Astronomers rewrite life story of stars

[glenc]

Astronomers have new information on how stars age and die - and it's not what they thought. The findings from an Australian-led study overturn conventional thinking that the "red giant phase" is the last stage in a star's life.
http://www.armidaleexpress.com.au/story/1538457/astronomers-rewrite-life-story-of-stars/?cs=12
http://theconversation.com/a-stellar-mid-life-crisis-why-do-some-cluster-stars-die-early-14598

[astroron]

Very interesting Glen.
The commentary on the video leaves a bit to be desired:(
Quote) a galaxy like ours with a hundred million stars. (end quote)
Also showed the Helix PNe as a supernova :rolleyes:.
Cheers:thumbsup:

[madbadgalaxyman]

No asymptotic giant branch (= AGB) phase for some of the solar mass stars in NGC 6752.

If this major revision of the existing theory of stellar evolution is confirmed, it will have major consequences.

The paper was published online, in 'Nature'(which is not a nudist magazine!) on the 29th of May this year. The lead author is Simon W. Campbell of Monash University.

They want me to pay $32 for this paper, which is totally unreasonable.....

Aha!! Here is the preprint of this article!!

140625

I hope that Dana will comment, as he seems to be very good on the Hertzprung-Russell diagram (color vs magnitude) and the theory of stellar evolution.

P.S.
the popular-level writeup of this work at the ESO website is reasonably clear and understandable.

[Weltevreden SA]

06/02/2013:

Hi Robert, Glen, Ron, & everybody who commented on the Campbell et al paper re AGB findings in NGC 6752 Pavo. I just got into the city after a week at my no-Internet, no-cell network dark site.

I’d like to preface what I’ve found about N6752 by saying you fellows are really fortunate to have such hot, forward looking astronomy teams at Macquarie, Monash, Austr. National Univ., Parkes, & Mt. Stromlo. I’ve come across several very new, innovative, and even exciting papers in the literature by the Aussie research teams. One is a really interesting March 2013 paper by Daniela Carollo, Ken Freeman, and Sarah Martel (http://arxiv.org/abs/1303.4168). (See their other papers at Carollo (http://arxiv.org/find/all/1/au:+Carollo%5fDaniela/0/1/0/all/0/1?skip=0&query_id=33876e763cc27a89), Freeman (http://arxiv.org/find/all/1/au:+Freeman%255fKenneth/0/1/0/all/0/1?skip=0&query_id=f79ab7abfbd37235), Martell (http://arxiv.org/find/all/1/au:+Martell%255fSarah/0/1/0/all/0/1?skip=0&query_id=d86092259261e4ca).) The Carollo et al paper provides an observational basis for rewriting the evolutionary history of inner and outer halo globulars. The first three pages are pretty important if you are a globular fan, and lucidly presented without too much astrojargon.

Second note: NC 6752 is an easy and very pleasing, well resolvable GC in Pavo. It’s now rising into good viewing angles after about midnight. Nearly every star we see sprinkled across the surface is a red giant originally the size of the Sun or smaller. Far away, yet close to home.

About the Campbell et al May 2013 paper, I notice it is strongly weighted towards model studies and deficient in observational data analysis detail. It states, ‘We obtained high resolution spectra (R - 24, 000) for a sample of 20 AGB stars and 24 red giant branch stars . . .’ The observatory data, telescope, spectrograph, and data reduction information is buried as a caption in Fig 2. In most papers that data is an entire section, usually Sec 3. The authors tell us that a FLAMES spectrograph was used, but don’t mention the spectra were direct optics-to-pixel or optical-fibre-fed. We don’t know if the resolution was 512 pixels, or 1024, or 4096. There’s no seeing quality or airmass data, no field star decontamination procedures identified, no mention of RMS scatter (and for that matter, RMS at all), and there’s no mention of isochrone fitting. Instead we are presented with not particularly compelling CMD data points based on 20 stars, while the discussion is weighted towards mathematical models. Twenty stars is an awfully small sample on which to claim overturning a long and very well researched body of data.

