Long wave UV imaging of galaxies is tedious due to the very long exposures required to get a good S/N ratio.
I have estimated a 70 hr exposure in UV will have a similar S/N ratio to a standard 1.5 luminance image for Centaurus A.
Rather than spending the rest of my life imaging Centaurus A I took a short cut by doing a 10 hr exposure and using this as a B channel image with existing L, R and G data.
A comparison between a conventional and UV image of Centaurus A is in the attachment with the standard image on the left.
The blue/bluish areas that are highlighted in UV B-channel image are regions of intense star formation.
What makes this image particularly interesting (for me at least) is that the star formation extends into the disk which is not seen in conventional visible images.
Perhaps Madbadgalaxyman has some comments.
I think this technique might be particularly useful in bringing out details in starburst galaxies such as NGC 1808.
Good stuff Steven. Gee that's a lot of integration time. Is that related to the low QE of the camera in that wavelength range or lack of signal from the target to start with?
Good stuff Steven. Gee that's a lot of integration time. Is that related to the low QE of the camera in that wavelength range or lack of signal from the target to start with?
Marc,
Low QE is the main problem.
Other issues is the type of optical system used and atmospheric conditions.
Refractors or reflectors that use corrector lenses will further reduce performance.
Long wave UV is much more sensitive to atmospheric extinction and reduced transparency than the visible spectrum.
Sounds very challenging Steven & unique seeing the star birthing areas you point out, now im curious of the yellow/green signal and its interpretation or reference of ? Amaizing stuff
Do you bin the camera 2x2 or even 3x3 to speed up the process? It doesn't look like it has a lot of detail in the UV so no point in not using binned?
Greg.
Greg,
Since my original intention was to use the UV images for luminance they are unbinned as I would never sacrifice resolution on luminance images.
Quote:
Originally Posted by astronobob
Sounds very challenging Steven & unique seeing the star birthing areas you point out, now im curious of the yellow/green signal and its interpretation or reference of ? Amaizing stuff
Hello Bob,
Since the UV data has been mapped to the Blue channel, hot stars emit more UV light than cooler stars hence the hotter the star the more blue it appears. Conversely the cooler the star the more yellow it appears in the image. If you look at the high resolution images for the visible light and UV images you will also notice that for the foreground stars, the redder the star in the visible light image, the yellower it appears in the UV image.
In effect the image is showing the blackbody radiation effect of temperature versus colour.
The exception is the red region at the centre of galaxy which is caused by the scattering of light by dust rather than blackbody radiation.
Incidentally that is a fine Centaurus image you have produced.
As you can see it requires far more integration time than the ten hours spent on the object.
Of course it throws a spanner into the works with my reply to Bob when I suggested the yellow regions were due to less UV radiation when clearly the L image shows otherwise.
It supports my theory that combining monochrome images to form colour is more voodoo than science.
I'll try merging the UV data with existing B data to see if I can come up with something more realistic.
Quote:
Originally Posted by Larryp
Interesting comparison, Steven, and both great images
Nice.
It might be interesting to do an image shifting all of the colours. Use U for the blue channel and B for the green channel and the G data for the red channel. This would just shift all the colours to the blue and then compare that to the usual tricolour image.
Great work Steven, really interesting stuff here! A lot of UV showing in and around the dust lanes, I guess showing star formation at the "shock front" of the colliding galaxies.
As you can see it requires far more integration time than the ten hours spent on the object.
Of course it throws a spanner into the works with my reply to Bob when I suggested the yellow regions were due to less UV radiation when clearly the L image shows otherwise.
It supports my theory that combining monochrome images to form colour is more voodoo than science.
I'll try merging the UV data with existing B data to see if I can come up with something more realistic.
Thanks Larry.
Regards
Steven
Thank you Steven, that is one interesting looking frame, very cool.
It actually seems to confirm my result from a while back where I tried to create a synthetic UV image for the Carina nebula, by subtracting L and NIR from a Clear frame. The results were pathetic to say the least and I just discarded it as unworkable. But based on your results I think it might have worked, the S/N was just so low that I thought it wasn't right.
