Prof. Elif Ertekin, doctoral student Channyung Lee and colleagues at the University of Utah have been analyzing atomic-scale simulations of monoclinic gallium oxide, a semiconductor material with promising applications in power generation and transmission. A vacancy -- an empty spot within the material's crystalline pattern -- can move like an inchworm through the material, a long-hypothesized phenomenon known as a "crowdion." Identifying how and when crowdions form could improve the manufacturing process for gallium oxide, making it capable of withstanding higher temperatures and voltages.
Written by Evan Lerner, University of Utah
The semiconductors at the heart of high-tech applications are some of the most carefully scrutinized materials in the world. Even miniscule changes to their chemical compositions can lead to major improvement to their performance.
So it is rare for materials scientists to find something entirely surprising when analyzing these materials, but a team of researchers from the University of Utah and the University of Illinois have discovered the molecular equivalent of a Sasquatch.
Proving the existence of semiconductor crowdions gives the researchers a new direction when it comes to engineering these materials.
Called a “crowdion,” it is a long-hypothesized phenomenon within crystalline lattices
featuring a rapid chain reaction of atoms crowded together repositioning, much like the suspended spheres in a Newton’s Cradle. The sequence of shifting bonds moves the location of an extra atom called an interstitial (or in this case, a vacancy — a spot in the lattice where an atom is missing) within the material.
The researchers observed this phenomenon in a variety of gallium oxide, a semiconductor eyed for its use in high-current, high-voltage applications. Controlling the formation of crowdions could make this extreme material even more durable. These moving defects might even be a useful feature of other semiconductors, with outside atoms riding these molecular conveyor belts from one part of the material to another.
MechSE Associate Professor Elif Ertekin and her PhD student Channyung Lee collaborated with Michael Scarpulla, a professor in the departments of Electrical & Computer Engineering and Materials Science & Engineering at the University of Utah.
In this animation (slowed down billions of times), the crowdion (highlighted in light blue) "moves" through the material by shifting a sequence of bonds between gallium ions (blue) and different sets of oxygen atoms (red). This is enabled by the presence of vacancies (white), or spots in the crystalline lattice that are missing a gallium atom.
Materials get their properties from the way their component atoms line up, which in turn determines how those atoms behave. Crystalline materials have their atoms tightly bonded in neat rows and columns; a piece of pure iron, for example, consists of iron atoms in a cubic lattice.
Elif Ertekin
“The crowdion was first postulated in metals and silicon, when people were trying to figure out what would happen if a crystalline material had too many atoms in one spot — hence the ‘crowd’ in ‘crowdion,’” says Scarpulla. “It’s relatively easy to imagine in pure metals, but for semiconductors, the crowdion was thought to exist only under extreme conditions, if at all.”
Compound semiconductors have both metallic and nonmetallic elements in their lattices; the ionic oxygen in gallium oxide, for example, was thought to be too brittle to allow for the kind of sequential repositioning that more easily occurs in pure metals.
This quality is also one of the reasons the material is eyed for applications in power electronics: gallium oxide can withstand currents, voltages, and temperatures that would cause standard semiconductor materials like silicon to break down and fall apart. The researchers were therefore interested in getting a more comprehensive picture of the kinds of defects that can form in gallium oxide, with the ultimate goal of making it even more resistant to extreme conditions.
Given the difficulties of conducting experiments at the scale of individual atoms, simulations are invaluable to research on point defects in semiconductors. Lee, the study’s lead author, first made an exhaustive list of the kind of defects that could conceivably occur within gallium oxide’s crystalline structure, then painstakingly simulated the necessary conditions to see if they would actually form.
When Lee pulled back the virtual ball representing a gallium atom and let it collide with the rest of the lattice, the researchers were shocked to find that it set off a crowdion chain-reaction that synchronously moved multiple atoms, resulting in the vacancy defect moving deeper into the bulk of the simulated material.
Proving the existence of semiconductor crowdions gives the researchers a new direction when it comes to engineering these materials.
“It turns out that, because of these inch-worm-like defects, the atoms inside this material can move around really fast, and that can lead to the kind of breakdown we’re trying to prevent,” says Ertekin. “So if we want to make these materials stronger, we need to know more about why these crowdions are so mobile in the first place, and how we could engineer the material to limit that mobility”
Power electronics are not the only applications for gallium oxide, and there are many other semiconductor materials with the same basic structure that are likely to exhibit this phenomenon. More research into crowdions could unlock new abilities in entirely different fields.
“Sometimes you want materials with ions that are easy to move. We might be able to transport other atoms deeper into the materials using those moving defects, and that could be useful for things like batteries and fuel cells.”
Channyung Lee, doctoral student in MechSE
"Investigation of Ga interstitial and vacancy diffusion in 𝛽−Ga2O3 via split defects: A direct approach via master diffusion equations." Channyung Lee, Michael A. Scarpulla, and Elif Ertekin
The authors acknowledge the funding provided by the Air Force Office of Scientific Research under Award No. FA9550-21-0078. This work used PSC Bridges-2 at the Pittsburgh Supercomputing Center through allocation MAT220011 from the Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program, which is supported by National Science Foundation Grants No. 2138259, No. 2138286, No. 2138307, No. 2137603, and No. 2138296.