogen), temperature class T1 (450 °C auto-ignition temperature of hydrogen). This is the same electrical classification required for oxygen-enriched atmospheres in tanker cargo handling spaces, and the hardware is available, but the classification cost adds to the overall system price.

Well-to-wake efficiency: fuel cells, battery-electric, and hydrogen-ICE compared

The three hydrogen propulsion pathways

Hydrogen reaches a ship through three distinct propulsion pathways, each with different energy chain efficiency. The fuel cell pathway converts hydrogen to electricity electrochemically. The hydrogen internal combustion engine (H-ICE) pathway burns hydrogen in a modified two- or four-stroke engine. The indirect battery-electric pathway uses hydrogen as a grid-scale storage medium, regenerating electricity ashore to charge the ship’s batteries. Each pathway’s efficiency matters because hydrogen is an energy carrier, not a primary energy source: every conversion step multiplies the upstream renewable electricity demand.

Fuel cell pathway efficiency

A PEM fuel cell converts hydrogen to electricity at 50 to 60 percent electrical efficiency (LHV basis) at optimal load. After the DC/DC converter (98 percent), the inverter (97 percent), and the propulsion motor (95 to 97 percent), the full tank-to-propeller efficiency is approximately 0.55 × 0.98 × 0.97 × 0.96 = 0.50, or 50 percent. Against a diesel-electric baseline of roughly 38 percent tank-to-propeller efficiency, the fuel cell provides a meaningful thermodynamic advantage per unit of fuel carried.

For well-to-wake analysis using green hydrogen (electrolysis efficiency 65 to 70 percent at the electrolyser level), compression to 700 bar or liquefaction reduces chain efficiency further. Compression at 700 bar consumes about 1.3 kWh/kg, which relative to hydrogen’s 33.3 kWh/kg LHV represents a 4 percent energy deduction. Liquefaction consumes 10 to 12 kWh/kg, a 30 to 36 percent energy deduction before the hydrogen reaches the ship’s tank. The total well-to-propeller efficiency for green compressed hydrogen through a PEM fuel cell is approximately 0.68 × 0.96 × 0.50 = 0.33, or 33 percent of the original renewable electricity input. For green liquid hydrogen the figure drops to roughly 0.68 × 0.64 × 0.50 = 0.22, or 22 percent.

Battery-electric pathway efficiency

Battery charging from shore (assumed grid: 95 percent charger efficiency, 99 percent battery round-trip, 97 percent motor, 96 percent drivetrain) gives a tank-to-propeller efficiency of approximately 0.95 × 0.99 × 0.97 × 0.96 = 0.87, or 87 percent. If the shore power is from the same renewable source as the green hydrogen electrolyser, the battery-electric pathway uses roughly 2.6 to 4 times less electricity per nautical mile than compressed hydrogen through a fuel cell, and 5 to 6 times less than liquid hydrogen.

This arithmetic advantage is why battery-electric ferries have expanded much faster than hydrogen fuel cell ferries on short routes where daily charging is practical. The limitation is energy density: current lithium-ion batteries store approximately 0.25 to 0.30 kWh per kilogram (cell level), versus hydrogen’s 33.3 kWh/kg LHV. A battery-electric vessel on a long route with no intermediate charging would need a battery mass that is structurally and commercially prohibitive. For routes exceeding 80 to 120 nautical miles between charge points, liquid hydrogen’s volumetric advantage over batteries begins to shift the calculation, despite the efficiency penalty.

Hydrogen-ICE comparison

A hydrogen internal combustion engine burns hydrogen in a conventional combustion cycle. Thermal efficiency for a large two-stroke marine H-ICE is approximately 42 to 49 percent at optimal load, slightly higher than a diesel on the same cycle because hydrogen’s high flame speed allows leaner combustion. Published hybrid configurations combining H-ICE with a fuel cell show an average system efficiency of approximately 49 to 54 percent, close to a fuel cell alone, but the H-ICE produces NOx emissions because combustion creates nitrogen-oxygen bonding at high cylinder temperatures.

