Helium He · 2

The only element cold enough to keep an MRI magnet superconducting, and the only one light enough to escape Earth's gravity for good once released.

Pourquoi est-ce important ?

Helium is extracted as a by-product of natural gas. When it is vented it is gone from the planet permanently.

Where it is in the Earth

Helium is not a mineral in any conventional sense. It does not crystallise, it does not bond with other elements, and it leaves no trace in rock chemistry the way copper or iron does. Instead, it accumulates as a gas, and it does so because of radioactive decay deep in the Earth's crust. Uranium and thorium, both present in granites and certain other ancient rocks, shed alpha particles as they decay over geological time. An alpha particle is simply a helium nucleus — two protons, two neutrons — and once it captures a pair of electrons from its surroundings it becomes a helium atom. That atom is chemically inert and extraordinarily small, so it migrates slowly upward through rock until something stops it.

What stops it is the same thing that traps natural gas: an impermeable cap rock — typically a dense shale, evaporite, or tight carbonate — arching over a porous reservoir. Helium collects in the pore spaces of that reservoir alongside methane and other gases. The key geological requirement is that the cap rock must be tight enough to hold a molecule even smaller than hydrogen, and the geological setting must be old enough for meaningful volumes to have accumulated. This is why the most productive helium provinces sit on ancient stable platforms — the mid-continent United States, the Siberian craton, the great gas fields beneath the Qatar peninsula — where thick sequences of source rock have had hundreds of millions of years to generate helium and where structural geology has provided the right traps. Younger, tectonically disturbed terrains rarely hold it in economic concentrations because the cap rocks have been fractured and the gas has long since seeped away.

A useful recent addition to the picture is the discovery that the East African Rift system can deliver large helium accumulations through a different mechanism: volcanic heat accelerates degassing from ancient crustal rocks, and the helium migrates into shallow sedimentary traps. Tanzania's discovery showed that not all prospective provinces are already known, and that geologists are still refining their understanding of where the gas can concentrate.

Getting it out

Helium is never mined in isolation. Because it occurs as a component of natural gas — usually a small fraction of the total gas stream — it is recovered at whatever point natural gas is being produced and processed. There is no helium mine in the way there is a copper mine. The primary operation is a gas well, drilled and completed using conventional petroleum techniques, and the helium is extracted as a by-product of that operation. The natural gas field is developed for its methane content; the helium comes along because it is there.

The concentration of helium in the raw gas stream — its grade, in the language of mining — varies considerably between fields. Ordinary natural gas contains only trace amounts, far too little to be worth separating. The fields that produce commercial helium contain materially higher concentrations, though the exact percentages differ by location and are not uniform even within a single reservoir. What matters economically is that the concentration must be high enough that the cost of separating helium from the other gases is recovered in the value of the helium sold. Fields that sit just below that threshold may become viable if the gas is being processed anyway for other reasons, or may never be developed specifically for helium at all.

Because helium production is tied to natural gas production, the rate at which helium is extracted is not set by helium demand alone. When natural gas output at a host field falls — because the reservoir pressure drops, because methane prices make production uneconomic, or because a facility is shut for maintenance — helium output falls with it. This dependency is structural and cannot be engineered away without developing dedicated helium wells, which remain uncommon.

What pulls on it

The single largest use of helium is cooling the superconducting electromagnets inside magnetic resonance imaging (MRI) scanners. A superconducting magnet carries electrical current with no resistance, but only when it is kept below a critical temperature that, for the alloys used in MRI, requires immersion in liquid helium. The global installed base of MRI machines is large and growing, particularly as middle-income countries expand their health infrastructure, so this demand is persistent and relatively insensitive to price. Hospitals cannot simply switch off their magnets or substitute another coolant without replacing the machine entirely.

The semiconductor industry is the second major consumer, in two distinct ways. Helium is used in the process chambers where silicon wafers are etched and deposited upon, because its inertness and thermal conductivity help control temperature precisely without contaminating the substrate. It is also used to detect leaks in equipment and pipelines, a role that exploits its small atomic radius — helium passes through the finest imperfection that a larger molecule would not. As semiconductor manufacturing has moved toward ever smaller feature sizes and more demanding process conditions, the gas-phase requirements of fabrication have grown. The processing plants listed in the accompanying table — TSMC's Fab 18 in Taiwan and Samsung's Pyeongtaek campus — are representative of the high-volume chip fabs that depend on a reliable helium supply.

