arXiv subjects
Publications and source records attributed to Falk Herwig.
While the basic properties of AGB stellar evolution are well established, comprehensive observational studies of late phases of intermediate mass stars continue to generate puzzles for current stellar models. Here, I review current techniques to model AGB stars, and I discuss important aspects of current research of AGB (and post-AGB) stellar evolution with a particular focus on how these interrelate.
(abridged) We model the nucleosynthesis during the thermal pulse phase of a rotating, solar metallicity AGB star of 3M_sun. Rotationally induced mixing during the thermal pulses produces a layer (~2E-5M_sun) on top of the CO-core where large amounts of protons and C12 co-exist. We follow the abundance evolution in this layer, in particular that of the neutron source C13 and of the neutron poison N14. In our AGB model mixing persists during the entire interpulse phase due to the steep angular velocity gradient at the core-envelope interface. We follow the neutron production during the interpulse phase, and find a resulting maximum neutron exposure of tau_max =0.04 mbarn^-1, which is too small to produce any significant s-process. In parametric models, we then investigate the combined effects of diffusive overshooting from the convective envelope and rotationally induced mixing. Models with overshoot and weaker interpulse mixing - as perhaps expected from more slowly rotating stars - yield larger neutron exposures. We conclude that the incorporation of rotationally induce mixing processes has important consequences for the production of heavy elements in AGB stars. Through a distribution of initial rotation rates it may lead to a natural spread in the neutron exposures obtained in AGB stars of a given mass - as appears to be required by observations. Our results suggest that both processes, diffusive overshoot and rotational mixing, may be required to obtain a consistent description of the s-process in AGB stars which fulfils all observational constraints. Finally, we find that mixing due to rotation within our current framework does increase the production of N15 in the partial mixing zone, however still falling short of what seems required by observations.
We present full stellar evolution and parametric models of the surface abundance evolution of O16, Ne22, Na23 and the magnesium isotopes in an extremely metal-poor intermediate mass star M_ZAMS=5M_sun, Z=0.0001. O16 and Ne22 are injected into the envelope by the third dredge-up following thermal pulses on the asymptotic giant branch. These species and the initially present Mg24 are depleted by hot bottom burning (HBB) during the interpulse phase. As a result, Na23, Mg25 and Mg26 are enhanced. If the HBB temperatures are sufficiently high for this process to deplete oxygen efficiently, Na23 is first produced and then depleted during the interpulse phase. Although the simultaneous depletion of O16 and enhancement of Na23 is possible, the required fine tuning of the dredge-up and HBB casts some doubt on the robustness of this process as the origin of the O-Na anti-correlation observed in globular cluster stars. However, a very robust prediction of our models are low Mg24/Mg25 and Mg24/Mg26 ratios whenever significant O16 depletion can be achieved. This seems to be in stark contrast with recent observations of the magnesium isotopic ratios in the globular cluster NGC6752.
We model the nucleosynthesis during a radiative interpulse phase of a rotating 3Msun Asymptotic Giant Branch (AGB) star. We find an enhanced production of the neutron source species C13 compared to non-rotating models due to shear mixing of protons and C12 at the core-envelope interface. We estimate that the resulting total production of heavy elements by slow neutron capture s-process is too low to account for most observations. This due to the fact that rotationally induced mixing during the interpulse phase causes a pollution of the C13 pocket layer with the neutron poison N14. As a result we find a maximum neutron exposure of tau_max=0.04mbarn^{-1} in the s-process layer of our solar metallicity model with rotation. This is about a factor of 5 to 10 less than required to reproduce the observed stellar s-process abundance patterns. We compare our results with models that include hydrodynamic overshooting mixing, and with simple parametric models including the combined effects of overshooting and mixing in the interpulse. Within the parametric model a range of mixing efficiencies during the interpulse phase correlates with a spread in the s-process efficiency. Such a spread is observed in AGB and post-AGB stars as well as in pre-solar SiC grains.