H-ICE is therefore not a zero-emission pathway at the stack; NOx aftertreatment (selective catalytic reduction, SCR) is required under MARPOL Annex VI Tier III to meet the 3.4 g/kWh NOx limit in emission control areas. The fuel cell pathway produces zero NOx, zero SOx, and zero particulates at the stack, which simplifies exhaust treatment and satisfies future zero-emission berth requirements already legislated by California Air Resources Board (CARB) for vessels calling at California ports.

Summary comparison table

Cost structure and commercial outlook

Fuel cell system capital cost

Marine-qualified PEM fuel cell systems carry a capital cost that reflects both low production volumes and the marine certification premium. Stack-level costs for large marine PEM systems were estimated at USD 500 to 1,200 per kW in 2023 to 2025 publications, but the total installed system cost, including balance of plant, DC/DC converters, cooling, hydrogen fuel processing, and system integration, runs USD 2,400 to 3,400 per kW. By comparison, a marine diesel generator of equivalent output costs USD 300 to 700 per kW installed.

Ballard has stated publicly that it targets a fuel cell system cost of approximately USD 1,200 per kW for bus-scale production volumes (defined as 25,000 hours stack lifetime at 100 systems per year production rate). At marine volumes, which reached fewer than 50 MW of installed capacity globally in 2025, the cost trajectory is far slower than bus or automotive applications. DOE analysis published in 2023 found that at 50,000 systems per year production, PEM fuel cell system cost could fall to approximately USD 1,650 per kW, but maritime demand cannot approach that volume unilaterally.

SOFC systems for marine use (Bloom Energy’s marine module received ABS type approval in 2024) carry a similar or higher cost per kW at current volumes, with the additional burden that high-temperature thermal cycling accelerates stack aging and requires more careful operational management than PEM.

Fuel cost: the green hydrogen premium

The production cost of green hydrogen via proton-exchange membrane electrolysis from renewable electricity was EUR 4 to 8 per kilogram at the production facility in 2024 to 2025, with European projects in high-renewable-fraction grids achieving the lower end of that range. Grey hydrogen from natural gas SMR without carbon capture costs EUR 1.5 to 2.5 per kilogram. At 120 MJ/kg LHV, green hydrogen at EUR 6/kg equates to an energy cost of approximately EUR 50 per GJ. VLSFO at 2024 average bunker prices of roughly USD 600 per tonne (approximately EUR 550 per tonne), with 40 MJ/kg LHV, equates to approximately EUR 13.75 per GJ. The green hydrogen energy cost is therefore approximately 3.6 times higher per unit of energy at the production gate, before distribution and bunkering margin.

The EU FuelEU Maritime GHG intensity compliance framework, in force from 1 January 2025, creates a mechanism to close part of this gap for ships using green hydrogen. Renewable hydrogen qualifying as an RFNBO under the EU RFNBO rules receives a 2x multiplier credit toward the FuelEU GHG compliance target between 2025 and 2034. That multiplier reduces the effective compliance cost for operators using green hydrogen relative to operators using fossil fuels, but it doesn’t change the physical energy cost.

Stack lifetime and replacement cost

A PEM fuel cell stack degrades primarily through membrane thinning, platinum catalyst dissolution, and water management degradation. Ballard targets 25,000 hours stack lifetime for transit applications (buses, ferries); marine duty cycles are in a similar range. At a 5,000 to 10,000-hour overhaul interval and 25,000-hour replacement interval, a MF Hydra-scale installation running 4,000 hours per year faces stack replacement roughly every six years. Published industry estimates suggest stack replacement at this scale costs approximately USD 200 to 400 per kW, meaning a 200 kW installation costs USD 40,000 to 80,000 per replacement cycle. This is broadly comparable to a diesel engine’s major overhaul cost, but diesel engines typically run 60,000 to 80,000 hours between overhauls at marine duty.

Barriers to commercial scale

Four barriers dominate the commercial outlook through the late 2020s. First, fuel cost: green hydrogen’s EUR 4 to 8/kg production cost plus EUR 2 to 6/kg distribution premium means bunkering prices that are 3 to 5 times the energy equivalent of VLSFO in 2025. Without a mandatory carbon price high enough to close that gap, operators in unregulated trades have no economic incentive to switch. The EU Emissions Trading System’s extension to shipping from 2024 adds a cost of roughly EUR 1 to 3 per GJ for VLSFO combustion (at EUR 60 to 80 per tonne CO2), reducing but not closing the gap.

