The impact of the 18F(a,p)21Ne reaction on fluorine production in AGB stars
Amanda I. Karakas, Hye Young Lee, Maria Lugaro, J. Goerres, Michael Wiescher
aa r X i v : . [ a s t r o - ph ] D ec The impact of the F( a , p ) Ne reaction on fluorineproduction in AGB stars
Amanda I. Karakas ∗ a , H. Y. Lee b , d , Maria Lugaro c , J. Görres d and M. Wiescher d a Research School of Astronomy & Astrophysics, Mt Stromlo Observatory, Weston Creek ACT2611, Australiab Physics Division, Argonne National Laboratory, Argonne, IL 60439-4843c Sterrenkundig Instituut, University of Utrecht, Postbus 80000, 3508 TA Utrecht, TheNetherlandsd Department of Physics and Joint Institute for Nuclear Astrophysics, University of Notre Dame,Notre Dame, IN 46556E-mail: [email protected] , [email protected] , [email protected] [email protected],[email protected]
The recent experimental evaluation of the F( a , p ) Ne reaction rate, when considering its associ-ated uncertainties, presented significant differences compared to the theoretical Hauser-Feshbachrate. This was most apparent at the low temperatures relevant for He-shell burning in asymptoticgiant branch (AGB) stars. Investigations into the effect on AGB nucleosynthesis revealed that theupper limit resulted in an enhanced production of F and Ne in carbon-rich AGB models, butthe recommended and lower limits presented no differences from using the theoretical rate. Thiswas the case for models spanning a range in metallicity from solar to [Fe/H] ∼ − .
3. The resultsof this study are relevant for observations of F and C-enriched AGB stars in the Galaxy, and to theNe composition of mainstream silicon carbide grains, that supposedly formed in the outflows ofcool, carbon-rich giant stars. We discuss the mechanism that produces the extra F and summarizeour main findings. ∗ Speaker. c (cid:13) Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlikeLicence. http://pos.sissa.it/ he F( a ,p) reaction and AGB stars Amanda I. Karakas
1. Introduction
Until 2006 the only rate for the F( a ,p) Ne reaction was the theoretical estimate availablein the Brussels nuclear reaction-rate library [1]. The first experiment aimed at determining the F( a ,p) Ne rate over a large range of stellar temperatures was carried out by Lee [2]. This exper-imental evaluation, when considering its associated uncertainties, presented significant differencescompared to the theoretical rate, especially at the low temperatures relevant for He-shell burning inasymptotic giant branch (AGB) stars ( T ≈ . F( a ,p) Ne rate was not present in previous studies e.g.,[3], although we had included the species F because of its important role in the reaction chain N( a , g ) F( b + n ) O, leading to the production of O in the He shell. It was found that theinclusion of the F( a ,p) reaction resulted in an increase in the production of the stable F; see§3 for more details on the production mechanism. This is of interest because AGB models do notsynthesize enough F to match the [F/O] abundances observed in AGB stars [4]. Also, the cosmicorigin of fluorine is still uncertain, with massive stars [5] playing a significant role in producingfluorine alongside AGB stars. However, AGB stars and their progeny (e.g., post-AGB stars, plane-tary nebulae) are still the only confirmed site of fluorine production thus far [6, 7]. Observations ofan enhanced F abundance ([F/Fe] = 2.90) in a carbon-enhanced metal-poor halo star [8] is furthermotivation to understand the details of F production in AGB stars [9].In this proceedings, we summarize the results of calculations published in Karakas et al. [10].
2. The F( a , p ) Ne reaction rate
The measurement of the F( a ,p) Ne reaction cross section is made difficult by the shorthalf-life ( T / ∼
109 min) of F. The time-reversed reaction of Ne( p , a ) F was investigatedat the Nuclear Science Laboratory in the University of Notre Dame [2]. The cross section wasmeasured in the energy range of 2.3 MeV to 4.0 MeV using the activation method. The lowerlimit of the cross-section measurement is mainly determined by the statistical uncertainty of theactivation data, while the upper limit is based on the uncertainty associated with the O inducedbackground. We refer the reader to [10] for further details.
3. Results
We computed the stellar structure first using the Mt Stromlo Stellar Structure code, and thenperformed post-processing on that structure to obtain abundances for 77 species, most of which arenot included in the small stellar-structure network. See [10] and references therein for more details.This technique is valid for studying reactions not directly related to the main energy generation.This is certainly the case for studying the effect of the F( a , p ) Ne reaction on AGB nucleosyn-thesis. We included models in our study with masses between 1.9 to 5 M ⊙ , with initial metallicitiesfrom solar ( Z = .