In the biblio, it cites two papers with R. C. Gratton et al, but not the most relevant paper to a N6752 discussion, ‘Precise reddening and metallicity of NGC 6752 from FLAMES spectra’ (A&A 440, 901-908 (http://www.aanda.org/index.php?option=com_article&access=bibcode&Itemid=129&bibcode=2005A%252526A...440..901GFU L)). Peruse this paper to see how Campbell et al might have been structured and elaborated. Campbell et al makes no mention of peer-review process or an anonymous referee.

The Monash press release makes some questionable statements, e.g., ‘A new study shows only 30% of stars in a globular cluster will reach old age and become planetary nebulae.’ Of course they don’t: the missing 70% are either still shining, or they evolved along the horizontal branch of the CMD into white dwarfs a long time ago. Only those first-generation red giants with core masses >0.58 solar masses and total masses of <1.2 solar masses evolve to the asymptotic giant stage, in which case the 30% figure is far from a compelling case about cluster longevity.

There are some questionable statements made in the paper itself.

The paper is inconsistent in the first two paragraphs. Paragraph 1 states, ‘It is expected, and predicted by stellar models, that the majority of cluster stars with masses around the current turn-off mass will evolve through the AGB phase. However this has never been confirmed observationally.’ Paragraph 2 then states, ‘We obtained high resolution spectra (R = 24, 000) for a sample of 20 AGB stars and 24 red giant branch stars in the Galactic globular cluster (GC) NGC 6752.’ Hmm, ‘it’s never been confirmed by hundreds of astronomers in dozens of observatories, but we just did’? I’ll be generous and attribute that self-inflation to an editing oversight.

But there’s another contradiction. In the first paragraph they write, ‘we found that every AGB star in our sample has a low sodium abundance, with [Na/Fe] 0.18 dex‘ and the next paragraph gives the opposite impression: ‘The spectral coverage included the strong Na I doublet at 5680 °A’.

This is not news, either. In the 2005 Gratton et al paper cited above, we have:

‘The O-Na anticorrelation in NGC 6752 is well known from previous observations of both TO and subgiant stars (Gratton et al. 2001; Carretta et al. 2005), as well as of red giants (Carretta 1994; Norris & Da Costa 1995; Yong et al. 2003). It is fully confirmed by the present data for RGB bump stars (see Fig. 3). Figure 3 also collects all data available up to now, which clearly shows how an extensive O-Na anticorrelation can be seen along all evolutionary phases.’

There are no other comparative abundance ratios in the paper, nor are there any references to other spectral line strengths beyond Na 5680 nm. The paper leaves aside other fundamental ratios of hydrogen or iron to alpha elements (slow-process neutron-capture elements such as carbon, oxygen, nitrogen, calcium, aluminium, or magnesium). The authors set a fiducial (assumed for the basis of the study) iron-to-hydrogen metallicity −1.54 dex where N6752’s [Fe/H] metallicity is actually measured at -1.62 dex, a low-to-medium metallicity where <-2.0 is typical of first-generation globular stars. Since dex is a logarithmic exponent, the difference between -1.62 and -1.58 metallicity ratio is 10 ∧0.08 greater. That doesn’t sound like much but when it comes to a question like the proportion of stars that evolve into asymptotic red giants, it’s a figure where exactitude makes a difference

I can’t understand why the authors are so preoccupied with [Na/Fe] (sodium to iron). That’s not a particularly noteworthy abundance ratio in first-generation globulars, compared with anomalous CO, NO, or [H/He]. When the authors say, ‘a low sodium abundance, with [Na/Fe] 0.18 dex, indicating they are exclusively first generation stars’, they are stating the already well understood. The Figure 3 data (http://www.aanda.org/index.php?option=com_article&access=bibcode&Itemid=129&bibcode=2005A%252526A...440..901GFU L) in the Gratton paper cited above has this caption:

‘Figure 3: [Na/Fe] ratio as a function of [O/Fe] for stars in NGC 6752. Red filled circles are our RGB bump stars in the present study. Green-filled and open circles are subgiant and turn-off stars, respectively, from Carretta et al. (2005). Literature data are as follow: green diamonds with crosses inside are bright red giants from the extensive study by Yong et al. (2003), and open green triangles are red giant stars from Norris & Da Costa (1995; 6 stars) and Carretta (1994; 4 stars).’

Compared with the above, the Campbell Fig. 2 doesn’t look so very exceptional.

