Astron. Astrophys. 337, 714-720 (1998)

1. Introduction

The origin and evolution of light elements (, , , , and ) is an important chapter in the development of nuclear astrophysics. In the 70's and 80's the problem of Li, Be and B (hereafter LiBeB) nucleosynthesis has been considered essentially solved by Galactic Cosmic Ray (GCR) spallation (Meneguzzi, Audouze & Reeves 1971, Reeves 1994). The constituent nuclei of the GCRs, protons and particles, as well as C, N and O (hereafter CNO), form LiBeB via spallation on stationary nuclei in the interstellar medium (ISM). which has additional sources of production is an exception, as is since the / isotopic ratio is not correctly predicted by GCR nucleosynthesis. An artificial solution for the B isotopic ratio had been proposed, based on a non-observable low energy spike in the GCR energy spectrum, the so-called "carrot" (Meneguzzi & Reeves 1975).

New observations in the late 80's prompted a reassessment of the question as to the origin of LiBeB. Be abundance measurements in halo stars were achieved down to [Fe/H] = -1.5 (Rebolo et al. 1988, Ryan et al. 1990). As is generally the custom, square brackets will denote logarithmic abundance ratios by number relative to the solar value. A good fit of the Be evolution was obtained within the limited range of these observations (Vangioni-Flam et al. 1990) by considering the progressive CNO enrichment of the ISM due to stellar production and injection throughout the lifetime of the Galaxy, and supposing that the GCR flux is proportional to the SN rate (SN shocks serving only to accelerate particles out of matter of the same metallicity as that of the interstellar medium). At that time, these evolutionary effects on both GCR nucleosynthesis and the ISM, were sufficient to explain the behavior of Be vs. Fe. Subsequently, however, data were obtained at even lower metallicities for beryllium (Gilmore et al. 1992, Ryan et al. 1994, Boesgaard & King 1993) and a few boron abundance measurements were made over a wide metallicity range (Duncan et al. 1992, Edvardsson et al. 1994). These observations indicated a quasi linear relationship between both Be and B vs. Fe, instead of the quadratic relationship expected if the GCR were accelerated out of the ISM. This increased the general perplexity of potential solutions (Pagel 1991) and gave rise to a new wave of research (Duncan et al. 1992, Walker et al. 1993, Feltzing & Gustaffson 1994, Vangioni-Flam et al. 1994, Cassé et al. 1995, Fields et al. 1995, Bykov 1995, Tayler 1995, Ramaty et al.  1996, Vangioni-Flam & Cassé 1996). The primary origin of beryllium and boron (i.e. the fact that the production rate is independent of the ISM metallicity) indicates that these elements result from the spallation of fresh products of nucleosynthesis (primarily from C and O), rather than nuclei accumulated in the ISM. Thus, we are presented with the challenge to find an appropriate mechanism different from the traditional GCR picture which has become problematic for two reasons. First, as we noted above, if the cosmic rays are accelerated out of the ISM and interact in the ISM, the rising CNO/H abundance in the ISM leads to cumulative Be and B abundances which depend quadratically on the ISM metallicity, and thus is in disagreement with the observations. In addition, Be production in the early Galaxy by GCR accelerated out of the ISM requires the supply of extraordinarily large amounts of energy to the cosmic rays (Ramaty et al. 1997; Ramaty, Kozlovsky & Lingenfelter 1998).

The carrot of Meneguzzi & Reeves (1975) (introduced to explain ), having the same GCR composition, is also problematic because the B production by low energy cosmic rays should give rise to a quadratic relationship instead of a linear one exactly as in the high energy case. Moreover on theoretical grounds, a low energy spike throughout the galactic history would lead to an overproduction of Be and B (Lemoine et al. 1998). Finally, Li would be overproduced in the early galaxy by the reactions, spoiling the observed (primordial) Li plateau.

Accelerated particle reactions are not the only sources of boron since carbon spallation by neutrinos in core collapse supernovae (Types II and Ib, hereafter SNII) can also contribute significantly to production (Woosley et al. 1990; Olive et al. 1994, Woosley &Weaver 1995, Vangioni-Flam et al.  1996; Ramaty et al. 1997). This mechanism is particularly interesting because it yields mainly ¹¹B, making -induced spallation important for the explanation of the meteoritic B isotopic ratio, ¹¹B/¹⁰B = 4.050.2 (Chaussidon & Robert 1995). Vangioni-Flam et al. (1996) found that the neutrino contribution to the total B production should amount to at most 30%. If the cosmic ray spectrum extends to high energies, production by neutrino spallation is always required since such cosmic rays are not capable of reproducing the meteoritic ratio, but again a 30% contribution to the total B production from neutrinos appears sufficient (Ramaty et al. 1997).

