Why Airlines Cap Dry Ice on Every Flight

If you ship pharmaceuticals by air, you have run into the dry ice limit. A booking comes back capped at fewer kilos than you asked for. One carrier lets you tender 200 kg per shipment, another 100, another wants a separate approval above some threshold. It reads like red tape, invented in different offices by people who never talked to each other. So most shippers treat the number as an annoyance to route around: split the load, book earlier, call the account rep.

The number is not arbitrary. It comes out of a physics problem with exactly one variable that anyone gets to choose. Once you see the problem, the limits stop looking like bureaucracy and start looking like what they are: a budget, spent on a resource nobody actually measures. This post walks the safety logic from the ground up, then gets to why that unmeasured resource is the interesting part.

Dry ice is a dangerous good because of what it does to air

Solid carbon dioxide sublimates. It goes straight from solid to gas at -78.5°C (-109°F), no liquid in between, releasing carbon dioxide gas the entire time it exists. On a warehouse floor that gas dissipates into the room and nobody thinks about it. Inside the sealed, pressurized hull of an aircraft at altitude, there is nowhere for it to dissipate to.

That is the whole reason dry ice is a regulated Class 9 dangerous good, UN 1845, under both the US Department of Transportation rules and the IATA Dangerous Goods Regulations. It is not toxic, not flammable, not corrosive. The hazard is simpler and more physical than any of those: the CO2 it gives off displaces oxygen in an enclosed space. Enough of it, in a small enough volume, and the air the crew breathes is no longer the air they were trained to breathe. The cargo hold and the cabin on most aircraft share a ventilation environment, so the gas from a pallet of frozen biologics is not somebody else's problem.

So the airline's job is to make sure the CO2 concentration in the occupied volume never climbs to a level that matters. That turns a vague safety worry into an equation.

The exposure math, and where 0.5% comes from

The ceiling is a concentration limit. FAA guidance (Advisory Circular 91-76A) puts the acceptable carbon dioxide concentration in the occupied cabin at 0.5 percent by volume, which is 5,000 parts per million. That same 5,000 ppm figure is the OSHA permissible exposure limit for CO2 averaged over an 8-hour workday, so it is not an aviation invention. It is the broadly accepted line for how much CO2 people can breathe, all day, without harm. Below it, normal. Above it, headaches, drowsiness, impaired judgment, the last thing anyone wants in a flight deck.

Pressurization is the second piece. Under the airworthiness standard FAR 25.841, the cabin pressure altitude cannot exceed 8,000 feet under normal operations, even when the aircraft itself is cruising at 40,000. That is why your ears pop on climb and why a bag of chips puffs up: the hull holds the interior at a thicker, lower-altitude air pressure than the thin air outside. That pressurized air is a fixed quantity of atmosphere, and every gram of subliming CO2 mixes into it. The 8,000-foot ceiling defines the volume and density of air that the CO2 has to share.

The loading formula, and the one number that is assumed

Here is how the airline turns those limits into a weight allowance. The maximum dry ice an aircraft can carry works out to:

X = (CO2 concentration limit) × (aircraft volume) × (air exchanges per hour) ÷ (sublimation rate)

Read it left to right and every term is something you can defend. The concentration limit is the 0.5 percent ceiling. The aircraft volume is the ventilated cubic footage of that specific type. The air exchanges per hour is how many times the environmental control system flushes and replaces the cabin air, carrying CO2 out with it. Multiply those three and you get how much CO2 the aircraft can shed per hour while staying under the line. Then you divide by the sublimation rate: how fast the dry ice on board is turning into gas. The faster it sublimates, the less of it you can carry.

Look at what that division does. Allowable dry ice is inversely proportional to the assumed sublimation rate. Assume the ice sublimates at 1 percent per hour and you get one number. Assume 0.5 percent per hour and the aircraft can carry roughly twice as much for the exact same safety margin. Same hull, same ventilation, same 5,000 ppm ceiling. The entire capacity of the aircraft for this cargo hinges on a single assumed rate.

Which is exactly why the FAA went and measured it. The legacy assumption, written into the original Advisory Circular 91-76 back in 1963, was 1 percent per hour per 100 pounds of dry ice. That figure governed loading math for decades. In 2024 the FAA published a study (DOT/FAA/TC-24/24) that put real containers on test and measured average sublimation rates of 0.53 to 0.71 percent per hour, meaningfully below the 1 percent legacy number. Container makers, for their part, advertise rates under 1 percent per hour. International carriage follows ICAO Doc 9284, the Technical Instructions that IATA's rules are built on. The point is not that any one number is right. The point is that a term driving real payload capacity has been an estimate for sixty years, and estimates are conservative for a reason.

What this means when you are the one tendering the box

Now the operational reality falls out of the math, and the annoyances start making sense.

Per-carrier caps and approval steps differ because the inputs differ. A carrier flying a larger aircraft with a higher ventilation rate really does have more CO2 budget than one flying a smaller type. Each also picks its own assumed sublimation rate and its own safety factor on top of it, which is conservative and defensible and also why two carriers quote you two different numbers for the same pallet. There is no single correct answer they are all failing to reach.

Summer biologics volume competes for a fixed budget per aircraft. The CO2 allowance is per flight, not per shipper. In peak season, when everyone is moving temperature-sensitive product at once, your dry ice is competing with everyone else's for the same fixed number of allowable kilos on that tail. That is why a booking that cleared in March gets capped in July. Nothing changed about your box. The aircraft filled up its CO2 budget with other people's boxes.

The declared dry ice weight at tender is a static number that is wrong within hours. You declare, say, 25 kg of dry ice on the dangerous goods paperwork. From the moment it was packed, that number started falling. By the time the shipment is loaded, it is carrying less. The system runs on the declared figure because it is the only figure available, not because it is accurate at the moment of flight.

Freighters and belly cargo carry different budgets. A dedicated freighter with no passengers and a different ventilation setup has a different CO2 allowance than the belly hold under a passenger cabin. Same cargo, different aircraft, different limit. This is not inconsistency. It is the formula, run on different inputs.

The turn: a measured number beats an assumed one, for everybody

Step back and look at what the entire system rests on. On the safety side, the airline protects the crew using an assumed sublimation rate, because the actual rate of the actual ice in the actual box is unknown to them. On the capacity side, the shipper is limited by that same assumption, forced to plan around a conservative worst case because there is no better number to plan around. Both sides are working from the same guess, in opposite directions.

That is the gap worth noticing. The assumed rate is conservative on purpose, and conservative is the right call when you have no data. But conservative also means the aircraft is often carrying less dry ice than it safely could, and the shipper is often accepting a tighter cap than the physics required, all to cover the case where the ice is subliming faster than expected. Nobody knows which shipments those are, so the penalty is spread across all of them.

Per-shipment measured sublimation data serves both ends of that trade at once. If the real CO2 output of a shipment is known rather than assumed, safety improves, because the number reflects the box on the aircraft instead of a 1963 average. And capacity improves, because there is no longer a reason to assume the worst case for a shipment you can actually see. A measured rate is not an argument for carrying more dry ice or less. It is an argument for knowing, instead of guessing, which is the one thing the current system cannot do.

That is the layer we work on at CryoTrak: the actual quantity of dry ice in the actual box, tracked over time, which is exactly the sublimation behavior the whole loading formula assumes and never sees. The caps will always exist, and they should. The air only holds so much CO2. But a limit built on a measured number is a better limit than one built on a sixty-year-old estimate, and it is better for the airline and the shipper on the same day. The next time a booking comes back capped, it is worth remembering that the binding constraint is not the airline. It is a number about your cargo that nobody has ever measured.