APOD: Spiral Galaxy NGC 2841 Close Up (2011 Feb 19)

[Ann]

As for the globular clusters, my impression is that most of them are single-generation objects. The reason is that they are so very metal-poor, and all their stars appear to have the same metallicity. The first O stars in them must have exploded as supernovae, thereby enriching the gas in the globulars with heavy elements. If a new generation of stars were born out of that enriched gas, then those more metal-rich stars in the cluster should have been detectable. This is particularly true if the globulars had kept producing O-stars and supernovae and kept producing star-birth out of still more metal-rich gas.

This is a color-magnitude diagram of a typical ancient metal-poor cluster, M55:

Click to view full size image

You can see the main sequence of stars, beginning at the lower right corner and extending upward and to the left. These stars shine by converting helium to hydrogen in their cores. As long as the stars are on the main sequence, they are brighter and bluer the more massive they are, and fainter and redder the more light-weight they are.

You can see that the main sequence suddenly ends. This happens at the turn-off point, where the broad "line" of stars suddenly turns to the right. What happens here is that the main sequence stars have exhausted the hydrogen in their cores, and can no longer produce energy by fusing hydrogen in their cores. Now their cores start shrinking and heating and their outer parts, their atmospheres, start swelling. Their atmospheres get bigger and cooler, and the stars get brighter and redder.

The stars then manage to turn on their stellar engines again when their cores have grown hot enough to fuse helium into carbon and oxygen. This is the point when the metal-poor stars reach the horizontal branch. Their outer atmospheres shrink and get hotter. If the stars are very metal-poor, their atmospheres will keep shrinking and heating, and the stars will grow progressively bluer but also fainter. But even the bluest of the blue horizontal branch stars are brighter than the brightest of the main sequence stars.

After the blue horizontal branch stars exhaust even the helium in their cores, their cores will shrink once again, their atmospheres will swell again, and they will become even larger, brighter and redder than before. They are not massive enough to start fusing the carbon and oxygen in their cores, but they will turn on and off hydrogen fusion and helium fusion in shells around their cores. The stars will start pulsating, and eventually they will shed their atmospheres, turning into white dwarfs. You can see the white dwarfs in the chart in the lower left corner.

You can see that the main sequence appears to continue even after the turn-off point. The stars here are the so-called blue stragglers. By some mechanism or another, they have managed to get a new helping of hydrogen into their cores, so that they can go on fusing hydrogen in their cores and stay on the main sequence even though almost all other stars of the same mass have left it.

This is the color-magnitude chart of 47 Tucanae:

Click to view full size image

I have frankly no idea what the colors symbolize, but you can see the various populations clearly enough. In particular, note the short red horizontal branch. Because 47 Tucanae is more metal-rich than most globulars of the Milky Way, the stars never shrink much and their photospheres never heat much when they start burning helium in their cores.

Both M55 and 47 Tucanae appear to be more or less single-generation clusters. It could be that they processed, at most, a few generations of supernovae and a few generations of new star birth, but since the metallicity of both clusters is so "similar within themselves", they can't have kept making more and more stars out of ever more metal-rich gas. The metallicity, measured as the amount of iron versus hydrogen (Fe/H) is -1.82 for M55, but "only" -0.71 for 47 Tuc. The difference is significant, and it shows that these two globulars formed out of very different gas. And M55 can't have kept producing many generations of new stars, because then its metallicity would have risen, or so I think anyway.

As for why the early globulars didn't keep on forming new stars, I have no idea. They must have produced a lot of supernovae more or less simultaneously, and since they never contained more than a few million solar masses worth of gravity, they may have been unable to hold on to the gas that was violently driven outwards by all the supernova explosions. Why that gas didn't fall back onto the globulars and started off more star formation, I have no idea.

Ann

[Ann]

I might add something. Very many globulars, and certainly all globulars with a blue horizontal branch, are twelve billion years old or more. But star formation in the universe did not peak back then. An article from Nature by Alan Heavens, Benjamin Panter, Raul Jimenez and James Dunlop, claims that star formation peaked about five billion years ago:

The determination of the star-formation history of the Universe is a key goal of modern cosmology, as it is crucial to our understanding of how galactic structures form and evolve. Observations1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 of young stars in distant galaxies at different times in the past have indicated that the stellar birthrate peaked some eight billion years ago before declining by a factor of around ten to its present value. Here we report an analysis of the ‘fossil record’ of the current stellar populations of 96,545 nearby galaxies, from which we obtained a complete star-formation history. Our results broadly support those derived from high-redshift galaxies. We find, however, that the peak of star formation was more recent—around five billion years ago. We also show that the bigger the stellar mass of the galaxy, the earlier the stars were formed, which indicates that high- and low-mass galaxies have very different histories.

If the article from Nature is correct, then star formation peaked a considerable time after most globulars were formed, and certainly after the blue globulars were formed. There might have been something about the conditions in the very early universe, at the time when the blue globulars were born, that made it hard for these globulars to keep forming stars after their first O stars had exploded as supernovae.

