Is There a Commercially Available Filter for Collimation by Filtering?

[stg]

Hello!

If one were to push uncollimated light through a tube of inner radius R and length L, coated with matte black - the maximum incidence angle it will allow through would be ATAN2(R, L). On the other end of the tube, only the "most collimated" fraction of light would exit - a very small fraction for a narrow tube.

Is this principle available as an optical filter in sheet form? For example, simply drilling, etching, etc. a large number of very small holes (fraction of sheet thickness) in a thin sheet of absorbtive material would suffice and it would seem to be this is something that ought to be available - possibly going by a name i simply cannot figure out.

Any ideas or tips would be most welcome.

 

[Tom.G]

Sounds like the 'Privacy Screens' you can put on your computer display so they can be seen only when you are directly in front of them. I think 3M makes some.

 

[stg]

Thank you!
I feel silly now, not having thought of those!

I'm finding a lot of such privacy screens cut to size, including from 3M - but no mention of what this type of material/filter/grating is called or how to go about finding the material, not a finished cut to size product with a frame or self-adhesive backing etc. Also, in it's productified form there are pretty much no useful data at all on their optical characteristics.

At the very least you've taken me one step closer, but if anyone has further information I'd be plenty happy to partake :)Davey

 

[blue_leaf77]

Is this principle available as an optical filter in sheet form? For example, simply drilling, etching, etc. a large number of very small holes (fraction of sheet thickness) in a thin sheet of absorbtive material would suffice and it would seem to be this is something that ought to be available - possibly going by a name i simply cannot figure out.

Such idea might be useful for XUV optics to collimate XUV beam. But the mechanism of collimation would be through the zeroth order diffraction order (remember, light is wave, you cannot overlook diffraction effect when you pass a beam of light through an obstacle), instead of by blocking or absorbing rays with directions above certain angle. By taking only the zeroth order diffraction and blocking the higher orders, you may get a collimated light of reduced power because you actually take out only a part of the original beam.
I don't see any useful application for visible light though, because there have been sophisticated parabolic mirrors which is able to do the same task with high precision.

 

[stg]

A parabolic mirror or lens is not quite practical in this case.

The setup is a linear array visible light sensor on one end, and a led illuminated backlight on the other with a gap between them.
The setup detects where and how much light is blocked by items within the gap, along the sensing line.
We're talking about a small unit - the gap is 1cm and the sensing length is 3cm.

I am scanning and counting objects of various sizes and shapes by moving them through this gap.
Because the backlight emits light that is not collinear and the gap is sometimes large in relation to the size of the object the light sensor produces a blurred image of the object blocking the gap between the light source and sensor.

By filtering out the least collinear light, and increasing the intensity of the light source I would get a much clearer image that would be much easier to accurately process.
Even an off the shelf privacy screen would likely increase the spatial resolution, but I would like to stick with fully specified components.

 

[blue_leaf77]

To me it sounds like you are trying to pinpoint the rays coming from which LED unit(s) are blocked by which part of the object in the gap. You somehow want the detection to be more accurate by letting only forward rays to reach the object, it's like you want to transform the LED arrays into small laser arrays. In this case, your own idea in the first post may work if the holes in the array are not bigger then the size of a LED but also not too small so as to approach the wavelength of the LED in order to suppress the effect of diffraction. Probably make the diameter at least $100\lambda$.

 

[stg]

@blue_leaf77 Sounds like you got it just right! And since I have the ability to amp up the light source about 50x, a filter that would only allow down to 2% of mostly collinear light through would be perfect and it would also only have to have filtering in one direction (which might be the case for privacy screens?). The diameter of the holes would not have to be that small at all since the sensor resolution (10µM) is not big enough to warrant such a high resolution "mesh" so I am not worried at all about approaching the wavelength - 1µM holes with as low as 50% coverage in a hexagonal pattern simulates quite nicely.

I still have hope there is already a commercial application for such filters so that they can be found without having to resort to custom fabrication as this would be too expensive. My other option is to use a much faster light sensor and "scan" the light source using DSP to recover a higher resolution image, but this is also at the very edge of what is economically reasonable.

If I could find privacy screen material with a much narrower angle this would likely work fine in this particular instance.