Diffusion is believed to occur whenever the outer parts of a star are quiet without large-scale velocity fields and are then not well mixed. Gravity and temperature will tend to concentrate the heavy elements towards the center of the star (Salaris, Groenewegen & Weiss 2000). Diffusion acts very slowly with time scales ~109 yr, so it is most important on the main sequence, and is particularly important for metal-poor stars. Diffusion of He is important in the Sun and affects helioseismology models. Diffusion is also important in precision distance determinations based on main sequence fitting, since the abundances adopted for the model isochrone must agree with those of the stars, while the abundances deduced for the stellar atmosphere may not be those of the interior. Diffusion may also be the solution to a disagreement of a factor of ~2 between the Li abundances derived for halo turn off stars, assumed to be the primordial lithium abundance, and the (lower) value predicted by standard Big Bang nucleosynthesis models that adopt the baryonic density inferred from the current concordance cosmology of WMAP (Melendez & Ramirez 2004; Korn et al. 2006; Bonifacio et al. 2007).
A key project to observe diffusion in action is to compare the elemental abundances for heavy elements near the Fe-peak of main sequence versus red giant and sub-giant stars in metal-poor globular clusters. Stars within a particular globular cluster, assuming its mass is not too high so that gas cannot have been retained beyond the first generation of star formation, are believed to have the same initial chemical inventory, and are sufficiently old that diffusion should have had time to act. The RGB stars have convective envelopes, and thus whatever diffusion might have occurred on the main sequence, the surface helium and heavy elements will have been restored to very close to their initial value, while the main sequence stars will be subject to the predicted larger effects of diffusion for metal-poor stars over their entire lifetime, i.e. that of the Galactic halo, ~13 Gyr. Such observations require high spectral resolution spectroscopy of main sequence stars in globular clusters with SNR difficult or impossible to achieve with existing 8 – 10 m telescopes. The net efficiency gain of HROS on TMT relative to current facilities will enable the required observations of these faint stars.
Massive stars contribute a large fraction of all the heavy elements through SN explosions. They provide the UV ionizing flux for the ISM and their supersonic winds help shape the ISM. They are probably linked to the re-ionization of the early universe (Bromm, Kudritzki & Loeb 2001) and perhaps to the GRB phenomenon. Massive stars of very low metallicity were, until quite recently, believed not to lose very much mass through radiatively driven winds due to their weaker absorption features. This would mean that such stars might often end up as black holes, locking up their heavy element chemical inventory in perpetuity, while a solar metallicity star of similar mass would lose enough mass to end up as a neutron star. However, Meynet et al. (2007) have suggested that these stars are rapid rotators that lose up to 50% of their initial mass through a rotationally driven wind. The mass loss rates they predict for the main sequence phase of a 60 Mʘ star are more than 20 times larger than if rotation were ignored. Coupled with rotationally driven instabilities that transport both angular momentum and chemical species (Zahn 1992), this means that massive low metallicity stars can significantly enrich the ISM in H- and He-burning products.
This theoretical development is very attractive as it solves a number of problems, but the observational evidence to support this development it is essentially non-existent at present. The low metallicity required to produce main sequence rotation rates of ∼600 km/sec, close to the breakup rate at which the effective surface gravity becomes zero, is less than that seen at any place in the Milky Way where star formation is still underway. The metal depletion of the ISM in the LMC and in the SMC is also not sufficient. A population of young massive stars in a low metallicity galaxy is required, i.e. a very metal-poor star-forming dwarf galaxy. There are none close enough to obtain the required spectroscopy with existing 8 – 10 m telescopes. The most metal-poor dwarf galaxy known in the nearby universe, AGC 198691, has a distance of approximately 8 Mpc and an oxygen abundance ~1/47 that of the Sun (Hirschauer et al. 2016). The galaxy SBS 0335–052W with a distance of 22 Mpc is almost as metal-poor (Izotov et al. 2009), with oxygen ~1/36 the solar value. With WFOS on TMT, it will be possible to observe the brightest supergiants in the nearest very metal-poor dwarf galaxies to determine the mass loss rate as a function of metallicity.