Also note that the authors of that article in Nature says that

the bigger the stellar mass of the galaxy, the earlier the stars were formed, which indicates that high- and low-mass galaxies have very different histories.

So huge elliptical galaxies like M87 may not even have come close to producing a thousand generations of O stars. And smaller galaxies may not have formed stars continuously, and so they, too, may not have produced a thousand generations of O stars.

Ann

[Ann]

Just wanted to point out that the brilliant quasars that we see in the distant universe are probably the violent creation of the kind of black holes that quenched star formation in their host galaxies and turned them into ellipticals. The reason why the quasars are so bright is because they are hungrily feeding and growing. Interaction and mergers probably play an important role when it comes to "feeding" the black holes. Later on, when much of the gas has been expelled from the host galaxies, and there are few concentrated clouds left for the black holes to feed on, the holes go "quiet" and the quasars disappear. But the galaxies are on their way to becoming ellipticals. It isn't certain that all host galaxies of quasars become ellipticals, but I think that many do.

http://astronomy.ua.edu/keel/agn/qsohosts.html

Ann

[owlice]

Ann, I'd like to point out that you have a PM. Please open it. Thanks.

[dougettinger]

As for the globular clusters, my impression is that most of them are single-generation objects. The reason is that they are so very metal-poor, and all their stars appear to have the same metallicity. The first O stars in them must have exploded as supernovae, thereby enriching the gas in the globulars with heavy elements. If a new generation of stars were born out of that enriched gas, then those more metal-rich stars in the cluster should have been detectable. This is particularly true if the globulars had kept producing O-stars and supernovae and kept producing star-birth out of still more metal-rich gas.

This is a color-magnitude diagram of a typical ancient metal-poor cluster, M55:

Click to view full size image

You can see the main sequence of stars, beginning at the lower right corner and extending upward and to the left. These stars shine by converting helium to hydrogen in their cores. As long as the stars are on the main sequence, they are brighter and bluer the more massive they are, and fainter and redder the more light-weight they are.

Ann, are the detection methods for metallicity sensitive enough to detect each generation of O and B type stars ? I would think, but do not actually know, that Population II stars would range from one to 100 generations of O and B type stars producing higher metals. Population I stars would be the result of metallicity being produced by the nucleosynthesis of O and B type stars from 100 to as many as 1000 generations of star birth/death cycles. Population III stars have not been discovered, but, I would imagine, more than serveral generations of O and B type stars would be required to measure any amount of metallicity. Please enlighten me ; I have very little knowledge of measurement techniques.

You speak of the blue horizontal branch of stars that globular clusters possess. Is this branch the group of blue stars that I observe in the upper righthand corner of the color-magnitude diagram ? Should not these stars have exploded a long time ago being inside a globular cluster ?

3/29/2011

[Ann]

You speak of the blue horizontal branch of stars that globular clusters possess. Is this branch the group of blue stars that I observe in the upper righthand corner of the color-magnitude diagram ? Should not these stars have exploded a long time ago being inside a globular cluster ?

I'll start with this question first, since I feel I know the answer quite well.

Yes, the blue horizontal stars are the ones you see in the upper lefthand corner of the color-magnitude diagram. And they have not exploded as supernovae, because they have never been nearly massive enough to explode.They are thought to have started out as stars that were only about as massive as the Sun. Of course a massive cluster like M55 originally contained O stars, but the O stars have indeed exploded billions of years ago. As the original solar-mass stars exhausted their nuclear hydrogen, they swelled to red giants at the same time as their cores shrunk and heated. But once they got hot enough in their cores to turn on helium fusion there, their outer atmospheres shrank. Why did their atmospheres shrink? Don't ask me, Doug. I don't have the faintest idea. I just know that astronomers agree that this is what happened. Oh, and I do know one more thing about it: astronomers claim that this shrinking of the atmosphere is strongly tied to the metal-poor nature of these stars. Personally I would assume that stars which are very metal-poor may have more transparent atmospheres, so that the gamma-ray photons that are generated in the cores of the stars don't lose so much energy on their way to the photospheres of the stars as the gamma-ray photons do in a star like the Sun.

Well, the atmospheres of these stars kept shrinking. And because their cores were hotter now than they had been when the stars were on the main sequence, their shrinking atmospheres could also get hotter, and therefore bluer, than they were when these stars were on the main sequence. It is almost certain that the bluest of the blue horizontal stars are bluer at this stage of their evolution than they ever were when they were on the main sequence.

An interesting thing about the blue horizontal branch is that the not-so-blue part of it contains the so called RR Lyrae stars. These stars pulsate very regularly in a way that is tied to their absolute luminosity. By detecting RR Lyrae stars in globular clusters, astronomers have been able to determine how far away the globulars are, and thus how bright they are. Wikipedia writes about the RR Lyrae stars:

RR Lyrae variables are periodic variable stars, commonly found in globular clusters, and often used as standard candles to measure galactic distances.

