Origin of Isotopic Diversity among Carbonaceous Chondrites

The ungrouped and anomalous carbonaceous chondrites of this study together with samples from the major groups of carbonaceous chondrites plot on a single δ^128/126Te–1/[Te] correlation line (Figure 1). For the major groups of carbonaceous chondrites, δ^128/126Te is also correlated with the mass fraction of matrix, indicating that the correlated variations of δ^128/126Te and Te content reflect variable amounts of CI chondrite-like matrix (Hellmann et al. 2020). This, in turn, implies that chondrules are characterized by light Te isotopic compositions (i.e., low δ^128/126Te), which is approximately the same for the major carbonaceous chondrite groups. The low δ^128/126Te of the chondrules may result from Te isotope fractionation during chondrule formation but may also be inherited from the precursor material of the chondrules (see Hellmann et al. 2020 for details). For the samples of this study, the volume fractions of matrix are known from the petrographic classification of the meteorites, but since the bulk densities of the samples are unknown, these volume fractions cannot easily be translated into matrix mass fractions. Thus, it is difficult to assess whether these samples also plot on a correlation line between δ^128/126Te and matrix mass fraction, but we note that the elevated δ^128/126Te values of these samples are consistent with their generally high matrix volume fractions.

Zoom In Zoom Out Reset image size

Hellmann et al. (2020) showed that for CV, CO, CM, and CI chondrites, as well as for Tagish Lake, δ^128/126Te is also correlated with nucleosynthetic μ⁵⁴Cr anomalies (Figure 1(b)). These correlations likely reflect mixing between volatile-rich, isotopically heavy (i.e., higher δ^128/126Te), and⁵⁴Cr-rich CI chondrite-like matrix with volatile-poor and isotopically light chondrules/chondrule precursors. This is consistent with the observation that chondrules from CV, CO, and CM chondrites, despite the large variations among individual chondrules from each group, are characterized by the same average μ⁵⁴Cr of ∼60 (Hellmann et al. 2020). This indicates that for a given chondrite group the offset of a bulk chondrite's μ⁵⁴Cr from the mean μ⁵⁴Cr of its chondrules also correlates with the amount of matrix in each chondrite. The new Te and Cr isotope data reveal that most ungrouped and anomalous carbonaceous chondrites also plot on the μ⁵⁴Cr–δ^128/126Te mixing line defined by the major groups (Figure 1(b)), indicating that the Cr and Te isotope variability among most carbonaceous chondrites can be accounted for by variable abundances of the same two components, namely chondrules/chondrule precursors and CI chondrite-like matrix.

The only carbonaceous chondrites clearly deviating from this general trend are the CR chondrites, which plot well below the μ⁵⁴Cr–δ^128/126Te mixing line defined by other carbonaceous chondrites. The CR1 chondrite GRO 95577 also plots below this mixing line but has a distinctly higher δ^128/126Te value and Te content compared to the characteristic composition of CR chondrites (Figure 1). Of note, GRO 95577 plots on the δ^128/126Te–1/[Te] mixing line between chondrules and matrix, indicating that its higher Te content is due to a higher portion of CI-like matrix compared to other CR chondrites and not solely the result of Te redistribution during parent body alteration. In the μ⁵⁴Cr–δ^128/126Te diagram, both CR chondrites and GRO 95577 plot on a mixing line between CR chondrules and CI chondrite-like matrix (Figure 1(b)), suggesting that the distinct isotopic composition of CR chondrites and GRO 95577 compared to other carbonaceous chondrites predominantly reflects the incorporation of a distinct population of chondrules with more elevated μ⁵⁴Cr values compared to chondrules in CV, CM, and CO chondrites (Figure 1(b)). This interpretation is consistent with a model in which CR chondrules formed by recycling of an earlier chondrule generation after mixing with CI chondrite-like dust (Marrocchi et al. 2022).

The three ungrouped C2 chondrites Essebi, Tarda, and Acfer 094 also appear to plot slightly below the μ⁵⁴Cr–δ^128/126Te mixing line between CV–CM–CO chondrules and CI chondrite-like matrix although this offset is not clearly resolved. If it exists, the small offset may reflect a slightly more elevated average μ⁵⁴Cr of the chondrules in these chondrites compared to chondrules in CV, CO, and CM chondrites. As for the CR chondrites, this higher μ⁵⁴Cr may reflect a slightly higher fraction of CI chondrite-like material among the precursors of these chondrules.

In summary, the Cr and Te isotopic variability among carbonaceous chondrites can be accounted for by mixing between chondrules/chondrule precursors and CI chondrite-like matrix. Whereas the latter appears to be contained in all carbonaceous chondrites, the⁵⁴Cr isotopic composition of the chondrule component may vary among the individual groups. This is most obvious for CR chondrites, whose chondrules have substantially more elevated μ⁵⁴Cr values compared to chondrules from most other carbonaceous chondrites but may also account for small deviations from the μ⁵⁴Cr–δ^128/126Te mixing line that may exist for some ungrouped carbonaceous chondrites.

