'Baked Alaska' cooked up in liquid helium.

Baked Alaska seems an unlikely term to encounter in physics, but this culinary surprise, consisting of meringue baked around ice cream, serves as an apt description of an exotic, theoretical model accounting for a curious aspect of liquid-helium behavior. The model, proposed in 1984 by Anthony J. Leggett of the University of Illinois at Urbana-Champaign, suggests that high-energy particles produced by cosmic rays can trigger the otherwise inexplicable formation of one form of superfluid helium-3 at the expense of another.

In the baked Alaska scenario, high-energy electrons, created by the passage of cosmic-ray-generated muons through the supercooled liquid, deposit significant amounts of energy in spots less than a micron in diameter. Each intensely heated microball of liquid helium expands into a hot shell, leaving behind a pocket of cold superfluid helium. Isolated from the rest of the liquid, this cold core provides a protected environment in which a bubble of a different type of superfluid helium-3 can nucleate and start to grow.

Now researchers have obtained experimental evidence establishing the plausibility of Leggett's scenario. "Our results are certainly consistent with the [baked Alaska] model, though there are still some unanswered questions," says Peter E. Schiffer of Stanford University.

"The Stanford results show that at least the idea of nucleation by high-energy particles isn't totally crazy," Leggett notes.

Schiffer, Douglas D. Osheroff, and co-workers report their findings in the July 6 PHYSICAL REVIEW LETTERS.

Helium-3, a rare isotope of helium, becomes a superfluid--a liquid that flows without friction--at temperatures below 2.5 millikelvins. In this chilly state, helium atoms tend to form pairs. Because these pairs can arrange themselves in two different ways, helium-3 has two distinct superfluid states. Depending on the pressure and the magnetic field applied to a sample, the so-called A phase is more stable than the B phase at higher temperatures, whereas the B phase takes over at lower temperatures.

In 1977, Osheroff (then at AT&T Bell Laboratories) and co-worker Michael Cross showed that the superfluids had characteristics implying that the A phase, even when supercooled well below the temperature at which a trasition from the A to the B phase should occur, cannot by itself spontaneously make the change. Because such phase transitions actually do occur, this puzzling feature led to a search for a mechanism that would explain how the transition happens.

"I played around with various ideas, and it eventually sank into my mind that no mechanism based on a thermal equilibrium distribution of energy was going to explain this," Leggett recalls. His baked Alaska model emerged out of this line of reasoning.

"I had to convince myself you couldn't apply the normal laws of hydrodynamics of thermal transport under these conditions because you're so far from equilibrium," he says. "It really matters how the heat spreads out."

To check whether radiation can indeed trigger the nucleation of the B phase within the A phase of superfluid helium-3, the Stanford group used a specially designed, long, thin, silica glass tube with microscopically smooth surfaces. Within this tube, the team discovered it could dramatically supercool samples of the A phase to temperatures as low as 0.37 millikelvins, much lower than temperatures achieved by other groups.

In addition, by placing sources of radiation near the sample cell, they discovered that they could greatly reduce the length of time before nucleation occurs in the supercooled A phase. Both gamma rays and neutrons produced comparable effects.

"It's clear that radiation does play a part," Osheroff says.

These findings indirectly suggest that the presence of surface irregularities or defects also has a strong influence on the nucleation of phase B. This factor may have thwarted previous attempts to detect radiation-induced nucleation.

Moreover, the Stanford experiment demonstrates the conditions necessary for observing the A phase at lower temperatures and lower magnetic fields than previously possible. "Now that we've got it pinned down, I think there's going to be a burst of activity," Leggett says. "A lot of people would love to have [A-phase] helium-3 in low magnetic fields at low temperatures. There are all sorts of things you can do with it."

Precisely how surface roughness and the presence of minute traces of such impurities as radioactive tritium contribute to the nucleation of phase B remains unclear. Osheroff and his team are now discussing the design of sample containers specially fabricated to have a certain roughness. The researchers would also like to observe nucleation at different pressures and magnetic fields.

"Helium-3 is an ideal system for understanding physics that would be completely masked in any other system," Osheroff says.

To Leggett, the A-B transition in superfluid helium-3 represents a particularly clear example of how locally concentrated energy that can't dissipate through normal channels can induce events that by any other, reasonable, statistical measure would seem astronomically improbable. --I. Peterson.