First generation stars are primordial H/He. Sodium doesn’t enter into their elemental make-up until their cores reach ~0.56 solar-masses in the horizontal branch—and very little forms even then, during a process called ‘second dredge-up’ that occurs when stars turn from helium burning to carbon-oxygen burning. Only stars with core masses are greater than 0.58 solar-masses become asymptotic red giants, and their total mass is not very much greater than that, and most 1.2. They begin as hot (30,000K to 10,000K) red-end HB stars covered with convective envelopes much thinner than their envelopes were when ascending to red giants. If a fading red giant’s total mass is >1.2 solar masses, the star evolves to an asymptotic red giant, shedding what remains of its envelopes, leaving behind only a white dwarf and attenuated planetary nebula debris. Next time you see one of those wonderfully coloured photos of a planetary nebula in which the caption lists the elements that make those pretty colors—usually hydrogen, oxygen, and a third element like sulphur or calcium—how often do you see sodium called out for special notice there?

Horizontal branch and AGB stars commence life somewhere between 0.6 and 2.0 solar masses. Up to and through the red giant stage, they lose mass till they are between 0.5 and 1.2 solar masses. We are taught that most white dwarfs are the end result of the red giant/asymptotic red giant process. The word ‘asymptotic’ is used because red giants which reach a total mass of about 1.2 solar masses after arriving at the horizontal branch then contract and ‘go blue’ as their cores start to burn carbon and oxygen, and the envelope (now <1.0 solar mass) re-ascends the CMD on a track very close to the red giant branch. Many think this is where all white dwarfs come from. Not so. Blue horizontal branch stars also can end as white dwarfs. Have a look at M15’s blue HB (http://gclusters.altervista.org/cluster_4.php?ggc=NGC+7078) and blue droop.

The blue HB is short lived, hot (10,000 K to 30,000 K surface temps), and drops almost straight down (What ho! Stars have the blue droop, too) till its component stars’ already-thin envelopes have been ejected and the bright stellar core shows through what resembles a thin cloud layer here on Earth as you look down from a jet airplane and see surface features beneath a cirrus layer. Here is an image plus text (http://spider.seds.org/spider/MWGC/Add/n6752_uit.html) showing only the HB stars in N6753, all of which radiate sharply in the UV.

Figure 1 of the Campbell paper shows only the RGB-AGB sector. N6752 has virtually no red end to the horizontal branch, while 47 Tuc’s red HB is so dominant there’s hardly a blue end at all. A perusal of the CMDs on The Globular Cluster Database (http://gclusters.altervista.org/) reveals that not unusual for globulars to have unique morphologies; CMD anomalies abound. The causes of the anomalies are varied, and many. Stellar rotation speed, the rotational velocity of star’s core with respect to its envelope, and binary mass transfer all affect a star’s temperature and heat loss rate.

To set the larger stage, one of your Aussie authors, Sarah Martell AAO, North Ryde, began a 2013 paper with this: ‘The process of chemical self-enrichment in stellar systems can be affected by the total mass of the system and the conditions of the large-scale environment. Globular clusters are a special dark matter-free case of chemical evolution, in which the only self-enrichment comes from material processed in stars, and only two bursts of star formation occur.’

Campbell et al have selected one cluster and one anomaly, and have neglected two other important anomalies, i.e., carbon-nitrogen and magnesium-aluminium, which could lead to different conclusions than what the authors present. The question of whether any of the study’s stars are binaries, what gas transfer during two-body encounters, and the effect of the cluster’s early relaxation (collapse effects which eject stars) on its initial mass have not been addressed.

In other words, the thesis needs work. And NEVER give a complex paper to a PR department to make a press release.

=Dana in S Africa

[madbadgalaxyman]

Thank you for your interesting comments, Dana.

I can always rely on you for incisive and opinionated commentary about stellar evolution and H-R diagrams!

I am still cogitating and pondering about this paper.
(currently I am just a little distracted with about a million other scientific and technological topics!)

Best Regards
Robert Lang

Good point about the very small numbers of stars used in their study of N6752. Also, reading your comments below, I see you have made the important point that it is quite possible that their derived Color-Magnitude Diagram (Hertzprung-Rusell Diagram) is in error and/or their model is in error.
Stellar evolution models have a habit of being 'further refined' with time; isochrones tend to be quite approximate and model-dependent.