I might give it a try again, given that I don't have a UV filter.
Great work Steven, really interesting stuff here! A lot of UV showing in and around the dust lanes, I guess showing star formation at the "shock front" of the colliding galaxies.
Interesting idea, David.
The idea that the peculiarity of the dust lane is in part due to some sort of shock has not really been considered in the professional literature.
In general, the majority of the professional astronomers who have studied this galaxy tend to think of the dust lane as being the visual manifestation of a somewhat-settled disk structure that bisects the galaxy along its shortest axis.
(in "real 3-D space", NGC 5128 is a triaxial ellipsoid, which could be almost prolate in shape (in other words, it is slightly 'hot dog bun' shaped, in three dimensions), and the disk of gas/dust/stars that was the cannibalized galaxy occupies the shortest axis of the spheroid)
Shock waves tend to occur in those galaxy mergers where both of the progenitor galaxies have a substantial Interstellar Medium..... so most of the people who have so far modelled the accretion of the small disk galaxy by the giant old spheroid of NGC5128 tend to (perhaps naively) think that because N5128 probably didn't have much of an ISM before the merger event, the role of shocks can be ignored.
cheers,
bad galaxyman
The observed dust lane in N5128 is thought to be due to a merger with a small, low mass, spiral galaxy. However, the kinematics and stellar ages of the primary spheroidal component indicate that the entire structure of NGC 5128 originated in a very-large-scale galaxy merger event that occured previous to the current instance of galaxy merger.
Keep up the great work!
I think your results are excellent!
Well done!
Thanks Ken.
Quote:
Originally Posted by Terry B
Nice.
It might be interesting to do an image shifting all of the colours. Use U for the blue channel and B for the green channel and the G data for the red channel. This would just shift all the colours to the blue and then compare that to the usual tricolour image.
Terry,
Thanks for the advice. One aspect I'll be checking is whether the colour balance has been effected by OII emissions which is the UV component of skyglow.
Quote:
Originally Posted by David Fitz-Henr
Great work Steven, really interesting stuff here! A lot of UV showing in and around the dust lanes, I guess showing star formation at the "shock front" of the colliding galaxies.
Thanks David.
Quote:
Originally Posted by SkyViking
Thank you Steven, that is one interesting looking frame, very cool.
It actually seems to confirm my result from a while back where I tried to create a synthetic UV image for the Carina nebula, by subtracting L and NIR from a Clear frame. The results were pathetic to say the least and I just discarded it as unworkable. But based on your results I think it might have worked, the S/N was just so low that I thought it wasn't right.
I might give it a try again, given that I don't have a UV filter.
Interesting concept Rolf. I wonder where the cut off point of your L filter is.
The Astrodon L filter for example cuts off at around 370nm. You might have very little data to play with.
Amongst the many pitfalls of long wavelength UV imaging is one factor I failed take into consideration, the night sky is "aglow" with single ionized oxygen (OII) emissions. While this has no effect on visual observation or even visible light imaging as the filters cut out the emission, it is a problem for UV imaging as the emission line sits in the middle of the UV pass bandwidth.
It also helps to explain why there is no contrast gain using a UV pass filter.
Due to the faintness of Centaurus A in UV, I decided to sum the images for the 10 hour exposure without realizing at the time I was adding up the background noise from the OII emission. The end result was that it completely messed up the colour balance.
The solution was to use sigma reject and normalize the background during the stacking process instead of summing.
The result is a far better colour rendition.
A comparison between the UV and visible image is in the attachment.
The advantages and disadvantages of UV astronomy are highlighted in this interesting site. http://www.astro.virginia.edu/class/...stron-f01.html
The article focusses more on Extreme, Far and Mid UV which requires satellite imaging but it summarizes the problems of long wave UV or "near UV" imaging quite well.
Quote:
"Near-UV": 3200-4000 Å. Usually considered part of "optical" band, but often compromised by poor reflectivity or transmission of optical elements, atmospheric opacity.
Gee that's a lot of integration time. Is that related to the low QE of the camera in that wavelength range or lack of signal from the target to start with?
