- Email: [email protected]
The s process in massive stars
Progress in Particle and Nuclear Physics 59 (2007) 174–182 www.elsevier.com/locate/ppnp
Review
The s process in massive stars M. Heil a,∗ , F. K¨appeler a , E. Uberseder a , R. Gallino b , M. Pignatari b a Forschungszentrum Karlsruhe, Institut f¨ur Kernphysik, D-76021 Karlsruhe, Germany b Dipartimento di Fisica Generale, Universit`a di Torino, via P. Giuria 1, 10125 (To), Italy
Abstract In the past decades considerable progress has been made in the understanding of the various processes responsible for the nucleosynthesis of the elements in stars. This holds especially for the main s process in low-mass AGB stars, which is largely responsible for the production of about half of the elemental abundances in the mass range 90 ≤ A ≤ 209. The weak s process, which produces elements with A ≤ 90, however, is much less understood. A better characterization of the weak s component would help to disentangle the various processes contributing to element production in this region. In this respect, more accurate neutron capture cross sections in the mass range 56 ≤ A ≤ 90 are indispensable. Also, neutron captures on abundant light elements with A < 56 play an important role, since they act as neutron poisons and affect the stellar neutron balance. In this context, the impact of new results for neutron capture cross sections on light and medium mass nuclei is discussed. c 2007 Elsevier B.V. All rights reserved.
Keywords: Stellar nucleosynthesis; Neutron capture reactions; s process
1. Introduction Since 1957, after the pioneering work by Burbidge, Burbidge, Fowler and Hoyle [1], it has been known that the elements heavier than iron are mainly produced by two neutron capture processes, the s (slow) and the r (rapid) process, both contributing about half of the observed solar abundances between Fe and U. A third process, the so-called p (photodissociation) process, is responsible for the origin of about 30 rare, proton-rich nuclei, but does not contribute significantly
∗ Corresponding author. Tel.: +49 6159 71 2433; fax: +49 6159 71 2902.
E-mail address: [email protected] (M. Heil). c 2007 Elsevier B.V. All rights reserved. 0146-6410/$ - see front matter doi:10.1016/j.ppnp.2006.12.013
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
175
to the synthesis of the elements in general (<1%). Since 1957 most progress has been made in the field of the s process. It became apparent that a single s process was not sufficient to explain the observed solar abundances. At least two components, the main and the weak s process, were called for and could be connected with the corresponding stellar objects and sites. A third component, the strong s process, which had to be considered responsible for part of the observed 208 Pb and 209 Bi abundances, has meanwhile been superseded, since it was found that the main s process in low-metallicity stars can account for the missing 208 Pb and 209 Bi [2]. The main s process, by far the most studied process, occurs in the He-rich intershell of thermally pulsing AGB stars and produces predominantly nuclei with mass numbers A > 90. These evolved Red Giants have already burnt all the H and He in their core to carbon and oxygen. The energy generation occurs by alternate episodes of H and He burning, which are separated by a thin He-rich intershell. The main s process takes place in this intershell, where neutrons are liberated by the 13 C(α, n)16 O reaction at temperatures of T8 ∼ 1 and at neutron densities of about 107 cm−3 . Since there are not many seed nuclei in this thin shell, the neutron/seed ratio is high and the s process operates very efficiently over a long period of time. During the subsequent convective He flashes, the freshly synthesized material is mixed and diluted with the He intershell and is again exposed to neutrons liberated by the 22 Ne(α, n)25 Mg reaction at temperatures T8 ≥ 2.5. The second neutron exposure is rather weak and not sufficient to produce s isotopes on a grand scale but strong enough to alter the isotope ratios of s-process branchings. After the He flash, where peak neutron densities of 1010 cm−3 are reached, part of the freshly synthesized material is mixed with the envelope and brought to the surface of the star, where it is detectable by spectroscopy. The weak s component, which is responsible for the production of nuclei between iron and yttrium (56 < A < 90), takes place during convective core-He burning in massive stars (M > 8), where for a short time temperatures of (2.2–3.5) × 108 K are reached and neutrons are liberated by the activation of the 22 Ne(α, n)25 Mg reaction. Since the neutron exposure is small, the s-process flow cannot overcome the bottleneck at the closed neutron shell N = 50. However, most of the material in the core is reprocessed by the following burning stages and only a small part survives in the outer layers and is ejected during the supernova explosion. Recent stellar models also discuss the possibility of a second neutron exposure during convective carbon shell burning of massive stars [3,5]. There, neutrons are produced by various reactions, e.g. 22 Ne(α, n)25 Mg, 17 O(α, n)20 Ne and 13 C(α, n)16 O. The high temperatures during carbon burning of T9 ∼ 1 cause high neutron densities, which start at ∼1011 –1012 cm−3 and then decrease exponentially. The nucleosynthesis yields of the