APOD: Hubble's Messier 5 (2014 Apr 25)

[geckzilla]

I would wager that they are image processing artifacts. They appear adjacent to all the larger redder stars and similarly oriented throughout the image. Also, I noticed that several of the blue stars along the left side of the image have a slightly offset grouping of red diffraction spikes, similar in color to the Ghost images (probably on the same color channel)

I wouldn't wager them on being introduced during processing. Processing is usually what removes artifacts, not what adds them. Also, these are common and they appear at the time of imaging. Classic example. Filter ghost and charge bleed around bright star haven't been removed for this old image. The two sets of diffraction spikes are from using both WFC3 and ACS data which were done at different times with the telescope oriented differently. Anyway, the ghosts aren't there in the combined images available in the HLA and I didn't think that such artifacts were removed. They must be there somewhere in some rawer data... or I could be totally wrong. I asked someone who should easily be able to explain it and am waiting patiently for a response.

[Anthony Barreiro]

Last week professor Graeme Smith of UC Santa Cruz and Lick Observatory gave our astronomy club a lecture on the distribution of globular clusters within the Milky Way, their ages, chemical compositions, and the types of stars found within them. The take away was that globulars are fossil remnants of the very earliest evolution of the Milky Way Galaxy. For instance, the stars in globular clusters in the galactic halo (like M5) have much lower concentrations of elements heavier than helium than do those in globulars that orbit in the galactic plane, while stars in the spiral arms have higher concentrations of heavy elements than those in any globular cluster. From this we can deduce that our galaxy must have collapsed into a spiral over a relatively short time, only about 100 million years. He explained the difference between asymptotic branch blue giants and main sequence blue stragglers, and why they give observational astronomers like him headaches. It was a fascinating lecture, and Smith very kindly stuck around after the lecture to answer questions and tell stories of observing globulars from the southern and northern hemispheres.

[Anthony Barreiro]

By the way, M5 is high in the sky after midnight these days, and will be well placed for evening viewing in June and July. You can see M5 as a smudge of light in binoculars, and it's glorious in even a small telescope.

[Chris Peterson]

I wouldn't wager them on being introduced during processing. Processing is usually what removes artifacts, not what adds them.

Well, that's certainly the goal of processing. But I'd say more artifacts are introduced by processing than by imaging.

This appears to me to be a processing artifact. My first guess would be a dark frame corrupted by a residual bulk image, just on the IR data. But there are other things you can do wrong calibrating with dark frames and flat frames that produce similar artifacts. I don't think this is an optical artifact, since that would be unlikely to result in the darker core of the ghost images.

[geckzilla]

I wouldn't wager them on being introduced during processing. Processing is usually what removes artifacts, not what adds them.

Well, that's certainly the goal of processing. But I'd say more artifacts are introduced by processing than by imaging.

This appears to me to be a processing artifact. My first guess would be a dark frame corrupted by a residual bulk image, just on the IR data. But there are other things you can do wrong calibrating with dark frames and flat frames that produce similar artifacts. I don't think this is an optical artifact, since that would be unlikely to result in the darker core of the ghost images.

Maybe it does have something to do with that. Hubble's WFC3 filter ghosts do look like donuts, though. PDF on the matter

[Chris Peterson]

Maybe it does have something to do with that. Hubble's WFC3 filter ghosts do look like donuts, though. PDF on the matter

You can get rings two ways: any out-of-focus internal reflection (ghost) will show a ring because of the obstructed aperture of the telescope, and any saturated star can result in rings during post processing if there's an alignment problem. Not sure which we have here. On close examination, the distance between the bright stars and their related artifacts isn't the same, which doesn't really make much sense for either a filter ghost or a post processing artifact.

According to the paper, the IR (814W) and V (606W) filters have about the same amount of intrinsic ghosting. Looking at the channels, it's clear that the IR is a deeper exposure, with greater star saturation. Without knowing the actual exposure times (from the FITS headers) I don't know much more than that. But the artifact is so strong in the IR channel, that if it were filter ghosting in the 0.3-0.4% range, I'd sort of expect it to show up in the V channel as well. But even when I blow up the contrast hugely, I'm not seeing anything.

