Microscopy – resolution comparison between visible and UV light

I built my UV microscope to help with my sunscreen research, but in doing so I quickly realized that UV offered the chance of improved resolution due to the shorter wavelength of light being used. While not so important for looking at sunscreen emulsions, this has proved very interesting for examining diatoms and looking at their structures. But how much of an improvement does it actually provide, especially when compared with more modern optics, and different ways of lighting a sample? This will probably be my last post of 2022, and I’ll show a comparison of a diatom imaged using UV at 313nm and with visible light (450nm) but with a different setup. It’s not quite a ‘apples with apples’ comparison, but as close as I can get. Sort of ‘apples with oranges’.

The subject will be a Surirella gemma diatom as this has some really fine structures. Here it is imaged with 313nm UV light with a 100x NA 1.2 objective.

313nm UV image of Surirella gemma diatom

And now an image taken using 450nm light with an NA 1.4 objective.

450nm light image of Surirella gemma diatom

As you’ll have no doubt noticed the images have a different field of view. So a better comparison is to look at a crop from each image (both have been cropped and then resized to 1200 pixels across). First the 313nm light image.

Crop of the 313nm light image

And second a crop from the visible light image done at 450nm.

Crop of the visible light image taken using 450nm light

Both images seem to offer similar resolution, although the 313nm image is a little noisier. There were many differences between these images, including objective used, condenser, camera, photoeyepiece, the 450nm image was stacked while the 313nm image was not, the slides are different as are the mounts for the diatoms. As I said, ‘apples with oranges’ rather than a true ‘apples with apples’. These differences occur as optics optimized for use in the UV are not ideal for visible light, and visible light optics wont work at 313nm. However the aim with this is to compare a visible light with my best imaging conditions with UV light with a fairly basic imaging setup. The comparison between the two imaging setups is show below.

Comparison between the setups at each wavelength

With the 450nm light I was able to use a higher NA objective and condenser compared with 313nm light. I used oblique light at 450nm, which helps resolve small features, and it was a stack of 9 images. The slide mount was also different – a nice high refractive index one of piperine, vs a dry mount in Debe’s for the UV image.

The interesting thing though is when you look at the theoretical resolution based on the wavelength of light used and the NA’s of the condenser and objective. They are very similar, less then 10% apart, with the UV objective being only slightly better. This explains why the images look pretty similar in terms of how they are resolving the small features. ‘But’, I hear you say, ‘you said that UV offers better resolution, so what is going on?’. The NA’s of the objective and condenser for the UV image were lower than the ones used for the visible light image, and this is offsetting the effect from the shorter wavelength. Visible light images can absolutely provide amazingly high resolution images of diatoms.

So, is there no point to UV imaging? Well, it’s not quite a simple as that. The NA of the condenser I used for the UV image was only 0.85. I do have a NA 1.25 one as well, I just didn’t use it as until recently I didn’t have a suitable mount for it. If I used that at 313nm, with the NA 1.2 objective, the theoretical resolution then goes down to 128nm. If I used the NA 1.2 objective and NA 1.25 condenser with 254nm light, the theoretical resolution improves again to 104nm.

Of course there are various other technique to improve visible light images, such as differential interference contrast, cross polarization and phase contrast which I haven’t looked at here. As a result modern techniques can certainly provide very very high resolutions, and the advent of these techniques was a contributing factor to the decline in the use of UV imaging for microscopy. However UV can offer high resolutions in a very simple optical setup – no need or DIC, cross polarization – and without complex image process. UV can offer these high resolutions with simple brightfield microscopy, and this is why it still has a place in imaging today.

As always, I hope you enjoyed this post, and thanks for being part of my research journey for 2022. Here’s to a better 2023, and if you’d like to know more about this or any aspect of my work I can be reached here.

