All Technical Articles

Detection Implications of Ammonia-Ready PCTCs

By Vignesh Durai · August 31, 2026 · 4 min read

Ammonia does not interfere with the hydrogen channel. The interference that matters is between H2 and CO — and both are thermal-runaway markers.

An ammonia-fuelled hull puts two gas-detection systems in adjacent compartments, and the intuitive worry — ammonia bleeding into the vehicle-deck hydrogen channel and producing false alarms — does not appear to be the real one. Published electrochemical cross-interference data puts the ammonia-to-hydrogen response at zero. The interference that does matter is between hydrogen and carbon monoxide, and both of those are lithium-ion runaway products.

What ammonia-ready means, precisely

A notation covering a future conversion, not a fuel currently burned. Aurora-class and comparable multi-fuel PCTCs are LNG dual-fuel ships carrying DNV ammonia-ready and methanol-ready notations — the first in the PCTC segment to hold them. That distinction matters for a detection specification, because the ammonia gas-detection duty attaches when the conversion happens, potentially years after the vehicle-deck system was installed and surveyed.

The thresholds that duty will carry are already published. The IMO's Interim Guidelines for the Safety of Ships Using Ammonia as Fuel, MSC.1/Circ.1687, issued 26 February 2025, set detection at 25 ppm in enclosed spaces, 110 ppm in the secondary enclosure and 220 ppm to trigger alarms and shutdowns. Those are toxicity numbers, not flammability numbers — ammonia's lower flammable limit is around 15% by volume, so the enclosed-space alarm sits some three orders of magnitude below anything that can burn.

0% / 0%
NH3 sensor response to H2, and H2 sensor response to NH3 — no mutual interference
22% / 20%
H2 sensor response to CO, and CO sensor response to H2 — both runaway products
25 / 110 / 220 ppm
MSC.1/Circ.1687 ammonia thresholds — enclosed space, secondary enclosure, shutdown
New sensors only
The interference table's own caveat — values vary as cells age

The cross-sensitivity that matters is not the obvious one

Ammonia turns out to be a well-behaved neighbour, and the vehicle-deck gas channel is not. A published cross-interference table for electrochemical sensors gives an ammonia sensor a 100% response to ammonia and zero to every other gas listed, and gives a hydrogen sensor zero response to ammonia. On that data the two systems do not contaminate each other, which removes the failure mode this post previously asserted.

The same table gives a hydrogen sensor a 22% response to carbon monoxide, and a carbon monoxide sensor a 20% response to hydrogen. Read that against what a stressed lithium-ion cell actually vents — hydrogen and carbon monoxide together, not in isolation. A gas channel using hydrogen as its primary runaway marker is therefore reading a signal that carbon monoxide is inflating by roughly a fifth, and a carbon monoxide channel is reading one that hydrogen inflates by about the same. The two markers are not independent measurements of the same event; they partially measure each other.

This corrects the previous version of this post, which stated that sensor housings must tolerate trace ammonia "without false H2 signals". The cross-interference data does not support an ammonia-to-hydrogen response. The real coupling is H2 to CO, inside the runaway signature itself — a more consequential problem, because it cannot be solved by separating compartments.

What ammonia does to the hardware

The exposure risk is chemical rather than electrical. MSC.1/Circ.1687 names material compatibility as one of ammonia's specific hazards alongside toxicity and explosive atmospheres, and the practical consequence for instrumentation in or near a fuel space is corrosion of susceptible alloys rather than spurious readings. A sensor that survives a salt-mist regime has not thereby been qualified for an ammonia environment; they are different exposures with different failure modes.

The interference figures carry their own decay warning, and it is easy to miss. The table states plainly that its cross-interference numbers apply to new sensors only and may vary with time. Every figure above therefore describes a commissioning condition, not a service condition — which matters on a vehicle deck where electrochemical cells sit for years between replacements, and matters more on a hull whose fuel is scheduled to change mid-service.

Two systems, two regimes, one hull

The architectural problem is regulatory before it is technical. The vehicle-deck system answers to SOLAS Chapter II-2 and the FSS Code, where heat detection has been mandatory in vehicle and ro-ro spaces since 1 January 2026. The ammonia system answers to non-mandatory interim guidelines and class rules, because the IGF Code covers gases and low-flashpoint oils and does not cover ammonia at all — which pushes novel arrangements toward SOLAS II-2 Regulation 17 alternative design, an engineering case submitted for Administration approval rather than a certificate held against a clause.

