[UVC SAFETY WARNING] Microscopy at 254nm - build and images

[JMC]

First a UVC warning - UVC is dangerous. The author does not recommend trying to replicate these experiments without knowledge and use of the proper safety equipment and procedures.

 

This thread will cover information about an experiment I'm doing which is at an early stage. With my UV microscope I've been using it at 365nm and 313nm and it has been fine with those wavelengths. When I built it, I made sure components were used which would be usable down to 250nm, and perhaps even a little lower. However at that time, I didn't have a suitable light source, and my filtering and even camera choice made any experiments down at 254nm impractical to attempt.

 

A couple of weeks ago I was chatting with a couple of folks on here, to try and get some advice with regards to a UVC LED source I was thinking of buying. I was steered away from my original idea, and towards trying a 3W UVC lamp as it was potentially small enough to use with my microscope. Thanks to Andy and David for taking the time to discuss things with me. The aim of this thread is to provide a bit of an update to what came of those discussions, and share some very early results of UVC microscopy at 254nm.

 

The light source is a 3W low pressure mercury lamp source from China over ebay, cost about 8USD. I ended up mounting this in a spare Olympus lamp holder I had, after removing the internal glass lens and IR blocking filters. As result the 3W lamp was used as is, no collimating, no filtering, no anything. The camera I used was a MaxMax monchrome converted Nikon d850 with a fused silica window. I have used this before for UVC photography so know it is sensitive that far down. Filter was a 254nm one from a Sirchie forensics cameras. Again I have used this before for UVC photography.

 

Some images of the setup etc. First, the lamp spectrum, at a distance of 10cm from the side of the lamp.

 

The microscope with the lamp in place and switched on.

 

And a first sample image using the diatom slide (fused silica/quartz for the slide and coverslip) using a 10x Zeiss Ultrafluar objective.

 

I also got an image with a Schott WG305 2mm thick filter in place, the idea being to let through everything which isn't UVC (i.e. leakage from the 254nm filter). Looking at the RAW files for both images, I reckon about 90% of the image above is from the UVC and 10% of contribution of other wavelengths. This just goes to show how insensitive the camera is down there.

 

This is very, very early days, and the method presents some very extreme challenges (in addition to safety). I need a way of focusing and collimating the light source better, as I am losing a lot at the moment. I need another 254nm filter to trying and better isolate that region. As with my 313nm work, I'd probably stack them together for better blocking. Live view focusing is currently not possible with the camera, as there is just so little sensitivity, so focusing is guesswork and trial and error. The image above was 30s at ISO1000, so it is not a fast process. I've got some 265nm LEDs on the way, so will check against the 3W low pressure mercury lamp to see how much light they are producing - maybe they will be better. Personally, I doubt they will be better, but only testing will tell.

 

This is a bit of an unfunded side project for me, more scientific curiosity to see if it can be done than anything else (after all, the initial UV microscopy pioneers back in the late 1900s and early 20th century were doing this with plate cameras), and will update as and when I have new data/images to share.

 

And again - just to emphasize, this is UVC work, do not try this at home.......

[Andrea B.]

It is nice to see UVP's inventive minds at work on problems of photographing the more extreme end of the UV range in spite of its dangers. Thank you for writing about this. I always enjoy seeing what you are up to. 

[colinbm]

This is great to see Johnathan, beaut results.
Have you thought about Far UVC Excimer 222nm lights, they are not harmful to humans or animals, if they are filtered to only pass 222nm.

[JMC]

Thanks, am happy to share. Colin, 222nm may be an option in future but not at the moment. Issues with 222nm for me is getting one I can focus down to a small spot size, filters for the camera (to filter out any fluorescence along the way) and camera sensitivity. Also, that far down, I'll start to see other losses in the system. I have no idea whether my mirror (underneath the field lens) is still working below 250nm, and the higher magnification objectives start to drop in transmission below 250nm.  All of this makes 222nm even more of a challenge than 254nm.

[JMC]

As a quick follow up, the '265nm' LEDs actually arrived yesterday evening. However despite being sold as 265nm, and having 265nm written on them, they are not 265nm, but nearer to 285nm, and they are much dimmer than the 3W 254nm lamp. Comparison at the same distance from the lamp shown below.

 

I really must stop buying cheap LEDs from ebay - about half of what I have received over the last few years has not matched specifications. I have fed this back to the supplier, and got a hilarious response this morning, "we expect to see +/- 5% on the wavelength compared to what is specified", and of course this is not mentioned in the advert anywhere. Not +/- 5nm, but +/-5%. Caveat emptor. On the plus side the LED was a fairly pure source with no obvious peaks in the visible (although they could have been down in the noise as it was heard to pick up the light at all). But it is just so low in power that I am not going to spend any more time with it for now, and will concentrate on the 254nm lamp and trying to make that better.

