• Ultraviolet Photography
  •  

Solar panel in IR, verification

Filters Infrared SWIR
11 replies to this topic

#1 Andy Perrin

    Member

  • Members+G
  • 3,156 posts
  • Location: United States

Posted 18 January 2020 - 04:04

Stefano's thread on solar panels glowing in IR (when you push current through them the "wrong way" rather than shine light on them) occasioned some skepticism from me (since indirect bandgap semiconductors are supposed to make very poor LEDs). It seems that "very poor" is not the same is "you don't get anything at all" because I can reproduce the effect.

The equipment was:
-TriWave germanium-on-CMOS camera (QE >0 between 350nm and 1600nm, roughly)
-Solar panel I ripped out of an old IKEA solar powered lamp
-9V battery and alligator clips
-1100nm, 1150nm, 1200nm, and 1500nm long pass filters (all from ThorLabs). The first two are blocked to OD4+, the latter two are blocked to OD5+.
- f=100mm NIR-AR-coated Thorlabs lens (25mm diameter)

The results were as follows.

With no filter on the camera at all, and power applied to the solar panel:
Attached Image: No filter solar cells on.jpg

With the 1100nm long pass:
Attached Image: 1100nm+ solar cells on.jpg

With the 1150nm long pass:
Attached Image: 1150nm+ solar cells on.jpg

I didn't continue on to the 1200 and 1500nm long pass since I'd already lost the signal at this point. It seems pretty clear that most of the signal is below 1150nm (regardless of what graphs on the internet might say...).

I put my 980BP10 filter (Thorlabs, naturally, one day I will own their whole catalog and die happy) on the Sony A7S and took a photo:

EL-Nikkor 80mm/5.6 F/5.6, ISO10000, 5 min exposure (bulb)
Attached Image: _DSC4844.jpg

Still to try: putting the 1100nm long pass and the 1150nm long pass on my Sony and seeing what I get.

Edited by Andy Perrin, 18 January 2020 - 04:06.


#2 dabateman

    Da Bateman

  • Members+G
  • 1,911 posts
  • Location: Maryland

Posted 18 January 2020 - 06:02

Andy n=1, doesn't apply to all. But interesting result for your pannel.
I still need to image the one I found in a box. Connected it to a 123 battery and got crazy hot. So quickly disconnected it.

I then shawn light on it to make sure it worked. That was what I thought was more interesting part. A white Led, drove the fan, but not a 365nm led. My UV lights would not power anything. Needed more than 400nm, in the light. I didn't test my IR light to see if that would work.

Edited by dabateman, 18 January 2020 - 06:05.


#3 Andy Perrin

    Member

  • Members+G
  • 3,156 posts
  • Location: United States

Posted 18 January 2020 - 06:15

Dabateman, maybe in biology n=1 doesn’t apply to all, but it agrees perfectly with the bandgap of silicon, 1.12 eV = 1107nm, so I’m gonna go out in a limb and say for the same kind of solar cell it’s probably valid. Amorphous silicon would give something different probably, and there may be temperature variation also.

Edited by Andy Perrin, 18 January 2020 - 06:16.


#4 Stefano

    Member

  • Members(+)
  • 815 posts
  • Location: Italy

Posted 18 January 2020 - 10:03

So almost all signal is above 1100 nm, right? There isn't much below that.

#5 Andy Perrin

    Member

  • Members+G
  • 3,156 posts
  • Location: United States

Posted 18 January 2020 - 17:00

View PostStefano, on 18 January 2020 - 10:03, said:

So almost all signal is above 1100 nm, right? There isn't much below that.
Huh, no, there's quite a bit below, as shown by the big drop between the "no filter" and the "1100+" photos. We don't know quite how much, but I am not done yet, I plan to try some bandpass filters soon. Also, the signal is gone by 1150nm. So you definitely are not seeing the 1200nm water peak.

Edited by Andy Perrin, 18 January 2020 - 17:01.


#6 Andy Perrin

    Member

  • Members+G
  • 3,156 posts
  • Location: United States

Posted 19 January 2020 - 00:45

I repeated the experiment with the same lens and filters on my Sony NEX-7 (CMOS APS-C sensor) and got better resolution but no difference to the results. Evidently, the solar panel is emitting with all wavelengths <1200nm and within the range of the silicon sensor. Still to be done is to try various band pass filters on the same camera to see where the start of the illumination is.

1100nm+ filter
Attached Image: _DSC9666 1100nm+.jpg

1150nm+ filter (same exposure as 1100nm+ photo)
Attached Image: _DSC9667. 1150nm+.jpg

I could not see anything at 1200+.

Edited by Andy Perrin, 19 January 2020 - 00:50.


#7 Stefano

    Member

  • Members(+)
  • 815 posts
  • Location: Italy

Posted 19 January 2020 - 01:42

View PostAndy Perrin, on 18 January 2020 - 17:00, said:

Also, the signal is gone by 1150nm. So you definitely are not seeing the 1200nm water peak.
Yes, definitely. 1200 nm is surely out of range both for the camera and the solar panel.
What I initially thought when I did this experiment was that I could see the beginning of the 1205 nm water peak, not the peak itself.
Also I would have expected more signal at 1150+ nm, since the emission should peak there.