This type of variable is named after the prototype, the variable star RR Lyrae in the constellation Lyra.

RR Lyraes are pulsating horizontal branch stars of spectral class A (and rarely F), with a mass of around half the Sun's. They are thought to have previously shed mass and consequently, they were once stars with similar or slightly less mass than the Sun, around 0.8 solar masses.

RR Lyrae stars pulse in a manner similar to Cepheid variables, so the mechanism for the pulsation is thought to be similar, but the nature and histories of these stars is thought to be rather different. In contrast to Cepheids, RR Lyraes are old, relatively low mass, metal-poor "Population II" stars. They are much more common than Cepheids, but also much less luminous. (The average absolute magnitude of an RR Lyrae is 0.75, only 40 or 50 times brighter than our Sun.[citation needed]) Their period is shorter, typically less than one day, sometimes ranging down to seven hours.

As for how metal-poor stars are detected, well, I think spectral analysis is needed. Basically metal-poor stars will have fewer and shawoller spectral lines. I found this image on the net:



Sorry about the "blurry" quality of the image. You can see a very wiggly line at top. That is a part of the spectrum of the Sun, which is a fairly metal-rich star. Below the wiggly line of the Sun is a line with only four notable dips. This is a part of a spectrum of a star which is much, much more metal-poor than the Sun. Below this second line is another one, which only has two notable dips.This third star is even more metal-poor. The bottom line is a never observed but theoretically modeled spectrum of a Population III star, which is made up of absolutely pure hydrogen and helium, so it could be described as "metal-less".

So what do those "dips" in the spectrum represent? They are absorption lines. A star that contain a lot of metals in its atmosphere will absorb a lot of photons of specific energies. Take a look at this picture of the "all the colors of the Sun", plus its absorption lines:

Click to view full size image

All those little black lines that you can see everywhere represent colors that have been absorbed by various elements in the Sun's atmosphere. A completely metal-less star would not lack such lines altogether, because hydrogen and helium also cause absorption lines, at least if the star is hot enough. But most absorption lines here are due to other elements than hydrogen and helium, and they would not be there in a metal-less star.

Ann

[dougettinger]

Hello Ann,

I am beginning to understand your explanation: " As for globular clusters my impression is that most of them are single-generation objects. The reason is that they are so very metal poor, and all their stars appear to have the same metallicity."

At least one, perhaps two, generations of O and B type stars are required to produce some metallicity and create Population II stars. Any more generations would start to cause some mixing and variation in the metallicity which is not the case. Thanks for being patient with me. I now understand. And the thought is that some globular clusters that did not have any O to B type stars should have only Population III stars.

My other question was about the horizontal branch of blue stars which is still a mystery to me. I always thought blue stars were always massive, young stars whose lives were very short. And white dwarfs are dim and found in the lower left-hand corner of the chart. I am guessing that these blue stars are the end-of-life yellow stars that have not fully evolved into white dwarfs. And the RR Lyrae stars are in the transition stage between going to white dwarfs from red giants. Please clarify my understanding when you get a chance. Thank you.

3/30/2011

[Ann]

Hello again, Doug!

And the thought is that some globular clusters that did not have any O to B type stars should have only Population III stars.

I don't think there are globulars that are made up of Population III stars, because no such globulars have been found. Please note that the longest-living red dwarfs in such a cluster, which would have to be less than 14 billion years old (because the universe is thought to be less than 14 billion years old), would be alive and well. Indeed, the main sequence K stars would also be left, except perhaps the most massive K0 main sequence stars. And there would be red giants and blue horizontal branch stars. Such a cluster would be easy to spot if it belonged to the Local Group (that is, to us or Andromeda or M33 or a number of local dwarfs), and I think that Hubble could see them in the Virgo Cluster too, which is about 60 million light-years away. I think you could analyze such a cluster spectroscopically "as a single body" so to speak, and its extreme lack of metals would just jump out at you.

I'm not saying that there are no globular clusters made up of Population III stars anywhere in the entire universe, but I am saying that astronomers have not found a single Population III star yet.

I like the suggestion that an ionization front swept all over the universe and made metal-poor (but not metal-less) gas clouds collapse all over the universe to form those massive clusters that still survive today as globulars. But there must probably have been a generation of super-massive stars that came first and which enriched the gas in the universe by exploding as supernovae. These stars would have got the ionization front going, too.

I always thought blue stars were always massive, young stars whose lives were very short.

Yes, that's what I used to think, too. But that's not the whole truth.

Fifteen years ago or so I bought Sky Catalogue 2000.0, which lists 50,000 stars by color and magnitude. Sky Catalogue estimates the absolute magnitude of each star by looking at its spectrum. This, however, turns out to be an uncertain method.

