What filter for UV < 320 nm?

[dabateman]

1 hour ago, lukaszgryglicki said:

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.

 

Well you could also order the bare Raspberry pi HQ sensor from MaxMax with your 50R. Only an extra $600 or so and its a BSI stacked sensor.  You could then compare it to the front side illuminated gfx50r.

Mine does seem wickedly sensitive.  Its just a small sensor. 

[JMC]

Both cameras I have used for UVC imaging (monochrome Nikon d850 and Sony A7III with Bayer filter in place but the sensor coverglass replaced with fused silica) did a better job than I expected with imaging at 254nm. However I have not tested a FSI sensor camera (both of these were BSI). I'm not really sure what to expect any more with UV imaging - rarely have results been what I expected.

[photoni]

2 hours ago, Daniel Csati said:

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

 

 

thanks @Daniel Csati for sharing this simple graphic.

I wonder if 365nm disco Led lights or Wood lights hurt

[ulf]

3 hours ago, dabateman said:

Well you could also order the bare Raspberry pi HQ sensor from MaxMax with your 50R. Only an extra $600 or so and its a BSI stacked sensor.  You could then compare it to the front side illuminated gfx50r.

Mine does seem wickedly sensitive.  Its just a small sensor. 

David, you make it sound like the deeper UV-sensitivity Raspberry pi HQ sensor from MaxMax amazing.

I get that it is true, compared to the sensitivity in UV of other sensors.

However I have gotten the impression that if you instead compare the UV-sensitivity to the sensitivity in VIS it is not that great due to other limitations.

Am I wrong?

 

It sounds like Lukasz is expecting a steady increase of sensitivity for shorter wavelengths due to the higher energy of the photons, as long as the bayer matrix and micro-lenses are removed and the cover glass is replaced by fused silica.

I suspect that is not the whole truth.

 

If that was the case there would be no need for special scientific sensors with extra high UV sensitivity, made possible with a layer structure to enhance the UV sensitivity. 

 

Please help me to understand and clarify what you compare with when saying "wickedly sensitive".

 

Even normal simple photodiodes have to get enhancement to be good at UV their peak sensitivity is normally in the lower NIR-band.

That is exactly the reason to having a BG-glass in a non modified camera to decrease the red part of the spectra, beside the sharp cutting dichroic second filter that cut UV and NIR

[dabateman]

@ulf,

The bare Raspberry pi HQ monochrome sensor is more sensitive to UVC at 254nm than my forensics UVC imager, which is designed to image fingerprints at 254nm. Also, its much much more sensitive than my Olympus Em5mk2 full-spectrum camera,  but that camera still has color filter array,  is front side illuminated sensor and has microlenses intact. 

 

I haven't compared the visible to UVC sensitivity,  but that may actually be hard. My lights aren't calibrated. So I can image with a low pressure Mercury lamp, which has both visible and UVC wavelengths,  but I couldn't tell you how low the visible output of the bulb is. So not a fair test. Sunlight doesn't have UVC. I could try similarly matched wattage of a compact fluorescent bulb, which contains Mercury to the low pressure Mercury UVC bulb, but still not a real known good test.

Can you think of a way to roughly know how many visible wavelength photons and UVC wavelengths photons I might be provided the sensor with? Than I could tell you how the quantum efficiency compares.

 

Dan also told me that a researcher was able to use the Raspberry pi HQ monochrome converted sensor to detect x-rays. 

[ulf]

Unfortunately I have no good idea to test this well.

 

The best idea is what you suggested, but the light from the visual tube is spread over a wide range of wavelengths, while your UV-C tube has one narrow main peak containing all the energy.

Also we do not know how efficient the conversion by the "phosphor" coating is, assuming that the internal UV-efficiency in both tubes are similar. That makes the comparison difficult.

Still it might give some indication to compare needed exposure times  or ISO between UV-C and VIS illumination

 

The difficulty of generating strong UV illumination is the next problem that Lukasz will see. 

 

I think the Teledyne quantum efficiency graph with the 220nm peak that Colin posted is for a specialist image sensor for combustion research

The cameras we get converted, sometimes to monochrome are optimised for best performance and SNR in the VIS range. I find it unlikely that such an optimisation would also make them very good at shorter UV-wavelengths.

I might be a pessimist.

 

[lukaszgryglicki]

Dan also told me that a researcher was able to use the Raspberry pi HQ monochrome converted sensor to detect x-rays. 

- wow - this is something...

 

The difficulty of generating strong UV illumination is the next problem that Lukasz will see. 

I guess I'll be facing that issues and learning by trial & error.

 

A joke: if p HQ mono could see X-rays, maybe it can also see alpha/beta/gamma rays? I have a piece of uranium, and a thorium/atom Nikkor lens....

In reality D600 mono used with that lens doesn't seem to see any radiation :p But I guess 50-80 uSv/h isn't enough to be detected :P

 

[Andy Perrin]

Alpha radiation is helium nuclei. There is no way it could be detected by a sensor that sees electromagnetic radiation, not nuclei.

[lukaszgryglicki]

This is why I said "joke" - and beta radiations is an electron or positron.

 

[ulf]

16 hours ago, Daniel Csati said:

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

As EN 60825 is for laser related  radiation safety it is not the best standard to use. 

 

IEC/EN 62471 or ANSI/IESNA RP-27 are the correct standards to use for all non laser illumination sources.

 

I have posted a weighting function table from the IEC/EN 62471 before:

 

When looking at 220nm in that table it seams not that safe either, just about 1/4 as dangerous as the 254nm peak, but still more than 1000 times as dangerous as the 365nm radiation.

 

It might be that recent research has found 220nm less dangerous than in this standard that is more than 10 years old.

 

 

[Daniel Csati]

5 minutes ago, ulf said:

As EN 60825 is for laser related  radiation safety it is not the best standard to use. 

 

IEC/EN 62471 or ANSI/IESNA RP-27 are the correct standards to use for all non laser illumination sources.

 

I have posted a weighting function table from the IEC/EN 62471 before:

 

When looking at 220nm in that table it seams not that safe either, just about 1/4 as dangerous as the 254nm peak, but still more than 1000 times as dangerous as the 365nm radiation.

 

It might be that recent research has found 220nm less dangerous than in this standard that is more than 10 years old.

 

 

Thanks for quoting the more relevant standard. Still pretty dangerous. If he decides to use a ArF excimer laser for illumination he will still find good use of the laser standard. ;)

[Andy Perrin]

8 hours ago, lukaszgryglicki said:

This is why I said "joke" - and beta radiations is an electron or positron.

 

At least there would be some chance of seeing an electron or a positron.

[lukaszgryglicki]

When looking at 220nm in that table it seams not that safe either, just about 1/4 as dangerous as the 254nm peak, but still more than 1000 times as dangerous as the 365nm radiation.

 

It might be that recent research has found 220nm less dangerous than in this standard that is more than 10 years old.

I want to research that wavelength anyway, I will need skin & eye protection.