Also, the ghosting shown in the paper for the 814W filter looks nothing like the ghost seen in today's image. But it's possible we're seeing the difference between a test bench result and the filter in actual operation in the HST.

[geckzilla]

Also, the ghosting shown in the paper for the 814W filter looks nothing like the ghost seen in today's image. But it's possible we're seeing the difference between a test bench result and the filter in actual operation in the HST.

That's why it's so odd to me. I've seen plenty filter ghosts which look exactly like they do in the paper but never like these. FYI, here are some real filter ghosts in a 700 second F814W exposure from a different dataset for this same object. The galaxy in this picture is the same obvious galaxy near the brightest red donut from the APOD image. (North isn't up, can't be bothered)

[raschumacher]

Some fun things in that image:
The eye loves to find patterns where there are none. Note the many apparent strings of stars.
The brightest objects have red ghosts to their left. Some funny defect in Hubble's optics.
One real thing is the distant spiral galaxy nearly edge-on faint in the background near the lower-right corner.

[BDanielMayfield]

Today (April 25) in my mailbox was a timely discovery; the June 2014 issue of Sky and Telescope magazine, wherein there is a Cover story article Observing Cepheids in M5. Cepheids are stars that vary in magnitude over time, and the article provided a link to a very interesting web page by Robert Vanderbei. It shows wider veiws of M5, but the two bottom images show blink comparisons of photos taken 11 months apart. This dramatically shows that M5 contains many variable stars.

http://www.princeton.edu/~rvdb/images/NJP/m5.html

[Ann]

Today (April 25) in my mailbox was a timely discovery; the June 2014 issue of Sky and Telescope magazine, wherein there is a Cover story article Observing Cepheids in M5. Cepheids are stars that vary in magnitude over time, and the article provided a link to a very interesting web page by Robert Vanderbei. It shows wider veiws of M5, but the two bottom images show blink comparisons of photos taken 11 months apart. This dramatically shows that M5 contains many variable stars.

http://www.princeton.edu/~rvdb/images/NJP/m5.html

Click to view full size image

Source: http://zebu.uoregon.edu/~soper/MilkyWay/cepheid.html

Interesting, Bruce. I would have used the term RR Lyrae stars for the variable stars in globulars.

The word Cepheids typically refers to variable stars that are brighter, redder, more massive, far more metal-rich and much, much younger than RR Lyrae stars. This wikipedia article talks about RR Lyrae variables.

Thank you so much for the "blink comparison image" that dramatically brings out the RR Lyrae stars of M5! Fantastic!

Ann

[BDanielMayfield]

Yes, the great majority of the variability seen in the blink comparison images is from the RR Larids. There are in fact only two very bright Cepheids in M5, but these happen to be (when they are at thier brightest) among the brightest stars in the cluster, and therefore the changing magnitudes of these two stars can be tracked with amateur sized telescopes, which is what the S&T article was about.

[Ann]

Fascinating!

http://en.wikipedia.org/wiki/Type_II_Cepheid wrote:
Type II Cepheids are variable stars which pulsate with periods typically between 1 and 50 days.[1][2] Like all Cepheid variables, Type IIs exhibit a relationship between the star's luminosity and pulsation period[3][4] Type II Cepheids are population II stars and are thus old, typically metal-poor, and low mass objects.[1]

The Cepheids in M5 are type II Cepheids, and type II Cepheids are old! I must admit that I wasn't aware of that at all.

Type II Cepheids are fainter than classical Cepheids:

Type II Cepheids are important standard candles since they obey a period-luminosity relationship,[2] although they are fainter than their classical Cepheid counterparts for a given period by about 1.6 magnitudes.

