Darned difficult diatom dotting, decisively done!

[Beatsy]

I've been building a UV-capable photomicrogaphy setup based on an Olympus BX61 fluorescence microscope - to increase available resolution when photographing diatoms. The project is ongoing with a few Heath-Robinson aspects left to improve, but serious attempts at first-light images became viable a couple of days ago. So I had a go with Amphipleura pellucida, a well-known tough species to dot. Note: "dotting" means resolving closely spaced rows of punctae (pores) in the frustules (shells).

I was waiting for a UV-pass filter to arrive so had to compromise by using DPX (RI 1.52) as the mountant. My preferred mountants, Zrax (RI 1.7) or Speedax (RI 1.72), both fluoresce under 365nm UV light and wash out the UV image (which a UV-pass filter fixes by blocking the fluorescence). These higher-RI mountants drastically improve contrast too, but DPX worked surprisingly well...

Cropped single image (no stacking) using a Sony A7riv, full spectrum conversion, in APS-C crop mode with 4-shot pixel-shift. Direct projection onto the sensor (C-mount sized circle). Heine condenser and Oly UPlanSApo 100/1.4, both oiled to the slide. A Convoy UV torch with a 365nm bandpass filter provided illumination. The Heine was at a dark field setting but that doesn't work at NA1.4 so it was providing circular-oblique illumination instead - sort of. Here's a deeper crop into the first image to better show the dots.

Considering the compromises, I was extremely pleased with this outcome! But while I was editing these pics the UV-pass filter arrived. So I had to have a go with that using a different A. pellucida mounted in zrax. Everything else the same as above, except I used 16-shot pixel-shift mode to put 4x more megapixels over the same field of view. OMG!

This is a 100% pixel-for-pixel crop from the resultant image. Slightly contrast stretched, but no sharpening or other editing. I can't recall ever using a 1 micron scale bar before. Wow!

An extended happy dance with lots of woohoos ensued. I didn't expect this much improvement over my past efforts. I'm not breaking any new ground as this has all been done before, but not by me. So now it's back to completing the project, with enormously enhanced enthusiasm.

Thanks for looking.

[dy5]

Wow! Spectacular success. Congratulations! Looks like your resolution is 300nm or less. Isn't that getting close to the theoretical limit?

Cheers, David

[Beatsy]

Thanks.

Resolution is way better than 300nm. The rows of pores (striae) are about 250-260nm apart and the pores themselves are spaced at roughly 220-230nm.

Using the formula 0.61*wavelength/NA for resolution, theoretical max is 0.61*365/1.4 = 159nm. I still have work to do on the illumination path to tidy up edges and remove double reflections etc, so I expect a little more improvement - in looks if not in actual resolving power. But theoretical max is rarely achievable (and I think the Heine might be limiting illumination NA a bit below 1.4) so I'll be dead chuffed if get a nice clear 180nm res and I might be getting that already. Not formally tested yet - not sure how I will either.

[jmc]

Very nicely done. If your oil is a source of fluorescence, Thorlabs currently have Cargille very low fluorescence oil for what looks to be a pretty good price (although keep in mind shipping costs with Thorlabs can soon offset that, so best to buy as part of a larger order). Here's the link - https://www.thorlabs.com/thorproduct.cf ... er=OILCL30

[Smokedaddy]

Wow, excellent results. I sure wish I had the expertise to do that.

[Lou Jost]

This is a great rabbit hole that I am also entering. Nice to see these good results. I'll be closely following your progress since you are way ahead of me....

You probably know the Ultraviolet Photography forum. Worth a look. In particular Jonathan Crowther's posts about his homebuilt deep UV microscope is interesting.
https://www.ultravioletphotography.com/ ... /index.php

He also posts on his own website, for example here:
https://jmcscientificconsulting.com/uv- ... ium-light/

I bought a couple of inexpensive ancient Zeiss quartz UV objectives optimized for 257nm....I have a long way to go before I can figure out how to use them...

[Beatsy]

Thanks Jonathan. I saved the Cargille link in case I want to tinker with the scope's (significant) fluorescence capabilities some day, but I only need the 365nm output and block all the other wavelengths (for now). The oil can fluoresce flashing disco lights for all I care, it won't get through

Smokedaddy: IMO, once you get your regular photomicrography chops up to scratch (which I believe you have) the main barrier to using UV is having or getting the kit. Expertise, not so much - and you can develop that "on the job". UV imaging has it's foibles of course, some quite tricky, but what doesn't?

Lou: Actually, it was Jonathan that "planted the seed" and got me thinking about UV (I've read most of his blog, but not visited the forum - thanks for the link). However, my main motivation was solely directed at resolving diatoms better when photographing them, and remains so, for now. But having novel, underutilised gear on hand does tend to lead one down other avenues, so I'm sure I'll "digress" before long.

Cheers

[Lou Jost]

I see now that Jonathan is the one who responded to you just before my post!!!

[Scarodactyl]

It is an impressive setup and some great resolution. I'll admit I'm curious what the bottleneck is on the camera that it's projecting a c-mount sized circle--did you also have to use a more transmissive tube lens?

[Beatsy]

The scope came with a C-mount adapter fitted to the trinocular. Also, a peltier-cooled C-mount camera, but no cables, s/w or manual. Since I already had a C-Mount to Sony adapter I just used that with my converted A7riv "to start with". I'd need a different Olympus camera mount and a projection eyepiece to fill full frame. Those are hard to find and the current setup clearly works, so I'm "managing" with that for now. I don't know what the effect would be on UV absorption either. Image scale is a bit small, approaching under-sampled (using 16-shot pixel shift), but I think full frame projection would be significantly more over-sampled - so I suspect this is the permanent arrangement really. If a different adapter turns up at a sensible price, I might give it a go - but I'm not holding my breath. BX61 parts aren't often offered cheaply.

[75RR]

Congratulations! Wonderful to see such resolution. Nice clean and clear image as well

[Scarodactyl]

An olympus head with an empty adapter will cover aps-c (with some minor corner degradarion if you don't have an uw head) as long as there isn't something like a c mount physically blocking the light.

[Beatsy]

Hmm, I think what I have is probably better. The projected image circle exactly matches the FoV through the eyepieces. In APS-C crop mode the left and right edges of the circle touch the sides of the crop frame but top and bottom are clipped. If I use full-frame with the 1:1 aspect ratio setting, the image circle fits the square exactly, almost like it was made for it. This probably belongs in equipment discussions, but here's a phone snap of the adapters it came fitted with.

[Smokedaddy]

Smokedaddy: IMO, once you get your regular photomicrography chops up to scratch (which I believe you have) the main barrier to using UV is having or getting the kit. Expertise, not so much - and you can develop that "on the job". UV imaging has it's foibles of course, some quite tricky, but what doesn't?

My biggest concern would be buying the proper parts to be able to image at 365nm and not wasting money collecting parts that I don't need. I don't mind going down another rabbit hole but don't want it to turn into another experience like I am having with the Leica DMIRB/E.

-JW:

[Scarodactyl]

Hmm, I think what I have is probably better.

If you always want to crop the image down it's a nonissue, just seems like a shame to discard a large chunk of your available FoV.