Ammonia's path toward becoming a mainstream marine fuel has moved from being purely a concept to something with real operating hours behind it, but the seminar's nine presentations, taken together, tell a more precise story than either "ammonia is ready" or "ammonia is still years away." Readiness is improving unevenly across a genuine ecosystem of separate, interdependent pieces, supply, engines, safety frameworks, training, ports, and each piece is progressing at its own pace.
The State of Readiness, According to GCMD
The Global Centre for Maritime Decarbonisation opened the seminar with a readiness assessment spanning eight distinct categories: ammonia supply, engine development, safety and risk, crew training, regulatory readiness, bunkering, environmental impact, and environmental attributes. Their assessment credits genuine progress on several fronts specifically: existing ammonia production and distribution infrastructure, built over decades for the fertiliser industry, can kick-start scaling of low-emission ammonia as a fuel rather than requiring entirely new infrastructure; engines are now type-approved or already in service; IMO's interim safety guidelines were approved in December 2024 and again in May 2026; and key ports are actively readying themselves for ammonia bunkering specifically.
GCMD's timeline of industry demonstrations is the clearest evidence that this has moved beyond planning. Since the first pipeline-to-ship transfer in Singapore in March 2024, at least eleven distinct bunkering and transfer demonstrations have taken place globally, spanning ship-to-ship transfers, truck-to-ship bunkering, and transfers at anchorage and at berth, across locations including South Africa, Gibraltar, Japan, Australia, Rotterdam, China, Oman and South Korea. Three more are planned or tentatively scheduled for 2026 and 2027, including a second GCMD-led demonstration in Singapore.
Two Gaps GCMD Flags as Still Genuinely Unresolved
Despite that progress, GCMD was specific about two categories of problem the industry has not yet solved.
The first is a risk gap: limited understanding of how ammonia actually behaves in enclosed spaces onboard vessels. GCMD points directly to the 2017 Elk Grove Village ammonia release, a real-world industrial incident, as the reference case illustrating why further validation of ventilation requirements and additional dispersion data are still needed to refine detection and mitigation strategies specifically for the shipboard environment, which differs meaningfully from the industrial and agricultural settings where most existing ammonia handling experience comes from.
The second is an operations gap: there is currently no prescribed operating procedure for achieving a "dry-disconnect" when uncoupling a bunkering line, the point in any fuel transfer where residual product in the hose or coupling is most likely to escape. Current standards remain high-level and goal-based rather than prescriptive. Connector hardware already exists, quick-connect/quick-disconnect couplings, dry-disconnect couplings, bolted flanges, and clamp couplings among the options, but GCMD notes each comes with its own operational limitations and trade-offs, and no consensus has formed yet around which approach, or combination of approaches, the industry should standardise on.
What GCMD's Cost Analysis Shows About E-Ammonia's Economics
Working with BCG, GCMD modelled the levelised cost of hydrogen (LCOH) as the dominant driver of e-ammonia's fuel-use cost, both now and looking toward 2050. In their 2030 scenario, e-ammonia costs an estimated USD 67 per gigajoule, of which 60% is LCOH itself, 12% is conversion cost to ammonia, 6% is hydrogen storage and logistics, and the remaining 22% covers downstream storage, logistics, bunkering and opportunity costs. By 2050, GCMD projects total cost falling to USD 52 per gigajoule, with LCOH's share actually holding roughly steady at 58%, driven by projected declines in electricity cost (from USD 39-46 to USD 30-36 per MWh) and electrolyser capital cost (from roughly USD 1,220 to USD 700 per kW). The practical implication: whatever policy or investment intervention is aimed at making e-ammonia cost-competitive, the single highest-leverage point remains the cost of low-carbon hydrogen itself, not the downstream logistics that get comparatively more attention in bunkering-specific discussions.
Why GCMD Frames Trust, Not Technology, as the Next Gap to Close
GCMD's closing argument was more structural than technical: to scale ammonia responsibly, the industry needs to prove its emissions reductions are credible, traceable and comparable across different certification schemes, not just technically achievable. They break this into three specific requirements: life-cycle accounting that actually captures full life-cycle GHG emissions rather than only combustion-stage figures; a harmonised monitoring, reporting and verification (MRV) framework that allows interoperability between different certification schemes, platforms, regulators and technologies, rather than each operating as a separate silo; and integration of that verified data into actual commercial value, since trusted, verified data is what unlocks green premiums, preferential financing and buyer confidence in practice. Without that trust layer, GCMD's argument runs, even a fully mature technical and safety ecosystem for ammonia won't translate into a functioning commercial market.
The IMO's Own Regulatory Picture, From the Secretariat
Camille Bourgeon, the IMO Secretariat's Technical Officer for Climate Action and Clean Air, gave the most granular walkthrough of where IMO's actual rulebook stands.
On fuel consumption reporting: the 2025 IMO Data Collection System (DCS) report, due for publication as MEPC 85/6, will show zero tonnes of ammonia fuel consumption reported for that year, an honest baseline reflecting that ammonia-fuelled ships have only just begun entering commercial operation. Several ammonia-fuelled and ammonia-ready vessels, spanning gas carriers, bulk carriers and containerships, are currently under construction or on order, meaning future DCS reports are expected to start showing real ammonia fuel consumption, uptake patterns across ship types, and ammonia's actual contribution to shipping's decarbonisation for the first time.
On air pollution rules: existing MARPOL Annex VI requirements already apply directly to ammonia-fuelled ships without needing new provisions specific to ammonia. Regulation 13's NOx emission limits apply to marine diesel engines above 130kW regardless of fuel type; Regulation 14's SOx rules effectively don't constrain ammonia itself, since ammonia combustion is sulphur-free, though any pilot fuel used alongside it remains subject to Regulation 14; and Regulation 18's fuel quality documentation requirements, including the Bunker Delivery Note and fuel supplier's declaration, apply as they would to any marine fuel.
