APOD: Hubble's Messier 5 (2015 Jun 20)

[Chris Peterson]

Well, they are only visible in the red channel, which makes me think they are out-of-focus reflections of the bright stars (donut shaped because that's what the aperture looks like) visible only with the 814W filter. Possibly the AR coating on that filter is less efficient in the IR? The offset could be caused by the tilt of the filter, either incidental or by design. In any case, they're clearly imaging artifacts.

Thanks Chris :). It was their donut shapes that particularly caught my interest. Mmmmm, I fancy a :doughnut: or two now.

We had a discussion about these donut artifacts the last time this image was run. I have looked all through the raw data to find out where they came from but never quite figured it out. If I had to guess I would say it came about by combining two F814W filters because all of the exposures have a lot of charge bleeds in them. So what you can do is combine a couple of them taken while the telescope was at different orientations and fill in the white streaks with real data from a second dataset. Somehow, some way, those annuli ended up there. I can't ask the image processor because ESA never credits them, much to my annoyance.

For the HST (or any telescope with a circular aperture and a circular central obstruction), donuts are the product of out-of-focus point sources. Dark donuts are usually dust shadows, light donuts are usually internal reflections.

In this case, I think the donuts are likely reflections off one of the 814W filter surfaces. I can't think of any way that using more than one image could cause them. See this reference for a discussion of the different surfaces in the WFC3 (including filters) that can produce these kinds of ghosts.

[DavidLeodis]

Thanks geckzilla (and Chris again) for your help . I had not thought to look up a previous use of the image, which I have now found was the APOD of April 25 2014 in the discussion of which there are comments about the image artefacts.

[geckzilla]

For the HST (or any telescope with a circular aperture and a circular central obstruction), donuts are the product of out-of-focus point sources. Dark donuts are usually dust shadows, light donuts are usually internal reflections.

In this case, I think the donuts are likely reflections off one of the 814W filter surfaces. I can't think of any way that using more than one image could cause them. See this reference for a discussion of the different surfaces in the WFC3 (including filters) that can produce these kinds of ghosts.

Yeah, but the thing is that I can't find the artifacts in the raw data. Either I'm looking at the wrong data or they were introduced. I've seen a lot of these before and none look quite like these ones. Not to say these can't be ghosts like the ones in the reference, but I'm leaning slightly toward them being introduced, at least partially, by some sort of post processing method.

[Chris Peterson]

Yeah, but the thing is that I can't find the artifacts in the raw data. Either I'm looking at the wrong data or they were introduced. I've seen a lot of these before and none look quite like these ones. Not to say these can't be ghosts like the ones in the reference, but I'm leaning slightly toward them being introduced, at least partially, by some sort of post processing method.

That's interesting. How sure are you that you have the correct raw data? I've had problems before tracking down things in the data archive. I wish that released images like this included that information. It's not always obvious.

[geckzilla]

Yeah, but the thing is that I can't find the artifacts in the raw data. Either I'm looking at the wrong data or they were introduced. I've seen a lot of these before and none look quite like these ones. Not to say these can't be ghosts like the ones in the reference, but I'm leaning slightly toward them being introduced, at least partially, by some sort of post processing method.

That's interesting. How sure are you that you have the correct raw data? I've had problems before tracking down things in the data archive. I wish that released images like this included that information. It's not always obvious.

I have looked at all of the data that is available in the HLA for F814W at the time of this image release (21 April 2014). If I recall correctly, I even downloaded the FITS files just to make sure it wasn't somehow invisible in the interactive previews. Another thing which makes me suspicious of these is that I am very familiar with Hubble's artifacts and I know these are a little strange. Impossibly strange? No, I could still be wrong. Another possibility is that data from another observatory was used to fill in the places where the charge bleeds were. That would help to explain their strange appearance and and absence from Hubble's raw data. It's a little sloppier than usual of them not to mention it, but I wouldn't say it's entirely uncharacteristic of these "pretty picture" image releases which aren't very important in a broader sense. My statements are very speculative. What I would really like to do is contact the person who processed the image. I might know who did it or could at least point me to who did it, so I can try asking them again.

[geckzilla]

Finally got an answer. An unlisted filter was used to compose the image. The ghosts are found in some narrowband data, this time F656N. No wonder it doesn't look anything like the filter ghosts from any broadband filter. I've learned something new and that is that filter ghosts for narrowband filters are more like lines which form circles while ghosts for broadband filters tend to be much more diffuse. Nature was supplying the clue all along.

[lbeckham]

Clusters demonstrate Newtonian physics. And you need nothing more than Newtonian physics to numerically model them, and the models behave just like what we observe.

So, "dark matter" (what the hell THAT is) has a no effect on globular clusters all?

[Ann]

Clusters demonstrate Newtonian physics. And you need nothing more than Newtonian physics to numerically model them, and the models behave just like what we observe.

So, "dark matter" (what the hell THAT is) has a no effect on globular clusters all?

Globular clusters are not generally thought to contain much dark matter. See this paper about NGC 2419, a distant Milky Way globular cluster that was the subject of an APOD recently.

But unexpectedly, astronomers have found what appears to be large quantities of dark matter in some of the globulars of Centaurus A, NGC 5128, recently.

Ann