02) to Z = . − . C pocket and free neutrons in the He-intershell via the C( a , n ) O reaction. Neutrons are necessary for the chain N( n , p ) C, where free protons are2 he F( a ,p) reaction and AGB stars Amanda I. Karakas
Table 1:
Results from the AGB models. For each mass and Z value, we show the mass of the partial mixingzone used in the computation, the yield ( y ) of F, and the multiplication factor ( X ) needed to obtain theupper limit F yield from the recommended-rate yield. All yields are in solar masses, and the multiplicationfactors are dimensionless quantities. The same information is also presented for Ne for each model.
Mass Z PMZ y ( F rec ) X ( F) y ( Ne rec ) X ( Ne)3.0 0.02 0.002 5.84( −
6) 1.526 1.25( −
6) 4.4233.0 0.012 0.002 5.66( −
6) 1.736 1.39( −
6) 5.3301.9 0.008 0.002 9.35( −
7) 1.178 1.60( −
7) 2.3403.0 0.008 0.002 1.71( −
5) 2.407 4.52( −
6) 9.6092.5 0.004 0.002 1.33( −
5) 2.061 2.81( −
6) 8.3645.0 0.004 0 1.45( −
7) 4.582 − − − −
5) 1.975 3.23( −
6) 8.551then used by O( p , a ) N. We include a PMZ of 0.002 M ⊙ for all lower mass cases. The protonsin the PMZ are captured by the abundant C to form a C pocket that is ≈ −
15% of the massof the He-intershell. We note that the extent in mass and the proton profile of the partial mixingzone are very uncertain parameters (see discussions in [12, 13]).From Table 1 it is evident that employing the upper limit of the F( a , p ) Ne reaction resultsin an increase in the production of F and Ne, compared to using the recommended rate. Thechange in the yield increases with decreasing metallicity, at a given mass, with the largest changefound in the 5 M ⊙ , Z = .
004 model. Note that the amount of F produced in the intermediate-mass models (at a given Z ) is much less than the amount produced in the lower-mass 3 M ⊙ model,by factors of ∼ F is destroyed by HBB in the 5 M ⊙ models.The enhanced abundance of F as a result of using the upper limit may be explained by con-sidering the O( p , a ) N( a , g ) F reaction chain. Including the F( a , p ) Ne reaction reducesthe abundance of O because it competes with O production via the F( b + n ) O decay. How-ever, the extra amount of protons from ( a , p ) enhances the O( p , a ) N reaction rate, even though O production has been deprived from the decay. In other words, the sum N O + N p (where N i isthe abundance by number of nucleus i ) remains constant, however, the product N O N p , on whichthe number of O+ p reactions depends, is maximized when N O is equal to N p . In [10] we an-alyzed the effect of the extra protons on the F production in the He-shell. It was found that theoverall F production increases as long as N N / N p >
1, where N p is the original number densityof protons without the inclusion of the F( a , p ) Ne reaction, and this condition is well satisfiedin the He-burning shell. During the network calculation a realistic N N / N p ≈ ; this ratio islarge enough to explain the enhanced fluorine production in the stellar models.
4. Discussion and conclusions
From Fig. 4, we see that the surface [ F/ O] ratios from the 3 M ⊙ , Z = .
008 model are afactor of ∼ . F( a , p ) Ne reaction. Wechose to show this model because it produces the largest F abundances, although the metallicity3 he F( a ,p) reaction and AGB stars Amanda I. Karakas
Figure 1:
Comparison of fluorine abundances observed by [4] and model predictions for the 3 M ⊙ , Z = . F abundance corresponds to the average Fabundance observed in K and M stars. Each symbol on the prediction lines represents a TDU episode. Solidlines represent calculations performed using no F( a , p ) Ne reaction, which are equivalent to using thecurrent lower limit, recommended value and Brussels library rate. Dotted lines are calculations performedusing the current upper limit of the rate. of this model is probably at the lower end of the distribution of Galactic carbon stars. Table 1and Fig. 4 shows that a match between the stellar models and the stars with the highest observed F abundances is possible, but only for very high C/O ratios of ∼ − M ⊙ , Z = .
