Incorporation of Short-Lived 10Be in a Calcium-Aluminum-Rich Inclusion from the Allende Meteorite

See allHide authors and affiliations

Science  25 Aug 2000:
Vol. 289, Issue 5483, pp. 1334-1337
DOI: 10.1126/science.289.5483.1334


Enrichments in boron-10/boron-11 in a calcium-aluminum–rich inclusion from the Allende carbonaceous chondrite are correlated with beryllium/boron in a manner indicative of the in situ decay of short-lived beryllium-10. Because this radionuclide is produced only by nuclear spallation reactions, its existence in early solar system materials attests to intense irradiation processes in the solar nebula. The particle fluence inferred from the initial beryllium-10/beryllium-9 is sufficient to produce other short-lived nuclides, calcium-41 and manganese-53, found in meteorites, but the high canonical abundance of aluminum-26 may still require seeding of the solar system by radioactive stellar debris.

When the oldest solar system materials, the Ca-Al–rich inclusions (CAIs) of primitive chondrite meteorites (1), first crystallized in the solar nebula, their major elements included small concentrations of short-lived radioactive isotopes. Evidence for the in situ decay of now extinct26Al [half-life (τ1/2) = 0.7 million years (My)] and 41Ca (τ1/2 = 0.1 My) in CAI minerals is provided by excesses of their respective daughter products, which are correlated with the relative abundance of a stable isotope of the parent element. The origins of these and other (2) relatively short-lived radionuclides found in meteorites are controversial: These isotopes can be produced by nucleosynthesis in different stellar environments (3), or they may be the products of nuclear reactions induced by the collisions of energetic particles with ambient dust or gas (4, 5). The astrophysical implications of these contrasting interpretations of extinct radioactivities are different regarding the processes and time scales inferred for the formation of the solar system (6). Here, we report evidence for the existence of an extinct radionuclide, 10Be (τ1/2 = 1.5 My), in CAIs that indicates that primitive meteorites preserve a record of intense irradiation processes that may have occurred in the early solar system.

Unlike other elements, Be, together with Li and B, is thought to be produced predominantly by spallation reactions between galactic cosmic rays (GCRs) and the interstellar medium (7). The inefficiency of this production, coupled with the destruction of these elements in stellar interiors by nuclear burning, accounts for the depleted cosmic abundances of Li-Be-B relative to neighboring low-mass elements (8). However, the discrepancies between the 7Li/6Li and 11B/10B ratios of the solar system (11B/10B = 4.0 ± 0.1 and 7Li/6Li = 12.1 ± 0.1) and those inferred for interstellar production by GCR (11B/10B ≈ 2.5 and7Li/6Li ≈1.5) require additional nucleosynthetic sources for these elements. Spallation reactions occurring at relatively low energy (< 100 MeV per nucleon) or several stellar sources producing pure 7Li and 11B can theoretically provide the additional Li and B necessary to match their respective chondritic isotopic compositions (9).

The observed B isotopic variations in chondrules (10) have been interpreted as the result of such low-energy irradiation processes occurring in the molecular cloud that gave birth to the solar system (11). Low-energy irradiation has also been invoked to explain the presence of live 26Al in CAIs. Such an irradiation could have taken place either during the quiescent phase of the protosolar nebula [e.g., (4)] or during the early phases of the sun's formation when energetic particles from magnetic reconnection flares irradiate materials at the inner edge of the protosolar accretion disk (5, 12). In addition to the stable isotopes of B and Li, the nuclear cross sections (13) are such that appreciable amounts of the radioactive isotope 10Be are also expected to be produced during any such low-energy irradiation processes. We used an ion microprobe to determine Li and B isotopic compositions and Be concentrations in CAIs to search for evidence of irradiation of early solar system materials.

Three coarse-grained type B CAIs (TS-34, 3529-30, and 3529-41) from the Allende carbonaceous chondrite were investigated (14). Previous ion microprobe measurements (15, 16) demonstrate that all three inclusions initially formed with close to canonical26Al/27Al ratios, thus indicating their early crystallization in the solar nebula. However, detailed petrographic and isotopic studies of the latter two inclusions (15) showed that some redistribution of radiogenic 26Mg occurred, probably during a secondary thermal event that resulted in alteration and recrystallization of melilite and/or anorthite. The degree of isotopic disturbance and of mineralogic alteration appears somewhat greater in 3529-30 than in 3529-41. The7Li/6Li, 11B/10B, and9Be/11B compositions as well as the atomic concentrations of Li, Be, and B were determined in multiple spots of each sample (Table 1) with a CAMECA ims 1270 ion microprobe (17).

