arXiv subjects
Publications and source records attributed to Georges Meynet.
In these lecture notes, we present the equations presently used in stellar interior models in order to compute the effects of axial rotation. We discuss the hypotheses made. We suggest that the effects of rotation might play a key role at low metallicity.
Globular clusters as $ω$ Cen and NGC 2808 appear to have a population of very He-rich stars. From a theoretical point of view, one expects the presence of He-rich stars in all globular clusters showing an oxygen-sodium (O-Na) anticorrelation. In this paper, we briefly recall how fast rotating massive stars could be the main source of the material from which He-rich stars have formed. We speculate that the UV-upturn phenomenon observed in all elliptical galaxies might be due to He-rich stars. If this hypothesis is correct then fast rotating massive stars may have played in the early evolutionary phases of these systems a similar role as the one they played in the nascent globular clusters.
We study the effects of rotation on the outer convective zones of massive stars. We examine the effects of rotation on the thermal gradient and on the Solberg--Hoiland term by analytical developments and by numerical models. Writing the criterion for convection in rotating envelopes, we show that the effects of rotation on the thermal gradient are much larger and of opposite sign to the effect of the Solberg-Hoiland criterion. On the whole, rotation favors convection in stellar envelopes at the equator and to a smaller extent at the poles. In a rotating 20 Msun star at 94% of the critical angular velocity, there are two convective envelopes, with the bigger one having a thickness of 13.2% of the equatorial radius. In the non-rotating model, the corresponding convective zone has a thickness of only 4.6% of the radius. The occurrence of outer convection in massive stars has many consequences.
We recall the basic physical principles governing the evolution of stars with some emphasis on the role played by the nuclear reactions. We argue that in general it is not possible from observations of stars to deduce constraints on the nuclear reaction rates. This is the reason why precise measurements of nuclear reaction rates are a necessity in order to make progresses in stellar physics, nucleosynthesis and chemical evolution of galaxies. There are however some stars which provides useful constraint on nuclear processes. The Wolf-Rayet stars of the WN type present at their surface CNO equilibrium patterns. There is also the particular case of the abundance of $^{22}$Ne at the surface of WC stars. The abundance of this element is a measure of the initial CNO content. Very interestingly, recent determinations of its abundance at the surface of WC stars tend to confirm that massive stars in the solar neighborhood have initial metallicities in agreement with the Asplund et al. (2005) solar abundances.
Mass loss plays a dominant role in the evolution of massive stars at solar metallicity. After discussing different mass loss mechanisms and their metallicity dependence, we present the possibility of strong mass loss at very low metallicity. Our models at Z=1e-8 show that stars more massive than about 60 solar masses may lose a significant fraction of their initial mass in the red supergiant phase. This mass loss is due to the surface enrichment in CNO elements via rotational and convective mixing. Our 85 solar mass model ends its life as a fast rotating WO type Wolf-Rayet star. Therefore the models predict the existence of type Ic SNe and long and soft GRBs at very low metallicities. Such strong mass loss in the red supergiant phase or the Omega-Gamma limit could prevent the most massive stars from ending as pair-creation supernovae. The very low metallicity models calculated are also very interesting from the nucleosynthesis point of view. Indeed, the wind of the massive star models can reproduce the CNO abundances of the most metal-poor carbon-rich star known to date, HE1327-2326. Finally, using chemical evolution models, we are able to reproduce the evolution of CNO elements as observed in the normal extremely metal poor stars.
We review some important observed properties of massive stars. Then we discuss how mass loss and rotation affect their evolution and help in giving better fits to observational constraints. Consequences for nucleosynthesis at different metallicities are discussed. Mass loss appear to be the key feature at high metallicity, while rotation is likely dominant at low and very low metallicities. We discuss various indications supporting the view that very metal poor stars had their evolution strongly affected by rotational mixing. Many features, like the origin of primary nitrogen at low metallicity, that of the C-rich extremely metal poor halo stars, of He-rich stars in massive globular clusters, of the O-Na anticorrelation in globular clusters may be related to the existence of a population of very fast rotating metal poor stars that we tentatively call the {\it spinstars}. A fraction of these {\it spinstars} may also be the progenitors of GRB in very metal poor regions. They may avoid pair instability explosion due to the heavy mass loss undergone during their early evolutionary phases and be, dependent on their frequency, important sources of ionising photons in the early Universe.