 

[colinbm]

UV LEDs & China are not a reliable association unfortunately.
I thought I had a reliable supplier for over a year, then one big spend by me went belly up, but after resorting to Ali's big stick it was righted, but I don't feel like taking any more chances.

You really have to 'know' what you are buying & have the equipment to test & check it, thankfully that I had.

[JMC]

3 minutes ago, colinbm said:

UV LEDs & China are not a reliable association unfortunately.
I thought I had a reliable supplier for over a year, then one big spend by me went belly up, but after resorting to Ali's big stick it was righted, but I don't feel like taking any more chances.

You really have to 'know' what you are buying & have the equipment to test & check it, thankfully that I had.

Indeed - having the means to check what you're getting is important with these suppliers. Scary thing is you have to check everything. I had a batch of 10x 365nm LEDs a couple of years ago. Despite them all being connected together ready to snap off into individual ones, 2 out of the 10 were not 365nm. It is just a lottery.

[colinbm]

Jonathan can you tell me more about this 3W low pressure mercury lamp & its power supply, & the 254nm filter from Sirchie forensics, I can't find either Please ?

 

[JMC]

Hi Colin,

 

The lamp was a 3W UVC sterilization lamp from ebay. This is the one I got - https://www.ebay.co.uk/itm/325200769479 - but there seem to be loads of them for sale on there. I then just bought an E17 base to mount it and give me somewhere to attach the wires. They claim to be 10V, but don't say whether that is 10V DC or AC. I ended up just using a DC power supply, connected the positive to the middle electrode and the negative to the outside electrode and gradually turned up the voltage and current limiter. In the end it took about 11V and 350mW to get it to fire and run stable.

 

The filter is a little more complex. A while back a few of us bought Sirchie forensics cameras and lenses off ebay - https://www.ultravioletphotography.com/content/index.php?/topic/3402-60mm-f35-c-mount-uv-lenses-on-ebay/. The filter came with that, and I don't think they'd sell them separately. Best option might be to contact Omega filters and see if they have something suitable. It had very good blocking of the out of band regions, but only about 20% transmission at 254nm. The low max transmission is not uncommon with these filters unfortunately.

 

[colinbm]

Thanks Johnathan
I have bought a few of these 3W UVC sterilization lamp from ebay, down under but they have all been faulty & I got refunded.
I had tried 10v DC too. Most were faulty with loose wires inside the metal base shorting out. Perhaps I could try again, at least I now know it is not any special driver ?
This filter just squeezes in 254nm....
https://www.asahi-spectra.com/opticalfilters/detail.asp?key=ZUV0325

[JMC]

Colin,

I needed to go slightly above 10V to get mine working, but it was definitely a DC powersupply I used. Mine had voltage and current control, but is just a normal benchtop powersupply. Perhaps I got lucky this time with these lamps (karma for the LEDs).

The Ashai one doesn't offer blocking above about 750nm, although it does have good UV transmission, and these low pressure mercury lamps will emit light in that 700-800nm region. Even a small amount there getting through could well wipe out the UVC due to the camera being much more sensitive in the 700-800nm region. I think mine is similar to the Edmund one as that offers blocking over a wider range, and the transmission is about the same - https://www.edmundoptics.co.uk/p/2537nm-25nm-fwhm-25mm-dia-first-surface-uv-bandpass-filter/40733/

As I say though, Omega might be able to offer something for less.

[colinbm]

I just found this extra info on the 10v 3w UVC bulbs....
https://www.ebay.com.au/itm/183793074264?_trkparms=amclksrc%3DITM%26aid%3D1110006%26algo%3DHOMESPLICE.SIM%26ao%3D2%26asc%3D240036%26meid%3Df4cb7e63cfb3426e971329b9a9ca1497%26pid%3D101195%26rk%3D2%26rkt%3D12%26sd%3D183793081755%26itm%3D183793074264%26pmt%3D1%26noa%3D0%26pg%3D2544595%26algv%3DSimplAMLv9PairwiseUnbiasedWebWithSearchFilterPromotion%26brand%3DUnbranded&_trksid=p2544595.c101195.m1851&amdata=cksum%3A183793074264f4cb7e63cfb3426e971329b9a9ca1497|enc%3AAQAHAAABEDNdRuflTEDzBfTKX%2BqmXH8pf%2B8V9puPczI5H7qtyLlQg%2FPd3LbQ9nGE%2BOPvEM%2FBknx3ncDmquETDGzeBRhRT3AMwWgNjXcX%2BhzRJh61%2BijcXj%2BMx9djGxlgvJ%2Fml0fHH%2F7%2BSUoS8TW6C3RRcgSZHTxGsOMHaakqHNEmlfqtpw964QAYbuKU0T6TBUIfCG0mj%2BJ2uOxUQGo2KO0ISaDP4ZilVNvC8UmhA2CRLIujzZ26njA8e0%2FF%2F%2BeYKA%2F1OLYKWTvKGuPB4iyHq5PpQy0fUpmSYXffE%2FDIvABYXye9xMdT1a4gdemeqS9Xq7sZiGhQ%2BX2epu8FxRMp7SI0Qrk2nhfPu%2B3XB86N%2F%2B3yPsAUfTem|ampid%3APL_CLK|clp%3A2544595

[dabateman]

Colin I got mine as an AC set like this one, but not this one specifically.  So don't know if this one will work:

https://www.ebay.com/itm/394116399696?mkcid=16&mkevt=1&mkrid=711-127632-2357-0&ssspo=qy-e8rbet6q&sssrc=2349624&ssuid=RxqKLNPYQr-&var=662642305832&widget_ver=artemis&media=COPY

[colinbm]

Good thanks Dave the 220v AC needs the correct capacitor in line.