weak component in massive stars are also important for the r process, since they determine the composition of a star before the supernova explosion. In addition, since the s-process abundances can be determined reliably on the basis of experimental (n, γ ) cross sections, the r abundances are commonly inferred by the r-residual method, that is by subtracting the s abundances from solar values [4]: Nr = N − Ns . The r abundances obtained in this way are then used for testing r-process models. This is of special interest, since recent observations of ultra-metal-poor halo stars suggest a second, so-called weak r process, which contributes to element production below barium. These very old stars do not yet show an s contribution, since the s process is a secondary process, which requires seed nuclei that were not present at early times. Correspondingly, these stars are expected to exhibit a pure r-process abundance distribution. Indeed, comparison of the observed abundances and the scaled solar r abundances, which were derived by the r-residual
176
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
method, are in excellent agreement for elements with atomic numbers Z ≥ 56, suggesting a very robust and unique r process [6]. This observation is meanwhile confirmed for a number of metal-poor halo stars like CS22892-052 [6], HD 115444 [7], BD + 17◦ 3248 [8] and CS 31082001 [9]. Below barium (Z = 56), however, the observed abundances fall systematically below the solar r distribution. This could be interpreted as a missing r-process component, but this argument relies necessarily on the r-residual method, and hence on reliable s-process abundances. However, the s abundances in the mass region between Fe and Ba are more difficult to determine, because the weak s process contributes significantly to the medium-heavy elements. In addition, current s-process models cannot explain the high observed abundances of the typical s elements Sr, Y and Zr in halo, thick disk and thin disk stars with low metallicity [10]. Again, since the s process is a secondary process, significant contributions are expected only for metallicities [Fe/H] > −1.5. Therefore, a new primary neutron capture process, the lighter element primary process (LEPP), was suggested by Travaglio et al. [10] and model calculations showed that intermediate conditions between typical r- and s-process environments are needed in this case (e.g. neutron densities of 1014 cm−3 ). Recently, a totally new process, the νp process, was introduced [11] which is a promising candidate for the missing process. These nucleosynthesis calculations explored the so far neglected effect of neutrino interaction and found that it is possible to produce neutrons via antineutrino captures on protons in the innermost proton-rich ejecta of core-collapse supernovae. Neutron densities of 1014 −1015 cm−3 could be obtained in this way for several seconds, when the temperatures are in the 1–3 GK range. These appear to be adequate conditions for the postulated LEPP process. The νp process offers also a natural explanation for the production of the pprocess nuclei 92,94 Mo and 96,98 Ru. So far, p-process models [12] have had enormous problems in describing the light p nuclei with A < 100, especially the relatively large abundances of 92 Mo, 94 Mo, 96 Ru and 98 Ru. It is clear from the above discussion that several processes contribute to the nucleosynthesis of medium-heavy nuclei from Fe to Ba. In order to disentangle the various processes and to identify possible astrophysical sites the individual contributions have to be identified. This contribution focuses on the weak s process. The cross sections of light elements are also important in this respect because they affect the neutron balance inside stars during the s process. Although their cross sections are small, these elements are much more abundant than those in the mass region above Fe. Therefore, light elements constitute potential neutron poisons and may consume neutrons, which are then not available for s-process nucleosynthesis. Especially important in this respect are neutron captures on the CNO elements and on the neon and magnesium isotopes, but also other light isotopes up to iron contribute as well. For many of these isotopes the neutron capture cross sections are not known with sufficient accuracy since they are small and difficult to measure. Neutron capture cross sections of light isotopes also play an important role in analyses of presolar grains, which can provide stringent constraints on the s-process models [13]. Because these grains are only a few µm in size and because the abundances of heavy elements are rather low, their isotopic abundance components in the grains are difficult to analyze. Lighter elements are more abundant, and therefore easier to detect. In this contribution we report on a series of neutron capture measurements of light and medium-heavy nuclei relevant for the weak s process in massive stars. The experimental method and the results are described in Section 2, and the consequences for the weak s process are discussed in Section 3. Section 4 concludes with a summary and outlook.
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
177
Fig. 1. The experimental setup for activation measurements (left) and a comparison of the produced neutron spectrum and a thermal neutron spectrum at kT = 25 keV (right).
