Visible, UVA and UVB transmission microscopy of a diatom

[JMC]

Last year I shared a post about some initial attempts at imaging a diatom using the UV microscope I've been building - https://www.ultravioletphotography.com/content/index.php?/topic/4883-uv-365nm-and-visible-546nm-microscopy-of-a-diatom/

 

With that initial test I used a commercial test slide from Diatoms Labs which is a nice slide, but is made from glass and the mountant used quite strongly absorbs the UV. As such the test slide is not usable in the UVB, so was of limited use for me.

 

Before Christmas I started asking around for someone who could make me a test slide using the quartz slides and coverslips I have, and with a mountant that would let UVB through. Quite quickly I tracked down a Diatom mounter who was intrigued by what I was after and offered to help.

 

After a lot of experimenting on his behalf (for which I will be forever grateful) the first test side arrived today, and I wanted to share some very preliminary images. This was of one of the diatoms on the slide.

 

These were done with with my UV transmission microscope, and with a 32x NA0.4 Zeiss Ultrafluar objective. Images captured at 546nm (visible), 365nm and 313nm using a monochrome converted Nikon d800 camera. Images processed similarly, and collected as stacks in Zerene. They have been reduced in resolution for sharing here.

 

First, at 546nm.

 

Next, at 365nm.

 

And finally, at 313nm.

 

As expected the resolution improves as the wavelength decreases.

 

Also the 313nm image needed loads more retouching - the shorter wavelength improves the resolution, but obviously makes all the imperfections and dirt in the image more visible as well. Hence more retouching. 313nm with my 100x NA0.85 Ultrafluar will be 'fun', or should I say, 'character building'.

 

Funnily enough the chap who made the slide for me wasn't happy with it at all, as the gelatin he used to mount the diatoms is too visible. He says he will try again with other techniques (what a nice man). I was just happy to have something that was transparent in the UVB.

 

[nfoto]

Theory is one thing, seeing it manifested in this way is amazing.

[Fandyus]

I made a trichrome from your images.

[Andy Perrin]

WOW, this is some amazing stuff, Jonathan. I am full of envy. The diffraction effect is certainly dramatic.

[Stefano]

The effect is clearly visible. I agree with Birna, experiencing science in the real world is much better than simple theory.

[photoni]

Jonathan ... the improvement in quality is amazing ... compared to the first two images

[JMC]

Cheers all.

 

Fandyus - which colour is which image in your trichrome?

 

I was surprised to see such a big jump in resolution between 365nm and 313nm. While there was improvement between 546nm and 365nm, it didn't look as obvious. Using Abbe's equation, the theoretical resolution with that objective is;

 

546nm light - 683nm resolution

365nm light - 456nm resolution

313nm light - 391nm resolution

 

However due to the other optical components in there I doubt I am getting close to that.

[Stefano]

If resolution improved with shorter wavelengths, does it mean that you are diffraction-limited? If so, I think you are close to the theoretical limit.

[JMC]

I don't claim to be an expert on that Stefano - microscopy along with photography is still something I am learning about every day. But any calculations of theoretical limits don't take into account other components in the optical setup. I'm using a mix of components from different manufacturers (and even some I have made myself) which will all have an impact of the actual resolution I can reach.  Abbe's theory states that as wavelength gets shorter, then for a given NA value of the objective the resolution will improve, which is what I am seeing in practice.

 

 

[Andrea B.]

Altogether a lovely (very cool!!) experiment, Jonathan. Thank you for posting this. As Birna noted, seeing the actual display of what we know theoretically is amazing indeed. 

 

I see this in every botanical series I make in UV, Vis and IR. I sometimes get frustrated with my Visible botanical photos because I know so much detail is "missing" as compared to the accompanying UV version. It would be fun to try two narrowband UV-pass filters also on a flower. 

[Andy Perrin]

Stefano, Jonathan: if the other effects were blurring the image more, I don't think you'd be able to see the diffraction effects so distinctly. Despite the homemade setup, I do think you're either diffraction limited or near to it.

[colinbm]

4 hours ago, Andrea B. said:

Altogether a lovely (very cool!!) experiment, Jonathan. Thank you for posting this. As Birna noted, seeing the actual display of what we know theoretically is amazing indeed. 

 

I see this in every botanical series I make in UV, Vis and IR. I sometimes get frustrated with my Visible botanical photos because I know so much detail is "missing" as compared to the accompanying UV version. It would be fun to try two narrowband UV-pass filters also on a flower. 

I thought the differences in clarity in IR & UV was the penetration of the IR to below the surface with organic subjects & not with UV ?

[Andy Perrin]

21 minutes ago, colinbm said:

I thought the differences in clarity in IR & UV was the penetration of the IR to below the surface with organic subjects & not with UV ?

