What filter for UV < 320 nm?

[lukaszgryglicki]

I need a filter that passes below 320nm and blocks > 320 nm (or stack) with 52mm thread.

Is there anything reccomended?

 

[Andy Perrin]

You need a stack for that because the blocking needs to be better than almost any single filter on the market. Omega probably sells a 320nm short pass, or can make you one, see their website. However you will want to stack it with another shortpass to increase the OD. For that purpose you could use an Omega 330WB80. 
 

See dabateman’s and Bernard’s posts for ideas. 
 

By the way 52mm will cost a fortune. Suggest going with 25mm. 
 

If you use an artificial light like an LED with no infrared you may be able to avoid stacking. But not for sunlight. 

[lukaszgryglicki]

No, I need 52mm, I've already piad fortune for UV-Nikkon and debayred Fuji GFX 50R, so Ican pay extra for a filter.

[colinbm]

Jonathan has a 312nm filter.

[JMC]

Do you need bandpass or do you want a 320nm short pass which blocks everything above 320nm?

 

Bandpass, best option is probably Edmund (I know Edmund are not always liked, but I use them for some jobs as they have good transmission at the specified wavelength). However with normal lights - sunlight, flash etc - the blocking is not enough so you'll need to combine it with something else, either another one the same, or U340 4mm etc.

 

For my UVB microscopy, I use 2x 313nm Edmund OD4 filters stacked together, and that is with a mercury xenon lamp which has nice sharp 313nm peak (along with 365nm and some visible). The camera is a d800 mono conversion.

 

Short pass which cuts off at 320nm, and blocks everything up to say 1200nm, I don't know of any one filter that would do that.

 

What is your lighting? The amount of other wavelengths will help you understand how much blocking you need. Basically thought for imaging <320nm you'll need more blocking than you think you need. Even with a mono conversion the sensitivity is very very low down there compared to longer wavelengths.

[Andy Perrin]

Omega sells a 305nm short pass. 
 

https://www.omegafilters.com/product/4086

 

It is only OD4-ish blocking up to 600nm so you will need to stack. Assume massive IR leakage. 
 

Omega will let you custom design filters so if you are going all in on this, they could make you something. 

[Stefano]

For 310 nm (bandpass) I use two Chinese 310 nm bandpass filters stacked together, which work for me, so I too think you will need a stack. If you use sunlight I think you will need at least OD 7 or more.

[dabateman]

I am surprised no one mentioned the Invisible vision 308nm filter.

https://www.invisiblevision.com/products/accessories/

 

You may need to stack it with a zwb1 or Ug11 filter to ensure no IR leakage. 

Jonathan and I think Colin own it.

Its in 52mm thread.

[colinbm]

I mentioned it above but as 312nm, my mistake but it has a widish half width.

[lukaszgryglicki]

I will need to research all of them once I have my UV-Nikkor (and also with 50R mono quartz), I will also need some lighting, I need UV-B lighting (around 300-320) - I can skip it first, because I can hope for a slightest bit of that light in sunlight, then I really need UV-C light (ideally below that most dangerous, say 190-230 nm)?

I want to try UV-C < 240 nm just to see if I'm able to get something: UV-Nikkor lens, then naked/debayered sensor with quartz 150+nm glass.

I wonder what would be the ideal filtering stack & lighting for < 240 nm imaging (I mean macro, don't even hope for any lighting allowing me to do normal photography of anything 5 meters away - which would be a dream).

 

BTW, how dangerous is, say 220 nm UV-C? I know that 240-270 range is most dangerous, but 220?

[colinbm]

Far UVC Excimer 222nm with the correct filter is harmless to humans & animals.
 

[lukaszgryglicki]

Oooo that looks ideal for me, need to research that - do you have any quick link to that UVC Eximer 222nm - is this laser or what? Gonna research tomorrow.

[colinbm]

Both these need the filter to block the small leak at 254nm & have no ozone.
The filters are not perfect at blocking visible & IR light, but are OK for induced visible fluorescence.
These need to be tested & published out 1000nm.
https://www.bbtuv.com/shop/
https://www.suntechleds.com/far-uvc-lamp-light/
 

[lukaszgryglicki]

OK I've found similar, but I'm having a problem to calculate/conver those irradiations in mJ/uJ / cm^2 into anything that would give me an idea of how much "light" it gives.

