By ShipCalculators.com · Published 8 May 2026

Per-fuel well-to-wake intensity: hydrogen

Hydrogen (H2) is the lightest molecule and the most demanding marine fuel on storage and bunkering, yet its well-to-wake (WtW) intensity ranges across more than two orders of magnitude depending on how the molecule is produced and conditioned for shipboard use. Burned in a marine fuel cell or in a hydrogen ICE, the molecule itself releases zero tank-to-wake (TtW) carbon dioxide because there are no carbon atoms in the fuel. The lifecycle outcome is set entirely by two upstream chains. The first is the production pathway that delivers gaseous hydrogen at the plant gate, and the second is the conditioning chain (compression to 350 or 700 bar, or liquefaction to minus 253 degrees Celsius) that delivers the molecule to the bunker manifold. Under MEPC.391(82) Annex 1 and FuelEU Annex II, conventional grey hydrogen from steam-methane reforming carries a WtW intensity of approximately 104 gCO2eq/MJ before liquefaction, materially worse than VLSFO at roughly 92 gCO2eq/MJ. Blue hydrogen with carbon capture on the syngas chain drops to 25 to 50 gCO2eq/MJ, and green hydrogen from electrolysis with renewable electricity reaches 1 to 15 gCO2eq/MJ depending on grid intensity at the electrolyser site. Green hydrogen qualifies as a renewable fuel of non-biological origin (RFNBO) under RED III and earns the FuelEU 2x multiplier when production satisfies the additionality, temporal and geographical correlation rules. The liquefaction step adds an energy penalty of approximately 35 percent of LCV and shifts the apparent WtW intensity sharply when the liquefier draws on a fossil grid. The three grades are the same molecule with the same fuel cell or ICE at the engine boundary, but they are not the same fuel for compliance under the IMO net-zero framework or the marine GHG fuel standard methodology. Operators size the gap with /calculators/fuel-wtw-hydrogen, price grey-green blends with /calculators/fuel-wtw-blend, and check RFNBO uplift with /calculators/fueleu-rfnbo-multiplier against the ammonia-grade and methanol-grade analogues.

Contents

Background: hydrogen as a marine fuel

Hydrogen is a colourless, odourless, non-toxic gas with the chemical formula H2 and a molecular weight of 2.016 g/mol. At standard atmospheric pressure it is a gas down to approximately minus 253 degrees Celsius, where it liquefies to a cryogenic fluid with a density of approximately 70.85 kg/m3. The lower heating value (LCV) of approximately 120 MJ/kg is the highest of any chemical fuel on a mass basis, roughly 2.8 times the LCV of VLSFO, 6 times the LCV of methanol, and 6.5 times the LCV of ammonia. The volumetric energy density tells the opposite story. Liquid hydrogen at minus 253 degrees Celsius delivers approximately 8.5 GJ/m3, which is roughly 24 percent of VLSFO and roughly 60 percent of LNG on the same volumetric basis. Compressed gaseous hydrogen at 700 bar and 15 degrees Celsius delivers approximately 4.7 GJ/m3, and at 350 bar approximately 2.8 GJ/m3. The bunker-tank size penalty for a hydrogen-fuelled ship is therefore in the band of 4 to 12 times the equivalent VLSFO tank for the same range, depending on the storage mode chosen.

The case for hydrogen as a marine fuel rests on four structural advantages. First, the molecule contains no carbon, so the TtW CO2 is zero by chemistry, and there is no carbon atom for the engine to oxidise. Second, the proton-exchange-membrane (PEM) and solid-oxide (SOFC) fuel cells operate at electrochemical efficiencies of 50 to 60 percent (PEM) and 55 to 65 percent (SOFC) at the cell level, materially higher than the 45 to 50 percent thermal efficiency of a marine diesel engine, which means a hydrogen-fuel-cell vessel converts a larger fraction of the bunkered LCV into shaft or hotel energy. Third, hydrogen is the universal precursor for ammonia, e-methanol, e-LNG and e-diesel, so a renewable-hydrogen value chain that scales for industrial offtake supplies the carrier-fuel pathways at the same time. Fourth, the by-product of fuel-cell oxidation is liquid water, which has zero local air-quality impact and zero global warming impact.

The case against rests on four considerations. The volumetric energy density is the worst of any practical marine fuel, and the bunker-tank size penalty is severe enough to displace cargo on a deep-sea ship. The cryogenic storage burden at minus 253 degrees Celsius is unprecedented in marine fuel handling: the boil-off rate, the materials hydrogen-embrittlement envelope, and the insulation thickness drive the tank capital cost above any liquid fuel currently used at sea. The liquefaction energy penalty is 30 to 40 percent of the LCV, which means almost a third of the renewable electricity used to make green hydrogen is consumed before the molecule reaches the bunker manifold. The bunkering supply chain at marine scale does not yet exist outside the Norwegian and Northern European pilot ports, and the IGF Code amendments for hydrogen are still under development at the IMO Sub-Committee on Carriage of Cargoes and Containers (CCC).

The shipboard architecture is materially different from any prior marine fuel. A hydrogen-fuel-cell vessel uses a high-pressure or cryogenic storage system, a pressure-let-down or vaporiser stage, a fuel-cell stack with a balance-of-plant for water and air management, an inverter for shaft or hotel-load conversion, and a fuel-cell-cooling circuit. A hydrogen ICE vessel adds a high-pressure gas injection skid, a pilot-fuel storage and injection system, and a substantial NOx after-treatment stack. The fuel-cell architecture has no moving combustion components, so maintenance intensity is lower than a diesel engine, but the stack life of 30,000 to 50,000 hours is substantially shorter than a marine diesel and represents a periodic capital event during the ship’s operational life.

H2 Council classification and the three primary grades

The Hydrogen Council, an industry consortium of more than 150 producers, technology vendors and offtakers, maintains a classification of hydrogen production pathways by colour code that has become the de-facto reference vocabulary for marine fuel discussions. The three primary grades for marine bunker discussions are grey, blue and green, and several adjacent colours describe minority pathways that surface in lifecycle accounting.

Grey hydrogen is produced from natural gas through steam-methane reforming (SMR) without carbon capture. The pathway is the dominant industrial process today, accounting for approximately 75 percent of global hydrogen production at roughly 70 million tonnes per year. The carbon dioxide produced in the reforming reaction and the water-gas-shift reaction is vented to atmosphere, which sets the WtW intensity at approximately 104 gCO2eq/MJ before any liquefaction or compression conditioning.

Blue hydrogen is produced from natural gas through SMR or autothermal reforming (ATR) with carbon capture and storage (CCS) applied to the synthesis-gas chain. Capture rates of 70 percent on legacy SMR retrofits, 90 percent on purpose-designed ATR plants with high-pressure CO2 capture, and 95 to 97 percent on the most modern integrated plants set the WtW intensity at 25 to 50 gCO2eq/MJ. The grade label requires a permanent geological storage commitment for the captured CO2, typically through a saline aquifer or depleted-field injection contract.

Green hydrogen is produced from water electrolysis using renewable electricity. The pathway is structurally different from grey and blue: there is no fossil feedstock, and the only inputs are water, electricity and a stoichiometric amount of energy to drive the electrolytic dissociation. The WtW intensity tracks the carbon intensity of the electricity supply, which is approximately 0 gCO2eq/kWh for direct-coupled wind or solar with curtailment management, 30 to 60 gCO2eq/kWh for a modern Northern European grid, and 400 to 800 gCO2eq/kWh for a coal-heavy grid. Green hydrogen qualifies as an RFNBO under RED III when the renewable-electricity supply satisfies the additionality, temporal and geographical correlation rules.

