Oxygen isotope systematics of chondrules in the Paris CM2 chondrite: indication for a single large formation region across snow line

Abstract

In-situ oxygen three-isotope analyses of chondrules and isolated olivine grains in the Paris (CM) chondrite were conducted by secondary ion mass spectrometry (SIMS). Multiple analyses of olivine and/or pyroxene in each chondrule show indistinguishable Δ¹⁷O values, except for minor occurrences of relict olivine grains (and one low-Ca pyroxene). A mean Δ¹⁷O value of these homogeneous multiple analyses was obtained for each chondrule, which represent oxygen isotope ratios of the chondrule melt. The Δ¹⁷O values of individual chondrules range from −7‰ to −2‰ and generally increase with decreasing Mg# of olivine and pyroxene in individual chondrules. Most type I (FeO-poor) chondrules have high Mg# (~99) and variable Δ¹⁷O values from −7.0‰ to −3.3‰. Other type I chondrules (Mg# ≤97), type II (FeO-rich) chondrules, and two isolated FeO-rich olivine grains have host Δ¹⁷O values from −3‰ to −2‰. Eight chondrules contain relict grains that are either ¹⁶O-rich or ¹⁶O-poor relative to their host chondrule and show a wide range of Δ¹⁷O values from −13‰ to 0‰.

The results from chondrules in the Paris meteorite are similar to those in Murchison (CM). Collectively, the Δ¹⁷O values of chondrules in CM chondrites continuously increase from −7‰ to −2‰ with decreasing Mg# from 99 to 37. The majority of type I chondrules (Mg# >98) show Δ¹⁷O values from −6‰ to −4‰, while the majority of and type II chondrules (Mg# 60-70) show Δ¹⁷O values of −2.5‰. The covariation of Δ¹⁷O versus Mg# observed among chondrules in CM chondrites may suggest that most chondrules in carbonaceous chondrites formed in a single large region across the snow line where the contribution of ¹⁶O-poor ice to chondrule precursors and dust enrichment factors varied significantly.

1. INTRODUCTION

Among carbonaceous chondrites (CCs), CM (Mighei-like) chondrites are the most abundant group (Weisberg et al., 2006). They are recognized as xenoliths in numerous other groups of CCs and meteorite classes and thus may correspond to the most abundant and/or widely dispersed material in the main belt (e.g., Zolensky et al., 1996; Gounelle et al., 2003; Bischoff et al., 2006; and references therein; Briani et al., 2012), hence their importance to deciphering the formation and evolution of the early Solar System. In addition to low to mild thermal metamorphism (e.g., Nakamura, 2006; Kimura et al., 2011; Tonui et al., 2014), CM chondrites experienced intense parent body aqueous alteration (e.g., Sears and Dodd, 1988; Brearley, 2003; Busemann et al., 2007; Schrader and Davidson, 2017). Most CM chondrites are of petrologic type 2 (e.g., McSween 1979; Kallemeyn and Wasson, 1981; Zolensky et al., 1993) and display various degrees of aqueous alteration and have been divided into petrologic subtypes from 3.0 to 2.0, where numbers decreases with increasing alteration (e.g., Zolensky et al., 1997; Rubin et al., 2007; Rubin, 2015; Kimura et al., 2020). Some CM chondrites are almost completely altered and assigned to be subtype 2.0, which was previously classified as CM1 (Zolensky et al., 1997; Rubin e al., 2007).

Paris is one of the least altered CM chondrites and has been used to investigate the early stages of the parent body aqueous alteration of CM chondrites (Hewins et al., 2014; Marrocchi et al., 2014; Rubin, 2015; Leroux et al., 2015; Pignatelli et al., 2016; Vacher et al., 2016, 2017; Verdier-Paoletti et al., 2017). Based on a detailed petrographic and mineralogical survey, Marrocchi et al. (2014) classified Paris as a CM2.7. However, Paris contains both highly and less altered lithologies (Hewins et al., 2014). Based on the PCP (Poorly Characterized Phases) index defined by Rubin et al. (2007), the less altered lithology of Paris is of petrologic subtype 2.9 (Hewins et al., 2014), which is consistent with the observation of a significant amount of Fe-Ni metal blebs and the presence of a pristine matrix (Leroux et al., 2015; Rubin, 2015). The oxygen isotope ratios of the less altered lithologies are as low as δ¹⁷O = −2.1‰o and δ¹⁸O = 2.4‰, which is at the lower end of a linear trend defined by Paris subsamples and bulk CM2 chondrites (Hewins et al., 2014). The variation of oxygen isotope ratios among bulk CM chondrites is ascribed to heterogeneity in the extent of secondary aqueous alteration processes (Clayton and Mayeda, 1999; Hewins et al., 2014). The less altered lithology of Paris thus offers a unique opportunity to investigate the origin and petrogenesis of CM chondrules.

