The White Dwarf Cooling Sequence of NGC 6397

[doubletruncation]

There was a paper out today on the white dwarf cooling sequence of the globular cluster NGC 6397 (the second closest globular to us, after M4). I don't know if there has been a press release for this (I wouldn't be surprised if someone puts one out for it since it was done with Hubble), but the reason I mention it is because this is really quite a beautiful result. <br /><br />Just to summarize, there had been a controversy for a long time that the oldest star clusters in our galaxy (the globular clusters) seemed to be about 16 billions years old or so, which is older than the estimated age of the universe based on cosmological models. This is obviously a problem, but one that seemed to be resolved at the end of the 90s when better distance estimates for the star clusters were obtained. The ages of these clusters has really come from only one measurement - the measurement of the brightest main sequence stars left in the cluster (over time the brighter, more massive stars burn out and so the main sequence seems to peel away, you can use that fact as an age indicator but it relies on you trusting your models of how stars evolve and so it'd be very nice to get another independent way to verify those ages). The bright stars will die and turn into white dwarfs which in turn cool down and become fainter. In principle the faintest white dwarf in a cluster could also be used to determine the age since it tells you how long ago the first white dwarf was made, but it's an incredibly difficult thing to measure in practice since white dwarfs are very very faint. But that's just what a group of astronomers from UCLA and other places have done. By using the Hubble telescope they were able to see objects as faint as 30th magnitude (The Sun is to Vega as the faintest stars you could in the suburbs by eye are to a 30th magnitude star), and convincingly show that the faintest white dwarf is about 27.6 magnitudes - which gives an 11.47 +- 0.47 billion years which is in agreement with <div class="Discussion_UserSignature"> </div>

 

[alkalin]

HR is about brightness and color which can be measured. It seems to me age is inferred and can be any scale you like because it is not directly measurable. It makes me nervous about how they then pin age on something from this data. Many stars could be far older than what shows on these diagrams due to the acquisition of nearby fuel that can keep them within a certain temperature range far longer than we think.

 

[alkalin]

There perhaps is good theory and ways to test it in binary systems, but there are a great many stars that are not in binary systems yet display remarkable variability that is not cyclical. Our sun seems to do the same in very low levels that is independent of the sun spot cycle. This to me is an indicator of what might be going on. The problem is how to measure star fuel that may not be visible. Have we ever seen a comet enter another star? This implies we cannot measure star fuel, and so astronomy simply ignores this possibility.

 

[doubletruncation]

I think this hypothesis (that stars' lives may be prolonged by the acquisition of additional fuel or that stellar variability might trick us into assigning the wrong luminosity that you'd expect from evolution) can effectively be tested. The fact that most stars tightly follow a sequence including all sorts of twists and turns (like the Hertzsprung gap, the red giant branch, the horizontal branch and the assymptotic giant branch) means that whatever additional effect that is not considered must somehow affect all the stars or at affect them in such a way that these sequences are kept tight and in the correct positions relative to each other. <br /><br />As eburacum mentioned the only stars that would really be affected by significant mass acquisition are those in close binaries. In fact, you can see that effect on a HR diagram for a cluster as a trickle of stars that keep on following the main sequence above the cluster turn-off. These are the so-called blue stragglers, they're stars that have accreted additional mass from a binary companion that has since died and were spun up so that they don't burn their fuel quite so rapidly as you'd expect, as a result they're artificially bigger and live longer than you'd expect for stars of that spectral type. Most stars, however, are not in close binaries (you can tell that because close binaries will either have spectroscopic or photometric signatures that you can see), so most stars do follow the main sequence turn-off and you can measure it quite easily. <br /><br />Now suppose we're talking about other types of mass transfer (like comets) that we might not have considered. I would first argue that comets, and even planets, are so miniscule in size compared to a star that the star would have to eat an incredible number of them to really affect its total mass at a level that we could measure. Even if it ate several Jupiters worth of mass, the affect on the mass is within our error bars. People do models of stars eating plan <div class="Discussion_UserSignature"> </div>

 

[alkalin]

