Main

The favoured explanation for the origin of 16O enrichments in calcium-aluminium-rich inclusions (CAIs) in meteorites has been that 16O was carried into the solar nebula by refractory oxide grains (for example, aluminium oxide, Al2O3) from supernovae5. Although our discovery of T84 indicates that micrometre-sized refractory oxide grains do indeed form in supernova ejecta, the abundance of such grains in meteorites is probably too low to account for the 16O excesses observed in CAIs.

We compare the oxygen-isotope ratios of T84 with those of 102 other presolar oxide grains from meteorites3, together with the ranges of these ratios in red giants6 and in the interstellar medium7 (Fig. 1). Like most of the grains, T84 was identified by isotopic imaging in an ion microprobe3. Its isotopic composition is: 17O/16O = 3.0 ± 1.8 × 10−5,18O/16O = 3.4 ± 1.3 × 10−4.

Figure 1: Oxygen-isotopic ratios of meteoritic grains.
figure 1

Filled circles, 103 presolar oxide grains from meteorites; light grey, range observed in red giants; dark grey, range in the gaseous interstellar medium (ISM). Grain T84 is unusually rich in 16O, and probably formed in the ejecta of a type II supernova.

Full size image

Unfortunately, T84 was completely consumed during measurement because of its small size (about 0.5 μm) and its mineralogy cannot now be determined. We estimate that the abundance of grains such as T84 in Tieschitz is less than 0.25 parts per billion.

The composition of T84 indicates that it formed in a stellar environment very rich in 16O. Massive stars synthesize copious amounts of 16O by burning helium, carbon and neon deep in their interiors4. The freshly made 16O is expelled into the interstellar medium either by type II supernova explosions or, in stars whose masses are more than 30 times that of the Sun,by strong winds during a so-called WO phase8. Type Ia supernovae also eject material rich in 16O (ref. 9).

Either supernovae or WO stars are potential sources of T84, but WO stars are extremely rare in the Galaxy whereas supernova explosions are relatively common. Furthermore, type II supernovae eject far more 16O than do type Ia, so a type II origin for T84 is more likely.

The discovery of T84 shows that oxide grains from supernovae were present in the early Solar System, but the abundance of such grains in meteorites is surprisingly small. Galactic dust production rates, calculated supernova yields, and the observed fraction (1%) of presolar SiC grains from supernovae suggest that Al2O3grains produced by supernovae should make up 10-70% of presolar oxides3,10 rather than the amount observed, which is less than 1%.

The dust produced by supernova 1987A (ref. 11) is much smaller than the grains described here and T84 is at the small end of what can be currently analysed in the ion microprobe. Supernovae may thus produce much smaller oxide grains than do red giants, and they could have been systematically excluded from the presolar oxide data set. Furthermore, detailed analysis of SN1987A's spectra indicates that the dust formed there was not oxygen-rich, but was possibly metallic iron11. Perhaps oxides do not easily condense from supernova ejecta, despite the extremely oxidizing conditions.

The very low abundance of presolar 16O-rich oxide grains seems to rule out supernova Al2O3as the prime carrier of 16O enrichments in CAIs. Instead, 16O may have been carried into the early Solar System by supernovae-derived 16O-rich silicates rather than oxides. Alternatively, the Solar System's oxygen-isotope heterogeneity may have resulted largely from non-mass-dependent isotopic fractionation12.

L. R. Nittler, C. M. O'D. Alexander, J. Wang Department of Terrestrial Magnetism, Carnegie Institution of Washington, 5241 Broad Branch Road NW, Washington, DC 20015, USA

X. Gao McDonnell Center for the Space Sciences, Washington University, 1 Brookings Drive, St Louis, Missouri 63130, USA