In situ SIMS (secondary ion mass spectrometry) oxygen 3-isotope analysis of individual chondrules is a powerful tool to constrain the conditions of their formation (e.g., Kita et al., 2010; Ushikubo et al., 2012; Schrader et al., 2013; Marrocchi et al. 2018; 2019). Many chondrules in CCs contain olivine grains with heterogeneous oxygen isotope ratios (e.g., Kunihiro et al., 2004, 2005; Jones et al., 2004; Wasson et al., 2004; Connolly and Huss, 2010; Rudraswami et al., 2011; Ushikubo et al., 2012; Schrader et al., 2013; 2017; Tenner et al., 2013; Marrocchi et al., 2018; 2019). They are considered as “relict” and interpreted as unmelted material that survived the final high-temperature event of chondrule formation. Because of the slow diffusivity of oxygen isotopes in olivine (e.g., Chakraborty, 2010), these relict grains preserved their initial oxygen isotope ratios and would provide important knowledge about precursor solids that formed chondrules. However, multiple high precision SIMS analyses of olivine and/or pyroxene in each chondrules are mostly indistinguishable (e.g., Rudraswami et al., 2011; Ushikubo et al., 2012; Tenner et al., 2013; 2015; 2017; Hertwig et al., 2018; 2019a; Chaumard et al., 2018). Ushikubo et al. (2012) showed plagioclase and glass in chondrule mesostasis from Acfer 094 are in agreement with those of olivine and pyroxene phenocrysts of the same chondrules. Internally homogeneous oxygen isotope ratios within individual chondrules represent those of the final chondrule melt from which the “non-relict” olivine and other minerals crystallized.

The degree of mass independent isotope fractionation of oxygen 3-isotopes, commonly expressed as Δ¹⁷O (= δ¹⁷O – 0.52 × δ¹⁸O), determined for individual chondrules in CCs, systematically vary against the Mg# (= MgO/[MgO+FeO] in mol.%) of olivine and pyroxene (Ushikubo et al., 2012; Tenner et al., 2013; 2015; 2017; Hertwig et al., 2018; 2019a). Variation of Mg#s among chondrules indicates that the redox state of the environments they formed in were variable. Such variations were probably due to metal-silicate equilibria (Zanda et al., 1994) which were influenced by varying proportions of anhydrous dust, H₂O ice, and organic matters relative to the solar-composition nebula gas in the outer disk regions where CC chondrules formed (e.g., Wood and Hashimoto, 1993; Grossman et al., 2008). Heterogeneous oxygen isotope ratios among chondrule precursor phases, such as ¹⁶O-rich anhydrous dust and ¹⁶O-poor H₂O-ice (Krot et al., 2006; Sakamoto et al., 2007), could explain the negative correlation between Δ¹⁷O values and Mg# among chondrules in CCs (e.g., Connolly and Huss, 2010; Schrader et al., 2013; Tenner et al., 2015; Hertwig et al., 2018).

In contrast, Δ¹⁷O values among chondrules in non-carbonaceous chondrites, such as ordinary and enstatite chondrites that likely derived from inner disk material (Kruijer et al., 2017), show narrower ranges and are ¹⁶O-depleted relative to those in CCs (e.g., Kita et al., 2010; Weisberg et al., 2011; Libourel and Chaussidon, 2011; Schneider et al., 2020). Kita et al. (2010) argued that solid precursors of ordinary chondrite chondrules were fractionated in δ¹⁸O as a result of condensation of solids from ¹⁶O-poor nebula gas at high temperatures. It is likely that the oxygen isotope ratios of precursor phases were homogenized in the inner disk region prior to the formation of chondrules.