Second, infrastructure absence: no commercial port has committed general-access LH2 bunkering infrastructure. Until at least a handful of major ports offer bunkering, network effects cannot develop and hydrogen fuel cell vessels are locked to point-to-point routes served by purpose-built bunkering.

Third, vessel economics: the capital cost premium of a hydrogen fuel cell vessel over a conventional diesel ferry remains large. MF Hydra cost substantially more than a conventional diesel ro-pax of equivalent capacity; Norwegian government ENOVA grants covered the premium. In trades without public subsidy, the business case for hydrogen fuel cell ferries cannot close at current technology and hydrogen costs.

Fourth, regulatory maturity: MSC.1/Circ.1647 remains an interim circular, not a mandatory Code chapter. Flag state approval processes differ between maritime administrations, vessel insurance underwriters require special arrangements, and port-state control officers have limited standardised training for hydrogen fuel cell vessel inspections.

Near-term commercial signals

Despite these barriers, several market signals point to continued development through the late 2020s. Ballard’s FCwave module received DNV type approval in 2022 and ABS approval in 2023, and had secured orders for cumulative capacity exceeding 6 MW for marine applications by late 2024. Corvus Energy’s Pelican system, type-approved by DNV in September 2024, brings a second commercial marine fuel cell product from an established marine energy storage supplier to market. The EU’s Clean Hydrogen Partnership (formerly FCH JU) continues to fund demonstration projects under Horizon Europe, including follow-on programmes to FLAGSHIPS targeting larger vessels. Norway’s E39 ferry procurement for two hydrogen ferries with delivery in 2026-2027 will be the first large-scale procurement follow-on to MF Hydra.

The International Energy Agency’s 2024 Global Hydrogen Review observed that while total global electrolyser capacity reached 1.4 GW in 2023, maritime hydrogen demand remained negligible at a global scale. The report projected that maritime hydrogen demand could reach 2 to 5 million tonnes per year by 2050 under ambitious decarbonisation scenarios, but noted that achieving this would require both carbon pricing and direct regulatory mandates that were not yet in place through 2025.

Regulatory framework

The IGF Code and its scope

The IGF Code (IMO resolution MSC.391(95), in force 1 January 2017, as amended by MSC.481(102)) is the primary IMO instrument for ships using low-flashpoint fuels. The Code currently has fully developed provisions for natural gas (Part A, which is chapter-complete) and framework provisions for other fuels (Part A-1 for LPG, with other fuels under development).

Hydrogen has not yet been absorbed into the IGF Code’s main chapters. Instead, hydrogen fuel cell systems are governed by a separate interim instrument while the Code develops.

IMO MSC.1/Circ.1647: Interim Guidelines for Fuel Cell Power Installations

IMO MSC.1/Circ.1647, issued by the Maritime Safety Committee in June 2022, is the operative interim guideline for ships installing fuel cell power systems. It replaces the earlier MSC.1/Circ.1hydrogen (2009) version and reflects experience from the Viking Lady, MF Hydra, and other projects.

The circular covers:

• Design requirements: fuel cell room construction including A-60 fire boundaries or equivalent, location relative to accommodation spaces, drainage, and structural fire protection.
• Ventilation: continuous ventilation of fuel cell spaces to limit hydrogen concentration below 25 percent of the lower flammability limit (1 percent H2 in air). Ventilation must operate before fuel supply valves can open.
• Gas detection: fixed gas detection covering fuel cell rooms, adjacent spaces, and fuel processing areas. Detector voting logic for automatic fuel cutoff.
• Fuel system safety: double-wall or equivalent protection for hydrogen piping, manual and automatic shutoff valves, pressure relief sizing, and leak-before-break design philosophy.
• Electrical safety: DC bus isolation, insulation monitoring, and earthing requirements specific to fuel cell DC output.
• Safety management: fuel cell system operation manual, crew training, emergency procedures, and interaction with the ship’s safety management system (SMS) under ISM Code.

