Jump to content
UltravioletPhotography

Strange spectrums with a diffraction grating


Recommended Posts

I have thousands of photos, so I will have a lot to post for a lot of time. This time I will focus on some of the spectrums that I have been able to capture, in various ranges.

For all shots the camera was a modified Panasonic DMC-F3, which doesn't have anything above the sensor except the lens (and filters). The focal length is always (apparently) 5 mm. I always WB in camera.

The source of illumination for the first photographs was a 12 V, 5 W tungsten bulb (used for cars, the 5 W/21 W type), ran at 24 V. I usually intentionally overload incandescent sources, to make them emit light with a spectrum more similar to sunlight. Of course using it like this dramatically shorts its lifespan, and a few minutes were enough to deposit tungsten on the inner surface of the glass envelope.

 

Full spectrum (no filters), f-stop: f/7.1, ISO 80, 1/1000 s exposure. "Sunny" WB.

post-284-0-84580700-1576869631.jpg

 

You can easily see the infrared region, but the ultraviolet one is just obscured. Tungsten lamps, although very inefficiently, do emit usable amounts of UV, I even used them for reflected ultraviolet photography.

 

Then, infrared only. I still don't have a proper IR longpass filter, and I used one that I made with black pen ink (not the "liquid" one, made with a black marker, of my first IR photos). The quality is what it is, at the moment is the best filter I have.

 

Infrared (black pen ink filter), f-stop: f/2.8, ISO 400, 1/500 s exposure, 5 mm focal length. Same WB as before.

post-284-0-97394600-1576870567.jpg

 

And now the infrared white balanced version, f/2.8, ISO 400, 1/800 s exposure.

post-284-0-07915700-1576872769.jpg

 

We can already note two things. First, when white balanced, the shorter IR wavelengths look yellow, and the longer ones look blue. By coincidence, it is the same in UV. And then there is a mysterious absorption line, roughly at something like 960 nm, and I cannot explain it. It is typical of gases an vapours to absorb at a specific, narrow line, but in this case the only gases are inside the bulb and in the air.

 

Taking a photo at the spectrum from a 940 nm LED, we see the same thing. I estimated the wavelength knowing that, from a previous experimental measurement, this LED peaks at 946 nm.

 

940 nm LED spectrum, f-stop: f/2.8, ISO 80, 1/8 s exposure.

post-284-0-93530700-1576874813.jpg

 

This one, instead, is just very strange. Underexposing (in camera, not digitally) an image similar to the third one (white balanced infrared) I obtained a rainbow.

 

f-stop: f/2.8, ISO 100, 1/640 s exposure.

post-284-0-94076500-1576874900.jpg

 

The next ones were shot with sunlight. Since I don't have a PTFE target, I used a paper tissue instead. It is one of the closest thing there is, together with snow (in my opinion) to a flat reflectance object.

 

In the next image there are the two 1st order diffraction "rainbows" of the sun, in UV. This is a proof of how narrow and limited the spectrum visible with a UV camera is.

 

f-stop: f/2.8, ISO 80, 1/30 s exposure. Filter: ZWB2 (2 mm) + chinese BG39 (2 mm).

post-284-0-94806400-1576876273.jpg

 

Here there is another mystery: beyond the yellow portion there is an orange one. It appears only in some photos and I think that light in that region should appear green, not orange.

 

f-stop: f/6.2, ISO 1600, 1/8 s exposure. Filter: ZWB2 (2 mm) + chinese BG39 (2 mm).

post-284-0-65657400-1576876593.jpg

 

And, finally, my best image of the solar spectrum. It came out very well. I think it was a 3rd order diffracion. I left the "clean" version too, so if someone wants to write on it he/she can.

 

f-stop: f/4.4, ISO 80, 1/8 s exposure. Filter: ZWB2 (2 mm) + chinese BG39 (2 mm).

post-284-0-70927600-1576877407.png

 

The same image, but with the Fraunhofer lines I could identify.