Chaumard et al. (2018) studied 29 chondrules in the Murchison CM2 chondrite and found that the distribution of Mg#s and Δ¹⁷O of individual chondrules are similar to those in Acfer 094 and CO3 chondrites. Co-existing olivine and pyroxene in individual Murchison chondrules show indistinguishable oxygen isotope ratios excluding the minor occurrence of relict grains, which further indicate that chondrules solidified from numerous melt droplets with homogeneous oxygen isotope ratios. Chondrule oxygen isotope systematics in Murchison further indicated that the CO and CM chondrite parent bodies collected similar populations of chondrules, while other studies suggest that the timing of the CM parent body accretion may have been delayed relative to the CO parent body accretion (e.g., Sugiura and Fujiya, 2014).

In Murchison, chondrules have an extremely limited range of Mg#s ranging from 99.6 to 99.0 for type I chondrules (with one exception with Mg#~96) and a slightly larger range, from ~65 to ~70, for type II chondrules. The Δ¹⁷O values among Murchison chondrules show a hint of systematic increase with decreasing Mg#s. Here we report in situ high precision SIMS oxygen 3-isotope measurements of olivine and pyroxene in chondrules from the less altered lithology of the Paris CM2 chondrite. The Paris meteorite provides an opportunity to study a diverse range of less altered chondrules in CM chondrites in petrologic context, including chondrule textures, mineralogy, and mineral chemistry (Hewins et al., 2014; Rubin, 2015; Stephant et al., 2017). This will enable us to further constrain the nature of pristine chondrules from carbonaceous chondrites in general.

2. ANALYTICAL PROCEDURES

2.1. Sample and chondrule selection

Within the less altered lithology of the Paris CM chondrite (type 2.9; Hewins et al., 2014), we selected and analyzed 29 chondrules from one polished section allocated by the Muséum national d’Histoire naturelle of Paris (MNHN 4029-11). We intended to obtain the most diverse selection of chondrules based on their size, texture, and mineralogy. Selected chondrules include 25 type I (Mg#≥90) chondrules (from ~200 μm to ~2 mm in diameter), three type II (Mg#<90) chondrules (~250–600 μm in diameter), and one fragment of a type I chondrule (~600 μm long). We also analyzed two isolated grains of FeO-rich olivine (~250–400 μm in size), one amoeboid olivine aggregate (AOA) (~500 μm long), and one FeO-poor olivine-bearing object (~400 μm long).

Nine type I chondrules are porphyritic olivine-pyroxene (POP; 20–80% modal olivine), three are porphyritic olivine (PO; >80% modal olivine), and eight are porphyritic pyroxene (PP; <20% modal olivine). One type I chondrule is barred olivine (BO) and two are composed of a BO core surrounded by a POP rim. We also analyzed one type I granular olivine (GO) chondrule and one granular olivine-pyroxene (GOP) chondrule. The fragment of the type I chondrule displays a BO texture. The three type II chondrules have a PO texture.

In all groups of CCs, chondrules are predominantly type I (e.g., Jones, 2012). It has been reported that ~95% of chondrules in CM chondrites are porphyritic, 10–40% of these porphyritic chondrules being type II (Jones, 2012; and references therein). Approximately 80% of the chondrules analyzed are porphyritic and ~13% of them are type II. This selected population is thus roughly representative of chondrules in CM chondrites. Moreover, PO, POP, and PP chondrules in our selection represent approximately 10%, 31%, and 28%, respectively, of the entire population of chondrules analyzed in this study.

2.2. Scanning electron microscopy and electron microprobe analysis

Backscattered electron (BSE), secondary electron (SE) imaging and energy dispersive X-ray spectrometry (EDS) analyses of chondrules were performed using a Hitachi S-3400N scanning electron microscope (SEM) at the University of Wisconsin-Madison. The accelerating voltage was set to 15 kV. The locations of SIMS analyses were selected for olivine and pyroxene grains in each chondrule that are free of cracks and other phases as identified in BSE and SE images.