EDIT 26/3/23. The slide used for the 450nm light here is labelled ‘piperine’. Piperine is an alkaloid found in black pepper, and it has always confused me as to why it would be used as a microscope slide mountant. However its use is mentioned in Microscope & Entomological Monthly, SH Meakin, The study of diatoms, 1939, Pages 209-212.

Mini UV camera – diatoms at 313nm including Amphipleura pellucida

Yesterday I had an exciting parcel arrive – a set of microscope slides that I have had made for me by a very nice gentleman who I met during an ebay purchase (Opticsman0127, check out his slides here). I sent him some quartz slides and fused silica coverslips and he made me a set of slides with different diatoms on them for my UV work. These have been mounted dry using a bit of Debe’s mountant (gelatin) so as not to have the mountant absorb too much UV. Today I am sharing some images from one of the slides – a diatom strew – taken using 313nm light.

For these images I used a mercury xenon lamp and filtered the light to image at 313nm. The camera was my new Matrix Vision one with the UV sensitive Sony IMX487 sensor (see here for some images done at 254nm with it). I decided to use 313nm here as the light source I have which covers that wavelength is more powerful than my 254nm lamp, and I wanted to try a higher NA objective – a 100x Leitz NA 1.2 objective (see here for some information on that). Condenser was my antique Zeiss quartz one (NA 0.85) and the photoeyepiece is a Lomo quartz 8x one. The images are not stacks, but just single images taken in brightfield. I noticed that at this magnification there was some left/right movement in the images which would have complicated stacking. The images have been reduced in resolution for sharing.

One of the really hard resolution tests for any microscope is the diatom called Amphipleura pellucida, as it has features of the order 200nm. This makes it extremely hard to image with normal light microscopy, and techniques such as cross polarized and annular oblique imaging are needed to see them even with objectives and condensers of NA 1.3 or more. I was trying this at 313nm with an NA 1.2 objective and 0.85 condenser (both using glycerine as immersion fluid) and normal brightfield, with a slide which just had a dry mount of gelatin. It’s a bit of a white whale for me as I have yet to get good images of it. Although I now have a slide I can do UV imaging with, this was always going to be fun.

First a set of 3 images done at different heights within the A. pellucida diatom. The field of view is so small that only part of the diatom can be seen and it is close to one end of it.

A. pellucida, focus position 1
A. pellucida, focus position 2
A. pellucida, focus position 3

In the first two images above, there are faint vertical lines which become more visible in the second image. These are I believe called striae. In the third image, these lines start to break up and form a dot pattern. These are the punctae in the structure coming into focus. Let’s go in closer to the third image and have a better look. This is a crop from the image above, taken before the image was resized for sharing.

Crop from the Image 3 of the A. pellucida

The regular dot pattern can be seen more clearly now. What is the spacing of these features? I put the image into ImageJ and did a couple of measurements. I have chosen to measure the distances between the dark spots, but it would be the same if I had chosen the white spots. First the distance between the striae.

Distance between the striae on A. pellucida (296nm)

Then the distance between the punctae on a given stria.

Distance between the punctae on A. pellucida (199nm)

From this it looks like the striae are about 300nm apart and the punctae about 200nm, which is in keeping with SEM data I’ve seen on this species of diatom. This shows why they are such a challenge for the light microscopist. Even with the 313nm light being used here, I’d expect the maximum resolution obtainable with my setup to be between 135nm and 150nm depending on which calculation I use, so this is getting close to that.

Have I finally caught my white whale? While it is lovely to see the features on it, I wont stop trying to get better images, so I do not consider it caught just yet.

For the second diatom, I moved around the strew and found what I think is a diatom from the Synedra family, although I don’t know which one (if I find out I will update this). At this point it is worth highlighting one of the challenges I have found with using this new camera. The image below shows what I can see through the eyepiece of the microscope and was taken with the camera phone, with the 100x objective.

Camera phone image through the eyepiece with the 100x objective

In the image above is a blue box. This is the approximate size of the image that the Matrix Vision camera sees when used with the 8x Lomo photoeyepiece and is about 16 microns square. As you can imagine it is a challenge even finding the subject with the camera as even the smallest touch on the stage causes it to disappear.

