Evolution and Nucleosynthesis of Extremely Metal Poor & Metal-Free Low- and Intermediate-Mass Stars I: Stellar Yield Tables and the CEMPs
aa r X i v : . [ a s t r o - ph . S R ] J a n Astronomy&Astrophysicsmanuscript no. 9597 c (cid:13)
ESO 2018October 29, 2018
Evolution and Nucleosynthesis of Extremely Metal Poor& Metal-Free Low- and Intermediate-Mass StarsI: Stellar Yield Tables and the CEMPs ⋆ S. W. Campbell , and J. C. Lattanzio Academia Sinica Institute of Astronomy and Astrophysics, P.O. Box 23-141, Taipei 10617, Taiwane-mail: [email protected] Centre for Stellar and Planetary Astrophysics, School of Mathematical Sciences, Monash University, Melbourne, Australia 3800e-mail: [email protected]
Received 18 February 2008 / Accepted 23 August 2008
ABSTRACT
Context.
The growing body of spectral observations of the extremely metal-poor (EMP) stars in the Galactic Halo provides constraintson theoretical studies of the chemical and stellar evolution of the early Universe.
Aims.
To calculate yields for EMP stars for use in chemical evolution calculations and to test whether such models can account forsome of the recent abundance observations of EMP stars, in particular the highly C-rich EMP (CEMP) halo stars.
Methods.
We modify an existing 1D stellar structure code to include time-dependent mixing in a di ff usion approximation. Usingthis code and a post-processing nucleosynthesis code we calculate the structural evolution and nucleosynthesis of a grid of modelscovering the metallicity range: − . ≤ [Fe / H] ≤ − . = . ≤ M ≤ . ⊙ , amounting to 20 stars in total. Results.
Many of the models experience violent nuclear burning episodes not seen at higher metallicities. We refer to these eventsas ‘Dual Flashes’ since they are characterised by nearly simultaneous peaks in both hydrogen and helium burning. These events havebeen reported by previous studies. Some of the material processed by the Dual Flashes is dredged up causing significant surfacepollution with a distinct chemical composition. We have calculated the entire evolution of the Z = / H] . − .
0) we find the yields to contain ∼ / H] ∼ − . / Fe] values consistent with those observed in the most C-rich CEMPs. However it is only the low-massmodels that undergo the Dual Shell Flash (which occurs at the start of the TPAGB) that can best reproduce the C and
N observations.Normal Third Dredge-Up can not reproduce the observations because at these metallicities intermediate mass models ( M & ⊙ )su ff er HBB which converts the C to N thus lowering [C / N] well below the observations, whilst if TDU were to occur in the low-mass( M ≤ ⊙ ) models (we do not find it to occur in our models), the yields would be expected to be C-rich only, which is at odds withthe ‘dual pollution’ of C and N generally observed in the CEMPs. Interestingly events similar to the EMP Dual Flashes have beenproposed to explain objects similarly containing a dual pollution of C and N – the ‘Blue Hook’ stars and the ‘Born Again AGB’ stars.We also find that the proportion of CEMP stars should continue to increase at lower metallicities, based on the results that some of thelow mass EMP models already have polluted surfaces by the HB phase, and that there are more C-producing evolutionary episodes atthese metallicities. Finally we note that there is a need for multidimensional fluid dynamics calculations of the Dual Flash events, toascertain whether the overproduction of C and N at ultra-low metallicities found by all studies is an artifact of the 1D treatment. Key words.
Stars: evolution – Stars: interiors – Galaxy: halo – Stars: AGB and post-AGB
1. Introduction
In terms of chemical pollution the Extremely Metal Poor (EMP, [Fe / H] . − .
5) Galactic Halo stars are the most ancient objectscurrently known in the Universe. Observations show that their metallicities reach as low as [Fe / H] = − . α systems for example. Studying the EMP stars is thus crucial to understandingthe chemical evolution of the early Universe. They provide the best link we have to the elusive first generation of stars (Pop III).Indeed, the surface abundances of individual EMP stars are enriched to such a small degree that they may reflect the compositionof the ejecta of only a few (or even one) Pop III supernovae. The chemical information from these stars should eventually aid inour understanding of the First Stars, providing indirect evidence for the chemical evolution of the Galaxy and, importantly, the firststellar mass function (FMF), which is still very uncertain. Deducing the FMF is also important in terms of the Epoch of Reionisation.For these reasons this field is often referred to as ‘Near Field Cosmology’. The nearness of the EMP stars (relative to ancient objectsat high redshift) also means that extremely detailed information can be collected from their spectra using current instruments. ⋆ Tables 1 to 6 are only available in electronic form at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or viahttp: // cdsweb.u-strasbg.fr / cgi-bin / qcat?J / A + A / S. W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars
The discovery of Galactic Halo EMPs has naturally led to a renewed interest in the theoretical modelling of Population III andlow-metallicity stars. In particular the subset of these ancient stars that is observed to contain large amounts of carbon, the C-richEMPs (CEMPs, which we define here as stars with [C / Fe] > + .
