• Ultraviolet Photography
  •  

Candle Flame with UV Imager

31 replies to this topic

#21 dabateman

    Da Bateman

  • Members+G
  • 2,699 posts
  • Location: Maryland

Posted 24 October 2020 - 08:37

Thanks Colin I think you gave us the answer. That a candle flame is the best UV filter stress test. You should be able to see the IR leakage through the BaaderU filter with this test.

So if we assume incomplete ionizing radiation at 250nm for the 2020BSI sensor, then a pixel will have 28% qe. Assumeing 100% ionizing radiation and its 62%.
At 850nm the qe is 50%.
So for 850nm candle with an OD3 blocking filter is 0.001x0.5x0.005 mW/cm2 = 2.5x10-6.
At full ionizing radiation 0.3x0.62x0.00001 = 1.86 x10-6

So there is twice as much detected 850nm light going through the filter than 250nm light.

So lets not use candles as UV light sources.
Best to see if you can get your hands on a 254nm low pressure mercury bulb. The cost has come down. Also ypu can buy a portable one cheap, that are typically used for short wave rock hunting.

#22 Stefano

    Member

  • Members(+)
  • 1,719 posts
  • Location: Italy

Posted 24 October 2020 - 08:52

View Postcolinbm, on 24 October 2020 - 05:20, said:

Shouldn't the flame spectrum represent the atoms being consumed ?
Maybe I understood what you meant. If you exite atoms on a flame they will emit light according to the energy level gaps between different electron energy states. I tried a famous experiment: take some cotton or a piece of paper towel, soak it into salt water and put it on a flame. If you look at the spectrum of the flame (I tried with an "eye spectrometer" I made with a slit and a diffraction grating) you will see the 589 nm sodium yellow line. They are actually two very close lines, at 588.995 and 589.5924 nm. The difference is basically spin up/spin down. Maybe with a very narrow slit and a good eye you can see the doublet.

#23 colinbm

    Member

  • Members+G
  • 2,287 posts
  • Location: Australia

Posted 24 October 2020 - 09:02

I have this UVC + UVvacuum lamp, but it produces ozone, but when I wrapped the tube with 5 thousands/inch clear Mylar it was eliminated, but please treat this carefully & do your own tests to be satisfied you are free of ozone & protected from the UVC radiation.

https://www.ebay.com...872.m2749.l2649

Uncovered bare tube....
Attached Image: 202007108WUVfluoro-Ref.png

Covered tube with Clear Mylar.....
Attached Image: 20200712UVCFlupr+Mylar-Ref.png

#24 colinbm

    Member

  • Members+G
  • 2,287 posts
  • Location: Australia

Posted 24 October 2020 - 09:06

View PostStefano, on 24 October 2020 - 08:52, said:

Maybe I understood what you meant. If you exite atoms on a flame they will emit light according to the energy level gaps between different electron energy states. I tried a famous experiment: take some cotton or a piece of paper towel, soak it into salt water and put it on a flame. If you look at the spectrum of the flame (I tried with an "eye spectrometer" I made with a slit and a diffraction grating) you will see the 589 nm sodium yellow line. They are actually two very close lines, at 588.995 and 589.5924 nm. The difference is basically spin up/spin down. Maybe with a very narrow slit and a good eye you can see the doublet.

Yes I was mistaken, you are correct.

#25 colinbm

    Member

  • Members+G
  • 2,287 posts
  • Location: Australia

Posted 24 October 2020 - 11:07

The Butane Lighter flame spectrum is slightly different....

Bare flame...
Attached Image: 20201024ButaneLighterFlame-Ref.png

Flame with S8612...
Attached Image: 202024ButaneLighterFlame+S8612-Ref.png

PS, unfortunately the data gets truncated at the ends as it is too small, so on the UV end it will continue along further at the lower end, tapering away....

Edited by colinbm, 24 October 2020 - 11:09.


