# Synopsis: New Recipe for Optics-Friendly Silicon

#### Genetic-Algorithm Discovery of a Direct-Gap and Optically Allowed Superstructure from Indirect-Gap Si and Ge Semiconductors

Mayeul d’Avezac, Jun-Wei Luo, Thomas Chanier, and Alex Zunger

Published January 12, 2012

Silicon is the darling of the microelectronics industry, but it has an Achilles’ heel: silicon can’t absorb or emit light without the help of phonons. This so-called indirect band gap makes it an inefficient option for light-emitting diodes and solar cells.

Researchers have tried to engineer a direct band gap in silicon by introducing defects or casting it in nanostructure form, but such materials don’t necessarily have an optically active band gap. Writing in Physical Review Letters, Mayeul d’Avezac at the National Renewable Energy Laboratory, Colorado, and his colleagues report numerical predictions that a material consisting of atomically thin layers of silicon and germanium—a superlattice—should have an optically active, direct band gap.

Superlattices of silicon and germanium are known to have direct band gaps that allow weak optical absorption, but to find the optimum mixture of the two materials requires sorting through all possible combinations. For this, D’Avezac et al. adopt an algorithm originally developed in the 1950s to simulate evolutionary random selection. In their case, they combine two parent superlattices into a new and possibly better “offspring”. A superlattice becomes “fitter” as it develops an optical transition, and “mutations” occur by randomly swapping germanium and silicon layers.

According to the team’s simulations, a magic formula of ${\text{SiGe}}_{2}{\text{Si}}_{2}{\text{Ge}}_{2}\text{Si}$ (the subscript denotes the number of monolayers) grown on a silicon-germanium alloy and topped with a germanium buffer should be 50 times more efficient at absorbing light than existing silicon-germanium superlattices. In principle, such structures could be prepared with molecular beam epitaxy. – Jessica Thomas