On lifecycle GHG accounting: because ammonia contains no carbon atoms, its combustion generates no direct CO2 emissions, but Bourgeon was explicit that future accounting will need to capture other greenhouse gases, particularly N2O, across both upstream and downstream stages, per the 2023 IMO GHG Strategy. The 2024 LCA Guidelines already establish a general well-to-wake methodology covering all marine fuels including ammonia. The specific numeric default emission factors that methodology will use are still being scientifically reviewed by GESAMP's LCA working group, which has already reviewed several proposed well-to-tank default factors representing different ammonia production pathways, including steam methane reforming and autothermal reforming with and without carbon capture. Separately, draft guidelines for sustainable fuel certification schemes and a fuel lifecycle label are under development, intended to let operators eventually use actual certified emissions values rather than only generic defaults.
On safety regulation specifically: MSC 111, meeting in May 2026, approved amendments to SOLAS II-1 clarifying that the IGF Code applies to ships using gaseous fuels regardless of flashpoint, amendments expected to enter into force in July 2028 following adoption at MSC 112 in December 2026. A separate but related clarification, described as the "one ship, one code" principle, confirmed that the IGF Code does not apply to gas carriers using ammonia as fuel; those vessels instead fall under the IGC Code, the code already governing gas carriers generally. In the interim, before those SOLAS amendments take effect, two sets of interim guidelines are already active: MSC.1/Circ.1687, approved December 2024, covers ammonia-fuelled ships other than gas carriers, using a goal-based approach covering fuel containment, bunkering, machinery spaces, ventilation, gas detection and emergency shutdown systems, with an explicit objective of matching the safety level of comparable conventional oil-fuelled installations; MSC.1/Circ.1702, approved May 2026, covers gas carriers specifically, including use of cargo boil-off as fuel, with its objective set against comparable LNG-fuelled installations rather than oil-fuelled ones.
On crew training: STCW.7/Circ.27, also approved at MSC 111 in May 2026, sets non-mandatory interim training guidelines ahead of formal STCW Code amendments. Bourgeon flagged a specific technical distinction driving the guidelines' design: ammonia's primary hazard is toxicity, not just flammability, which requires a genuinely different training approach than the fire-focused training built for LNG or methanol. The guidelines establish a two-tier structure, basic training for any crew member with a designated ammonia-related safety duty, and advanced training specifically for masters, engineer officers, and those directly responsible for ammonia fuel systems, built around a competence-based framework of defined knowledge, understanding and proficiency standards.
On liability and compensation, the area Bourgeon flagged as most clearly unresolved: the IMO's Legal Committee, at its April 2026 session (LEG 113), agreed that a genuine regulatory gap exists covering liability and compensation for incidents involving alternative fuels used for ship propulsion. The existing Civil Liability Convention and Bunkers Convention, which cover spills of hydrocarbon mineral oil, simply don't apply to ammonia. The HNS Convention will apply once it enters into force in November 2027, but only for ammonia carried as cargo, not as fuel. Work has now begun at the Legal Committee to determine whether existing liability regimes need to be adapted or extended to adequately address alternative fuels generally, including LNG, ammonia, hydrogen and biofuels, with an intersessional correspondence group established to report back to LEG 114 in spring 2027.
Bourgeon's presentation also confirmed the IMO's own seminar programming sequence: this was the fifth in the FFT project's technical seminar series, following onboard carbon capture (September 2025), the broader energy transition of shipping (November 2025), biofuels (February 2026), and methane-based fuels (May 2026). The sixth, tentatively scheduled for early 2027, is expected to cover ethanol and methanol specifically.
Yara's View From the Production Side
Yara's presentation grounded the seminar in industrial-scale reality: global ammonia production runs around 200 million tonnes annually, with only about 9-10% of that actually traded internationally, since most production is consumed close to where it's made, upgraded into urea and other finished fertiliser products. China alone produces 57 million tonnes annually, more than three times the next-largest producer, India, at 18.7 million tonnes.
Yara's demand projection is the more consequential figure for marine fuel specifically: total ammonia demand is projected to rise from 186 million tonnes in 2023 to roughly 470 million tonnes by 2050, with new applications, including marine fuel, power generation co-burning and hydrogen carrier use, accounting for the majority of that growth beyond conventional fertiliser and chemical demand. That is a market Yara frames as needing to develop essentially from scratch on the transport side, since current marine ammonia use is negligible against that projected 2050 scale.
On production pathways, Yara laid out the standard three-tier framework: conventional "grey" ammonia, made from natural-gas-derived hydrogen via steam methane or autothermal reforming; low-carbon "blue" ammonia, the same process but with the resulting CO2 captured and stored; and renewable "green" ammonia, made using hydrogen from electrolysis powered by renewable electricity. Yara's own renewable ammonia project in Porsgrunn, Norway, uses a 24 MW PEM electrolyser to replace roughly 3% of the plant's natural gas feedstock, yielding an annual renewable ammonia capacity of 20,000 tonnes, RFNBO-certified as of August 2026, intended to supply the container vessel Yara Eyde.
On the specific marine fuel demonstration project: Yara Eyde, developed in partnership with North Sea Container Line and CMB Tech as vessel owner, is a 1,417 TEU ammonia-fuelled container vessel running a weekly roundtrip between the Oslo Fjord and continental Europe, trading for Yara and other customers, intended as an extension of the emission-free shipping model Yara already runs on the Yara Birkeland route between Herøya and Brevik. The vessel is expected to run on renewable and low-carbon ammonia from Yara's own upstream project portfolio, with delivery from the shipyard estimated for the end of 2026.