So a single newbuild can carry two detection systems approved by two different mechanisms, alarming on two different physical principles, at thresholds three orders of magnitude apart. Treating them as one procurement is how a specification ends up applying fire-detection logic to a life-safety system, or a toxicity threshold to a fire one.

What to specify

  • Cross-interference data for the specific sensor models proposed, at end-of-life as well as new — the published tables are commissioning-condition figures.
  • Whether the hydrogen channel's threshold accounts for concurrent carbon monoxide, given roughly a fifth of the apparent H2 reading can come from CO in a runaway plume.
  • Material qualification for ammonia exposure as a separate question from salt-mist qualification, for anything sited in or adjacent to a fuel space.
  • Which approval route each system sits on — FSS Code prescriptive, class rule, or Regulation 17 alternative design — because that determines what has to be re-opened at conversion.
  • What happens to the vehicle-deck system's approval when the ship converts, since the fuel change is scheduled inside the detection equipment's service life.
Conclusion

How RoRoSAFE helps

On an ammonia-ready PCTC, cargo fire detection and fuel toxicity detection are two systems with two regimes. RoRoSAFE is the cargo-side one: per-vehicle thermal and battery-vent gas sensing on the vehicle decks, fused so that a thermal-runaway signature is read on more than one channel. It runs alongside the fuel system's ammonia detection and does not replace it.

Pilot: one deck · installed alongside the berth · no drydock · 6 months of dashboard access

Sources

  • 1. IMO — MSC.1/Circ.1687, Interim Guidelines for the Safety of Ships Using Ammonia as Fuel, issued 26 February 2025 (approved MSC 109, December 2024): detection thresholds of 25 ppm in enclosed spaces, 110 ppm in the secondary enclosure and 220 ppm triggering alarms and shutdowns; ammonia's specific hazards identified as high toxicity, potential for explosive atmospheres, and material compatibility challenges. Non-mandatory, aligned with IGF Code and SOLAS principles.
  • 2. Industrial Scientific — published electrochemical gas sensor cross-interference table: NH3 sensor 100% to ammonia and 0% to the other listed gases, including 0% to hydrogen; H2 sensor 0% to ammonia, 22% to carbon monoxide and 30% to nitric oxide; CO sensor 20% to hydrogen. The table states that its figures apply to new sensors only, may vary with time, and are subject to change.
  • 3. IMO — IGF Code (International Code of Safety for Ships Using Gases or Other Low-flashpoint Fuels): mandatory criteria for fuels with a flashpoint below 60°C; does not cover ammonia, methanol, ethanol, LPG or hydrogen, which are handled through interim guidelines and class rules.
  • 4. Höegh Autoliners / Baird Maritime — Aurora-class vessel data: LNG dual-fuel with DNV ammonia-ready and methanol-ready notations, the first in the PCTC segment to hold them.
Frequently asked

Questions, answered

Does ammonia interfere with hydrogen detection on a vehicle deck?+

Published electrochemical cross-interference data says no — an ammonia sensor shows zero response to hydrogen and a hydrogen sensor zero response to ammonia. That removes the intuitive worry about a fuel leak contaminating the runaway gas channel. It is one manufacturer's data for its own cells rather than a universal property, so it should be confirmed against the datasheet for whatever model is actually specified.

Which gas cross-sensitivity actually matters for runaway detection?+

Hydrogen and carbon monoxide. The same table gives a hydrogen sensor a 22% response to carbon monoxide and a carbon monoxide sensor a 20% response to hydrogen — and a stressed lithium-ion cell vents both together. So the two markers are not independent measurements of one event; roughly a fifth of an apparent hydrogen reading can be carbon monoxide, and the reverse.

Why does ammonia detection alarm so far below the flammable limit?+

Because MSC.1/Circ.1687 sets those thresholds on toxicity. Detection is required at 25 ppm in enclosed spaces, 110 ppm in the secondary enclosure and 220 ppm for alarms and shutdowns, while ammonia's lower flammable limit is around 15% by volume — roughly 150,000 ppm. Crew are at risk long before ignition is possible, which inverts the usual fire-detection assumption.

What should an ammonia-ready newbuild specification make explicit?+

Which approval route each detection system sits on, because that determines what re-opens at conversion. The vehicle-deck system answers to the FSS Code, where heat detection has been mandatory since 1 January 2026; the ammonia system answers to interim guidelines and class rules, since the IGF Code does not cover ammonia. Also ask for cross-interference data at end-of-life, not just new.

Related reading

Continue the thread