There are indeed two type II Cepheids in M5.

http://arxiv.org/pdf/1310.0594v1.pdf wrote:
A search for variable stars using the DAOMASTER variability index enabled us to recover the three known bright variables (two Cepheids and one red, long period variable, or LPV)
...
In addition to its abundant RR Lyrae stars, M5 has two well-studied type II Cepheids, V42 and V84, which are the subject of recent CCD photometry by Rabidoux et al. (2010). The irregular LPV star V50 has been known since Bailey (1917)

The paper by Katie Rabidoux et al. from 2010 gives the following values for M5 type II Cepheids V42 and V84:

V42: B-V color: 0.63, similar to the Sun. Apparent V magnitude: 11.19 (similar to the apparent magnitude of Proxima Centauri). Absolute V magnitude: -3.35. I'm extremely bad at converting absolute magnitudes to "solar luminosities", but I think that the absolute V magnitude of V42 might be close to 1,800 times that of the Sun.

V84: B-V color: 0.69, slightly yellower than the Sun. Apparent V magnitude: 11.42, a bit fainter than V42. Absolute V magnitude, -3.12. That just might be close to 1,500 times that of the Sun.

V42 and V84 are strangely bright in view of the fact that they are old metal-poor globular cluster stars, I think. After all, the best-known classical (young and relatively metal-rich) Cepheid of them all, Delta Cephei, is "only" about 1,400 times as bright as the Sun, according to my software. But classical Cepheids "ought to be" brighter than type II Cepheids.

Ann

[DavidLeodis]

I often get lost when trying to get my mind round the time scale of the Universe and each time I think I have started to understand things something will crop up that causes me to again realise that I really know nothing! In the information brought up through the 'stunning close-up view' link it states "But Messier 5 is no normal globular cluster. At 13 billion years old it is incredibly old, dating back to close to the beginning of the Universe, which is some 13.8 billion years of age. It is also one of the biggest clusters known, and at only 24 500 light-years away, it is no wonder that Messier 5 is a popular site for astronomers to train their telescopes on".

When there are mentions such as the 13.8 billion years I take that to be in our Earth-year times (using a our present year length). The information would seem to imply that 24,500 light-years equates to 13 billion years? It confuses me (easily done I admit) that something that is 13 billion years old can be only 24,500 light-years away, as I'm sure I have seen it reported that many objects are very much further away in light-years yet are apparently very much less old. Apologies if I am asking something whose answer should be obvious, but could someone please help with my confusion over this.

[BDanielMayfield]

Yes Ann, this really is fascinating, and it was the stars V42 and V84 that were described in considerable detail in the hot off the presses June S&T article. It’s by Howard Banich, and the title inside the mag is M5 Surprise, with the subject line “The great globular cluster sports two stunningly bright variable stars.”

This quote from the article shows why you made your earlier comment:

… M5 contains many variable stars besides the two bright Cepheids. My 28 inch scope is big enough to show M5’s RR Lyrae variables, which shine around 15th magnitude. RR Lyraes are abundant in globular clusters – so much so that they used to be called “cluster variables.”

V42’s pulses are regular as clockwork, with the star ranging from, just eyeballing the chart, about mag. 11.2 at minimum up to about 10.3 with a steady period of 25.72 days. V84’s variability is itself variable, which “may be a sign that it is changing into an RV Tauri star, a special subclass of Cepheids – or perhaps a separate category.” It can range between about 12.3 and 10.8 with a period of about 26 days, plus or minus 0.5 days.

In the blink comparison photos V42 is the bright star blinking in the lower left edge of the main body of the cluster. V84 is the brightest star shown, an outlier in the lower right-hand corner. It’s brightness doesn’t change here because the two photos happened to both have been taken when it was near maximum.

That shows that M5 is even more variable that what we see in these blink images, because numerous stars would have been caught at about the same place in their light curves. It would be really nice to see a time lapse of many photos over an extended period that would show this cluster really pulsating…

Bruce

[Chris Peterson]

When there are mentions such as the 13.8 billion years I take that to be in our Earth-year times (using a our present year length).

A year, in the sense of the time for the Earth to make one orbit around the Sun, is not a fixed unit of time. Astronomers use a Julian year, defined as exactly 365.25 days of 86,400 seconds each (and a second is defined based on the physical properties of cesium-133). A light year is the distance traveled at a velocity of c in one Julian year. So it is independent of anything to do with the Earth, and independent of the age of the Universe.