We used a Cameca SXFive FE electron microprobe at the University of Wisconsin-Madison to obtain quantitative chemical analyses of olivine and pyroxene grains with an accelerating voltage of 15 keV and a beam current of 20 nA. Counting times for the peak and background were 10 and 5 s, respectively. Several standards were analyzed for matrix correction of individual elements: natural olivine and synthetic forsterite and enstatite (Mg, Si), jadeite (Na, Al), microcline (K), chromian augite (Ca), TiO₂ (Ti), synthetic Cr₂O₃ (Cr), fayalite (Fe), and synthetic Mn₂SiO₄ (Mn). We analyzed Mg, Al, Si, and Na with a LTAP crystal, Ca and K with a LPET crystal, Cr and Ti with a PET crystal, and Mn and Fe with a LLIF crystal. We used the Probe for EPMA™ (PFE) software (Donovan, 2015) for data reduction and matrix corrections (ZAF and fr(z)). For each analysis, we calculated the Mg# based on EPMA measurements.

2.3. SIMS oxygen three-isotope analysis

We performed in situ oxygen three-isotope analyses of olivine and pyroxene using the Cameca IMS 1280 at the WiscSIMS laboratory, University of Wisconsin-Madison. Analytical conditions and data reduction methods were generally similar to those of Kita et al. (2010) and Tenner et al. (2013, 2015) using multi-collector Faraday cups; ¹⁶O and ¹⁸O on multi-collector array and ¹⁷O on a fixed mono-collector (axial detector FC2). The analyses were performed in two sessions with different primary Cs⁺ beam conditions; a 15 μm (session #1) and 10 μm (session #2) diameters with intensities of ~3 nA and ~1 nA, respectively. During the analysis session #1 (15 μm diameter spot), secondary ion intensities of ¹⁶O, ¹⁷O, and ¹⁸O were ~3.5×10⁹, ~1.5×10⁶, and ~7.5×10⁶ counts per second (cps), respectively. FC amplifier resistors were 10¹¹ Ω for ¹⁷O and ¹⁸O, and 10¹⁰ Ω for ¹⁶O as in the previous studies (e.g., Kita et al., 2010), though a newer version of the FC amplifier board was used for the detection of ¹⁷O on the axial FC2 detector. This newer FC amplifier board from Cameca (for IMS 1280-HR) shows improved thermal noise that is close to the theoretical limit (1SD ~1,300 cps for 4 s integrations) compared to the original FC amplifier board installed to IMS 1280 (1SD ~2,000 cps for 4 s integrations). Taking advantage of lower noise level on ¹⁷O, the acquisition time for a single analysis was reduced from 200 s to 100 s and the total analysis time (including pre-sputtering and secondary beam centering time) was reduced from 7 min to 4 min compared to previous oxygen 3-isotope analyses at the WiscSIMS laboratory. Analyses session #2 (10 μm) was performed at the same time that some of chondrules from the Murchison were analyzed (Chaumard et al., 2018). Secondary ion intensities of ¹⁶O, ¹⁷O, and ¹⁸O decreased to ~1.5×10⁹, ~5.5×10⁵, and ~2.9×10⁶ cps, respectively. In order to reduce the noise level of the FC amplifier for ¹⁷O analyses below 10⁶ cps intensity, we replaced the resistor and capacitor pair on the original IMS 1280 FC amplifier board by a 10¹² Ω resistor and 1 pF capacitor from a Finnigan MAT 251 stable mass spectrometer (noise level was reduced to 1SD ~1,200 cps for 4 s integrations; Chaumard et al., 2018). The mass resolving power (MRP at 10% peak height) for both sessions was set to ~2200 for ¹⁶O and ¹⁸O, and 5000 for ¹⁷O. We measured ¹⁶OH⁻ at the end of each analysis to determine its contribution to the ¹⁷O⁻ signal following the methods described by Heck et al. (2010). The correction of the ¹⁷O⁻ signal from ¹⁶OH⁻ was negligible (<0.08‰) for both standards and unknowns.