7, also see Fig. 5) has attracted much stellar modelling work sincetheir abundance patterns are di ffi cult to explain with standard stellar evolution. These interesting objects also appear to comprise alarge proportion of the EMPs ( ∼ → = = ff er H-ingestion during the core He flash at the tip of the Red GiantBranch (D’Antona 1982). About ten years later this was confirmed by detailed simulations (Fujimoto et al. 1990; Hollowell et al.1990). The H-ingestion during these evolutionary events – which are peculiar to models of low-mass Z = ffi et al. 2001) and explored the possibility of initiating the s-process at zero metallicity (Goriely & Siess2001).Although some grids of EMP and Z = / or have not calculated all of the AGB evolution. In the current studywe aim to provide a homogeneous set of yields that include the nucleosynthetic e ff ects of the EMP H-ingestion events (which werefer to below as ‘Dual Flashes’), AGB Third Dredge-Up (TDU, the periodic mixing up of He-burning products after a thermalpulse) and AGB Hot Bottom Burning (HBB, the H-burning of the convective envelope). To this end we have undertaken a broadexploration of EMP ( − . ≤ [Fe / H] ≤ − .
0) and Z = . ≤ M ≤ . ⊙ ). We include (for the first time) evolutionary and nucleosynthesis calculations from ZAMS to the end of thethermally-pulsing AGB phase (TPAGB) involving 74 species, as well as chemical yield tables. With this homogeneous set of modelswe hope to shed some light on whether or not 1D stellar models can help to explain some of the EMP halo star observations, and inparticular the CEMP abundance patterns.
2. Method
Our simulations were performed utilising two numerical codes – a stellar structure code and post-processing nucleosynthesis code.The stellar structure code used was the Monash version of the Monash / Mount Stromlo stellar evolution code (MONSTAR, seeeg. Wood & Zarro 1981; Frost & Lattanzio 1996). The code is largely a standard 1D code that utilises the Henyey-matrix method(a modified Newton-Raphson method) for solving the stellar structure equations. For the present study the instantaneous convectivemixing routine was replaced by a time-dependent (di ff usive) mixing routine (similar to that described by Meynet et al. 2004). Thischange was necessary due to the violent evolutionary events (the H-flashes) that occur in models of Z = ff e 2006 for a comparison ofdi ff erent methods).Opacities have been updated to those from Iglesias & Rogers (1996) (for mid-range temperatures) and Ferguson et al. (2005)(for low temperatures). Convective boundaries were always defined by the Schwarzschild criterion – ie. the search for a neutralconvective boundary (see Frost & Lattanzio 1996) was not performed and no overshoot was applied.A key problem with modelling EMP stars is the unknown driver(s) of mass loss. We have used the empirical mass-loss formulaof Reimers (1975) during the RGB. We believe this to be acceptable because the mass lost prior to the AGB is very small due to theshort-lived giant branches at these metallicities (see eg. D’Antona 1982). For the AGB we use the formula of Vassiliadis & Wood(1993). As described below, all the models experience some self-pollution – and always before or at the very beginning of theTPAGB phase. We find that the surfaces of the AGB models usually have metallicities approaching that of the LMC or even Solar(as defined by Z = − X − Y rather than Fe – they are still Fe-poor). Thus, since the stellar surfaces have (some of) the ingredientsneeded to form grains, we argue that using a standard mass loss formula is warranted, at least as a first approximation. We notethat metallicity is also indirectly taken into account by the mass loss formulae, since they depend on bulk stellar properties (such asradius, luminosity, pulsation period), which vary significantly with metallicity.The nucleosynthesis calculations were made with the Monash Stellar Nucleosynthesis code (MONSOON), a post-processingcode which takes input from the MONSTAR code (eg. density, temperature, convective velocities). It solves a network of 506nuclear reactions involving 74 nuclear species (see eg. Cannon 1993; Lattanzio et al. 1996; Lugaro et al. 2004). The yield tablescontain a reduced number of species because we have excluded isotopes with very short lifetimes (and thus have negligible yields). . W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars 3 Fig. 1.