2. Neutron capture measurements The most important inputs for stellar models of the s process are Maxwellian averaged neutron capture cross sections (MACSs) and β-decay rates, but stellar enhancement factors (SEF) also have to be known. The weak s process is responsible for the production of elements in the mass range 56 ≤ A ≤ 90. In this mass range the (n, γ ) cross sections need substantial improvement for deriving a reliable description of the abundance contributions from massive stars. This is especially important since the local approximation (hσ iN = const) is not valid during the weak s process. Therefore, any change in the cross section of a light isotope, e.g. 62 Ni [14], can affect the abundances of all the heavier isotopes up to zirconium and maybe even higher up. This underlines that neutron capture cross sections in the mass range 50 ≤ A ≤ 90 have to be measured with significantly higher accuracy. Therefore, a measuring campaign was launched at Forschungszentrum Karlsruhe with the aim of improving the neutron capture cross sections in this region. A reliable and accurate approach to the derivation of Maxwellian averaged cross sections at kT = 25 keV is the activation method [15], where the 7 Li(p, n)7 Be reaction is used to produce a quasi-stellar neutron spectrum as sketched in Fig. 1. After irradiation in that spectrum the induced sample activity is counted in a low-background environment. The proton beam with an energy of E p = 1912 keV and typical intensities of 100 µA was delivered by the Karlsruhe 3.7 MV van de Graaff accelerator. The neutron production target consists of a metallic Li layer, which is evaporated onto a water cooled copper backing. The sample is placed inside the resulting neutron cone, which has an opening angle of 120 deg. The neutron flux is monitored throughout the irradiation by means of a 6 Li-glass detector, positioned at a distance of 1 m from the target. After the irradiation the total number of activated nuclei A is given by A = Φ · N · σ · fb ,
(1)
where Φ is the time integrated neutron flux, N the number of sample atoms per cm2 and σ the spectrum averaged neutron capture cross section. In order to determine the neutron flux, the sample is sandwiched between gold foils. Since the gold cross section is well known, the total
178
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
Table 1 MACS at kT = 30 keV (in mb) Target isotope
Previously recommended [16]
This work
23 Na
2.1 ± 0.2 12.1 ± 2 38 ± 4 8.7 ± 0.9 94 ± 10 41 ± 5 627 ± 42 313 ± 16 15.5 ± 1.5
1.5 ± 0.1 13.0 ± 0.6 38.7 ± 2 7.2 ± 0.4 54.2 ± 2 28.7 ± 2 615 ± 19 232 ± 9 15.9 ± 2
58 Fe 59 Co 64 Ni 63 Cu 65 Cu 79 Br 81 Br 87 Rb
number of neutrons can be obtained by measuring the 412 keV line from the decay of 198 Au with HPGe detectors. The factor f b accounts for the variation of the neutron flux and for the decay during activation. The cross section can then be calculated from the number of counts in a characteristic γ -ray line Cγ = A · K γ · εγ · Iγ · (1 − exp(−λtm )) · exp(−λtw ),
(2)
where K γ is a correction factor for γ -ray self-absorption, εγ the efficiency of the Ge detector, Iγ the line intensity, tw the waiting time between irradiation and counting and tm the duration of the activity measurement. In this way, we have measured the Maxwellian averaged capture cross section of several isotopes. The results are shown in Table 1 together with the previously recommended values [16]. For some isotopes the measured cross sections differ significantly from previous recommendations. It is conspicuous that many new results are systematically smaller than the recommended cross sections from [16], which are often based on time-of-flight (TOF) measurements performed with C6 D6 detectors in the 1970s and early 1980s [17–19]. In fact, this trend is confirmed by a general comparison between MACSs obtained with the activation method and the TOF method performed with C6 D6 detectors, which reveal large discrepancies on average. The cross sections from activation measurements are consistently lower, often in complete disagreement with the quoted uncertainties. This trend is illustrated in Fig. 2 by the comparison between MACSs at kT = 30 keV obtained with the TOF and the activation method. A possible explanation could be that the background due to sample-scattered neutrons was systematically underestimated in the older TOF experiments. Neutrons scattered in the sample and captured in the detector and/or in surrounding materials produce background events, which are difficult to distinguish from true capture events. This background can be as high as 50% for light and medium-heavy nuclei, where the scattering/capture ratios are large. The correspondingly large and uncertain corrections tend to give rise to large systematic errors. 3. Astrophysical consequences To explore the effect for the weak s process, stellar model calculations for a 25M star were performed with an updated post-processing code described in [20,21]. Fig. 3 shows the effect of the new 64 Ni cross section. Plotted are the nucleosynthesis yields at the end of carbon burning relative to the yields obtained with the previous 64 Ni cross section. Since the new cross section is
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
179
Fig. 2. Comparison of MACSs at kT = 30 keV obtained by the activation and TOF method. Note that most ratios are smaller than unity.
Fig. 3. Nucleosynthesis yields of stellar model calculations for a 25M star at the end of carbon burning with the new 64 Ni cross section relative to the yields obtained with the previous cross sections.
smaller, the calculated abundance of 64 Ni is accordingly higher. But one can also see, similar to the case of 62 Ni [14], a strong influence on the following heavier isotopes. Therefore, one has to conclude that a reliable abundance prediction of the weak s process is only possible if all neutron capture cross sections of the involved isotopes are known with high accuracy. Fig. 4 shows the combined effect of all new cross sections listed in Table 1 on the abundance distribution of the weak s process after core-He burning (top) and shell-C burning (bottom), respectively. It has already been mentioned that light nuclei can be important neutron poisons because of their very high abundances. Therefore, we have also measured the neutron capture cross section of 23 Na and found a significantly smaller value than was recommended previously [16]. The effect of this result for the s-process yields produced in a 25M star is illustrated in Fig. 5. Obviously, the reduction of this cross section leads to an increase in s-process efficiency, which partly compensates the effect of the smaller cross sections near the Fe seed. The neutron poison
180
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
Fig. 4. Abundance yields obtained for a 25M star after core-helium burning (top) and carbon-shell burning (bottom) using the neutron capture cross sections of this work relative to the standard case. 23 Na is especially important at the late stage of carbon-shell burning since it is strongly produced
via the 12 C(12 C, p)23 Na reaction. We note that the increase in s-process efficiency also implies a higher neutron density in this scenario as witnessed by the markedly smaller 64 Zn abundance. Similar indications can also be found in the previous Figs. 3 and 4. 4. Summary and outlook The present measurements have shown that the nucleosynthesis yields for the weak s process, calculated with stellar models for massive stars, still show large variations due to the uncertainties of the neutron capture cross sections involved. Unlike the main s process, flow equilibrium is not reached during the weak s process. Therefore, improved neutron capture cross sections do not only influence the yield of the respective isotope, but also the production of all heavier nuclei on the weak s-process path. Observations of metal-poor halo stars show that the abundances of the elements below Ba cannot be fully accounted for by the present nucleosynthesis processes. There seems to be a need for an additional s and/or r process, which is expected to add 10%–20% of the abundances for some elements between Fe and Ba. Before this problem can be settled, more
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
181
Fig. 5. Effect of the new 23 Na(n, γ )24 Na cross section on the nucleosynthesis yields for a 25M star at the end of carbon burning.
