Slashdot Mirror
Metamaterial Forms Near-Perfect Mirror
New submitter JMarshall writes: Researchers have made near-perfect reflectors out of a silicon metamaterial. These reflectors could offer a simpler, less expensive way to make high-performance mirrors for lasers or telescopes. Metamaterials typically use nanoscale patterning to create unusual properties not present in the bulk material. In this new method, researchers used off-the-shelf, nanosized polystyrene beads and allowed them to self-assemble into a monolayer with a hexagonal pattern. Using the monolayer as a photolithographic mask, the researchers etched an array of silicon cylinders, each a few hundred nanometers across, onto a wafer. The cylinders act like tiny resonators for a particular light frequency—analogous to the way a given sound frequency will make a tuning fork hum. The array reflected 99.7 % of incident light at their peak wavelength. These simple metamaterial mirrors might one day replace current high-performance reflectors, which are somewhat costly to make.
Sure, metal mirrors such as silver may only give you 95-97% reflection. However, dielectric mirrors are pretty common (aka distributed Bragg reflectors) and a quick check on ThorLabs shows some with efficiencies of at least 99.5%. The paper and summary over hype this result by suggesting that these new mirrors will be much cheaper over large scale. Distributed Bragg reflectors rely on multiple coatings of thin dielectrics, which can be scaled to large areas fairly easily and controlled precisely. The presented work uses microsphere lithography, where you need to get tiny spheres to pack closely on a substrate EXACTLY one layer thick. Having tried this process personally, that's a lot more difficult than these papers usually let on, and the frailness of getting the particles to settle over large areas makes scaling to telescope size unlikely.
Oh, and the authors took the easier route and demonstrated a mirror for telecom wavelengths, ~1500-1600nm. To make a mirror in the visible range requires smaller spheres which suffer from poorer packing due to a larger coefficient of variance in the diameter.