On infrastructure: Yara operates 18 ammonia terminals and 14 vessels globally as part of what it describes as its "world scale" ammonia system, moving 1.7 million tonnes in external sales alongside roughly 3.7 million tonnes of internal consumption in Europe and 0.5 million tonnes in the rest of the world. A new import terminal at Brunsbüttel, Germany, adds up to 3 million tonnes of annual import capacity specifically to supplement fertiliser production, positioned on the North Sea and the Kiel Canal. Separately, Yara's Sluiskil facility in the Netherlands is described as Europe's largest cross-border carbon capture and storage project for low-carbon ammonia production specifically.
On pricing: Yara's presentation included Argus Media price assessment data from 17 August 2026 showing the scale of the current cost gap between conventional and low-carbon ammonia bunker fuel across major hubs, expressed relative to VLSFO. In Northwest Europe, "grey" ammonia priced at USD 747 per tonne VLSFO-equivalent above conventional fuel, "blue" ammonia at USD 1,822 above, and "green" ammonia at USD 1,850 above. In Singapore, the gap was similarly stark: "grey" ammonia at USD 562 above VLSFO, "green" ammonia at USD 2,612 above. Yara's own read on this data is direct: conventional ammonia has established price benchmarks, but the low-carbon and renewable ammonia market specifically remains illiquid, with pricing today assessment-based rather than transaction-based, expected to improve toward genuine liquidity only as bunkering volumes and vessel deployment increase in practice.
Rotterdam's Bunkering Pilot: What Actually Happened, and What It Cost in Restrictions
The Maersk Mc-Kinney Møller Center for Zero Carbon Shipping, presenting jointly with the Port of Rotterdam under the EU-funded MAGPIE project, gave the seminar's most granular account of an actual completed bunkering operation.
The pilot, executed on 12 April 2025 with partners including OCI, James Fisher Fendercare, APM Terminals, Trammo and Victrol, transferred roughly 500 tonnes (approximately 800 cubic metres) of ammonia between the vessels Oceanic Moon and Gas Utopia, cooled to -33°C, over 2.5 hours, conducted outside normal working hours and designed specifically to test and validate international standards in practice. Rotterdam's own framing of the lesson learned was explicit: taking community stakeholders along in the development process is essential to building a long-term sustainable environment for the fuel, not merely a public-relations consideration.
Rotterdam's Port Readiness Level (PRL) framework, a nine-stage scale running from "fuel relevance assessed" through to "bunkering of specific fuel integrated in regular port operations," positions ammonia specifically at PRL 6, framework demonstrated in a protected environment, as of the pilot's completion, moving toward PRL 7, project-basis bunkering, and beyond that PRL 8-9, a complete, qualified and commercially available system, which the port expects only once EU funding is secured for dedicated bunker operators. By comparison, the same framework places methanol already at PRL 8-9, fully integrated, underlining how much further ammonia-specific bunkering infrastructure has to travel relative to methanol's more mature position.
The port's 11-part Bunkering Safety Framework, covering governance, port resilience and emergency response, external safety, control zones, terminal preparation, nautical safety, vessel safety, mooring safety, bunker safety, simultaneous operations (SIMOPS) and operational safety, is what the port describes as the basis for "due diligence" in ammonia bunkering: all 11 topics need to be satisfied for a specific bunker operation before the required level of safety for a licence to operate is considered achieved.
The additional, port-specific restrictions Rotterdam layered on top of that framework, beyond generic international requirements, are worth naming directly since they show how conservatively the port is currently approaching the fuel: cold bunkering only (not pressurised); large transfer rates require two lines above 1,000 cubic metres per hour; a zero-emission requirement via an after-release mitigation system; emissions permitted only during disconnection and capped at 25 parts per million; mandatory spill containment via a knock-out facility; a total prohibition on discharging ammonia-contaminated water into harbour waters, requiring a dedicated contaminated water containment system; and mandatory external supervision of operations at PRL 7. Rotterdam's stated ambition for "new fuels" generally is that first commercial bunkerings become possible around project-basis readiness (PRL 7) by 2027, with full system maturity (PRL 8) targeted by around 2029-2030.
The Belgium/Lloyd's Register Presentation: Real Vessels, Named and Specified
A joint presentation from Belgium's FPS Mobility and Transport and Lloyd's Register grounded the regulatory discussion in two specific, named, currently operating or delivered vessel classes.
The first: 210,000 DWT bulk carriers operated by CMB.TECH with Anglo-Eastern as technical manager, flagged in Belgium, classed by Lloyd's Register, built at CSSC Beihai in China, running WinGD 6X72-DF-A dual-fuel engines. The second: 46,000 cubic metre ammonia carriers operated by Exmar, also Belgian-flagged and Lloyd's Register-classed, built at HD Hyundai Heavy Industries in South Korea, running WinGD 6X52-DF-A engines, a vessel class directly connected to the EXMAR ANTWERPEN covered elsewhere in this series' maritime coverage. WinGD's own presentation, discussed further below, identifies this same Exmar/HD Hyundai vessel as the world's first ocean-going commercial ammonia vessel, giving both presentations a shared, verifiable reference point.
The presentation's regulatory pathway diagram traces a consistent through-line from existing gas industry knowledge (via the IGF and IGC Codes) to dual-fuel engine technology to reduced GHG emissions, built on a five-part regulatory foundation: IGC Code amendment, guidelines, engine approval, training requirements, and an operational fleet actually generating real-world data.