accurate neutron capture cross sections are needed to see whether the weak s process itself may provide the missing fraction. It is planned to further pursue activation measurements of isotopes relevant for the weak s process at Karlsruhe, but for isotopes which are not accessible by this method, time-of-flight experiments have to be performed. Especially important would be neutron capture measurements of 58 Fe, all Ge isotopes, all Ga isotopes and all Zn isotopes. Also the abundant light isotopes below Fe are important, since they may constitute crucial neutron poisons for the s process. The respective neutron capture cross sections still show large uncertainties, in particular for 12 C, 16 O and 22 Ne. References [1] E.M. Burbidge, G.R. Burbidge, W.A. Fowler, F. Hoyle, Rev. Modern Phys. 29 (1957) 547. [2] R. Gallino, C. Arlandini, M. Busso, M. Lugaro, C. Travaglio, O. Straniero, A. Chieffi, M. Limongi, Astrophys. J. 497 (1998) 388. [3] C.M. Raiteri, R. Gallino, M. Busso, D. Neuberger, F. K¨appeler, Astrophys. J. 419 (1993) 207. [4] E. Anders, N. Grevesse, Geochim. Cosmochim. Acta 53 (1989) 197. [5] T. Rauscher, A. Heger, R.D. Hoffman, S.E. Woosley, Astrophys. J. 576 (2002) 323. [6] C. Sneden, J.J. Cowan, J.E. Lawler, I.I. Ivans, S. Burles, T.C. Beers, F. Primas, V. Hill, J.W. Truran, G.M. Fuller, B. Pfeiffer, K.-L. Kratz, Astrophys. J. 591 (2003) 936. [7] J. Westin, C. Sneden, B. Gustafsson, J.J. Cowan, Astrophys. J. 530 (2000) 783. [8] J.J. Cowan, C. Sneden, S. Burles, I.I. Ivans, T.C. Beers, J.W. Truran, J.E. Lawler, F. Primas, G.M. Fuller, B. Pfeiffer, K.-L. Kratz, Astrophys. J. 572 (2002) 861. [9] V. Hill, B. Plez, R. Cayrel, T.C. Beers, B. Nordstr¨om, J. Andersen, M. Spite, F. Spite, B. Barbuy, P. Bonifacio, E. Depagne, P. Francois, F. Primas, Astron. Astrophys. 387 (5) (2002) 560. [10] C. Travaglio, R. Gallino, E. Arnone, J. Cowan, F. Jordan, C. Sneden, Astrophys. J. 601 (2004) 864. [11] C. Fr¨ohlich, G. Martinez-Pinedo, M. Liebend¨orfer, F.-K. Thielemann, E. Bravo, W.R. Hix, K. Langanke, N.T. Zinner, Phys. Rev. Lett. 96 (2006) 142502. [12] M. Arnould, S. Goriely, Phys. Rep. 384 (2003) 1. [13] E. Zinner, Ann. Rev. Earth Planet. Sci. 26 (1998) 147. [14] H. Nassar, M. Paul, I. Ahmad, D. Berkovits, M. Bettan, P. Collon, S. Dababneh, S. Ghelberg, J.P. Greene, A. Heger, M. Heil, D.J. Henderson, C.L. Jiang, F. K¨appeler, H. Koivisto, S. O’Brien, R.C. Pardo, N. Patronis, T. Pennington, R. Plag, K.E. Rehm, R. Reifarth, R. Scott, S. Sinha, X. Tang, R. Vondrasek, Phys. Rev. Lett. 94 (2005) 092504. [15] H. Beer, F. K¨appeler, Phys. Rev. C 21 (1980) 534. [16] Z.Y. Bao, H. Beer, F. K¨appeler, F. Voss, K. Wisshak, T. Rauscher, Atomic Data Nucl. Data Tables 76 (2000) 70.
182
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
[17] M. Kenny, B. Allen, R. Macklin, Aust. J. Phys. 30 (1977) 605. [18] R. Macklin, Nucl. Sci. Eng. 88 (1984) 129. [19] A.de L. Musgrove, B. Allen, R. Macklin, Neutron Physics and Nuclear Data for Reactors and Other Applied Purposes, 426, OECD, Paris, 1978, p. 1. [20] C.M. Raiteri, M. Busso, G. Picchio, R. Gallino, Astrophys. J. 371 (1991) 665. [21] F. K¨appeler, M. Wiescher, U. Giesen, J. Goerres, I. Baraffe, M. El Eid, C.M. Raiteri, M. Busso, R. Gallino, M. Limongi, A. Chieffi, Astrophys. J. 437 (1994) 396.
