UVA and UVB polariser - Moxtek UVD260A

[JMC]

Cross polarised imaging of sunscreens was what first got me into UV imaging. Cross polarisation - one on the light source, one on the lens, set at 90 degrees to each other - gets rid of specular shine, and allowed me to see the product film morphology after it was spread. At the time, I used some older glass linear polarisers, which knew weren't ideal for UV imaging, but as they did transmit at least some of the UV it was enough for my initial research.

 

However these old linear glass polarisers would be no use for UVB as they didn't transmit anything below 350nm. Since the initial work I have been on the look out for polarisers which would be more suited to UV and especially UVB imaging.There are some UV polarisers at Edmund Optics, but these seem to be aimed at around the 350nm to 365nm region, and not much lower so wont be good.

 

I came across Moxtek UV polarisers, http://moxtek.com/optics-product/ultraviolet-polarizers/ , and they have been kind enough to loan me a pair of them for a week to run some tests. They arrived today, and I'll be updating this thread as I run some tests with them.

 

 

 

Size would be an issue or most jobs - these only have a clear inside diameter of about 17mm. I'll be using them with my Rayfact and LC8 light source (which has a small opening). Cost is, well, quite frankly enormous, and I am happy to have them as loan ones - I just hope I dont damage them, otherwise I may have to sell a kidney to cover the cost. They can make them bigger (up to around 50mm diameter), but the cost follows suit.....

[dabateman]

These may work best with you UAT lens. I can get away with 12mm filters which have only 8mm opening using them close to the front element. Although this is on a M43rds camera which also has smaller sensor.

 

But since the front element on the UAT lens is 18mm. It will be your best option.

[JMC]

Good point David. If I get chance I will try that too. If the proposed work comes off though, it would be based around a Rayfact as I don't fancy selling one of my UATs, hence my interest in that one.

[dabateman]

Good point David. If I get chance I will try that too. If the proposed work comes off though, it would be based around a Rayfact as I don't fancy selling one of my UATs, hence my interest in that one.

 

Selling one of your UATs? How many do you own? There are only about 25 in the world. I use mine as a work horse, shifted to almost my most used lens. But I am afraid of loosing it. I actually had a dream last night that I broke it. So will be more careful in the coming weeks.

[JMC]

Selling one of your UATs? How many do you own? There are only about 25 in the world. I use mine as a work horse, shifted to almost my most used lens. But I am afraid of loosing it. I actually had a dream last night that I broke it. So will be more careful in the coming weeks.

I have 2 - one boxed with filters, and one as a bare lens, David. Having the bare lens one, makes me feel better about using it for work. The boxed one I am more paranoid about using, but is a backup in case the worst ever happens. I too have had nightmares about breaking one.

[JMC]

Experiment 1 - Flower, UVA (Baader U), multispectral colour camera

For these experiments I am using my Hamamatsu LC8 200W HgXe lamp. The spectra for that can be see here - http://www.ultraviol...dpost__p__25916 . This is useful as its output is only about 1cm diameter. The drawback is a slightly uneven light intensity distribution, however for the purposes of testing here that will not be considered.

 

Camera - EOS 5DSR, UV-Vis-IR conversion.

Lens - Rayfact 105mm UV Macro (at f16).

ISO - ISO 800.

Filter - Baader U on lens. Moxtek UV polarisers, on lens and light source as needed.

Processing - Whitebalanced using PTFE tile in Darktable. Images exported from Darktable, and then auto-contrast used in Photoshop, as they were all a bit dark (this flower seems to be very low UV reflectance, especially when compared to the black background).

Setup - light approx 70cm from the flower. Camera on tripod next to light source, focussed on central part of the flower. Background was cardboard painted with a coat of Semple Black 2.0 paint. A photo showing the setup and the flower used is shown below;

 

 

First image - no polarisers, 1/4s exposure

 

Second image - parallel oriented polarisers, 1s exposure

 

Third image - cross polarised. 10s exposure

 

As expected, parallel polarisation robbed about 2 stops of light and looked very similar to non-polarised. Cross polarisation definitely took the shine away (look at the glass vase reflection in the first 2 images which goes away when the image was cross polarised, as well as the petals). It also took about 5 stops of light compared to the non-polarised image. The image is quite monochrome, which I presume is due to the nature of the light source. Being HgXe it s very spikey emission spectrum and the peak at 365nm dominates the response when using the Baader U.