The presenters were candid about lessons drawn directly from operational feedback: crew mindset, work discipline, situational control and awareness, prioritising safety, ergonomics and personal protective equipment, and trust in the underlying system design, alongside the practical need for a genuinely reliable fuel supply. Their closing key points named revision of guidelines as an ongoing, active need, not a completed task, alongside continued emphasis on sharing operational experience from the established gas shipping industry into the newer ammonia-specific context, since much of the accumulated safety knowledge transfers directly rather than needing to be developed from zero.
The Engine Manufacturers' Own Data: EUROMOT's Three Presentations
EUROMOT, which has held consultative status at the IMO since 1993 representing engine designers and manufacturers, gave the seminar's engine-technology session over to three of its own member companies presenting independently: Wärtsilä, Everllence, and WinGD. Read together, they give the most granular, company-verified performance data in the entire seminar.
Wärtsilä's four-stroke platform. The Wärtsilä 25 Ammonia engine traces a development line from fuel-lab combustion testing in 2019 through full-scale research engine testing in 2021, product development from 2022, conversion of a laboratory engine from diesel to ammonia in 2023, ammonia type approval and first customer Factory Acceptance Test in 2025, and customer operation starting in 2026. The engine spans a 1.9-3.4 MW power range across 6-9 cylinder configurations. Wärtsilä reports ammonia supplies roughly 95% of the engine's energy share across a wide load range, with a small pilot fuel providing ignition, and states fuel efficiency of its low-pressure Otto concept has proven consistent with equivalent dual-fuel LNG engines. On emissions, Wärtsilä reports CO2-equivalent emissions reduced by roughly 90% on a tank-to-wake basis compared with a diesel engine running on light fuel oil, with the remaining emissions attributable to pilot fuel combustion and a small amount of N2O; testing shows an N2O catalyst can improve on this further but is not required to reach good levels, and NOx aftertreatment via Wärtsilä's existing NOx Reducer system has shown consistent Tier III compliance. Wärtsilä frames its safety approach as built on proven LNG dual-fuel safety principles, with automatic trip to diesel mode on any gas-mode failure, comprehensive personnel safety training backed by accessible PPE, and gas detection paired with planned ventilation. Three named commercial projects are underway: conversion of Norwegian shipowner Eidesvik's platform supply vessel Viking Energy; a newbuild for Grieg Skarv's cargo vessel; and two new gas carriers for Navigator Amon Shipping.
Everllence's two-stroke platform. Everllence's ME-LGIA engine, presented by Dorthe Jacobsen and Johan Hult, builds on what the company describes as a decade of dual-fuel experience across more than 2,100 combined orders spanning its methane, methanol, ethane and LPG dual-fuel engine families; the ammonia variant currently has 7 engines on order, a genuinely early-stage number set against that broader base. Everllence's design uses two entirely separate fuel injection systems for fuel oil and ammonia, with high-pressure liquid ammonia injection at roughly 650 bar, a small diesel pilot flame for ignition consuming about 5% of fuel at 100% load, and an explicit safety concept built around immediate switching to diesel operation on any failure in ammonia mode. Testing across a 4-cylinder unit at Everllence's Copenhagen Research Centre (2023-2026), plus separate test programmes with Mitsui in Japan and HHI in Korea, feeds two already-delivered first commercial engines: a 13 MW unit in Japan and a 12 MW unit in Korea. On emissions specifically, Everllence's data is detailed: NH3 slip measured below 10 ppm at the stack after SCR treatment; NOx held to roughly 70% margin below the Tier II limit without urea dosing, or 50% margin below Tier III with urea; and N2O emissions intrinsically low at 1-3 ppm, close to the detection limit of the FTIR measurement equipment used, meaning no additional N2O mitigation was found necessary. Everllence's own summary statement is direct: the toxicity challenge of ammonia as a marine fuel remains a crucial focus point requiring a responsible approach, a caveat offered alongside, not instead of, the company's confidence that its engines are now tested, certified, approved and ready to order.
WinGD's landmark commercial vessel. WinGD's presentation, delivered by Anirudh Purayil, covers the vessel this seminar's other presentations reference indirectly: the world's first ocean-going commercial ammonia vessel, powered by WinGD's X-DF-A engine, built for Exmar at HD Hyundai Heavy Industries (HHI-EMD and HD Hyundai Mipo) under Belgian flag and Lloyd's Register class, the same vessel class covered in the Belgium/Lloyd's Register presentation above. WinGD reports more than 40 X-DF-A engine orders now placed across gas carriers, bulkers, feeders and tankers, which the company frames as evidence of industry-wide movement rather than a single prototype. The vessel's official shop test achieved fuel consumption in line with estimates, roughly 5% pilot fuel consumption, and better fuel efficiency in ammonia mode than in diesel mode; emissions testing showed very low engine-out NH3 slip (near zero downstream of SCR), NOx emissions Tier II and Tier III compliant and under half of diesel-mode NOx, and very low N2O requiring no dedicated aftertreatment, together yielding greater than 90% GHG reduction on an engine-cycle, tank-to-wake basis. The vessel was delivered to its customer on 10 June 2026 following successful sea trials that logged over 100 cumulative hours in ammonia mode at stable operation up to 85% load. Its maiden voyage, from Nanjing, China to Kakinada, India, logged more than 230 cumulative hours in ammonia mode, with no ammonia leakage or odour detected in the engine room and post-voyage inspection confirming components and sealings remained in good condition; the only two trips back to diesel mode during that voyage were related to experience-building around cargo handling effects on the ammonia fuel supply system, not equipment failure. WinGD's own read on where regulatory work still needs to happen echoes a theme running through multiple presentations at this seminar: the "gas safe machinery space" concept currently carried over directly from the IGF Code was designed with flammability and explosion risk as the primary concern, while ammonia's defining hazard is toxicity, meaning standardised training and the human element deserve focus alongside the physical engineering safeguards already well underway.