#26 WiSi-Testpilot

    Member

  • Members(+)
  • 105 posts
  • Location: Germany

Posted 24 October 2020 - 18:05

Colinbm, thank you for your nice spectra. Now I have tested it with a new butane lighter. The result is basically the same as with a candle. The candle is a bit brighter. The daylight leak is not a problem for me. For the planned representation of the corona discharges, the leak should add the landscape.
So I'll stop at this point. I've learned a lot.
Best regards,
Wilhelm

Butane Lighter, 270 nm, 500 ms, with and without a 5 mm glass plate
Attached Image: Butane_Lighter_with&without_5mm_glass.jpg


Candle from #18, 270 nm, 500 ms, with and without a 5 mm glass plate
Attached Image: Candle_with&without_glass_plate.jpg

#27 Stefano

    Member

  • Members(+)
  • 1,719 posts
  • Location: Italy

Posted 24 October 2020 - 18:20

Is that your window in the background in the first image? You shouldn't be able to see through it at 270 nm. That's probably also a leak.

#28 WiSi-Testpilot

    Member

  • Members(+)
  • 105 posts
  • Location: Germany

Posted 24 October 2020 - 18:35

Yes, that is the daylight leak. But it wasn't clear to me that there is a stronger source in the system.

#29 dabateman

    Da Bateman

  • Members+G
  • 2,699 posts
  • Location: Maryland

Posted 24 October 2020 - 19:55

Using the numbers from Colin's low pressure mercury bulb is interesting.
At 800nm we have 0.007mW/cm2. The plot for the GSENSE 2020BSI has the exact same quantum efficiency at 250nm as at 800nm.
So that could divide out.
Thus assume OD3 blocking filter and you have the following:
@800nm, 0.007mW/cm2 x 0.001 x 0.62 = 4.34x10-6
@250nm (ionizing), 0.085mW/cm2 x 0.3 x 0.62 = 1.58x10-2

@250nm, 0.085mW/cm2 x 0.3 x 0.28 = 7.14x10-3.

So using the correct light you may not see any IR leak in your UVC images using your current filter.

I am just assuming its OD3 Blocking, it might be better, as your candle IR leak wasn't super bright, and it would be summed over more IR wavelengths.

I am ignoring the lens transmission as it doesn't matter much. Basically 84% in the IR and 88% at 250nm. I know it cuts off hard at 220nm though. So not the best lens for deep UV-c.
Here is a link to some of its spectrum, starting at 250nm into SWIR:
https://www.universe...trans_curve.pdf

Edited by dabateman, 24 October 2020 - 20:08.


#30 WiSi-Testpilot

    Member

  • Members(+)
  • 105 posts
  • Location: Germany

Posted 24 October 2020 - 20:30

Thank you for your calculations. I am very happy with the setup, but as you wrote, the candlelight is not optimal to test the UVC sensitivity.
Best regards,
Wilhelm

#31 Andy Perrin

    Member

  • Members+G
  • 3,995 posts
  • Location: United States

Posted 24 October 2020 - 21:31

Yeah, colin's spectrum says it all, really. The flame is weighted very strongly toward IR so needs fantastic blocking in that region if you truly want to see the UVC only. Wilhelm, your setup is fine if you just use a light source that is mainly UVC and not mainly IR with a tiny bit of UVC.

#32 dabateman

    Da Bateman

  • Members+G
  • 2,699 posts
  • Location: Maryland

Posted 25 October 2020 - 05:58

View PostWiSi-Testpilot, on 24 October 2020 - 20:30, said:

Thank you for your calculations. I am very happy with the setup, but as you wrote, the candlelight is not optimal to test the UVC sensitivity.
Best regards,
Wilhelm

Actually for testing I think you nailed the worst case scenario. You still see UVC on the bottom and sides. Its not totally blown out by the IR leak at the top. So this is a good way to stress test for leaks.
If I get a UVC capable sensor, this will be my first test.