IR, visible and UV of small blow torch flame

[JMC]

Inspired by GaryR's work on the Invisible Fire (http://www.ultraviol...dpost__p__26828), I thought I would try imaging the flame from the small blowtorch I have, using different filters and cameras.

 

Parameters were as follows;

Rayfact 105mm f4.5 lens (at f11), ISO800, used for all images.

All images shot in a dark room, with a black background.

All images shot as RAW, and loaded into RawDigger as RAW composite images. Then screenshotted and cropped. No further modification.

 

Physical setup as shown below;

 

Visible image - Longpass 420nm from Dr Klaus Schmitt, Normal EOS 5DSR, 2s exposure.

IR image - Hoya R72 filter, Multispectral converted EOS 5DSR, 2s exposure

UV 1 - Baader U filter, Multispectral converted EOS 5DSR, 30s exposure

UV 2 - 330nm band pass UV filter from Dr Klaus Schmitt, Multispectral 5DSR, 30s exposure

UV 3 - 308nm Invisible Vision filter, Monochrome converted EOS 5DSR, 30s exposure

 

Visible image

 

IR image

 

UV 1 (Baader U)

 

UV 2 (KS 330nm band pass)

 

UV 3 (Invisible vision 308nm, monochrome camera)

 

I'm definitely seeing light from the UV through to the IR, and even down at the 308nm end of the UV. This flame would be much hotter than outdoor fire in GaryR's work though, and I was able to eliminate ambient light.

 

I find it interesting that the inner core of the flame looks much brighter than the surrounding part of the flame, in the visible image, but also the 308nm one. No idea why that would be the case. The blow torch was set to the same setting each time, although I did turn it on and off while changing filters/cameras.

[Cadmium]

Did you try any with the lights on, other than the visible shot?

[GaryR]

Hi Jonathan,

You're definitely burning hotter than my outdoor fire pit. The 308nm image is quite interesting with it's bright core. I'm wondering if that part of the flame is emitting UV in that specific range.

 

Arc welding radiation would likely give you even brighter captures with all those filters.

[JMC]

Steve, no all the images of the flames were done with the external lights off, and in a dark room. I wanted the only light to be from the flame.

 

Gary, yes, the 308nm is intriguing. The core is the hottest part, so I presume that's most likely to have UV. I wonder if I dare getting the spectrometer near enough to see what is there?

[Cadmium]

Just wondering, of course, if the lights would make the UV flame shots invisible, as with Gary's outdoor fire shots...

[dabateman]

Jonathan,

I wonder if the 308nm filter is leaking IR, masking the true outer UV edge of your flame. You may want to stack it with an R72 or whatever IR filter you have to see if you are only seeing IR leak.

 

I am reminded of these threads:

http://www.ultravioletphotography.com/content/index.php/topic/2479-candlelight/page__p__18213__hl__candle__fromsearch__1#entry18213

http://photographyoftheinvisibleworld.blogspot.com/2016/12/deep-uv-ultraviolet-reflected-light.html?m=1

 

The IR output seems to be orders more than the UV, which will really challenge the IR suppression of our filters.

 

 

[ulf]

I wonder if I dare getting the spectrometer near enough to see what is there?

Small collimators give a beam only a few mm wide.

IF mounted on a suitable fiber the spectrometer will be quite safe.

[Andy Perrin]

 

Small collimators give a beam only a few mm wide.

IF mounted on a suitable fiber the spectrometer will be quite safe.

Which reminds me of a question I had— don’t fibers only admit certain modes/wavelengths that depend on their diameter? Is that only an issue for you guys with really thin fibers or do you get a cutoff wavelength that’s in the range you can measure with the spectrometer?

[JMC]

Ooh, loads of comments this morning.

 

Steve. I hadn't done shots with the lights on. However room lights are LED ceiling lights, and there are no windows. I doubt there is any ambient UV from the room lights. I can check though.

 

David. Good call. Will try that.

 

Ulf, Andy. Yes I can try using a collimator. Will see if I can make it work (safely). Using a blow torch in a dark room with the lights off, and expensive, delicate, equipment everywhere is always 'character building'. Not heard about that issue with the fibers. Mine is 600 microns diameter, and is good for measurements down to 250nm and up to 850nm which is the limits of my spectrometer.

 

EDIT - update with pictures.

 

Steve, with the lights on I can still see the UV pictures. If there is UV in the ambient light it is not high enough to impact the images. For example here's a picture using the Baader U, multispectral 5DSR camera, Rayfact f11, ISO800, treated the same way as above.

 

David, done a few new images for the 308nm Invisible vision filter. Same settings as original images above.

 

Firstly, Invisible Vision 308nm. Room lights switched on.

 

Secondly, Invisible Vision 308nm + R72, in the dark (room lights off)

 

And finally, Invisible Vision 308nm + R72, with the room lights switched on.

 

With the room lights switched on I can still see the flame, as if the room lights were off. With the IV308nm filter and the R72 filter stacked, with either the room lights on or off, I cannot see the flame. I therefore think the IV308nm filter is itself effectively blocking the IR in this circumstance.

[dabateman]

Wow, very cool. You have a strong UVB source there. Much stronger than the candle flame threads. Too bad you don't have any UVc filters.

So if I want to test the limits of my camera again with my 260 and 280 bp filters, it looks like I need a real torch. Not the British kind ;p

 

[ulf]

Which reminds me of a question I had— don’t fibers only admit certain modes/wavelengths that depend on their diameter? Is that only an issue for you guys with really thin fibers or do you get a cutoff wavelength that’s in the range you can measure with the spectrometer?