The μ⁵⁴Cr anomalies among the carbonaceous chondrites are not only correlated with δ^128/126Te but are also inversely correlated with μ⁵⁰Ti (Figure 2). This observation holds for the major groups of carbonaceous chondrites as well as for the ungrouped and anomalous samples of this and prior studies. This μ⁵⁰Ti–μ⁵⁴Cr correlation may reflect mixing between CI-like matrix and an isotopically distinct dust component characterized by elevated μ⁵⁰Ti and low μ⁵⁴Cr, but such a component has not yet been identified in carbonaceous chondrites (Alexander 2019). Instead, components with elevated μ⁵⁰Ti typically also have elevated μ⁵⁴Cr, such as for instance CAIs and AOAs, which are characterized by both positive μ⁵⁰Ti ∼900 (Trinquier et al. 2009; Davis et al. 2018; Render et al. 2019; Torrano et al. 2019) and μ⁵⁴Cr ∼600 (Papanastassiou 1986; Birck & Allegre 1988; Larsen et al. 2011). Moreover, given that CAIs are strongly enriched in refractory elements like Ti, they exert a strong control on the Ti isotopic composition of bulk carbonaceous chondrites (Trinquier et al. 2009) but contain too little Cr to significantly affect a carbonaceous chondrite's μ⁵⁴Cr. To assess the effect of CAI (and AOA) admixture on the Ti and Cr isotopic compositions of bulk carbonaceous chondrites more quantitatively, we have calculated mixing lines between various chondrule-matrix mixtures and refractory inclusions (Figure 2(b)). Note that Ti isotope data are only available for chondrules from the CV chondrite Allende (Gerber et al. 2017) and chondrules in other carbonaceous chondrite groups may have different μ⁵⁰Ti signatures. Nevertheless, the calculations reproduce the observed abundances of refractory inclusions in the various chondrite groups reasonably well, such as for instance ∼2%–4% CAIs/AOAs in CM chondrites and ∼3%–8% in CV and CO chondrites, respectively (e.g., Ebel et al. 2016; Kimura et al. 2020; Fendrich & Ebel 2021). Together, these calculations show that the variations in μ⁵⁰Ti among bulk carbonaceous chondrites are almost entirely governed by the abundance of refractory inclusions, while variations in μ⁵⁴Cr predominantly reflect the relative proportions of chondrules and matrix (Figure 2(b)). A corollary of these observations is that the covariation of μ⁵⁰Ti and μ⁵⁴Cr among carbonaceous chondrites appears to reflect coupled variations in the abundances of refractory inclusions and chondrules, which systematically decrease from a high abundance in CV and CO chondrites to essentially zero in CI chondrites.

Zoom In Zoom Out Reset image size

Whereas CAIs are the oldest dated solids of the solar system and, together with AOAs, are thought to have formed close to the Sun (e.g., Wood 2004), most chondrules in carbonaceous chondrites formed ∼2–4 Myr later (e.g., Kurahashi et al. 2008; Budde et al. 2016b) and further away from the Sun. The linked abundances of refractory inclusions and chondrules, therefore, require a dynamical process that led to a coupled enrichment of two distinct dust components originating from different regions of the accretion disk into the narrow formation location of specific carbonaceous chondrite parent bodies. Furthermore, the ∼2–4 Myr age difference between CAIs and chondrules implies that CAIs were stored in the accretion disk for at least this period of time before they were incorporated into their host chondrites (e.g., Cuzzi et al. 2003; Ciesla 2007). Because objects the size of CAIs are expected to be rapidly lost in the Sun by gas drag, it has been suggested that the inward drift of CAIs was blocked by a "pressure trap," i.e., a pressure maximum in the disk, which is also the location in the disk where carbonaceous chondrite parent bodies later formed (Desch et al. 2018; Scott et al. 2018; Brasser & Mojzsis 2020). This pressure maximum may have formed through the formation of Jupiter's core and the associated opening of a gap in the disk (e.g., Kruijer et al. 2017), but other origins of this structure are also possible (e.g., Izidoro et al. 2022). Regardless of its exact origin, a pressure trap provides an efficient mechanism not only for blocking the inward drift of CAIs (and AOAs) but also for producing local enrichments of these objects. This is consistent with the dearth of refractory inclusions in noncarbonaceous chondrites and their enhanced abundance in some carbonaceous chondrites (Desch et al. 2018). However, storage of refractory inclusions alone cannot account for the systematic covariation of CAI/AOA and chondrule abundances relative to matrix, which we propose reflects the trapping of not only refractory inclusions but also chondrules or their precursors in the pressure maximum. This would naturally result in coupled variations of CAIs/AOAs and chondrules because both components became enriched by the same process (Figure 3).