Adjacent grades surface in lifecycle conversations. Brown or black hydrogen comes from coal gasification (lignite or bituminous) and is the most carbon-intensive grade at 180 to 280 gCO2eq/MJ, principally a Chinese and Indian feedstock. Turquoise hydrogen is produced through methane pyrolysis with solid-carbon byproduct rather than gaseous CO2, currently at pilot scale through Monolith and HiiROC. Pink hydrogen is electrolytic hydrogen from nuclear-powered electricity, structurally similar to green at low WtW intensity but excluded from RFNBO eligibility under the current RED III text. Yellow hydrogen is electrolytic hydrogen from grid electricity without renewable certification, with WtW intensity tracking the grid mix and not eligible for RFNBO uplift. White hydrogen is naturally occurring geological hydrogen at exploration stage, currently a research curiosity rather than a commercial source. The three grades that dominate marine fuel discussions are grey, blue and green, and the rest of this article focuses on those three.

Grey hydrogen: SMR pathway

Steam-methane reforming is the workhorse of industrial hydrogen production. The two principal reactions are the reforming step (CH4 + H2O to CO + 3H2) at 700 to 1,000 degrees Celsius over a nickel catalyst, and the water-gas-shift step (CO + H2O to CO2 + H2) at 200 to 400 degrees Celsius over an iron-chromium or copper-zinc catalyst. The combined stoichiometry is CH4 + 2H2O to CO2 + 4H2, which delivers four moles of hydrogen per mole of methane consumed and one mole of CO2 as a co-product. The pressure swing adsorption (PSA) unit downstream of the shift reactor purifies the hydrogen to better than 99.9 percent and recycles the off-gas as a fuel stream to the reformer burners.

The carbon dioxide emissions from the SMR pathway come from three sources. The first is the stoichiometric CO2 from the shift reaction itself, at approximately 5.5 kg CO2 per kg H2 for a typical SMR plant. The second is the combustion CO2 from the reformer burners that supply the endothermic reforming heat, at approximately 1.5 to 2.5 kg CO2 per kg H2 depending on whether the burners run on the PSA off-gas, the natural-gas feed, or a mixed stream. The third is the upstream CO2-equivalent from natural-gas extraction, processing and transport, at approximately 0.5 to 2.0 kg CO2eq per kg H2 depending on the methane leakage rate of the gas supply chain. The aggregate WtT intensity is approximately 8 to 12 kg CO2eq per kg H2, or approximately 67 to 100 gCO2eq/MJ on the LCV basis of 120 MJ/kg.

The MEPC.391(82) Annex 1 default for grey hydrogen at the plant gate is approximately 104 gCO2eq/MJ on a WtT basis, which already includes a 0.5 to 1.5 percent upstream methane leakage rate at GWP100 of 28. The figure rises to approximately 130 to 145 gCO2eq/MJ once liquefaction at a fossil-grid liquefier (35 percent of LCV at 60 to 80 gCO2eq/kWh grid intensity) is added, which is materially worse than VLSFO at 92 gCO2eq/MJ. Grey liquid hydrogen as marine fuel is therefore not a decarbonisation pathway, and it does not appear in any current FuelEU or IMO net-zero compliance plan as a long-term solution.

The grey-hydrogen production base in 2026 is concentrated in the United States Gulf Coast, the Middle East, China and Northern Europe. The merchant market for hydrogen is small relative to the captive market: most hydrogen is consumed within the producing refinery or chemical plant for desulphurisation, ammonia synthesis or methanol synthesis. The marine bunker offtake of grey hydrogen as a fuel is currently zero on a deep-sea basis, with pilot operations limited to inland and short-sea shuttle vessels in Belgium, Norway and the Netherlands.

Blue hydrogen: SMR/ATR + CCS pathway

Blue hydrogen applies carbon capture and storage to the SMR or autothermal-reforming pathway. The capture envelope depends on the reformer technology, the capture solvent and the integration design. A legacy SMR plant with post-combustion capture on the reformer flue gas captures only the combustion CO2 (approximately 30 to 35 percent of total CO2) and leaves the larger stoichiometric CO2 stream uncaptured. A modern SMR plant with pre-combustion capture on the shift-reactor outlet captures the stoichiometric CO2 (approximately 60 to 65 percent of total CO2) but leaves the reformer-burner CO2 uncaptured, yielding an aggregate capture rate of approximately 60 percent. A purpose-designed ATR plant with oxygen-blown reforming concentrates the CO2 in a single high-pressure stream and achieves capture rates of 90 to 95 percent on the carbon-bearing streams, with the residual CO2 from minor heat-integration burners.

The most modern integrated blue-hydrogen plants combine ATR with full off-gas capture and a small electric heater for heat integration, achieving aggregate capture rates of 95 to 97 percent. Two reference designs at this performance level are the Equinor and BP Net Zero Teesside H2Teesside project in the United Kingdom and the Air Products NEOM facility in Saudi Arabia. The WtT intensity for the highest-capture blue-hydrogen plant is approximately 25 gCO2eq/MJ, which combines a residual stoichiometric and combustion CO2 of approximately 0.7 kg CO2 per kg H2 with the upstream methane-supply chain emissions and the parasitic energy demand of the capture and compression units.

The blue-hydrogen WtT intensity is highly sensitive to two parameters. The first is the upstream methane leakage rate of the gas supply: a 0.5 percent leakage supply produces blue hydrogen at approximately 25 gCO2eq/MJ, while a 2.5 percent leakage supply produces blue hydrogen at approximately 55 gCO2eq/MJ, which erodes most of the CCS benefit. The methane GWP100 of 28 is multiplicative on the leakage fraction, and a high-leakage gas supply can degrade blue hydrogen below the threshold for credible decarbonisation. The second is the permanence of the geological storage. The current accounting under MEPC.391(82) and FuelEU treats permanent geological storage as a 100 percent credit; if the storage facility leaks at a rate above 0.1 percent per year, the lifecycle credit is reduced through a re-balancing protocol that has not yet been finalised under the IMO methodology.

The MEPC.391(82) Annex 1 default for blue hydrogen at the plant gate is approximately 25 to 50 gCO2eq/MJ on a WtT basis, with the band reflecting capture rate (70 to 95 percent) and methane leakage (0.5 to 2.0 percent). After liquefaction at a low-carbon grid liquefier (35 percent of LCV at 30 to 60 gCO2eq/kWh), the WtW intensity for liquid blue hydrogen is approximately 35 to 70 gCO2eq/MJ. Blue hydrogen does not qualify as an RFNBO and does not earn the FuelEU 2x multiplier, but it does qualify as a low-carbon fuel under MEPC.391(82) and contributes positively to the GFI attained calculation.

Green hydrogen: electrolysis pathway (RFNBO)

Green hydrogen is produced from water electrolysis driven by renewable electricity. The two dominant electrolyser technologies are alkaline (AEL) at 60 to 70 percent stack efficiency on LCV basis, and polymer-electrolyte-membrane (PEM) at 65 to 75 percent stack efficiency, with solid-oxide electrolysis (SOEC) at 75 to 85 percent stack efficiency emerging at pilot scale through Topsoe and Sunfire. The system efficiency at the electrolyser balance-of-plant (rectifier, water-treatment, cooling, gas drying) is approximately 5 percentage points lower than the stack figure, so an AEL system delivers approximately 55 to 65 percent system efficiency, a PEM system 60 to 70 percent, and an SOEC system 70 to 80 percent. The net result is that approximately 50 to 55 kWh of electricity is consumed per kg of hydrogen produced on a state-of-the-art system, against the LCV reference of 33.3 kWh/kg.

The carbon intensity of the renewable electricity supply determines the WtT intensity of the green hydrogen. A direct-coupled wind or solar electrolyser with curtailment management runs at approximately 5 to 10 gCO2eq/kWh on a lifecycle basis (including manufacturing of the wind turbine or solar panel, balance-of-plant, transmission losses, and end-of-life). The corresponding WtW intensity for the green-hydrogen molecule at the plant gate is approximately 1 to 5 gCO2eq/MJ, before any liquefaction or compression conditioning. A green-hydrogen plant connected to a low-carbon grid (Norwegian, Quebec or French nuclear-and-hydro mix) at 30 to 50 gCO2eq/kWh produces green hydrogen at approximately 6 to 12 gCO2eq/MJ.