I read through all of Sky Catalogue and made a note of every star that had a negative B-V index (i.e., it was blue). I then checked all these stars against the Hipparcos Catalogue. I found that Sky Catalogue had generally overestimated the brightness of the blue stars. They were still bright, but generally not quite as bright as astronomers had thought. Some of them were very bright indeed, 10,000 or more times as bright as the Sun in visual light, but several of them were pretty run-of-the-mill hot stars that were only a thousand times as bright as the Sun, or even less, in visual light. Of course the O stars peak at very ultraviolet magnitudes, so their bolometric (total) energy output is always a lot more than a thousand times that of the Sun's - at least if they are "normal" O stars!

Amazingly, to me at least, I found some hot stars - early B stars or late O stars - that did not seem normal at all, because they were so puny. They were only about thirty times as bright as the Sun in visual light! How was that possible?

My answer is that these hot stars must have been small. The way I see it, that is the only reasonable explanation. There is no doubt that they were hot and blue, but if they were so faint they just had to be small and not very massive. Perhaps they were metal-poor blue horizontal stars? Why not? Thirty times the luminosity of the Sun is just what we can expect from blue hoizontal stars, if that color-magnitude chart of M55 is to be believed. And if those faint blue stars were really blue horizontal stars, then they were certainly not very massive at all: indeed, blue horizontal stars are less massive than the Sun! They just generate more energy than the Sun does in their evolved cores.

Or perhaps those faint blue stars had shed much of their atmosphere by some other means. Almost certainly they had not started out as O or B-type stars at all. Perhaps they had originally been, say, A- or F-type stars that had evolved, grown a hotter core, and shed much of their atmospheres for some reason. A star is always hottest in its center, and if a star's atmosphere gets thinner, then more of the central heat will reach the photosphere of the star as blue and ultraviolet light, and there will be much less of the central heat that is converted into red or infrared light before it reaches the photosphere. The star will simply look bluer, but its total light output will not be that impressive, because the star is small. A small star has a small photosphere, and it simply can't radiate that much light from such a small surface.

I'll give you a few examples. I'm afraid I can't find the really small blue stars now, because I don't have the energy to go through all of Sky Catalogue again. But take a star like V444 Cygni, for example. It is a binary star consisting of an O star and a Wolf Rayet star (an evolved blue star that has shed much of its atmosphere), and the total luminosity of this system was thought to be about 8,500 times that of the Sun in visual light. According to Hipparcos, its combined luminosity is 66 times that of the Sun! Well, that figure is wrong, because the system is strongly reddened by intervening dust, and that always makes stars look much fainter than they really are. Also the distance to the binary is uncertain, and if it has been underestimated the stars will be brighter than Hipparcos seems to show. But still, what a shock to find that figure - 66 times as bright as the Sun!

Let me compare V444 Cyg with a star that turned out to be surprisingly blue and bright even though it is only classified as an A-type star that is rich in silicon. The star is called IQ Aurigae. Bright Star Catalogue writes about IQ Aurigae: "Bluest known peculiar A-star: color characteristic of about B4V. Classified as A0p because He weak, but in other respects the color is more indicative of the temperature than the spectral class of Ap stars. One of the hottest Ap stars known, about 17000K." Note that IQ Aurigae is pretty bright, too: about 111 times as bright as the Sun in visual light, which is quite a lot for an A-type main sequence star. Take a star like Vega: it is not nearly as blue as IQ Aurigae, and it is "only" about fifty times as bright as the Sun. And even so Vega is rather bright and blue as A-type stars go.

IQ Aurigae is the bluest-looking star here. It is the brightest middle star in the grouping of five between the two red nebulae: http://www.nightsky.at/Photo/Neb/IC405_IC410_WN.jpg

One star that may be very big, bright and blue is Alnilam, the middle star of Orion's Belt. Its color index is -0.184, which is really quite blue, and it may be as bright as 30,000 times the Sun in visual light, and many times more in ultraviolet light. But I must point out that the distance to Alnilam is uncertain. Without a doubt, however, Alnilam is massive, and it is destined to end in a brilliant supernova.

Alnilam, the middle star of Orion's Belt, may be a true whopper: http://www.skyfactory.org/orionbelt/ori ... llview.jpg

My point is that not all hot blue stars are the same. At all. Some are big, and they are humongously bright. Some are small, and they are really faint as hot stars go. And these faint blue stars are usually not very massive.

Oh, and by the way, don't forget the smallest, bluest of all stars, the white dwarfs! Yes, they can come in all colors, depending on how much they have cooled, but many of them are very blue. This is particularly true of most central stars of planetary nebulae. Here is an example:

Click to view full size image

Planetary nebula Abell 43. The central star is very blue, but it is also very small.

Ann

[dougettinger]

I realize that no Population III stars have been found in globular clusters, but the search continues. Since the oldest structures in the universe, the globular clusters, have some metallicity is proof that the initial stars in the universe were mostly O and B type stars.

Yes, I am just learning that white dwarfs and RR Lyrae stars can be blue stars but not massive like O and B type stars. And the horizontal branch of blue stars in globular clusters is the result of the death pangs of some proportion of the long lived stars of that cluster. It is really astounding have much knowledge can be gained by looking at the night sky.