The guidelines apply to all fuel cell types (PEM, SOFC, MCFC) regardless of hydrogen storage method, and to both propulsion and auxiliary power applications. Flag state approval is required for each installation, with the flag granting equivalence to SOLAS requirements under SOLAS II-1/55 (equivalent design and arrangement).

IMO interim guidelines for hydrogen as fuel (CCC 11, 2024)

The IMO Sub-Committee on Carriage of Cargoes and Containers (CCC) at its 11th session (2024) completed draft interim guidelines for the safety of ships using hydrogen as fuel in direct combustion, covering both internal combustion engines and fuel cells. DNV reported that these guidelines were finalized in 2025 and circulated for adoption. They address hydrogen bunkering operations, cryogenic storage, and interactions between hydrogen fuel systems and vessel safety systems in more detail than the 2022 circular.

Class society rules

DNV’s Rules for Classification: Ships, Part 6 Chapter 2C (Fuel Cell Installations) entered force in 2020 and have been updated through 2024. DNV was the classification society for the MF Hydra and the Viking Lady SOFC project. The rules set detailed requirements for hydrogen piping class notation, fuel cell room layout, monitoring and control system performance, and crew certification.

ABS published its Guide for Fuel Cell Power Systems for Marine and Offshore Applications in 2020. It uses a risk-based approach aligned with MSC.1/Circ.1647 and covers PEM, SOFC, MCFC, and phosphoric acid fuel cells across propulsion and auxiliary applications. Bloom Energy’s SOFC module received ABS type approval in 2024, marking the first commercial marine SOFC product to pass a class society product certification process.

Lloyd’s Register, Bureau Veritas (BV), and ClassNK each publish guidance notes and rule amendments covering fuel cell vessels. LR certified the Energy Observer research vessel; BV and ClassNK have approved ferry designs in European and Japanese programmes respectively.

The STCW Convention (as amended at the Manila Conference, 2010) does not yet include a dedicated fuel cell competency table, but flag states and class societies require operators and maintenance crews to hold documented training for hydrogen fuel handling and fuel cell system management, typically from the OEM or from accredited maritime training providers.

Demonstration projects and commercial deployments

MF Hydra (Norway, 2023)

Norled’s MF Hydra entered scheduled commercial service in April 2023 on the Hjelmeland-Skipavik-Nesvik triangle in Ryfylke, western Norway. The 82.4-metre ro-pax ferry carries approximately 80 passengers and 30 cars, powered by two Ballard FCwave fuel cell modules delivering a combined continuous output of approximately 200 kW, with a 330 kWh battery buffer for peak manoeuvring loads.

Liquid hydrogen is stored in a 500-litre cryogenic tank mounted on deck, holding approximately 200 kilograms of LH2. Bunkering takes approximately one to two hours by road tanker at the Hjelmeland berth. The vessel was built by Westcon Yards and classified by DNV. MF Hydra is the first commercial passenger vessel to operate on liquid hydrogen in regular scheduled service anywhere in the world.

The project received funding support from the Norwegian government’s ENOVA program and the EU’s Fuel Cells and Hydrogen Joint Undertaking (FCH JU, now Clean Hydrogen Partnership). Construction and commissioning costs were substantially higher than a comparable battery-electric or diesel ferry, and the project has been studied extensively as a reference case for technology cost and regulatory pathway.

Energy Observer (international, 2017-present)

The Energy Observer is a 30.5-metre trimaran converted from a racing catamaran and relaunched in 2017 as a zero-emission research vessel. It combines PEM fuel cells with solar photovoltaics, wind turbines, and an onboard electrolyser that produces hydrogen from seawater. The vessel circumnavigated the world across multiple voyages and has visited more than 100 ports to advocate for hydrogen and renewable energy.

Energy Observer is not a commercial ferry; it carries a crew and research team and is funded by sponsorship. Its technical significance is as a long-endurance demonstration of integrated hydrogen energy management: produce hydrogen from renewable electricity, store it compressed, and run it through PEM stacks as needed, with solar and wind covering base load.