The MEPC.391(82) Annex 1 default for green hydrogen at the plant gate is approximately 1 to 15 gCO2eq/MJ on a WtT basis, depending on the electricity source. The corresponding WtW intensity for liquid green hydrogen, after a liquefaction step at a renewable-grid liquefier, is approximately 5 to 35 gCO2eq/MJ. The wide band reflects the dominant role of the liquefaction grid intensity in the lifecycle outcome: a green-hydrogen molecule that is liquefied on a coal grid carries the coal-grid intensity for the 35 percent of LCV consumed in liquefaction, which adds approximately 50 to 90 gCO2eq/MJ to the lifecycle. The liquefaction-grid attribution rule is therefore central to the green-hydrogen WtW outcome and is the subject of detailed treatment in the RED III delegated regulations.

Green hydrogen qualifies as an RFNBO under Directive (EU) 2023/2413 (RED III) when production satisfies the criteria of the additionality, temporal correlation, and geographical correlation rules in Commission Delegated Regulation (EU) 2023/1184. The renewable-electricity supply must come from a new asset (or one that is not yet 36 months old at the start of operation), the production must be temporally matched to the renewable generation on an hourly basis from 2030, and the renewable asset must be located in the same bidding zone as the electrolyser. Green hydrogen that satisfies the criteria earns the FuelEU 2x multiplier on the energy delivered, which doubles the compliance value of the molecule against the FuelEU GHG-intensity target.

MEPC.391(82) Annex 1 defaults per grade

The IMO MEPC.391(82) Lifecycle GHG Intensity of Marine Fuels Guidelines provide default emission factors per fuel and per production pathway, with the option to substitute a verified pathway-specific value when the producer maintains a recognised certification. For hydrogen, Annex 1 provides defaults for the principal SMR, ATR-CCS and electrolysis pathways, with the values expressed in gCO2eq per MJ of fuel delivered to the bunker manifold (LCV basis).

The MEPC.391(82) Annex 1 default for grey hydrogen at the plant gate is approximately 104 gCO2eq/MJ. The figure aggregates the upstream natural-gas extraction and processing emissions (approximately 8 to 15 gCO2eq/MJ), the SMR-and-shift stoichiometric and combustion CO2 (approximately 75 to 85 gCO2eq/MJ), and the PSA off-gas treatment (approximately 1 to 2 gCO2eq/MJ). A vessel that bunkers grey gaseous hydrogen at 350 or 700 bar from a pipeline source bypasses the liquefaction penalty, and the WtW intensity is approximately 105 to 110 gCO2eq/MJ including the compression energy at the bunker terminal.

The MEPC.391(82) Annex 1 default for blue hydrogen at the plant gate is approximately 25 to 50 gCO2eq/MJ, with the band reflecting capture-rate variation (70 to 95 percent) and methane leakage variation (0.5 to 2.0 percent). The default for the highest-performance ATR plant with 95 percent capture and 0.5 percent leakage is approximately 25 gCO2eq/MJ. The default for a legacy SMR plant with 70 percent capture and 1.5 percent leakage is approximately 50 gCO2eq/MJ. The producer can substitute a verified pathway-specific value when the carbon-capture rate and the methane leakage rate are documented under a recognised certification scheme such as CertifHy, the Hydrogen Council scheme, or the proposed ISO 14687-3 protocol.

The MEPC.391(82) Annex 1 default for green hydrogen at the plant gate is approximately 1 to 15 gCO2eq/MJ on the WtT basis. The default for a direct-coupled wind or solar plant with curtailment management is approximately 1 to 5 gCO2eq/MJ. The default for a low-carbon grid plant in the Northern European mix is approximately 6 to 12 gCO2eq/MJ. The default for a renewable-PPA plant in a coal-heavy regional grid is up to approximately 15 gCO2eq/MJ, even with the renewable certificates, because of the lifecycle emissions of the wind or solar manufacturing and the residual grid balancing.

The MEPC.391(82) framework requires that the liquefaction step be accounted separately when the molecule is delivered as cryogenic liquid to the bunker manifold. The default attribution is the grid intensity at the liquefier site, applied to the 35 percent of LCV consumed in liquefaction. For a green-hydrogen plant in Norway with a Norwegian-grid liquefier at 20 gCO2eq/kWh, the liquefaction adds approximately 7 gCO2eq/MJ to the WtT intensity. For a grey-hydrogen plant in the United States Gulf Coast with a Texas-grid liquefier at 400 gCO2eq/kWh, the liquefaction adds approximately 130 gCO2eq/MJ, raising the grey-LH2 WtW to approximately 235 gCO2eq/MJ, which is more than double VLSFO.

FuelEU Annex II treatment

Regulation (EU) 2023/1805 (FuelEU Maritime) Annex II implements the lifecycle GHG-intensity methodology for the EU/EEA voyage scope. The treatment of hydrogen mirrors the MEPC.391(82) structure but adds the RFNBO multiplier for green hydrogen and the FuelEU GHG-intensity target trajectory. The Annex II default WtW emission factors for hydrogen are aligned with the IMO Annex 1 values, with minor adjustments for the EU-specific upstream methane attribution and the renewable-electricity statistical methodology in Annex VI of RED III.

The FuelEU Annex II default for grey hydrogen is approximately 104 gCO2eq/MJ at the plant gate, rising to 130 to 200 gCO2eq/MJ once liquefaction is added at a fossil-grid liquefier. The FuelEU Annex II default for blue hydrogen is 25 to 50 gCO2eq/MJ at the plant gate, rising to 35 to 90 gCO2eq/MJ for liquid blue hydrogen including liquefaction at a low-carbon grid liquefier. The FuelEU Annex II default for green hydrogen is 1 to 15 gCO2eq/MJ at the plant gate, rising to 5 to 35 gCO2eq/MJ for liquid green hydrogen including liquefaction at a renewable grid liquefier. The defaults can be replaced by certified producer-specific values where the chain of custody and the production-pathway audit are accepted by the verifier.

The energy basis of FuelEU is the LCV of the fuel delivered to the bunker manifold, regardless of the storage state. A vessel that bunkers 1 tonne of liquid hydrogen logs 120 GJ of LCV, the same as a vessel that bunkers 1 tonne of compressed gaseous hydrogen at 700 bar. The volumetric burden differs by a factor of approximately 3 between the two storage modes, but the FuelEU energy accounting is on mass-LCV basis and is not affected. The volumetric burden affects the operational range and the bunker-tank size, not the compliance arithmetic.

The FuelEU GHG-intensity target trajectory begins at 2 percent reduction below the 2020 baseline of 91.16 gCO2eq/MJ (so a target of approximately 89.34 gCO2eq/MJ) in 2025, and tightens to 6 percent in 2030, 14.5 percent in 2035, 31 percent in 2040, 62 percent in 2045 and 80 percent in 2050. A vessel that bunkers green liquid hydrogen at 25 gCO2eq/MJ delivers a compliance surplus of approximately 64 gCO2eq/MJ against the 2025 target, and the surplus widens through the trajectory. The same vessel running grey hydrogen at 130 to 200 gCO2eq/MJ delivers a compliance deficit at all trajectory points, which forecloses grey hydrogen as a FuelEU-compliant marine fuel.

RED III sustainability criteria

Directive (EU) 2023/2413 (RED III) sets the sustainability and GHG-saving criteria for renewable fuels of non-biological origin (RFNBOs), the category that captures green hydrogen. The principal sustainability requirement is the 70 percent GHG-saving threshold against the fossil-fuel comparator of 94 gCO2eq/MJ, which translates to a maximum allowed RFNBO intensity of 28.2 gCO2eq/MJ on a WtW basis. A green-hydrogen molecule with a WtW intensity above 28.2 gCO2eq/MJ does not qualify as an RFNBO and is not eligible for the FuelEU 2x multiplier or the RED III renewable-energy target counting.