[dabateman]

Jonathan,

Your polarized image is amazing. Looks like you have some fluorescence coming out in the UV. This is really interesting and makes me want to test out the polarizers I have.

 

Have you heard of this company:

https://www.polarprofilters.com/products/quartzline-filters?variant=12201172271188

 

May work, but I don't know if they are using standard PVA iodine film or silver substrate in the glasss or are true quartz polarizers like in microscopes. I sent them an email, hopefully they can let me know the spectral range of their filters.

 

I also came across this UK company:

https://www.knightoptical.com/stock/optical-components/uv-optics/uv-polarisers/uv-grade-sheet-polarizers/

 

Theirs seem promising but the transmission is low in uvb range compared to visible.

 

[JMC]

David, I hadn't seen those Quartzline camera filters before. The issue I found with polarisers though have been the following - they often use polymeric polarising film in between glass, the polymer is highly absorbing in the UV (either the polymer or the adhesive used). Also they often have anti reflective coatings, again not good for UV. Also I found I could get the best extinction of specular reflection with linear polarisers. I found old glass linear ones to be best for cross polarisation. I'd be interested to hear what they say if you hear back. If they have a transmission spectra, even better. The Knightoptical ones, as you say aren't good for UVB.

 

Tell me more about the fluoresence. The light source was only filtered with the first polariser. The lens was filtered first by the second polariser, and then using a Baader U.

[JMC]

Experiment 2 - Flower, UVA (Baader U) and UVB (Invisible Vision 308nm), monochrome 5DSR camera

For the second experiment I am using my Monochrome 5DSR and the Hamamatsu LC8 200W HgXe lamp. I am using the Monochrome 5DSR as my standard multispectral 5DSR doesn't have UVB sensitivity as it still has the OLPF attached.

 

The aim here is to image in UVA (with the Baader U) and UVB (with the Invisible vision 308nm), and look at how cross polarisation effects the images in both regions.

 

Camera - EOS 5DSR, Monochrome conversion.

Lens - Rayfact 105mm UV Macro (at f16).

ISO - ISO 800 (higher for cross polarised UVB images).

Filter - Baader U or Invisible Vision 308nm on lens. Moxtek UV polarisers, on lens and light source as needed.

Processing - Custom in camera whitebalance, saved as jpegs, then cropped in Photoshop (no further contrast boosting as I upped the exposure from the previous test).

Setup - light approx 70cm from the flower. Camera on tripod next to light source, focussed on central part of the flower. Background was cardboard painted with a coat of Semple Black 2.0 paint.

 

First image - UVA (Baader U) no polarisers, 1/2s exposure

 

Second images - UVA (Baader U) cross polarised, 15s exposure

 

Third image - UVB (Invisible Vision 308nm) no polarisers 2s exposure

 

Fourth image - UVB (Invisible Vision 308nm), cross polarised, 30s exposure, ISO1600

 

Fifth image - UVB (Invisible Vision 308nm), cross polarised, 30s exposure, ISO3200

 

I've noted with previous work using the monochrome camera and LC8 light source, that the Invisible Vision filter needed about 2 stops more light than the Baader U, to keep exposure consistent, so I kept that ratio here. I gave the cross polarised image 5 stops more in the UVA region with the Baader U. In the UVB region, 5 stops more leads to quite a dark image, so I also tried it at 6 stops more (fifth image). This is because the polarisers transmission is lower at 308nm than 365nm, hence longer exposure needed. As you can also see I ran out of exposure time on the the camera, so needed to up the ISO to increase the exposure.

 

As before the reflection in the glass shows the effects of cross polarisation quite well. I am aligning these by eye, so may not be 100% on each time which is leading to a small amount of reflection in the glass - 1 degree out in the alignment will cause this. The glass is also nice and black in the UVB image.