Example of one of the Dual Core Flash events. Convective regions are shown by grey shading whilst the mass location of theedge of the H-exhausted core is shown by the solid line (blue).Initial composition for the Z = ⊙ Z = / H] values (for example 10 M ⊙ of Big Bang material was required for [Fe / H] = − . ff erence between thiscomposition and that of a scaled-solar composition is an underabundance of N, since the supernova calculation did not producemuch of this element. This was found to have little e ff ect however since whenever the CNO cycle operates it quickly converts muchof the C to N anyway ([C / Fe] is ∼ M = . , . , . , . ⊙ and the metallicity range: [Fe / H] = − . , − . , − . , − . =
3. Results and Discussion
As mentioned above it has long been known that theoretical models of Z = / H] . − .
5, seeFujimoto et al. 2000). In this event the normal flash-driven convection zone breaks out of the He-rich core. Thus H-rich material ismixed down to regions of high temperature, producing a secondary flash: a H-flash. This flash reaches luminosities comparable tothe core He flash itself and exists concurrently with the He flash. Thus we refer to the combination of these events as a ‘Dual CoreFlash’ (DCF). These events have been given rather cumbersome names in the literature to date: eg. Helium Flash Induced Mixing(HEFM, Schlattl et al. 2002) and Helium Flash-Driven Deep Mixing (He-FDDM, Suda et al. 2004). We propose the simpler nameDCF to illustrate the essential nature of the phenomenon.A similar event occurs in stellar models of higher mass and higher (although still very low) metallicities. In these cases it isthe normal flash-driven convection zone present during the first pulse (or first few pulses) of the TPAGB phase that breach theH-He discontinuities. Again a H-flash results, concomitant with the He shell flash, so we refer to this event as a ‘Dual
Shell
Flash’(DSF). Cassisi et al. (1996) appear to be the first to have reported the occurrence of a DSF, although they were unable to followthe evolution of the event. DSFs have since been reported and modelled by a number of groups (eg. Chie ffi et al. 2001; Siess et al.2002; Iwamoto et al. 2004).Both the DCF and DSF events are driven by the same phenomenon: ingestion of protons into a hot region caused by expansionsof He convective zones into H-rich regions. Thus our proposed nomenclature unifies them as being “Dual Flashes” and then dis- S. W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars
Fig. 2.
Time evolution of the surface composition in the [Fe / H] = − .
45, 3 M ⊙ model, for selected species. This model is a memberof our self-pollution Group 3 (see text and Fig. 3 for details). The rich nucleosynthesis arising from TDU and HBB is seen in theright-hand panel. This chemical signature by far dominates that of the DSF in this case (see bottom left panel).tinguishes them again by referring to the driving event: be it the core or shell flash. Furthermore, the term only applies to H-flashevents occurring at low metallicities.Both the DCF and DSF have consequences for the surface composition of the star since, in both cases, the convective envelopesubsequently deepens and mixes up the (processed) material overlying the H-burning shell. In Fig. 1 we display an example DCFevent (our 1 M ⊙ , [Fe / H] = − . . W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars 5 Table 1.
Part of the yield table for the Z = Nuclide A Initial 0.85 M ⊙ ⊙ ⊙ ⊙ H 1 7.548E-01 7.014E-01 6.597E-01 6.596E-01 5.807E-01 He 4 2.450E-01 2.976E-01 3.295E-01 3.367E-01 4.066E-01 Li 7 3.130E-10 4.704E-10 1.263E-09 4.491E-10 4.887E-11 C 12 0.000E +
00 2.598E-05 1.844E-03 1.309E-04 4.882E-04 C 13 0.000E +
00 7.778E-06 3.619E-04 3.034E-05 1.152E-04 N 14 0.000E +
00 2.437E-04 3.919E-03 3.432E-03 1.166E-02 O 16 0.000E +
00 5.034E-04 4.333E-03 4.885E-05 1.516E-04 F 19 0.000E +
00 1.848E-09 6.225E-06 2.879E-10 1.188E-09 Ne 20 0.000E +
00 2.485E-07 1.726E-06 2.737E-05 1.386E-04 Na 23 0.000E +
00 1.291E-09 1.131E-05 1.294E-05 9.539E-05 Mg 24 0.000E +
00 3.838E-11 1.362E-06 1.865E-07 2.630E-07 Mg 25 0.000E +
00 1.459E-08 3.166E-07 1.756E-06 1.562E-05 Mg 26 0.000E +
00 2.475E-08 4.065E-08 8.159E-06 6.889E-05 Al 26 0.000E +
00 3.182E-11 3.364E-10 3.085E-07 1.487E-06 Si 28 0.000E +
00 1.002E-07 1.649E-11 6.170E-07 1.315E-06 P 31 0.000E +
00 2.200E-08 2.071E-12 5.219E-07 1.117E-06
32 32 0.000E +
00 4.381E-09 1.224E-12 1.870E-07 2.665E-07 ∼ − M ⊙ ) than in the EMP / Z = ∼ − M ⊙ or higher).We discuss the chemical composition of the surface pollution resulting from the Dual Flash events when we compare the C andN yields with observations below ( § In Table 1 we present a sample of the yields for the models. A total of five tables with the same format as these areavailable in their entirety online at the CDS (Tables 1 to 5; via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or viahttp: // cdsweb.u-strasbg.fr / cgi-bin / qcat?J / A + A / ). Each table contains yields for models with initial stellar masses of 0.85, 1.0, 2.0and 3.0 M ⊙ at each of the metallicities: [Fe / H] = − . , − . , − . , − .