The 70 percent threshold is the dominant constraint on the liquefaction stage of green hydrogen. A green-hydrogen plant at the Norwegian coast with 5 gCO2eq/MJ at the plant gate, liquefied on the Norwegian grid at 20 gCO2eq/kWh, delivers liquid green hydrogen at approximately 12 gCO2eq/MJ, which satisfies the threshold. The same green-hydrogen plant liquefied on a Texas-grid liquefier at 400 gCO2eq/kWh delivers liquid green hydrogen at approximately 135 gCO2eq/MJ, which fails the threshold by a factor of nearly 5. The geographical placement of the liquefier is therefore a binary determinant of RFNBO eligibility for the resulting liquid hydrogen.

The RED III sustainability framework also requires demonstration of additionality (the renewable electricity must come from a new asset that would not otherwise have been built), temporal correlation (the renewable generation must be temporally matched to the electrolysis on a monthly basis from 2024 and an hourly basis from 2030), and geographical correlation (the renewable asset must be in the same bidding zone as the electrolyser, or in an adjacent zone with documented capacity to import the renewable electricity). Each criterion is documented in the bunker certificate and audited by the FuelEU verifier under Article 14 of FuelEU Maritime.

The RED III GHG-calculation methodology in Commission Delegated Regulation (EU) 2023/1185 specifies the default values for the renewable-electricity emission factors, the electrolyser efficiency, the liquefaction energy demand, and the conditioning chain. The default values can be replaced by certified producer-specific values where the chain of custody is documented under a recognised voluntary scheme such as ISCC EU, REDcert or CertifHy. The methodology requires mass-balance accounting under the voluntary scheme, with the physical chain of custody documented through commingled storage and transport.

RFNBO multiplier and correlation rules

Article 5(7) of FuelEU Maritime applies a multiplier of 2 to the energy delivered as RFNBO from 2025 to 2033, which doubles the compliance value of green hydrogen against the FuelEU GHG-intensity target. The multiplier reduces the effective WtW intensity of the RFNBO portion of the bunker, so that a 1 GJ green-hydrogen bunker at 10 gCO2eq/MJ counts as 2 GJ of compliance energy at 5 gCO2eq/MJ in the FuelEU calculation. The arithmetic is detailed at /calculators/fueleu-rfnbo-multiplier and at /wiki/fueleu-rfnbo-multiplier.

The RFNBO multiplier is conditional on the fuel satisfying all four criteria. The first is the 70 percent GHG-saving threshold against the 94 gCO2eq/MJ fossil-fuel comparator, which sets the maximum WtW intensity at 28.2 gCO2eq/MJ for the RFNBO portion. The second is the additionality criterion, which requires a renewable-electricity supply from a new asset under one of the four pathways in Article 4 of Commission Delegated Regulation (EU) 2023/1184: direct-coupled installation, PPA-supported new asset, biddingzone import with PPA, or grid offtake in a bidding zone with renewable share above 90 percent and grid intensity below 18 gCO2eq/MJ. The third is the temporal-correlation criterion, which requires monthly matching from 2024 and hourly matching from 2030. The fourth is the geographical-correlation criterion, which requires the renewable asset and the electrolyser to be in the same bidding zone or in interconnected zones with documented capacity.

A vessel that bunkers green liquid hydrogen with full RFNBO certification therefore earns the 2x multiplier on the green-hydrogen energy fraction. A vessel that bunkers blue hydrogen does not earn the multiplier because the hydrogen feedstock is fossil-derived, even when the WtW intensity is below 28.2 gCO2eq/MJ. A vessel that bunkers green hydrogen without full RFNBO certification (for example, electrolytic hydrogen from a grid offtake without PPA support) does not earn the multiplier even if the WtW intensity is below the 70 percent threshold.

The RFNBO multiplier expires at end of 2033 under the current text of Article 5(7). The expiry date is set in primary legislation and requires a co-decision amendment to extend, which is unlikely to be adopted in the next FuelEU revision. Operators planning hydrogen-fuelled vessels for delivery in 2027 to 2032 should size the multiplier benefit against the multi-year operational period and avoid over-counting the multiplier value on bunkers delivered after 2033.

On-board storage: cryogenic LH2, compressed CH2, LOHC, methanol-as-carrier

The on-board storage of hydrogen is the principal engineering constraint on hydrogen as a marine fuel. Four storage modes are technically viable at marine scale, each with a different volumetric and gravimetric envelope.

Cryogenic liquid hydrogen (LH2) at minus 253 degrees Celsius and approximately 4 bar absolute is the highest volumetric energy density storage mode at approximately 8.5 GJ/m3, or roughly 24 percent of VLSFO on the same basis. The tank is typically a vacuum-jacketed double-walled vessel with multi-layer insulation, similar in principle to an LNG IMO Type C tank but with materials selected for the hydrogen-embrittlement envelope and the much lower temperature. The boil-off rate is approximately 0.5 to 1.5 percent per day for a well-insulated marine LH2 tank, which is materially higher than the 0.1 to 0.3 percent per day for an LNG tank. The boil-off gas requires either re-liquefaction (rare on board), use as fuel in the propulsion plant (the dominant approach), or vent through a high-altitude vent stack (a safety fallback).

Compressed gaseous hydrogen (CH2) at 350 or 700 bar and ambient temperature is the simplest storage mode but the lowest volumetric energy density at approximately 2.8 GJ/m3 (350 bar) or 4.7 GJ/m3 (700 bar). The tank is typically a Type IV composite-overwrapped pressure vessel with a thermoplastic liner and a carbon-fibre overwrap, sized for the working pressure and a safety factor of 2.25. The 350 bar pressure level is dominant for inland and short-sea ferries, while 700 bar is dominant for high-density passenger vessels and small fuel-cell installations. The compression energy is approximately 8 to 12 percent of LCV at 350 bar and 13 to 18 percent of LCV at 700 bar, materially less than the 35 percent liquefaction penalty.

Liquid Organic Hydrogen Carriers (LOHC) are reversible hydrogenation systems where a liquid carrier (typically dibenzyltoluene or methylcyclohexane) is hydrogenated at the production site, transported as an ambient-temperature liquid, and dehydrogenated at the use site to release the hydrogen. The carrier is recovered and returned to the production site for re-use. The hydrogenated LOHC has a volumetric hydrogen density of approximately 6 to 7 weight percent and approximately 2.0 to 2.2 GJ/m3 of usable hydrogen energy, materially worse than LH2 but with the advantage of ambient-temperature liquid storage. The dehydrogenation is endothermic and requires approximately 25 to 30 percent of the LCV of the released hydrogen as heat input at 250 to 350 degrees Celsius, which is a substantial parasitic load on the propulsion plant. Marine LOHC pilot installations have been demonstrated at small scale through Hydrogenious LOHC Technologies and Chiyoda.

Methanol-as-carrier is the use of methanol as a hydrogen carrier with on-board reforming to release hydrogen for fuel-cell use. The methanol-reforming reaction (CH3OH + H2O to CO2 + 3H2) at 250 to 300 degrees Celsius over a copper-zinc-aluminium catalyst delivers approximately 18 percent hydrogen by mass on the methanol feedstock, and the hydrogen is fed to a low-temperature PEM fuel cell. The CO2 produced in the reforming step is vented (cancelling the chemical advantage of using hydrogen at the fuel-cell side), so the overall TtW CO2 is the same as direct methanol use. The methanol-as-carrier approach is appropriate when the on-board fuel-cell stack benefits from the higher purity of reformed hydrogen relative to direct-methanol fuel cells.

For deep-sea hydrogen-fuelled vessels, the LH2 mode is the dominant choice for newbuilds at scale, with CH2 reserved for short-sea and inland operations and LOHC and methanol-as-carrier reserved for niche applications. The MS Topeka Ro-Ro under the HySHIP project is the reference newbuild for marine LH2 at approximately 1,000 m3 of LH2 storage.