Bureau Veritas: The Certification Details Still Being Worked Out Engine by Engine
Patrick Jan of Bureau Veritas gave the seminar's most technically granular account of what actually has to be resolved, case by case, before an ammonia engine can be certified, working from IACS's Unified Requirement M78 Revision 4, which member classification societies have agreed to apply uniformly from July 2027, with engine makers agreeing to apply it ahead of that formal date.
UR M78's core technical content addresses risk analysis (the interaction between off-engine and on-engine systems and the specific hazard of toxic releases), crankcase pressure ventilation, and a maximum 110 ppm ammonia concentration threshold in exhaust gases during normal operation. A structural distinction Bureau Veritas flagged explicitly: engines built to the IGF Code (for ammonia-fuelled ships generally) operate under MSC.1/Circ.1687, with a 110 ppm alarm threshold and 220 ppm triggering the Ammonia Release Mitigation System, while engines on IGC Code vessels (gas carriers) fall instead under MSC.1/Circ.1702, which removes those specific ppm thresholds in favour of referring more generally to the engine's own safety concept, a genuine divergence between the two regulatory tracks this seminar's IMO Secretariat presentation also touched on.
On crankcase and lubricating oil specifically, Bureau Veritas listed several questions still being worked through on a case-by-case basis rather than settled by a single uniform answer: whether crankcase venting in ammonia mode should route to atmosphere, to an ammonia treatment system, or be redirected to the engine exhaust line for dilution; whether gas detection and redundancy are required at each of those points; and whether double-wall containment is needed for crankcase ventilation paths. On lubricating oil, the presentation flagged that ammonia absorption into lube oil needs to be analysed and documented, that some engine makers recommend a separate gas detector above the lube oil separator specifically in case of separator failure, that sludge from that separator should be treated as toxic since it will contain ammonia-water, and that all of this needs to be reflected both in the engine's own safety concept and in the product guide and installation instructions given to customers.
On double-wall piping specifically, Bureau Veritas walked through Everllence's own approach as a worked example of how alternative-design certification actually happens in practice: because Everllence's specific gas-detection monitoring approach for its double-wall ME-LGIA piping deviates from the parameter monitoring set out in MSC Circular 1687's own table, the design has to be validated on a case-by-case basis by the flag administration as a genuine alternative-design process, demonstrating an equivalent level of safety through both design documentation (a hazard evaluation letter specifying Everllence's detection threshold at 5,500 ppm and shutdown at 11,000 ppm, based on extensive testing and running experience) and a Type Approval Test that demonstrates the claimed safety level in practice rather than only on paper. A related open question Bureau Veritas raised directly: where a high-pressure ammonia fuel booster pump is located in the engine room itself, as a pressure-boosting unit rather than strictly an injection pump, it remains an open question whether that arrangement receives classification approval under flag acceptance in the same way IACS's draft rules for methyl/ethyl alcohol-fuelled vessels already accept injection pumps as part of the engine system.
On the broader maturity of the technology, Bureau Veritas was direct: ammonia engines are an innovative product, not yet a mature technology, which is precisely why close, early communication between engine makers and classification societies about internal test results, including any remarks or observations from overhaul inspections, matters specifically for informing type approval certification, and why IACS's existing Stage A internal testing requirements under UR M71 may themselves need updating to reflect the lower current maturity of this specific engine category, ensuring sufficient cumulative running hours are logged to properly validate critical component design robustness, injection systems specifically named as an example, before certification is finalised.
Beyond Carbon: What the Environmental Science Adds
Dr. Sofia Esquivel Elizondo's presentation, drawing on her own peer-reviewed research alongside co-authors including Gilliam, Kershaw, Sartzetakis and Cabbia Hubatova, deliberately shifted the seminar's frame away from ammonia's carbon-free combustion chemistry and toward its broader footprint as a source of what she terms reactive nitrogen, or Nr: any form of nitrogen capable of readily interacting with living systems and the environment. Her presentation's opening framing is stark and worth stating plainly: around 80% of all human-produced reactive nitrogen is lost to the environment annually, contributing to climate change, air and water pollution, and biodiversity loss, and ammonia as a marine fuel adds a new source to that existing global burden rather than existing in isolation from it.
The specific reactive nitrogen species relevant to ammonia marine fuel span the full value chain, not just the point of combustion: N2O (a potent, long-lived greenhouse gas and, notably, an ozone-depleting substance not controlled under the Montreal Protocol), NOx (an air pollutant and short-lived climate forcer contributing to smog, respiratory problems and acidification), and NH3 itself (a fertiliser input in small quantities but toxic and an aerosol precursor at scale). Sources span production (combustion used for hydrogen production, process heat and electricity), fugitive emissions (tank breathing, pressure relief events, loading and unloading operations, pipeline and transfer system leaks, and emergency venting), boil-off gas from heat ingress into cryogenic storage, ammonia slip from incomplete combustion, and both direct N2O from production and engines and indirect N2O formed after NH3 and NOx are deposited and subsequently transformed in the natural environment, plus direct releases to water through accidental spills or nitrogen-bearing effluent, a mechanism this seminar's Korean Register presentation examined in independent technical detail.