Excellent remarks, but I find it difficult to believe that the theories are always correct.<br /><br /><font color="yellow">I think this hypothesis (that stars' lives may be prolonged by the acquisition of additional fuel or that stellar variability might trick us into assigning the wrong luminosity that you'd expect from evolution) can effectively be tested. The fact that most stars tightly follow a sequence including all sorts of twists and turns (like the Hertzsprung gap, the red giant branch, the horizontal branch and the assymptotic giant branch) means that whatever additional effect that is not considered must somehow affect all the stars or at affect them in such a way that these sequences are kept tight and in the correct positions relative to each other. </font><br /><br />I agree, but this may not clearly indicate age of the star, only luminosity and color. The parameter we cannot directly measure is age.<br /><br /><font color="yellow">Eburacum mentioned the only stars that would really be affected by significant mass acquisition are those in close binaries. In fact, you can see that effect on a HR diagram for a cluster as a trickle of stars that keep on following the main sequence above the cluster turn-off. These are the so-called blue stragglers, they're stars that have accreted additional mass from a binary companion that has since died and were spun up so that they don't burn their fuel quite so rapidly as you'd expect, as a result they're artificially bigger and live longer than you'd expect for stars of that spectral type. Most stars, however, are not in close binaries (you can tell that because close binaries will either have spectroscopic or photometric signatures that you can see), so most stars do follow the main sequence turn-off and you can measure it quite easily</font><br /><br />Here I agree also that stars in clusters seem to be about the same age. Assuming that a binary has different data that needs some such different c

 

[doubletruncation]

I agree with you that we don't really have a direct indicator of age for most stars, and that it is inferred from the theory. If we could get a direct age indicator (like a radioactive dating) that would provide an excellent test of these theories - so I certainly don't think we should stop looking. For the solar system we actually do have ages from radioactive dating of meteorites and they point to an age which agrees with the agree from modeling the Sun, while this may not absolutely prove that everything is right I think it does strongly point out that people are on the right track. <br /><br />Perhaps I'm a little less skeptical of the theory, that's because the color magnitude diagrams for clusters seems to agree so well with what is predicted by theory, and because when new regions of the diagram are opened up (like the white dwarf cooling sequence) it still agrees with the theoretical predictions. But you're right, we should still keep looking to test the theories. That's really one of the reasons why I am excited about the cooling sequence result. <br /><br />Re. the swarms of comets that might be ingested by stars, it is a fair hypothesis that I think can and should be tested. We can do a quick order of magnitude calculation to put some constraints on the hypothesis. So for the sun's mass to have changed by even 1% over the course of its lifetime it would have had to ingest 2E31 grams / 4.5E9 years * 1 comet / (8E16 grams) = 50000 comets/year. (Note the comet mass estimate comes from http://www.science.edu.sg/ssc/detailed.jsp?artid=2093&type=6&root=6&parent=6&cat=64 for Halley's comet). SOHO has detected roughly 1100 sun-grazing comets (not all of which are actually destroyed) in ten years (e.g. http://ares.nrl.navy.mil/sungrazer/i <div class="Discussion_UserSignature"> </div>

 

[doubletruncation]

While dust is a big problem with estimating distances, it's not really true that people just ignore it. Usually what you do is you measure the difference in color for the object in question from what you'd expect, call this E(B-V) = (B-V)_observed - (B-V)_expected. You then assume some conversion to go from this number to the fraction of light that is blocked by the dust. Of course this is a tricky business to find out what that conversion is, but you can calibrate it locally for stars where you know their distance from parallax. The intervening dust doesn't just sit there absorbing the light, it also reemits it in the far-infrared. By observing in the far infrared you can get a good idea for how must dust is there and then correct your observations for it. Indeed there are maps of the total amount of dust that you see through (in the galaxy) for any direction in the sky, (e.g. http://irsa.ipac.caltech.edu/applications/DUST/ ) the paper which presented these maps (Schlegel, Finkbeiner, & Davis 1998) is actually the most cited paper in astronomy. There certainly are assumptions that go into this procedure, but it's not something that people just ignore completely. Also, people have started making observations in the near-infrared which is absorbed much less by dust, and hence systematic errors in correcting for the dust will have a smaller effect on your distance measurement. <div class="Discussion_UserSignature"> </div>