[JMC]

Interesting. I could only find the bulbs when I searched, so it's interesting to see the power supplies are available too. Thanks.

[JMC]

Bit of an update.

 

I bought a few bits from Thorlabs to help with the light source, including a fused silica condenser lens, and a UV enhanced mirror.  The lamp house has now been modified to include the condenser lens (but not yet the mirror) and I am getting about 6x the light intensity at 10cm as I was before.  Yes, this means and even more intense and dangerous UVC light source - YOU HAVE BEEN WARNED.....

 

I also got a Semrock 260/16 Brightline filter (through a company called Laser2000 here in the UK). This was a bit of an expensive gamble, as it isolates the 254nm line quite nicely and has good transmission there while blocking visible light, however it offers little if any benefit for the IR blocking. The aim was to use this in combination withe the Sirchie 254nm filter, to help remove unwanted wavelengths better. Initial work looks positive for this, but I need to do a more thorough investigation.

 

Anyway, the fun stuff, some new images of the diatom slide again with the 10x NA 0.2 Zeiss Ultrafluar objective. All images at ISO1000 with the monochrome converted Nikon d850 which has a fused silica window. Images processed in Darktable as monochrome files. These are full frame images, but heavily reduced in resolution for sharing.

 

To start with a picture of the slide with white LED light (unfiltered).

 

And the same area with 254nm light.

 

Some of the diatoms are quite strongly absorbing at 254nm, which is in keeping with what I have seen in the literature, and there are some differences between the images which was good to see.

 

A couple of other images of the slide at 254nm.

 

 

Finally a couple of crops from the last image, at the original pixel resolution (these are obviously noisy at this size - they were ISO1000 and 30s exposures).

 

 

Overall I am happy with the direction this is going, but is far from optimized at the moment. Working at such short wavelengths continues to be a challenge. I think the 'scratches' on the images are from my camera and are either on the sensor itself or the coverglass.

 

When it's properly setup, I think I'll work out a way to do long 'bulb' exposures using ISO200 which should help with the noise.

[colinbm]

Great improvements Jonathan

[Andy Perrin]

Excellent, Jonathan! Why don’t you try image averaging instead of bulb mode? That way you don't get as much “amp glow” or risk saturating the sensor. You can use any astro software to average images (or Photoshop). Also if you take a shot without any diatoms you may be able to subtract off the scratches. 

[dabateman]

Wow this is really awesome.  I clearly see the difference.  

Now look at some insect wings. If a fly dies on your window sill or if you catch one in your window.  They seem to transmit UVC and I think you will see some nice structures. 

I had some luck with a cicada wing.

[JMC]

Cheers all and for the advice. I think it is coming along well, but there is a still a lot to do. For instance I still can't use live view to help with focusing, so there is a lot of trial and error there.

[colinbm]

I have just received one of these 10v3w UVC bulbs & I am surprised at how much 254nm they output compared to the the rest of the spectrum down into IR.

[JMC]

4 hours ago, colinbm said:

I have just received one of these 10v3w UVC bulbs & I am surprised at how much 254nm they output compared to the the rest of the spectrum down into IR.

Yeah they are pretty good. There are a few more bumps above 750-800nm, but the largest seem to be in the visible. I cannot go above 800nm for irradiance measurements so haven't been able to check up there. The low sensitivity of the camera at 254nm though means very good blocking of the out of band regions are needed if you are planning on trying to image it. For fluorescence it's likely less of an issue, so filtering it with something like UG5 or ZWB3 should be ok.

[colinbm]

Yes, thanks Johnathan
Bang-for-Buck, these are good value, as most Mercury vapour tube lights are only about 1watt per inch of tube.
These are 3w at 1-1/4 inch, & you can get in nice & close to the subject.

[colinbm]

Looking for reflectors for these 10v3w UVC bulbs.
I had no success with polished aluminium, UVC need a special Anolux-MIRO® UVC.
I did try these solar garden path lights & the reflectors seen to be working good.
I just cannibalized the reflectors, the hole in the base need to be opened up to 3/4" & they slide snuggly over the bulbs.
I made a plastic base with a 1/4" thread for mounting on camera accessories.

 

[Andy Perrin]

Surprised the aluminum didn’t work.