Say, is there a way (other than buying & testing) to say, hom much exposure will I need if I shine with this into something else from 1 meter away?

Assumin UV-Nikkor (70% transmission), quartz coverglass (probably similar) and no Bayer Filter....

 

[colinbm]

I have no experience with Far UVC photography,
only some visible fluorescence.
It is absorbed by the atmosphere, but I don't know what the transmittance in air is ?

 

[lukaszgryglicki]

Transmitance in air should be OK above 200nm. I'm speaking about 1-2 meters no more.

 

 

[ulf]

Your main problem here will likely the diminishing sensor sensitivity for short wavelengths.

 

1-2m is a looong distance to spread the illumination energy from a lamp.

As I am sure you know the power density decays with the square of the distance.

[lukaszgryglicki]

Yes I know power density ~ 1/d^2. Hmm sensor sensitivity is so low at 200nm ? My understanding was that sensor sensitivity drops mainly because of Bayer array which becomes almost black in UV-B, UV-C, having everything quite transparent - the bare sensor also drops at 200nm? Strange, photon energy is higher when wavelength is shorter.... so, say, if I shine photons with wavelength, say, 50nm directly onto sensor (skipping air being opaque at that wavelength) it won't detect it?

I know that Si sensors are less and less sensitive to IR > 1um, because photon energy goes down, but I expected the opposite in UV - everythings seems to absorb UV better than visible, so I expected the same for silicon pixels.... but of course I know this isn't that simple, say, deeper into UV - then comes X-ray which stops interacting with *some* matter and goes through (hence X-ray skeleton scans) - then goes gamma rays that can penetrate lead...

 

[Daniel Csati]

I'm not an expert on the topic but ..

at 200nm the most of the light will get absorbed in the first few nanometers. There is in the first few nanometers either a passivation layer or a transparent electrode on the top of the silicon. Plus the Bayer filter, microlenses etc.

..Now the light either doesn't reach the Si at all or gets absorbed only at the surface and can't generate mobile (or long enough lifetime) hole-electron pairs. Without electrons there is no voltage on the pixel, so no signal.

 

I can imagine that the process is not this simple but that's probably most of it.

[colinbm]

I this what you expect your new camera to be able to achieve ?

https://www.photometrics.com/learn/imaging-topics/quantum-efficiency

[dabateman]

Humidity matters. You need a dehumidifier near where your imaging.

Be as close to your subject as you can.

Lukas will be using a monochrome stripped sensor,  so only thing between the sensor and the photons maybe a thin sheet of fused silica.  If he is getting that added by Dan of MaxMax. 

So he should be able to see quite deep.

I also have a feeling using this specific Sony sensor in the Gfx cameras, that it might be quite sensitive. The camera also has this magical ability to not care about ISO settings and you can easily under expose by many stops without worrying. 

[Daniel Csati]

Some information on how dangerous are the different UV wavelengths: very dangerous. Protect skin (hands, face, legs etc) and eyes.

Before using UV light sources I would do a safety assessment based on maybe DIN EN 60825

[lukaszgryglicki]

I ordered fused silica supposed to transmit from 150 nm... and I've read (a bit) about those eximers UVC 222nm. I'm starting to think this might work... otherwise I would probably need mercury low pressure lamp, but it uses a lot more dangerous 254-255nm and also 193nm (which is OK but is generating ozone).

 

From @colinbm chart I see another peak in 220-250 which would be perfect for me - just wonder if this is just one particular sensor or a generic propery of silicon sensors...

 

I would like to see that efficiency for GFX sony sensor... naked sensor.

 

At 1 micron for example response is much lower than on 200nm, and I can make photos with 950 nm IR filter easily - even without debayering (which makes no difference above 850-900nm IMHO - all R, G, B pixels are transparent in IR).

 

 

 

[lukaszgryglicki]

Maybe BSI/FSI type also matters, because it seems tht in Front side iluminated sensors (50R) gate is made of thin material that can block UV-C a bit (or a lot?) while in BSI gate is on the opposite side... correct me if I'm wrong... seems like I will be the first one to try.

 

[Andy Perrin]

Don’t expect any QE chart to be “generic” out in UV-C. Every case will be different. You have already identified some of the reasons.