Metalenses at visible wavelengths: development of flat lenses

[lynnb]

See PetaPixel news and link to Science magazine paper.

This seems to be a genuine breakthrough in optical design.

“In my opinion, this technology will be game-changing,” Federico Capasso, professor at Harvard University and senior author on the Science paper, told The BBC. “The quality of our images is actually better than with a state-of-the-art objective lens.” These are real, and they’re already generating exceptional results.

The "metalenses" use a layer of Titanium Dioxide on a planar surface, to focus the light, effectively this is a flat lens which can be fabricated in computer chip factories. The paper cites "diffraction-limited focusing demonstrated at wavelengths of 405, 532, and 660 nm", i.e. within the visible light spectrum of 400-700nm.

 

[michaelwj]

If I remember correctly, the same group previously demonstrated this but at a single wavelength and with not so good image quality. Nice to see they got continued funding 😀 Hopefully more to come...

 

[sevo]

We'll see whether it is universally applicable, and whether its results are well suited to postprocessing. There already are quite a few alternatives to conventionally refracting lenses, from pinholes and fresnel lenses over holograms to phase plates, but all of them so far only fill irrelevant niches where conventional photography is concerned.

I predict that any alternative to the traditional multi-element glass lens will require electronic processing to reconcile the resultant images with the limitations of human vision - the eye is a refracting lens/camera system, and our cameras inherently deliver results that are not too alien. Other optical systems can be jarringly different in some aspect.

 

[popavvakum]

Many thanks for the reference, very interesting indeed. I cannot get the access to the full article but from the abstract the principle may be easily seen. Two questions come straight: 1) focussing through an array of sub-wavelength reflective surfaces means that the inclination of those surfaces needs to be adjusted to the angle, and possible the wavelength of the coming light. 2) the edges may well require different inclinations and/or density of those little "pillars" reflecting the light. So that means that to make a practical lens producing an image large than say a couple of mm, one would need a way of actuating the little pillars. I guess it is possible by inducing a variable electric field on the surface but it is unlikely to happen within a size of a mobile phone any time soon.

Another thing - "the diffusion-limited" focussing - I presume, from the wavelengths they mention, that they focussed highly coherent laser beams. How would that compare with focussing of a random "usual" light?

Ouph, my Jupiters will have to serve a little longer... But the concept is very interesting nonetheless.

 

[Spanik]

Nice to see that there are still people thinking. Might not make it to largescale production for large sensors in my lifetime but they have said that before of other things.

Even if they are strongly wavelength dependent they could put such an optimised lens before each pixel. Once you start with semiconductor manufacturing tech such things are possible.

We'll see whether it is universally applicable, and whether its results are well suited to postprocessing. There already are quite a few alternatives to conventionally refracting lenses, from pinholes and fresnel lenses over holograms to phase plates, but all of them so far only fill irrelevant niches where conventional photography is concerned.

While they don't look like conventional lenses, Fresnel lenses are conventional refracting lenses. They just cut a lot of glas away that was not contributing to the working.

I expected to see at least some of those liquid lenses around now. They were announced years ago but it has become very quiet around them.

 

[giganova]

I see two problems right away: these silicon "pillars" create only one reflecting surface, which creates massive aberration. You need multiple stacked reflecting surfaces to correct for these errors. Even if you manage to correct this, you only found a solution for one wavelength!

BTW, we have been using this technology at NASA (where I work) for many space telescopes that operate at X-ray wavelengths since the 1950s, so this is not new at all.

 

[Keith]

Way over my head! 😀

Interesting though because optics theory is something that I always assumed was cast in stone ... not so it seems!