16O Excesses in Olivine Inclusions in Yamato-86009 and Murchison Chondrites and Their Relation to CAIs

See allHide authors and affiliations

Science  05 Feb 1999:
Vol. 283, Issue 5403, pp. 828-831
DOI: 10.1126/science.283.5403.828


In situ ion microprobe analyses of oxygen isotopes in Yamato-86009 and Murchison chondrites show that they contain abundant olivine-rich inclusions that have large oxygen-16 (16O) excesses, similar to those in spinel grains in calcium-aluminium–rich inclusions in Allende and other carbonaceous chondrites. The existence of16O-enriched olivine-rich inclusions suggests that oxygen isotopic anomalies were more extensive in the early solar system than was previously thought and that their origin may be attributed to a nebular chemical process rather than to an unidentified16O-rich carrier of presolar origin.

Calcium-aluminium–rich inclusions (CAIs), most commonly observed in carbonaceous chondrite groups (CV, CO, and CM chondrites), have anomalous oxygen isotopic compositions, with δ17O and δ18O values down to ∼−40 per mil (‰) (1–3) relative to the standard mean ocean water (SMOW) δ17O and δ18O values (4). The origin of these isotopic anomalies, however, is still under debate—whether they are derived from a carrier having extreme oxygen isotopic compositions (almost pure 16O) (1, 5) or from mass-independent fractionation caused by some kind of chemical process (6–9). It has generally been considered that such oxygen isotopic anomalies are characteristic of highly refractory components (CAIs) and not of less refractory Mg-rich ones, although some olivine phases rarely observed inside CAIs are found to have relatively high oxygen isotopic anomalies (10). In the course of our in situ ion microprobe study of oxygen isotopes, however, we discovered abundant olivine-rich inclusions (OIs) in two chondrites that have excess 16O, with δ17O and δ18O values down to ∼−50‰, which is similar to the highest anomalies observed in Allende and Murchison CAIs (2, 3). We report here details of our findings and discuss the origin of the oxygen isotopic anomalies on the basis of our results.

Recent improvements in ion microprobe techniques have enabled in situ analyses of oxygen isotopes with a spatial resolution of ∼10 μm and a precision of ∼1‰ (11). These technical improvements allow us to study the micro-distribution of oxygen isotopes in different components in meteorites, which is essential to understanding the origin of the oxygen isotopic anomalies. We performed in situ oxygen isotope analyses of Yamato-86009 (CV3) and Murchison (CM2) chondrites using a CAMECA ims-6f ion microprobe. A Cs+primary beam ∼15 μm in diameter, with a beam intensity of 0.1 to 0.3 nA and an impact energy of 19.5 kV, was used for the analyses (except for position 19; see Table 1). Negative ions of oxygen isotopes were accelerated at −9.5 kV, separated at a mass resolution of ∼5000, and detected with a Faraday Cup (16O) or an electron-multiplier–based ion counting system (17O,16O1H, and18O). The contribution of the16O1H peak was negligible (≪1‰) at the center of the 17O peak for all the analyses. A normal incidence electron gun was used for charge compensation. Repeated analyses were performed on a San Carlos olivine standard (12) before and after the periods of sample analyses without changing the analytical conditions, and all the data were normalized to the average of the San Carlos olivine data. The reproducibility of the analysis was ∼±1.5‰ for17O/16O and ∼±2.2‰ for18O/16O (1σ) for all analyses.

Table 1

Analyzed phases and their oxygen isotopic compositions. Ag, spinel-pyroxene-feldspar aggregate; Ch, chondrule.

View this table:

The meteorite samples were prepared in thin sections, carbon-coated, examined with a scanning electron microscope (SEM) equipped with an energy-dispersive x-ray spectrometer (EDS) for petrographic studies, and used for the ion microprobe analysis. We selected four OIs from Yamato-86009 (Y86009-A, -B, -C, and -D) and four OIs from Murchison (MC5a-26, -47, -52, and -94) for the ion microprobe analysis. The OIs are composed predominantly of Mg-rich olivine (Table 1) and of variable amounts (∼1 to ∼50% by volume) of interspersed Ca-Al–rich domains (mostly diopsidic or fassaitic pyroxene with or without spinel) enclosed by the olivine. The OIs are irregular but lumpy in shape, and many of them have numerous irregularly shaped voids (typically submicrometer sized to a few micrometers in diameter) both in olivine and in Ca-Al–rich phases (Fig. 1) (13). The irregular shape of the OIs and the voids suggests that OIs never completely melted but formed by the sintering of small forsterite grains at subsolidus temperatures. The typical size of OIs is 50 to 200 μm, but some exceed 300 μm and they occupy more than 1% by volume of these meteorites (14).