3/30/2011

[Ann]

And the horizontal branch of blue stars in globular clusters is the result of the death pangs of some proportion of the long lived stars of that cluster.

I don't think that the blue stars in globular clusters are the result of death pangs either of long lived stars or of any stars in that cluster. I think the blue stars in metal-poor globular clusters were born at the same time as the longest-living small red stars in that cluster. I like the idea that there was at least one generation of stars before the globular clusters, which produced truly massive stars that exploded as supernovae and enriched the gas of the very early universe. These primoridal stars, the Population III stars, would also have reionized the universe, both with their ferocious stellar winds and blastingly energetic ultraviolet light and with their own explosions as supernovae.

But the blue horizontal stars are blue precisely because they are so metal-poor, so they don't need metral-enriched gas to be blue - rather the opposite. But clearly, since no Population III globular clusters have been detected, but thousands and thousands of 12 to 13 billion-year-old metal-poor globular clusters with blue horizontal stars have been detected, we can say that back at that time conditions were just perfect for making globular clusters, but when the fist Population III stars existed, the conditions in the universe were not "globular cluster-friendly". But I think you are definitely right that the Population III stars played a crucial role for giving the universe the properties it needed, back then, to make all those metal-poor globulars.

Ann

[Ann]

Why haven't we found any Population III stars? That is really a mystery.

I think we can draw the following conclusions:

1) The Population III stars didn't form the kind of clusters that will stay gravitationally bound for more than 13 billion years. If such Population III clusters had existed in our own galaxy or anywhere in the Local Group, I think it is almost certain that we would have detected them by now.

2) The Population III stars also didn't form the sort of rich clusters that will leave obvious star streams behind as the clusters disperse.

Click to view full size image

Star streams orbiting galaxy NGC 5907.

There are several star streams orbiting our own Milky Way, and at least one such stream has been shown to originate from a globular cluster in the process of being shredded: http://www.skyandtelescope.com/news/330 ... page=1&c=y

The Population III stars almost certainly didn't form the kind of clusters that would result in those star streams, because I really think we would have detected them by now. Admittedly there could be some very faint streams that have not been detected. What a sensation if astronomers really found a stream of Population III stars orbiting the Milky Way!

3) The really massive Population III stars have exploded as supernovae billions of years ago. All the O stars have exploded, and the B stars and A stars have also disappeared. Indeed, so have the F stars. What is left is faint main sequence red and yellow dwarfs that may be very well mixed with much, much numerous and more metal-rich younger populations.

Click to view full size image

Imagine trying to find a faint Population III star in this crowd. Note that the diffuse yellow light comes from billions of mostly small stars.

There ought to be brighter Population III stars left too, however. There ought to be Population III red giants and AGB stars left, and they can be several hundred times as bright as the Sun. You should be able to spot them. Of course, they will not immediately look any different than other read giants and AGB stars, and there are many of them, too, so you still have to do a lot of searching. But it should be possible to find them, if they exist.

What if stars didn't form in the same way at all when the first stars were born as they do today? Today, star formation always creates mostly small and low-mass stars. What if it was not like that when the first stars were born? What if most of them were massive, so that the number of low-mass long-lived Population III stars was really small? Then it is no wonder that we can't find them.

Ann

[dougettinger]

You can see that the main sequence suddenly ends. This happens at the turn-off point, where the broad "line" of stars suddenly turns to the right. What happens here is that the main sequence stars have exhausted the hydrogen in their cores, and can no longer produce energy by fusing hydrogen in their cores. Now their cores start shrinking and heating and their outer parts, their atmospheres, start swelling. Their atmospheres get bigger and cooler, and the stars get brighter and redder.

The stars then manage to turn on their stellar engines again when their cores have grown hot enough to fuse helium into carbon and oxygen. This is the point when the metal-poor stars reach the horizontal branch. Their outer atmospheres shrink and get hotter. If the stars are very metal-poor, their atmospheres will keep shrinking and heating, and the stars will grow progressively bluer but also fainter. But even the bluest of the blue horizontal branch stars are brighter than the brightest of the main sequence stars.

After the blue horizontal branch stars exhaust even the helium in their cores, their cores will shrink once again, their atmospheres will swell again, and they will become even larger, brighter and redder than before. They are not massive enough to start fusing the carbon and oxygen in their cores, but they will turn on and off hydrogen fusion and helium fusion in shells around their cores. The stars will start pulsating, and eventually they will shed their atmospheres, turning into white dwarfs. You can see the white dwarfs in the chart in the lower left corner.

You can see that the main sequence appears to continue even after the turn-off point. The stars here are the so-called blue stragglers. By some mechanism or another, they have managed to get a new helping of hydrogen into their cores, so that they can go on fusing hydrogen in their cores and stay on the main sequence even though almost all other stars of the same mass have left.