The information would seem to imply that 24,500 light-years equates to 13 billion years? It confuses me (easily done I admit) that something that is 13 billion years old can be only 24,500 light-years away, as I'm sure I have seen it reported that many objects are very much further away in light-years yet are apparently very much less old. Apologies if I am asking something whose answer should be obvious, but could someone please help with my confusion over this.

Then you must really be bothered by the fact that the Sun is nearly 5 billion years old and only light minutes away!

Cosmologically speaking, the closer something is to us the older it is. Our own galaxy (and the globular clusters around it) is around 13 billion years old (although it now contains many stars that are much younger than that). The most distant galaxies we can see are very young, less than a billion years old. Because of long light travel times, all distant objects must be young. The cosmic microwave background is only a few hundred thousand years old (the photons themselves, of course, were emitted more than 13.7 billion years ago, but the objects that emitted them- which is what we "see" when we record those photons, were just 378,000 years old).

[Anthony Barreiro]

I often get lost when trying to get my mind round the time scale of the Universe and each time I think I have started to understand things something will crop up that causes me to again realise that I really know nothing! In the information brought up through the 'stunning close-up view' link it states "But Messier 5 is no normal globular cluster. At 13 billion years old it is incredibly old, dating back to close to the beginning of the Universe, which is some 13.8 billion years of age. It is also one of the biggest clusters known, and at only 24 500 light-years away, it is no wonder that Messier 5 is a popular site for astronomers to train their telescopes on".

When there are mentions such as the 13.8 billion years I take that to be in our Earth-year times (using a our present year length). The information would seem to imply that 24,500 light-years equates to 13 billion years? It confuses me (easily done I admit) that something that is 13 billion years old can be only 24,500 light-years away, as I'm sure I have seen it reported that many objects are very much further away in light-years yet are apparently very much less old. Apologies if I am asking something whose answer should be obvious, but could someone please help with my confusion over this.

A year is a measure of time, based on the time it takes the Earth to complete one orbit around the Sun. A light year is a measure of distance, the distance a photon of light travels through a vacuum in one year. A rough analogy: I might say that Sacramento is a two-hour drive from San Francisco (if there's no traffic!). I'm using my car, an object traveling at a constant speed, as a way to measure distance. While my statement of distance ("a two hour drive") contains a time unit, what I'm really interested in is the distance traveled during that time.

You can be very close to something that is very old. For instance, the Sun is about four and a half billion years old and about eight light minutes distant from the Earth. (The Earth herself is also about four and a half billion years old, and we're standing, or sitting, or lying down on her.) You can also be far away from something that is relatively young, e.g. 10,000 light years distant from the remnant of a supernova that exploded 1000 years ago.

But remember that we can only see things whose light has had enough time to get to us. So the observable universe has an edge, from our location in space, even though there's probably stuff beyond the edge that we can't see yet. And remember that the universe is expanding, stretching the light as it travels through space ... . Then we get into relativity, four dimensional spacetime, and my brain starts to hurt.

I hope this helps.

[Chris Peterson]

But remember that we can only see things whose light has had enough time to get to us. So the observable universe has an edge, from our location in space, even though there's probably stuff beyond the edge that we can't see yet.

Not "yet". "Never". The boundary of the observable universe is defined by the surface beyond which things are moving away from us at greater than c. And that velocity is only increasing with time. What is outside the observable universe will be so forever.

[Ann]

DavidLeodis wrote:
The information would seem to imply that 24,500 light-years equates to 13 billion years? It confuses me (easily done I admit) that something that is 13 billion years old can be only 24,500 light-years away, as I'm sure I have seen it reported that many objects are very much further away in light-years yet are apparently very much less old.

When we talk about incredibly distant galaxies, they invariably have very high redshifts. We know that they are old because the spectral lines in the light that is only now reaching us from these galaxies have been shifted enormously to the red by the expansion of the universe. Often we only see these galaxies in infrared light, even if they originally emitted a lot of energetic shortwave light.

