On 31 December 2025 a deadline arrived that had been twenty years in the making. From that date, under European law, no aircraft registered in the EU was allowed to carry a halon portable fire extinguisher in its cabin. Large parts of the fleet missed it. Nothing happened.

Spain’s aviation regulator published the official position three weeks before the deadline, and it is a remarkable document. Aircraft that have not replaced their extinguishers, it says, may carry on flying — while in what it calls an inherent infringement of the regulation. They are breaking the law and they are airworthy, and those two facts simply sit next to each other.

That contradiction runs through the whole story of halon in aviation. The gas was banned worldwide before most of today’s pilots were born. It is one of the most destructive ozone-depleting substances ever manufactured. And it is still sitting, under pressure, in steel bottles plumbed into the cargo hold of very nearly every airliner in the sky — because after three decades of looking, nobody has found anything that works as well.

The ban that was never a ban

Almost every account of this subject gets the central fact wrong, and the error is worth clearing up before anything else. People say aviation has a special exemption from the halon ban. It does not. It has never held one, and it has never needed one.

The Montreal Protocol controls production and consumption — manufacture, import and export. Article 2B drove new halon production to zero in developed countries from 1 January 1994 and in developing countries from 1 January 2010. What the Protocol does not do, anywhere in its text, is restrict the use of halon that was already made. The gas in an airliner’s cargo-bay bottles today was almost certainly manufactured before 1994, and no treaty reaches it.

The UN Environment Programme’s Ozone Secretariat states this plainly: the control measures apply to newly manufactured halon, not to recycled halons, and recycling remains the sole source for applications where alternatives are still under development, as in aviation.

There is a mechanism for permitting new production for an essential use, and aviation has tried to qualify. It fails on a specific limb of Decision IV/25, which allows an essential-use exemption only when the substance is not available in sufficient quantity and quality from existing stocks of banked or recycled controlled substances. Recycled halon is available. So aviation fails the test — not because its need is small, but because the second-hand market is still supplying it.

What the European Union does, uniquely, is restrict use — going beyond the Protocol. Its ozone regulation sets two kinds of date for each aircraft application. A cut-off date is the point after which halon may not be designed into new equipment. An end date is the point after which it must be taken out of service altogether.

Those dates have moved. Regulation (EC) No 1005/2009, as amended in 2010, set the cargo-compartment cut-off at the end of 2018, in the stated expectation that further research and development would deliver alternatives in time for installation in new aircraft. It did not. In February 2024 the regulation was repealed and replaced, and the cargo cut-off was quietly pushed to 31 December 2024. The end date for cargo, engine nacelles, APUs, fuel-tank inerting and dry bays is 2040.

The aerosol can that defeated everything

If halon is so dirty, why has nothing replaced it in a cargo hold? The answer is one test, and it is not the one most people would guess.

To certify a cargo-compartment suppression system, an agent must pass the US Federal Aviation Administration’s Minimum Performance Standard. That means beating four fire threats: a deep-seated fire in bulk-loaded cargo, a fire inside a cargo container, a flammable-liquid fire, and — the one that matters — the explosion of an aerosol can.

A suitcase full of burning clothes will eventually burst a deodorant can, and the resulting deflagration has to not blow the compartment liner apart. Halon manages this. The candidates do not. The technical committee’s 2022 assessment is blunt: every single-component vaporising liquid agent put through the exploding-aerosol-can test — HFC-125, 2-BTP and FK-5-1-12 — caused an undesired pressure rise in the test compartment when discharged below its fire-suppressing concentration. On that basis, all the single halogenated agents tested so far have been found unacceptable.

That is the whole impasse in one sentence. Not that the replacements cannot put a fire out, but that on the way to putting it out they make the aerosol can worse.

A short primer on how cargo compartments are classified, and what each class is required to carry.

Weight is the other wall. For engine nacelles, the same committee tabulates what it would take to match Halon 1301 by equivalent concentration: CF3I needs 1.54 times the mass, Novec 1230 needs 2.25 times, HFC-125 2.70 times and carbon dioxide 2.35 times the mass and five times the cylinder volume. On a cargo system sized in hundreds of kilograms, the two cargo solutions that have passed the standard — a 2-BTP and carbon-dioxide blend, and a water-mist and nitrogen system — add somewhere between 100 and 500 kg per aircraft, according to the industry’s own paper to ICAO.

Neither is in service on anything. The blend is at technology readiness level 6, and the committee notes drily that it is toxic at its design concentration.

