One of the curious aspects of a new prediction that hydrogen can be a high-temperature superconductor under pressure1 is that the same conclusion was voiced in the same journal forty years ago. In 1968, Neil Ashcroft at Cornell University argued2 that the conventional description of superconductivity in terms of the pairing of electrons mediated by phonon lattice vibrations — the so-called BCS picture of John Bardeen, Leon Cooper and Robert Schrieffer — made solid hydrogen a good candidate for a superconductor with an unprecedentedly high transition temperature (Tc).

Ashcroft pointed out that the lightness of hydrogen atoms gives the solid a high phonon frequency, and that the lack of core electrons in hydrogen promotes strong coupling between electrons and phonons; both factors favour a high Tc. He held back from making explicit predictions of Tc (which would be pressure-dependent) although one can deduce that his lower limit was about 50 K. Ashcroft proposed that Jupiter might have a hydrogen-rich interior warm and dense enough for part of it to be superconducting.

This story can be traced back still further to the original notion that hydrogen might be metallic. That idea is normally thought to originate with Eugene Wigner and his colleague Hillard Huntington3, although they attributed the notion that many substances should become metallic under pressure to the crystallographer J. Desmond Bernal, a protégé of William Bragg. This prediction wasn't verified until 1996, using shock compression of hydrogen4.

Ashcroft's hypothesis has been elaborated several times since. In 1989, for example, Marvin Cohen and his co-workers at the University of California at Berkeley carried out quantum-mechanical calculations of the strength of electron–phonon coupling in a high-pressure phase of hydrogen, from which they deduced that the superconducting Tc at around 400 GPa should be about 230 K (ref. 5). At that stage, such high pressures weren't accessible experimentally.

Ashcroft himself returned to the problem in 1997 (in fact, it seems never to have been too far from his mind), when he and C. F. Richardson showed that the superconducting behaviour depends crucially on whether the high-pressure phase is monatomic or retains the diatomic character of molecular hydrogen6. In the latter case, there are strong correlations between the electrons and holes in the electron bands that arise from overlapping orbitals, which enhance the electron pairing on which superconductivity depends. That boosts Tc even more, perhaps even to values of 400 K or so.

So what's new now? Pierluigi Cudazzo, at the Università deglie Studi dell'Aquila in L'Aquila, Italy, and co-workers, have looked again at the electron–phonon coupling in metallic hydrogen using ab initio quantum methods to figure out which are the main factors driving superconductivity1. They confirm that the electron–phonon interaction is strong, and say that the complex Fermi surface — the 'shape' of electron bands in momentum space — offers plentiful opportunities for electrons to couple. Most strikingly, their predicted Tc of 242 K at 450 GPa matches closely the cruder, earlier predictions, affording some confidence that, if only the experiments can be arranged, we'll find that superconductivity near room temperature is truly possible.