The class of light elements, namely Li, Be and B, sets itself apart from any other by its interstellar origin (except for part of the 7Li, produced in the Big Bang ages, and perhaps part of the 11B, produced in supernova (SN) explosions by neutrino-spallation). Concentrating on the most representative isotope, the abundance of 9Be in stars of increasing metallicity can be regarded as the witness and tracer of the nuclear spallation efficiency during Galactic chemical evolution. Indeed, virtually every atom of Be observed in the atmosphere of stars must have been produced by the spallation of a larger nucleus, most probably C or O, induced by the interaction of energetic particles (EPs) with the interstellar medium (ISM).
Since the first measurement of Be in a very metal-poor star at the beginning of the decade (Gilmore et al. 1991), increasing evidence has been gathered showing that the abundance of Be and B in the early Galaxy (until the ambient metallicity is 10% that of the sun, say) kept increasing jointly and linearly with ordinary metallicity tracers, such as Fe or O, as if they were actually primary elements (Duncan et al. 1992, 1997; Edvardsson et al. 1994; Gilmore et al. 1992; Kiselman & Carlsson 1996; Molaro et al. 1997; Ryan et al. 1994). Now they are not, since as we just recalled C and O nuclei have to be produced first in order that they can be spalled by EPs into light elements. The observations therefore suggest that some process must act to ensure that, on average, an equal amount of Be is synthesized each time a given mass of Fe or O is ejected into the ISM. It should be clear, however, that this statement relies on the assumption that the abundances of O and Fe are proportional to one another, at least during the early evolution stages in which we are interested here.
This assumption has long been used with high confidence level based on both theoretical and observational arguments, but new observations seem to contradict it dramatically (Israelian et al. 1998; Boesgaard et al. 1998). Although an independent confirmation of these observations would be welcome, they have recently been used to reappraise the alleged `primary behavior' of 6LiBeB Galactic evolution (Fields & Olive, 1999). Indeed, if the O/Fe abundance ratio is not constant but actually decreases with metallicity, then the observed approximate constancy of the Be/Fe ratio implies an increasing Be/O ratio. Fields & Olive (1999) find a Be-O logarithmic slope in the range 1.3-1.8, which seems to contradict both the primary scenario (slope 1) and the secondary scenario (slope 2), in which the spallation reactions producing the light elements are induced by standard Galactic cosmic rays (GCRs) accelerated out of the ISM. However, the current lack of Be and O abundance measurements in the same very metal-poor stars (with [O/H] , say) makes the data marginally compatible, within error bars, with both scenarii.
While the situation should be soon clarified, notably by the accumulation of data at lower metallicity and independent measurements of Be, B, O and Fe in the same set of halo stars, we (Parizot & Drury, 1999; Paper I) choose to investigate the Be production in the ISM from the other direction, i.e calculate the Be yield associated with the explosion of an isolated supernova (SN) in the ISM, according to current knowledge about supernova remnant (SNR) evolution and standard shock acceleration, and compare this Be yield with the value required to explain the observed Be/Fe ratio in metal-poor stars. We identified two different mechanisms leading naturally to a primary evolution of Be in the early Galaxy. In the first mechanism, particles from the ambient ISM (i.e. metal-poor) are accelerated at the forward shock of the SN and confined within the SNR until the end of the Sedov-like evolution phase. There, they interact with the freshly synthesized C and O nuclei, and therefore produce Be by spallation at a much higher rate than in the (secondary) GCR nucleosynthesis scenario in which they merely interact with the ambient, metal-poor ISM. In the second mechanism, particles from the enriched SN ejecta are accelerated at the reverse shock and again confined within the SNR during Sedov-like phase, where they suffer adiabatic losses through which they lose between 30% and 70% of their initial energy, depending on the ambient density. After the end of the Sedov-like phase, these particles diffuse out in the ISM where the energetic C and O nuclei can be spalled by the H and He atoms at rest in the Galaxy.
We have shown in Paper I, through approximate analytical calculations, that the total Be yield obtained by processes 1 and 2 depends on the ambient density, and that this third mechanism is actually the most efficient (for light element production) in most cases, though not efficient enough to account for the observed Be/Fe ratio of . If each SN ejects on average 0.1 of Fe in the ISM, then the average Be yield per SN must be atoms (cf. Ramaty et al. 1997), which exceeds even our most optimistic calculated yields by about one order of magnitude. We concluded that another mechanism or source of energy should be invoked, and argued that a model based on superbubble acceleration (involving the collective effect of SNe rather than individual SN shock acceleration) is a quite natural and promising candidate. In this paper, we confirm the results of Paper I by performing time-dependent numerical calculations, and discuss in more details their implications for Galactic chemical evolution scenarii. The reader is referred to Paper I for a more detailed description of the mechanisms considered here, and a discussion of their motivation and theoretical justification.
© European Southern Observatory (ESO) 1999
Online publication: May 21, 1999