I am trying to better understand the stars of the blue horizontal branch in globular clusters as you have explained in the above paragraphs. I incorrectly term red giants as dying stars when they are really just a phase in the life of a star. I did not realize that there are other phases between the red giant and the white dwarf phases of which these blue stars are one. Once the star is fusing helium in its core, it contracts and no longer is considered a red giant. The core helium fusing phase of a star's life is called the "horizontal branch" in metal-poor stars, so named because these stars lie on a nearly horizontal line in the H-R diagram of many star clusters. Another point that I never realized is that metal-rich helium stars instead lie on the so-called "red clump" in the H-R diagram.

I am guessing that the metal-rich stars have more mass and are able to fuse more higher metals in their cores. The higher metals give the star a more reddish color.

3/31/2011

[Ann]

I am trying to better understand the stars of the blue horizontal branch in globular clusters as you have explained in the above paragraphs. I incorrectly term red giants as dying stars when they are really just a phase in the life of a star. I did not realize that there are other phases between the red giant and the white dwarf phases of which these blue stars are one.

Yes, there are two "red giant branches". The first starts when the stars run out of hydrogen in their cores. Then their atmospheres swell and get cooler and therefore redder. But like you said, when the cores get sufficiently hot, the stars can fuse helium in their cores. Then, if the stars are very metal-poor, their atmostpheres shrink very much, and the heat from their now much hotter cores reach their photospheres, and they become blue. The smaller they get, the bluer they will be.

But the stars will deplete the helium in their cores too, and when that has happened they will leave the blue giant branch. Again their cores will heat, and again their atmospheres will swell. I think their atmospheres will swell even more this second time than they did the first time the stars ran out of fuel in their cores. But because these old stars are all low-mass, they can never get hot enough in their cores to start fusing other elements than helium. Therefore they are now dying. They are on their way to becoming white dwarfs. But just before they die, they get bigger, brighter and redder than they ever were before. This "second red giant branch" is called the "Asymtotic Giant Branch", or the AGB branch, and the stars here are called AGB stars.

Another point that I never realized is that metal-rich helium stars instead lie on the so-called "red clump" in the H-R diagram.

Exactly, that is what the "red clump" is. You could see that the second brightest globular of the Milky Way, 47 Tucanae, doesn't have a blue horizontal branch, only a red clump. Therefore 47 Tucanae doesn't have any RR Lyrae stars either.

I am guessing that the metal-rich stars have more mass and are able to fuse more higher metals in their cores. The higher metals give the star a more reddish color.

This is partly right and partly wrong, I think. Yes, metal-rich stars are typically redder than metal-poor stars of the same mass. So a metal-poor star of spectral class G2V, like our Sun, would have less mass than our Sun. Generally stars will have more spectral lines the cooler they are, but they will also have more spectral lines the more metal-rich they are. Therefore a metal-poor star will have to be cooler than the Sun in order to display the same or similar spectral lines. Conversely, a metal-poor star of the same mass as the Sun will be bluer than the Sun, and it will display a different spectrum seemingly characteristic of a hotter star. So, to summarize: If you find a metal-poor blue star whose spectral lines suggest that the star is an A0 star like Vega, don't expect it to have the same physical characteristics as Vega. The metal-poor star will surely be more light-weight than Vega, which is about three times more massive than the Sun.

But wait a minute. If a star is cooler than the Sun, then surely it will also look redder than the Sun? Isn't the color of the star indicative of its temperature? Yes, but that only refers to the temperature of the photosphere of the star. Indeed the color of the photosphere corresponds to the temperature of the photosphere. But the temperatures of the cores of a metal-rich and a metal-poor star will be different if the color of their photospheres is the same. The core of the low-mass star will be cooler. That is because the atmosphere of the metal-poor star is more transparent, and the gamma rays generated in the metal-poor star's core don't lose as much energy on their way to the photosphere as the gamma rays generated in the core of a metal-rich star do.

The temperature of the core of the Sun is about 15 million degrees Kelvin, but a low-mass star of the same color and spectral class than the Sun will have a cooler core. The metal-poor star of spectral class G2V will simply not generate as much energy in its core as our metal-rich Sun does. Therefore the metal-poor star of spectral class G2V will also be fainter than the Sun.

So it is true that metal-rich stars will be more massive than metal-poor stars of the same spectral class. It isn't automatically true that metal-rich stars will be able to fuse higher metals in their cores. I think the Sun, too, will only be able to fuse helium in its core. Once it has run of of helium in its core, the fusion processes will stop in there, and the Sun will start to die.

Ann

[dougettinger]

If we see and measure the characteristics of the photosphere, how do we know the temperature of the core ? You indicated that the color or spectral class of the photosphere reveals the temperature of the core. The cooler the core the hotter the photosphere becomes because the transparency of the atmosphere is more than a star with a hotter core. The gamma rays from the hotter core are more energetic, bounce around more in the intervening atmosphere thereby losing more energy before they reach the photosphere.

I am still having difficulties distinguishing between poor-metal and rich-metal stars. Our Sun has 1 % higher metals than H and He in its composition. How can that minute amount of higher metals make a difference in determining the spectral class of a star. Does a poor-metal star have 1/2 % higher metals compared to the Sun with its 1% ? How does such little matter as the higher metals create different spectral classes? I do understand how different temperatures create difference spectral classes.