0, and Z =
0. We list the yields for all the stable species usedin the nucleosynthesis calculations, which range from H to S plus a small iron group (Ni, Co, Fe isotopes). We also provide yieldsfor the important radionuclides Al (which decays to Mg) and Fe (which decays to Ni). Yields are given in mass fraction ofeach species in the total ejecta. In the tables we also give the initial compositions for each metallicity. The final masses (remnantcore masses) are given in Table 6 (also at CDS). Using this information it is easy to convert the yields to any format. The yields arecalculated by integrating the mass of each species lost by the star over its lifetime ( τ ∗ ): M toti = Z τ ∗ X i ( t ) dMdt dt (1)where X i is the mass fraction of species i . The total mass of each species lost to the ISM, M toti , is then scaled with M e j , the totalmass lost by the star, to give the mass fractions.In some cases our models failed to converge towards the end of the AGB. This is a common problem with stellar codes. Oftenthere was very little mass left in the envelope so this was just added to the yields. However in some cases there was enough massleft that it would not have been lost in one interpulse period. In these cases we performed a short synthetic evolution calculationfor the remaining thermal pulses (including third dredge-up and core growth) to complete the evolution, following the method ofKarakas (2003). Yields were then calculated taking into account this extra mass loss. In most models the number of thermal pulsescalculated in this way was .
8. This represents between ∼ M = ⊙ and [Fe / H] = − .
45 model. Here it isTDU and Hot Bottom Burning that are the main contributors to the yield of the star. Indeed, the chemical signature arising from theDSF occurring at the start of the AGB is totally erased by these normal AGB evolutionary episodes. In particular the CN cyclingproduct N dominates the surface composition during most of the AGB (in terms of metallic species). The C / C and C / N ratiosquickly approach equilibrium values once the (strong) HBB starts.We summarise the self-pollution episodes over the whole grid of models by dividing them into three categories, defined by theevolutionary events / phases that dominate the chemical signature in the yields: – Group 1 yields are dominated by the DCF events – Group 2 are dominated by DSF events – Group 3 are dominated by TDU + HBBMembers of the DCF group have polluted surfaces during the horizontal branch phase onwards whilst the members of the DSFgroup have polluted surfaces after the first few pulses of the AGB (see Fig. 3). In the low mass models ( M ≤ . ⊙ ) this pollution S. W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars occurs despite the lack of TDU. Thus our models predict a greater proportion of C-rich stars at extremely low metallicity, sincethese Dual Flash events do not occur at higher metallicities.In Fig. 4 we show the self-pollution groups in a mass-metallicity diagram. Here it is clear that the yields of all the intermediatemass models ( M ≥ . ⊙ ) that su ff er surface pollution resulting from Dual Shell Flashes are actually dominated by the pollutionoccurring during the AGB phase. In other words the AGB pollution erases the DSF pollution. It can also be seen that the DCF eventsare limited to ultra metal poor low-mass models, whilst the DSF events occur at higher metallicities (for low mass models). We notethat Fujimoto et al. (2000) also provide a mass-metallicity diagram for their low-metallicity study (their Fig. 2). Our diagram isqualitatively similar to theirs. One point of di ff erence is that our boundary for the AGB-DSF models (filled circles with open circlesaround them in Fig. 4) is at a lower metallicity. This may be due to the fact that we adopt a ‘hard’ Schwarzschild convectiveboundary in our models, although we are unsure if Fujimoto et al. 2000 used any overshoot or not. A further small di ff erence is thatour diagram shows a mass dependence in addition to the metallicity dependency for the pollution event boundaries (diagonal linesin Fig. 4). In Fig. 5 we compare the carbon yields from our entire grid of models with the observed [C / Fe] abundances in EMP halo stars. It canbe seen that the yields are universally C-rich. Thus there is a qualitative agreement between the models and the observations in termsof C, as found by previous studies. Another interesting feature of this diagram is that the model yields predict [C / Fe] to continueincreasing towards lower and lower metallicities (stars of such low metallicity may show up in future surveys). Furthermore, takinginto account the evolutionary stage at which the surface pollution is gained in the lower mass models ( M = .