Review
The s process in massive stars M. Heil a,∗ , F. K¨appeler a , E. Uberseder a , R. Gallino b , M. Pignatari b a Forschungszentrum Karlsruhe, Institut f¨ur Kernphysik, D-76021 Karlsruhe, Germany b Dipartimento di Fisica Generale, Universit`a di Torino, via P. Giuria 1, 10125 (To), Italy
Abstract In the past decades considerable progress has been made in the understanding of the various processes responsible for the nucleosynthesis of the elements in stars. This holds especially for the main s process in low-mass AGB stars, which is largely responsible for the production of about half of the elemental abundances in the mass range 90 ≤ A ≤ 209. The weak s process, which produces elements with A ≤ 90, however, is much less understood. A better characterization of the weak s component would help to disentangle the various processes contributing to element production in this region. In this respect, more accurate neutron capture cross sections in the mass range 56 ≤ A ≤ 90 are indispensable. Also, neutron captures on abundant light elements with A < 56 play an important role, since they act as neutron poisons and affect the stellar neutron balance. In this context, the impact of new results for neutron capture cross sections on light and medium mass nuclei is discussed. c 2007 Elsevier B.V. All rights reserved.
Keywords: Stellar nucleosynthesis; Neutron capture reactions; s process
1. Introduction Since 1957, after the pioneering work by Burbidge, Burbidge, Fowler and Hoyle [1], it has been known that the elements heavier than iron are mainly produced by two neutron capture processes, the s (slow) and the r (rapid) process, both contributing about half of the observed solar abundances between Fe and U. A third process, the so-called p (photodissociation) process, is responsible for the origin of about 30 rare, proton-rich nuclei, but does not contribute significantly
∗ Corresponding author. Tel.: +49 6159 71 2433; fax: +49 6159 71 2902.
E-mail address: [email protected] (M. Heil). c 2007 Elsevier B.V. All rights reserved. 0146-6410/$ - see front matter doi:10.1016/j.ppnp.2006.12.013
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
175
to the synthesis of the elements in general (<1%). Since 1957 most progress has been made in the field of the s process. It became apparent that a single s process was not sufficient to explain the observed solar abundances. At least two components, the main and the weak s process, were called for and could be connected with the corresponding stellar objects and sites. A third component, the strong s process, which had to be considered responsible for part of the observed 208 Pb and 209 Bi abundances, has meanwhile been superseded, since it was found that the main s process in low-metallicity stars can account for the missing 208 Pb and 209 Bi [2]. The main s process, by far the most studied process, occurs in the He-rich intershell of thermally pulsing AGB stars and produces predominantly nuclei with mass numbers A > 90. These evolved Red Giants have already burnt all the H and He in their core to carbon and oxygen. The energy generation occurs by alternate episodes of H and He burning, which are separated by a thin He-rich intershell. The main s process takes place in this intershell, where neutrons are liberated by the 13 C(α, n)16 O reaction at temperatures of T8 ∼ 1 and at neutron densities of about 107 cm−3 . Since there are not many seed nuclei in this thin shell, the neutron/seed ratio is high and the s process operates very efficiently over a long period of time. During the subsequent convective He flashes, the freshly synthesized material is mixed and diluted with the He intershell and is again exposed to neutrons liberated by the 22 Ne(α, n)25 Mg reaction at temperatures T8 ≥ 2.5. The second neutron exposure is rather weak and not sufficient to produce s isotopes on a grand scale but strong enough to alter the isotope ratios of s-process branchings. After the He flash, where peak neutron densities of 1010 cm−3 are reached, part of the freshly synthesized material is mixed with the envelope and brought to the surface of the star, where it is detectable by spectroscopy. The weak s component, which is responsible for the production of nuclei between iron and yttrium (56 < A < 90), takes place during convective core-He burning in massive stars (M > 8), where for a short time temperatures of (2.2–3.5) × 108 K are reached and neutrons are liberated by the activation of the 22 Ne(α, n)25 Mg reaction. Since the neutron exposure is small, the s-process flow cannot overcome the bottleneck at the closed neutron shell N = 50. However, most of the material in the core is reprocessed by the following burning stages and only a small part survives in the outer layers and is ejected during the supernova explosion. Recent stellar models also discuss the possibility of a second neutron exposure during convective carbon shell burning of massive stars [3,5]. There, neutrons are produced by various reactions, e.g. 22 Ne(α, n)25 Mg, 17 O(α, n)20 Ne and 13 C(α, n)16 O. The high temperatures during carbon burning of T9 ∼ 1 cause high neutron densities, which start at ∼1011 –1012 cm−3 and then decrease exponentially. The nucleosynthesis yields of the weak component in massive stars are also important for the r process, since