We normalized the measured ¹⁸O/¹⁶O and ¹⁷O/¹⁶O ratios to the VSMOW scale (δ¹⁸O and δ¹⁷O expressed as a deviation from standard mean ocean water in the unit of 1/1000; VSMOW-scale, Baertschi 1976). The external reproducibility has been determined by intermittent measurements of a San Carlos olivine (SC-Ol) standard (δ¹⁸O = 5.32‰; Kita et al., 2010). We bracketed 4 to 19 unknown chondrule analyses with 8 SC-Ol analyses, 4 before and 4 after (Kita et al., 2009). External reproducibility is calculated as the 2SD of the SC-Ol brackets, with average values during session #1 (15 μm) of 0.2‰, 0.3‰, and 0.3‰ for δ¹⁸O, δ¹⁷O, and Δ¹⁷O, respectively. For session #2 (10 μm), the 2SD average values are 0.4‰, 0.5‰, and 0.4‰ for δ¹⁸O, δ¹⁷O, and Δ¹⁷O, respectively. These values represent the spot-to-spot reproducibility and were thus assigned as the uncertainties of each individual spot analysis (Kita et al., 2009). Measurements of four olivine (Fo_0.6-100), three low-Ca pyroxene (En_70-97), and one diopside (Wo₅₀) standards with known oxygen isotope ratios (Eiler et al., 1997; Kita et al., 2010) were used to estimate corrections for instrumental biases of unknown olivine and pyroxene analyses (EA1). The compositional ranges of standards cover those of unknowns measured.

We obtained multiple SIMS analyses per chondrule (n=4 to 12, typically 8) to examine the homogeneity of the isotope ratios. As in the previous studies, a specific analysis in a single chondrule is identified as a relict when its Δ¹⁷O value deviates more than 0.5‰ and 0.7‰ (3SD limits of bracket standard analyses in each analysis session) from the chondrule mean (Ushikubo et al., 2012; Tenner et al., 2013, 2015; 2017; Hertwig et al., 2018; 2019a; Chaumard et al., 2018). To identify all chondrules containing rare relict grains, a multitude of analyses per chondrules (e.g., ~50; Marrocchi et al., 2018; 2019) is required; however, a total of ~8 analyses per chondrule is sufficient to identify the Δ¹⁷O value of most phenocrysts that in turn represents the value of the chondrule melt, if the chondrule consists largely of minerals with homogeneous oxygen isotope ratios. Thus, we can calculate mean oxygen isotope ratios (referred to as “host chondrule” oxygen isotope ratios) from multiple analyses within each chondrule excluding relict grains. The uncertainties of host δ¹⁸O and δ¹⁷O values is the propagation of (i) the 2 standard error of the mean of multiple analyses that constitute the host chondrule value (=2SD/√number of analyses; 2SD of the bracket standard analyses or 2SD of multiple analyses within a chondrule, whichever is larger), (ii) the 2SE of associated SC-Ol bracketing analyses for instrumental bias correction, and (iii) the uncertainty due to the sample topography and/or sample positioning on the SIMS stage as well as uncertainty of instrumental bias corrections, estimated to be 0.3‰ for δ¹⁸O and 0.15‰ for δ¹⁷O (Kita et al., 2009). Because (iii) is mass-dependent and does not affect Δ¹⁷O, we only use (i) and (ii) for the propagated uncertainty of Δ¹⁷O. The uncertainties of Δ¹⁷O in the relict grains are the spot-to-spot reproducibility (2SD) as determined by bracketing analyses of San Carlos olivine.

We inspected each SIMS spot by obtaining BSE and SE images using SEM after each SIMS session. Ten of 252 pits were rejected from our final dataset because they either overlapped cracks, imperfections, or display a contribution of ¹⁶OH to the ¹⁷O signal of ~0.1‰ or higher.

3. RESULTS

3.1. Texture, petrography, and mineralogy

Chondrules display a wide diversity of sizes and textures, as previously observed by Hewins et al. (2014) for the Paris CM chondrite. In the following summary, textural features and chemistry of chondrule minerals are described. BSE images of individual chondrules are shown in Figs. 1-4 and EA2, in which locations of EPMA and SIMS analyses are annotated. EPMA major element data for olivine and pyroxene are shown in EA3. Additional petrologic descriptions of individual chondrules can be found in EA4. Within 19 of the 29 chondrules analyzed, coexisting olivine and low-Ca pyroxene have similar Mg#s. For the other 10 chondrules, we were only able to obtain quantitative analyses either for olivine or low-Ca pyroxene. In type I chondrules where both olivine and pyroxene were analyzed, Mg#s of olivine and low-Ca pyroxene are indistinguishable, similar to type I chondrules in other pristine chondrites (e.g., Jones, 1994; Tachibana et al., 2003; Ushikubo et al., 2012; Tenner et al., 2013, 2015; Schrader and Davidson, 2017; Schrader et al., 2017; Chaumard et al., 2018; Hertwig et al., 2018; 2019a).