The oldest stars hide information about the chemical evolution of the early universe. In this contect new models of the AGB stellar evolution phase at very low and zero metallicity are presented. Due to the deficiency or absence of CNO catalytic material hydrogen burning operates at significantly higher temperatures than in stars of higher metallicity. As a result convective mixing plays a very important role since the nuclear burning shells are not well separated by an entropy barrier. These stars can host flash-like burning events induced by mixing of C12 and protons at the core-envelope interface. Three different mixing events with proton-capture nucleosynthesis can be encountered. One of them, the hot dredge-up, is reported here for the first time. All these specific modes of nucleosynthesis are relevant for the production of CNO material in the first intermediate mass stars.
[abridged] We study the s-process in AGB stars using three different stellar evolutionary models computed for a 3Msun and solar metallicity star. First we investigate the formation and the efficiency of the main neutron source. We parametrically vary the number of protons mixed from the envelope into the C12 rich core. For p/C12 > 0.3, mainly N14 is produced, which represent a major neutron poison. The amount of C12 in the He intershell and the maximum value of the time-integrated neutron flux are proportional. Then we generate detailed s-process calculations on the basis of stellar evolutionary models constructed with three different codes. One code considers convective hydrodynamic overshoot that depends on a free parameter f, and results in partial mixing beyond convective boundaries, the most efficient third dredge up and the formation of the C13 pocket. For the other two codes an identical C13 pocket is introduced in the post-processing nucleosynthesis calculations. The models generally reproduce the spectroscopically observed s-process enhancements. The results of the cases without overshoot are remarkably similar. The code including hydrodynamic overshoot produces a He intershell composition near to that observed in H-deficient central stars of planetary nebulae. As a result of this intershell dredge up the neutron fluxes have a higher efficiency, both during the interpulse periods and within thermal pulses. The s-element distribution is pushed toward the heavier s-process elements and large abundances of neutron-rich isotopes fed by branching points in the s-process path are produced. Several observational constraints are better matched by the models without overshoot. Our study need to be extended to different masses and metallicities and in the space of the free overshoot parameter f.
H-deficient post-AGB objects, e.g. PG1159 type star K1-16 and born-again AGB star Sakurai's object, have been reported to be significantly iron-deficient. We find that the iron deficiencies expected due to neutron-capture nucleosynthesis during either the progenitor AGB evolution and/or the neutron burst that occurs as a result of the rapid burning of protons during a post-AGB He-flash are generally in line with observations.
The carbon and nitrogen isotopic ratios of pre-solar SiC grains of type A+B suggest a proton-limited nucleosynthetic process as encountered, for instance, during the very late thermal pulse of post-AGB stars. We study the nuclear processes during this phase and find carbon and nitrogen isotopic ratios which can reproduce those of A+B grains. These results are still preliminary because they depend on uncertain factors such as the details of mixing during the post-AGB thermal pulse, the rates of some nuclear reactions, and the assumptions on mixing during the progenitor AGB phase.
We present a stellar evolution model sequence of the past and future evolution of the post-AGB born again star Sakurai's object (V4334 Sgr). In order to match the short evolutionary time scale we have to assume that the convective ingestion of hydrogen-rich envelope material into the He-flash convection zone proceeds slower than predicted by the mixing length theory. For the future we predict a swift second evolution through the central star region before a second born-again evolution occurs.
We discuss the occurrence of the s-process during the radiative interpulse phase of rotating AGB stars. Due to differential rotation, protons are mixed into C12-rich layers after thermal pulses, in the course of the so called third dredge up episode. We follow the time evolution of key isotope abundances in the relevant layers with a post-processing code which includes time dependant mixing and nucleosynthesis. In rotating AGB models, the mixing persists during the entire interpulse phase due to the steep gradient of angular velocity at the envelope-core interface. As the layers containing protons and C12, which are formed this way, become hotter, a C13-pocket is formed in a natural way. However, in this situation also N14 is formed and spread over the entire C13-pocket. We include the neutron consuming N14(n,p) reaction in our network and determine to what extent it reduces the production of trans-iron elements. We propose that rotation may be responsible for the spread of efficiencies of the C13 neutron source as required by observations.
