Bismuth telluride (Bi₂Te₃) and gallium arsenide (GaAs) make quite a pair, combining to form a highly efficient electronics-cooling thermoelectric material. Although scientists have been using this material pair for years, how the two materials stick together has remained a mystery. Now, with the aid of high-powered, highly sensitive imaging systems, researchers at North Carolina State University have found a definitive answer.

Most inorganic compounds are held together with chemical bonds, says James LeBeau, a professor of materials science and engineering at North Carolina State University. But for some time, scientists have known that Bi₂Te₃ and GaAs do not follow this pattern. Instead, the two materials combine through van der Waals interactions that are considerably weaker than a chemical bond. Just how the bond is formed and how it manages to hold the system together was the real mystery, says LeBeau.

Solving the mystery was a difficult task. Identifying the van der Waals connection between two thin sheets of materials required instrument resolution higher than most standard systems could provide. LeBeau and his team combined an atomic-resolution aberration-corrected scanning transmission electron microscope (STEM) with a high-powered atomic resolution x-ray spectroscopy system. Such a combination would theoretically image not only the physical structure of the Bi₂Te₃ and GaAs, but also the chemical structure at the binding interface.

Scientists had some suspicions as to what was occurring at the interface. GaAs would have dangling bonds that would prevent van der Waals epitaxy. Therefore, either bismuth or tellurium had to interact. To pinpoint which element was reacting with the GaAs, LeBeau and his colleagues along with researchers from RTI International used a vapor deposition technique. Using GaAs as a substrate, Bi₂Te₃ was grown layer by layer.

The results, which were published in Applied Physics Letters, were surprising, says LeBeau. The STEM images show that the interface between Bi₂Te₃ and GaAs is very bright, suggesting that a heavy element must be settled there. The interpretation of the image, says LeBeau, points to the bismuth interacting with the arsenide at the interface.

However, x-ray spectroscopy told a different story. The chemical makeup at the interface included tellurium, not bismuth.

An ultrathin layer of gallium telluride—less than a quarter the width of a normal GaTe structure—solves the long-standing mystery as to how Bi₂Te₃ grows through van der Waals epitaxy into remarkably pristine lattices.

“The crystallography tells us that these materials should not grow nicely on top of one another, but they do,” says LeBeau. “And it’s ultimately because there are no strong bonding forces directly between the substrate and the film that enable that.”

Electron micrograph of the Bi₂Te₃/GaAs interface. Image credit: James LeBeau.