Esquivel Elizondo's most consequential point is that the same quantity of deposited reactive nitrogen produces very different ecological outcomes depending on where it lands, since shipping routes cross marine ecosystems with fundamentally different sensitivities. Coral reefs face nitrogen enrichment that disrupts the coral-zooxanthellae symbiosis, favours macroalgae over coral, and increases vulnerability to thermal stress and bleaching, with a correspondingly low N2O conversion yield (under 0.1% of deposited nitrogen). Oligotrophic ocean gyres, naturally nutrient-limited environments, see increased primary production that disrupts that natural limitation and shifts phytoplankton communities in ways that affect the ocean's own carbon sequestration capacity, also at a low N2O yield. Oxygen-minimum zones see a markedly higher N2O conversion yield, 1-4% of deposited nitrogen, directly amplifying the climate impact of any given quantity of reactive nitrogen emitted in those specific waters. Coastal regions and marginal seas see both a comparably elevated N2O yield (1-3%) and worsened eutrophication, oxygen depletion, and degradation of seagrass meadows and the ecosystem services they provide. The practical implication Esquivel Elizondo draws from this: uncontrolled reactive nitrogen emissions from ammonia bunkering and combustion could disproportionately threaten specific regions critical to marine biodiversity and fisheries production, depending on where along a given shipping route those emissions actually occur, rather than posing a uniform risk regardless of geography.
On the scale of the problem going forward, Esquivel Elizondo's own published modelling projects that total reactive nitrogen emissions from the ammonia marine fuel value chain could become substantial within the next 15 to 25 years as ammonia production scales, estimated at roughly 0.5% to 4.4% of the total nitrogen supplied to the value chain, with the actual outcome within that range determined directly by how effective emissions mitigation proves to be, not a fixed or unavoidable figure. Her prescribed controls follow a four-part structure: prevent emissions at source through low-emission production and minimising leaks and losses; capture what can't be prevented, recovering and reusing vapours and containing and treating nitrogen-bearing residual streams (the same effluent-management problem Korean Register's presentation addressed from an engineering perspective); control jointly, optimising after-treatment systems for NH3, NOx and N2O together rather than addressing each pollutant in isolation; and verify, through ongoing monitoring of both emissions and environmental risk, reporting transparently, and continuously improving practice over time. Her closing framing is a direct challenge to a narrower, carbon-only view of ammonia's climate credentials: strong lifecycle controls can genuinely preserve ammonia's climate benefits and prevent pollution, but doing so requires the industry to address impacts more broadly than greenhouse gas accounting alone, explicitly including indirect climate warming through indirect N2O, direct reactive nitrogen discharges to water, and downstream effects on marine ecosystems as part of what "getting ammonia right" actually means.
Lloyd's Register: Building the Organisational Readiness Layer
Tom Chitseko, Head of Risk & Safety at Lloyd's Register's Maritime Decarbonisation Hub, presented on behalf of a genuinely broad coalition, developed jointly with the Maersk Mc-Kinney Møller Center for Zero Carbon Shipping and a roster of industry partners spanning V.Group, Maersk, Cargill, NYK Line, Stolt Tankers, Anglo-Eastern, Yara Clean Ammonia, CF Industries, Everllence, Equinor, Hapag-Lloyd, TotalEnergies, SGMF, Seaspan, the Ammonia Safety Training Institute, Maersk Training and Rio Tinto, addressing a level of readiness distinct from what most of this seminar's other presentations covered: not the ship itself, and not the individual seafarer, but the operating organisation as a whole.
Chitseko's framing is that IMO's own regulatory work is already progressing on two of three necessary layers: the ship, covered by the interim safety guidelines (MSC.1/Circ.1687 and 1702, approved December 2024 and May 2026 respectively), and the seafarer, covered by the interim training guidelines (STCW.7/Circ.27, approved May 2026). The third layer, how an operator actually organises itself day-to-day to control major-accident hazards, sits outside what those two regulatory tracks address directly, and is the specific gap this presentation's resources target.
The underlying reference-design assessment behind this work rated four human-factors domains by risk level: competency and training was rated low-maturity across the industry currently, since new skills and knowledge will be needed across many roles that don't yet have them; work process and procedures and occupational health hazards were both rated medium, reflecting the need for substantially changed and more heavily relied-upon functional work practices, alongside practical measures such as restricting access to and time spent in spaces containing ammonia equipment and increased use of remote monitoring; and process safety hazards were rated high, reflecting that managing ammonia safely to a high standard will require companies to further develop their planning, emergency preparedness and maintenance practices specifically as guided by the existing ISM Code framework, rather than assuming existing conventional-fuel safety management systems transfer over unmodified.
Two concrete tools, both released in August 2026, operationalise this organisational-readiness layer. The first, a Process Safety Management (PSM) Maturity Self-Assessment Tool, borrows an approach with a genuine track record controlling major hazards in other industrial sectors, built around the core concept of "keeping the process in the pipe": concerned specifically with major hazards capable of causing catastrophic events, covering the whole hazard-control arc from prevention through preparedness, mitigation, response and restoration, and deliberately integrative across design, technical, operational and leadership dimensions functioning together rather than as separate silos. The tool assesses 16 elements drawn from the established CCPS Risk-Based Process Safety Management framework, each rated across five maturity bands from Ad-hoc through Reactive, Standardised, and Proactive to Generative. Chitseko was explicit about what the tool is not: it's intended to help an organisation understand its own maturity level, not to produce a single competitive "score," and while the ultimate decision that a given operation is ready for ammonia remains with operators and regulators rather than the tool itself, higher PSM maturity does correspond to a higher likelihood of genuine organisational readiness. Maturity is also explicitly not meant to be averaged across all 16 elements uniformly; different elements can carry different priority depending on context, and the tool's real value lies in the structured evaluative process of making that assessment, producing specific, evidence-based next steps and priority ratings for each element rather than a single overall verdict.