I'm sure that doesn't help either, but a lot of it is also diffraction. The subsurface scattering is what makes skin look waxy in IR, though.

[Fandyus]

On 3/4/2022 at 10:13 AM, JMC said:

Cheers all.

 

Fandyus - which colour is which image in your trichrome?

 

I was surprised to see such a big jump in resolution between 365nm and 313nm. While there was improvement between 546nm and 365nm, it didn't look as obvious. Using Abbe's equation, the theoretical resolution with that objective is;

 

546nm light - 683nm resolution

365nm light - 456nm resolution

313nm light - 391nm resolution

 

However due to the other optical components in there I doubt I am getting close to that.

If I didn't make a mistake, it should go from red to blue as the wavelength decreases.

546nm red

365nm green

313nm blue

[Cadmium]

Jonathan, Very well done work. Great you found someone to make you what you needed. Thanks!

[JMC]

Cheers all.

 

Interesting Andy. I do still think think certain aspects of my setup are limiting me. I have yet to properly control vibration, both from the camera and just the general vibration from the ground (as I live next to a busy road). I've tried using vibration isolation feet but they actually compounded the problem, so I need to think more about how to deal with it all. However as you say, maybe the other optical components aren't so much of an issue. I am in the process of trying to find a suitable, calibrated resolution test target, that I can use in the UVB. I think I have something sorted, but have a few remaining questions before taking the plunge and buying it (as it costs nearly 1000GBP and I want to be certain it'll be useful before committing to that).

 

Thanks Fandyus for explaining the colours.

[Andy Perrin]

Jonathan, it could be that diffraction and the other effects are of similar size. In that case improving vibration or the other components will help the image, but you will continue to see the trend in the wavelengths. What I don't think is happening is that the other effects are much larger than the diffraction, or you would not see a difference at different wavelengths (unless all the other effects ALSO get better with wavelength reduction for totally different reasons).

[JMC]

Yes Andy, makes sense.

[dabateman]

This is great Jonathan.

Yes you might be able to optimize the cone of light entering your objective lens.  But that may not improve the image as much as just locking everything down. 

For stabilization there are typically two methods:

1. Place everything on a 6 inch slab of marble. 

2. Get an aluminum air table that floats the table on a cushion of air.

Method 2 is the best, but really expensive. But really cool to play with. My Phd 2-photon microscope was on an air table and it was amazing. This on a major street in downtown Toronto with streetcars going by.

[JMC]

Thanks David. Serious options like you mentioned are not an option at the moment - small place and a multifunctional space for the microscope. The good thing is I think the effects of movement are small contribution to the overall image. Also when the astro camera is eventually modified for UV, I'll be able to use that as well which should help.

[Adrian]

Great images  Jonathan! Very cool!

[JMC]

Cheers Adrian.

 

Bit of an update, but using a 6mm Zeiss Monochromat quartz lens from about a 100 years ago - https://jmcscientificconsulting.com/uv-microscopy-6mm-zeiss-monochromat-imaging-of-a-diatom/

[Andy Perrin]

That’s a beautiful lens Jonathan. 
 

will you be able to test it at its design wavelength?

[JMC]

25 minutes ago, Andy Perrin said:

That’s a beautiful lens Jonathan. 
 

will you be able to test it at its design wavelength?

I wish Andy. At the moment I have no light source that'll give me 275nm (at least not much of it), nor do I have any optical filters that will block everything else. Camera sensitivity will be questionable there as well. Testing it at its design wavelength will have to wait for now....

[JMC]

Geeky update today. I was looking at my diatom slide at 313nm with some of my objectives - Zeiss Ultrafluar 32x NA 0.4 (using glycerine immersion fluid), Leitz UV 40x NA 0.65 (again with glycerine immersion) and an old Beck Reflecting objective 36x NA 0.5 (no immersion). The objectives are shown below (32x Zeiss Ultrafluar, 40x Leitz, 36x Beck, from left to right).

 

I was using the 32x Ultrafluar and the 40x Leitz ones yesterday without immersion fluid (the 32x Ultrafluar can be used without) and was a bit disappointed with the image quality I see seeing with the Leitz one. This prompted me to try both with glycerine today. The quality of the image from the Leitz one when used with the glycerine immersion fluid blew me away.  This is what the image with the Leitz 40x looked like (full frame, no cropping, and from a stack of 13 images, and reduced in resolution for sharing on here);

 

Taking a crop from the original one which wasn't reduced in resolution, I put it into Image J and got the following;

 

That little yellow line is 449nm long. In theory the resolution with that lens at 313nm should be 240nm according to Abbe's equation, and it certainly looks like it could be capable of that.

 

I thought the best UV objectives I had were the Zeiss Ultrafuar ones. However it looks like the Leitz 40x UV is a bit of special one, so must try that again for other samples.