Artistic impression of light emission from SiGe. [Photo of CompSOC by K.G.W. Goossens and Bart van Overbeeke.]
Artistic impression of light emission from SiGe. [Photo of CompSOC by K.G.W. Goossens and Bart van Overbeeke.]

Silicon dominates the electronics industry despite not having a direct band gap, which means that the material cannot emit light efficiently. But by altering the crystal structure of silicon, researchers show that hexagonal silicon alloyed with germanium can be induced to emit light efficiently [Fadaly et al., Nature (2020), https://doi.org/10.1038/s41586-020-2150-y].

“The performance of electronic chips can be enormously enhanced if signals could be sent optically instead of electrically, since light has no resistance,” explains Erik P. A. M. Bakkers, who led the work. “But as Si and Ge cannot emit light it has not (really) been possible to integrate optics into Si electronics.”

Bakkers and his colleagues at the Eindhoven University of Technology, Johannes Kepler University, Eurofins Materials Science Netherlands, Technische Universität München, and Jena University calculated the band structures of hexagonal SiGe in varying proportions using ab initio density functional theory. At a certain value of x for Si1-xGex, the calculations predict a direct bandgap is tunable over the range of 0.3−0.7?eV. Indeed, the researchers were able to detect short bursts of light emission from hexagonal Ge-rich alloys as predicted.

“We observed very bright emission with short lifetimes exactly at the energies where the emission was expected based on band structure calculations,” says Bakkers. “We measured the light emission by photoluminescence, as a function of temperature, and we also looked at the specific lifetimes of the emission.”

Moreover, the observed emission decreases with increasing temperature, as would be expected for a direct-bandgap semiconductor. The researchers believe that the light emission from the direct bandgap arises because the cubic structure of SiGe alloys is subtly altered to the hexagonal form. The difference lies in the stacking of the atomic layers. In the usual cubic form, atomic layers are stacked in an ABC… repeating pattern, where the third layer does not sit directly above the first layer. In the hexagonal form, however, the stacking pattern is ABA, with the third layer directly above the first.

To produce the hexagonal form, the researchers grew Ge-rich Si1-xGenanowires around thin Au-catalyzed GaAs cores lattice-matched to Ge. The thin GaAs core layer reduces lattice strain and the corresponding defects, while the Au catalytic particles are removed via wet chemical etching.

“The evidence for the hexagonal structure comes from electron microscopy, which provides unique signals for the hexagonal structure,” points out Bakkers.

The results indicate that using hexagonal Si1-xGex could enable the integration of optical functionality with Si technology, increasing the speed of operation of chips while simultaneously reducing energy consumption.

For now, admits Bakkers, their approach to fabricating hexagonal Si1-xGeis not compatible with Si processing and the GaAs substrate needs to be replaced with an alternative platform. The researchers are now looking for more practical and scalable methods, he says.

“Ultimately, [this could lead to] a revolution in the electronics industry,” Bakkers says. “We are now trying to get this material to lase [which] would be spectacular, but also show its usefulness,” he adds.

This article originally appeared in Nano Today 34 (2020) 100951.