The second tool, a set of three Competency and Training Practitioner Handbooks covering Leak Detection, Management, Isolation and Repair (18 competencies across 7 groups), Maintenance (28 competencies across 6 groups), and Emergency Response (116 competencies across 10 groups), maps role-specific competencies across both the ship and shore sides of the ammonia value chain, covering shipboard management, operations and support roles alongside shore-based corporate, port authority, terminal, bunker barge, reception facility and salvor roles. Each individual competency entry specifies four elements: the competency itself (what the person needs to be able to do), the specific training standard that satisfies it, the required level of competence (weighted across knowledge, understanding and proficiency), and explicit role applicability tagged across both ship and shore, with training requirements scaled deliberately in proportion to a given role's actual responsibility and exposure to risk, rather than applying uniform training requirements regardless of role.
Chitseko's closing framing named the tools' intended users directly: for operators, a structured but non-prescriptive way to answer "where do we begin" in maturing their safety management organisation toward ammonia readiness; for port authorities and terminals, shore-side competencies for ammonia already identified and mapped, ready to use rather than requiring separate development; and for training providers, role-differentiated competency sets that can support building modular training programmes now, rather than waiting for formal regulatory training requirements to be finalised.
NYK's Operational Track Record: Two Years, Zero Incidents
NYK Line's presentation offered the seminar's longest continuous operational dataset, and its most candid first-person crew testimony.
NYK has been developing two ammonia-fuelled vessels since 2021 under Japan's Green Innovation Fund, led by NEDO, METI and MLIT: the ammonia-fuelled tugboat A-Tug (delivered August 2024, converted from LNG to ammonia fuel by NYK subsidiary Keihin Dock), and the ammonia-fuelled medium-size gas carrier AFMGC (targeted delivery November 2026, to be chartered by Yara after delivery). The A-Tug, known as SAKIGAKE, has now completed more than two years of commercial operation without a single incident, achieving a nitrogen co-firing rate of up to 95.2% at full main engine load and a corresponding GHG reduction of up to 94.0% on a tank-to-wake basis compared with fuel oil, figures that hold reasonably consistent even at partial engine loads down to 25%.
The AFMGC's propulsion architecture, developed by a consortium including NYK, Nihon Shipyard (NSY), Japan Engine Corporation and IHI Power Systems, uses a two-stroke Japan Engine Corporation main engine burning liquid ammonia, paired with three IHI Power Systems four-stroke auxiliary engines running on gas ammonia, alongside a dedicated ammonia gas abatement system. Additional safety measures specific to this vessel, going beyond baseline requirements, include a Machinery room safety (MRS) notation from ClassNK, awarded specifically for vessels demonstrating excellent ammonia safety measures in the machinery room; a semi-enclosure around the main engine, containing any leak to a limited area rather than allowing dispersion through the wider engine room, backed by gas detection, visible and audible alarms and dedicated ventilation; and design responses to specific failure-mode questions NYK's own engineers raised directly during development, including what happens under a double failure (simultaneous machinery failure and safety system malfunction), how crews can safely evacuate high ammonia concentrations from the engine room, and whether routine maintenance work near a 5-10 MPa ammonia pipe carries acceptable splash risk. NYK's own framing of this process, that its engineers' hands-on operational recommendations were directly incorporated into vessel design and manuals rather than the design being handed down from naval architects alone, is presented as central to why the resulting vessel has operated incident-free.
The most direct evidence in the entire seminar comes from NYK's own crew. Captain Kawai, the A-Tug's master since delivery, previously captained Japan's first LNG-fuelled tug in 2015, reported that toxicity concerns have been alleviated through adequate education and training, that engine output and load response are comparable to conventional diesel engines, and that day-to-day operations have proceeded in the same manner as diesel or LNG-fuelled tugboats, while still flagging that as a development vessel, unforeseen challenges have emerged and additional equipment requirements have increased onboard workload relative to a conventional tugboat. Chief Engineer Takahashi, who took part in the A-Tug's design and commissioning and has served as chief engineer since delivery, offered a specific technical comparison: as a single-component fuel, ammonia is comparatively straightforward to manage operationally, in contrast to LNG, a mixed fuel whose composition can vary with handling in ways that affect engine performance; on the vessel's previous LNG-fuelled configuration, knocking caused frequent sensor failures, while no knocking has been observed on ammonia under fluctuating engine load. On toxicity specifically, Takahashi reported no hazardous exposure to date given appropriate safety measures and PPE, and no ammonia odour detected even during operation in ammonia mode, while still describing toxicity as a genuine ongoing concern requiring due care during maintenance work wherever ammonia contact is possible, rather than a solved problem to be dismissed.
NYK's own key takeaway, stated directly: a serious accident on an ammonia-fuelled vessel could set back wider industry adoption meaningfully, which is precisely why building confidence in ammonia's safe use, through practical, reasonable standards the wider industry can actually rely on, matters as much as any single technical milestone.
Korean Register: The Overlooked Problem of What Comes Out of the Ship
Korean Register's presentation addressed a specific, narrower technical question the rest of the seminar's presentations largely assumed away: what happens to the water used to safely control gaseous ammonia releases, once that water has absorbed ammonia and become a liquid effluent that itself now needs managing.
The underlying mechanism: as global shipping adopts ammonia as fuel, gaseous ammonia releases occur during normal operations such as purging, venting, pressure relief and depressurisation. Existing IMO safety guidance already calls for water-based absorption or scrubbing to control those releases safely, a genuine and necessary safety function, but that absorption process itself generates a new stream, ammonia-containing liquid, that the current IGF Code and existing interim guidelines don't yet specify how to manage through to a defined final destination. This gap is now the subject of formal work at the IMO's Sub-Committee on Pollution Prevention and Response (PPR), first considered at PPR 13, with further consideration at PPR 14, targeting completion in 2028; the presentation itself represents Korean input feeding into that international process via a Korean consultative group spanning government, classification and shipbuilding interests, alongside informal international technical exchange among member states, NGOs, class societies, industry, academia and technical experts, explicitly not itself an IMO decision-making body.
