GALACTIC CHEMICAL EVOLUTION: THE IMPACT OF THE 13C-POCKET STRUCTURE ON THE S-PROCESS DISTRIBUTION 7 S. Bisterzo1 and C. Travaglio2 1 0 2 INAF - Astrophysical Observatory Turin, Turin, Italy n a J [email protected]; [email protected] 4 ] M. Wiescher R S Joint Institute for Nuclear Astrophysics (JINA), Department of Physics, University of . h p Notre Dame, IN, USA - o r t F. Ka¨ppeler s a [ Karlsruhe Institute of Technology, Campus Nord, Institut fu¨r Kernphysik, Karlsruhe, 1 v Germany 6 5 0 and 1 0 . 1 R. Gallino2 0 7 Department of Physics, University of Turin, Italy 1 : v i X Received ; accepted r a 1Department of Physics, University of Turin, Italy 2B2FH Association-c/o Strada Osservatorio 20, 10023 Turin, Italy – 2 – ABSTRACT The solar s-process abundances have been analyzed in the framework of a GalacticChemical Evolution (GCE) model. Theaimofthiswork istoimplement thestudybyBisterzo et al.(2014), whoinvestigated theeffect ofoneofthemajor uncertainties of asymptotic giant branch (AGB) yields, the internal structure of the13Cpocket. WepresentGCEpredictionsofs-processelementscomputedwith additional tests in the light of the suggestions provided in recent publications. The analysis is extended to different metallicities, by comparing GCE results and updated spectroscopic observations of unevolved field stars. We verify that the GCE predictions obtained with different tests may represent, on average, the evolution of selected neutron-capture elements in the Galaxy. The impact of an additional weak s-process contribution from fast-rotating massive stars is also explored. Subject headings: Stars: AGB - Galaxy: evolution, abundances – 3 – 1. GCE Solar s-process Predictions AGB stars with low initial mass are the major responsible for the nucleosynthesis of solar s isotopes with A > 90 (Busso, Gallino & Wasserburg 1999). The main neutron source of low-mass AGB models is the 13C(α, n)16O reaction, which burns radiatively during the interpulse in a thin layer of the He intershell, the so-called 13C pocket (Straniero et al. 1995). The formation of the 13C pocket requires an unknown mixing mechanism that allows partial mixing of a few protons from the convective envelope into the top layers of the radiative He- and C-rich intershell. This is assumed to occur at the quenching of a Third Dredge Up (TDU) episode. When the star contracts, the H shell reignites and protons in the intershell are captured by the abundant 12C nuclei to yield primary 13C. If more protons than 12C nuclei are diffused in the outer layers, a region of primary 14N may form by further proton captures on 13C. Subsequently, the temperature in the 12C pocket increases to ∼1×108 K, and neutrons are released radiatively within the pocket via 13C(α, n) reactions at quite low neutron densities. Various physical mechanisms have been explored for the formation of 13C pocket (e.g., overshooting, rotation, magnetic fields, gravity waves; Herwig et al. 1997; Langer et al. 1999; Denissenkov & Tout 2003; Siess, Goriely, & Langer 2004; Straniero, Gallino & Cristallo 2006; Piersanti, Cristallo, & Straniero 2013; Nucci & Busso 2014). The details of how the 13C pocket forms are still debated, making its mass extent and the H profile largely uncertain. Spectroscopic observations provide key information to constrain theoretical models: chemically peculiar s-rich stars have evidenced a dispersion of the s abundances for a given metallicity (e.g., MS, S, C(N), Ba, CEMP-s and post-AGB stars, planetary nebulae; see the recent review by Ka¨ppeler et al. 2011 and Karakas & Lattanzio 2014). This dispersion has been recognized since first studies by Busso et al. (2001) and Abia et al. (2002), but – 4 – the reason(s) are not definitely identified. A variation in the stellar rotational velocity may be regarded as a possible explanation (see Piersanti, Cristallo, & Straniero 2013, and references therein). Owing to the present uncertainties, the 13C pocket is artificially introduced in our