Table 1

Li and B isotopic compositions and Li-Be-B concentrations in Allende CAIs.

View this table:

Concentration ranges for Li [3 to 750 parts per billion (ppb)] and Be (50 to 3000 ppb) are comparable to, but more variable than, previous determinations in CAIs (18). Boron concentration is even more variable among the three CAIs. In two samples, B is depleted (6 to 150 ppb, with one spot at ∼900 ppb) relative to CI concentrations, as expected on the basis of its volatility (19), but in 3529-30, B is abundant [1 to 10 parts per million (ppm)] in all spots analyzed. Beryllium is found to be preferentially concentrated in melilite in accord with previous ion probe analyses and partitioning experiments (20, 21), resulting in high Be/B ratios in melilite from 3529-41.

Among the three CAIs, Li isotope variability is limited to <50 per mil (‰), and all analyses overlap the range previously observed in chondrules (22). In contrast, B shows large δ11B anomalies ranging to –410‰ in sample 3529-41, greatly exceeding B isotope variations reported in meteorites (22, 23). The large negative δ11B values cannot be due to Rayleigh fractionation resulting from volatilization of B during CAI formation because this would result in the preferential loss of 10B, opposite to the observed10B excesses. The magnitude of the anomaly also cannot be the result of an in situ spallation by GCR in the Allende meteoroid that would produce correlated 6Li and 10B excesses that are not observed.

The 10B excesses in 3529-41 are correlated with Be/B (Fig. 1) in a manner indicating that short-lived 10Be was incorporated into this CAI during crystallization and was preferentially partitioned into melilite where it subsequently decayed, producing radiogenic 10B. The slope of the correlation line implies an initial10Be/9Be ratio of (9.5 ± 1.9) × 10−4, and the intercept yields an initial11B/10B ratio = 3.93 (δ11B = −28 ± 9‰), which is close to the lowest value found in chondrules (22) and lower than the ratio of 4.03 (δ11B = −3‰) for CI chondrites (23). The other CAIs, which do not exhibit strong Be/B fractionation (24), lack significant 10B enrichments and have δ11B values in agreement with the inferred initial 11B/10B ratio of 3529-41.

Figure 1

Boron isotopic composition of individual minerals from Allende CAIs as a function of Be/B ratio in the same material. Error bars are 2σ. The10B/11B values from various spots of CAI 3529-41 show 10B excesses that are correlated with the Be/B ratio in a manner indicative of the in situ decay of10Be in the CAI. The slope of the correlation line (weighted fit to the data of 3529-41 only) corresponds to an initial 10Be/9Be = 0.00095 ± 0.00019 (2σ) at the time of crystallization. The intercept indicates10B/11B = 0.254 ± 0.002, which is higher than 10B/11B for CI chondrites (23), as shown by the horizontal line. The inset shows the same data at an expanded scale; data for CAIs 3529-30 and TS-34 are consistent with the Be-B isotope systematics of 3529-41.

Because 10Be is not produced by stellar nucleosynthesis, the existence of this radioisotope in a CAI demonstrates that some early solar system materials preserve a record of irradiation processes. That the excess 10B follows the crystal-chemical behavior of Be indicates that the radioactive 10Be existed in the CAI melt, similar to the situation for 26Al and41Ca. The presence of live 10Be raises three questions related to the astrophysical implications of how CAI precursors obtained their short-lived radioactivity: What are the intensity and energy spectrum of the incident particles, where did this irradiation take place, and what are the collateral consequences of such an irradiation, e.g., with respect to other short-lived radionuclides present in the early solar system. A rigorous calculation of the fluence (Φt) necessary to yield10Be/9Be = 9.5 × 10−4in CAI precursors requires detailed knowledge of the target composition and the energy spectrum of incident particles; however, a first-order estimate may be obtained with cross sections (13) for the nuclear reaction 16O(p,x)10Be and assuming a power law (E−3) distribution in proton energy. The proton fluence (> 20MeV) is in the range 1018 to 1019 cm−2 for irradiation of targets having CAI compositions. Irradiation of solar composition material results in fluence estimates about one order of magnitude lower. All these estimates are below the upper limits set by previous Li isotope measurements of an Allende CAI (18) and slightly lower than the fluence inferred from the B isotopic variations observed in chondrules (22).

