Beating the diffraction limit

[pwnell]

I just read this - what do you guys think?

http://www.gizmag.com/optical-lens-one- ... ick/41588/

It is always thought that the diffraction limit places a hard stop on the smallest detail we can resolve using photons*. Is there any absolute limit that would prevent even above optics to not go any further?

*any particle really that exhibits wave-particle duality

[Asha]

Maybe for monochromatic light. As a diffractive optical element, it would probably have a grating effect, causing chromatic dispersion in broad-band light. That would have to be compensated, somehow, if it were to be put to general use.

[rjlittlefield]

The gizmag article does not accurately represent the original paper, which can be read HERE.

In the original article, the authors consistently speak of coming close to diffraction limited performance.

--Rik

[Asha]

Yes, they don't claim to exceed the diffraction limit that I could see. They did address the issue of dispersion--they used a broadband laser to etch the features in the lens, and also did some testing to see if the flexible mounting material could be controlled to improve focus.

Interesting tidbit that was buried in the article--the smallest feature they can create on the film determines the finest resolution of the image. So...I'm assuming they are using traditional optics to get the fabrication laser to focus or collimate to this small point? I wonder if they are using something like what is used in the chip manufacture industry to get very tiny traces on Silicon.

[mawyatt]

Yes, they don't claim to exceed the diffraction limit that I could see. They did address the issue of dispersion--they used a broadband laser to etch the features in the lens, and also did some testing to see if the flexible mounting material could be controlled to improve focus.

Interesting tidbit that was buried in the article--the smallest feature they can create on the film determines the finest resolution of the image. So...I'm assuming they are using traditional optics to get the fabrication laser to focus or collimate to this small point? I wonder if they are using something like what is used in the chip manufacture industry to get very tiny traces on Silicon.

I think the Silicon processing folks routinely defy diffraction limits by working the problem backwards. Basically they know what the shape of the exposure they need and work back to what the masks would need be to produce the desired shape, including utilizing multiple masks. I guess you can think of these "masks" as very thin lenses, specifically designed to image one subject! These mask computations are very involved and require Super Computers to work through.

Maybe someone on here in silicon processing fabrication (I'm in the circuit design side) can shed some light on this (pun intended :>)

Cheers,

Mike

Best,

Mike

[Asha]

mawyatt, when I posted, I couldn't remember the name, but it is a micro-lithography lens. It is basically like a giant microscope objective used in reverse, and for very small features, they tend to use the blue to UV spectrum. I'm not even sure if the materials used for high UV transmission would pass anything in the red to NIR regime, which the paper stated was used. So, this may be a bit of a red herring, lol. I was just wondering out loud what kind of setup the researchers would need to create very small blazing on their diffractive elements.

[Asha]

Rik, a couple more thoughts about the diffraction limit...at my work, we routinely use infrared wavelengths. Since diffraction physics is wavelength dependent, these systems reach the diffraction limit quite easily, even with optical mirrors that have visible machining marks which would destroy image quality in the visible regime. My point is that we never have systems that exceed the diffraction limit, even with very long wavelengths, because the line is a moving target. In addition, diffraction limited performance is easily achieved in the visible if the criteria is such that there are few cycles per radian / line pairs per mm.

[mawyatt]

mawyatt, when I posted, I couldn't remember the name, but it is a micro-lithography lens. It is basically like a giant microscope objective used in reverse, and for very small features, they tend to use the blue to UV spectrum. I'm not even sure if the materials used for high UV transmission would pass anything in the red to NIR regime, which the paper stated was used. So, this may be a bit of a red herring, lol. I was just wondering out loud what kind of setup the researchers would need to create very small blazing on their diffractive elements.

Asha,

The feature size in advanced silicon CMOS processes is very small. Intel has been shipping 14nm CMOS chips for a couple years, these are in Apple's high end iPhones, iMacs, Mabooks, & iPads, and also some high end PCs & laptops. Samsung has been working on 10nm CMOS for a couple years and has had 14nm available in products for some time also, 7nm is right around the corner!!

10nm wavelength is the transition from UV to X-Ray, I find this absolutely amazing that they can create features on silicon this small. This feature is the minimum CMOS FET gate length, or distance between the FET drain and source and there are well over 1 billion CMOS FETs on a single silicon chip!!

Best,

Mike

[Asha]

mawyatt, I looked at the article again--it shows the measured profile in Figure 1, with the rings at <1um FWHM. They probably can get away with a less stringent setup than what we've been discussing (the same figure also shows a "High NA objective")

[bobfriedman]

The gizmag article does not accurately represent the original paper, which can be read HERE.

--Rik

interesting.. thx rik