post-process AGB models, following the observational constraints. The shape of the 13C and 14N profiles and the mass involved in the pocket are regulated by a free parametrization. The internal structure of the 13C pocket adopted so far has been calibrated to represent the solar main component (Arlandini et al. 1999): it is a three-zone pocket (each zone has defined X(13C) and X(14N) abundances) with a total mass of about 0.001 M (see ⊙ Table 1; first group of data). A range of 13C-pocket strengths is assumed to reproduce the spectroscopic s-process dispersion: we parametrically vary the concentration of 13C (and 14N) of each zone given in Table 1 by different factors, leaving the mass of the pocket constant. We refer to Bisterzo et al. (2010, 2014) for a detailed discussion. This systematic approach appears justified by the present uncertainties: the formation of diverse 13C pockets may result from the interplay between different physical processes in stellar interiors. Post-process models should be considered useful tests to address full stellar evolutionary models (and in general multidimensional/hydrodynamical simulations) against observational constraints. The solar s-process abundances must account for the complex chemical evolution of the Galaxy, which includes AGB yields of various masses and metallicities. The chemical evolution model adopted to reproduce the solar s distribution has been exhaustively described by Travaglio et al. (1999, 2004). In the framework of GCE, we showed that the solar s distribution of isotopes with 130 < A ≤ 208 can be accurately reproduced once we consider a proper weighted average among – 5 – the 13C-pocket strengths1. This is consistent with the observed spectroscopic s-process dispersion. A deficit (of about 25%) between GCE predictions of s-process elements and the solar abundances was found for isotopes with 90 < A < 130 (solar LEPP-s). Bisterzo et al. (2014) have investigated a possible connection between this deficit and the 13C-pocket structure. On the basis of their tests, solar GCE predictions of s-process elements are marginally affected. The aim of this work is to implement the analysis carried out on the 13C-pocket structure by Bisterzo et al. (2014), (Section 2). In Section 3, we consider the sensitivity of AGB yields to metallicity, focusing on the contribution by metal-rich AGB stars to the light elements (see discussion by Maiorca et al. 2012 for open clusters). Although the solar composition is fundamental to constrain AGB yields, it only provides a single piece of information about the Galactic history. The reliability of the 13C-pocket tests needs to be verified by considering the complex framework of Galactic chemical enrichment. The GCE predictions of selected neutron-capture elements versus metallicity are compared with updated spectroscopic observations in Section 4. Recently, rotation-induced mixing in low-metallicity massive stars has been proposed as an explanation of the observed [Sr/Ba] dispersion in extremely metal-poor stars, being efficient primary producers for s isotopes heavier than Sr, up to the Ba neutron-magic peak (Frischknecht, Hirschi, & Thielemann 2012; Pignatari et al. 2013; Cescutti et al. 2013). In this context, we investigate the impact of recent weak s-process yields by Frischknecht et al. (2016), available from a large grid 1Tothispurpose, theunbranched s-onlyisotope150Smistaken asreference nuclide forthe whole s-process distribution. The high production of 208Pb in low-mass low-metallicity AGB stars plays another indicative GCE constraint (Travaglio et al. 2001). The solar abundance of 208Pb is matched once the s-process occurring in low-metallicity AGB stars is properly considered in the context of the chemical evolution of the Galaxy. – 6 – of rotating massive stars (Z from 10−5 to solar), on the Galactic chemical enrichment (Section 4.1). In Section 5 our results are briefly summarized and future outlook are discussed. 