Ammonia is a toxic release problem wearing a fire scope

Three of the ten days to 21 September produced an ammonia announcement in the Gulf. That is where a large share of the region’s next process capacity is going, and ammonia is the one major derivative whose protection scope is routinely specified against the wrong consequence.

Ten days of ammonia

On 10 September an EPC contract worth USD 3.465 billion was awarded for a grassroots fertiliser complex in eastern Saudi Arabia: one ammonia train of 1.2 million tonnes a year, two urea trains totalling 2.6 million, a carbon capture unit and supporting infrastructure, commissioning targeted for Q3 2030. It lifts the owner’s urea capacity from 4.8 to 7.4 million tonnes a year, around 54 per cent.

On 15 September the urea technology licence and proprietary equipment package for that same complex was awarded separately at about EUR 125 million, covering two units of 3,850 tonnes per day.

On 11 September an Abu Dhabi producer — the largest seaborne exporter of urea and ammonia — signed a memorandum of understanding with a German polymer producer and a German terminal developer to develop an ammonia import terminal at Hamburg, with the producer evaluating both supply into that terminal and an equity position in it. The stated commercial driver on the buyer’s side is feedstock and hydrogen supply, not fertiliser.

Set against the Ruwais chemicals zone, scheduled to produce 4.7 million tonnes a year from 2028 with low-carbon ammonia among its products and a dedicated chemicals port in development, the pattern is clear. This is not a fertiliser story but a bulk liquefied toxic gas logistics story: new inventory, new refrigerated storage, new jetties, new transfer operations, on sites and in free zones that have mostly handled hydrocarbons.

The design basis inverts

Ammonia is flammable. Its lower explosive limit in air is 15 per cent and its upper limit 28 per cent. That is a narrow band with an unusually high floor, and in open air ammonia does not readily sustain a fire.

Its NIOSH immediately-dangerous-to-life-or-health value is 300 ppm over 30 minutes. That is 0.03 per cent, against a lower explosive limit of 150,000 ppm — five hundred multiples of lethality between the concentration that harms a person and the one that burns.

That ratio is the whole design problem. On a hydrocarbon plant the numbers governing layout, occupied-building siting, escape routes and muster points come from radiation contours and overpressure. On an ammonia plant they come from a toxic dispersion study, and that distance is larger by a wide margin. A scope built around fire produces the smaller envelope and looks complete doing it.

The cloud also behaves against the instinct. Dry ammonia vapour has a relative density of about 0.6, so it is lighter than air and expected to rise and disperse. A liquid release does something else: it flashes, entrains atmospheric moisture and forms a visible aerosol denser than the surrounding air that travels along the ground. Detection located for a buoyant plume, and escape provided at grade, are both wrong for that behaviour.

Where a conventional scope quietly fails

Detector technology and setpoints. Catalytic bead and infrared hydrocarbon detectors are calibrated on hydrocarbons and will not serve here. Ammonia needs ammonia-specific detection with alarm levels set against a toxic endpoint in parts per million. A philosophy that alarms at a percentage of LEL — the normal convention on a hydrocarbon unit — raises its first alarm at a concentration that has been lethal for some time.

Detector placement. Low level, and biased toward transfer points, pump seals and flanged connections, on the basis of a cold dense aerosol rather than a warm rising plume.

Water is a scrubbing medium, not a fire-fighting one. Ammonia is highly soluble in water, dissolution is exothermic, and the product is ammonium hydroxide — a corrosive alkaline solution. Water curtains and monitors are therefore genuinely effective at knocking a cloud down, which is why they appear in most mitigation strategies. The consequence is rarely followed through: every cubic metre applied leaves as alkaline contaminated effluent that must be contained and segregated. Containment volume and drainage routing for that duty are a civil scope item, sized from a mitigation strategy that lives in the fire safety philosophy — two documents usually written by different parties months apart.

Occupied buildings. If the credible major accident is a cloud, the building failure mode is ingress through HVAC intakes, and the controls are intake gas detection, automatic damper closure and a defined shelter-in-place duration. Control rooms are frequently specified for overpressure and left entirely permeable to gas.

Emergency response. A response built around fire water puts people at grade and downwind. Any credible ammonia response is a respiratory protection and exclusion-zone problem first, and a water application problem second.

Which major-hazard regime the contract actually names

Jurisdictional thresholds for ammonia diverge sharply. Process safety management obligations attach above roughly 4.5 tonnes in the United States, above 50 tonnes in the United Kingdom and above 200 tonnes in Australia. Facilities of the scale announced this month sit far above all three, so the question is never whether a major-hazard regime applies. It is which one the contract names, what consequence study that regime requires, and whether the study is issued before layout and occupied-building siting freeze. In the Gulf the applicable framework depends on the authority having jurisdiction and, on free-zone sites, on the zone — so it has to be established in writing, not assumed from the last project.

What to check now

Six questions separate an ammonia scope from a hydrocarbon scope with ammonia written on the cover:

Are gas detector alarm levels set against a toxic endpoint in ppm, or against a fraction of LEL?
Are ammonia detectors positioned for a dense ground-hugging aerosol, or on the same drawing convention as the plant’s hydrocarbon detection?
Is water used for vapour knockdown accounted for in containment volume and drainage segregation, and by whom?
Are occupied buildings specified against gas ingress, with intake detection and damper closure, or only against overpressure?
On export and import terminals, what is the defined response to a release at a loading arm with a vessel connected, and who can order emergency disconnection?
Which major-hazard regime does the contract name, and does the toxic consequence study it requires exist before layout freeze?

The governing event on these facilities will be visible, quiet and at knee height, moving in whichever direction the wind happens to be going. Nothing in a fire scope is aimed at it.

Mainstream Petrochemical Instruments FZE provides fire safety engineering for high-hazard industrial facilities. This article is general technical commentary and is not a design recommendation for any specific facility.