Ten Minutes on a Heavier Gas: What a Second Species Adds to the Case for Fluorocarbon Decompression

Dark teal-and-amber science illustration: a cut-away block of tissue at depth with small fast helium molecules streaming away to the left and larger carbon tetrafluoride molecules drifting in from the right, and a single gas bubble in the middle visibly shrinking as the breathing gas changes

Decompression medicine has spent a century arguing about schedules: how deep to stop, how long to stop, how fast to come up. Two papers published this year argue about something more basic and stranger — what a diver breathes during the last ten minutes before the ascent.

The newer of the two appeared in the European Journal of Applied Physiology on 31 October. Li Jiaqi and colleagues at the Naval Medical Center in Shanghai, with co-authors at the PLA General Hospital in Beijing and Shanghai Ocean University, compressed rats on air to 6 ATA or on heliox to 9 ATA, then switched them — ten minutes before decompression began — from the dive gas to a blend of 79 percent carbon tetrafluoride and 21 percent oxygen. Decompression sickness in the air protocol fell from 19 of 28 animals to 3 of 28. In the heliox protocol it fell from 15 of 22 to 4 of 22.

Those are 68 percent down to 11 percent, and 68 percent down to 18 percent. Run the published counts through a Fisher exact test and the two-sided p values come out at roughly 0.00002 and 0.0019. Mortality in the air arm went from 8 of 28 to none, which the authors put at P = 0.004 — a figure that reproduces almost exactly when the same test is applied to the same cells. Echocardiography found fewer or no intracardiac bubbles after the switch, and lung tissue showed less inflammation, less endothelial injury and less histological damage in both protocols. A separate 28-day repeated-exposure arm reported no rises in serum alanine aminotransferase, aspartate aminotransferase, creatinine or urea.

None of that makes carbon tetrafluoride a medicine. What it does is put a second species behind an idea that until this year had mostly been tested in pigs, and it does so under a different background gas, at a different depth, with the injury endpoints that decide whether a bubble is a nuisance or a lesion.

A different lever from the one everyone pulls

Prolonged decompression works. It also costs bottom time, breathing gas, fuel and exposure, and that trade-off sits at the front of the Shanghai paper: every extra minute of ascent buys safety with something the operation cannot spare. Their proposal is the alternative — if you cannot buy safety with time, change what the animal is breathing.

Divers already switch gases. Oxygen at six metres is standard for decompression stops, air breaks are routine, and travel mixes change with depth. A dedicated decompression gas is therefore not a radical suggestion. The radical part is choosing one whose molecules barely fit through tissue.

Diffusivity first, everything else second

Carbon tetrafluoride is small only by the standards of organic chemistry: one carbon, four fluorines, molar mass 88, against 28 for nitrogen and 4 for helium. Graham’s law, applied to molar mass, puts its diffusion rate at about 0.56 times nitrogen’s and 0.21 times helium’s, so tissue loaded with CF₄ equilibrates slowly — and, just as usefully here, unloads slowly.

What counterdiffusion actually buys you

The trick lives in what happens at constant ambient pressure. A rat or a diver who has spent fifty minutes on heliox is carrying helium in tissue, a fast and poorly water-soluble gas that leaves quickly once the gradient reverses. Swap the breathing gas to CF₄/O₂ without changing depth and the helium keeps exiting at helium’s speed while CF₄ enters at a fraction of it. For a while the total dissolved inert gas pressure inside the tissue sits below ambient, and a bubble in that tissue loses gas faster than it gains it. That is isobaric counterdiffusion, the phrase in the title of the earlier of the two papers, and it explains why the switch is timed minutes before ascent rather than hours.

The rat work also suggests the benefit is not only mechanical. Lungs and endothelium were better preserved in the switched animals, which is what you would expect if fewer and smaller bubbles reach the pulmonary circulation and the vessel wall — but the paper does not separate the two mechanisms, and its authors do not claim to.

The density bill that comes with it

Molar mass is also a liability. A 79/21 CF₄/O₂ mix has a mean molar mass of about 76 against 9.9 for 79/21 He/O₂, so gas density multiplies by roughly 7.7 at the instant of the switch, and CF₄/O₂ ends up about 2.6 times as dense as air breathed at the same pressure. At the equivalent of 200 feet of seawater, gas density stops being a trivia question — it is one of the main reasons deep mixes are built around helium in the first place.

Neither published summary reports work of breathing or exercise tolerance as an endpoint, which is precisely what a gas swap made at depth would need to survive before anyone straps it to a working diver.

Duke got there first, with pigs

The nucleus of this line of work is at Duke, in the Department of Anesthesiology and the Division of Pulmonary, Allergy and Critical Care Medicine, with collaborators at the University of Texas Medical Branch’s aerospace medicine division. Their paper, in PNAS in August, is the study the Shanghai group is extending. Sedated swine weighing about 20 kg breathed 79/21 helium-oxygen at a pressure equivalent to 200 feet of seawater for fifty minutes, then were randomised either to keep breathing it or to switch to 79/21 CF₄/O₂ for ten minutes before decompression.