4/01/2011

[Ann]

If we see and measure the characteristics of the photosphere, how do we know the temperature of the core ?

If you ask me, that is revealed by the total (bolometric) energy output of the star.

Consider red supergiant star Betelgeuse, whose surface temperature is about 3,500 degrees Kelvin (3,650 Kelvin according to Professor emeritus Jim Kaler). According to Kaler, the bolometric luminosity of Betelgeuse is somewhere between 85,000 and 105,000 times the luminosity of the Sun. Most of that energy will be emitted as infrared light, precisely because the photosphere of Betelgeuse is so cool. But whether the energy is emitted in the infrared or in another part of the spectrum, the amount of energy generated is huge.To generate that much energy the core of this star has to be very hot indeed.

Now compare Betelgeuse with our own nearest stellar neighbour, M5.5 dwarf Proxima Centauri. Proxima is indeed a little cooler on the surface than Betelgeuse, because its outer temperature is only 3,040 Kelvin according to Kaler, but the surface temperature of these two stars is nevertheless comparable. Their energy output is not. The bolometric luminosity of Proxima Centauri is about 1/600 the luminosity of the Sun, or 0.001666...% the luminosity of the Sun. So what is the luminosity difference between Betelgeuse and Proxima? It seems to me that it would be about 100,000 x 600, or 60,000,000 - sixty million. These two stars have comparable surface temperatures, but one star is sixty million times brighter than the other one. How can that be? It must be because the "stellar engine" of Betelgeuse is sixty million times more efficent that the stellar engine of Proxima Centauri at generating energy. Much of that difference in efficiency will be due to the temperature of the cores of these two stars.

And of course Betelgeuse is a humongously big star, which radiates at an average of 3,650K all over its monstrously large surface. Proxima, by comparison, is a tiny candle, radiating at 3,040 K from a surface that is like the surface of a pin-head by comparison.

Ann

[Ann]

I am still having difficulties distinguishing between poor-metal and rich-metal stars. Our Sun has 1 % higher metals than H and He in its composition. How can that minute amount of higher metals make a difference in determining the spectral class of a star. Does a poor-metal star have 1/2 % higher metals compared to the Sun with its 1% ? How does such little matter as the higher metals create different spectral classes?

I wish I could explain better, but I have not taken any sort of detailed interest in this.

I know I have heard that our Sun is described as relatively metal-rich. I have also heard that it was probably formed out of gas that was recently compressed and, probably, enriched by a supernova. Gas that has recently been given a fresh helping of elements that have just been forged in a supernova can be expected to be somewhat metal-rich. Sveral stars that are much younger than the Sun, such as the Pleiades, are much more metal-poor than the Sun. Still, they are many times more metal-rich than the stars in globular clusters.

According to the article about metallicity in Wikipedia (a good article, you should check it out) metallicity is determined as the the logarithm of the ratio of a star's iron abundance compared to that of the Sun. If a star's iron abundance compared to the Sun's is -1, then the star has only a tenth as much iron relatively to hydrogen as the Sun does. If the iron abundance is -2, the star has only a hundredth as much iron as the Sun. A very common metallicity for globular clusters is about -1.50, but several clusters are more metal-poor than that. NGC 5053 has an estimated metallicity of -2.58, which would make its iron content several hundred times lower than that of the Sun's.

Click to view full size image

NGC 5053, a very metal-poor cluster.

As to why a lower metal content changes a star's spectral class, well, the spectral class is determined by the absorption lines in the star's spectrum. The higher the temperature of the photosphere of the star, the fewer absoprtion lines there will generally be. But a lower metal content also affects the spectral lines. Some lines will simply disappear because the star barely contains any amounts of this particular metal. Other metal lines will still be there, but they will be much "weaker", shallower and thinner, than expected. This affects the classification of the star, so that a metal-poor star of a given mass gets another spectral classification than a metal-rich star of the same mass. But if the stars are the same mass and at the same stage of evolution, they will nevertheless have the same bolometric luminosity, or so I think anyway.

Ann

[Ann]

Why is a metal-poor star bluer than a metal-rich one? Well, take a look at the solar spectrum that I posted a little while ago. You can see that most of the absorption lines are especially numerous in the blue part of the spectrum. A disproportionate part of the blue light of the Sun is absorbed by those spectral lines, and then that energy that was absorbed as blue light will be re-emitted at longer, redder wavelengths. So a lot of absorption lines will simply shift the Sun's overall light curve slightly to the red.

A star that doesn't have many spectral lines at all will not absorb very much of its own blue light at all. Most of the blue light will be emitted, and very little of it will be shifted to longer, redder wavelengths. To me, that seems like a reasonable explanation for the bluer color of metal-poor stars.

Click to view full size image

Typical spectral lines of stars of different spectral types.