Figure 1

Back-scattered electron images of two OIs: (A) Y86009-A and (B) MC5a-26. Fo, forsterite; Sp, spinel; Px, pyroxene (fassaitic to diopsidic); Di, diopside; Fe(Ni), metallic iron containing some Ni; FeS, troilite. Calcium-aluminium–rich domains (Px + Sp) are enclosed by forsterite in Y86009-A. Many irregularly shaped voids (typically less than a few micrometers) and a large void in the center are visible in MC5a-26. This OI contains only a small amount (∼1 volume %) of diopside grains less than a few micrometers in size, interspersed in forsterite.

For comparison, one CAI (Y86009-E) and one chondrule (Y86009-F) in Yamato-86009 and one chondrule (MC5a-85) in Murchison were also analyzed. Y86009-E is a spinel-pyroxene-feldspar aggregate (15). Because of its fine-grained texture (typical grain size <20 μm), individual phases could not be analyzed without some overlap to other phases, except for one spinel grain (position 12). Y86009-F (∼900 μm in diameter) consists of enstatite, forsterite (rounded grains of 10 to 50 μm in size, poikilitically enclosed by enstatite), diopside (euhedral crystals rimming enstatite and protruding into the glass), and glass. MC5a-85 (∼100 μm by ∼170 μm in size) consists predominantly of enstatite and forsterite with minor (less than a few percent) amounts of diopside and ferrosilite.

All 13 analyses on olivine grains inside eight OIs showed16O enrichments with δ17O and δ18O ranging from ∼−42 to ∼−51‰. (One datum for position 23 is omitted from the discussion because of its large errors, caused by the instability of the secondary ion beam.) Ca-Al–rich phases in Yamato-86009 OIs showed relatively smaller 16O enrichments with δ17O and δ18O ranging from ∼−32 to ∼−42‰. Spinel and spinel-rich phases of the CAI (Y86009-E) exhibited 16O enrichments with δ17O and δ18O ranging from ∼−41 to ∼−49‰, which are consistent with the previous results for spinel in Allende and Murchison (2, 3). The two chondrules (Y86009-F and MC5a-85) exhibited much smaller oxygen isotopic anomalies, with δ17O and δ18O ranging from ∼−13 to ∼−4‰ (16).

It is known that physical and chemical processes such as evaporation, condensation, diffusion, and chemical reactions in general, accompany mass-dependent isotopic fractionations. Oxygen is highly mobile in many circumstances in nature and is expected to show large isotopic fractionations. If isotopic compositions of oxygen in both the start and end materials generated in those processes are expressed in the so-called three-isotope diagram (Fig. 2), they should form a line, with slope ∼1/2, according to the mass differences for the 17O-16O pair (one mass unit) and the 18O-16O pair (two mass units). An example is the terrestrial fractionation (TF) line for terrestrial materials. Mineral separates from CAIs in Allende and other carbonaceous chondrites, on the other hand, form a correlation line with slope ∼1 called carbonaceous chondrite anhydrous mineral (CCAM) line. They exhibit up to 4 to 5% (or 40 to 50‰) of16O enrichments relative to SMOW, which is much different from the bulk isotopic compositions of most of the solar system materials (meteorites, Earth, the moon, and Mars; see the insert inFig. 2). The CCAM line (or any correlation line with a slope different from 1/2) is inexplicable by the processes that accompany a mass-dependent isotope fractionation. The correlation line with slope ∼1 may require mixing of an isotopically distinct16O-rich material that has an isotopic composition along the lower left extension of the line, to an isotopically “normal” oxygen reservoir with an oxygen isotopic composition located somewhere on the upper right extension. Alternatively, an extraordinary chemical reaction may have enacted mass-independent fractionation of oxygen isotopes in the solar nebula (6–9). The fact that OIs, like spinel in CAIs, plot in the lower end of the CCAM line suggests a common oxygen source or a similar chemical process for both OIs and CAIs.

Figure 2

Oxygen isotopic compositions of various phases in Yamato-86009 and Murchison chondrites. Also plotted for comparison (insert) are bulk oxygen isotopic compositions of different classes of meteorites and Earth and moon rocks (26). H, L, and LL, ordinary chondrites; CI, CM, CO, and CV, carbonaceous chondrites; HED, a class of differentiated meteorites (howardite, euclite, and diogenite); SNC, martian meteorites; Ure, ureilite; EL and EH, enstatite chondrites; Aub, aubrite; TF line, terrestrial fractionation line (which has a slope of ∼0.5); CCAM line, carbonaceous chondrite anhydrous mineral line, defined by many analyses of Allende CAI minerals (2,3). All the Murchison and Yamato-86009 data are normalized to the average of San Carlos (SC) olivine data. Individual data for SC olivine are also shown after normalization to their average, demonstrating the reproducibility of the present ion microprobe analyses. Error bars are not shown except for position 23 for clarity. The definition of δmO is given in (4).

