A refrigerator that cools without helium-3

The machines that cool quantum computers have a geopolitical problem hiding inside them. The workhorse technology for reaching temperatures below one kelvin is the dilution refrigerator, and dilution refrigerators need helium-3, a rare isotope that is produced in meaningful quantities almost nowhere except as a by-product of nuclear weapons programmes. Supplies are tight, prices are volatile, and every lab in the world that wants to run a superconducting qubit or a single-photon detector is bidding on the same scarce gas. A paper in the 11 February issue of Nature offers a way out: a metallic alloy that cools to 106 millikelvin using the magnetocaloric effect, with no helium-3 anywhere in the loop.

The principle is as old as the nineteenth century and as modern as a quantum processor. Magnetocaloric materials heat up when magnetised and cool down when the field is removed, because the magnetic degrees of freedom do work on the lattice. Repeated magnetise-and-demagnetise cycles can therefore pump heat, in exactly the same way a compression-expansion cycle does in an ordinary refrigerator. The sub-kelvin version of this idea, adiabatic demagnetisation refrigeration, has existed for a century, and its problem has always been the same one: the best magnetic coolants are poor heat conductors, insulating salts that behave like a wooden block, cold in the middle but unable to move heat out quickly enough to be useful. The material that cools well does not conduct, and the material that conducts does not cool well.

The new alloy resolves that trade-off by being a metal. The authors, from the Hefei Institutes of Physical Science of the Chinese Academy of Sciences, the Institute of Theoretical Physics, and Shanghai Jiao Tong University, designed a three-dimensional alloy whose magnetic order arranges itself into what they call a metallic spin supersolid, a state in which the magnetic moments sustain the kind of correlated order normally associated with a supersolid while the material retains genuinely metallic heat transport. The result is a substance that behaves as both a heat-absorbing sponge and a heat pipe: it reaches 106 millikelvin, the lowest temperature any metal magnetocaloric material has reached without helium-3, and it conducts heat fifty to a hundred times better than conventional magnetocaloric salts.

The implications go beyond a new record. If solid-state refrigeration can be made to work reliably at these temperatures, the dependence of quantum research on a scarce isotope begins to loosen, and with it the fragile supply chain that currently connects every low-temperature laboratory to a handful of ageing nuclear facilities. A cooling technology based on ordinary, abundant elements can be built where it is needed, scaled in whatever form factor the experiment demands, and controlled electronically rather than by managing a gas inventory. The authors are careful not to overstate the distance travelled: reaching 106 millikelvin is one thing, building a refrigerator that maintains it with a usable cooling power is another, and engineering the magnetising field cycles at scale is a substantial task.

What makes the result more than a curiosity is that it is a paradigm shift in name and in practice. Spin supersolids have been studied for decades as a purely theoretical curiosity, a state of matter too fragile to imagine using. This work takes that exotic state and puts it to work, and it does so in the exact place where quantum technology’s cost curve is steepest. The helium-3 problem is not solved, but it now has a credible rival, and rival technologies have a way of making scarce resources less scarce than they once appeared.