The light from globular clusters like M5 is not redshifted. The Milky Way globulars orbit the center of our galaxy, and although they are on wide orbits, their distance from us doesn't change all that much. Certainly not in our lifetimes!

Click to view full size image

Source: http://www.eso.org/public/usa/news/eso0228/

But it is important to understand that we can't use redshift to determine the age of the Milky Way globular clusters. To find out how old these globulars are, astronomers measure the metallicity of the stars that belong to the clusters. Take a look at the following diagram of stars of different metallicity:

The top spectrum belongs to the Sun. As you can see from the very "squiggly line" of the solar spectrum, the Sun has lost a lot of light due to absorption by elements in its outer atmosphere. This "squiggly line" tells us that the Sun is a moderately metal-rich star. Although it certainly consists mostly of hydrogen and helium, it also contains quite a bit of of heavier elements such as oxygen, carbon, calcium, iron etcetera.

But you can see that the spectra below the spectrum of the Sun become increasingly straight. The bottom spectrum, which is almost perfectly straight, is the spectrum of a perfect "first-generation star", made of pure hydrogen and helium. Such a star has never been found. The third spectrum from the top belongs to a real star, however. This is likely a second-generation star. The gas it was made of had been "contaminated" by heavier elements from a previous generation of stars, which produced heavy elements in their cores during their lifetimes and scattered these heavy elements when they exploded as supernovas.

The stars of M5 are quite metal-poor. Their spectra are very unlike the spectrum of the Sun. Because of that, we know that M5 is a very old cluster.

Click to view full size image

But we can also tell the age of a globular cluster by the shape of its Herzsprung-Russell diagram. A Herzsprung-Russell diagram plots the luminosity of the stars versus their color and thus temperature. The color and luminosity of a star is a function of its mass and evolution. In the diagram on the left, you can see a slightly wavy diagonal line that is blue at the upper left and red at the lower right. The stars that fall on the main sequence all fuse hydrogen to helium in their cores, but their color and luminosity depends on their mass. The brighter and bluer they are, the more massive they are.

But the most massive stars quickly evolve off the main sequence. They turn into supergiant stars of different colors. Stars of intermediate mass evolve slower, but they too will eventually use up their core hydrogen and turn into giant stars. Most of the giant stars are red or orange giants. In the diagram on the left, you can see the variously colored supergiants scattered at top and the red, orange or yellow giants below the supergiants. These giants form the so-called giant or red giant branch. In this black and white HR diagram, you can see that the giant branch appears to be connected to the main sequence.

In a very young cluster like NGC 2264, only the most massive stars have even reached the main sequence. The less massive stars are located to the right of it, still contracting towards the main sequence. In young cluster NGC 2362, the most massive star is a blue supergiant, whereas most of the other stars are still located on the main sequence.

Click to view full size image

Source: https://www.e-education.psu.edu/astro80 ... l7_p4.html

What about the HR diagram of the Pleiades? This brilliantly blue cluster contains no bright red stars, and therefore it lacks a red giant branch. It also has no supergiant branch. But interestingly, it also has no O-type stars. Therefore the top of the main sequence is empty for the Pleiades. Were there ever any stars there in that spot, when the Pleiades were younger? Well, in my amateur opinion, the Pleiades is such a rich cluster that it is likely that it was born with at least one O-star. But if so, that star has exploded and disappeared a relatively long time ago. Astronomers agree that the Pleiades cluster is at least a hundred million years old, and perhaps several million.

Here you can see the HR diagrams for various clusters. The older the cluster is, the shorter is its main sequence. As the cluster ages, the main sequence shrinks from the top down as more and more stars turn into red giants.

This is a typical globular cluster HR diagram. The main sequence is very short, and the so-called turn-off point has "eaten its way down" to late F and early G-type stars. Only a very old cluster could have an HR diagram like this one. And because the cluster is so metal-poor (again because it is old), it has a blue horizontal branch, which is unique to very metal-poor clusters.

Astronomers use the position of the the turn-off point to calculate how old a cluster is. For M5, the estimated age is 13 billion years.