The discovery that isotopically anomalous OIs are abundantly present in the two chondrite thin sections (14) demonstrates that the oxygen isotopic anomalies with δ17O and δ18O ranging from ∼−40 to ∼−50‰ are characteristic not only of Ca-Al–rich components but also of the most common meteoritic and planetary silicate (olivine), implying that such oxygen isotopic anomalies were rather extensive in the early solar system. In terms of the mineral stability relationships applied to the nebulae (17), forsterite has a high condensation temperature among other condensable phases, except for the CAI minerals such as spinel and gehlenite, which have higher condensation temperatures than that of forsterite. The less refractory nature of OIs as compared to CAIs suggests that the largest oxygen isotopic anomalies (∼−50‰) are not limited to the highest temperature condensates. Yet its refractory nature compared to the rest of the condensable phases reinforces a previously held view that the oxygen isotopic anomalies are somehow linked to high temperatures.

A presolar component model, in which a carrier (seed crystals) of an 16O-rich oxygen (almost pure 16O) is assumed, has been one of the preferred models for the origin of the oxygen isotope anomalies in CAIs. In this model, the16O-enriched refractory oxide grains of presolar origin provided nucleation sites for CAI-forming minerals to condense from a hot nebular gas (1, 18). However, such a model cannot explain the present observations. After the condensation of CAIs, the seed crystals would disappear before forsterite condensed. Even if they survived, forsterite would have to incorporate them in exactly the same dilution as for the CAIs to explain the unique δ17O and δ18O values (namely, −40 to −50‰). In fact, presolar oxides are rare (using current recovery procedures), and most do not show 16O excesses (19) except for one oxide grain (∼0.5 μm in diameter) discovered recently (20) in the Tieschitz ordinary chondrite. It has an extreme 16O enrichment, but the inferred abundance of such grains [<0.25 part per billion (ppb)] in the meteorite is too low for them to be plausible carriers of 16O enrichments in CAIs.

Further evidence against the presolar component model comes from recent discoveries of large oxygen isotopic anomalies (δ17O and δ18O down to ∼−45‰) in diopside (21) in the Wark-Lovering rims (22) and in olivine (23) in accretionary rims (19) of CAIs. Because the Wark-Lovering rims and accretionary rims are considered to be features added later to main CAI bodies (22,24), the observations that the cores and rims of CAIs have similar oxygen isotopic anomalies are difficult to reconcile with the idea of16O-rich, refractory seed crystals of presolar origin.

From another angle, the longtime effort by many researchers to characterize the isotopic signatures of CAIs, chondrules, and bulk chondrites has revealed that all the other elements do not show major mass-independent shifts from their canonical isotopic compositions when oxygen gives ∼−40‰ variation from the terrestrial ensemble. This would appear to exclude a nucleosynthetic source for the16O excess in CAIs and OIs.

Alternatively, some particular sorts of chemical processes may have generated mass-independent isotopic fractionations before, during, or after the formation of these objects. Thiemens (8) argues that a mass-independent isotopic fractionation effect is restricted neither to ozone (O3)-oxygen (O2) reactions nor to photochemical reactions but should occur more generally. He postulated the possibilities of three different types of reactions: recombination reactions, isotopic exchange, and thermal dissociation. For example, Wen and Thiemens (7) have experimentally demonstrated that atomic oxygen was enriched in 16O up to ∼40‰ relative to CO2 through isotopic exchange between the two gaseous species. This particular process, however, is unlikely to be a cause of the oxygen isotopic anomalies in CAIs or OIs because atomic oxygen and CO2 are extremely minor constituents (25) as compared with the major oxygen-bearing gaseous species (such as CO and H2O), which prevents their becoming effective oxygen reservoirs. Furthermore, in order to explain the oxygen isotopic anomalies in the condensed phases such as CAIs or OIs, the anomaly in the gas phase would have to be transferred into the lattice-forming oxygen; no effective process for this has been postulated.

We cannot presently determine the origin of the oxygen isotopic anomalies in CAIs and OIs. However, the observation that OIs, CAIs, and CAI rims all show a unique oxygen isotopic composition (δ17O and δ18O ∼−40 to −50‰) seems to suggest that the oxygen isotopic anomalies are more plausibly attributed to some kind of nebular chemical process that occurred during the formation of these primordial objects rather than to an exotic carrier with 16O enrichment.

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


Stay Connected to Science

Navigate This Article