they determine the composition of a star before the supernova explosion. In addition, since the s-process abundances can be determined reliably on the basis of experimental (n, γ ) cross sections, the r abundances are commonly inferred by the r-residual method, that is by subtracting the s abundances from solar values [4]: Nr = N − Ns . The r abundances obtained in this way are then used for testing r-process models. This is of special interest, since recent observations of ultra-metal-poor halo stars suggest a second, so-called weak r process, which contributes to element production below barium. These very old stars do not yet show an s contribution, since the s process is a secondary process, which requires seed nuclei that were not present at early times. Correspondingly, these stars are expected to exhibit a pure r-process abundance distribution. Indeed, comparison of the observed abundances and the scaled solar r abundances, which were derived by the r-residual
176
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
method, are in excellent agreement for elements with atomic numbers Z ≥ 56, suggesting a very robust and unique r process [6]. This observation is meanwhile confirmed for a number of metal-poor halo stars like CS22892-052 [6], HD 115444 [7], BD + 17◦ 3248 [8] and CS 31082001 [9]. Below barium (Z = 56), however, the observed abundances fall systematically below the solar r distribution. This could be interpreted as a missing r-process component, but this argument relies necessarily on the r-residual method, and hence on reliable s-process abundances. However, the s abundances in the mass region between Fe and Ba are more difficult to determine, because the weak s process contributes significantly to the medium-heavy elements. In addition, current s-process models cannot explain the high observed abundances of the typical s elements Sr, Y and Zr in halo, thick disk and thin disk stars with low metallicity [10]. Again, since the s process is a secondary process, significant contributions are expected only for metallicities [Fe/H] > −1.5. Therefore, a new primary neutron capture process, the lighter element primary process (LEPP), was suggested by Travaglio et al. [10] and model calculations showed that intermediate conditions between typical r- and s-process environments are needed in this case (e.g. neutron densities of 1014 cm−3 ). Recently, a totally new process, the νp process, was introduced [11] which is a promising candidate for the missing process. These nucleosynthesis calculations explored the so far neglected effect of neutrino interaction and found that it is possible to produce neutrons via antineutrino captures on protons in the innermost proton-rich ejecta of core-collapse supernovae. Neutron densities of 1014 −1015 cm−3 could be obtained in this way for several seconds, when the temperatures are in the 1–3 GK range. These appear to be adequate conditions for the postulated LEPP process. The νp process offers also a natural explanation for the production of the pprocess nuclei 92,94 Mo and 96,98 Ru. So far, p-process models [12] have had enormous problems in describing the light p nuclei with A < 100, especially the relatively large abundances of 92 Mo, 94 Mo, 96 Ru and 98 Ru. It is clear from the above discussion that several processes contribute to the nucleosynthesis of medium-heavy nuclei from Fe to Ba. In order to disentangle the various processes and to identify possible astrophysical sites the individual contributions have to be identified. This contribution focuses on the weak s process. The cross sections of light elements are also important in this respect because they affect the neutron balance inside stars during the s process. Although their cross sections are small, these elements are much more abundant than those in the mass region above Fe. Therefore, light elements constitute potential neutron poisons and may consume neutrons, which are then not available for s-process nucleosynthesis. Especially important in this respect are neutron captures on the CNO elements and on the neon and magnesium isotopes, but also other light isotopes up to iron contribute as well. For many of these isotopes the neutron capture cross sections are not known with sufficient accuracy since they are small and difficult to measure. Neutron capture cross sections of light isotopes also play an important role in analyses of presolar grains, which can provide stringent constraints on the s-process models [13]. Because these grains are only a few µm in size and because the abundances of heavy elements are rather low, their isotopic abundance components in the grains are difficult to analyze. Lighter elements are more abundant, and therefore easier to detect. In this contribution we report on a series of neutron capture measurements of light and medium-heavy nuclei relevant for the weak s process in massive stars. The experimental method and the results are described in Section 2, and the consequences for the weak s process are discussed in Section 3. Section 4 concludes with a summary and outlook.
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
177
Fig. 1. The experimental setup for activation measurements (left) and a comparison of the produced neutron spectrum and a thermal neutron spectrum at kT = 25 keV (right).