There is still no concensus for the formation of low-mass stars from metal-free clouds. Model calculations with extremely low metallicity ([Fe/H] < -5) with excess or no excess of light elements, like carbon suggest the roles of fine structure lines of these elements and of dust emission (Norris et al. 2013). Formation of multiple systems might be a key to understanding the formation processes of low-mass stars in first stars (Machida 2008). Such study requires large sample of extremely metal-poor stars and radial velocity monitoring for them to obtain statistics. Metal-poor stars are extremely rare objects, which makes finding them a great challenge. Large-scale systematic searches began with the HK survey by Beers et al. (1985, 1992), followed by the Hamburg/ESO Survey (Wisotzki et al. 1996, Christlieb et al. 2008) that covered ~ 1000 square degrees of the southern sky collecting data of some 4 million point sources. More recent searches use medium resolution spectroscopy using large multi-object spectrographs (SDSS, SEGUE, LAMOST survey) or selecting metal-poor candidates from photometric survey data. The SkyMapper telescope is photometrically surveying the southern sky in specific filter sets (e.g. Bessell et al. 2011) that allow candidate selection in a very efficient way. Follow-up high resolution spectroscopy is still required for confirmation and detailed chemical abundance studies. The oldest and the most Fe deficient ([Fe/H] < -7.0) star known so far, SMSS J031300.36-670839.3, is a discovery from this survey. Similar metal-poor objects to be found from this and other on-going surveys will provide observational constraints to test theoretical models for low-mass star formation at extremely low metallicity. TMT/WFOS would be useful not only to enhance the extreme metal-poor stars database but also for conducting detailed chemical abundance studies.
TMT equipped with a high resolution spectrograph will provide a unique opportunity to understand the origin and evolution of rare exotic faint stellar objects through detailed chemical abundance studies. For example, observational evidence of the presence of extraordinarily strong Sr features in the hydrogen-deficient carbon (HdC) star HE 1015-2050 from medium resolution spectroscopy (Goswami et al. 2010, Goswami & Aoki 2013) brought the issue of Sr synthesis in stellar interiors to the forefront. To acquire high resolution spectra required for detailed abundance analysis to understand the origin and evolution of such faint objects (Vmag of HE 1015-2050 ~ 16.3) is extremely difficult and time consuming with the existing 8 -10 m class telescopes particularly if such objects happen to undergo significant brightness variations. Similarly, enhancement of fluorine abundance observed in the Extreme helium (EHe) and R Coronae Borealis (RCB) stars (Pandey et al. 2008) by factors of 800 – 8000 relative to its likely initial abundance raises questions about F synthesis mechanisms and the prevailing conditions in the stellar interiors. Measurement of fluorine abundance from the HF band in the near IR in a large sample of K and M giants covering a wide range in metallicities and ages would be essential to understand the origin and evolution of fluorine. Many such potentially interesting objects will form important targets for TMT due both to the large collecting area and more efficient spectrograph.
Since about 2004, the Keck Observatory has spearheaded an effort to measure one of the most fundamental relations of stellar astrophysics, the initial-final mass relation – IFMR v1.0.This relation connects the mass of stellar remnants, white dwarfs, to the mass of their progenitor hydrogen-burning main-sequence stars, and therefore provides a direct way to understand stellar mass loss.
The initial-final mass relation is constrained by spectroscopically measuring the masses of white dwarfs that are members of rich star clusters. The spectroscopy yields the masses of the remnants (Balmer line fitting - Bergeron et al. 1992), and the initial masses are known given the ages of the clusters. Despite the tremendous progress, two of the biggest uncertainties in the relation today are:
The WFOS instrument on TMT offers a unique opportunity to answer these two questions. With Keck/LRIS, our current sample of target star clusters for which spectra of white dwarf members suitable for mass determination can be obtained is limited to the nearest open star clusters within ~3 kpc, and the single nearest globular cluster within ~2 kpc. Within this volume, there is a lack of very young clusters with massive turnoffs to address the first challenge, and a lack of a metallicity spread within the clusters to address the second challenge. With TMT's light collecting area, white dwarf spectroscopy can be pushed to systems at 10 kpc. This volume includes three dozen globular clusters with a wide range of metallicities, as well as numerous rich open clusters with ages from 30 to 100 Myr (i.e., present day evolving mass ranging from 5 to 10 Mʘ). Searching for white dwarfs in these clusters, and linking them to their progenitor masses, will answer these fundamental questions. The key requirements for this program that would directly impact many areas of galactic and extragalactic astrophysics are ultra high throughput at UV wavelengths down to 3600 Angstroms (covering the high order Balmer lines), MOS capability and low resolution spectroscopy with R = 2000 - 4000.