We present and discuss evolutionary synthesis models for massive stellar populations generated with the Starburst99 code in combination with a new set of stellar evolution models accounting for rotation. The new stellar evolution models were compiled from several data releases of the Geneva group and cover heavy-element abundances ranging from twice solar to one fifth solar. The evolution models were computed for rotation velocities on the zero-age main-sequence of 0 and 300 km/s and with the latest revision of stellar mass-loss rates. Since the mass coverage is incomplete, in particular at non-solar chemical composition, our parameter study is still preliminary and must be viewed as exploratory. Stellar population properties computed with Starburst99 and the new evolution models show some marked differences in comparison with models obtained using earlier tracks. Since individual stars now tend to be more luminous and bluer when on the blue side of the Hertzsprung-Russell diagram, the populations mirror this trend. For instance, increases by factors of two or more are found for the light-to-mass ratios at ultraviolet to near-infrared wavelengths, as well as for the output of hydrogen ionizing photons. If these results are confirmed once the evolution models have matured, recalibrations of certain star-formation and initial mass function indicators will be required.
After a brief review of the observational evidences indicating how the populations of Be stars, red/blue supergiants, Wolf-Rayet stars vary as a function of metallicity, we discuss the implications of these observed trend for our understanding of the massive star evolution. We show how the inclusion of the effects of rotation in stellar models improves significantly the correspondence between theory and observation.
We study the effect of wind anisotropies on the stellar evolution leading to collapsars. Rotating models of a 60 M$_\odot$ star with $Ω/Ω_{\rm crit}=0.75$ on the ZAMS, accounting for shellular rotation and a magnetic field, with and without wind anisotropies, are computed at $Z$=0.002 until the end of the core He-burning phase. Only the models accounting for the effects of the wind anisotropies retain enough angular momentum in their core to produce a Gamma Ray Burst (GRB). The chemical composition is such that a type Ic supernova event occurs. Wind anisotropies appear to be a key physical ingredient in the scenario leading to long GRBs.
After a review of the many effects of metallicity on the evolution of rotating and non-rotating stars, we discuss the consequences of a high metallicity on massive star populations and on stellar nucleosynthesis. The most striking effect of high metallicity is to enhance the amount of mass lost by stellar winds. Typically at a metallicity $Z=0.001$ only 9% of the total mass returned by non-rotating massive stars is ejected by winds (91% by supernovae explosion), while at solar metallicity this fraction may amount to more than 40%. High metallicity favors the formation of Wolf-Rayet stars and of type Ib supernovae. It however disfavors the occurrence of type Ic supernovae. We estimate empirical yields of carbon based on the observed population of WC stars in the solar neighborhood, and obtain that WC stars eject between 0.2 and 0.4% of the mass initially locked into stars under the form of new synthesized carbon. Models give values well in agreement with these empirical yields. Chemical evolution models indicate that such carbon yields may have important impacts on the abundance of carbon at high metallicity.
Big Bang nucleosynthesis produces only light elements and the very first generation stars are thus formed from metal-free clouds. They start the production of heavy elements during their life, and enrich the interstellar medium through their explosive death. Stellar evolution models show that the treatment of rotation has important effects on the evolution of those metal-free stars: for example, rotating models produce up to five orders of magnitude more primary nitrogen than non rotating models, due to internal mixing. This will have an impact in the composition of the second generation stars, some of which may now be observed in the Galactic halo. In the case Population III stars were very massive and would end up as direct black holes, rotation again have an interesting effect of enhancing mass loss through centrifugal force and surface enrichment. CNO composition patterns observed in ultra metal-poor halo stars may be explained by a 'wind only' contribution.
Some indirect observations, as the high fraction of Be stars at low metallicity, or the necessity for massive stars to be important sources of primary nitrogen, seem to indicate that very metal poor stars were fast rotators. As a consequence of this fast rotation, these stars, contrarily to current wisdom, might lose large amounts of mass during their lifetime. In this paper, we review various mechanisms triggered by rotation which may induce strong mass loss at very low metallicity. The most efficient process comes from surface enrichments in CNO elements which then drive mass loss by stellar winds. Due to this process, a fast rotating 60 M$_\odot$ with metallicities in the range of $Z=10^{-8}$ and $10^{-5}$, can lose between 30 and 55% of its initial mass. This rotationally wind ejected material participates to the chemical evolution of the interstellar medium, enriching it exclusively in H- and He-burning products. In particular, metal poor fast rotating stars may play a key role for explaining the origin of the peculiar abundance pattern observed at the surface of the extremely metal-poor C-rich stars, for explaining the chemical inhomogeneities observed in globular clusters, and the presence of stars in $ω$ Cen with a very high helium content .