All in all, this is how we know that M5 is about 13 billion years old.

Ann

[Chris Peterson]

When we talk about incredibly distant galaxies, they invariably have very high redshifts. We know that they are old because the spectral lines in the light that is only now reaching us from these galaxies have been shifted enormously to the red by the expansion of the universe.

This is the source of much confusion. High redshift galaxies are not old. They're young.

[Anthony Barreiro]

But remember that we can only see things whose light has had enough time to get to us. So the observable universe has an edge, from our location in space, even though there's probably stuff beyond the edge that we can't see yet.

Not "yet". "Never". The boundary of the observable universe is defined by the surface beyond which things are moving away from us at greater than c. And that velocity is only increasing with time. What is outside the observable universe will be so forever.

Thanks for the correction. But what if "dark energy" proves to be a transient phenomenon and the universe starts collapsing? Then we would be able see new stuff, right?

My brain is starting to hurt.

[Anthony Barreiro]

When we talk about incredibly distant galaxies, they invariably have very high redshifts. We know that they are old because the spectral lines in the light that is only now reaching us from these galaxies have been shifted enormously to the red by the expansion of the universe.

This is the source of much confusion. High redshift galaxies are not old. They're young.

In the sense that they were young when they emitted the light we're seeing today. But the light itself is very old.

Now my brain is really hurting.

[geckzilla]

Thanks for the correction. But what if "dark energy" proves to be a transient phenomenon and the universe starts collapsing? Then we would be able see new stuff, right?

Transient is a bad word to use here. Dynamic might be a better word. To me you could ask the same question except replace "dark energy" with "gravity" and then understand that the idea of it being transient doesn't make sense.

[Anthony Barreiro]

Thanks for the correction. But what if "dark energy" proves to be a transient phenomenon and the universe starts collapsing? Then we would be able see new stuff, right?

Transient is a bad word to use here. Dynamic might be a better word. To me you could ask the same question except replace "dark energy" with "gravity" and then understand that the idea of it being transient doesn't make sense.

If there were cosmologists more than five billion years ago, they would have observed that the expansion of the universe was decelerating, and would have reasonably predicted an eventual "big crunch." Extrapolation is always an uncertain enterprise.

[Chris Peterson]

This is the source of much confusion. High redshift galaxies are not old. They're young.

In the sense that they were young when they emitted the light we're seeing today.

But we always look at things as having the age where we observe them. Thus, if we see a supernova tonight that's a million light years away, we consider the supernova brand new, not a million years old. The same is true for redshifted galaxies. What makes them interesting is they provide a window into how galaxies formed, because we can see them very young.

But the light itself is very old.

Whatever that means. Light is just energy, and all the energy in the Universe appeared at the instant of the Big Bang.

[Ann]

Anthony Barreiro wrote:
If there were cosmologists more than five billion years ago, they would have observed that the expansion of the universe was decelerating, and would have reasonably predicted an eventual "big crunch."

Sorry about standing in for Nitpicker here, but since I am so relieved that we are apparently not headed for a Big Crunch, I have to defend the open universe. (I was about to write "the open university"!)

If the universe was already decelerating, then we would undoubtedly be headed for a Big Crunch. But the universe has never been decelerating. It was the acceleration of the universe that was slowing down for a while, not the universe itself. When the two teams headed by Saul Perlmutter and Brian Schmidt and Adam Riess were chasing distant supernovas, their aim was to find out how much the acceleration of the universe was decelerating. By finding out how much the acceleration of the universe was being slowed down by the effects of gravity, the teams wanted to find out if the universe was closed or not. They took it for granted that the acceleration of the universe was slowing down, but they didn't take it for granted that the universe was closed. It depended on how much the universe was slowing down. Of course, what they found was that the acceleration of the universe was speeding up.

Now imagine that there were astronomers and cosmologists more than five billion years ago. If they measured the acceleration of the universe at that time, they would indeed find that the acceleration was slowing down. But they would also find that the slowing down of the acceleration was not enough to ever "turn the universe around" and eventually make it collapse. They, too, would have concluded that they lived in an open universe.

Ann