 

Key difference I can see for the flower is in the pollen in the middle - it shows up as being bright in the UVA image, but dark in the UVB image.

[dabateman]

After looking at your monochrome images. I now don't think you had fluorescence in the original. But rather was spectral reflection leak due to the angle of the polarizers in that setup.

I was looking at the bright, fourish pollen points near the center of the flower. If they were indeed fluorescencing. Then they should have appeared as brighter dots in the monochrome, compared to their immediate surroundings. But they don't.

However, still very interesting the details I am loosing without polarizers. May still need to try this in the UVA region.

 

The fused silica filters, I now have little hope in. Their website says all the filters have 16 layer anti-scratch coatings. This for sure will rob us of uv, and may only be 400nm to 700nm. But I will see what they come back with next week. If magically they use oriented silver in fused silica, then they would be great.

[JMC]

David, I would recommend partial polarisation for pictorial imaging. Complete cross polarisation can make things look a little flat. Part polarisation gets rid of some of the shine and allows the underlying colour to be seen more easily.

[dabateman]

Well I was surprised by the number of polarizers I have. I have 10 circular polarizers, three linear polarizers and one 100mm x 100mm linear polarizer sheet. My Em1 came back from repair today so I tested some.

Shocking discovery. No polarizer with 2 reptile bulbs with 313nm filter, the exposure was f8, ISO 800, 8seconds. With my Quantaray L.P 52mm polarizers I got an equal image at F8 ISO 6400, 25 seconds. Only 5 stops less light. I have 3 Quantaray filters and one was really good, next was less than 1/3 stop worse and the third was at least a full stop worse. So there is variability.

The other winner was my Hoya circular polarizer with 7 stops down at 313nm. I am very excited by this. My 0.15mm linear polarizer sheet was horrible, no uv light below 400nm passes through it.

These filters are totally usable like regular in UVA range with BaaderU filter. As in you clearly see in live view and can adjust them for effect. The Hoya at max polarization is only 2 stops slower than without polarizer. The linear Quantaray are 1 to 1.3 stops slower.

So now I can play around.

 

I did get an email back from Quartz line and the tech claims the polarizer can see 300nm to 1100nm. I don't believe it. But maybe Jonathan you could contact them, ask for a sample and run a proper spectrum. If true this would definitely help their sales. They are the only ones with very large filters.

[JMC]

Thanks David, I will follow up with them.

 

As an example here are a few different circular and linear polarisers transmission spectra, measured as I was dong my original research in the area.

 

 

These were measured using both lamps on my light source so there will be some stray light effects at the shorter wavelength end, but as you can see there is pretty good transmission in the UVA for most of them. But down by 310nm it is essentially zero. I haven't measured the Quantanary ones though, but they sound interesting.

[dabateman]

Jonathan here are some images.

 

This is the flower I shot to see the normal colors:

 

This is the flower shot using two reptile bulbs 3 inches from the area. On the UAT is the 313bp25 and 330WB80 improved filters. The exposure is F8, ISO 800, 10 seconds:

 

This is the shot using one of my Quantaray L.P. 52mm filters I call it green as I colored the L.P. on the filter green differentiate if from my other good one that I colored red. Exposure is F8, ISO 800, 60 seconds:

 

This is the same using the other L.P. filter I call red. The exposure here is F8, ISO 1600, 60 seconds:

 

I was also playing round with cross polarization using my 365nm LED bulbs, as I have 52mm filters threads on them. I was able to get some interesting effects using the Baader Venus filter. What was fun was I could place the red on a 365nm bulb, the green on a 365nm bulb and my Hoya circular polarizer on the Camera. It actually worked and could turn the Hoya to the cross point as you see the image dark. HOwever, I don't know if the two linear polarizer were outputting the same light. But was fun to play around.

 

It doesn't show up well here, but I was able to knock down some highlights in the flower. I am only showing the infocus flower that I used. This will be fun with better live subjects. This flower in the image is actually pressed, flat and dried.

[JMC]

Very cool David. Looks like you're getting a good image without IR contamination (because of the dark green appearance).

