Dodging the infinity hump

[Beatsy]

I've had my main scope, a Zeiss ICM405 inverted, for nearly twenty years now and it's still in almost daily use. I'm perfectly content with it for all my general microscopy needs and it isn't going anywhere.

But diatoms (a special interest) cause a problem. I still chase maximum resolution for examining and photographing some of those and often hit the limits of what my kit can resolve (or reveal). That's been a recurring irritant for a few years now.

I know I'm working close to the physical limits already, for sure, (even using blue/violet light etc) but I've been aware of a "potential improvement" for many years. Namely PlanAPO objectives with NA greater than the usual 1.4. Sadly, the highest (over 1.5) require sapphire cover slips and poison for immersion oil. Which leaves the NA1.46 ones by Zeiss. These work with normal immersion oil and glass and they've started appearing on the second hand market at £6000 and up. In short, drifting within financial reach...

Trouble is, they're infinity objectives. My Zeiss is a finite tube-length system and the condenser only reaches NA1.4 anyway. My Nikon Labophot is infinity, but kitted out like an Epiphot and pretty much dedicated to diatom inspection and manipulation. There are no objectives >NA1.4 available for the Nikon anyway. So another plan is needed.

I looked into replacing my ICM405 with a newer infinity system, maybe a second hand Axio or similar from another main brand. They're super expensive in the UK with phase + DIC capabilities monstrously so. The equivalent objectives to the finite ones I use most aren't cheap either. It would be way too costly and could take longer than I have left on the planet to find the parts on the second hand market anyway. So I discounted that.

Next, I considered making a dedicated, single-objective frankenscope, purely for photographing diatoms in bright-field (keeping the ICM for everything else). Objective on a tube lens, old focus mech etc. But when you consider the physical support needs for "convenient" focus-stack shooting, and the precision needed at such high resolution, it has to be a proper microscope stand really. I do use DIC with mix-matched sliders for diatoms at the highest magnifications on my ICM - I know I'd eventually wish I had that too. So not an attractive route really. Far too many compromises.

Which leaves just one option, I think. Build a 365nm UV capability around the existing Zeiss scope - which is what I now plan to do.

It took me a while to reach this decision as I was thinking of replacing the ICM405, so the new kit would need to support everything else I use the ICM for or discontent would ensue. I also got a bit fixated on my pre-supposed solution (an objective) but once I realised the "problem" is restricted to resolution of diatoms, only, I figured wavelength is the one variable I *can* flex affordably and to better effect. Reading this paper sealed the deal ---> http://www.mikroskopie-ph.de/UV-diatoms-2011.pdf Essentially I'll get *way* more resolution improvement (for diatoms) by going this way than I would with the higher NA objective. Of course, UV *and* a hi-res objective would be even better - but that's future GAS

I need to get a camera converted to full spectrum, and might need to buy one for that (my older A7rii is a candidate but possibly not suitable, not sure yet). But everything else remains the same, with DIC and phase contrast capabilities still available. I don't have room for another full-sized scope anyway - so this seems a very good solution for me and removes the need to get over the infinity hump at all. Woohoo!

Only posted for conversation really. Anyone else going through similar "switch to infinity" issues?

[Scarodactyl]

I wonder how inaccessible solid immersion systems are to the average person https://www.ebay.com/itm/LEICA-TWI-350X ... 1285473379
I assume it's even worse than the ultra-HR liquid objectives.

I'd also note Zeiss isn't unique in edging over 1.4, and I would guess eg a Nikon system would cost less to assemble and be a bit less proprietary in general. No way to avoid that kind of system being expensive in general though.

[Lou Jost]

Beatsy, are you sure your desired objectives freely transmit 365nm UV?

One advantage of UV or other monochrome light, as I'm sure you know: if you use a barrier filter tuned to your light source's wavelength, you don't need apo objectives. In fact apo objectives may be unhelpful, as they have more UV-blocking elements, and resolution can be favored over color correction in non-apo objectives.

[Beatsy]

Beatsy, are you sure your desired objectives freely transmit 365nm UV?

My Neofluars and a couple of others do, but they're lower NA - might still be suitable though given the resolution-boost from using shorter UV wavelengths. I know my other objectives, including the high-NA Plan APOs don't "freely" transmit as they all attenuate a fair bit, to varying degrees. DIC sliders are quartz (no problem) and the two polarisers absorb about 30% each. The internal projection lens to the camera nabs a bit too. But a 365nm torch shone directly into the condenser at the lamp end still causes (weak) fluorecence at the front photoport. So yes, it'll be like working in the dark, but it looks like there's enough light for imaging. I don't mind if it has to be relatively long exposures, all the stack shooting is automated.