Sakurai's object (V4334 Sgr) is a born again AGB star following a very late thermal pulse. So far no stellar evolution models have been able to explain the extremely fast evolution of this star, which has taken it from the pre-white dwarf stage to its current appearance as a giant within only a few years. A very high stellar mass can be ruled out as the cause of the fast evolution. Instead the evolution time scale is reproduced in stellar models by making the assumption that the efficiency for element mixing in the He-flash convection zone during the very late thermal pulse is smaller than predicted by the mixing-length theory. As a result the main energy generation from fast proton capture occurs closer to the surface and the expansion to the giant state is accelerated to a few years. Assuming a mass of V4334 Sgr of 0.604Msun -- which is consistent with a distance of 4kpc -- a reduction of the mixing length theory mixing efficiency by a factor of ~ 100 is required to match its evolutionary time scale. This value decreases if V4334 Sgr has a smaller mass and accordingly a smaller distance. However, the effect does not disappear for the smallest possible masses. These findings may present a semi-empirical constraint on the element mixing in convective zones of the stellar interior.
Sakurai's object is a born again AGB star of the very late thermal pulse flavor. In this contribution I will discuss new models of stellar evolution and nucleosynthesis models of this phase. Two most intriguing properties of Sakurai's objects have so far not been understood theoretically: the peculiar chemical appearance, in particular the high lithium abundance and the short time scale of only a few years on which the transition from the dwarf configuration into the born again giant appearance has occurred. A new nucleosynthesis mode of hot hydrogen-deficient He3 burning can explain the extraordinary lithium abundance. During the thermal pulse He3 is ingested from the envelope together with the protons into the hot He-flash convection zone. The first network calculations show that due to the large C12 abundance protons are rather captured by carbon than destroy newly formed Be7 and ultimately Li7. Moreover, the short evolution time scale has been reproduced by making the assumption that the convective efficiency for element mixing is smaller by two to three orders of magnitude than predicted by the mixing-length theory. As a result the main energy generation from fast convective proton capture will occur at a larger mass coordinate, closer to the surface and the expansion to the giant state is accelerated to a few years in agreement with Sakurai's behavior. This result may represent an independent empirical constraint on the poorly known efficiency of element mixing in convective zones of the stellar interior.
I discuss recent new models of post-Asymptotic Giant Branch stellar evolution. These models aim to clarify the evolutionary origin and status of a variety of hydrogen-deficient post-AGB stars such as central stars of planetary nebulae of Wolf-Rayet spectral type, PG1159 stars or Sakurai's object. Starting with AGB models with overshoot such stars can evolve through one of four distinct channels. Each of these channels has typical abundance patterns depending on the relative timing of the departure from the AGB and the occurrence of the last thermal pulse. I discuss the responsible mechanisms and observational counterparts.
A thermal pulse during the post-AGB phase of stellar evolution may lead to a unique mode of light element nucleosynthesis. The stage is set by the ingestion of the unprocessed envelope material into the hot He-flash convection zone below. If the temperature is sufficiently large and the C12 abundance high enough (e.g. T_8 > 0.8, X(C12) ~ 0.4 and X(H) ~ 1E-3) protons react faster with C12 and form C13 than destroying Be7 The latter forms by alpha-capture of He3 after an initial reduction of the He3 abundance to about 3E-5 X(He4) by the ppI reaction He3(He3,2p)He4 (for T_8 ~ 1). All He3 is burned within minutes to weeks depending on the temperature. Be7 is now present at about the previously mentioned level of He3. Its further fate is determined by the reactions Be7(e-,nu)Li7 and the alpha-capture reactions of Be7 and Li7. These captures lead to the production of Be11 which in turn is finally destroyed by Be11(alpha,n)N14$. The details of this mechanism of light element production in real stars is expected to be fairly dependent on the description of mixing.