The chemistry driving management complexity: ammonia in water exists in equilibrium between free ammonia (NH3, un-ionised, more volatile, associated with acute toxicity and vapour risk) and ammonium (NH4+, ionised, less volatile, but still representing reactive nitrogen that remains in the system after any neutralisation). Higher pH and temperature shift that balance toward more free ammonia; Korean Register's own illustrative research found pH readings of 11.9 to 12.7 in tested liquid, well into the range where free ammonia dominates. Total ammonia nitrogen (TAN) alone, the presentation stresses, does not by itself define either the hazard or which management route is actually appropriate; salinity, co-contaminants and the liquid's broader composition all factor into route suitability.
On how much liquid this actually generates: Korean Register's illustrative example, based on a very large ammonia carrier with a 10 MW reciprocating engine, an ammonia load of roughly 3,600 kg per hour (about 86 tonnes per day) and a water absorption rate of 600 kg per hour, gives a sense of realistic scale, though the presentation is explicit these are illustrative figures dependent on the specific ship, fuel system, operating profile and absorption basis, not universal sizing values. Different generation scenarios carry very different volumes: planned operational streams such as change-overs or shutdowns might generate 0.5-2 cubic metres per changeover or 2-4 cubic metres per shutdown; foreseeable routed events such as pressure relief valve activation might generate 0.5-5 cubic metres per event; and response-generated water from leakage mitigation or firewater use could range from several cubic metres up to more than 20 cubic metres within roughly eight minutes in an emergency scenario.
On management pathways: Korean Register's framework follows a three-stage pollution-prevention logic, prevent and minimise the liquid at source without compromising safety, control all outputs rather than simply transferring pollution between different phases (gas, liquid, solid), and verify environmental acceptability before any release. Where feasible, recovering ammonia directly, by condensing and recovering it for reuse in the fuel tank or elsewhere, is preferred over generating a liquid effluent stream at all; where recovery isn't feasible, water-based absorption remains a necessary safety function but creates a liquid stream requiring further management, either through stabilisation (acid dosing and pH control, shifting free ammonia to ammonium and reducing re-vaporisation risk, though total nitrogen remains in the stream regardless) or through reuse and recovery as a qualified process feed or recovered product. Whatever residual stream remains after that, Korean Register argues, needs a defined, verified final destination among three genuine options: to air (verified atmospheric release, compliant with applicable NOx, N2O and ammonia slip requirements), to sea (conditional marine discharge, explicitly not currently authorised under existing rules and still subject to IMO PPR consideration), or to shore (retained onboard and transferred to an authorised shore reception facility with confirmed subsequent treatment).
The presentation's core, repeated argument is that treatment performance alone is not sufficient evidence of pollution prevention; what needs to be demonstrated is the complete pathway, tracking nitrogen through every gas, liquid and solid output the treatment process generates, each with a verified, environmentally acceptable, and genuinely available final destination, backed by a ship system capable of safely containing, monitoring and, where the preferred route is temporarily unavailable, retaining that liquid without forcing an uncontrolled release.
Reading All Nine Presentations Together
No single presentation at this seminar claimed ammonia as a marine fuel is a solved problem, and the picture that emerges from reading all nine together is consistent with that restraint, while also being more textured than any single presentation alone would suggest.
Supply-side infrastructure and engine technology have moved furthest, with genuine, company-verified operational data now measured in years for some vessel classes and in hundreds of cumulative operating hours for others: NYK's A-Tug at two-plus years incident-free, WinGD's Exmar vessel at over 230 hours across its maiden commercial voyage, Everllence's two engines already delivered and operating in Japan and Korea. But "ready" engine technology and "ready" certification are not quite the same claim, and Bureau Veritas's presentation is the clearest evidence of that gap: even engines already tested, certified and delivered are still working through case-by-case alternative-design approval for specific components, crankcase venting routes and double-wall piping monitoring among them, precisely because ammonia engines remain, in Bureau Veritas's own words, an innovative product rather than a mature technology.
Bunkering infrastructure, port-level safety frameworks and crew training guidelines are actively being built out and tested through real pilots, but remain earlier-stage and more geographically concentrated than the engine and production side, and GCMD's own operational and dry-disconnect gaps sit alongside this. Liability and compensation regulation, harmonised MRV and certification interoperability, and specific operational procedures such as dry-disconnect protocols and ammonia-containing effluent management remain genuinely open questions still working through IMO's committee process, with target completion dates extending out to 2027 and 2028 rather than settled today.
What the four additional presentations covered here add is a layer the seminar's earlier engineering-and-regulation-focused sessions don't fully capture on their own: readiness is not only a question of ship design and fuel chemistry. Sofia Esquivel's research reframes the entire discussion around reactive nitrogen rather than carbon alone, showing that ammonia's environmental case depends on emissions controls the industry hasn't yet universally adopted, not on the fuel's carbon-free combustion chemistry by itself. Tom Chitseko's presentation, in turn, makes the case that even a fully mature ship design and a fully trained individual seafarer still leave a third readiness gap unaddressed, the operating organisation's own capacity to manage major-accident hazards day to day, a gap current IMO regulation doesn't directly target and industry has only this year begun building shared tools to close.
Taken as a whole, the seminar reads less as a victory lap for ammonia and more as a working status report from an industry that has moved past pure concept work into live pilots, real vessels and named companies, while being unusually candid, across nearly every presentation, about precisely which pieces, technical, regulatory, environmental and organisational, still need to close before the fuel can scale broadly rather than remain a small number of carefully managed first movers.
Source: IMO Technical Seminar on the Use of Ammonia as Marine Fuel, 17 September 2026 (part of the IMO's Future Fuels and Technology (FFT) project seminar series).