Fig 1.

BSE images of type I porphyritic chondrules analyzed in Paris that are homogeneous in terms of oxygen isotope compositions (C18: POP, C30: PO, C21 and C11: PP). SIMS oxygen three-isotope analysis points are shown by the vertex of the triangles, color-coded for mineral phases (olivine: blue, relict olivine: white, low-Ca pyroxene: green). Δ¹⁷O values of individual analyses are indicated.

Fig. 4.

BSE images of a type II chondrule (C15: PO), two isolated Fe-rich olivine grains (G32 and G33), and an isolated Fe-poor olivine grain (G24) analyzed in Paris. SIMS oxygen three-isotope analysis points are shown by the vertex of the triangles, color-coded for olivine (blue). Δ¹⁷O values of individual analyses are indicated.

3.1.1. Type I porphyritic chondrules

Olivine in type I porphyritic chondrules is present as large anhedral phenocrysts and euhedral grains (up to ~300 μm) (Fig. 1, 2). In several chondrules (e.g., C4, C8, C7; Fig. 2a for C7), olivine is located in the cores while the pyroxene is more abundant at the periphery. In type I POP chondrules, low-Ca pyroxene often displays a poikilitic texture with numerous cracks and pores (e.g., C7; Fig. 2a). Low-Ca pyroxene was also observed as large euhedral grains, i.e., up to ~150 μm length in chondrule C7 (Fig. 2a). In type I PP chondrules, low-Ca pyroxene display less cracks and pores (e.g., C21; Fig. 1c). High-Ca pyroxene is also present in many type I porphyritic chondrules as small grains (from a few μm to ~70 μm length), either in association with Low-Ca pyroxene or olivine (e.g., C30; Fig. 1b). Olivine grains in type I porphyritic chondrules are chemically homogeneous, with an average Mg# of 99.0±0.8 (2SD) (97.6–99.8). They contain up to 0.44 wt% Al₂O₃, 0.84 wt% CaO, 0.67 wt% Cr₂O₃, and 0.46 wt% MnO. Low-Ca pyroxene grains have Mg#s ranging from 92.5 to 99.2 and compositions of En_89-98Fs_1-7Wo_0-4. We measured 0.17–2.13 and 0.30–1.07 wt% Al₂O₃ and Cr₂O₃, respectively.

Fig 2.

BSE images of type I chondrules analyzed in Paris displaying various textures (C7: POP, C5: POP + BO core, C16: BO fragment, C25: PP, and C9: GO). SIMS oxygen three-isotope analysis points are shown by the vertex of the triangles, color-coded for mineral phases (olivine: blue, relict olivine: white, low-Ca pyroxene: green, high-Ca pyroxene: orange). Δ¹⁷O values of individual analyses are indicated.

The two POP chondrules C5 (Fig. 2b) and C14 (Fig. 3) contain fragments of BO chondrules in their cores with sizes of ~200 μm and ~600 μm, respectively. Olivine and pyroxene compositions do not show significant differences between BO core and POP rim (see analysis locations on BSE images in EA2 and EPMA data in EA3). There are two type I granular chondrules (C1, GOP; C9, GO) that are composed of evenly sized grains. Olivine grains in the chondrule C9 display linear trails of micron-sized inclusions of metal (Fig. 2e). Mg# of olivine and low-Ca pyroxene in these type I chondrules are all close to ~99.

Fig 3.

BSE images of the type I chondrule C14 analyzed in Paris. SIMS oxygen three-isotope analysis points are shown by the vertex of the triangles, color-coded for mineral phases (olivine: blue, relict olivine: white, low-Ca pyroxene: green). Δ¹⁷O values of individual analyses are indicated.

