Helium vs Hydrogen

For lifting things, use helium, and the flammability is not really why. For everything else these two gases are not competitors at all — they are used in completely separate industries, for reasons that have nothing to do with each other.

The comparison is worth making anyway, because it contains one genuinely counterintuitive fact: the lighter gas is not much better at lifting, and the heavier one is the one the world is running short of.

The lift difference is almost nothing

Lift is not about how light the gas is. It is about the difference between the density of the gas and the density of the air it displaces, and air is much heavier than either.

A cubic metre of air at sea level weighs about 1.29 kg. Fill that volume with hydrogen and you carry roughly 0.09 kg of gas, leaving about 1.20 kg of lift. Fill it with helium and you carry about 0.18 kg, leaving 1.11 kg. So although hydrogen is half the density of helium, it delivers only about 8% more lift.

Eight per cent is not nothing on a large airship, but it is nowhere near enough to justify a flammable lifting gas, and that arithmetic settled the question decades ago. What has changed since is that the argument has quietly inverted: helium is now scarce and expensive enough that its cost, rather than hydrogen's fire risk, is the live issue. Meteorological services in many countries fill their radiosonde balloons with hydrogen generated on site, because the balloon is unmanned, disposable, and cheaper to fill with a gas you can make from water.

One can be made; the other has to be found

This is the deepest difference between them and it is a geological one.

Hydrogen is the most abundant element in the universe and there is essentially none of it free in Earth's atmosphere. Every kilogram used industrially is manufactured, overwhelmingly by reforming natural gas with steam. Supply is a question of how much gas and energy you are prepared to spend; it is not a question of whether the hydrogen exists.

Helium ranks second only to hydrogen in cosmic abundance, and it cannot be manufactured on any useful scale at all. Terrestrial helium is radiogenic — alpha particles emitted by decaying uranium and thorium deep in the crust, thermalised into helium atoms and trapped over hundreds of millions of years beneath the same seals that trap natural gas. It is extracted only where a gas field happens to contain enough of it to be worth separating, principally in the United States, Qatar, Algeria and now Russia's Amur plant.

And unlike almost every other resource, helium that escapes is gone permanently. It is light enough to reach escape velocity from the upper atmosphere, so vented helium does not enter a recoverable pool. The United States Federal Helium Reserve, established in 1925 and wound down under legislation passed in 1996, had spent decades acting as the world's buffer stock; its final disposal, together with a run of plant outages, produced the recurring price spikes that have made helium conservation a serious line item in laboratory budgets.

Below twenty kelvin, only one is still useful

Helium's second irreplaceable property is that it stays liquid further down the temperature scale than anything else. It boils at 4.2 K and, alone among substances, will not freeze at all under its own vapour pressure however cold it gets — solidifying it takes about twenty-five atmospheres. Below 2.17 K it becomes a superfluid with zero viscosity, which is a laboratory curiosity and also a practical nuisance, since superfluid helium creeps up and over container walls.

That low boiling point is why an MRI scanner has a helium supply chain behind it. The magnet only superconducts while it sits in a liquid helium bath, nothing else gets that cold, and medical imaging is consequently one of the world's largest helium markets. Magnet designs using a fraction of the traditional charge, sealed and recondensing, exist precisely because the supply has become unreliable.

Liquid hydrogen boils at 20 K, cold enough for many purposes, but it is flammable and it has an awkward habit: the two spin isomers, ortho and para, convert slowly into one another, releasing enough heat in the process to boil off a stored tank unless a catalyst is used to force the conversion during liquefaction.

Helium's inertness is the other half of its usefulness. It forms no compounds under any ordinary condition, which is why it purges rocket propellant tanks and blankets sensitive manufacturing. Its only well-known molecular species, the helium hydride ion, was finally detected in space in 2019 in the planetary nebula NGC 7027, and is thought to have been the first molecule to form in the universe.

Hydrogen is a feedstock long before it is a fuel

Almost all discussion of hydrogen concerns fuel, and almost all hydrogen is not used as fuel. Roughly half of world production goes into making ammonia for fertiliser, and most of the rest into oil refining — removing sulfur from fuels and breaking heavy fractions into lighter ones. Methanol and, increasingly, direct-reduced iron take much of the remainder. Transport is a rounding error.

Where hydrogen does work as a fuel, its numbers are split. Per kilogram it stores about three times the energy of diesel, which is why it is attractive for rockets and, in principle, for aviation. Per litre it is dismal even compressed to 700 bar, and liquefying it costs a large fraction of the energy it contains. Hydrogen's problem was never the energy; it was always the volume and the tank.

The Hindenburg argument, honestly

The 1937 Hindenburg fire is invariably cited as the reason airships use helium, and the causes are genuinely disputed. The former NASA engineer Addison Bain argued from the 1990s that the airship's doped outer fabric was highly flammable and that the fire spread through the skin rather than the gas cells. Other investigators have tested the fabric and concluded it could not have burned fast enough to account for the film footage, and that hydrogen was the principal fuel. The reasonable position is that the covering contributed and the hydrogen dominated. Either way, the airship had been designed for helium and was flying on hydrogen because the United States would not export helium — which makes the disaster as much a supply story as a chemistry one.

Lifting, cooling, or feeding a reactor

• Lifting anything with people aboard — helium, with the 8% penalty accepted.
• A disposable unmanned balloon — hydrogen, generated where it is needed.
• Cooling a superconducting magnet — helium, with no alternative below 20 K.
• Leak testing a sealed system — helium, small and inert.
• Making fertiliser or cleaning up fuel — hydrogen, and that is where most of it goes.
• Storing energy by mass — hydrogen.
• Storing energy by volume — almost anything else.