"Light holds the key to some of nature's deepest secrets, but it is very challenging to confine it in small spaces," says Antonio Badolato, professor of physics at the University of Rochester and corresponding author of the Applied Physics Letters paper. "Light has no rest mass or charge that allow forces to act on it and trap it; it has to be done by carefully designing tiny mirrors that reflect light millions of times."
Nanocavities are key components of nanophotonics circuits and Badolato explains that this new approach will help implement a new-generation of highly integrated nanophotonics structures. Researchers are interested in confining light because it allows for easier manipulation and coupling to other devices. Trapping light also allows researchers to study it at its fundamental level, that is, at the state when light behaves as a particle (an area that led to the 2012 Nobel Prize in Physics).
Until now, researchers have been using educated-guess procedures to design the light-trapping nanostructures. However in this case, the team of researchers – which included lead author and Badolato's Ph.D. student, Yiming Lai, and groups from the Ecole Polytechnique Federale de Lausanne, Switzerland, and the Universita di Pavia, Italy– perfected a numerical technique that lead to the design improvement. Their computational approach allowed them to search for the optimal combination of parameters among thousand of realizations using a "genetic" (or "evolutionary") algorithm tool.
The principle behind the genetic approach is to regard each new nanocavity as an individual in a population. The individuals mutate and "breed," meaning that two single structures combine to create a new one that is a cross between the two "parents." As new generations succeeded one another, the algorithm selected the fittest ones in each generation, in this case, the ones that exhibited the longest trapping time (i.e. highest quality factor).