[Beatsy]

...
One advantage of UV or other monochrome light, as I'm sure you know: if you use a barrier filter tuned to your light source's wavelength, you don't need apo objectives. In fact apo objectives may be unhelpful, as they have more UV-blocking elements, and resolution can be favored over color correction in non-apo objectives.

Yes, I was thinking about that. In the near term I plan to use the objectives I have, hence APOs at the top end. But once the system is up and running, it will be relatively cheap to experiment with high-NA plan objectives in the hope of getting more light. If needed. See if it's a problem first though.

Edit to add: I think lack of correction for UV wavelengths will be the main bugbear with cheaper lenses. I already have a few objectives that can't focus violet (405nm) to the same plane as the other colours. I've a horrible feeling that might be the Neofluars, but I need to check.

[Lou Jost]

I already have a few objectives that can't focus violet (405nm) to the same plane as the other colours

But you don't need to care about this, if you are only using 405nm light.

If you do need to care about it, the VC line of Nikon objectives is explicitly corrected for 405nm light.

[Beatsy]

I already have a few objectives that can't focus violet (405nm) to the same plane as the other colours

...
But you don't need to care about this, if you are only using 405nm light.

I plan to use 365nm, but saw this at 405nm already. I assume it can only get worse as wavelength gets shorter. On top of that, I further assumed that aberrations kick-in or increase away from the designed focus distance (spherical, for instance). Hence why I *think* I need to care. Nice if I don't need to though.

[Lou Jost]

I further assumed that aberrations kick-in or increase away from the designed focus distance (spherical, for instance). Hence why I *think* I need to care.

Yes, you're surely right, I hadn't thought of that.

[Beatsy]

I wonder how inaccessible solid immersion systems are to the average person https://www.ebay.com/itm/LEICA-TWI-350X ... 1285473379
I assume it's even worse than the ultra-HR liquid objectives.

I'd also note Zeiss isn't unique in edging over 1.4, and I would guess eg a Nikon system would cost less to assemble and be a bit less proprietary in general. No way to avoid that kind of system being expensive in general though.

I read about this many years ago but assumed it was a surface-only contact technique at the time. Your post prompted me to follow a few more links and it seems it can extend past the contact plane (the "solid immersion lens" is called an "NA increasing lens" in that scenario). I think the lens you linked to will comprise a normal objective with a SIL mounted at the front (inside). In principle, it sounds easy as pie to do. Mount a diatom in the centre of the flat side of a small (1mm ?) hemispheric lens and then view it from the other side with a normal EPI objective (even low NA and/or LWD - like a Mitty). The numerical aperture of the image coming out of the hemisphere is the same as the RI of the material it's made of - 1.5 for glass, 1.76 for sapphire etc. Sapphire are the cheapest - which is handy. There are some available at RI 2.0, but they cost around £100 each, for a 1mm diameter hemisphere!

If this is as "doable" as it appears, then my EPI-equipped Labophot might support it. It certainly look a trivially easy experiment to do - in principle. Just need to get my grubby paws on a few sapphire hemispheres...

Thanks for mentioning this!

And yes, Nikon would be my first choice of "other manufacturer". In fact, if I wasn't so invested in Zeiss objectives and such, Nikon would likely be my first choice overall these days. Along with the Labophot, my working stereos are all Nikon too.

[Scarodactyl]

If you ever get an SIL objective working you will be a true hero. I think any enthusiast who has run across them has been curious what those marvelous NAs might reveal, but so far nobody's had the gumption as far as I know. I sure haven't anyway.

[rjlittlefield]

Mount a diatom in the centre of the flat side of a small (1mm ?) hemispheric lens and then view it from the other side with a normal EPI objective (even low NA and/or LWD - like a Mitty). The numerical aperture of the image coming out of the hemisphere is the same as the RI of the material it's made of - 1.5 for glass, 1.76 for sapphire etc.

Those words sound overly optimistic. My understanding is that the NA of the normal objective is only multiplied by the RI of the hemispheric lens. If that's correct, then I think you'll have trouble exceeding by very much the NA of an ordinary purpose-designed oil immersion lens.