2. Impact of New 13C-pocket Tests on Solar s Abundances We focus on specific additional tests carried out on the basis of recent advice available in literature. The internal 13C-pocket structure adopted in each test is given in Table 12. • Starting from the three-zone 13C-profile adopted so far, we investigate the impact of a substantially extended 13C-pocket mass than assumed in our previous computations (up to four times larger, corresponding to a total mass of M (pocket)∼4×10−3 M ; tot ⊙ see tests described as CASE 1 and CASE 2 in Table 1). In this regard, Maiorca et al. (2012) have proposed such a 13C-pocket mass to represent the abundances of neutron-capture elements in young open clusters. Magnetic buoyancy (or other forced mechanisms) are suggested as good candidates to form such a 13C reservoir (Trippella et al. 2014). Rotation models by Piersanti, Cristallo, & Straniero (2013) indicate that low-metallicity AGB stars and fast-rotating metal-rich stars might produce such an extended 13C pocket. Comparison between theoretical models and the strontium and barium isotopic signatures measured in mainstream SiC grains require 13C pockets with M (pocket)≥1×10−3 tot M (Liu et al. 2015). ⊙ 2We remind that all tests are performed on low-mass AGB models because the effect of the 13C pocket in AGB stars with intermediate mass (4 ≤ M/M < 8) is negligible for ⊙ GCE solar predictions (see Section 2 by Bisterzo et al. 2014; Straniero, Cristallo & Piersanti 2014). – 7 – • In CASE 3 (see Table 1), we test the effect of an additional parametrized 14N-pocket. Asanticipated inSection1, anouter14N-richlayer mayforminthepocket once enough protons are mixed in the external zone of the He intershell (Goriely & Mowlavi 2000; Cristallo et al. 2009, 2011; Karakas et al. 2010; Lugaro et al. 2012; Trippella et al. 2014). The presence of rotation may widen the 14N pocket by inducing partial mixing of the 14N-rich region with the inner 13C-rich zone. • In CASE 4, we have performed a set of 1.5, 2 and 3 M models computed with a ⊙ more efficient Reimers mass loss than in previous models. This implies that AGB models experience about half of the TDU episodes compared to CASE 3. The 13C pocket structure is the same as adopted in CASE 3 (see Table 1). This test simulates the recent prescriptions of updated FRUITY3 models by Cristallo et al. (2009, 2011). FRUITY models include an efficient AGB mass-loss rate, which has been calibrated using the infrared luminosity function of Galactic carbon stars, and improved radiative C-enhanced opacity tables. Accordingly, these new models with a reduced number of TDU episodes are in better agreement with observations (Guandalini & Cristallo 2013 and references therein). • Finally, in CASE 5 (see Table 1), we test the impact of an additional extended inner zone of the pocket with a mass of 2×10−3 M and a correspondingly lower 13C ⊙ abundance (X(13C)=2.75E−3). Recently, Cristallo et al. (2015) found that a different convective/radiative boundary condition allows a deeper penetration of protons with a very low mixing efficiency during TDU episodes. The resulting 13C pocket displays an extended tail with a smooth decrease of the 13C profile (see Tail model in their Fig. 7). CASE 5 roughly approximates the Tail model by Cristallo et al. (2015). 3web: fruity.oa-teramo.inaf.it/. – 8 – The resulting solar GCE predictions of s-process elements are displayed in Fig. 1. A proper weighted average among the various 13C-pocket strengths must be adopted for each test to reconcile GCE predictions with 100% of solar 150Sm (see Bisterzo et al. 2014). This approach allows us to reproduce the solar 150Sm within 5% uncertainty (Lodders, Palme & Gail 2009). The s-only isotopes with atomic mass A & 90 show variations smaller than ∼10%, thus confirming the need of a solar LEPP-s mechanism in order to increase the solar s abundances in the range 90 < A < 130, as predicted by Travaglio et al. (2004). Note that AGB yields computed with a single 13C-pocket