As before a series of single images showing part of the diatom taken at different focus heights.

Synedra diatom, focus position 1
Synedra diatom, focus position 2
Synedra diatom, focus position 3
Synedra diatom, focus position 4

As the diatom is moved up and the focus position varies, there are dramatic changes in the images, and the features which are visible. In the third image a series of bright dots appears in lines at 90 degrees to the diatoms central axis. As we go further, the dots reveal themselves to be what look like pits in the structure. Presumably these are holes in the structure, and with the focus at one position the light gets through, and when it is moved it cannot. Perhaps this is an interference effect given the wavelength being used and the size of the features?

As with the A. pellucida image, I took a crop of one of the images and put it into ImageJ to get some measurements on the distance between the dots.

Distance between the dots (302nm)

On this diatom the dots look at be about 300nm apart.

Where has this new work taken me? The Matrix Vision camera continues to impress me, and live view at 313nm enabled me to focus down to fractions of a micron to get the images (the camera is so sensitive, I was able to do live view imaging at about 1s exposure to get the focus right). The field of view is tiny though, so I need to go back and re-evaluate some of the my other quartz photoeyepieces to see if I can find something which presents a wider image to the camera. At the moment the pixel resolution with this setup would be about 178 pixels per micron (5.6nm per pixel). Given the max resolution I would expect is about 130nm, that means I am wasting a lot of the resolution. A photoeyepiece which gives a wider field of view would help here, giving me both a larger area being imaged, while hopefully not compromising actual resolution. More work needed here in the future. 254nm illumination should be even more impressive than 313nm, but my current homemade light source is not powerful enough to use with this objective so that will be another job for 2023.

Before I go, a quick photo of the slide itself.

The slide itself….

I look forward to spending plenty of time on this and the other slides in 2023. As always, thanks for reading, and if you’d like to know more about my work I can be reached here.

Mini UV camera – Initial diatom images using 254nm light

My new UV camera uses a Sony IMX487 sensor, and as I showed recently certainly offers improved UV sensitivity, and its behaviour at 280nm is very impressive (as I have shown here). With my spectral sensitivity testing though, I can only get down to 280nm. With the UV microscope I built, one of the goals was to use it for imaging at 254nm (see here for some early work on that), however at the time, my cameras had very little sensitivity that far into the UV and I was dealing with very long exposure times, often 10 minutes or longer. Another issue was that live view didn’t work so focusing was a nightmare. Today I will share some initial images taken with the new Matrix Vision camera with the Sony IMX487 UV sensor.

The slide imaged was my custom made diatom slide (made with quartz and fused silica instead of glass for the slide and coverslip) on my modified Olympus BHB microscope, and using the 254nm lamp I built. The objective was a 40x Leitz NA 0.65 UV objective with glycerine immersion and the condenser an antique Zeiss quartz one. The optical filter was a 254nm one from a forensics camera. The images are single shots (not stacks) and show parts of two diatoms on the slide. They are shown as full frame images and are not cropped, although I dropped the resolution of the images from 2848×2848 to 1600×1600 for sharing here. They have been sharpened slightly, and I played with the curves a bit as well.

First one, part of a Pleurosigma angulatum diatom.

And second one is part of another diatom (I am not sure what this one is).

The images are showing very high resolution – which is what would be expected given the short wavelength – especially given that the NA of the objective is only 0.65. There are some artifacts in the images (banding for instance) and I am not sure how much this is due to the camera settings, the light source itself or other optical components in the setup as I have seen similar effects before with another camera at this wavelength. I shall have to dig more into that. On the second image there are some very bright white points of light which are appearing in the structure of the diatom. I’ve not seen these before, but they seem to be ‘real’ features and not artifacts. This is what happens when you look at things with different wavelengths…..