The blue Main Sequence (bMS) of $ω$ Cen implies a ratio of helium to metal enrichment $ΔY/ΔZ \approx 70$, which is a major enigma. We show that rotating models of low metallicity stars, which account for the anomalous abundance ratios of extremely metal poor stars, are also useful for understanding the very high $ΔY/ΔZ$ ratio in $ω$ Cen. Models of massive stars with moderate initial rotation velocities produce stellar winds with large He-- and N--excesses, but without the large C-- (and O--) excesses made by very fast rotation, in agreement with the observed chemical abundance ratios in $ω$ Cen. It is still uncertain whether the abundance peculiarities of $ω$ Cen result from the fact that the high velocity contributions of supernovae escaped the globular cluster, usually considered as a tidally stripped core of a dwarf galaxy. Another possibility is a general dominance of wind ejecta at very low $Z$, due to the formation of black holes. Some abundance and isotopic ratios like $Mg/Al$, $Na/Mg$, $Ne/N$, $^{12}C/^{13}C$, $^{16}O/^{18}O$ and $^{17}O/^{18}O$ may allow us to further discriminate between these scenarios and between the AGB and massive star contributions.
The aim of this paper is to determine the initial rotational velocities required on the ZAMS for single stars to reach the critical velocity, sometimes called the break-up velocity, during the Main-Sequence (MS) phase. Some useful relations between $Ω/Ω_{\rm crit}$, $\upsilon/\upsilon_{\rm crit}$ ($\upsilon$ is the velocity at the equator), the moments of inertia, the angular momenta, the kinetic energy in the rotation and various other basic physical quantities are obtained.
At very low metallicity, the effects of differential rotation have a more important impact on the evolution of stars than at high metallicity. Rotational mixing leads to the production of great quantities of helium and of primary $^{14}$N by massive stars. Rotation induces important mass loss and allows stars to locally strongly enrich the interstellar medium in CNO elements. Stars formed from interstellar clouds enriched by the winds of fast rotating massive stars would present surface abundances similar to those of C-rich extremely metal-poor stars. C-rich stars can also be formed by mass accretion in a binary system where the primary would be a fast rotating intermediate mass star in the early-AGB phase. Fast rotation may also lead to the formation of collapsars even at very low metallicity and make the most massive stars avoid the pair instability.
Using single star models including the effects of shellular rotation with and without magnetic fields, we show that massive stars at solar metallicity with initial masses lower than about 20-25 M$_\odot$ and with an initial rotation above $\sim 350$ km s$^{-1}$ likely reach the critical velocity during their Main-Sequence phase. This results from the efficient outwards transport of angular momentum by the meridional circulation. This could be a scenario for explaining the Be stars. After the Main-Sequence phase, single star in this mass range can again reach the critical limit when they are on a blue loop after a red supergiant phase \citep{HL98}. This might be a scenario for the formation of B[e] stars, however as discussed by Langer & Heger (1998), this scenario would predict a short B[e] phase (only some 10$^4$ years) with correspondingly small amounts of mass lost.
The effects of rotation on low-metallicity stellar models are twofold: first, the models reach break-up during main sequence and may lose mass by mechanical process; second, strong internal mixing brings freshly synthesized elements towards the surface and raises the effective metallicity to higher values, so that the initially very low radiative winds are enhanced. Those two effects are also found in Z=0 models, but to a lesser degree because of structural differences. This weak mass loss becomes interesting though in the case of the black hole-doomed stars (25-140 M_\sun and > 260 M_\sun) because it allows these stars to contribute to the enrichment of the medium by stellar winds.
Rotating massive stars at $Z=10^{-8}$ and $10^{-5}$ lose a great part of their initial mass through stellar winds. The chemical composition of the rotationally enhanced winds of very low $Z$ stars is very peculiar. The winds show large CNO enhancements by factors of $10^3$ to $10^7$, together with large excesses of $^{13}$C and $^{17}$O and moderate amounts of Na and Al. The excesses of primary N are particularly striking. When these ejecta from the rotationally enhanced winds are diluted with the supernova ejecta from the corresponding CO cores, we find [C/Fe], [N/Fe],[O/Fe] abundance ratios very similar to those observed in the C--rich extremely metal poor stars (CEMP). We show that rotating AGB stars and rotating massive stars have about the same effects on the CNO enhancements. Abundances of s-process elements and the $^{12}$C/$^{13}$C ratio could help us to distinguish between contributions from AGB and massive stars. On the whole, we emphasize the dominant effects of rotation for the chemical yields of extremely metal poor stars.