I think this only is an issue for long and thin fibers mostly used for data distribution.

Naturally there is some transmission variations in fibers for spectrometers too, but not that much.

They come in different materials, often optimised for UV-VIS or VIS-IR and with a fiber diameter of 100-1000µm.

Most of the ones I use are of UV-VIS type and with a 400µm core

[JMC]

David. Yes I'll get the spectrometer out again when I have a few minutes, and see what is being given out by it. Even with lower wavelength filters my cameras are only good down to about 300nm, as they have Schott WG280 over the sensors, so 308nm is about as low as I can reliably go.

[Cadmium]

I thought the inspiration for this was to explore the 'missing flames' in Gary's UV shots of his outdoor fire.

[JMC]

I thought the inspiration for this was to explore the 'missing flames' in Gary's UV shots of his outdoor fire.

Nope, inspiration for me was imaging a flame and seeing what was there. I shall leave the missing flame to Gary.

[ulf]

The emissions are dropping fast for shorter wavelengths.

 

Assuming the light come from glowing carbon particles emitting a black body radiation and reaching the max temperature of 1977°C, referring to the table at this site.

http://www.derose.ne.../flametemp.html

https://astrogeology...nce-calculator/

[JMC]

Got some spectra from the flame, using my Ocean Optics FX, set up for Irradiance. Measured at about 1cm from the flame with a small collimator and with a 1s capture time. There are 4 lines on the graph - 2 from the 'core region' at the heart of the flame, 1 from the sheath of the flame between 5mm and 1cm from the front tip of the core, and a baseline with the torch switched off.

 

Two plots - firstly, full scale of irradiance.

 

Secondly, same plot but rescaled to show more detail below 700nm.

 

There's definitely some complex emission going on in the core of the flame even down into the UVB, and it explains why there is such a bright 308nm image. Also explains why there is much less difference between the core and sheath in the UVA region, as there isn't much going on between 380nm and 350nm. The sheath though is relatively devoid of emission structure.

 

Very sharp double peak in both parts of the flame in the IR, at 765nm and 769nm, only a few nm width.

[Andy Perrin]

Hah, yeah no way is that blackbody radiation. That’s a line spectrum from the gases. In a blowtorch I don’t think you have any carbon particles (like the soot in a candle flame). It’s just straight gas combustion converting one gas to other gases.

[Alaun]

You can look up the spectra at the NIST database: https://www.nist.gov/pml/atomic-spectra-database

[dabateman]

Jonathan,

That is one cool strong peak between 300nm and 320nm. Looks like you found a great UVb source. Do you know what the fuel is?

 

I think you may have a plasma being formed, and your ionizing some metal impurities in the fuel causing that peak. Just not sure what it would be, maybe iron? Or some other elemental impurity.

 

Out of curiosity I tested my plasma plate. Similar to a plasma ball. I see a green dot at the center using a 313nm filter and weakish UV with a Baader venus U filter.

But mine has thick glass. If anyone has a plasma ball, might be fun to image.

[Andy Perrin]

I think you may have a plasma being formed, and your ionizing some metal impurities in the fuel causing that peak. Just not sure what it would be, maybe iron? Or some other elemental impurity.

If this is just a simple butane or propane torch, it's really unlikely to contain any metals. I don't know why any would be needed to make that line anyhow. It looks like most of the lines, including the spike at 308nm (although the intensity seems much lower?) are on this graph:

https://upload.wikim...e_flame.svg.png

 

Based on the above, it would seem to be the OH line.

 

Jonathan's lines are at 308, 429, 515, and 766nm. The OH line is supposed to be 306.357 nm.

[dabateman]

 

If this is just a simple butane or propane torch, it's really unlikely to contain any metals. I don't know why any would be needed to make that line anyhow.

 

Too often, my first gut reactions come from past experiences. I have seen lots of inductivly coupled plasma (ICP) ionization emmission, which takes things to the elemental level. But you may have it. An OH peak makes much more sense.

[ulf]

Too often, my first gut reactions come from past experiences.

 

Mine too. I have a friend that did a dissertation about combustion and soot creation processes in flames and have read some parts of the document some years ago. Not my field at all.

[Cadmium]

I use to have cats that liked to get under the big burners of my kiln when it was firing.

[JMC]

This is a butane gas torch, filled from a little butane aerosol can. As for why the 'OH' peak is high on mine, the spectrometer is setup for irradiance measures, which corrects for the drop off in sensitivity at the extremes of the measurement range. A lot of the spectra you see when it just has the axis labelled as 'intensity' have not had this correction, which would explain why the OH peak looks small on some. Also, I doubt the butane in that aerosol is particularly pure - a bit of water in there with it would lead to a bigger OH peak.

 

Steve, when I was growing up we had coal fires at home, and one of the fires had a back boiler for the water. In the morning our cat would climb up inside the chimney and sleep on the ledge on top of the boiler because it was nice and warm. We had to check the chimney every day before lighting the fire, to make sure that cat wasn't up there.

[colinbm]

JMC, interesting work......In a previous life I did gas welding as part of my trade.

The hottest part of the flame is not the centre core, but just outside the centre core at its tip.

The inner cone is oxygen.

You can test this by placing a small piece of wood or bare match-stick across the flame & you will see that the two outside edges burn first & leave the centre un-burnt the longest.