 

post-284-0-38825400-1576877461.png

 

from this I can estimate that my camera, with these filter stack, can roughly see from 365 to 400 nm.

 

I know that this post came out pretty long, but I didn't even put everything I had.

 

Anyway, this is the diffraction grating I used.

post-284-0-56602400-1576878828.jpg

post-284-0-53879700-1576874726.jpg

Link to comment
Hm. Anything past the mid-800nm mark in infrared should be monochrome. Which is what we see in your 940nm LED's case. Not sure what is going on in those color ones, although perhaps it's just tinted monochrome in that region due to the choice of white balance.
Link to comment

I don't have experience with ink filters, but white balance might be the issue with the IR colors.

Here is a grating/Sparticle comparison in with UV. Sorry, no lines show up in this grating.

 

post-87-0-18917200-1576884026.jpg

 

post-87-0-66161100-1576884046.jpg

Link to comment

Hm. Anything past the mid-800nm mark in infrared should be monochrome. Which is what we see in your 940nm LED's case. Not sure what is going on in those color ones, although perhaps it's just tinted monochrome in that region due to the choice of white balance.

 

I forgot to mention that I took the 940 nm LED directly in B&W. There is no reason to use a custom WB, because everything is monochromatic there. Also (speaking about the same image) the only "filter" was the diffraction grating, so maybe that is what absorbs at that specific wavelength. I remember someone publishing the measured transmission spectrum of my same diffraction grating, but I may be wrong.

Link to comment

I don't have experience with ink filters, but white balance might be the issue with the IR colors.

Here is a grating/Sparticle comparison in with UV. Sorry, no lines show up in this grating.

 

post-87-0-18917200-1576884026.jpg

 

post-87-0-66161100-1576884046.jpg

 

Mine spectrum is very similar to the last one. It's interesting how that region can appear either green or orange. I will probably experiment with a 340 nm LED in the future (they aren't very efficient, but can be found pretty easily). I expect it to appear green.

Link to comment

I forgot to mention that I took the 940 nm LED directly in B&W. There is no reason to use a custom WB, because everything is monochromatic there. Also (speaking about the same image) the only "filter" was the diffraction grating, so maybe that is what absorbs at that specific wavelength. I remember someone publishing the measured transmission spectrum of my same diffraction grating, but I may be wrong.

The light source probably has a gas inside and that is making the absorption line. I doubt it's the diffraction grating. (You also have the LENS absorbing too, don't forget, but that wouldn't make a line spectrum.)

Link to comment

Incandescent tungsten bulbs are filled with mostly argon gas and nitrogen.

Nitrogen does have lines at 938 and 939.

That could be what you are seeing.

Argon has a strong line at 912nm.

 

Tungsten and lead, also present in your bulb don't have any strong lines in that region. So you maybe overloading the bulb and seeing something from the Nitrogen.

 

Link to comment

Incandescent tungsten bulbs are filled with mostly argon gas and nitrogen.

Nitrogen does have lines at 938 and 939.

That could be what you are seeing.

Argon has a strong line at 912nm.

 

Tungsten and lead, also present in your bulb don't have any strong lines in that region. So you maybe overloading the bulb and seeing something from the Nitrogen.

 

I did an experiment that proves that gases in the bulb are not the responsible for the absorption line (in this particular case). I took a photo of the spectrum of both the incandescent light and the 940 nm LED at the same time, and there is the same exact line in both. So the only "candidates" here are the diffraction grating, the camera lens glass and the air. The atmosphere has a set of very strong lines in the near infrared, mostly caused by water vapour and oxygen, but I think that just ~1 meter of air isn't nearly enough to absorb that strongly. Still the mystery remains. I remember seeing a weak line with the naked eye (and a grating) in the yellow-orange region with a halogen car headlight bulb.

