Chemical enrichment by Wolf–Rayet stars: non-solar metallicities

Following on from our recent Paper I, we present theoretical models of Wolf–Rayet (WR) stars for non-solar metallicities from   Z = 0.03  to 0.0001 by mass fraction with different mass-loss rate assumptions. We find that some single WR stars may still form even at the lowest metallicities, but whether this occurs or not depends critically on the upper cut-off point of the initial mass function used. As at solar metallicity, a population of binaries is required to fully reproduce WR star observations. For most scenarios, these binaries dominate the low-metallicity WR population but probably not the enrichment. We find comparable carbon enrichment from single WR stars to that from asymptotic giant branch stars at all metallicities for which data are available, but which of them is the dominant source of carbon depends strongly on the set of asymptotic giant branch yields adopted and the assumed initial mass function. We find an increase in carbon enrichment with increasing metallicity but a decrease in oxygen enrichment, as confirmed by observation.     

Chemical enrichments by massive stars and the effects of rotation

All model output of massive star evolution, i.e. the paths in the HR diagram, the lifetimes, the surface and wind composition, the chemical yields and the nature of the progenitors of supernovae strongly depend on both mass loss and rotation. We examine here specifically the effects of mass loss and rotation on the chemical yields of CNO elements. A major effect is that the importance of mass loss depends on the metallicity Z of the galaxies. This may also be the case for stellar rotation in massive stars, in the sense that average rotation may be faster at lower metallicities.     

Outflows from massive blue stars

We present in this contribution a revision of the origin, main properties and open issues in the field of winds of massive blue stars, with a particular emphasis in the ultraviolet observations     

Synthetic HR-diagrams for massive stars

Observations of luminous stars are compared to synthetic HR diagrams generated from three sets of evolutionary models with different treatment of mass loss, convective overshoot, and considering or not variations in the opacity due to heavy elements. We take into account also different values for the slope of the initial mass function, and the possibility that a fraction of the earliest spectral type stars is missing either because of photometric incompleteness or because they spend part of their Main-Sequence lifetime being not visible as still embedded in their parental cloud (shortly abbreviated as phase of invisibility). The statistical comparison between theory and observations is performed by means of a two-dimensional Kolmogorov-Smirnov test. Models computed taking into account a moderate opacity enhancement in the region of the CNO ionization, in addition to mass loss and convective overshoot, can be accepted at the 95% significance level. However, even in this case, the distribution of luminous stars in the HR diagram can be fitted only adopting an initial mass function slightly steeper than the Salpter one, and supposing that a fraction of the luminous stars are still not included in our counts either by photometric incompleteness of the data or intrinsic invisibility. On the contrary, the other models that incorporate mass loss and convective overshoot only, can be rejected at better than 95% confidence under any assumption for the IMF slope and fraction of missing stars.     

Presupernova structure of massive stars

Issues concerning the structure and evolution of core collapse progenitor stars, and stellar evolution in general, are discussed with an emphasis on interior evolution. We discuss some recent results that address quantifying the uncertainties inherent in modern stellar evolution calculations, and we describe a research effort aimed at investigating the transport and mixing processes associated with stellar turbulence, which is arguably the greatest source of uncertainty in supernova progenitor structure, besides mass loss, at the time of core collapse. We highlight the important role played by precision observations of stellar parameters in constraining theoretical models, as well as the physical insight that can be garnered from three-dimensional hydrodynamic simulation.     

Tidal disruption of stars by a massive black hole

The hydrodynamics of stellar models in the tidal field of a massive black hole is calculated numerically under certain simplifying assumptions. It is found that solar type stars are totally disrupted only when falling toward a black hole whose mass does not exceed 10⁶–10⁷ M . Red giants lose their entire envelope for all black hole masses studied.Our findings strengthen the view that tidal disruption is not an important source of gas in AGN's.     

The production of circumstellar26Al by massive stars

We investigate the production of²⁶Al during hydrogen burning and its ejection by massive single and binary stars. Effects of convection and rotation are studied. We discuss the importance of RSGs, LBVs and WR stars to the total Galactic²⁶Al production, and the detection probability of the²⁶Al decay in individual objects as P Cygni, Velorum and Carinae.     

Mass-luminosity relation for massive stars

A catalog of massive (⩾10 M _⊙) stars in binary and multiple systems with well-known masses and luminosities has been compiled. The catalog is analyzed using a theoretical mass-luminosity relation. This relation allows both normal main-sequence stars and stars with peculiarities: with clear manifestations of mass transfer, mass accretion, and axial rotation, to be identified. Least-squares fitting of the observational data in the range of stellar masses 10M _⊙ ⩽ M ≲ 50 M _⊙ yields the relation L ∼ M ^2.76. An erratum to this article is available at .     