I checked out the metal-poor F4 type star. Guess what? It is not particularly blue for an F4V star, 0.498 ±0.013, but I do think it is rather faint for a star of that spectral class. Its luminosity is 1.17 ± 0.14 times the luminosity of the Sun. Main sequence F stars are supposed to be a bit more massive and therefore noticably more luminous than main sequence G stars like the Sun. Compare it with well-known "solar twin" 18 Scorpii, whose color is similar to the Sun's (0.652 ± 0.009 for 18 Scorpii versus 0.656±0.005 for the Sun) and the luminosity of 18 Scorpii is also similar to the Sun: 1.042 ± 0.026 for 18 Scorpii, versus, of course, 1.000 for the luminosity of the Sun. Frankly I think that HD 94028 is most like a slightly blue G0 star, which is what you would expect from a metal-poor star: faint for its spectral class, blue for its luminosity. I hunted around for a star to compare HD 94028 with, and I found 71 Orionis. This star is classified as an unevolved F6 star, so it ought to have been redder and fainter than HD 94028. It is neither. It is slightly bluer than HD 94028, 0.430 ± 0.000, and it is also brighter, 3.09 ± 0.11 times the luminosity of the Sun. I repeat what I said, this metal-poor F4 star is faint for its class, probably low-mass for its spectral class, and when it comes to its mass and energy production, it is more like a G star like the Sun than it is like a more massive F star.

Ann

[Ann]

Guess what? I decided to check out that O6.5 star too, HD 12993. It is classified as a main sequence star, so it is still burning hydrogen in its core. Its core has not noticably heated because the star is evolving, and its atmosphere has not noticably expanded. I'm sure it is a "normal", massive O star.

Guess what, though? According to my software, HD 12993 is exceedingly faint for an O star, only 70 times the luminosity of the Sun! Well, that's impossible for a normal O star which is not exceedingly metal-poor, which it is not. But the star is definitely reddened by intervening dust: its B-V index is +0.18, when -0.30 would be expected. Also the distance to it is uncertain. However, the distance measurement, while uncertain, is probably not altogether off. Not so far from HD 12993 in the sky is another star, 5 Persei or 5 Per, whose parallax and proper motion is almost the same as HD 12993. I therefore think that these two stars are located at much the same distance from us, and that the distance measurements are not so bad, since they are so similar for these two stars. But 5 Per is much brighter than HD 12993, and my software tells me that the estimated luminosity of 5 Per is 1400 times that of the Sun, which would make it twenty times brighter than O star HD 12993! Actually 5 Per is clearly much, much brighter than 1400 times the Sun, because it is classified as a supergiant star of spectral class B5Ia. Such stars aren't 1400 times as bright as the Sun, but more likely at least 20,000 times as bright as the Sun in visual light. 5 Per is also reddened by intervening dust, because its B-V is +0.298, where something between -0.05 and -0.10 could be expected. So 5 Per may look about 14 times too faint to us for its brightness and distance, due to intervening dust. If HD 12993 is also about 14 times too faint-looking, it would be about 980 times brighter than the Sun in visual light, which would make it a run-of-the-mill O star. Of course, because HD 12993 is such a hot star, it has a large "bolometric correction", because it emits most of its energy at other wavelengths than the yellow-green part of the spectrum. It could be that its bolometric luminosity is a 100,000 times the bolometric luminosity of the Sun, although this is not certain.

Ann

[Ann]

Couldn't HD 94028, the metal-poor F4V star, also be dust-reddened and therefore actually brighter than it appears to be? If we could remove any dust in the line of sight between us and the star, wouldn't HD 94028 appear like a more "normal" F4V star, both when it comes to its color and when it comes to its luminosity?That is a good question, since HD 94028 is actually a little red for its spectral class.

But I don't think HD 94028 is noticeably dust-reddened. The star is in the constellation Leo, relatively far from the galactic plane. Leo is not a very dusty constellation. Also HD 94028 is not all that far away. The distance to it is estimated to be 169.6 ± 10 light-years, and this distance shouldn't produce a lot of reddening in a relatively dust-poor constellation. The seemingly shockingly faint O6.5V star HD 12993 is located in the constellation Perseus, square in the middle of the plane of the Milky Way, and the star may even be located in the Perseus arm of our galaxy, outside our own local "Orion spur". The distance to HD 12993 is estimated by my software to be 1900 ± 1400 light-years, but with this level of uncertainty the distance to the star could be more than 3300 light-years. In any case, any star that is more than 500 light-years away in the constellation Perseus is likely to be quite reddened by dust.

So I think that the seemingly abnormally red color and faint luminosity of HD 12993 are both produced by intervening dust, whereas the slightly red color and faint luminosity of HD 94028 are produced by the fact that this is a star with a mass characteristic of a G0V star, which, thanks to its own metal-poor nature, is mimicking as an F4V star.

Ann

[dougettinger]

You explained exceedingly well the correspondence of a star's spectral class, luminosity, and metallicity. And now I know much more detail about the branches of the Hertzsprung-Russell chart. Thank you so much, Ann.