Sea Change (San Francisco, 2022)

The Sea Change is a 70-foot passenger ferry built for the Water Emergency Transportation Authority (WETA) in San Francisco Bay, delivered by All American Marine in 2022. It carries up to 75 passengers, powered by two Ballard FCwave fuel cell modules (totalling approximately 360 kW continuous) with a battery buffer and compressed hydrogen storage in roof-mounted cylinders at 350 bar.

Sea Change was the first hydrogen fuel cell passenger ferry to enter service in the United States. It operates Bay Area commuter and tourist routes and has accumulated several thousand hours of operational experience on compressed hydrogen, informing fuel cell system reliability data for subsequent US programmes.

FLAGSHIPS project (Europe, 2020-2023)

The FLAGSHIPS project (Fuel ceLl and hydrogen vessels for Green sHipping In European Port regionS) was an EU Horizon 2020-funded programme running from 2020 to 2023. It supported the deployment of two hydrogen fuel cell vessels in commercial operation in Europe:

• PA.X.ELL2: a push-boat operating on the Rhine at Duisburg, Germany, with a 100 kW PEM fuel cell system supplied by Ballard, operated by Schifffahrtsgesellschaft der Lys.
• Viking Energy: not the offshore vessel of the same name, but a separate river passenger vessel on the Seine in Paris, with a 50 kW PEM fuel cell, operated in connection with the city’s SNCF-Transilien network during trials.

The FLAGSHIPS project demonstrated commercial fuel cell vessel operation under EU inland waterway regulations, producing public reports on reliability, bunkering logistics, and cost structure that are referenced by subsequent European port hydrogen programmes.

HySeas III (Scotland, 2020-2022)

The HySeas III project (Hydrogen-Powered Zero-Emission Ferry for Scotland) was an EU-funded feasibility and design study targeting a roll-on/roll-off passenger ferry for the Scottish island routes operated by CalMac Ferries. The project (2018-2022) developed a concept design for a 120-passenger, 14-car ferry using approximately 1 MW of PEM fuel cell power with liquid hydrogen storage.

HySeas III did not result in a built vessel; the project concluded with a design package and regulatory pathway assessment. Funding was from the EU Interreg NW Europe programme. The design work is being referenced by successor Scottish and UK programmes studying hydrogen ferry deployment on the western island routes.

Norled and Norwegian ferry programme

Norway’s ferry procurement agency (Statens vegvesen / Riksvegen) has issued tenders requiring hydrogen-powered ferries on several routes. Norled won a contract for two hydrogen ferries for the E39 route programme, with deliveries targeted for 2026-2027. Several other Norwegian operators (Color Line, Bastø Fosen, Fjord Line) have run concept or design studies for hydrogen-fuelled vessels.

The Norwegian programme is the most mature national hydrogen ferry programme in the world, with actual vessels in operation (MF Hydra) and follow-on procurement in contracting. ENOVA subsidies have supported the above-market cost of the first vessels while the technology matures.

Japanese hydrogen vessel programmes

Kawasaki Heavy Industries (KHI) has been developing hydrogen fuel cell vessels in parallel with its liquid hydrogen supply chain project. KHI’s Suiso Frontier, a dedicated liquid hydrogen carrier for importing Australian hydrogen, uses hydrogen for its own auxiliary power. KHI and several Japanese shipyards are developing hydrogen-powered ferries and tugboats under the NEDO programme, with several projects targeting commercial deployment in the late 2020s.

ClassNK published its guidelines for hydrogen-fuelled vessels in 2021 and has been involved in approving Japanese hydrogen vessel designs. Japan’s National Maritime Research Institute (NMRI) is coordinating several research programmes on hydrogen fuel cells for coastal shipping.

Summary of key projects

Efficiency and well-to-wake analysis

Electrical efficiency and the tank-to-propeller chain

A PEM fuel cell system converts hydrogen to DC electricity at 50 to 60 percent electrical efficiency (LHV basis) at rated load. Efficiency drops at low load (below 20 percent rated power) due to parasitic losses in the balance-of-plant equipment: air compressors, cooling pumps, hydrogen recirculation blowers, and control systems. A practical design target for a marine PEM system over a full duty cycle is 45 to 55 percent average electrical efficiency.