The CF₄/O₂ animals had less decompression sickness and lower mortality, less hypoxaemia and pulmonary oedema, fewer venous gas emboli, less gait disturbance, less rash and less central nervous system haemorrhage. The abstract does not report group sizes, which makes the size of those effects impossible to weigh from the summary alone. The authors’ stated ambition is broader than any single result: deeper dives, longer dives, and rapid decompression without surrendering safety.

What changed between August and October is not the concept but the evidence base — a second species, a second background gas, a deeper exposure at 9 ATA on the heliox arm, and a first set of tissue-level endpoints.

The safety file is real, narrow, and entirely animal

CF₄ has been characterised as a breathing gas at least since the late 1960s, when its anaesthetic potency was measured in dogs alongside sulphur hexafluoride. The contemporary toxicology comes from the same Duke group: Sprague-Dawley rats exposed to 8 ATA air or 8 ATA 79% CF₄/21% O₂ for thirty minutes a day across five days showed normal behaviour, no gross or microscopic pathology, organ weights and haematology near published normal ranges, no renal or hepatic signals on clinical chemistry, a negative polychromatic erythrocyte micronucleus test and no DNA damage in lung on comet assay. One result cut the other way. Females exposed to CF₄ had 2.5 times the percentage of tail DNA in liver tissue that controls did (P = 0.009), which the authors describe as still inside the range of their local negative controls; a bacterial reverse mutation assay across five strains was nonmutagenic.

Every one of those sentences is about rodents, on five exposures, across a finite list of endpoints. The new paper is unusually candid about where its own safety data stop. The 28-day arm, it says, “did not include post-exposure histopathology or haematology, so safety conclusions are limited to the measured endpoints.” No published protocol has put a human being under a CF₄ atmosphere and reported the result.

There is a second ledger that no diving paper is obliged to open. The Environmental Protection Agency lists carbon tetrafluoride with an atmospheric lifetime of 50,000 years and a 100-year global warming potential of 7,380. Research chambers vent trivial quantities. A fleet would not, and operational adoption would eventually have to answer for that too.

What this does not change inside a chamber

Decompression illness is treated by recompression on oxygen, which is where hyperbaric medicine began and where it remains most uncontested. The Undersea and Hyperbaric Medical Society lists decompression sickness as indication 05 of fifteen; Medicare’s national coverage determination names decompression illness among the conditions it pays to treat; the FDA clears hyperbaric oxygen devices for it as Class II devices through the 510(k) pathway under product code CBF. Moon and Mitchell’s 2025 update on decompression sickness recommendations and Mitchell and colleagues’ 2022 New England Journal review are the places a clinician should be reading about that.

The CF₄ work is prophylaxis, which is a different act in a different place. It happens to a diver before ascent, not to a patient in a chamber, and the authors are explicit that they did not test anything faster: “Shorter decompression protocols were not tested, and shortening decompression time remains a potential application rather than a demonstrated outcome.”

Nothing about a fluorocarbon mixture belongs in the wellness aisle either. There is no trial, animal or human, in which CF₄ improves fatigue, recovery, cognition or ageing; the only indication anywhere in this literature is decompression illness, and the only cleared treatment for it is oxygen under pressure. The FDA’s August 2025 letter to health care providers is still the document that governs anyone running a chamber — it records fires in hyperbaric devices that caused serious injuries and deaths, and asks for cotton clothing, adequate training and monitoring, and none of the prohibited electrical or static-generating items inside the shell. Whatever gas a diver breathes on the way up, the chamber has its own rules.

The tests that would have to happen

A human hyperbaric exposure study comes first: escalating CF₄ fractions at depth, with pulmonary function, work of breathing under exercise and neurobehavioural testing, run by a group with the toxicology capacity to look at the endpoints the rat work skipped. Gas supply and purity come next — carbon tetrafluoride is an industrial chemical, not a diving commodity — and the regulatory route for a new breathing gas runs through frameworks a 510(k) clearance for an oxygen chamber does not touch.

Then there is the question this field has never solved cleanly: how to test a decompression intervention in people who cannot ethically be injured on purpose. The animal designs work because the endpoint is induced. Human evidence would have to come from occupational cohorts, from registries, and from sham-controlled designs that hyperbaric researchers have repeatedly attempted and rarely brought off — as Lansdorp and van Hulst’s review of double-blind trials in hyperbaric medicine describes at length.

For now the honest summary is narrow. In rats, under two protocols, a ten-minute switch to a slow, heavy gas cut decompression sickness by half to nearly two-thirds and left lungs and endothelium better off than controls. In pigs, at a comparable depth, the same switch did something similar. In people, nothing has been published at all — and the last century of breathing-gas research is a long record of animal results that never made it into a diver’s cylinder.

Sources

← Back to blog