85 and 1.0 M ⊙ ) – ie.the DCF events rather than the AGB – the models also predict a higher proportion of C-rich stars at lower and lower metallicities.This is due to the fact that these stars already have self-polluted surfaces during the HB stage – which has a lifetime roughly 1 orderof magnitude longer than the AGB phase.Previous studies have also attempted to quantitatively reproduce the abundances observed in some CEMP stars, in particularC and N. However they have universally found that the 1D models produce by far too much of these elements. The early studyby Hollowell et al. (1990) reported their resultant post-DCF surface abundance of nitrogen to be ∼ times that observed in theEMP star CD −
38 245. Schlattl et al. (2002) find about the same two orders of magnitude overproduction of both N and C, whilstIwamoto et al. (2004) find their [Fe / H] = − . ∼ / H] = − .
5, the current lower limit of the observations, our models produce ∼ ⊙ ), which have yields dominated by AGB products, that are the closest to the observations – their yieldsare only ∼ . / N] ratios well below the CEMP observations (Fig. 6).At [Fe / H] = − . ⊙ model still produces ∼ / H] ∼ − . / H] = − . Shell
Flashes atthe start of the AGB. Since they do not show TDU the abundance patterns in their yields are primarily from the DSFs. The highermass models do not undergo any Dual Flashes at all, they experience normal AGB TDU and strong HBB (even at 2 M ⊙ ). Howeverthe low [C / N] ratios in the yields from these IM models are again too low compared to the observations. Thus, based on the twoconstraints of [C / Fe] and [C / N], it is our low mass models which undergo DSFs that provide the best fit to the observations. We notethat our results are in contrast with the models of Iwamoto et al. (2004), who found their [Fe / H] = − . ∼ / N] ratios from EMP Galactic Halo star observations along with our yield abundances in Fig. 6. An in-teresting feature of the observations is that the bulk of the CEMPs have [C / N] values ranging from Solar to ∼ / N ∼ ff er HBB are heavily enriched in N – so much so that N rather than C dominates their composition –in contrast to the DF-polluted yields. Thus the yields from intermediate mass AGB stars can not reproduce the bulk of the CEMPobservations (we note that there are a few CEMPs located below the C / N = / N] values in the yields dominated by the DF events (ie. the low-mass models)are reasonably consistent with the spread of the CEMP observations. Again at the lowest metallicities ([Fe / H] = − . , − .
0) we aredealing with very small samples (2 or 3 stars). The main discrepancy is HE 0107 − / N] ratio significantly abovethat of our models of comparable metallicity. We note that Venn & Lambert (2008) have recently suggested that some of the most . W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars 7 metal-deficient Halo stars, such as this one, may actually be low-metallicity analogues of the ‘chemically peculiar’ stars (whichshow dust depletion of particular elements, such as Fe) and thus may be intrinsically more metal-rich than currently thought.The yields from our higher metallicity low-mass models ([Fe / H] = − .