Liquefaction energy penalty (~35% of LCV consumed)

The liquefaction of hydrogen from gas at ambient temperature to liquid at minus 253 degrees Celsius requires the removal of approximately 14 MJ of heat per kg of hydrogen liquefied. Adding the inefficiency of the cryogenic refrigeration cycle (Carnot efficiency at minus 253 degrees Celsius is approximately 8 percent, and real-system second-law efficiency is approximately 25 to 30 percent), the actual electrical energy demand for hydrogen liquefaction is approximately 12 to 14 kWh per kg of hydrogen liquefied. The LCV of hydrogen is 33.3 kWh/kg, so the liquefaction energy demand represents approximately 36 to 42 percent of LCV. The conventional rule of thumb in the hydrogen industry is 35 percent of LCV consumed in liquefaction, which is the value used in the MEPC.391(82) and FuelEU calculations.

The liquefaction energy demand is supplied as electricity to the cryogenic refrigeration cycle. The carbon intensity of that electricity is the determinant of the liquefaction-stage WtT contribution. A green-hydrogen plant in Norway with a Norwegian-grid liquefier at 20 gCO2eq/kWh consumes 12 kWh per kg, releasing approximately 0.24 kg CO2eq per kg H2 from the liquefaction stage, or approximately 2.0 gCO2eq/MJ on the LCV basis. A grey-hydrogen plant in Texas with a Texas-grid liquefier at 400 gCO2eq/kWh consumes the same 12 kWh per kg, releasing approximately 4.8 kg CO2eq per kg H2, or approximately 40 gCO2eq/MJ on the LCV basis.

The liquefaction-stage attribution is therefore the swing variable in the WtW intensity of liquid hydrogen. A green-hydrogen molecule liquefied on a coal grid carries the coal-grid intensity for 35 percent of its LCV, which can transform an otherwise 5 gCO2eq/MJ green molecule into a 100 to 150 gCO2eq/MJ liquid product that fails the RFNBO threshold. The liquefier-grid attribution rule under RED III requires the liquefier to be supplied by the same renewable-electricity certificate as the electrolyser, or by a separately certified renewable-electricity supply, for the resulting liquid hydrogen to retain RFNBO status.

The liquefaction stage is concentrated at coastal hub locations where renewable-electricity supply, large-scale electrolysers and bunker-port infrastructure can be co-located. The two reference hub designs at scale are the NEOM Green Hydrogen project in Saudi Arabia (4 GW of renewable electricity feeding a 600 tonne/day electrolyser and liquefier complex with offtake to Asia and Europe) and the Aukra Hydrogen Hub in Norway (1.5 GW of renewable electricity feeding a 100 tonne/day liquefier with offtake to the European Northern Range bunker ports). At smaller scale, the Hamburg, Rotterdam and Oslo bunker terminals are developing 1 to 5 tonne/day pilot liquefiers for short-sea and ferry offtake.

The compression alternative for shipboard storage avoids most of the liquefaction penalty. Compression to 350 bar consumes approximately 8 to 12 percent of LCV (3 to 4 kWh per kg), and compression to 700 bar consumes approximately 13 to 18 percent of LCV (4.5 to 6 kWh per kg). The compressed-gas mode is therefore the preferred choice when the volumetric energy density penalty is acceptable, which it is for short-sea, inland and harbour-craft operations where the bunker frequency is high and the tank volume is not the binding constraint.

Fuel-cell propulsion vs hydrogen ICE

Marine hydrogen propulsion can use either an electrochemical fuel cell or an internal-combustion engine (ICE), with materially different efficiency, emissions and architecture profiles.

Proton-exchange-membrane (PEM) fuel cells operate at 60 to 80 degrees Celsius with a Nafion-style ion-conducting polymer membrane, a platinum catalyst layer, and a gas-diffusion-layer assembly. The cell-level efficiency is approximately 50 to 60 percent on LCV basis, and the system efficiency at the stack and balance-of-plant boundary is approximately 45 to 55 percent. The PEM fuel cell is the dominant marine fuel cell for installations below 5 MW, with reference suppliers Ballard Power Systems, Plug Power, Cellcentric and PowerCell. The PEM stack life is approximately 30,000 to 50,000 hours of operation, after which the membrane and catalyst layer require renewal. The PEM operates only on high-purity hydrogen (better than 99.97 percent on the ISO 14687 grade D specification), so the bunker hydrogen must be delivered at fuel-cell-grade purity.

Solid-oxide fuel cells (SOFC) operate at 700 to 1,000 degrees Celsius with a yttria-stabilised zirconia electrolyte, a nickel-zirconia anode, and a perovskite cathode. The cell-level efficiency is approximately 55 to 65 percent on LCV basis, and the system efficiency including waste-heat recovery can reach 70 to 80 percent in a combined-heat-and-power configuration. The SOFC is the emerging marine fuel cell for installations above 5 MW, with reference suppliers Bloom Energy, Solid Power and Mitsubishi Heavy Industries. The SOFC accepts impurity tolerance and can operate on reformed natural gas, ammonia or methanol in addition to pure hydrogen, which makes it the leading candidate for ammonia-fuelled and methanol-fuelled SOFC ship designs. The high operating temperature creates thermal-cycling stress that limits stack life to 40,000 to 80,000 hours and requires careful start-up and shutdown protocols.

Hydrogen internal-combustion engines are conventional spark-ignition or compression-ignition diesel engines modified to combust hydrogen as the principal fuel. The thermal efficiency is approximately 38 to 45 percent on LCV basis, comparable to a marine diesel engine. The TtW CO2 from a hydrogen ICE is zero (no carbon in the fuel), but the high combustion temperature produces NOx in the same range as a diesel engine, requiring a full SCR after-treatment stack. The hydrogen ICE is the preferred choice when the propulsion power exceeds the practical fuel-cell range (above 10 MW shaft power) or when the operational profile includes transient load changes that strain a fuel-cell stack. The reference suppliers for marine hydrogen ICE are MAN Energy Solutions (with a hydrogen-diesel dual-fuel two-stroke under development), Wartsila (with the Wartsila 25H hydrogen four-stroke), and CMB.TECH (with retrofit hydrogen-diesel dual-fuel conversions).

The choice between fuel cell and ICE depends on the operational profile. A short-sea ferry with a 20-minute crossing and high transient load benefits from the fuel-cell efficiency at part-load (the fuel-cell efficiency curve is flat from 20 to 100 percent of rated load, while the diesel engine peaks at 80 percent and falls steeply at part-load). A deep-sea bulk carrier with a steady-state cruise at 70 percent MCR and infrequent manoeuvring benefits from the lower capital cost of the hydrogen ICE and the longer overhaul interval. The fuel-cell mode is the long-term decarbonisation pathway for hydrogen-fuelled vessels, but the hydrogen ICE is the bridge technology for the 2027 to 2035 newbuild window where fuel-cell stacks at marine power level remain capital-intensive.

Pilot fuel needs in dual-fuel ICE

A hydrogen-diesel dual-fuel ICE requires a small pilot-fuel injection of conventional fuel (typically MGO) to ignite the main hydrogen charge in the cylinder. The pilot fuel is needed because hydrogen has a high autoignition temperature of approximately 585 degrees Celsius and a wide flammability range (4 to 75 percent by volume in air), which makes spark-ignition controllable but compression-ignition challenging. The pilot fuel acts as the ignition source in the diesel-cycle dual-fuel configuration, with the hydrogen injected at high pressure into the cylinder shortly after the pilot ignites.

The pilot-fuel fraction in marine hydrogen-diesel dual-fuel engines is typically 3 to 7 percent of total energy input on energy basis. The MAN B&W ME-LGIH two-stroke design under development uses a 3 to 5 percent MGO pilot, while the Wartsila 25H four-stroke uses a 5 to 7 percent MGO pilot. The CMB.TECH retrofit dual-fuel conversion of medium-speed diesels uses a 7 to 10 percent MGO pilot, reflecting the larger ignition margin needed in a retrofit installation versus a purpose-designed engine.