2. Neutron capture measurements The most important inputs for stellar models of the s process are Maxwellian averaged neutron capture cross sections (MACSs) and β-decay rates, but stellar enhancement factors (SEF) also have to be known. The weak s process is responsible for the production of elements in the mass range 56 ≤ A ≤ 90. In this mass range the (n, γ ) cross sections need substantial improvement for deriving a reliable description of the abundance contributions from massive stars. This is especially important since the local approximation (hσ iN = const) is not valid during the weak s process. Therefore, any change in the cross section of a light isotope, e.g. 62 Ni [14], can affect the abundances of all the heavier isotopes up to zirconium and maybe even higher up. This underlines that neutron capture cross sections in the mass range 50 ≤ A ≤ 90 have to be measured with significantly higher accuracy. Therefore, a measuring campaign was launched at Forschungszentrum Karlsruhe with the aim of improving the neutron capture cross sections in this region. A reliable and accurate approach to the derivation of Maxwellian averaged cross sections at kT = 25 keV is the activation method [15], where the 7 Li(p, n)7 Be reaction is used to produce a quasi-stellar neutron spectrum as sketched in Fig. 1. After irradiation in that spectrum the induced sample activity is counted in a low-background environment. The proton beam with an energy of E p = 1912 keV and typical intensities of 100 µA was delivered by the Karlsruhe 3.7 MV van de Graaff accelerator. The neutron production target consists of a metallic Li layer, which is evaporated onto a water cooled copper backing. The sample is placed inside the resulting neutron cone, which has an opening angle of 120 deg. The neutron flux is monitored throughout the irradiation by means of a 6 Li-glass detector, positioned at a distance of 1 m from the target. After the irradiation the total number of activated nuclei A is given by A = Φ · N · σ · fb ,
(1)
where Φ is the time integrated neutron flux, N the number of sample atoms per cm2 and σ the spectrum averaged neutron capture cross section. In order to determine the neutron flux, the sample is sandwiched between gold foils. Since the gold cross section is well known, the total
178
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
Table 1 MACS at kT = 30 keV (in mb) Target isotope
Previously recommended [16]
This work
23 Na
2.1 ± 0.2 12.1 ± 2 38 ± 4 8.7 ± 0.9 94 ± 10 41 ± 5 627 ± 42 313 ± 16 15.5 ± 1.5
1.5 ± 0.1 13.0 ± 0.6 38.7 ± 2 7.2 ± 0.4 54.2 ± 2 28.7 ± 2 615 ± 19 232 ± 9 15.9 ± 2
58 Fe 59 Co 64 Ni 63 Cu 65 Cu 79 Br 81 Br 87 Rb
number of neutrons can be obtained by measuring the 412 keV line from the decay of 198 Au with HPGe detectors. The factor f b accounts for the variation of the neutron flux and for the decay during activation. The cross section can then be calculated from the number of counts in a characteristic γ -ray line Cγ = A · K γ · εγ · Iγ · (1 − exp(−λtm )) · exp(−λtw ),
(2)
where K γ is a correction factor for γ -ray self-absorption, εγ the efficiency of the Ge detector, Iγ the line intensity, tw the waiting time between irradiation and counting and tm the duration of the activity measurement. In this way, we have measured the Maxwellian averaged capture cross section of several isotopes. The results are shown in Table 1 together with the previously recommended values [16]. For some isotopes the measured cross sections differ significantly from previous recommendations. It is conspicuous that many new results are systematically smaller than the recommended cross sections from [16], which are often based on time-of-flight (TOF) measurements performed with C6 D6 detectors in the 1970s and early 1980s [17–19]. In fact, this trend is confirmed by a general comparison between MACSs obtained with the activation method and the TOF method performed with C6 D6 detectors, which reveal large discrepancies on average. The cross sections from activation measurements are consistently lower, often in complete disagreement with the quoted uncertainties. This trend is illustrated in Fig. 2 by the comparison between MACSs at kT = 30 keV obtained with the TOF and the activation method. A possible explanation could be that the background due to sample-scattered neutrons was systematically underestimated in the older TOF experiments. Neutrons scattered in the sample and captured in the detector and/or in surrounding materials produce background events, which are difficult to distinguish from true capture events. This background can be as high as 50% for light and medium-heavy nuclei, where the scattering/capture ratios are large. The correspondingly large and uncertain corrections tend to give rise to large systematic errors. 3. Astrophysical consequences To explore the effect for the weak s process, stellar model calculations for a 25M star were performed with an updated post-processing code described in [20,21]. Fig. 3 shows the effect of the new 64 Ni cross section. Plotted are the nucleosynthesis yields at the end of carbon burning relative to the yields obtained with the previous 64 Ni cross section. Since the new cross section is
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
179
Fig. 2. Comparison of MACSs at kT = 30 keV obtained by the activation and TOF method. Note that most ratios are smaller than unity.
Fig. 3. Nucleosynthesis yields of stellar model calculations for a 25M star at the end of carbon burning with the new 64 Ni cross section relative to the yields obtained with the previous cross sections.
smaller, the calculated abundance of 64 Ni is accordingly higher. But one can also see, similar to the case of 62 Ni [14], a strong influence on the following heavier isotopes. Therefore, one has to conclude that a reliable abundance prediction of the weak s process is only possible if all neutron capture cross sections of the involved isotopes are known with high accuracy. Fig. 4 shows the combined effect of all new cross sections listed in Table 1 on the abundance distribution of the weak s process after core-He burning (top) and shell-C burning (bottom), respectively. It has already been mentioned that light nuclei can be important neutron poisons because of their very high abundances. Therefore, we have also measured the neutron capture cross section of 23 Na and found a significantly smaller value than was recommended previously [16]. The effect of this result for the s-process yields produced in a 25M star is illustrated in Fig. 5. Obviously, the reduction of this cross section leads to an increase in s-process efficiency, which partly compensates the effect of the smaller cross sections near the Fe seed. The neutron poison
180
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
Fig. 4. Abundance yields obtained for a 25M star after core-helium burning (top) and carbon-shell burning (bottom) using the neutron capture cross sections of this work relative to the standard case. 23 Na is especially important at the late stage of carbon-shell burning since it is strongly produced
via the 12 C(12 C, p)23 Na reaction. We note that the increase in s-process efficiency also implies a higher neutron density in this scenario as witnessed by the markedly smaller 64 Zn abundance. Similar indications can also be found in the previous Figs. 3 and 4. 4. Summary and outlook The present measurements have shown that the nucleosynthesis yields for the weak s process, calculated with stellar models for massive stars, still show large variations due to the uncertainties of the neutron capture cross sections involved. Unlike the main s process, flow equilibrium is not reached during the weak s process. Therefore, improved neutron capture cross sections do not only influence the yield of the respective isotope, but also the production of all heavier nuclei on the weak s-process path. Observations of metal-poor halo stars show that the abundances of the elements below Ba cannot be fully accounted for by the present nucleosynthesis processes. There seems to be a need for an additional s and/or r process, which is expected to add 10%–20% of the abundances for some elements between Fe and Ba. Before this problem can be settled, more
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
181
Fig. 5. Effect of the new 23 Na(n, γ )24 Na cross section on the nucleosynthesis yields for a 25M star at the end of carbon burning.