The tank-to-propeller chain on a fuel-cell-electric vessel looks like this: hydrogen LHV in → fuel cell at 52% efficiency → DC bus (98% conductor efficiency) → DC/AC inverter (97%) → propulsion motor (96%) → propeller shaft. The combined tank-to-propeller efficiency is approximately 0.52 × 0.98 × 0.97 × 0.96 = 0.47, or about 47 percent. A direct comparison with a diesel-electric vessel (diesel generator at 42% efficiency, same drivetrain downstream) gives approximately 0.42 × 0.98 × 0.97 × 0.96 = 0.38, or 38 percent. The fuel cell gains roughly 9 percentage points in drivetrain efficiency.

SOFC systems in CHP mode on a continuous-power auxiliary application can deliver 70 to 80 percent total energy utilisation if the high-temperature exhaust is used for space heating, cargo hold conditioning, or hotel steam generation.

Well-to-wake CO2 intensity

Well-to-wake intensity for hydrogen fuel cells depends entirely on production pathway. On grey hydrogen from natural gas steam methane reforming (SMR) without carbon capture, the WtW intensity is approximately 100 to 120 g-CO2eq/MJ, comparable to HFO combustion and providing no decarbonisation benefit.

On green hydrogen from renewable-electricity electrolysis, the WtW intensity drops to approximately 5 to 10 g-CO2eq/MJ, the lowest achievable for any marine fuel. This comparison is covered in detail in the well-to-wake intensity article and in the per-fuel well-to-wake hydrogen article.

The EU FuelEU Maritime Regulation (Regulation (EU) 2023/1805) counts hydrogen-fuelled vessels under its GHG intensity compliance framework. Renewable hydrogen qualifying as an RFNBO under the EU RFNBO rules receives a 2x multiplier credit toward the GHG compliance target between 2025 and 2034.

Safety characteristics of hydrogen at sea

Flammability and dispersion

Hydrogen has a wide flammability range in air: 4 to 75 percent by volume, compared to methane’s 5 to 15 percent and gasoline vapour’s 1.4 to 7.6 percent. The lower flammability limit of 4 percent is only modestly lower than methane’s 5 percent, so the principal concern is the wide upper flammability limit, which means a larger range of concentrations can sustain ignition.

Minimum ignition energy for hydrogen is approximately 0.017 millijoules, roughly 10 times lower than methane. A static discharge from clothing or a metal tool can theoretically ignite a hydrogen-air mixture. Electrical equipment in hydrogen-risk zones must therefore be rated to the relevant IEC Ex standards for hydrogen (temperature class T1; explosion group IIC).

The mitigating factor is hydrogen’s buoyancy. Hydrogen is 14 times lighter than air and diffuses approximately 3.8 times faster. In an open or ventilated space, released hydrogen rises and disperses rapidly upward. The natural consequence is that a hydrogen leak in a well-ventilated space rarely accumulates to ignitable concentration at floor level, unlike propane, LPG, or petrol vapour, which pool near the deck.

Cryogenic hazards for liquid hydrogen

Liquid hydrogen at minus 253 degrees Celsius presents contact hazards (cryogenic burns, cold embrittlement of structural steels), oxygen condensation on uninsulated cold surfaces (creating an oxygen-enriched atmosphere locally), and boil-off gas that must be managed and vented safely.

Class rules require that LH2 vent outlets are positioned clear of ignition sources, accommodation air intakes, and escape routes. DNV rules specify vertical upward-discharge vent stacks at a minimum height above deck and minimum horizontal distance from air intakes. The IGF Code provisions for cryogenic fuels (relevant primarily to LNG but applied by analogy to LH2 pending dedicated rules) specify double-walled piping or equivalent protection for cryogenic fuel lines.

CO2 and inert gas interaction

Fuel cell rooms require ventilation, and the ventilation system design must consider interaction with the CO2 fixed firefighting system. CO2 flooding of a fuel cell room suppresses the electrical fire risk but must be coordinated with hydrogen fuel supply isolation. MSC.1/Circ.1647 addresses this coordination requirement, specifying that CO2 release must be sequenced after confirmed fuel cell shutdown and hydrogen supply valve closure.

