Proponent of supersolid helium joins sceptics

Physicist Moses Chan, who in 2004 published evidence for supersolidity, explains how he came to doubt his own discovery.

Moses Chan

Moses Chan reported¹ in 2004 that ultracold solid helium-4 might become a quantum superfluid, and flow through itself. A solid that could flow had been predicted decades before, and the suggestion that the strange substance might have been created thrilled condensed-matter physicists. But some were sceptical, including Chan’s former PhD supervisor, John Reppy of Cornell University in Ithaca, New York, who was profiled in Nature in 2010 as “The supersolid's nemesis”. Now Chan, who works at Pennsylvania State University in University Park, has joined the sceptics, as he declares in a paper published in Physical Review Letters². Nature asked him how he came to change his mind.

What first caused you to doubt that supersolidity was the right explanation for what you and Eunseong Kim published in 2004?

In 2004, we reported that some of the moment of inertia of solid helium-4 in a rotating porous Vycor glass cell — an apparatus called a torsional oscillator — went missing when it was cooled down below 0.2 degrees above absolute zero. We suggested that was evidence of a supersolid forming. But in 2007, physicists James Day and John Beamish at the University of Alberta in Edmonton, Canada, showed³ that the shear modulus — the stiffness — of helium-4 increases as you decrease the temperature. Since then, I think, everyone in the field has been grappling with the possibility that when the temperature drops in a torsional oscillator, the solid helium becomes stiffer, and that mimics the supersolid signal.

How did you go about testing that?

In solid helium-4 you have dislocation lines — places where the crystal structure doesn’t quite line up — that become stiffer at lower temperatures. In the initial experiment, the solid helium formed in the pores of the Vycor glass, which are a few nanometres (billionths of a metre) across, and so could not accommodate dislocation lines that are tens of micrometres long (one micrometre is one millionth of a metre). But we realized that there was a gap between the Vycor glass and the top metal plate, where a layer of bulk solid was forming that could contain dislocation lines. So, with Duk Young Kim at Penn State, I designed a new cell, in which the Vycor glass is painted and sealed with epoxy resin, so that there is no space for any bulk solid. And we absolutely don’t see any signal with this cell. So our conclusion is that the signal we saw in 2004 was not due to the helium in the Vycor glass, but to the very thin bulk-solid layer.

Were you disappointed?

Sure, certainly it is disappointing after almost 10 years. We thought we had done all the right control experiments and that there was supersolidity. But, well, it didn’t turn out that way. It’s like being a detective for 10 years and as it turns out the culprit is not the most glamorous. On the other hand, it’s a sort of a relief because we have been at this for so long, and been so puzzled.

What are the human relationships like in this small field, where you are all critiquing each others’ experiments?

John Reppy, who this year also published⁴ an experiment with a Vycor glass cell suggesting that the signal was not due to supersolidity, is my former supervisor and my mentor. And Duk Young Kim is the former student of Eunseong Kim, with whom I published the 2004 paper. So there are four generations of us involved. And we’re all good friends.

Is everyone convinced now that there is no supersolid?

That is not something that we have measured directly in bulk helium; it is something that we infer. In 2010, Eunseong Kim — who is now at the Korea Advanced Institute of Science and Technology in Daejon — reported⁵ on an experiment in which bulk helium in a torsional oscillator acts as if it contains rotating quantum vortices, which is also a signal of supersolidity. That is very exciting. But between 2007 and now, many experiments in bulk solid helium have shown different, contradictory things. Knowing what we know now, it is reasonable for Eunseong and others to do the vortex experiment again at two different frequencies, because if the signal doesn’t depend on the frequency then it would be a supersolid signal and we would be excited.

Even though you would be wrong again?

That would be really nice. I would rather be wrong, and hear that there is supersolidity in bulk solid helium-4 after all. But at this point, there’s no other conclusion I can draw from my recent experiment than that there isn’t.