The pilot fuel is documented separately on the bunker note and contributes its own WtW intensity to the energy-weighted average. A hydrogen-fuelled vessel with a 5 percent MGO pilot at the conventional MGO Annex II default of 91.2 gCO2eq/MJ adds approximately 4.5 gCO2eq/MJ to the engine-side blended intensity. A vessel that switches to bio-MGO or HVO pilot at approximately 25 to 35 gCO2eq/MJ reduces the pilot contribution to approximately 1.5 to 2.0 gCO2eq/MJ, which is the next decarbonisation step beyond the main fuel switch.

The fuel-cell mode does not require a pilot fuel, so the engine-side WtW intensity is the unblended hydrogen value. A vessel that runs on PEM or SOFC fuel cells with green hydrogen at 5 to 15 gCO2eq/MJ delivers an engine-side intensity at the same value, with no pilot-fuel uplift. The fuel-cell mode therefore retains the full RFNBO benefit on the entire fuel input, which is one of the structural advantages of the fuel-cell architecture for hydrogen-fuelled vessels.

Pilot projects: HySHIP, Hydroville, Energy Observer, FCS Alsterwasser

The current global fleet of hydrogen-fuelled vessels is small, principally short-sea and inland demonstrators with fuel-cell or dual-fuel ICE propulsion. Five reference projects illustrate the technology envelope at 2026.

HySHIP is a Horizon Europe-funded consortium led by Wilhelmsen and Norwegian partners, with the MS Topeka Ro-Ro as the flagship vessel. The MS Topeka has approximately 1,000 m3 of LH2 storage, a 1 MW PEM fuel cell, and a Wartsila hybrid diesel-electric propulsion system that runs the fuel cell as the principal energy source with diesel backup. The vessel started commercial operation in 2024 on the Aukra-Bergen-Stavanger short-sea route in Norway, supported by the Aukra Hydrogen Hub liquefier. The HySHIP project is the global reference for marine LH2 at scale and the basis for the Norwegian short-sea hydrogen value chain.

Hydroville is a 14-passenger commuter ferry operated by CMB.TECH on the Antwerp port, with a hydrogen-diesel dual-fuel medium-speed diesel engine providing approximately 440 kW of propulsion power. The vessel uses 350 bar compressed hydrogen storage and runs a 7 to 10 percent MGO pilot. Hydroville started commercial operation in 2017 and is the reference installation for hydrogen ICE marine propulsion at small scale. CMB.TECH has extended the platform to a 200-passenger Hydrocat workboat for the European offshore wind market and is developing the Hydrobingo passenger ferry for the Japanese and Australian markets.

Energy Observer is an autonomous research catamaran demonstrating an integrated renewable-energy system: solar panels, wind kites, a vertical-axis wind turbine, and a PEM electrolyser that produces hydrogen on board from sea water and renewable electricity. The hydrogen is stored at 350 bar and consumed in a Toyota PEM fuel cell that supplies the electric propulsion system. Energy Observer started its global voyage in 2017 and has visited more than 50 countries on a multi-year demonstration campaign. The vessel is the reference installation for fully renewable shipboard hydrogen production and consumption, and the project has spun out a separate hydrogen-fuel-cell newbuild design for inland and short-sea operators.

FCS Alsterwasser was the first marine PEM fuel cell passenger vessel, operated on the Hamburg Alster lake from 2008 to 2013. The vessel had a 100 kW Proton Motor PEM fuel cell, 350 bar compressed hydrogen storage at 50 kg, and an electric propulsion system delivering passenger service for up to 100 passengers. The Alsterwasser was decommissioned in 2013 after the demonstration campaign, with the lessons fed into the subsequent Norwegian and Danish ferry programmes. The vessel is the historical reference for marine PEM fuel cells.

Hydrogen Challenger is the working name for several hydrogen-fuelled support vessels and ferries currently in design or construction, including the Norled MF Hydra (the first liquid-hydrogen ferry, in service since 2023 on the Hjelmeland-Skipavik route), the ABB-built Sea Change (a 75-passenger fuel-cell ferry on the San Francisco Bay), and the Wilhelm Tham (a Swedish hydrogen passenger vessel under refurbishment). The pilot fleet at end of 2025 totals approximately 30 to 40 hydrogen-fuelled vessels globally, almost all under 100 metres LOA, with the first deep-sea hydrogen-fuelled vessel expected for delivery in 2027 to 2028 under the EU-funded H2RES and EverLoNG programmes.

IGF Code amendments under development

The IMO IGF Code (International Code of Safety for Ships using Gases or other Low-flashpoint Fuels) was adopted in 2017 and currently provides Part A (general requirements) and Part A-1 (specific requirements for natural gas as fuel). Part B is the framework for additional fuels through Interim Guidelines, with separate circulars for methanol (MSC.1/Circ.1621), ammonia (MSC.1/Circ.1687) and LPG. The IGF Code amendments for hydrogen are under development at the IMO Sub-Committee on Carriage of Cargoes and Containers (CCC), with adoption targeted for MSC 110 in May 2027 and entry into force on 1 January 2028.

The principal technical scope of the hydrogen IGF amendments covers the cryogenic LH2 storage envelope, the Type IV composite-overwrapped pressure vessel (COPV) standard for compressed CH2 storage, the materials qualification for hydrogen-embrittlement resistance, the boil-off gas management and venting protocol, the bunkering connection standards (SAE J2601 for compressed gaseous hydrogen, ISO 13985 for liquid hydrogen), the leak-detection and ventilation requirements, the fuel-cell-room and engine-room safety case, and the crew training and fire-fighting protocol. The amendments draw on the IMO MSC.420(97) IGC Code for liquefied gas carriers and on the SAE and ISO automotive hydrogen standards, adapted to marine duty cycles and operational envelopes.

The interim guidelines under MSC.1/Circ.1671 (under development) provide a transitional framework for ships using hydrogen as fuel before the formal IGF Code amendments enter force. The interim guidelines are based on the alternative-design approach under SOLAS Regulation II-1/55, with case-by-case approval by the flag administration and the classification society. The principal class societies (DNV, ABS, Lloyd’s Register, Bureau Veritas, ClassNK, RINA) have published hydrogen rules sets that align with the interim guidelines and provide a basis for newbuild approval ahead of IGF Code entry into force.

The bunkering safety case is a particular focus of the IGF Code amendments. Hydrogen has a wide flammability range (4 to 75 percent by volume), a low ignition energy (0.02 mJ versus 0.25 mJ for natural gas), and a high flame propagation speed, which create a more demanding ventilation and detection envelope than natural gas. The hydrogen flame is invisible to the naked eye and emits weakly in the visible spectrum, so flame detection requires UV or IR sensors, not the visible-spectrum CCTV used for natural-gas leak monitoring. The bunkering procedure includes a mandatory inert-gas purge of the transfer line before and after the connection, a high-density leak-detection grid in the bunkering manifold area, and an emergency-shutdown chain integrated with the bunker-barge or terminal control system.

Boil-off gas management

The cryogenic storage of liquid hydrogen at minus 253 degrees Celsius produces continuous boil-off gas (BOG) due to the inevitable heat ingress through the tank insulation, the support structure and the bunker-line connections. The boil-off rate for a marine LH2 tank is approximately 0.5 to 1.5 percent per day for a 1,000 m3 tank with multi-layer insulation, materially higher than the 0.1 to 0.3 percent per day for an LNG tank at minus 162 degrees Celsius because of the larger temperature differential between the cargo and the ambient environment. A laden vessel with 1,000 m3 of LH2 at 70.85 kg/m3 produces approximately 350 to 1,000 kg of BOG per day, which represents a substantial energy and emissions stream that must be managed.