Limitations

Energy density constraint. Liquid hydrogen at minus 253 degrees Celsius stores approximately 8.5 MJ/litre, compared to approximately 35 MJ/litre for VLSFO. A vessel burning 10 tonnes of VLSFO per day at sea would need to carry approximately 3.7 tonnes of hydrogen per day (hydrogen’s mass advantage: 120 MJ/kg vs 40 MJ/kg LHV). At liquid hydrogen density of approximately 71 kg/m3, 3.7 tonnes requires roughly 52 cubic metres of tank volume, before structural allowance and insulation thickness. For a tanker or bulk carrier burning 50 to 80 tonnes of HFO per day on a two-week voyage, the required LH2 tank volume is impractical within current hull designs.

Bunkering infrastructure scarcity. Green hydrogen bunkering at port scale does not exist at most commercial shipping ports as of mid-2026. The few operating hydrogen bunker facilities serve specific ferry pilot projects. Expanding to general commercial shipping would require new production capacity, pipeline or road transport distribution, and cryogenic or high-pressure bunkering equipment at hundreds of ports worldwide.

Fuel cell cost. Marine-qualified PEM fuel cell systems cost substantially more per kilowatt than diesel generators. Total installed system costs for large marine PEM installations were USD 2,400 to 3,400 per kW in 2023 to 2025, compared to USD 300 to 700 per kW for a diesel generator of equivalent output. Stack lifetime is approximately 20,000 to 25,000 hours before replacement at current technology levels, adding ongoing cost at roughly USD 200 to 400 per kW per replacement event.

High-temperature fuel cell start-up time. SOFC and MCFC systems require hours to reach operating temperature and cannot tolerate rapid thermal cycling. This makes them unsuitable for vessels with intermittent duty cycles or frequent shutdowns, limiting their marine application to continuous-power roles like auxiliary power generation underway.

Crew training and operational complexity. Hydrogen fuel cell systems are mechanically and electrically more complex than diesel generator sets, and the crew competency requirements are more demanding. The absence of a dedicated STCW fuel cell competency table means that crew training varies by flag state and OEM, creating inconsistency in operational knowledge across the global fleet.

Regulatory maturity. MSC.1/Circ.1647 is an interim guideline, not a mandatory Code provision. Flag state discretion means that approval processes differ substantially between maritime nations. A vessel receiving equivalence approval from one flag may need to re-demonstrate compliance if reflagged. The full integration of hydrogen fuel cell requirements into the IGF Code, expected progressively through the late 2020s, will reduce this uncertainty.

Green hydrogen premium. The production cost of green hydrogen via electrolysis from renewable electricity remains substantially above grey hydrogen from SMR. In 2024 to 2025, green hydrogen production costs were estimated at EUR 4 to 8 per kilogram at the production site, before distribution and bunkering margin, compared to grey hydrogen at EUR 1.5 to 2.5 per kilogram. At 120 MJ/kg LHV, green hydrogen at EUR 6/kg equates to an energy cost of approximately EUR 50 per GJ, compared to VLSFO at approximately EUR 20 to 25 per GJ at 2024 bunker prices.

Scalability to large commercial vessels. The hydrogen fuel cell niche sits firmly at short-route passenger ferries, harbour craft, and research vessels. Scaling to the energy demands of a 300-metre containership or a suezmax tanker would require onboard hydrogen storage volumes and fuel cell capacities that are not technically or economically practical with technology available in 2026. Ammonia direct combustion and methanol direct combustion are far better placed for large commercial shipping decarbonisation through the 2030s.

See also

• Hydrogen as Marine Fuel
• IGF Code: Low-Flashpoint Fuel Ships
• Ammonia as Marine Fuel
• Methanol as Marine Fuel
• Battery-Electric Ferries: Technology, Fleet, and Operations
• Well-to-Wake Intensity
• Green Shipping Corridors
• RFNBO EU Rules
• Per-Fuel WtW: Hydrogen Grades
• Marine Electrical Generation and Distribution

Related calculators:

• Marine Fuel Cell Power Calculator
• Hydrogen Volumetric Energy Density
• Hydrogen Well-to-Wake: Grey / Green
• Hydrogen (liquid): Core Properties
• Alt Fuel Bunker Premium
• Voyage Fuel and CO2