The Helium-3 That Runs Out With the Bombs

As the shortage hit, a researcher trying to buy helium-3 for a low-temperature physics lab was quoted $600 a liter — five times what he'd paid in years past. Others were quoted up to $2,000, roughly a tenfold jump from a historical baseline near $100 [S1]. The Department of Energy's own published price, $350–400 per liter, was already a multiple of what the gas used to cost, and even at that price the DOE could barely fill the orders coming in [S1].

The weird part isn't the price. It's the supply chain. Nearly every liter of civilian helium-3 in the United States comes out of the Savannah River Site in South Carolina, where the National Nuclear Security Administration separates it from aging tritium — the hydrogen isotope used to boost the yield of thermonuclear warheads [S3]. Tritium decays into helium-3 with a half-life of 12.32 years [S3]. The bombs, sitting in their bunkers, are slowly turning themselves into the gas.

That is the entire civilian pipeline. No mine, no well, no chemical plant. The country's stockpile of a strategically critical isotope is, structurally, the off-gas of the nuclear weapons enterprise [S3].

Helium-3 is a stable isotope with two protons and one neutron, and it is vanishingly rare on Earth. What the United States has, it has because it built hydrogen bombs and then waited.

Which was fine, until people noticed what helium-3 could do.

It does, roughly, four things. In a dilution refrigerator — the only practical way to get matter down to a few thousandths of a degree above absolute zero — a helium-3/helium-4 mixture is the working fluid. That's the platform most quantum computers and millikelvin physics experiments run on [S1]. In a hyperpolarized lung MRI, patients inhale helium-3 and a scanner watches it move through their airways in real time; by 2010 the National Institutes of Health was supporting roughly 25 active projects requiring helium-3, with cumulative US investment in helium-3 MRI estimated at $60–100 million over the prior decade [S7]. It's used in neutron scattering instruments at major physics facilities [S1]. And — the one that wrecked everything — it is the active medium in the neutron detectors used to spot smuggled fissile material at ports and borders [S2].

After 9/11, DHS started buying those detectors at scale. The agency has deployed over 1,400 radiation portal monitors domestically, the kind of installation that sits at a truck inspection lane and pings if a shielded chunk of plutonium rolls past [S2]. Each one is hungry. By 2008, US demand for helium-3 had climbed to as high as 70,000 liters a year against a federal supply that couldn't keep up [S5]. In that single year, the country burned through more than half of the existing stockpile [research].

The scale of the mismatch is hard to overstate. Annual US production from tritium-decay helium-3 is now about 8,000 liters a year [S3]. Demand at the peak was nearly an order of magnitude above that. The gap had to close somehow, and it closed by rationing.

In 2009 the National Security Council stood up an Interagency Policy Committee to decide who got helium-3 and at what price [S3]. The committee set a two-tier price — one for medical users, one for everyone else — and began allocating the country's annual production by hand [S3]. Starting in fiscal 2010 it cut off new allocations to domestic radiation portal monitors entirely, on the grounds that alternative neutron detectors were good enough [S2]. The agency tasked with preventing nuclear terrorism had been told it could no longer have the gas its detectors needed, because it had already taken too much.

The committee allocating helium-3 to lung-MRI researchers reports up the same chain that oversees nuclear strategy [S3]. It is hard to think of another commodity where the rationing board and the weapons-policy board are the same room.

There is a second supplier the standard story tends to skip. From 2004 to 2008, the United States had been importing roughly 25,000 liters of helium-3 a year from Russia, also harvested from Russian weapons tritium [S5]. Then, in August 2008, Russia halted exports to keep the gas for itself — the same year US demand spiked [S5]. The Russian cutoff wasn't a downstream complication of the crisis; it was one of the things that precipitated it. The "US monopoly" framing erases the fact that two nuclear powers were quietly supplying the global market with the slow exhaust of their warheads, and then one of them stopped.

The Government Accountability Office, looking at all this in 2011, declined to blame DHS alone. Its report faulted DOE for not tracking demand or inventory, and for not warning users of the looming shortage until July 2009 — more than a year after it had recognized the imbalance [S2]. The proximate cause of the federal panic, in other words, was that the agency holding the stockpile didn't know how fast it was emptying.

The gas hadn't all been used, either. Reuter-Stokes, the largest industrial maker of helium-3 neutron detectors, has argued that more than 120,000 liters of helium-3 is sitting inside already-deployed radiation portal monitors, physically intact and theoretically recoverable [S4]. By that accounting, the United States didn't burn its stockpile so much as mislay it — sealed inside the same DHS hardware that drove the run on the gas in the first place [S4]. Reuter-Stokes has a commercial interest in the recycling pitch, but the gas is real and the detectors haven't gone anywhere.

Which leaves the moon.

Whenever helium-3 comes up in public, someone proposes mining it from lunar regolith — the dust on the moon's surface, which has been bombarded by the solar wind for billions of years and contains traces of the gas Earth's magnetic field shields us from. The numbers are not kind. The highest helium-3 concentration measured in returned lunar samples is about 10 parts per billion, with an average around 4 ppb in the regolith [S6]. Roughly 150 tons of moon dust must be processed to recover a single gram [S6]. To fuel a fusion reactor on this stuff, you would need to strip-mine and bake an area the size of a small country.

And there are no working deuterium–helium-3 fusion reactors to feed [S6]. The planetary scientist Ian Crawford is among the academics who have pointed out that the economics don't survive contact with the measured concentrations, and that the technology that would consume lunar helium-3 doesn't exist [S6]. The most cited supply solution for the helium-3 shortage is a multibillion-dollar mining program to feed a reactor nobody has built.

Meanwhile, the federal demand curve has been bent down to meet the supply curve, rather than the other way around. Projected federal demand is now under 6,000 liters a year, against ~8,000 liters of annual production [S3]. The shortage has been "solved" the way droughts are solved in a city that bans lawn-watering: the gas that runs the quantum computers, the lung MRIs, the neutron experiments, and the port scanners is parceled out a few liters at a time by a committee that reports to the National Security Council [S3].

The most sustainable known source of helium-3 on Earth remains the slow radioactive decay of the warheads we are trying not to build. Every gram that comes out of Savannah River is, in a literal sense, a tritium atom that didn't get used in a bomb. Disarmament shrinks the stockpile; weapons production refills it. The portal monitors at the docks, scanning for the next bomb, are running on the residue of the last one.