Where there is no system at all

There is a part of this story that almost nobody outside the industry knows: a great many aircraft fly every day with cargo compartments that have no fire-suppression system whatsoever, entirely legally.

US airworthiness rules sort cargo holds into classes. A Class C compartment — the lower hold of a passenger airliner — must have a detector and a built-in extinguishing system controllable from the cockpit. A Class E compartment, found only on all-cargo aircraft and typically comprising the entire main deck of a freighter, must have a liner and a detection system and a means of shutting off the ventilating airflow. It is not required to have a built-in fire suppression system. The strategy instead is to starve the fire of air, depressurise, and land as fast as possible.

Class D no longer exists. It was a compartment with neither detection nor suppression, relying on being sealed tightly enough to smother a fire. On 11 May 1996 a ValuJet DC-9 went into the Everglades with 110 people aboard after undeclared oxygen generators ignited in a Class D hold. The FAA eliminated the class outright with Amendment 25-93, effective 19 March 1998 — but with a carve-out: operators had to upgrade Class D compartments to Class C standard, unless the operation was all-cargo, in which case Class E would do.

That carve-out has a body count. On 3 September 2010 a UPS Boeing 747-400 freighter took off from Dubai; twenty-one minutes later a main-deck fire warning sounded. Smoke filled the cockpit within three minutes and obscured the instruments for the rest of the flight. Both crew were killed. The accident board’s finding 46 is one line long.

Less than a year later, on 28 July 2011, an Asiana 747-400 freighter was lost in the East China Sea after a main-deck fire. Two crew killed. Same aircraft family, same deck, same compartment class, same absence.

Mentour Pilot’s account of UPS Flight 6, the accident that put Class E cargo compartments under scrutiny.

The fire halon cannot put out

Halon works by chemistry rather than smothering. Bromine radicals interrupt the chain reaction of combustion, which is why so little of it does so much — and why nothing else comes close on a weight budget.

But chain-breaking has a blind spot, and in 2006 an FAA researcher found it. Testing rechargeable lithium-ion cells, he reported that Halon 1301 easily extinguishes the electrolyte fire at both the 5 per cent knockdown concentration and the 3 per cent suppression concentration — and then added the sentence that matters: Halon 1301 has no cooling effect and did not prevent the release of electrolyte from heated cells. Cells will continue to vent due to the air temperature, but will not ignite in the presence of Halon 1301.

Read that carefully. The flame goes out and stays out. The heat does not. Thermal runaway inside a lithium cell is a self-sustaining exothermic reaction that needs no outside oxygen, so an agent that works by interfering with flame chemistry cannot reach it. The cells keep cooking, keep venting flammable gas, and keep pushing adjacent cells into runaway — a cell in thermal runaway can reach 1,100 degrees Fahrenheit or higher, according to FAA research cited by the US National Transportation Safety Board.

The NTSB spelled out the consequence in a 2016 recommendation letter: the pressure pulse from cells in runaway can activate pressure-relief features and damage cargo covers and liners, which may result in a reduced suppressant agent concentration and rapid fire escalation. The halon leaks out through the holes the fire made.

The FAA’s own short briefing on the lithium battery risk.

Ninety tonnes

Here is the number that should worry anyone who flies. The civil aviation industry’s own dedicated stockpile of Halon 1301, as tabulated by the Montreal Protocol’s technical committee at the end of 2022, is 90 tonnes.

It survives on other people’s. Of a worldwide bank of roughly 34,310 tonnes, most is locked away — Japan holds 16,455 tonnes, militaries hold just over 4,200 tonnes installed and in reserve, oil and gas 1,500, nuclear almost 500, and a further 4,001 tonnes is already installed in aircraft and therefore unavailable. What aviation can actually buy is the 7,331 tonnes still sitting in the world’s computer rooms, plus 199 tonnes at sea, plus its own 90.

When those server-room systems are decommissioned, their halon is reclaimed and sold on. That is the supply chain. There is no global halon bank, no strategic reserve, no central authority — just, as the Ozone Secretariat puts it, just-in-time commercial trade in recovered halon, with a global supply, distribution and demand that is unbalanced. Shipments are sometimes held up at borders because customs authorities misclassify recovered halon as hazardous waste under the Basel Convention.

Modelled run-out dates vary a great deal and the honest answer is a range, not a number. The 2022 assessment ran eight scenarios and got anywhere from 2030 to 2049 depending on supply and emission rates — two to five years earlier than the same committee had projected in 2018. Its 2026 modelling narrowed to roughly 2031–2039, with a central estimate around 2035. The industry’s own paper to ICAO warns that reserves may become insufficient potentially starting around 2030.

