High-contrast lenses could find use in artificial neural networks

Cambridge University researchers have created microscale Fresnel lenses that could have applications in sensors, optical communications and even artificial neural networks.

Compared with a conventional lens design of comparable aperture and focal length, a Fresnel lens requires less mass and volume, allowing it to be thinner and flatter, capturing more oblique light from a light source.

The first Fresnel lens was installed in 1823 in the Cordouan lighthouse, where its beam was visible for 32km. Since then, this lens design has been used in lighthouses, traffic lights, automobile headlights, magnifying glasses and cameras.

One limitation, however, is that reflections from the opaque zones in a Fresnel lens can degrade the focusing and lensing properties. Dr Tim Wilkinson and colleagues at the Department of Engineering at Cambridge University are said to have overcome this limitation by using the darkest man-made material ever — low-density, vertically aligned carbon nanotube arrays.

According to a statement, the team created arrays of carbon nanotube Fresnel lenses, each 77 micrometres in diameter and with 15 zones. These lenses have exceptional optical properties, with high contrast and efficient focusing.

The team said these fabricated carbon nanotube Fresnel lenses offer new possibilities in designing highly flexible and efficient interconnection networks. Potential applications include efficient focusing, deflecting and collimating tasks in optical sensor systems, optical computers, optical data transfer, optical communication and even integration into 2D source arrays for neural network architectures.

Their findings are published in an article in the journal Advanced Optical Materials.