Zoom In Zoom Out Reset image size

The coupled trapping of refractory inclusions and chondrules raises the question of why the abundances of these components vary among the different groups of carbonaceous chondrites. Provided that all carbonaceous chondrites formed in the same pressure maximum, this requires that the enrichment of refractory inclusions and chondrules in this structure varied in time and/or space. This could be caused by the varying supply of these components due to differences in their radial drift speed caused by their distinct aerodynamical properties defined by their grain size and internal density. Models of planetesimal formation via the streaming instability show that this mechanism prefers grains of high Stokes numbers, that is, the largest and/or densest dust aggregates (Bai & Stone 2010). This may lead to preferential incorporation of refractory inclusions and chondrules over matrix dust into chondrite parent bodies as long as these more refractory components are available. This is consistent with the observation that the parent bodies of the matrix-rich CI and CM chondrites tend to have younger inferred accretion ages than those of matrix-poor chondrites such the CV and CO chondrites (Figure 4). The only carbonaceous chondrite group plotting off this trend are the CR chondrites, which are characterized by a relatively young accretion age of ∼3.6 Myr after CAI formation (e.g., Schrader et al. 2017; Budde et al. 2018) but have the lowest matrix mass fraction among the carbonaceous chondrites (Figure 4). There is some uncertainty in the matrix mass fraction of CR chondrites because different members of this group have somewhat variable matrix fractions and metal abundances. The latter exerts a strong control on the conversion of matrix volume to mass fractions and accounts for at least some of the variability in reported matrix fractions. Nevertheless, even when using the higher CR matrix mass fraction reported by Patzer et al. (2022), the CR chondrites plot off the trend of matrix mass fraction versus accretion age (Figure 4). This offset may indicate that the CR chondrites formed from distinct precursor material than the other carbonaceous chondrites or that most of their CI chondrite-like material has been processed into a second generation of chondrules (Marrocchi et al. 2022).

Zoom In Zoom Out Reset image size

Despite the good correlation between accretion age and matrix content, it is important to recognize that particularly for the matrix-rich chondrites some inconsistencies exist in the inferred accretion ages. For instance, carbonates in the matrix-rich carbonaceous chondrite Flensburg and in some samples from asteroid Ryugu (which resemble CI chondrites) have ⁵³Mn-⁵³Cr ages of ∼<1.8–2.6 Ma after CAI formation (Bischoff et al. 2021; McCain et al. 2023). This implies parent body accretion before that time and hence, earlier than inferred for the matrix-rich CI and CM chondrites. Given that both Flensburg and Ryugu are characterized by similarly high matrix fraction as these two groups of chondrites, it appears that they would not plot on the trend of accretion age versus matrix content shown in Figure 4. Clearly, more work is needed to better establish the accretion ages of especially the matrix-rich carbonaceous chondrites and to more reliably assess whether they formed later than their matrix-poor counterparts.

Based on their chemical and isotopic composition, several studies have argued that the accretion region of carbonaceous chondrites is in the outer solar system, beyond the orbit of Jupiter (e.g., Warren 2011; Kruijer et al. 2017). Some other studies, however, proposed that chondrules (e.g., Williams et al. 2020) and perhaps even their host carbonaceous chondrites (e.g., Van Kooten et al. 2021) formed in the inner solar system. This proposal is based primarily on the observation that some chondrules from CV chondrites have ⁵⁴Cr isotopic compositions overlapping with those of NC chondrites, which are thought to represent the inner disk. Moreover, while the cores of some CV chondrules are characterized by ⁵⁴Cr deficits (i.e., the characteristic composition of the NC reservoir), their rims display ⁵⁴Cr excesses (i.e., the characteristic composition of the CC reservoir). This has been interpreted to indicate that CV chondrules in general formed in the inner disk, were subsequently transported outwards, and were mixed with CI chondrite-like dust, which drifted inward toward the Sun (Van Kooten et al. 2021).

In this model, refractory inclusions would be added to the accretion location of carbonaceous chondrites together with CI chondrite-like dust by inward transport from the outer disk, while chondrules would be added by outward transport from the inner disk. However, there is no reason why in this scenario the abundances of refractory inclusions and chondrules should be coupled because the outward-drifting region in a vicinity of the pressure trap is much narrower than the inward-drifting outer disk, and it gets depleted much faster (e.g., Pinilla et al. 2012). Thus, the linked abundances of refractory inclusions and chondrules argue against the massive outward transport of carbonaceous chondrite chondrules inferred based on NC-like ⁵⁴Cr signatures in some of these chondrules. Instead, these signatures, and the overall ⁵⁴Cr variability among chondrules, are best understood as reflecting isotopic heterogeneities in their precursor dust (Schneider et al. 2020). Within this framework, NC-like ⁵⁴Cr isotopic compositions of chondrules indicate that their precursor material—but not the chondrules themselves—originated in the inner disk. Similar to CAIs, this precursor material was likely transported outwards during the early viscous expansion of the disk (Nanne et al. 2019) and later drifted back toward the Sun by gas drag. This is consistent with chemical and isotopic evidence showing that CAIs have partially been incorporated into chondrules (e.g., Gerber et al. 2017), indicating that CAIs were part of the chondrule precursor material and that, therefore, the coupled abundances of CAIs and chondrules were established prior to chondrule formation.