0) are at the high end of the observations ([C / N] ∼ / Fe in a binary mass-transfer scenario. Anotherpossibility is that the Dual Shell Flash events – which we stress are the only events that come close to simultaneously reproducingthe [C / Fe] and [C / N] observations – may in reality produce a distribution of C (and N) pollution that is not reflected in the models.It is interesting to note that other objects also show surfaces simultaneously enriched in C and N. Recent observations of potential‘Late Hot Flasher’ stars (see § / Fe] and [C / N] abundances in the yields compare reasonably with the observations, the lifetimes ofthese stars also need to be taken into account. We find models with initial masses of 0.85 M ⊙ (ie. the low-mass edge of our grid) havelifetimes of ∼
10 Gyr. This is comparable to the age of the Galactic Halo Globular clusters so these stars may still be ‘living’ today.Conversely the 1, 2 and 3 M ⊙ models, which have main sequence lifetimes of ∼ . ∼ . ∼ . ff e et al. 2007; Lugaro et al. 2008).Finally we note that here we have only discussed the chemical signatures of C and N. We shall provide further comparisons andanalysis of other elements and abundance patterns in future papers in this series. The results presented here naturally contain many uncertainties, especially since they are the first attempt at detailed yields in thismass and metallicity range, and particularly because EMP stellar evolution is so challenging.Some sources of uncertainties include: the unknown mass-loss rates, uncertain nuclear reaction rates and the treatment of con-vection and mixing. Testing all of these uncertainties is outside the scope of this study, and indeed would require an extremely largeamount of work (it is somewhat more tractable using synthetic stellar models, see eg. Marigo 2001).Determining some bounds to AGB model results has however been attempted by various studies. For instance our groupexplored the e ff ects of varying mass-loss rates and (some) nuclear reaction rates in low and intermediate mass stars of lowmetallicity (Fenner et al. 2004). A more detailed study into the mass-loss dependence of AGB yields was recently reported byStancli ff e & Je ff ery (2007), and Herwig et al. (2006) investigated the e ff ect of reaction rate uncertainties on TDU e ffi ciency and theresultant AGB yields. A further source of uncertainty arises because detailed AGB models are currently lacking opacity tables thattake into account the C and N enrichment that occurs as a result of 3DU and HBB. In particular low temperature opacity tablesvariable in H, He, C and N (at least) are needed. By approximating the opacity of a few key molecules in synthetic stellar models,Marigo (2002) showed how important including this opacity source is. She found, amongst other things, that it a ff ects the e ff ectivetemperature and lifetime of C-rich models. We note that Lederer & Aringer (2008) have just completed detailed calculations toproduce tables for use in AGB stellar models. Some preliminary results have recently been used in a study by Cristallo et al. (2007).We are currently updating our structure code to include these opacity tables.In the case of the Z = ff usion. They did howeverconclude that DFs are a robust prediction of 1D stellar models. They also noted that the degree and type of pollution arising fromtheir models is very similar to previous studies. It is reassuring that all 1D DCF simulations appear to give very similar results,despite the various numerical methods employed between codes. However, seeing as most of the models overestimate the amountof C and N production by orders of magnitude (see § ∼
4. Summary
We have modified an existing stellar code and used this to calculate the evolution and nucleosynthesis of a grid of low- andintermediate-mass EMP and Z = / H] . − . ∼ S. W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars mass transfer or better numerical modelling may resolve this discrepancy. We also note that the very small observational data setsat these extreme metallicities make comparisons uncertain.At higher metallicities ([Fe / H] ∼ − . / H] = − . / Fe]and [C / N]. The models that have their yields dominated by TDU and HBB produce comparable amounts of C but have too muchN (generated from the HBB), which pushes [C / N] to much lower values than seen in the observations. Moreover, if the lower massstars were to experience TDU (and no DFs), the pollution expected would be (mainly) C-rich, whereas the CEMPs are enrichedin N as well as C. Thus TDU – with or without HBB – does not seem to be a viable solution for the CEMP abundance patterns.We stress that the Dual Shell Flash events only occur at low metallicities. We note that our result is in contrast with the findings ofIwamoto et al. (2004) who found their [Fe / H] = − . ∼ / Fe] to increase at lower and lower metallicities. Furthermore, the proportion of CEMPstars should also continue to increase at lower metallicities, based on the results that some of the low mass EMP models alreadyhave polluted surfaces by the HB phase (which has a long lifetime compared to the AGB), and that there are more C-producingevolutionary episodes at these metallicities.Finally we note that all these calculations contain many uncertainties. These include the unknown mass-loss rates, uncertainnuclear reaction rates, the treatment of convection and opacities. In the case of the Dual Flash events we believe this warrantsmultidimensional fluid dynamics calculations, which a number of groups have started working on.The models and yields from this study will be described in more detail in future papers in this series.
Acknowledgements.
This study utilised the Australian Partnership for Advanced Computing (APAC) supercomputer, under Project Code g61 . SWC thanks theoriginal authors and maintainers of the stellar codes that have been used in this work – Peter Wood, Rob Cannon, John Lattanzio, Maria Lugaro, Amanda Karakas,Cheryl Frost, Don Faulkner and Bob Gingold. SWC was supported by a Monash University Research Graduate School PhD scholarship for 3.5 years.