Edited by Stefano, 19 January 2020 - 01:45.


#8 Stefano

    Member

  • Members(+)
  • 815 posts
  • Location: Italy

Posted 19 January 2020 - 01:45

View PostAndy Perrin, on 19 January 2020 - 00:45, said:

I repeated the experiment with the same lens and filters on my Sony NEX-7 (CMOS APS-C sensor) and got better resolution but no difference to the results. Evidently, the solar panel is emitting with all wavelengths <1200nm and within the range of the silicon sensor. Still to be done is to try various band pass filters on the same camera to see where the start of the illumination is.

1100nm+ filter
Attachment _DSC9666 1100nm+.jpg

1150nm+ filter (same exposure as 1100nm+ photo)
Attachment _DSC9667. 1150nm+.jpg

I could not see anything at 1200+.
This is interesting. It seems like your camera can actually see past 1100 nm, since I imagine the filters you are using have a very sharp, almost vertical cut on and the 1150 nm one is blocked to OD 4+.

Both the TriWave camera and your Sony NEX-7 behaved in a seemingly same way.

Edited by Stefano, 19 January 2020 - 01:46.


#9 Andy Perrin

    Member

  • Members+G
  • 3,156 posts
  • Location: United States

Posted 19 January 2020 - 02:19

Quote

This is interesting. It seems like your camera can actually see past 1100 nm, since I imagine the filters you are using have a very sharp, almost vertical cut on and the 1150 nm one is blocked to OD 4+.

Both the TriWave camera and your Sony NEX-7 behaved in a seemingly same way.
Yeah, my NEX-7 seems to cut off between 1150 and 1200nm (and is very very weak at 1150nm, so probably only a few nm beyond that point). There does seem to be a little sensitivity to 1100-1150nm, based on this experiment.

The TriWave and the Sony NEX-7 behaving the same way indicates that the limitation is not the cameras, it's the solar panel. There simply is nothing to see past 1150nm or so, and the NEX-7 also goes to zero there, but because there is no light either, it's irrelevant.


Quote

Also I would have expected more signal at 1150+ nm, since the emission should peak there.
Yeah, except that graph seems to have been inaccurate, at least for my solar panel, and the CMOS cameras can't see much past 1150nm. Next on the list: where does the solar panel peak actually start?

Edited by Andy Perrin, 19 January 2020 - 02:22.


#10 Andy Perrin

    Member

  • Members+G
  • 3,156 posts
  • Location: United States

Posted 19 January 2020 - 04:42

So I tried using the following bandpass filters to image the solar panel using the TriWave to avoid camera complications (and also because it's way way more sensitive than my NEX-7 or even my A7S):
780BP30, 830BP40, 905BP30, 980BP20, 1064BP25.

In ever single case it was a bust:

780nm:
Attached Image: 780bp30.jpg

830nm:
Attached Image: 830bp40.jpg

905nm:
Attached Image: 905bp30.jpg

980nm:
Attached Image: EyeTVSnapshot.jpg

1064nm:
Attached Image: 1064bp25.jpg

But take out the filters and get:
Attached Image: no filter.jpg

The 980nm case coming up black was interesting because I tried that with the NEX-7 and got something. What that means is that I was seeing an out of band signal -- probably near 1100nm since we know the camera is sensitive there and the 1100+ long pass shows a strong image there on the TriWave.

Conclusions so far:
- The solar panel does glow, and the results look like Stefano's.
- The glow starts after 1064nm+/-25.
- The glow ends by 1200nm and is almost gone by 1150nm.
- So it's in the range 1090 - 1150nm.
- The NEX-7 definitely ends by 1200nm, and is shaky beyond 1150nm, but one can still get something from 1100-1150nm. However it's easy to punch through OD4 filters. I would always use the TriWave in this region, just to be sure of what I'm seeing, since the TriWave isn't struggling there.

Edited by Andy Perrin, 19 January 2020 - 04:54.


#11 dabateman

    Da Bateman

  • Members+G
  • 1,911 posts
  • Location: Maryland

Posted 19 January 2020 - 06:04

So your solar panel maybe a 1100bp50 light.
Interesting that you found that the Sony sensor can see to 1150nm.
I wonder if you test your 980bp10, 1064, Lp1100, lp1150 filters with the Sony and a glass of water with a hallogen lamp if you miss the dark water point.
Similar to what I was trying to get a year ago here:
https://www.ultravio...h__1#entry23674

#12 Andy Perrin

    Member

  • Members+G
  • 3,156 posts
  • Location: United States

Posted 19 January 2020 - 06:12

Quote

So your solar panel maybe a 1100bp50 light.
Nah, it's 1125bp50 if I understand the nomenclature right. That is, it's 25nm on either side of the center wavelength.

Edited by Andy Perrin, 19 January 2020 - 06:14.