Back to forums

Green stars...!?

Processing Deep Sky M4 Messier Objects Globular Cluster
13 replies•536 views
Willem Jan Drijfhout avatar
Willem Jan Drijfhout avatar
While processing M4, something strange happened. There appear to be two distinctly green stars. I've tried different color calibration technique, but all give the same result. Also looking at signal in individual R, G and B stacks, it is clear that these specific stars are much greener than anything else in the image.

As far as I know, green stars don't exist, so it is probably some kind of error. But where and how was that error made? Does anyone have any idea?

Well Written Engaging
Lorenzo Siciliano avatar
Lorenzo Siciliano avatar
Check out if they are pulsating stars. SX phe type stars in globular clusters seem to have a green tint when you do an RGB integration.
Ciao
Lorenzo
Victor Van Puyenbroeck avatar
Victor Van Puyenbroeck avatar
Both stars are listed as RR Lyrae Variables in SIMBAD:

NGC 6121 SAW V16
NGC 6121 SAW V9
Willem Jan Drijfhout avatar
Willem Jan Drijfhout avatar
Thank you both! So it does not seem to be an error, but these are variable stars with short periods.
Well Written Respectful Supportive
Oscar avatar
Oscar avatar
wow, didn't know green stars existed
Habib Sekha avatar
Habib Sekha avatar
@Willem Jan Drijfhout , your question reminded me of a recent post on a Dutch forum.

Apparently also a ‘green’ star (albeit a different object) and he came up with an explanation, which might or might not be plausible.

It would be the text starting after the second picture of the first post in:

https://www.astroforum.nl/threads/sh2-104-en-een-mysterieus-voorwerp.1479976/
Victor Van Puyenbroeck avatar
Victor Van Puyenbroeck avatar
That's my post on the Dutch forum What I got from literature is that B[e] stars have an abnormal amount of H-alpha. Since I used narrowband data in an SHO color map, the excess H-alpha from a B[e] star made it appear green in such a color scheme. I assume B[e] stars would turn very red in a normal RGB image, but I'm not sure about that.


I checked the Hubble Space Telescope archive for this field. There was data available for 3 filters:
  • f685n (H-alpha filter, approx 655 - 662 nm)
  • f625w (Sloan Digital Sky Survey r filter, approx 555 to 700 nm)
  • f435w (Johnson B filter, approx 380 to 480 nm)

There is no obvious visual excess of H-alpha on these 2 green stars. They also don't appear to be double stars with different color temperatures. 


However, if I increase the color saturation of the original image you can see both green and magenta stars. Both of these colors are normally not possible. Could there be some form of interference or UV/IR leak in the filters?


Helpful Insightful Engaging
Thomas W avatar
Thomas W avatar
Did you continually loop filters like RGBRGB… or split them in to large blocks like 1hr R, 1hr G, 1hr B?
Well Written Engaging
Oscar avatar
Oscar avatar
I thought I might copy and paste this info I found on Quora:
…"'It is very remotely possible (but almost impossible) that a star could have vaporized copper(II) in its photosphere, but since copper is more dense that the hydrogen, helium and other gases likely to make up the photosphere, it would sink into the star.'

'But it gets worse. The photosphere is a plasma with free electrons. Plasmas tend to act like blackbodies (below a certain frequency) and so it would be extremely unlikely to have a color which is not a blackbody color even if ionized copper were in the photosphere. You would need to have a cooler gas above the photosphere that absorbed blue and red light to leave green light to pass through.'"

I remember burning copper wire before and seeing that greenish color, if that can be related at all to this phenomenon

I'm not sure if this is helpful, but I also remember clearly seeing green and purple stars in my SHO mapped images
Willem Jan Drijfhout avatar
Willem Jan Drijfhout avatar
Did you continually loop filters like RGBRGB... or split them in to large blocks like 1hr R, 1hr G, 1hr B?

In principle in blocks, but in this case, frames were collected over multiple days, so some days only one or two of each colour had been collected.
Willem Jan Drijfhout avatar
Willem Jan Drijfhout avatar
Victor Van Puyenbroeck:
That's my post on the Dutch forum What I got from literature is that B[e] stars have an abnormal amount of H-alpha. Since I used narrowband data in an SHO color map, the excess H-alpha from a B[e] star made it appear green in such a color scheme. I assume B[e] stars would turn very red in a normal RGB image, but I'm not sure about that.