With this new camera I was able to capture images in seconds rather than minutes (14s exposure and a gain of 20 were used for these images as opposed to 10-30mins and ISOs of 4000 and above) and I was able to focus using live view on the computer so it is definitely an improvement. Id like to be able to use lower gain for the images, but that means longer acquisition times than the software currently allows for, so is not a simple ask. The sensor is much smaller than I am used to with my SLR cameras (11mm x 11mm vs 36mm x 24mm) and finding the diatoms was a challenge even with this 40x objective. I dread to think how difficult it would be with a 100x one.

As far as I am aware these must be some of first images with this sensor (which aren’t from Sony) which are being shared. I hope to do more with this camera for imaging diatoms and also for my sunscreen work, and it continues to impress me with its capabilities. Before I go, a photo to compare the Matrix Vision camera with one of my SLRs (both with UV lenses on), just to show how small it really is.

Comparison of an SLR camera (on the left) with the Matrix Vision UV camera (on the left)

As always, thanks for reading and if you’d like to know more about this or my other work, I can be reached here.

Mini UV camera – Matrix Vision BlueCougar with the Sony IMX487 sensor, spectral sensitivity

Having received a Matrix Vision BlueCougar camera with a UV sensitive Sony IMX487 sensor (which I reported here) one of the first jobs is to understand just how sensitive it is, especially in the UV. To do this I ran it through a test with my spectral sensitivity measurement system (see here for details on that build). As comparisons I used two different cameras – a monochrome converted Nikon d850 with the Bayer filter and microlens array removed, and an astro camera (Altair 26M) which still has microlenses. Both of these have fused silica coverglasses and are optimized for my UV work. I normalized the output from each camera to the max value to more easily compare them, and here is what the comparison looks like.

Comparison of the IMX487 sensor with other monochrome cameras

And a more close up view of the UV region.

Comparison of the IMX487 sensor with other monochrome cameras closeup on the UV region

Here’s an image of the testing setup, on a suitably disorganized work bench.

Spectral sensitivity testing setup

Having seen a few different QE curves for this sensor, I wasn’t sure what to expect. This was why I needed to do the measurement myself, to allow for comparison with other meaningful cameras, and as of today (14th Dec 2022) as far as I am aware this is the only such comparison available. The Matrix Vision software for the camera was great for this test, as it allowed me to get the histogram data directly, without saving the files and then analyzing later, which saved me loads of time. Also I used binning to improve the overall sensitivity of the camera. As I am slightly limited by max exposure time (I used 5s for the Matrix Vision camera, vs 30s for the Nikon d850 and up to 4 mins for the astro camera) I did a 4×4 binning which reduced file size but drastically reduced the need to long acquisition times. This was kept the same for all wavelengths, so doesn’t change the shape of the curve and still allows for comparisons with other cameras.

After all that testing, what to make of this? I’m actually pretty amazed by how sensitive this camera is in the UV, especially below 350nm where it is dropping only very slowly. In fact for my microscopy at 313nm and 365nm, the sensitivity at these wavelengths is almost the same, while with my other cameras at 365nm they are about 4x as sensitive as they are at 313nm. This will make imaging at short wavelengths simpler, especially when combined with the live view capability of the Matrix Vision camera. While I cannot measure below 280nm with my setup, the sensor is supposed to offer good sensitivity down to 200nm, only dropping slightly between 300nm and 200nm, and this is where it really should shine when compared with the others. The only way I can test that though is to get it on the microscope and see how it behaves, and that is for another day. Sony have been pretty cagey about how they have made this UV sensitive. I have my suspicions, but more testing needed on that.

As always, thanks for reading, and if you’d like to know more about this or any other aspect of my work, I can be reached here.