Within the framework of the current models for the slow neutron capture (s) process in Asymptotic Giant Branch (AGB) stars we explore the uncertainties introduced by the assumptions made on stellar modeling. On the basis of star models constructed with three different evolutionary codes we generate detailed neutron-capture nucleosynthesis post-processing models. The main difference among the codes is that one of them includes an overshooting mechanism. As a result, the neutron fluxes are stronger both during the interpulse periods and within thermal pulses. The efficiency of the C13 source is also studied and we find that a linear relationship exists between the initial C12 in the intershell and the maximum number of neutrons produced.
(abridged) Observations clearly show that low-mass AGB stars can provide a nucleosynthesis site of the s-process. Recent stellar evolution models indicate that radiative burning of C13 between thermal pulses in low-mass AGB stars may indeed provide the needed neutrons. Some mixing between the proton-rich envelope and the carbon-rich core may lead to the production of C13. However, the responsible physical mechanism is not yet unambiguously identified. We present stellar model calculations with overshoot and rotation. Overshoot, with a time-dependent and exponentially decaying efficiency, leads to a partial mixture of protons and C12 during the third dredge-up. According to the depth-dependent ratio of protons and C12, a small C13-pocket forms underneath a N14-rich layer. Overshoot does not allow for any mixing during the interpulse phase. Rotation introduces mixing driven by large angular velocity gradients which form at the envelope-core interface in AGB stars, in particular after a thermal pulse. This leads to partial mixing after a pulse, as in the case of overshoot. However, rotation continues to mix the region of the C13-pocket with a diffusion coefficient of log D ~ 2...3 [cm**2 / s]. This does not only spread the C13-pocket, but also the more massive N14-rich layer, and finally leads to mixture of both layers. By the time when the temperature there has risen to about 9*10E7K and neutron production sets in, the N14 abundance exceeds the C13 abundance by a factor of 5...10. We analyze the role of N14 as a neutron poison by considering the recycling of neutrons via N14(n,p)C14 and C12(p,gamma)N13(beta+)C13 qualitatively. We find that as long as X(N14) << X(C12), the s-process will still be possible to occur under radiative conditions.
The influence of extended convective mixing (overshoot) on asymptotic giant branch stellar evolution is investigated in detail. The extended mixing is treated time-dependently, and the efficiency declines exponentially with the geometric distance from the convective boundary. It has been considered at all convective boundaries, including the He-flash convection zone in the intershell region which forms during the thermal pulses. Both the structural and the chemical evolution are affected by the inclusion of overshoot. The main results include a very efficient third dredge-up which leads to the formation of carbon stars of low mass and luminosity. A C13 pocket which may serve as a neutron source for the s-process can form after the third dredge-up has reached into the C12 rich intershell. Overshoot applied to the pulse-driven convective zone during the He-flash leads to a deeper penetration of the bottom of this convective zone into the C/O core below the He-burning shell. This in turn causes He4 to be less abundant in the intershell while C12 and O16 are more abundant compared to calculations without overshoot. We show that overshoot at the He-flash convection zone as well as at the base of the envelope convection enhance the efficiency of the third dredge-up. Characteristic properties of the structural and chemical evolution of AGB stars are presented.
We give a brief review on the properties of asymptotic giant branch models with overshoot. Then we describe new model calculations with overshoot. Initial masses are ranging from 1 to 3Msun and metallicities are Z=0.02, 0.01 and 0.001. Third dredge-up occurs efficiently for low masses and carbon stars are formed, with some at core masses as low as 0.58Msun. After the thermal pulse at which stars become C-rich the luminosities are in the range of the observed C-star luminosity function during the whole interpulse phase and for all C-star models. The dredge-up evolution depends mainly on the core mass at the first thermal pulse and on the metallicity. The Z=0.001 models of the 2 and 3Msun sequence become C-rich almost instantaneously after the onset of the first thermal pulses. For the 2Msun case the C/O ratio initially exceeds 4. During following dredge-up episodes the C/O ratio decreases.