Ryan Duncan froze. He had just performed a new experiment examining common graphite—the stuff of pencil lead—but the results seemed physically impossible: Heat, which typically disperses slowly, had traveled through the graphite at the speed of sound. That is like placing a pot of water on a hot stove and instead of counting down the long minutes until that water starts to simmer, watching it boil almost instantaneously.
It is no wonder that Duncan, a graduate student at the Massachusetts Institute of Technology, could not quite believe his eyes. To ensure that he had not made a mistake, he quadruple-checked everything within his set-up, ran the experiment again, and took a mental-health break. “I tried to get some sleep, knowing that I wouldn’t be able to tell if the experiment was successful or not for several more hours, but I was finding it pretty difficult to shut down for the night,” he recalls. When Duncan’s alarm went off the next morning, he ran to his computer (still in his pajamas) and crunched the new measurements only to confront the same result: Heat had still moved impossibly fast.
Duncan and his colleagues published their results last week in the journal Science. The phenomenon, known as “second sound,” has physicists in a state of euphoria—in part because it could pave the way for advanced microelectronics, but mostly because it is so deeply weird.
To understand why, just think about how heat is conducted through the air. It is carried via molecules, which constantly collide with each other and scatter the heat in all directions—forwards, sideways and even backwards. That fundamental inefficiency makes conductive heat relatively sluggish (radiant heat, by comparison, can travel at light speed as infrared radiation). The same sluggishness holds for heat moving through a solid. Here, phonons (packets of acoustic vibrational energy) carry the heat much like molecules in the air, allowing it to scatter in all directions and slowly disperse. “It's a little bit like, if you take a drop of food coloring and put it into water, it spreads,” says Keith Nelson, Duncan’s advisor at MIT. “It doesn’t just move straight as an arrow away from where you put the drop.” But that is precisely what Duncan’s experiment suggested. In second sound, the backscattering from phonons is heavily suppressed, allowing heat to shoot forward. “That’s the way wavelike motion behaves,” Nelson says. “If you’re in a pool and you launch a water wave, it will leave where you are.… But it’s just not normal for heat to behave that way.”
And for the most part, it does not. Second sound was first detected in liquid helium 75 years ago and later seen within three solids. “All indications early on were that this was something that would really be confined to very few materials and only at very low temperatures,” Nelson says. As such, scientists thought they had hit the end of the road. “It wasn’t super clear what [second sound] could be apart from a scientific statement,” says Nicola Marzari, a materials scientist at the Swiss Federal Institute of Technology, in Lausanne, who was not involved in this study. “So, the entire field went dormant for many years.”
But dramatic improvements in numerical simulations helped to revive the field roughly five years ago—allowing scientists to recognize that the phenomenon might be more widespread. Gang Chen, an engineer at MIT, for example, was able to predict that second sound might be visible within graphite at rather balmy temperatures. That prediction electrified Duncan, who tested it just as soon as he could—eventually putting the rest of his pursuits on the back burner, once the results proved to be so counterintuitive.
First, Duncan deposited heat into the graphite sample using two crossed laser beams to create an interference pattern—alternating bright and dark regions that correspond to crests and troughs in the colliding waves of light. At the outset, the crests heated up the graphite while the troughs remained cool. But once Duncan switched off the lasers, the pattern would begin to slowly diminish as heat flowed from the hot crests to the cool troughs. The experiment would reach its end once the entire sample reached a uniform temperature. Or at least that is what typically happens. But when the lasers stopped shining, the graphite had other plans, continuing to allow the heat to flow until the hot crests became cooler than the troughs. This is rather like a stove top that becomes ice-cold the instant you turn it off rather than gradually cooling to ambient temperature. “That’s weird,” Nelson says. “Heat isn’t supposed to do that!”
And it certainly is not supposed to do that at such high temperatures. Marzari, who predicted the phenomenon at almost the same time as Chen, was therefore fairly confident that it would prove valid. Even so, he was less certain that second sound would be seen at the foreseen high temperatures. “If you had asked me to bet my mortgage on the existence of this effect, I would have said yes,” Marzari says. “But the question is always does it happen at 100 Kelvin, 20 Kelvin or 0.1 Kelvin?” Duncan’s experiment found the effect at 120 Kelvin—more than 10 times higher than previous measurements. “Nobody ever thought that you would actually be able to do this at such high temperatures,” says Venkatesh Narayanamurti, a research professor of technology and public policy at Harvard University who was not involved in the study. “In that sense, it breaks some conventional wisdom.”
It also suggests that the finding might find a practical use in the future. Not only is the temperature far more practical than the cryogenic chill required to work with the previous findings, but graphite is a commonplace material—two characteristics that might help engineers overcome the daunting issue of heat management in microelectronics today. Just imagine if heat rushed away at the speed of sound, allowing materials and devices to cool much more rapidly. Such a feat surely would allow engineers to build smaller, more efficient microelectronics. With this in mind, Narayanamurti (who worked on second sound when he was at AT&T Bell Laboratories from 1968 to 1987) suspects that the field soon flourish once again. “If I were still at Bell Labs, I would have people doing experiments on it because it will be important 10, 15 years down the road.”