The principal BOG management approaches are use as fuel, re-liquefaction, and venting. Use as fuel is the dominant approach for hydrogen-fuelled vessels: the BOG is routed to the fuel-cell stack or the hydrogen ICE through a low-pressure gas-handling skid, with the molecule consumed for propulsion or hotel load. The approach is the most energy-efficient and has zero direct emissions, but it requires a continuous fuel-demand profile that can absorb the BOG rate. A vessel at anchor or in port with low fuel demand may produce more BOG than the fuel system can consume, which forces a fall-back to re-liquefaction or venting.

Re-liquefaction is the conversion of BOG back to liquid hydrogen through a cryogenic refrigeration cycle on board the vessel. The energy demand for re-liquefaction is approximately 12 to 14 kWh per kg of BOG re-liquefied, similar to the shoreside liquefaction figure. The re-liquefier is a substantial capital investment (typically 5 to 10 percent of the cargo-system capital) and requires a continuous electrical supply that is itself a fraction of the fuel input. Re-liquefaction is appropriate for ships with long port turnarounds or cargo carriers (LH2 carriers) where the cargo cannot be consumed as fuel.

Venting is the emergency fallback when the BOG production exceeds both the fuel consumption and the re-liquefaction capacity. The vent is routed through a high-altitude vent stack with a flame arrestor and an inert-gas purge, and the released hydrogen disperses into the atmosphere. The venting has zero direct GHG impact (hydrogen is not a direct GHG, although it is an indirect GHG with a GWP100 of approximately 12 through its impact on atmospheric methane lifetime), but it represents a complete loss of the energy and the emissions value of the vented molecule. The MEPC.391(82) framework treats vented hydrogen as a 100 percent loss of the WtT investment, and the FuelEU verifier requires the vented quantity to be excluded from the compliance accounting.

The BOG management problem is sharpest at port and during bunkering. The bunker-line cool-down before transfer requires the displacement of the warm gas in the line, typically vented through a return line back to the supply tank. The bunker-tank initial cool-down on a newbuild vessel requires approximately 5 to 10 percent of tank volume to be vented as BOG before the tank reaches operational temperature. The MEPC.391(82) and FuelEU frameworks accept the cool-down losses as part of the bunker-supply chain WtT intensity, with the figures included in the Annex 1 defaults.

Formula, assumptions, and limits

Formula

The well-to-wake intensity of a hydrogen grade in MEPC.391(82) and FuelEU Annex II accounting takes the form:

\text{EF}_{\text{WtW,H}_2,g} = \text{EF}_{\text{WtT,production},g} + \text{EF}_{\text{WtT,liquefaction}} + \text{EF}_{\text{TtW}}

where g indexes the production grade (grey, blue, green), EF_WtT,production,g is the well-to-tank intensity at the plant gate for that grade as certified or as the Annex 1 / Annex II default, EF_WtT,liquefaction is the additional WtT intensity from the liquefaction step (only when the bunker is delivered as cryogenic liquid), and EF_TtW is the tank-to-wake intensity at the engine boundary (zero for fuel-cell propulsion, small NOx-related contribution for hydrogen ICE).

The liquefaction-stage contribution is calculated as:

\text{EF}_{\text{WtT,liquefaction}} = 0.35 \cdot \text{LCV}_{\text{H}_2} \cdot \text{EF}_{\text{grid}}

where 0.35 is the liquefaction energy demand as a fraction of LCV, LCV_H2 is the lower heating value of hydrogen (33.3 kWh/kg or 120 MJ/kg), and EF_grid is the carbon intensity of the electricity supply at the liquefier site (gCO2eq/kWh on a lifecycle basis).

For a blend of grey and certified-green hydrogen, the blended WtW intensity is:

\text{Blend WtW} = (1 - x_{\text{green}}) \cdot \text{EF}_{\text{grey}} + x_{\text{green}} \cdot \text{EF}_{\text{green}}

where x_green is the energy fraction of certified green hydrogen in the bunker mix.

The pilot-fuel adjustment for a hydrogen-diesel dual-fuel ICE engine-side energy mix is:

\text{WtW}_{\text{engine,mix}} = (1 - p) \cdot \text{WtW}_{\text{H}_2} + p \cdot \text{WtW}_{\text{pilot}}

where p is the pilot energy fraction (typically 0.03 to 0.10).

Derivation

The WtT term aggregates upstream emissions from feedstock extraction or sourcing (natural gas for grey/blue, water and renewable electricity for green), hydrogen production (steam-methane reformer, autothermal reformer, or electrolyser), and the production-site balance-of-plant (PSA, dryers, compressors). Each step is quantified in gCO2eq per MJ of hydrogen delivered at the plant gate, and the values are summed.

The liquefaction term is separate from the production term because it can be sourced from a different electricity supply than the electrolyser. For green hydrogen, the RED III rules require the liquefier electricity to satisfy the same RFNBO criteria as the electrolyser electricity to retain the RFNBO status of the resulting liquid hydrogen. For grey and blue hydrogen, no such restriction applies, and the liquefier-grid intensity is taken at face value.

The TtW CO2 term for hydrogen is set to zero by chemistry. Hydrogen (H2, molecular weight 2.016) contains zero carbon atoms, so combustion or electrochemical oxidation produces only water as the principal product. The TtW intensity is dominated by NOx emissions in the ICE pathway (with indirect N2O and atmospheric chemistry contributions, typically 1 to 3 gCO2eq/MJ post-SCR) and is essentially zero in the fuel-cell pathway.

The pilot-fuel term is zero in fuel-cell propulsion (no ignition source needed) and 3 to 10 percent of total energy input in dual-fuel ICE propulsion. The pilot contribution to the energy-weighted WtW intensity is the pilot-fuel WtW intensity multiplied by the pilot energy fraction.

Assumptions

The framework assumes:

The bunker certificate accurately documents the production pathway and the WtT intensity, with full chain-of-custody traceability under a recognised voluntary scheme (CertifHy, ISCC EU, REDcert or equivalent).
The mass-balance accounting is implemented with verified physical chain of custody through commingled storage and pipeline transport.
The engine-side TtW value conforms to the type-test certified value, which is verified in service through periodic emission-monitoring requirements under the IMO NOx Technical Code.
The liquefier-grid attribution conforms to the RED III rules for RFNBO eligibility, with separate certified renewable supply for the liquefier when claiming RFNBO status on the liquid hydrogen.
The pilot-fuel stream is documented separately on the bunker note and contributes its own WtW intensity to the energy-weighted average.
The renewable-electricity supply for green hydrogen satisfies the RED III additionality, temporal and geographical correlation criteria.
The boil-off gas is consumed in the propulsion system or re-liquefied on board; vented BOG is excluded from the compliance accounting.

Worked example

A vessel bunkers 100 tonnes (12 TJ) of certified green liquid hydrogen at a WtW intensity of 18 gCO2eq/MJ (5 gCO2eq/MJ at the plant gate plus 13 gCO2eq/MJ from liquefaction at a Norwegian-grid liquefier), with a 5 percent MGO pilot at the conventional MGO Annex II default of 91.2 gCO2eq/MJ. The blended engine-side WtW intensity is:

\text{WtW}_{\text{mix}} = 0.95 \times 18 + 0.05 \times 91.2 = 21.66 \text{ gCO}_2\text{eq/MJ}

The vessel’s compliance value under FuelEU 2026 with a target of approximately 88.0 gCO2eq/MJ is 88.0 minus 21.66 = 66.34 gCO2eq/MJ of surplus per MJ of fuel consumed. With the RFNBO 2x multiplier applied to the green-hydrogen fraction, the effective surplus rises further for the hydrogen portion of the bunker, with the arithmetic detailed at /calculators/fueleu-rfnbo-multiplier.

A grey-hydrogen counterfactual at 145 gCO2eq/MJ (104 plant-gate plus 41 liquefaction at a fossil grid) with the same 5 percent MGO pilot yields a blended intensity of approximately 142 gCO2eq/MJ, which is in deficit by approximately 54 gCO2eq/MJ versus the 2026 target. The deficit triggers the FuelEU penalty mechanism per /wiki/fueleu-penalties-pooling-multipliers, at a substantially larger penalty per tonne than any current marine fuel.