choice do not provide accurate interpretations neither of solar s abundances nor of observations of peculiar s-rich stars. By working in a larger range of uncertainties, potential missing contributions are not necessarily highlighted. Accordingly, the result of this paper does not disagree with a recent study by Cristallo et al. (2015) that provides a meticulous discussion about the uncertainties affecting stellar models and the solar GCE distribution. Within the estimated uncertainties a LEPP mechanism is not necessarily required. However, the authors suggest how their representation of the solar distribution could be improved once models with different initial rotational velocities (or different prescriptions for convective overshoot during the TDU) will be included in GCE computations for an extended metallicity grid. These upcoming progresses in stellar models will assess whether additional contributions are needed. 3. The Stellar Yields versus Metallicity The complex dependence of s-process yields on the initial stellar metallicity is well known. In Fig. 2 (top panel), we display the AGB production factors of selected s-only isotopes. Starting from [Fe/H] = −0.5, which corresponds to t = 3.5 Gyr, the s Gal – 9 – Table 1. Internal structure of the 13C-pockets adopted in the tests displayed in Fig. 1. Zone0 ZoneI ZoneII ZoneIII ZoneIV ZoneV Standard choice: three-zonemodel;Mtot(pocket) =1.09E−3M⊙ Mass(M⊙) − 5.50E−4 5.30E−4 7.50E−6 − − X(13C) − 3.20E−3 6.80E−3 1.60E−2 − − X(14N) − 1.07E−4 2.08E−4 2.08E−3 − − CASE 1a: three-zonemodel;Mtot(pocket) =3×1.09E−3M⊙ Mass(M⊙) − 1.65E−3 1.59E−3 2.25E−5 − − X(13C) − 3.20E−3 6.80E−3 1.60E−2 − − X(14N) − 1.07E−4 2.08E−4 2.08E−3 − − CASE 2b: 1.3≤M ≤1.5M⊙ models;Mtot(pocket) =4×1.09E−3M⊙ Mass(M⊙) − 2.20E−3 2.12E−3 3.00E−5 − − X(13C) − 3.20E−3 6.80E−3 1.60E−2 − − X(14N) − 1.07E−4 2.08E−4 2.08E−3 − − CASE 3 and CASE 4c: five-zonemodelwithexternal 14N-richzones;Mtot(pocket) =3.24E−3M⊙ Mass(M⊙) − 1.20E−3 1.05E−3 3.30E−4 3.30E−4 3.30E−4 X(13C) − 3.20E−3 6.80E−3 1.60E−2 4.00E−2 4.00E−2 X(14N) − 1.07E−4 2.08E−4 2.08E−3 8.00E−2 1.49E−1 CASE 5d: four-zonemodelwithanextended innertailof13C;Mtot(pocket) =3.09E−3M⊙ Mass(M⊙) 2.00E−3 5.50E−4 5.30E−4 7.50E−6 − − X(13C) 2.75E−3 3.20E−3 6.80E−3 1.60E−2 − − X(14N) 5.73E−5 1.07E−4 2.08E−4 2.08E−3 − − a Themassofthepocketisincreasedbyafactorofthreewithrespecttoourstandardchoice(Bisterzoetal.2014). b Themassofthepocket isincreasedbyafactoroffourwithrespecttoourstandardchoicein AGBmodelsof1.3to1.5M⊙. Weleavethe13Cpocket unchanged forM >1.5M⊙ models. c Weincludetwoadditional externalzones withX(14N)muchhigherthanX(13C). d Weincludeanadditional internalzonewithX(13C)=2.75E−3. – 10 – ) un 13C-pocket standard choice: three-zone model; M (pocket) ≈ 0.001 M tot o s t = CASE 1 = three-zone model with M (pocket) ≈ 0.003 M tot o (t CASE 2 = three-zone model with M (pocket) ≈ 0.004 M for 1.3 - 1.5 M tot o o % CASE 3 = five-zone model with external 14N-rich zones; M (pocket) ≈ 0.003 M tot o s CASE 4 = CASE 3, but with a more efficient mass loss es CASE 5 = four-zone model with extended inner 13C tail; M (pocket) ≈ 0.003 M tot o c o100 r p - s n o i t u l o v E l a c i m e h C 10 60 80 100 120 140 160 180 200 220 Atomic mass Fig. 1.— Effect of the 13C-pocket uncertainties in low mass AGB models on GCE solar s- process predictions. The s-only isotopes (and 208Pb) obtained with our standard three-zone 13C-pocket choice are represented by filled circles. Stable nuclei are displayed by crosses. The results obtained by several tests have been displayed with different symbols (see label in the top panel of the figure). We have adopted a range of 13C-pocket strengths (Table 1) in order to reproduce 100% of solar 150Sm when changing the structure of the 13C pocket (see Bisterzo et al. 2014, their Fig. 4). Note that s-only isotopes with A < 90 (70Ge, 76Se, 80,82Kr, 86,87Sr) receive an additional contribution by the weak s-process in massive stars (see e.g., Pignatari et al. 2010).