3.1.2. Type II PO chondrules

The three type II chondrules analyzed display a porphyritic texture and are mainly composed of olivine phenocrysts. Most of the olivine grains in chondrule C15 have euhedral shapes (Fig. 4a). The chondrules C15 and C17 contain olivine grains with forsteritic cores (~20–80 μm, Fo_99-91). The ranges of olivine Mg#s excluding forsteritic core in chondrule C15 and C17 are 58.9–77.3 and 69.9–79.7, respectively. In chondrule C22, olivine displays Mg#s ranging from 60.3–79.0. Olivine grains in these three type II chondrules contain 0.20–0.60 wt% Cr₂O₃ and 0.02–0.40 wt% MnO.

3.1.3. Isolated olivine grains

The isolated FeO-rich olivine grains G32 and G33 (Fig. 4b-c) are 200-300 μm in sizes. The Fo contents of olivine grains show ranges of 67-75 and 30-49 for G32 and G33, respectively, which becomes more FeO-rich towards the margin of grains. Another isolated object G24 is FeO-poor olivine-bearing object (Fig. 4d) with Mg# of ~99.4±0.1. The interior of the object G24 is composed of olivine that contains numerous μm-sized Fe-metal inclusions. The texture of olivine is similar to that of dusty olivine grains, in which FeO was reduced to form Fe-rich metal (Nagahara, 1981; Rambaldi, 1981).

3.1.4. Amoeboid olivine aggregate (AOA)

The AOA I3 texturally resembles the amoeboid olivine inclusions (AOI) previously observed by Rubin (2015) in Paris, as well as AOAs in CV and CK chondrites (e.g., Grossman and Steele, 1976; Rubin, 2013). Small grains of diopside (<5 μm) are enclosed within large, porous, forsteritic olivine grains. The AOA I3 mostly contains chemically homogeneous olivine with Mg# 99.5±0.1 and 0.35±0.11 and 0.28±0.03 wt% of Cr²O₃ and MnO, respectively.

3.2. Oxygen isotope ratios

A total of 242 oxygen 3-isotope analyses were obtained in 2 SIMS sessions from 33 objects (including 29 chondrules) in the CM2 Paris, which are listed in EA5. Typically, 8 analyses were performed in each object at the same location as the EPMA analyses (EA2 and EA3), including multiple phases (olivine, low-Ca pyroxene, and high-Ca pyroxene) where available. Only four good analyses were obtained from BO chondrule C13 because several analyses were rejected due to significant surface roughness (EA2). Only four analyses each were taken from two FeO-rich isolated olivine grains that were relatively small (200-300 μm). We performed four analyses in AOA I13, though two were rejected after inspections of the SIMS pits. Our new data plot between the CCAM (carbonaceous chondrite anhydrous mineral; Clayton et al., 1977) and the Y&R (Young and Russell, 1998) lines, close to the PCM (primitive chondrule minerals; Ushikubo et al., 2012) line. The δ¹⁷O and δ¹⁷O data from chondrules range from −23‰ to +4‰ and from −25‰ to +2‰, respectively, for olivine, and from −11‰ to +2‰ and from −14‰ to −1‰, respectively, for pyroxene (Fig. 5). These ranges are very similar to those observed in chondrules from Murchison (Chaumard et al., 2018).

Fig. 5.

Oxygen 3-isotope diagram of individual spot analyses of olivine, Low-Ca pyroxene (Lpx), high-Ca pyroxene (Hpx), and relict grains in chondrules in Paris. Error bars, corresponding to the spot-to-spot reproducibility (2SD), are smaller than the symbol sizes. The CCAM (carbonaceous chondrite anhydrous mineral; Clayton et al., 1977), Y&R (Young and Russell, 1998), and PCM (primitive chondrule minerals; Ushikubo et al., 2012) lines are shown for reference. The terrestrial fractionation line (TFL) is also shown.