 

This is my setup. The bulb and the LED are aligned (as much as possible). Since the LED requires current limiting, I had to use a power supply with adjustable voltage and current for it, and so the bulb was lit up by two 12 V lead-acid batteries in series (it wasn't bright enough with just 12 V).

post-284-0-55591300-1577032609.jpg

 

I also used this homemade spectrometer, with which I saw the Fraunhofer lines in the visible spectrum (and even the calcium doublet in the 390 nm range).

post-284-0-06352200-1577033002.jpg

post-284-0-18389900-1577033010.jpg

 

It has a fairly narrow slit on the front made with black tape.

post-284-0-48819200-1577033022.jpg

 

...and the results I got.

 

Camera: full spectrum Panasonic DMC-F3, f-stop: f/2.8, ISO 1600, 1/8 s exposure, no filters.

post-284-0-30684900-1577033142.jpg

 

and a better version, same settings as before.

post-284-0-65272900-1577033402.jpg

Link to comment

Hi Stefano

Where do you get 340 nm LED's Please ?

Thorlabs sells one rated at 60 mW typical output power, but you can also find them on eBay. I never bought them, but probably they eBay ones are real. If you want there are LEDs with an even shorter wavelength, and you can find them all the way down to 265 nm.

Still Thorlabs sells LEDs at 250 nm, and although not commercially available (from what I know) researchers successfully made a crazy 210 nm LED with AlN (aluminum nitride). Keep in mind that UV LEDs with wavelengths shorter than 365 nm are quite inefficient, and below 265 nm they are very inefficient and weak.

Link to comment
Here is a grating/Sparticle comparison in with UV. Sorry, no lines show up in this grating. post-87-0-18917200-1576884026.jpg post-87-0-66161100-1576884046.jpg

Here is a clear display of the IR-leakage of the BaaderU. Nice!

 

I think the reason you do not see the lines is that you do not have a slit limiting the light source angle.

The sun's angular diameter is ca 0.53° and I think the slit Stefano is using create a smaller angle.

Link to comment

Yes... but I wonder if that might be related to the dichroic nature of the filter, rather than the 900nm range OD?

Right, I wasn't using a slit.

 

Baader doesn't have any visual leak, not until up around 800nm (ignore 460, it leaks, got rid of it). Sorry, not filter in the to define or separate the 900nm range.

post-87-0-11870900-1577080131.jpg

Link to comment

Hm. Anything past the mid-800nm mark in infrared should be monochrome. Which is what we see in your 940nm LED's case. Not sure what is going on in those color ones, although perhaps it's just tinted monochrome in that region due to the choice of white balance.

Yes, I am convinced the purple in the NIR-end is due to the white balance of all of the spectra.

I seldom bother to white balance in the camera as I only shoot RAW.

The preview-images of my 850nm all have a purple tint with the WB-setting (daylight?) happen to be.

Link to comment

Yes... but I wonder if that might be related to the dichroic nature of the filter, rather than the 900nm range OD?

 

Your'e right. I wasn't thinking it through and was wrong. Thanks for making me rethink.

Your inmages do not include the NIR, they are focussed on the UV and the ghost trails of the ones from BaaderU must sit in the blue or green range where we know the filter have a good OD

Link to comment

I did a last test. I have a chinese USB "32 Megapixel" camera (640x480), with plastic lenses and a sensor which is like 1x2 mm in size. Of course I modified it to be able to see IR and UV. I did the same thing, imaging the spectrum from the same 940 nm LED.

 

Everything is unknown, including the camera model.

post-284-0-60318700-1577085521.jpg

 

As you can see, there is no line. Since I used the same diffraction grating, it means that my camera lenses, somehow, are absorbing in that line. Probably they have a coating or something like that. Very strange.

Link to comment

Here is a clear display of the IR-leakage of the BaaderU. Nice!

 

I think the reason you do not see the lines is that you do not have a slit limiting the light source angle.

The sun's angular diameter is ca 0.53° and I think the slit Stefano is using create a smaller angle.