Have I misunderstood something?

--Rik

Added ref: https://www.researchgate.net/publication/228364651_A_case_study_for_optics_The_solid_immersion_microscope

[Beatsy]

If you ever get an SIL objective working you will be a true hero. I think any enthusiast who has run across them has been curious what those marvelous NAs might reveal, but so far nobody's had the gumption as far as I know. I sure haven't anyway.

Won't be me. The RI of the substrate (which would be the mountant in/around a diatom in my case) has to be at least as high as that of the SIL material or NA is limited to the RI of the substrate. Tough call to match the 2.35 of the lens linked! Probably meant for imaging deeper metal layers through silicon on ICs, or something related to that (RI of silicon is 3.4).

Zrax, a high-RI diatom mountant has an RI of 1.7+ when properly hardened. This is a good match for glass and sapphire SILs (hemispheres). There is a better shape for the SIL too, the super-sphere, which is simply a bit more than half a sphere, say 3/5ths. This increases NA by up to RI squared I can envisage a way of starting with a sphere and grinding/polishing a flat like that.

This is a very intriguing little rabbit hole that I'm going to continue exploring anyway - independently of the UV stuff.

[Beatsy]

Mount a diatom in the centre of the flat side of a small (1mm ?) hemispheric lens and then view it from the other side with a normal EPI objective (even low NA and/or LWD - like a Mitty). The numerical aperture of the image coming out of the hemisphere is the same as the RI of the material it's made of - 1.5 for glass, 1.76 for sapphire etc.

Those words sound overly optimistic. My understanding is that the NA of the normal objective is only multiplied by the RI of the hemispheric lens. If that's correct, then I think you'll have trouble exceeding by very much the NA of an ordinary purpose-designed oil immersion lens.

Have I misunderstood something?

--Rik

Added ref: https://www.researchgate.net/publication/228364651_A_case_study_for_optics_The_solid_immersion_microscope

More likely I have. But the "supersphere" mentioned above saves the day if I did. Lot of the info came from here https://www.researchgate.net/publicatio ... microscope

[rjlittlefield]

Added ref: https://www.researchgate.net/publication/228364651_A_case_study_for_optics_The_solid_immersion_microscope

More likely I have. But the "supersphere" mentioned above saves the day if I did. Lot of the info came from here https://www.researchgate.net/publicatio ... microscope

I think we're quoting the same reference there. I just formatted my URL tags so as show the whole url instead of ellipsis form.

I struggled a little with the terminology. I believe that paper uses the term "aplanatic focusing lens" for what you're calling "supersphere" and https://en.wikipedia.org/wiki/Solid_immersion_lens calls "superhemispherical".

In any case I agree, that's an attractive approach. The demo in Figure 2 of the paper is convincing, using a silicon lens to image buried features in a silicon integrated circuit at NA 3.3. The paper notes that "The increases in resolution and light collection are the primary reasons why these lenses have been used in commercial instruments for integrated circuit failure analysis".

--Rik

[Chris S.]

Next, I considered making a dedicated, single-objective frankenscope, purely for photographing diatoms in bright-field (keeping the ICM for everything else). Objective on a tube lens, old focus mech etc. But when you consider the physical support needs for "convenient" focus-stack shooting, and the precision needed at such high resolution, it has to be a proper microscope stand really. I do use DIC with mix-matched sliders for diatoms at the highest magnifications on my ICM - I know I'd eventually wish I had that too. So not an attractive route really. Far too many compromises.

Emphasis mine--have italicized the bits I question.

This is the route I'd take. I don't think the precision required--even if one adds DIC--is beyond the capability of a dedicated DIY-person.

Building the thing wouldn't be trivial, but sounds like fun. And the aspect of making it diatom-specific might reduce complexity, plus perhaps be an opportunity to add capability.

At first blush, not sure I'd build it for single-objective use, as a low-powered finder lens or two would make aiming easier.

As a rough first thought (really a "think-aloud"), I'd consider making a vertical rig with a heavy steel base and a Velmex BiSlide for the upright (I'm myself doing this now). Mount a microscope stage on the BiSlide, with the camera assembly on that. Condenser mounting is an interesting question--maybe put it on a second, shorter BiSlide? Or as a distant second thought, could the condenser perhaps be mounted on a second carriage on a single BiSlide?

Fun stuff to think about.

--Chris S.