A blue-liquid-hydrogen case at 50 gCO2eq/MJ (35 plant-gate plus 15 liquefaction at a low-carbon grid) with the same pilot yields approximately 52 gCO2eq/MJ blended, which is in surplus by approximately 36 gCO2eq/MJ versus the 2026 target. The blue-hydrogen case demonstrates that even without the RFNBO multiplier, a high-quality CCS-based hydrogen stream delivers strong FuelEU compliance value when the liquefaction is on a low-carbon grid.

A fuel-cell-only counterfactual on green liquid hydrogen at 18 gCO2eq/MJ (no pilot fuel) delivers an engine-side intensity of 18 gCO2eq/MJ unchanged, with a compliance surplus of 70 gCO2eq/MJ versus the 2026 target. The fuel-cell mode therefore extracts the full RFNBO benefit on the entire fuel input, which is one of the operational arguments for fuel-cell propulsion over hydrogen ICE on FuelEU compliance grounds.

Edge cases and limits

Coal-gasification (brown) hydrogen from China carries a WtW intensity of approximately 200 to 280 gCO2eq/MJ before liquefaction, and 270 to 380 gCO2eq/MJ as liquid product, which is more than three times the VLSFO baseline. The pathway is unviable under any current EU compliance framework and is excluded from RFNBO and certified-blue offtake markets.
Liquefier-grid mismatch for green hydrogen: a green-hydrogen plant in Norway whose liquefier is supplied by a grid connection rather than the electrolyser’s renewable PPA loses RFNBO status on the resulting liquid hydrogen, even when the upstream electrolytic hydrogen satisfies the criteria. The bunker certificate must document the liquefier-grid certification.
Methane leakage in blue-hydrogen gas supply: a blue-hydrogen plant on a 2.5 percent upstream methane leakage gas supply delivers a worse WtT intensity than a 0.5 percent leakage supply by approximately 25 to 30 gCO2eq/MJ, which can erode the entire CCS benefit. The certificate must document the gas-supply leakage value, and certain US Permian Basin and Russian gas streams have leakage rates that disqualify them from credible blue-hydrogen sourcing.
Boil-off gas venting: a vessel that vents BOG due to inadequate fuel-demand profile or re-liquefaction capacity is treated as having lost the WtT investment in the vented quantity, and the verifier excludes the vented mass from the compliance accounting. The operational protocol must minimise BOG venting through fuel-demand management or shoreside re-liquefaction.
Pilot fuel stream: a hydrogen-ICE vessel that bunkers green hydrogen but uses conventional MGO as pilot retains the fossil pilot intensity on the energy-weighted average. The transition to bio-MGO or HVO pilot fuel is the next decarbonisation step beyond the main fuel switch. The fuel-cell pathway avoids the issue entirely.
Mass-balance versus book-and-claim: blue and green hydrogen certification under RED III follows mass-balance, which requires physical chain of custody through commingled storage and pipeline transport. Book-and-claim allocation is not currently accepted under FuelEU Annex II, and a vessel that bunkers grey hydrogen with a green RFNBO certificate (decoupled certificate) cannot claim the green WtW value.
Hydrogen as indirect GHG: vented hydrogen has a GWP100 of approximately 12 through its impact on atmospheric methane lifetime and ozone formation. The current MEPC.391(82) and FuelEU frameworks do not yet incorporate the hydrogen GWP into the compliance accounting, but the next revision (under discussion at MEPC 84 and beyond) is expected to add a hydrogen-leakage term that would penalise vented BOG and bunker losses.
Fuel-cell stack life: a hydrogen-fuel-cell vessel with a 30,000 to 50,000 hour stack life faces a periodic capital event for stack replacement, which is typically 8 to 12 percent of the vessel capital cost and occurs every 5 to 8 years of operation. The lifecycle accounting should include the embodied emissions of the stack manufacture, which add approximately 0.5 to 1.5 gCO2eq/MJ to the WtW intensity over the vessel’s operational life.

Regulatory basis

IMO MEPC.391(82), 2023 Guidelines on Lifecycle GHG Intensity of Marine Fuels (LCA Guidelines), Annex 1 default emission factors per fuel and pathway including hydrogen production and liquefaction.
Regulation (EU) 2023/1805 (FuelEU Maritime), Annex II default WtW emission factors and Article 5(7) RFNBO multiplier.
Directive (EU) 2023/2413 (RED III), sustainability and GHG-saving criteria for RFNBOs including the 70 percent threshold.
Commission Delegated Regulation (EU) 2023/1184 (RFNBO additionality, temporal and geographical correlation rules).
Commission Delegated Regulation (EU) 2023/1185 (RFNBO GHG calculation methodology including liquefaction-grid attribution).
IMO IGF Code Part A and the under-development hydrogen amendments (target adoption MSC 110 in May 2027).
IMO MSC.1/Circ.1671 (Interim Guidelines for the safety of ships using hydrogen as fuel, under development).
IMO NOx Technical Code (engine type-test methodology for hydrogen ICE).
IPCC AR5 / AR6 GWP100 values, with methane at 28 and the emerging hydrogen GWP at 12.
ISO 14687 grade D specification for hydrogen fuel quality at the fuel-cell inlet.
ISO 13985 standard for liquid hydrogen bunkering connections.
SAE J2601 standard for compressed gaseous hydrogen bunkering protocol.

Common errors

Treating hydrogen as a zero-emission fuel because it contains no carbon. The TtW CO2 is zero (or near-zero in the ICE pathway), but the WtT intensity of grey hydrogen exceeds VLSFO and the liquefaction step can dominate the lifecycle outcome on a fossil grid.
Omitting the liquefaction stage from the WtW calculation. The 35 percent of LCV consumed in liquefaction at the liquefier-grid intensity is a substantial term and is the swing variable in liquid-hydrogen WtW outcomes.
Applying the RFNBO 2x multiplier to blue hydrogen. Blue hydrogen does not qualify as an RFNBO because the hydrogen feedstock is fossil-derived; only green hydrogen with full RED III RFNBO certification qualifies.
Conflating green hydrogen with green liquid hydrogen. The plant-gate value of 1 to 15 gCO2eq/MJ is the green-hydrogen WtT, but the liquid-hydrogen WtT can be 5 to 35 gCO2eq/MJ once the liquefaction stage is added.
Assuming book-and-claim allocation is accepted under FuelEU Annex II. The current rule requires mass-balance with physical chain of custody, and a decoupled certificate does not support a green-hydrogen WtW claim on a vessel that bunkers grey or blue molecules.
Using AR4 GWP of 25 for methane instead of AR5 / AR6 GWP of 28 in the upstream methane leakage calculation for grey and blue hydrogen.
Ignoring the pilot-fuel contribution in dual-fuel ICE propulsion. The 3 to 10 percent MGO pilot adds 3 to 9 gCO2eq/MJ to the engine-side blended intensity and is non-trivial against an otherwise low hydrogen WtW value.
Treating compressed hydrogen and liquid hydrogen as the same product on the WtW basis. Compression to 350 or 700 bar consumes 8 to 18 percent of LCV, while liquefaction consumes 35 percent of LCV. The compressed pathway has a materially smaller energy penalty and is the preferred mode for short-sea and inland vessels.
Applying the LCV of 120 MJ/kg uniformly. The MEPC.391(82) and FuelEU frameworks specify the LCV at the bunker manifold, which can include hydrogen-quality losses from the bunker chain (purity, water, oxygen, nitrogen contaminants) that reduce the effective LCV by up to 1 percent.
Counting the hydrogen GWP in the current FuelEU calculation. The hydrogen indirect-GHG effect is not yet incorporated in the FuelEU or MEPC.391(82) framework, although it is under discussion for the next revision.