The narrower your light source, the higher the resolution of your spectrum is. The sun is small in the sky, but not small enough to see the spectral lines. To see the Fraunhofer lines you have to use a narrow slice of your light source, like spectrometers do.

Link to comment

I wanted to add something - it's possible increase the size of this diffraction grating - if necessary for some purpose.

I tried to put this grating on the lens - removed the cardboard, then washed the glue (I was wearing gloves), wiped it with alcohol, it gets a size of 5 cm, square then. It seems not damaged, but I was very careful. Then I made a stack between two BC3 glasses, and with a tape - everything is glued to the adapter ring.

For rotation, I put the ring in the ring from the polarizing filter.

Attached are a couple of photos of how it looks, and a UV test.

 

post-242-0-44751800-1577093826.jpg

 

On the left - black light Uniel 345-400, on the right MTE 303 UV365.

Filters: S8612_2mm + Hoya360_2mm, both shots are pure UV.

 

post-242-0-52625900-1577093854.jpg

 

My UV gradient of this diffraction grating for S8612_2mm + Hoya360_2mm looks like this.

 

post-242-0-22604700-1577093960.jpg

Link to comment

I would say that the reddish color - is a leak in the infrared - about 760 nm + or close, a second-order spectrum (?).

 

However, I am puzzled by the Cadmium test, for BaaderU.. If it is possible that this very expensive filter shows any leak? I don’t know.

Anyway, I can confirm - that 800 nm can show reddish color on the diffraction grating, after setting the white balance. [upd] I don't mean visual red of course.

Link to comment

Red doesn't always mean visual red, it just means that the Bayer filters are transmitting more red than blue and green.

At about 850nm the Bayer filters transmit a more even amount of red/blue/green.

Any 830, 800, 780, or below all transmit below their 50% point well into the lower 700nm range,

You will see red in such gratings, meaning you are seeing more Bayer red from IR than Bayer IR from blue and green.

 

Keep in mind that 1000 filters will transmit some down into the 700 range.

Link to comment

However, I am puzzled by the Cadmium test, for BaaderU.. If it is possible that this very expensive filter shows any leak? I don’t know.

 

The Baader U has some leakage.

It sometimes can be visible against an otherwise UV-black background if the motif is reflecting NIR well.

 

I played around a bit with my measurement results of my Baader U (new type), in the Schott filter calculator:

https://www.ultravio...dpost__p__28458

Link to comment

Yes the BaaderU filter differently has an IR leak. Not a big problem normally.

But you can push it with a hallogen bulb, which has little UV and much more IR. Also use a camera which is very IR sensitive like a Sigma SD14 and you will see IR, not UV.

Link to comment

Hello,

UlfW, thank you very much, very interesting to see it!

 

But - In any case in the photo with Baader - this can not be any error of the white balance.

Because this lens doesn’t pass UV in this part of the gradient, is not it?

 

Front 'yellow' UV part, there cannot be any UV part, green or orange or any 'color' of UV part - there should be a 'black hole' only.

So this 'orange' in the gradient - only if any IR leak.

 

Ok, but then, why if it is IR, why it is in front of the UV part? Then it should be test photos a spectrum 2-order -

in the test photo from Cadmium, and from Stefano also? Unlikely..

No, I didn't understand this gradient with leak. ¯\_(:-|)_/¯

Link to comment

Hello,

UlfW, thank you very much, very interesting to see it!

 

 

No, I didn't understand this gradient with leak. ¯\_(:-|)_/¯

 

Yes it is quite strange.

The surface of the BaaderU is like a coloured mirror.

Could it be reflections from the main pattern that are bouncing between filter and the lens's front element?

Link to comment

What lens is being used in the Baader U grating showing orange on the yellow side?

See my gratings show orange also, on the yellow side, when using the two poor transmitting UV lenses, but no orange when using the good transmitting Kuribayashi lens.

Not sure about that.

I don't think the light on the blue side in the Baader grating shots is from IR, but it could be isolated perhaps.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...