References
Aoki, W., Beers, T. C., Christlieb, N., et al. 2007, ApJ, 655, 492Aoki, W., Bisterzo, S., Gallino, R., et al. 2006a, ApJ, 650, L127Aoki, W., Frebel, A., Christlieb, N., et al. 2006b, ApJ, 639, 897Asplund, M., Gustafsson, B., Lambert, D. L., & Kameswara Rao, N. 1997, A&A, 321, L17Asplund, M., Lambert, D. L., Kipper, T., Pollacco, D., & Shetrone, M. D. 1999, A&A, 343, 507Beers, T. & Christlieb, N. 2005, ARA&A, 43, 531Beers, T. C., Sivarani, T., Marsteller, B., et al. 2007, AJ, 133, 1193Brown, T. M., Sweigart, A. V., Lanz, T., Landsman, W. B., & Hubeny, I. 2001, ApJ, 562, 368Cannon, R. C. 1993, MNRAS, 263, 817Cassisi, S. & Castellani, V. 1993, ApJS, 88, 509Cassisi, S., Castellani, V., & Tornambe, A. 1996, ApJ, 459, 298Cassisi, S., Schlattl, H., Salaris, M., & Weiss, A. 2003, ApJ, 582, L43Chie ffi , A., Dom´ınguez, I., Limongi, M., & Straniero, O. 2001, ApJ, 554, 1159Christlieb, N., Gustafsson, B., Korn, A. J., et al. 2004, ApJ, 603, 708Coc, A., Vangioni-Flam, E., Descouvemont, P., Adahchour, A., & Angulo, C. 2004, ApJ, 600, 544Cohen, J. G., McWilliam, A., Shectman, S., et al. 2006, AJ, 132, 137Cohen, J. G., Shectman, S., Thompson, I., et al. 2005, ApJL, 633, L109Cristallo, S., Straniero, O., Lederer, M. T., & Aringer, B. 2007, ApJ, 667, 489D’Antona, F. 1982, A&A, 115, L1Ezer, D. 1961, ApJ, 133, 159Fenner, Y., Campbell, S., Karakas, A. I., Lattanzio, J. C., & Gibson, B. K. 2004, MNRAS, 353, 789Ferguson, J. W., Alexander, D. R., Allard, F., et al. 2005, ApJ, 623, 585Frebel, A., Aoki, W., Christlieb, N., et al. 2005, Nature, 434, 871Frebel, A., Christlieb, N., Norris, J. E., et al. 2006, ApJ, 652, 1585Frost, C. A. & Lattanzio, J. C. 1996, ApJ, 473, 383Fujimoto, M. Y., Iben, I. J., & Hollowell, D. 1990, ApJ, 349, 580Fujimoto, M. Y., Ikeda, Y., & Iben, I. J. 2000, ApJ, 529, L25Goriely, S. & Siess, L. 2001, A&A, 378, L25Goswami, A., Aoki, W., Beers, T. C., et al. 2006, MNRAS, 372, 343Herwig, F., Austin, S. M., & Lattanzio, J. C. 2006, Phys. Rev. C, 73, 025802Herwig, F., Bl¨ocker, T., Langer, N., & Driebe, T. 1999, A&A, 349, L5Hollowell, D., Iben, I. J., & Fujimoto, M. Y. 1990, ApJ, 351, 245Iben, Jr., I., Kaler, J. B., Truran, J. W., & Renzini, A. 1983, ApJ, 264, 605Iglesias, C. A. & Rogers, F. J. 1996, ApJ, 464, 943Iwamoto, N., Kajino, T., Mathews, G. J., Fujimoto, M. Y., & Aoki, W. 2004, ApJ, 602, 377Johnson, J. A., Herwig, F., Beers, T. C., & Christlieb, N. 2007, ApJ, 658, 1203Karakas, A. I. 2003, PhD thesis, Monash University, AustraliaLanz, T., Brown, T. M., Sweigart, A. V., Hubeny, I., & Landsman, W. B. 2004, ApJ, 602, 342Lattanzio, J., Frost, C., Cannon, R., & Wood, P. R. 1996, MmSAI, 67, 729Lederer, M. T. & Aringer, B. 2008, A&A, submittedLimongi, M., Chie ffi , A., & Bonifacio, P. 2003, ApJL, 594, L123Lucatello, S., Beers, T. C., Christlieb, N., et al. 2006, ApJL, 652, L37Lucatello, S., Tsangarides, S., Beers, T. C., et al. 2005, ApJ, 625, 825Lugaro, M., de Mink, S. E., Izzard, R. G., et al. 2008, A&A, 484, L27Lugaro, M., Ugalde, C., Karakas, A. I., et al. 2004, ApJ, 615, 934Marigo, P. 2001, A&A, 370, 194Marigo, P. 2002, A&A, 387, 507 . W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars 9 Marigo, P., Girardi, L., Chiosi, C., & Wood, P. R. 2001, A&A, 371, 152Meynet, G., Maeder, A., & Mowlavi, N. 2004, A&A, 416, 1023Picardi, I., Chie ffi , A., Limongi, M., et al. 2004, ApJ, 609, 1035Reimers, D. 1975, Memoires of the Societe