008 model. These high C/O ratios are likely not realistic, and the inclusion of carbon-rich, low-temperature opacities into the stellar models would cause the TP-AGB evolution to end before themodel star reached such C/O ratios. A solution to the mystery of the high F abundances at modestC/O ratios is still missing, but a re-evaluation of the F and C abundances in the sample of AGBstars considered by [4] may help, along with a detailed examination of model uncertainties.The F( a ,p) Ne reaction also affects the abundance of Ne in the He-shell of AGB stars.There is a long-standing puzzle concerning the isotopic composition of Ne measured in stellarsilicon carbide (SiC) grains extracted from meteorites, which formed in the envelopes of carbon-rich AGB stars e.g., [14, 15]. Models computed with the upper limit of the F( a , p ) Ne reactionrate show an increase in the Ne abundance, and hence in the Ne/ Ne ratio in the intershell of upto a factor of 6; see [10] for further details. We conclude that the F( a , p ) Ne reaction rate being4 he F( a ,p) reaction and AGB stars Amanda I. Karakas closer to its upper limit may be a promising explanation for the Ne/ Ne ratios in SiC grains. Themodeling uncertainties related to convection and mass loss do not affect the intershell compositionsand thus do not apply to the discussion of the Ne composition of stellar SiC grains. For this reason,the results for Ne are also a more reliable hint that the F( a ,p) N reaction is indeed closer toits upper limit than the comparison of F in AGB stars. However, more experimental data for thisreaction at temperatures below 0.4 GK are required to help verify this result.
References [1] M. Aikawa, M. Arnould, S. Goriely, A. Jorissen, and K. Takahashi. BRUSLIB and NETGEN: theBrussels nuclear reaction rate library and nuclear network generator for astrophysics.
A&A ,441:1195–1203, 2005.[2] H. Y. Lee.
The F( a ,p) Ne Reaction and Its Astrophysical Implications . PhD thesis, University ofNotre Dame, 2006.[3] A. I. Karakas, M. Lugaro, M. Wiescher, J. Goerres, and C. Ugalde. The Uncertainties in the 22Ne + alpha-capture Reaction Rates and the Production of the Heavy Magnesium Isotopes in AsymptoticGiant Branch Stars of Intermediate Mass. ApJ , 643:471–483, 2006.[4] A. Jorissen, V. V. Smith, and D. L. Lambert. Fluorine in red giant stars - Evidence fornucleosynthesis.
A&A , 261:164–187, 1992.[5] S. E. Woosley and T. A. Weaver. The Evolution and Explosion of Massive Stars. II. ExplosiveHydrodynamics and Nucleosynthesis.
ApJS , 101:181, 1995.[6] S. R. Federman, Y. Sheffer, D. L. Lambert, and V. V. Smith. Far Ultraviolet Spectroscopic ExplorerMeasurements of Interstellar Fluorine.
ApJ , 619:884–890, 2005.[7] K. Werner, T. Rauch, and J. W. Kruk. Fluorine in extremely hot post-AGB stars: Evidence fornucleosynthesis.
A&A , 433:641–645, 2005.[8] S. C. Schuler, K. Cunha, V. V. Smith, T. Sivarani, T. C. Beers, and Y. S. Lee. Fluorine in aCarbon-enhanced Metal-poor Star.
ApJ , 667:L81–L84, 2007.[9] M. Lugaro, S. E. de Mink, et al. Fluorine in carbon-enhanced metal-poor stars: a binary scenario.
A&A , 484:L27–L30, 2008.[10] A. I. Karakas, H. Y. Lee, M. Lugaro, J. Görres, and M. Wiescher. The Impact of the F( a ,p) NeReaction on Asymptotic Giant Branch Nucleosynthesis.
ApJ , 676:1254–1261, 2008.[11] A. I. Karakas and J. C. Lattanzio. Stellar Models and Yields from Asymptotic Giant Branch Stars.
Publ. Astron. Soc. Australia , 24:103–117, 2007.[12] F. Herwig. Evolution of Asymptotic Giant Branch Stars
ARA&A , 43:435–479, 2005[13] M. Lugaro, C. Ugalde, A. I. Karakas, J. Görres, M. Wiescher, J. C. Lattanzio, and R. C. Cannon.Reaction Rate Uncertainties and the Production of 19F in Asymptotic Giant Branch Stars
ApJ ,615:934–946, 2004.[14] R. S. Lewis, S. Amari, and E. Anders. Meteoritic silicon carbide - Pristine material from carbon stars.
Nature , 348:293–298, 1990.[15] E. Zinner, L. R. Nittler, R. Gallino, A. I. Karakas, M. Lugaro, O. Straniero, and J. C. Lattanzio.Silicon and Carbon Isotopic Ratios in AGB Stars: SiC Grain Data, Models, and the GalacticEvolution of the Si Isotopes.
ApJ , 650:350–373, 2006., 650:350–373, 2006.