Mini UV camera – Matrix Vision BlueCougar with the Sony IMX487 sensor

Good things come in small packages, and today’s arrival will hopefully fall into that description. With imaging in UV, sensor sensitivity is always a challenge, with the sensitivity getting worse as the wavelength gets shorter. With my UV microscopy, if I image down at 254nm, it is not unusual to have exposure times of 20 minutes or more, even with my monochrome converted cameras. What seems like an age ago now, Sony announced they were going to be making a new camera sensor which was specifically designed for imaging in the UV from 200nm to 400nm – the Sony IMX487 sensor. This is a small sensor, designed for C mount industrial cameras, but had an impressive 8Mp resolution, so it got me interested in the possibility of using it for my microscopy work on sunscreens and diatoms, or for other general UV imaging. I approached a couple of different vendors when they started announcing they were going to be making cameras with these sensors in and settled on Matrix Vision, who would be making it as part of the their Blue Cougar range (see here). The camera arrived today, so I thought I would share some image of it before the testing begins.

This thing is tiny. I thought it would be much smaller than my normal cameras, but I wasn’t quite ready for just how small it is (a diminutive 4x4x5cm and apparently they can make it smaller if needed). Here it is on a tripod with a 25mm f2.8 quartz lens attached (which is also dinky).

And a shot of an initial visible light image from the setup testing.

So far first impressions are good. The software is straightforward but comprehensive, and the camera is built very well. I did do some UV images at 365nm and 254nm, and it certainly gives images there, but I haven’t done anything systematic yet to see just how sensitive it is. The next step will be to do some spectral sensitivity measurements, which will give me the sensitivity from 280nm to 800nm. Then it’ll be time to get it on the microscope and see how it behaves. This should offer much more sensitivity than my monochrome converted cameras when imaging at 313nm and 254nm, without the drawbacks of the astro camera I tried (cooling fans causing vibration for instance). The ideal goal would be to be able to do videos in the UVA, UVB and even UVC if possible. While I have been able to get UVA and some UVB video files, the cameras I have had up until now just didn’t have the sensitivity for UVC video work.

I shall of course be updating with my findings on here, so check back in occasionally for new news. As always, thanks for reading and if you’d like to know more about this or any other aspect of my work, I can be reached here.

Royal Photographic Society – Good Picture 2022 Meeting

When I get the chance I make the time to give talks about the research that I do, and this last weekend I spoke at the Royal Photographic Societies Imaging Science Group meeting – Good Picture 2022.

The meeting brought together those of us with a passion for imaging, from widely different backgrounds, for a day of talks, and was held at the University of Westminster in London.

I gave a talk on the building of my UV microscope, and its use for looking at sunscreen emulsion structure and the imaging of the fine structure present in diatoms.

My time at the podium…..

Involvement with organizations such as this helps drive the science forward and is something I recommend to anyone at any stage in their career. Face to face meetings have been hard, if not impossible, for the last few years, but things are starting to ease a little now, and it has been great to get back and see people again.

Charity donation – Wildlife Aid Foundation

As part of running my company, I think it is important to give something back to the wider world. Wildlife is a bit of a passion of mine, and I often spend time photographing the birds in our garden. For a long time now I’ve been a volunteer with various wildlife related projects and organizations such as Surrey Wildlife Trust and the Woking Biodiversity Partnership, either helping out in person or through donations from my company. I also donate to charities which look after injured wildlife as humanity has a huge, and often detrimental, impact on the environment. I wanted to share a recent story as it highlights one of the reasons I do what I do.

A couple of months ago I went out into my front garden one sunday morning to go the bins. It was raining heavily, and as I came back towards the house I saw a bird under my car. Initially I thought it was a Sparrowhawk, but it turned out to be a Kestrel, a beautiful little bird of prey. I went towards it and it hopped off, unable to fly. My guess is that it was hit by a car as we live on a busy road. I went back inside and got a box and a blanket and went back out, spending about 5 or 10 mins gradually getting closer to the bird, and was eventually able to pick it up and put it in the box. I wouldn’t necessarily advise doing this – sharp talons and beak – but I had thick gardening gloves on.

