Portland, Ore. -The National Aeronautics and Space Administration (NASA) recently confirmed a 50-year-old hypothesis regarding the nucleation barrier-the process that causes ice to melt faster than it freezes.
The nucleation barrier was first cited by German physicist Gabriel Fahrenheit in the 18th century, but Fahrenheit was at a loss to explain the mechanisms that might be responsible. It wasn't until 1952 that physicist Charles Frank came up with a theory regarding crystal growth that explained the phenomenon. Frank hypothesized that an evanescent structure that formed an energy barrier caused the effect, but he didn't have the tools to prove it.
It took NASA and its advanced Electrostatic Levitator tool to confirm Frank's hypothesis. Using the ESL, researchers can suspend a sample inside a vacuum chamber and use a laser to study it. The agency used the tool to uncover the mechanism underlying the nucleation barrier, dubbed quasicrystals.
"While our experiments were with the change from liquid to solid phase, the nucleation barrier exists when forming thin films on chips too-it's just that it has to involve, say, 40 atoms for liquids and only four or five for vapor deposition," said Kenneth Kelton, a physics professor from Washington University in St. Louis who leads the NASA research team.
NASA's Electrostatic Levitator comes from the agency's Marshall Space Flight Center (Huntsville, Ala.).
NASA's Marshall Center researchers Jan Rogers, Tom Rathz and Mike Robinson were members of Kelton's research team, as were researcher Robert Hyers at the University of Massachusetts (Amherst) and Doug Robinson, a researcher at the U.S. Department of Energy's Ames Laboratory (Ames, Iowa). Kelton was assisted at Washington University by Geun Wu Lee, a graduate student, and Anup Gangopadhyay, a research scientist.Sticky situation
During vapor deposition onto a semiconductor substrate, adatoms (free atoms released in a vacuum) at first just dance around on the surface, despite the fact that they should crystallize into a solid at that temperature. In a process that until now was thought to be random, the adatoms eventually start sticking to one another when they collide, forming islands that grow, merge and coarsen.
Theoretically, the adatoms should immediately begin self-assembling into a thin film as soon as they hit the "freezing" temperature, but in practice semiconductor manufacturers can do little more than tweak parameters to hasten the onset of self-assembly. For liquids changing to solids, the process is three-dimensional, which complicates it further. Thus, instead of needing five adatoms to come together at once on a plane to form an island (as in chip making), in liquids the five adatoms just form one face of an icosahedron with 40 or more atoms, which loosely bond into a quasicrystal.
"People used to think that there was no structure in liquids, but we have shown that there is structure-based on the icosahedron-and that the structure is responsible for the nucleation barrier. We have to break down the quasicrystalline structure before it can go from liquid to solid; that's the barrier to be overcome," said Kelton.
To make their observations, the researchers suspended a drop of titanium zirconium nickel in midair and measured its structure with X-rays.
In the 1700s, Fahrenheit, was working on his now-famous temperature scale when he came across the nucleation barrier. He found that water began to melt immediately once the melting temperature was reached. Something unexpected happened when Fahrenheit cooled water to its freezing temperature, however. He had to wait an inordinate amount of time for crystals to condense out of the liquid.
The puzzle was reexamined in the 1950s, when General Electric Co. researchers David Turnbull and Robert Cech repeated the experiment with metals and discovered the same barrier existed during the phase when solids were turning into liquids.
The rediscovered barrier then piqued the interest of physicist Frank of the University of Bristol in England. In 1952, Frank speculated that atoms of liquid metal form icosahedrons with 20 faces and that those icosahedrons must be broken before a solid can crystallize, thus offering an explanation for the formation of the observed barrier.
"It takes energy to break the icosahedral formations so you can get over the barrier and into a new phase," said Kelton.
Unfortunately, Frank lacked Kelton's access to NASA's Electrostatic Levitator to confirm his predictions. Also, the 1984 discovery of quasicrystals, with NASA's experimental setup, enabled Kelton's research team to crack the case for good.
Quasicrystal less stable
NASA's levitator, which was moved to the Advanced Photon Source at Argonne National Laboratory in Chicago for the experiment, used a beam of X-rays focused on a suspended drop of titanium zirconium nickel to map the location of the atoms as they crystallized into a solid.
When they were done, the researchers observed the icosahedral structures breaking up as they rearranged into the crystals of a solid, just as Frank predicted. As the temperature was lowered, the icosahedral quasicrystals nucleated first, because quasicrystals have a lower energy than the crystals. But because the quasicrystal is less stable than a normal crystalline solid, the barrier is eventually overcome with the transition to the solid phase.
NASA's Materials Science Program. which is managed by the Marshall Center, is sponsoring the research in the hopes of creating new materials for space. An upcoming International Space Station experiment in space's zero-gravity environment, sponsored by NASA, may shed further light on the nucleation process.
The current research was funded by the Physical Science Research Program, which is part of NASA's Office of Biological and Physical Research in Washington and NASA's Marshall Center.