As the spallation of C, N and O is the only significant source of Be, the linear dependence of [Be/H] with respect to [Fe/H] (at least up to [Fe/H] = -1), or equivalently the approximate constancy of [Be/Fe], implies that a mechanism whereby C and O are accelerated above the spallation thresholds and impinge on the ambient H and He is operative. SNII's are the most plausible sources of accelerated C and O in the early Galaxy. Accelerated N, however, makes only a very minor contribution since it is highly underproduced in SNII's. Among the different scenarios proposed to explain the linear evolution of Be and B, we consider the following two, which shall subsequently be referred to as models (a) and (b).

In model (a) Be and B are produced both by low energy nuclei (hereafter LEN), highly enriched in C and O relative to H and He, and standard GCR accelerated out of the ISM (Cassé, Lehoucq & Vangioni-Flam 1995, Meyer et al., 1997, Ellison et al., 1997). The latter is only dominant at late times in the evolution of the Galaxy. This model was motivated by the observations of a linear dependence of Be and B on Fe which implies a primary source for their production and the observations of C and O deexcitation gamma ray line emission from Orion (Bloemen et al. 1994; 1997). It was suggested (Bykov 1995, Parizot et al. 1997) that the required population of C and O enriched LEN could result from the acceleration, by an ensemble of weak shocks in superbubbles, wherein the seed particles for acceleration originate from the winds of massive stars and the ejecta of supernovae from massive star progenitors. Only the most massive stars (M60), that is those which explode within superbubbles due to their very short lifetime, should be involved in this scenario. In the early Galaxy, these extended acceleration sites would be sustained essentially by SNII exploding in OB associations (Samland 1998). Later on, in the disk phase, WR stars would also participate, since the stellar winds intensify at increasing metallicities (Meynet et al. 1994). The scenario further assumes that the metallicity of the LEN component is independent of the average Galactic metallicity, thereby dominating the Be production in the halo phase ([Fe/H] -1 with the GCRs taking over in the disk phase (Vangioni-Flam et al. 1996, 1997).

In model (b) Be and B are produced by standard GCRs accelerated at all epochs of Galactic evolution from the ejecta of supernovae (Ramaty et al. 1997; 1998). This model, motivated by the observed, essentially constant [Be/Fe] in the early Galaxy, envisions the acceleration of the erosion products of high velocity refractory grains formed in a supernova ejecta (Lingenfelter, Ramaty & Kozlovsky 1998). These authors have shown that sufficient O is incorporated in refractory Al₂O₃, MgSiO₃, Fe₃O₄ and CaO to account for the GCR source O abundance. They have further argued that the GCR source C abundance could also be understood if the fraction of C ejecta incorporated in refractory grains (mainly graphite) is the same as that of the other main refractories, and they have shown that the standard arguments against the acceleration of the refractory metals out of supernova ejecta are model dependent and answerable in principle. It is thus possible that at all epochs of Galactic evolution the standard GCR would contain sufficient C and O to explain the linear Be evolution. In this scenario, individual SNII with progenitors of the same mass range as that responsible for Fe production (M8) participate in the production of Be and B. Note that the evolution of Be/Fe would be somewhat similar to that of O/Fe since Be, O and Fe are coproduced by SNII at least up to [Fe/H] = -1. If all SNII from 8 to 100 participate in oxygen production, the evolution of O/Fe will be similar to that of Be/Fe in model (b).

If the Be in the early Galaxy is indeed produced by particles whose acceleration is related to short-lived very massive stars, then the difference in the lifetimes of the progenitors of Be and Fe and the relative number of stars implied in each case, could also affect the evolution of Be/Fe. In the present paper we shall critically examine the evolution of B and Be in the early Galaxy, taking into account: i) the relative Be yields associated to each mass domain considered and ii) potential time dependent effects due to the mass dependence of the lifetimes of the stellar progenitors of the core collapse supernovae responsible for the production of B and Be. In the following, we reproduce the observed Be evolution through a Galactic evolutionary model and explore the correlated behavior of B considering three plausible B/Be production ratios, in agreement with the results of the nuclear spallation models of various compositions and energy spectra (Vangioni-Flam et al. 1996, Ramaty et al. 1997). The wide range of B/Be ratios explored leaves room for neutrino spallation.

In what follows, we will examine whether or not it is possible, using the existing data on B and Be, to distinguish between models a) and b). In Sect. 2, we will describe the current status of the B and Be data. In Sect. 3, we will describe and develop the proposed test to distinguish between the models and present the results of our calculations. Our conclusions are found in Sect. 4.

© European Southern Observatory (ESO) 1998

Online publication: August 27, 1998
helpdesk.link@springer.de