With respect to the astrophysical environment, three possible scenarios for this irradiation can be considered: (i) The 10Be may have been inherited from the steady-state galactic reservoir produced by GCR, (ii) the 10Be may have been produced just before collapse of the solar nebula from the interaction of energetic ejecta from a supernova explosion with the surrounding interstellar medium, or (iii) the 10Be could have been locally synthesized by irradiation within the protosolar nebula by intense stellar flares, similar to those inferred from x-ray observations (25) of young stellar objects (YSOs).

An upper limit on the 10Be/9Be ratio of interstellar matter in a molecular cloud core that collapsed to form the solar nebula can be constructed from the 9Be/H production rate by GCR 4.5 billion years ago (10−21 per year) derived from stellar observations (26), the production of 10Be relative to 9Be (∼0.1), and the rate of 10Be radioactive decay. For an initial solar 9Be/H (2.6 × 10−11), a steady state is reached after ∼7 My with10Be/9Be = 8.3 × 10−6, which implies that only ∼1% of the 10Be found in Allende 3529-41 should be from the galactic background. Lower energy cosmic rays (< 100 MeV per nucleon) are stopped within a molecular cloud core, in principle providing a mechanism for locally enhancing the proportion of young GCR-Be relative to old interstellar Be in the presolar molecular cloud compared with the average galactic mixing ratio discussed above. Because such GCRs have10Be/9Be = 0.1 (27), an enhancement by a factor of 100 of this mixing ratio would be necessary for local heterogeneities to develop with a10Be/9Be ratio of ∼0.001. However, mixing ratios of this order appear to be ruled out (28) to avoid overproducing other short-lived nuclides (e.g., 53Mn). A supernova explosion can provide another potential source of freshly synthesized 10Be from either spallation reactions occurring between C and O accelerated nuclei and ambient interstellar gas (9, 29) or nuclear reactions in shock waves (30). Although once promising, the first scenario now seems unlikely because the measured flux of γ-rays from nuclear deexcitation lines of C and O are low (31), and the second process yields calculated 11B/10B ≈ 135 (30), in disagreement with the CAI ratio.

It is thus unlikely that the CAI 3529-41 could have inherited much of its complement of radioactive Be from the presolar cloud, but is it plausible that the 10Be could have been produced locally within the solar nebula. Possible analogs for the early solar system are provided by YSOs, forming low-mass stars that produce copious x-rays with spectral characteristics similar to solar flares (25). The ubiquitous x-ray signature of YSOs indicates that it is possible that a fraction of the material in the solar nebula went through a period of intense irradiation in the vicinity of the sun (32). It has been proposed that CAIs formed at the inner edge of the solar accretion disk where they are heavily irradiated before being launched to greater heliocentric distance where they may accrete into chondrite parent bodies (5, 12). In YSOs, the flux of energetic protons accompanying flares should be at least 3 × 105 times the present-day solar flux, i.e., Φ = 2 × 109 H cm−2s−1 at 0.1 astronomical unit (33), and impulsive flares may provide an even higher flux (5). The corresponding irradiation time that yields the 10Be content in the CAI can be calculated to be on the order of decades to 100 years. This duration is also sufficient to produce other short-lived nuclides, 41Ca and 53Mn, at the abundance levels found in meteorites (34), but the canonical CAI abundance of 26Al cannot be reached unless other energetic particles (e.g., 3He) and special target shielding configurations are considered (5). Alternatively, it may be possible to obtain the correct relative levels of41Ca, 26Al, and 10Be in an irradiated Ca-Al–depleted silicate dust, such as refractory olivine, which is evaporated after, or during, the irradiation. A mechanism for quantitative distillation of radioisotopes into CAI melts needs to be elaborated before it can be concluded that such a scenario can explain all the important short-lived nuclides in CAIs by local production. Finally, we note that in a local irradiation scenario, the low initial δ11B accompanying live 10Be in an early formed CAI can be interpreted as reflecting the progressive production in the solar nebula of 11B-rich B by low-energy irradiation. If this is correct, higher initial δ11B values in CAIs should be associated with higher10Be/9Be ratios.

  • * To whom correspondence should be addressed. E-mail: kdm{at}


View Abstract

Stay Connected to Science

Navigate This Article