Royale des Sciences de Liege, 8, 369Renzini, A. 1990, 11, 549Schlattl, H., Cassisi, S., Salaris, M., & Weiss, A. 2001, ApJ, 559, 1082Schlattl, H., Salaris, M., Cassisi, S., & Weiss, A. 2002, A&A, 395, 77Shigeyama, T. & Tsujimoto, T. 1998, ApJL, 507, L135Siess, L., Livio, M., & Lattanzio, J. 2002, ApJ, 570, 329Sivarani, T., Beers, T. C., Bonifacio, P., et al. 2006, A&A, 459, 125Spite, M., Cayrel, R., Hill, V., et al. 2006, A&A, 455, 291Stancli ff e, R. J. 2006, MNRAS, 370, 1817Stancli ff e, R. J., Glebbeek, E., Izzard, R. G., & Pols, O. R. 2007, ArXiv Astrophysics e-printsStancli ff e, R. J. & Je ff ery, C. S. 2007, MNRAS, 375, 1280Suda, T., Aikawa, M., Machida, M. N., Fujimoto, M. Y., & Iben, I. J. 2004, ApJ, 611, 476Suda, T., Katsuta, Y., Yamada, S., et al. 2008, PASJ, 60, in pressSweigart, A. V. 1997, in The Third Conference on Faint Blue Stars, ed. A. G. D. Philip, J. Liebert, R. Sa ff er, & D. S. Hayes, 3Vassiliadis, E. & Wood, P. R. 1993, ApJ, 413, 641Venn, K. A. & Lambert, D. L. 2008, ApJ, 677, 572Weiss, A., Schlattl, H., Salaris, M., & Cassisi, S. 2004, A&A, 422, 217Wood, P. R. & Zarro, D. M. 1981, ApJ, 247, 247Woodward, P., Herwig, F., Porter, D., et al. 2008, 990, 300 Fig. 3.
Displayed in each HR diagram is a representative example from our grid of models for each ‘Self-Pollution Group’ (see textfor details on the groups). Black lines indicate phases of the evolution in which the surface is strongly polluted with CNO nuclides(from the DCF, DSF or 3DU events). Grey (light blue) lines indicate that the surface still retains the initial metal-poor composition.Evolutionary stages and self-pollution sources are marked, as are the mass and metallicity ranges of each group. Question marksindicate unknown upper boundaries (due to the limited mass range of the current study). . W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars 11
Fig. 4.
Mass-metallicity diagram summarising the dominant sources of pollution in the yields. Each symbol represents a singlestellar model. Crosses (red) represent the DCF self-pollution Group 1, filled triangles (blue) the DSF Group 2 and filled circles(green) the AGB pollution Group 3. The open circles (blue) around the filled circles (green) indicate intermediate mass models thatexperienced DSFs. Pollution from TDU (and HBB) easily dominates the pollution from the DSF events at IM mass so the yields ofthese models fall into the AGB group. The Z = / H] = − Fig. 5.
Carbon yields from our models and observations of EMP stars. Stars with [C / Fe] > + . / Fe] < + . / Fe] ∼
0, except for the Z = = / H] = −
8. The two most metal-poorstars can be seen at [Fe / H] ∼ − . / H] = + .
0. The yields from our models are colour- and shape-coded to highlight thedi ff erent episodes that produced the bulk of the pollution in each yield (see text for details). Numbers beside each yield markerindicate initial stellar mass, in M ⊙ . The 3 M ⊙ model at this [Fe / H] = − . . W. Campbell and J. C. Lattanzio: Yields from Low-Mass EMP Stars 13 Fig. 6.
Comparing the [C / N] ratios in the yields of our models with those of the observations. See Fig. 5 for the observational datasources and the definitions of CEMPs and ‘normal’ EMPs. We have marked in the CN equilibrium line (for HBB) at [C / N] = − / N = . / N] ∼ − . / N ∼
1, so that C dominates N above this line. For reference [C / N] = + / N ∼
50. Note that we have notplotted the Z ==