accurate neutron capture cross sections are needed to see whether the weak s process itself may provide the missing fraction. It is planned to further pursue activation measurements of isotopes relevant for the weak s process at Karlsruhe, but for isotopes which are not accessible by this method, time-of-flight experiments have to be performed. Especially important would be neutron capture measurements of 58 Fe, all Ge isotopes, all Ga isotopes and all Zn isotopes. Also the abundant light isotopes below Fe are important, since they may constitute crucial neutron poisons for the s process. The respective neutron capture cross sections still show large uncertainties, in particular for 12 C, 16 O and 22 Ne. References [1] E.M. Burbidge, G.R. Burbidge, W.A. Fowler, F. Hoyle, Rev. Modern Phys. 29 (1957) 547. [2] R. Gallino, C. Arlandini, M. Busso, M. Lugaro, C. Travaglio, O. Straniero, A. Chieffi, M. Limongi, Astrophys. J. 497 (1998) 388. [3] C.M. Raiteri, R. Gallino, M. Busso, D. Neuberger, F. K¨appeler, Astrophys. J. 419 (1993) 207. [4] E. Anders, N. Grevesse, Geochim. Cosmochim. Acta 53 (1989) 197. [5] T. Rauscher, A. Heger, R.D. Hoffman, S.E. Woosley, Astrophys. J. 576 (2002) 323. [6] C. Sneden, J.J. Cowan, J.E. Lawler, I.I. Ivans, S. Burles, T.C. Beers, F. Primas, V. Hill, J.W. Truran, G.M. Fuller, B. Pfeiffer, K.-L. Kratz, Astrophys. J. 591 (2003) 936. [7] J. Westin, C. Sneden, B. Gustafsson, J.J. Cowan, Astrophys. J. 530 (2000) 783. [8] J.J. Cowan, C. Sneden, S. Burles, I.I. Ivans, T.C. Beers, J.W. Truran, J.E. Lawler, F. Primas, G.M. Fuller, B. Pfeiffer, K.-L. Kratz, Astrophys. J. 572 (2002) 861. [9] V. Hill, B. Plez, R. Cayrel, T.C. Beers, B. Nordstr¨om, J. Andersen, M. Spite, F. Spite, B. Barbuy, P. Bonifacio, E. Depagne, P. Francois, F. Primas, Astron. Astrophys. 387 (5) (2002) 560. [10] C. Travaglio, R. Gallino, E. Arnone, J. Cowan, F. Jordan, C. Sneden, Astrophys. J. 601 (2004) 864. [11] C. Fr¨ohlich, G. Martinez-Pinedo, M. Liebend¨orfer, F.-K. Thielemann, E. Bravo, W.R. Hix, K. Langanke, N.T. Zinner, Phys. Rev. Lett. 96 (2006) 142502. [12] M. Arnould, S. Goriely, Phys. Rep. 384 (2003) 1. [13] E. Zinner, Ann. Rev. Earth Planet. Sci. 26 (1998) 147. [14] H. Nassar, M. Paul, I. Ahmad, D. Berkovits, M. Bettan, P. Collon, S. Dababneh, S. Ghelberg, J.P. Greene, A. Heger, M. Heil, D.J. Henderson, C.L. Jiang, F. K¨appeler, H. Koivisto, S. O’Brien, R.C. Pardo, N. Patronis, T. Pennington, R. Plag, K.E. Rehm, R. Reifarth, R. Scott, S. Sinha, X. Tang, R. Vondrasek, Phys. Rev. Lett. 94 (2005) 092504. [15] H. Beer, F. K¨appeler, Phys. Rev. C 21 (1980) 534. [16] Z.Y. Bao, H. Beer, F. K¨appeler, F. Voss, K. Wisshak, T. Rauscher, Atomic Data Nucl. Data Tables 76 (2000) 70.
182
M. Heil et al. / Progress in Particle and Nuclear Physics 59 (2007) 174–182
[17] M. Kenny, B. Allen, R. Macklin, Aust. J. Phys. 30 (1977) 605. [18] R. Macklin, Nucl. Sci. Eng. 88 (1984) 129. [19] A.de L. Musgrove, B. Allen, R. Macklin, Neutron Physics and Nuclear Data for Reactors and Other Applied Purposes, 426, OECD, Paris, 1978, p. 1. [20] C.M. Raiteri, M. Busso, G. Picchio, R. Gallino, Astrophys. J. 371 (1991) 665. [21] F. K¨appeler, M. Wiescher, U. Giesen, J. Goerres, I. Baraffe, M. El Eid, C.M. Raiteri, M. Busso, R. Gallino, M. Limongi, A. Chieffi, Astrophys. J. 437 (1994) 396.