 

I did speak with PolarPro about the Quartz line filters. When asked, the first response I got was that the polarisers transmitted 90% of the light down to 170nm. When I asked them to check whether that was actually true, I got back a response saying, "oh yeah, sorry that's just the bare quartz. The polarisers cut 1.6 stops of light", and no discussion about wavelength cut off. Hmm, not inspiring confidence, and I wouldn't believe any of it until actually tested. Does seem slightly odd to market a UV filter that doesn't block UV without plastering loads of coatings on it. Surely have a substrate for a UV blocking filter which transmits UV down to 170nm seems a bit weird.

[dabateman]

Yep,

Kind of why I hoped they would give you a sample to test. I am not sure they really know. Also with 16 layers to increase strength and scratch resistance, I think they may actually be good uv blockers.

 

But I was shocked to get anything out of my Quantaray filters, So I guess you never really know until you test it.

These were my only good transmiting ones. My KSM also seem to block uv.

[Guest]

This all made me wonder... just how (in)effective might my CPs be in this context? So I pulled out a handful of old CPs I've had forever and photographed one of my highly technical test scenes... shown here in VIS and UV. Included here is a paraffin candle, glass vial of stainless steel bearings, an ornament, plastic dice, a colorful eraser, and some kind of polished orange stone (I don't know what) which has some highly glittery content in it (mica maybe?).

 

 

Next I added one of each of the following circular polarizers to the lens in sequence, taking an image at the position of highest observed reflections (top row), then another at the position of lowest observed reflections (bottom row).

 

 

As expected, these CPs are quite ineffective in this UV range, as there is very little difference between the two positions. It was worth a try, and since I had low expectations going into this, I'm not exactly disappointed :).

[dabateman]

Mark,

I think your being too hard on them. Look at the two bright dots on the unknown orange rock. The F-pro 77mm cp kills the glare. The Hoya 55mm and Tiffen 58mm do an okish job, the Hoya looking little better than the tiffen. The F-pro 77mm also pulls out the best circle strain appearance from the one spot on your red die. Polarizers can see stress and strain points in plastics.

I would say the 77mm is the winner, with the Hoya 55mm a close second. How much did they knock down the exposure in the UV?

[JMC]

Mark. I had little luck in eliminating specular reflection with my circular polarisers. Also keep in mind my images were done with polarisation of the light source, and then setting a second polariser on the lens at 90 degrees to the one on the light. Did you polarise your incoming light? If not, then I wouldn't expect it to cut all the shine.

[Guest]

I don't recall off-hand what the difference in exposure was with-vs-without the polarizers, but it was significant - to say the least. And I didn't polarize the source - I see how that would be better, but this was just a curiosity test. I think it's tough enough to get enough UV in our images as it is, never mind making it even more difficult by cutting it further with polarizers (though your examples above (JMC) do make a strong case...). I may just splurge for some linear polarizers and have a real go at it sometime.

[JMC]

Ah we have flowers again in the garden.

 

Out of interest, I just got a Celandine and did a couple of quick images using my multispectral EOS 5DSR. Light source was a Hamamatsu LC8. Lens Rayfact 105mm (at f22). One image unpolarised, and one cross polarised with the Moxtek polarisers. UV filter was LaLaU.

 

Whitebalanced in Darktable.

 

Unpolarised (ISO800, 4s)

 

Cross polarised (ISO2000, 20s)

 

The cross polarisation gets rid of the shine in the unpolarised image, and lets the yellow come through.

 

However 20s (or even 4s) on a wooden bench which also has the light source attached is not good. There is a little shake in flower.

 

All I need now is a lottery win to buy some large versions of these polarisers suitable for studio flashes.

[Andy Perrin]

Definitely I like the cross polarization effect for the very shiny flowers. Try a buttercup next! I bet going to full cross-polarized is too much and it may be better to trade a little more light for a little less specular reflection blocking. But I know you were in testing mode here, not art mode.

[Stefano]

How does cross-polarization work? Wouldn’t blocking both orthogonal axis block everything?

[Andy Perrin]

Stefano: diffuse reflected light loses its polarization (to some degree, anyway).

[dabateman]

Stefano it you want your mind blown watch this description of Bell's theorem.