Big Bang nucleosynthesis in the first 20 minutes of the Universe is believed to have created deuterium, the two isotopes of He (3He and 4He) and a very small amount of lithium as well as 1H. Almost all other elements in the periodic table are synthesized in the stellar interiors and envelopes during hydrostatic and explosive burning.
Each stellar nucleosynthetic path has a different timescale and produces characteristic elemental abundance patterns. Interpretation of the observed abundance ratios and abundance patterns in the framework of a proper galactic chemical evolution model allows to trace back the nucleosynthetic origin and the prevailing astrophysical conditions. Chemical evolution in differing stellar populations traces the star formation history and age and provides insight into the chemical evolution of galaxies and their interstellar matter. For accurate abundance measurements high spectral resolution (R >20,000) with a good S/N (>50) ratio are required, especially to detect and analyze faint spectral features. As compared to low and moderate resolution, at higher resolution many important absorption lines become available for study. At resolution R >90,000 (S/N ~200) estimates of isotope abundances become possible. TMT equipped with a high resolution spectrograph will have a dramatic impact in nucleosynthesis studies as it will aid detection of faint spectral features and also open up hitherto unreachable classes of stars for study.
While elements are formed inside stars, some light elements such as D, Li, Be are also destroyed inside stars, under certain conditions. The detection and abundance estimate of these elements is important to understand the stellar structure, evolution and mixing process in stars and the amount of stellar processing. We discuss a few outstanding questions related to stellar nucleosynthesis, answers to which can be sought from TMT observations.
WMAP determination of the cosmic baryon density, combined with the Big Bang Nucleosynthesis (BBN) theory, tightly predicts a Li abundance value that differs from the value of `Spite Plateau' (Spite & Spite 1982) by a factor of two or more. Several possible solutions, such as stellar destruction and astration, nuclear uncertainties and new physics, atomic diffusion etc. have been proposed to explain this discrepancy. The Li abundances observed in hyper metal-poor stars that are still in the main-sequence turn-off phases (Aoki et al. 2006, Caffau et al. 2011) might provide a clue to this problem. High resolution and high sensitivity of TMT/HROS are required for measurements of low Li abundance and Li isotope ratios in hyper metal-poor stars as well as in extreme metal-poor stars. An understanding of the Li problem and depletion processes inside stars may be achieved better from a detailed study of metal poor globular clusters. Estimating the Li abundance in the faint main-sequence stars is very challenging with existing 8 to 10 meter telescopes but will be relatively easy with TMT. Correlating the Li abundance with other heavy elements will help understand the discrepancy between the observed Li abundance compared to WMAP results.
C, N and O are produced in all types of stars during hydrogen and helium burning. These elements play an important role in star formation by cooling the clouds through several fine structure lines. The observational evidence of the increase of C with decrease in the metallicity may possibly be an evidence of a gradual change in the characteristic mass of the IMF, which is likely to be mediated by CMB temperatures acting as a minimum temperature (Tumlinson 2007). With the TMT it will be possible to look for such correlations in the local dwarf galaxies and test the theoretical prediction.
Be abundances in the Galaxy and satellites of the Milky Way provide constraints on the pre-Galactic cosmic ray fluxes and cosmic magnetic fields. If Be and B are produced as secondary elements, one would expect the Be abundance to decrease quadratically with respect to metallicity. However, the observation of a small sample of metal-poor stars show that there is a linear decrease with metallicity (Primas et al. 2000), indicating a primary production of C, N, O and Be from the same mechanism. The TMT will enable such studies to fainter limits, extending the sample beyond the Milky Way to the local group dwarfs galaxies.