Against which, set this: the committee’s 2026 report states that current aircraft designs in production will rely on halon 1301 for at least 50 years, well beyond the projected run-out date. Aircraft rolling off the line today will still want halon in the 2070s. The halon is expected to run out in the 2030s. Those two sentences are in the same document.

A period promotional film for BCF — Halon 1211 — from the years when the industry was still selling the stuff rather than hoarding it.

What has actually changed

It is not all stalemate. Lavatory waste bins were the easy win: HFC-227ea and HFC-236fa went in as near drop-in replacements, and virtually all current production aircraft are fitted with them. Hand-held extinguishers are being solved too, slowly, with 2-BTP — sold as Halotron BrX — which EASA lists among four acceptable halon-free agents and which is gradually replacing Halon 1211 platform by platform.

Engines are stuck. The only approved agents for civil engine nacelles and APUs remain Halon 1301, plus HFC-125 on a military derivative of a commercial airliner — the Boeing 767-based KC-46 — and a customised approval for phosphorus tribromide on one model of business jet. Novec 1230 was developed for nacelles and then failed an FAA cold-soak live-fire test. After abandoning it in the 1990s on toxicity grounds, industry is now circling back to CF3I, which is both the closest thing to a drop-in and the only candidate that would escape a proposed European restriction on PFAS chemicals.

That PFAS problem is the cruel twist. The leading cargo-hold candidate contains 2-BTP, which counts as a PFAS under some definitions — so the replacement for the ozone-destroying chemical may be regulated out of existence before it is ever certified. And the committee notes it is not aware of any new agents under active development. If one were identified tomorrow, it would take at least ten years to commercialise.

The deadline that passed

Which brings us back to December 2025. EASA trailed the hand-held extinguisher end date in a notice that November, and then explained what non-compliance would actually mean. The ozone regulation, it pointed out, is not airworthiness legislation. An infringement would not affect an aircraft’s airworthiness, and there are no grounds obliging member states to ground an aircraft over it.

Spain’s regulator went further on 19 December, publishing the European Commission’s interpretation directly.

Operators were asked to submit a replacement timetable. Cost is a real part of why they had not: in a derogation request filed in January 2025 asking Poland’s ministers to push the end date to 2040, a general-aviation representative set out the arithmetic — a 2.5 lb Halon 1211 extinguisher at around 475 dollars against roughly 1,513 dollars for the 1.92 lb Halotron BrX equivalent, plus disposal costs for the old unit. His warning was that some owners would simply carry no extinguisher at all. It is an advocacy document and should be read as one, but the price gap is not in dispute.

The regulators have, meanwhile, been retreating. ICAO’s standard said that aircraft types applying for certification on or after 28 November 2024 must use something other than halon in the cargo hold. That date passed with no certifiable alternative in existence. At its 42nd Assembly in 2025, ICAO adopted Resolution A42-11, which removed the mandate. Industry had asked for 2035.

And the Montreal Protocol parties, in Decision XXXVII/4 of November 2025, asked their technical panel to go and find out how much halon is actually left and where it is — the first decision aimed squarely at aviation. Thirty-two years after production stopped, the treaty is asking the question that the industry has been quietly asking itself for a decade: is there enough left to see the fleet out?

Nobody yet knows. In the meantime, the safest cargo-hold fire suppression agent ever devised keeps doing a job it was supposed to have stopped doing in 1994, and every aircraft delivered this year commits another forty years of demand to a supply that is not being replaced.

Sources: UNEP Ozone Secretariat (Montreal Protocol Article 2B, Decision IV/25, Halon in Aviation); UNEP TEAP Fire Suppression Technical Options Committee 2022 Assessment Report and 2026 Progress Report; Regulation (EU) 2024/590 and Commission Regulation (EU) No 744/2010; EASA and European Commission, Halon Replacement in the Aviation Industry (February 2025); EASA Certification Memorandum CM-CS-013; EASA newsroom, 6 November 2025; AESA Spain, 19 December 2025; FAA Final Rule Amendment 25-93 (63 FR 8032); 14 CFR 25.857 and FAA AC 25.857-1; FAA reports DOT/FAA/AR-06/38, DOT/FAA/TC-TN12/11 and DOT/FAA/AR-11/31; UAE GCAA Accident Board Findings, UPS Airlines Flight 6; NTSB Safety Recommendations A-16-001 and A-16-002; ICAO A42-WP/74 and A42-WP/395; Flight Safety Foundation, AeroSafety World, September 2009.