Production of dust by massive stars at high redshift

The large amounts of dust detected in sub-millimeter galaxies and quasars at high redshift pose a challenge to galaxy formation models and theories of cosmic dust formation. At z>6 only stars of relatively high mass (>3 M_⊙) are sufficiently short-lived to be potential stellar sources of dust. This review is devoted to identifying and quantifying the most important stellar channels of rapid dust formation. We ascertain the dust production efficiency of stars in the mass range 3–40 M_⊙ using both observed and theoretical dust yields of evolved massive stars and supernovae (SNe) and provide analytical expressions for the dust production efficiencies in various scenarios. We also address the strong sensitivity of the total dust productivity to the initial mass function. From simple considerations, we find that, in the early Universe, high-mass (>3 M_⊙) asymptotic giant branch stars can only be dominant dust producers if SNe generate ≲3×10⁻³ M_⊙ of dust whereas SNe prevail if they are more efficient. We address the challenges in inferring dust masses and star-formation rates from observations of high-redshift galaxies. We conclude that significant SN dust production at high redshift is likely required to reproduce current dust mass estimates, possibly coupled with rapid dust grain growth in the interstellar medium.     

Radial pulsations of massive main-sequence stars

The evolution of Population I stars (X = 0.7, Z = 0.02) with initial masses 40M _⊙ ≤ M _ZAMS ≤ 120M _⊙ until core hydrogen exhaustion has been computed. Models of evolutionary sequences have been used as the initial conditions in solving the equations of radiation hydrodynamics that describe the spherically symmetric motion of a self-gravitating gas. Stars with initial masses M _ZAMS ≥ 50M _⊙ are shown to become unstable against radial oscillations during the main-sequence evolution. The instability growth rate and the limit-cycle oscillation amplitude increase as the star evolves and as its initial mass increases. The pulsational instability is attributable to the iron Z-bump κ mechanism (T ∼ 2 × 10⁵ K). Convection that transfers from 20 to 50% of the total energy flux and, thus, reduces the efficiency of the κ mechanism emerges in the same layers. The periods of the radial oscillations of main-sequence stars lie within the range from 0.09 to 8 days. The boundaries of the instability region of radial pulsations in the Hertzsprung-Russell diagram have been determined and observational criteria for revealing pulsating variable main-sequence stars have been proposed.     

On the formation of massive stars

We present a model for the formation of massive ( M ≳10 M⊙) stars through accretion-induced collisions in the cores of embedded dense stellar clusters. This model circumvents the problem of accreting on to a star whose luminosity is sufficient to reverse the infall of gas. Instead, the central core of the cluster accretes from the surrounding gas, thereby decreasing its radius until collisions between individual components become sufficient. These components are, in general, intermediate-mass stars that have formed through accretion on to low-mass protostars. Once a sufficiently massive star has formed to expel the remaining gas, the cluster expands in accordance with this loss of mass, halting further collisions. This process implies a critical stellar density for the formation of massive stars, and a high rate of binaries formed by tidal capture.     

Sensitivity in the models of massive stars

Two related problems are discussed in this article: The width of the Main-Sequence of massive stars and sensitivity peatures introduced into the evolutionary tracks of massive stars by mass loss and core-overshooting. It is suggested that core-overshooting may not necessarily be implied by the observations of the width of the Main-Sequence band. It is also noted that models evolved with both mass loss and/or core-overshooting reveal the presence of a large and unexplained expansion of the stellar models under certain conditions. This sensitivity feature would seem to be a fundamental feature inherent to the structure of massive stars.     

Rates of tidal disruption of stars by massive central black holes

There is strong evidence for some kind of massive dark object in the centres of many galaxy bulges. The detection of flares from tidally disrupted stars could confirm that these objects are black holes (BHs). Here we present calculations of the stellar disruption rates in detailed dynamical models of real galaxies, taking into account the refilling of the loss cone of stars on disruptable orbits by two-body relaxation and tidal forces in non-spherical galaxies. The highest disruption rates (one star per 10⁴ yr) occur in faint ( L ≲10¹⁰ L_⊙) galaxies, which have steep central density cusps. More luminous galaxies are less dense and have much longer relaxation times and more massive BHs. Dwarf stars in such galaxies are swallowed whole by the BH and hence do not emit flares; giant stars could produce flares as often as every 10⁵ yr, although the rate depends sensitively on the shape of the stellar distribution function. We discuss the possibility of detecting disruption flares in current supernova searches. The total mass of stars consumed over the lifetime of the galaxy is of the order of 10⁶ M_⊙, independent of galaxy luminosity; thus, disrupted stars may contribute significantly to the present BH mass in galaxies fainter than ∼10⁹ L_⊙.