I checked the Hubble Space Telescope archive for this field. There was data available for 3 filters:
  • f685n (H-alpha filter, approx 655 - 662 nm)
  • f625w (Sloan Digital Sky Survey r filter, approx 555 to 700 nm)
  • f435w (Johnson B filter, approx 380 to 480 nm)

There is no obvious visual excess of H-alpha on these 2 green stars. They also don't appear to be double stars with different color temperatures. 


However, if I increase the color saturation of the original image you can see both green and magenta stars. Both of these colors are normally not possible. Could there be some form of interference or UV/IR leak in the filters?



Yes, I too was of the impression that the answer was some kind of error on my or the systems behalf.

One thing to note is that this target was very low on the horizon, so all frames at an altitude between 22 and 26°. And several blue frames at the lower end of that range.
Total frames is 9 for each red and green, and 12 for blue. So the magenta cast at some stars would not ring too many alarm bells immediately. 

What would you mean with interference or UV/IR leak? Are you thinking about a filter not blocking the whole aperture of the holder in the filterwheel? Or some kind of coating degradation that would allow UV/IR through? And if so, why would that affect individual stars, rather than give a colourcast as a whole?
The filters used are a set of 2" unmounted LRGB filters from Chrome.
Helpful Engaging
Thomas W avatar
Thomas W avatar
Willem Jan Drijfhout:
In principle in blocks, but in this case, frames were collected over multiple days, so some days only one or two of each colour had been collected.


Ah I see, the fact that this was collected over multiple days complicated things but I reckon variability may be the cause. Like Victor pointed out, both those stars are RR Lyrae type variables, which vary a lot in amplitude (V band on the order of 0.5 to 1 mag) with short periods of roughly half a day. My guess is that the time/s when the green channel was taken happened to coincide with when those variable stars were near their brightest. Other variables which peaked when the red or blue channels were taken would appear redder or bluer, though as this puts them near to the blackbody colours of non variable stars, they do not stand out as much. Green on the other hand stands out like a sore thumb!

I wanted to demonstrate this last year and shot M3 with just an L filter. This globular cluster contains lots of RR Lyrae variables. The first two hours L were assigned as the red channel, next two the green and final two blue. By composing this as an RGB image, the hue of the RR Lyrae variables corresponds to the point in time when brightness peaked.

It may be an instrumental or processing issue, however for me this is the most likely cause. 

Well Written Helpful Insightful Engaging Supportive
Victor Van Puyenbroeck avatar
Victor Van Puyenbroeck avatar
Willem Jan Drijfhout:

What would you mean with interference or UV/IR leak? Are you thinking about a filter not blocking the whole aperture of the holder in the filterwheel? Or some kind of coating degradation that would allow UV/IR through? And if so, why would that affect individual stars, rather than give a colourcast as a whole?
The filters used are a set of 2" unmounted LRGB filters from Chrome.

I was trying to think of factors that would change how your filters capture the blackbody radiation profile. The explanation that Tom presented above makes much more sense for variable stars.
Well Written
Willem Jan Drijfhout avatar
Willem Jan Drijfhout avatar
Willem Jan Drijfhout:
In principle in blocks, but in this case, frames were collected over multiple days, so some days only one or two of each colour had been collected.


Ah I see, the fact that this was collected over multiple days complicated things but I reckon variability may be the cause. Like Victor pointed out, both those stars are RR Lyrae type variables, which vary a lot in amplitude (V band on the order of 0.5 to 1 mag) with short periods of roughly half a day. My guess is that the time/s when the green channel was taken happened to coincide with when those variable stars were near their brightest. Other variables which peaked when the red or blue channels were taken would appear redder or bluer, though as this puts them near to the blackbody colours of non variable stars, they do not stand out as much. Green on the other hand stands out like a sore thumb!

I wanted to demonstrate this last year and shot M3 with just an L filter. This globular cluster contains lots of RR Lyrae variables. The first two hours L were assigned as the red channel, next two the green and final two blue. By composing this as an RGB image, the hue of the RR Lyrae variables corresponds to the point in time when brightness peaked.

It may be an instrumental or processing issue, however for me this is the most likely cause. 


This sounds like a very plausible explanation. Thank you!
Interesting experiment you did there with M3, that makes explains it in a very visible way.