DIY apodized phase plates/objectives?

[Xiphosura]

I'm looking into a diy solution to convert preexisting 40x 0.65 NA and 100x 1.3 (oil) phase contrast objectives into apodized ones to reduce halo artifacts and to visually enhance the resulting image. The technique itself is remarkably simple and it shouldn't be too hard to implement, but I reckon it's going to be deceptively difficult to do well

Apodized PC concept and technical aspects:

• https://www.microscopyu.com/techniques/ ... e-contrast

• sci-hub.do/10.1007/s10043-000-0119-5

So, apodized phase plates consist of a classic phase ring and two secondary neutral-density filter (ND) rings with higher transmittance adjecent to the inner and outer bounds of the phase ring.

Materials:

• non-reflective ND film , 0.3 mm thick, 40% transmittance (best I could find)

Ideas:

1. Precisely positioning and gluing/cementing ND filter rings to the phase plate with glue (or canada balsam) for glass or optical instruments.
2. Cutting a glass disc to the appropriate size to be inserted into the objective and on top of the phase plate. ND film rings will be glued/applied to the glass disc.
3. Somehow creating a removable insert using a 3d printer, in which II will be positioned.
4. Start from a non-phase contrast objective of the same brand/series and construct a phase ring and both ND rings in tandem, based off the sizes of actual phase contrast objectives.

Also, the math on calculating the ND width is still a bit tricky, though, but I hope I'll figure it out soon.

Before I start tinkering I'd like to hear your thoughts on this concept and my ideas. Do you think it's infeasible or do you have concerns? Are there any improvements or better, alternative methods you can think of?

[abednego1995]

Why not use a 4f relay to externalize the objective BFP? It should be much easier to introduce fourier filtering elements in the intermediate plane there.

Cheers,
John

[Xiphosura]

That sounds like a phenomenal idea and might be way easier to test experimental setups with, though I currently don't have any experience nor the technical understanding to know how to install such relay system. How would one do that with finite scopes? I currently have an Olympus BH-2 with a phase contrast setup and a Zeiss Standard (with Optovar).

[BramHuntingNematodes]

I wouldn't go out and say it's impossible but were dealing with some pretty small margins for error here. Careful gluing stuff to your phase plate as I imagine you will not get it right the first time and then you are out a perfectly good phase plate. Also, maybe reflection should be the principle of the ND rings? I think you want to be careful that your apodizimg rings don't introduce additional phase shifts.

[Xiphosura]

I wouldn't go out and say it's impossible but were dealing with some pretty small margins for error here. Careful gluing stuff to your phase plate as I imagine you will not get it right the first time and then you are out a perfectly good phase plate. Also, maybe reflection should be the principle of the ND rings? I think you want to be careful that your apodizimg rings don't introduce additional phase shifts.

I was indeed wondering if the type of ND filter (absorptive vs. reflective) would make any difference, but I couldn't find any information on this. So I'm wondering what led you to think reflective ND filters should be used?

I've found a different type of ND film that's thinner, has 50% transmittance, and is made of a gelatin-derived material.
Based on exactly 0.1 mm thickness (= 100000 nm), 550 nm light, a refractive index of 1.5, and assuming uniform thickness, it shouldn't produce any significant destructive nor constructive phase interference. However, in the case that the thickness might be off by a few percent, then interference might become a problem, though it's hard to estimate the severity.

[BramHuntingNematodes]

Ideally you want to sputter some metallic thin film on a plate as this would not have an effect on the phase of light that that penetrates. I worry if you somehow manage to cut two concentric rings from this material with a cutting plotter or whatever it's going to do something to the phase.of the surrounding light which may have some undesirable effect in the phase contrast.

[patta]

Unfortunately I agree with the previous issue, if you glue an extra thick ring over the phase plate, it will destroy the phase coherence.
I think that 0.1mm does not mean exactly 100,000 nm, but has an error or plus minus 10,000 nm, so you never know where the phase will be. A plate exactly 100,000 nm thick is extremely difficult to manufacture, and will cost 100,000 US$.
Make the ND apodization by sputtering metal, may work; you may ask to those that prepare samples from electron microscope, they're sputtering gold all the time.

The other suggestion, put a relay lens and work at the aperture there, is nice and likely more feasible but still, pretty complicated; and will need to have some other film besides the ND, so the surface is flat and phase is preserved.
One way is to use polarizers /retarders to get the apodization; you use the same layer (so same thickness) but with different orientation (so retard or absorption), see f. ex
viewtopic.php?f=28&t=13237&start=30 post #42

My suggestion, first attempt, try to make an annulus for the condenser with the ND rings, it is much easier and may work as well. The condenser annulus is more easily accessible than the objective phase plate! Once you've mastered the art of cutting rings with 0.2 mm precision out of the ND gelatine, you can move to the objective...

You can also try "color apodization", maybe it works, see for example the last image of post #1 here; the green ring with red borders.
viewtopic.php?f=28&t=13095
The idea is to then apodize digitally, subtract the red channel from green.

[BramHuntingNematodes]

Still working on that but it's taking a little longer as I decided to build a clean hood for assembling the plates.

[Xiphosura]

Unfortunately I agree with the previous issue, if you glue an extra thick ring over the phase plate, it will destroy the phase coherence.
I think that 0.1mm does not mean exactly 100,000 nm, but has an error or plus minus 10,000 nm, so you never know where the phase will be. A plate exactly 100,000 nm thick is extremely difficult to manufacture, and will cost 100,000 US$.
Make the ND apodization by sputtering metal, may work; you may ask to those that prepare samples from electron microscope, they're sputtering gold all the time.

The other suggestion, put a relay lens and work at the aperture there, is nice and likely more feasible but still, pretty complicated; and will need to have some other film besides the ND, so the surface is flat and phase is preserved.
One way is to use polarizers /retarders to get the apodization; you use the same layer (so same thickness) but with different orientation (so retard or absorption), see f. ex
viewtopic.php?f=28&t=13237&start=30 post #42

My suggestion, first attempt, try to make an annulus for the condenser with the ND rings, it is much easier and may work as well. The condenser annulus is more easily accessible than the objective phase plate! Once you've mastered the art of cutting rings with 0.2 mm precision out of the ND gelatine, you can move to the objective...

You can also try "color apodization", maybe it works, see for example the last image of post #1 here; the green ring with red borders.
viewtopic.php?f=28&t=13095
The idea is to then apodize digitally, subtract the red channel from green.

The 'color apodization' method seems ingenious and surprisingly simple in concept, I'll certainly give that a try.

If I remember correctly, Zeiss Optovars correct to infinity and may already act as a relay lens? I'm in possession of a second unused Zeiss Optovar with horribly delaminated lenses in the 2x slot. Do you reckon I could use an empty slot to insert DIY filters/phase plates/polarizers (with or without a custom filter holder) to experiment with?