Uses in data centres and in emerging fields such as hard-disk drive manufacturing and certain aspects of AI infrastructure add further demand, though these remain smaller in aggregate than medicine and semiconductors. What would have to change for demand to fall sharply is either the commercial availability of MRI magnets that do not require liquid helium coolant — a technology that has been demonstrated at small scale but not yet deployed widely — or a significant shift away from superconducting magnets in medical imaging. Neither appears imminent. In the other direction, demand could rise sharply if novel superconducting technologies — for power transmission, quantum computing, or particle physics — move from laboratory to commercial scale.

Turning ore into product Niveau 3

Raw gas arriving at the surface contains methane, nitrogen, carbon dioxide, heavier hydrocarbons, water vapour, and — in helium-bearing fields — a small fraction of helium. The first processing steps remove water and acid gases such as carbon dioxide and hydrogen sulfide using standard gas-sweetening methods, typically amine absorption. What remains is a mix of hydrocarbons and the lighter non-condensable gases. Helium, because of its extremely low boiling point — lower than any other element — can be separated by a process of cryogenic fractionation: the gas stream is chilled in stages until the methane and nitrogen liquefy and can be drawn off, leaving a crude helium stream that is substantially enriched relative to the feed. This crude helium, sometimes called raw or Grade B helium, still contains significant nitrogen and small quantities of other impurities.

Upgrading crude helium to refined Grade A gas, or further to liquid helium, requires additional purification. Pressure swing adsorption — a process in which gas is cycled through beds of adsorbent material that preferentially retain nitrogen — is the most common route. The output is a gas that is very nearly pure helium. Liquefaction to the traded form used in MRI cooling requires chilling to approximately 4.2 Kelvin, which is just above absolute zero and colder than outer space. This demands large and energy-intensive liquefier plants. The cold chain between producer and end user must be maintained without interruption; liquid helium warms and boils away if the insulated transport vessel — a Dewar — loses its vacuum or is left without topping up for too long. Losses occur at every transfer, and a measurable fraction of all helium placed into commerce is simply lost to the atmosphere before it reaches its final application, at which point it is gone permanently from the recoverable resource base.

The location of processing plants is not always close to the producing field. Crude helium is sometimes transported by pipeline or in high-pressure tube trailers to centralised liquefaction or purification facilities. Qatar's production, for example, is liquefied at plants on the peninsula and then shipped in ISO-standard cryogenic containers to markets in Asia, Europe, and the Americas. The processing capacity available at any one hub therefore acts as a constraint independent of the underlying gas resource, and outages at a single large plant — as occurred with the Skikda facility in Algeria and with production disruptions in Qatar — can tighten global supply even when the gas itself is still flowing.

Substitution and recycling Niveau 3

For its primary application in MRI cooling, helium has no working substitute today at scale. The physical property being used is the temperature at which it boils under pressure: 4.2 Kelvin at one atmosphere, colder than any other substance in its liquid form. Liquid hydrogen boils at a higher temperature and introduces serious safety and handling complexity. Mechanical coolers — cryocoolers — can in principle recondense helium that would otherwise boil away, and so-called closed-cycle or zero-boil-off MRI systems have been developed that dramatically reduce helium consumption per scanner over its lifetime. These systems still require a helium charge at installation, but the ongoing consumption is much reduced. Their adoption is growing among new MRI purchases, which will gradually reduce the rate of demand growth, though not reverse aggregate demand while the installed base of older systems persists.

In semiconductor manufacturing, some uses of helium can be partially met by argon or nitrogen in applications where thermal conductivity is not the limiting factor, but helium's thermal properties are genuinely superior for wafer cooling, and substitution carries a performance cost that manufacturers are reluctant to accept given the value of the product being made. Leak detection with helium can in principle be done with hydrogen — which has similar permeation properties — but the additional safety requirements around hydrogen handling mean that the cost advantage is not straightforward, and adoption has been limited.

Recycling of helium is technically feasible and is practised in some research and industrial settings. Large physics installations — particle accelerators, for instance — often invest in recovery and reliquefaction systems that capture the gas that boils off and reuse it. In medical settings, recovery systems can be retrofitted to MRI installations, and some hospitals in regions with constrained supply have done so. However, the majority of helium used commercially is not recovered. The collection infrastructure does not exist at the point of use, the economics of small-scale recovery are unfavourable unless the gas price is high, and once helium escapes to the atmosphere it rises and, over geological timescales, is lost to space. The combination of no practical atmospheric recovery, limited recycling infrastructure, and inelastic end uses makes the loss rate a genuine long-term concern rather than a solvable short-term logistics problem.