I phoned the Wildlife Aid Foundation who are based about 15 miles away from me and they said ‘yes, bring it down to us’. I quick drive later and it was with them. It turns out that it had had a head injury, but they were hopeful of its chances. A few weeks later I was called up to say it had healed and was ready to be released and if I wanted to come and collect it so I could release it where I had found it.

Back I went and brought the Kestrel home and duly released it in the back garden (away from the traffic) where it was glad of its freedom – it was off as soon as the box lid was opened. Although I managed to grab a couple of quick photos of the release, it was all over very quickly and off it went.

The Kestrel, watching me as I got ready to open to box to let it go free
After release it hopped around the decking a bit before flying off

The Wildlife Aid Foundation is one of the organizations I donate to when I can, however this very personal experience really brought it home to me as to why these places need to keep going and carry on doing the work they do. We all have huge financial pressures at the moment with everything that is going on in the world. If you can though, remember the wildlife when it comes to giving something to charity. Research has shown the beneficial effects spending time in nature has on our well-being, and with the rate at which we are destroying it, anything to help is a good investment in our future.

Microscopy of gold coated diatoms

Coating of diatoms in a thin layer of metal or metal oxide has been reported for many years as a way of improving their visibility for microscopy (as examples, work I have discussed and published on slides from Horace Dall and John Dale. However it was never a common technique and is not widely used today. Recently I was sent a few slides from a fellow microscopist which contained diatoms which had had a thin (a few nm) coating of gold applied to them before mounting.

Her is an example image from one of the slides. It shows part of a diatom (likely Triceratium grande, or Triceratium favus). The resolution has been reduced for sharing here.

Gold coated diatom sample

And the slide itself.

The slide with the diatoms mounted on it

Imaging was done on my modified Olympus BHB microscope with a 60x Olympus Splan Apo NA 1.4 objective, and oblique illumination from below using white LED light. The image is a stack of 10 images (stacked using Zerene stacker) and photographed with a Canon R7 camera. The image shown has not been cropped – this was the full field of view.

Gold coating of of the diatoms certainly improves the visibility of the features, and I am looking forward to looking at the other samples. Thanks for reading, and if you’d like to know more about my microscopy or other aspects of my work, I can be reached here.

Award and presentation with the Quekett Microscopical Club

On saturday 15th October 2022, I attended the Quekex meeting of the Quekett Microscopical Club (of which I am a member). I had been asked to give a talk on my UV microscopy work on sunscreen and diatom imaging, and as it turned out I had also been awarded a Certificate of Technical Merit for one of my UV microscope images of a diatom (taken using 313nm light). A very successful and fun day and here’s a few pictures from it.

The title page from my talk to the club.
My diatom image of a Pleurosigma Angulatum taking at 313nm on my UV microscope.
The awards notice from the day (I’ve blurred other people for the sake of confidentiality).

If you’d like to know more about my microscopy or other work, I can be reached here.

Bausch and Lomb 365nm microscope objectives and equipment

Given my ongoing research interest in UV microscopy, I am always looking out for information on the subject. A few days ago I found a brochure published by Bausch and Lomb which talked about objectives and other items they had made which were designed for use at 365nm. As I hadn’t heard about these before, and I haven’t found any more about these, I thought I would share it here.

The document was called “Ultra Violet Photomicrography at 3650Å Optics and Other Accessories”. Here’s the cover page.

3650Å is 3650 Angstroms and is the same at 365nm. These are glass objectives and are corrected so the focus point at 365nm is the same as that at 546nm, so that they can be focused in visible light and then used for imaging at 365nm. This was done to improve resolution, as a ‘half way house’ imaging in the UV with glass optics but not going to the extremes of using quartz optics and imaging down below 300nm.

They give an example of the comparison between a 546nm image and a 356nm one of a Amphipleura pellucida diatom, which shows the improvement nicely.

Here’s the information about the objectives.

Thing is, I had not heard of these before, nor have I ever seen any examples of them. If anyone has any, I’d love to see what they look like, and please feel to contact me here.