On page 449 of this issue1, Pereverzev et al. report mechanical oscillations spontaneously generated by two weakly coupled volumes of superfluid 3He — providing a direct demonstration of its large-scale quantum behaviour.

At very low temperatures all materials tend to fall into highly ordered states. Most commonly, this order takes the form of perfect crystals, but there are exceptions. Most strikingly, the two isotopes of helium (3 He and 4 He) never solidify at all under their vapour pressures. Rather, they stay liquid to the lowest temperatures attainable, and are assumed to remain so right down to absolute zero.

This is baffling at first sight, because liquids are usually disordered, whereas the third law of thermodynamics says that, as the temperature tends to zero, the entropy of any system (that is, its disorder) should decrease towards a constant value that can be taken as zero. But ordering does also occur in the liquid heliums, albeit in momentum space rather than real (cartesian) space: they both undergo a form of Bose-Einstein condensation, in which the 4He atoms (or pairs of 3He atoms) all fall into the same quantum state, described by a macroscopic wavefunction that fills the entire volume of liquid.

By analogy with the microscopic worlds of atoms or electrons, conditions at any point in the liquid are determined by a wavefunction. Liquid motion is specified by the relative phases of the wavefunction at different points: if its phase is the same everywhere, the liquid is stationary; if in a particular direction the phase progressively advances, then the liquid is flowing in that direction. All of this makes good sense in relation to a single bath of liquid helium, or to the electron gas (also described by a macroscopic wavefunction) in a single piece of superconducting metal. But what happens when there are two separate baths, or two separate pieces of superconductor, and they are brought together?

Based on ideas introduced by Josephson2, Anderson3 and Feynman4, one can expect a variety of counter-intuitive and very striking phenomena to occur when two separately phase-coherent systems of this kind are weakly coupled together. One of these, the a.c. Josephson effect, consists of an oscillatory flow of particles to and fro through the weak link coupling the two systems. It occurs when the chemical potential μ (an average energy per particle) differs on each side. In the case of weakly coupled superconductors, the effect can be detected through the electromagnetic radiation that is created by the oscillatory current of negatively charged electrons. It is inherently much harder to demonstrate in liquid helium, partly because the oscillatory current is neutral and therefore creates no electromagnetic radiation, and partly because the weak link between the baths has to be made extraordinarily small.

Pereverzev et al.1 have created their weak link by using modern silicon microfabrication techniques, enabling them to construct 100-nm-diameter apertures in a 50-nm-thick sliver of silicon nitride. The small diameter and length are crucial to success: unless they are both comparable with, or smaller than, the healing length ξ of the superfluid (the minimum length over which the wavefunction can vary significantly), the aperture cannot be expected to function as a weak link. Given that ξ ˜ 0.1 nm for 4He, and ˜50 nm for 3He, it is clear that the stateof-the-art Berkeley technology is just good enough to see Josephson effects in superfluid 3He. Because oscillatory flow in such a tiny hole would be so slight as to be undetectable, they used 4,225 identical holes, hoping that this would amplify the effect.

The silicon chip with its tiny holes was placed in the pill-box-shaped cell sketched in Fig. 1b on page 4491. A ‘soft membrane’ divides the cell into inner and outer chambers, and can be moved by changing the voltage on an electrode, thereby applying an electrostatic force. Its position can be detected by means of a d.c. SQUID (a superconducting quantum interference device, used here as a position transducer of extreme sensitivity). Following a change in the position of the soft membrane, two things happen. First, the tiny change of pressure of the helium in the inner reservoir, as compared with that in the outer one, will change their relative values of μ, leading to oscillatory flow through the apertures. According to the Josephson relation, its frequency should be 183.7 kHz Pa−1 for 3He. Second, superimposed on the oscillation, there will be a steady flow through the apertures that reduces the pressure difference, leading to a continuous decrease towards zero in the frequency of the oscillatory flow.

The experimenters found that the amplified output of the displacement transducer, passed to audio headphones, could be heard as a ‘whistle’ of steadily decreasing pitch — precisely what was expected. Remarkably, it was the discrimination and flexibility of the human ear that led to the initial discovery. Only later were the authors able to make quantitative measurements of the frequency as a function of pressure difference between the reservoirs. Gratifyingly, these fall on a straight line whose gradient agrees with the theoretical prediction based on the Josephson relations.

The way is now open for seeking other mechanical macroscopic quantum phenomena and applications of the Josephson relations in superfluid 3He — the mechanical analogues of electrical effects in superconductors — and, perhaps, for developing a pressure transducer of extraordinary sensitivity.