3.2.1. Chondrule oxygen isotope analyses

Table 1 lists the host oxygen isotope ratio calculated for all chondrules except for two (C9, C15), as well as their texture, Mg#, the number of measurements per mineral, and the individual analyses that were not included in host chondrule calculations. In 18 chondrules, multiple analyses within a single chondrule display indistinguishable Δ¹⁷O values within the 3SD external reproducibility (0.5‰ and 0.7‰, for 15 μm and 10 μm spot analyses, respectively). These chondrules are considered to be internally homogeneous in oxygen 3-isotope ratios. Eight chondrules contain olivine grains showing Δ¹⁷O values that differ by more than the 3SD external reproducibility from the average values calculated using the remaining multiple (> 5) analyses. As described in section 2.3, these olivine grains with distinct oxygen isotope ratios are considered to be relict and were not used to calculate the individual host chondrule values. In the case of chondrule C1, one each of olivine and low-Ca pyroxene show Δ¹⁷O values of −9.5‰ and −7.0‰, respectively, which are significantly lower than remaining five analyses (two olivine and three low-Ca pyroxene grains) with the mean Δ¹⁷O = −5.0±0.6‰ (2SD) (Fig. 6). We considered the ¹⁶O-rich olivine and low-Ca pyroxene grains as relict grains. By excluding relict grain data, we calculate the host chondrule oxygen ratios from multiple (4–11) analyses for all but two chondrules. We consider the host chondrule Δ¹⁷O values to represent oxygen isotope ratios of the chondrule melt during their formation (e.g., Ushikubo et al., 2012; Tenner et al., 2013; 2015).

Table 1.

Mg#’s and O-isotope ratios of host chondrules and relict grains.

Chondrule	Type, texture	Mg# a	+/− b	n (ol, lpx, hpx)c	δ18O	unc.	δ17O	unc.	Δ17O	unc.	Δ17O 2SD	Beam (μm)
Homogeneous
C21	I, PP	99.0	0.2/0.3	2,5,0	−7.7	0.3	−10.9	0.3	−6.93	0.22	0.6	15
C6	I, PP	98.8	0.3/0.8	2,4,0	−6.5	0.4	−9.5	0.4	−6.10	0.34	0.3	15
C28	I, PO	99.6	0.1/0.2	8,0,0	−6.3	0.3	−9.3	0.3	−5.99	0.18	0.4	15
C23	I, POP	98.7	0.3/0.3	3,4,0	−5.7	0.6	−8.8	0.3	−5.84	0.17	0.4	15
C10	I, PO	99.2	0.1/0.2	5,1,0	−6.4	0.5	−9.1	0.3	−5.72	0.16	0.3	15
	relict ol	99.3	n.a.	1,0,0	−5.8	0.2	−9.5	0.3	−6.5	0.3	n.a.	15
	relict ol	99.0	n.a.	1,0,0	−7.3	0.2	−10.3	0.3	−6.6	0.3	n.a.	15
C27	I, PP	99.3	0.5/0.9	6,2,0	−4.5	0.4	−7.9	0.3	−5.57	0.13	0.2	15
C4	I, POP	99.0	0.2/0.3	6,2,0	−5.9	0.4	−8.7	0.4	−5.63	0.31	0.2	15
C26	I, POP	99.0	0.1/0.1	2,3,0	−4.7	0.4	−7.9	0.3	−5.51	0.25	0.4	15
	relict ol	99.3	n.a.	1,0,0	−6.8	0.2	−9.8	0.5	−6.3	0.4	n.a.	15
	relict ol	99.3	n.a.	1,0,0	−6.8	0.2	−10.1	0.5	−6.6	0.4	n.a.	15
	relict ol	99.2	n.a.	1,0,0	−6.3	0.2	−9.9	0.5	−6.6	0.4	n.a.	15
C29	I, PP	99.0	0.3/0.9	2,4,0	−4.8	0.4	−7.9	0.3	−5.43	0.19	0.4	15
C20	I, POP	98.9	0.4/0.2	4,5,0	−5.0	0.3	−7.9	0.2	−5.32	0.18	0.4	15
C13	I, BO	99.1	0.2/0.1	4,0,0	−4.2	0.4	−7.4	0.3	−5.18	0.19	0.3	15
C2	I, POP	98.9	0.3/0.4	4,4,0	−4.8	0.4	−7.7	0.2	−5.19	0.17	0.4	15
C19	I, POP	98.6	0.2/0.2	5,3,0	−4.8	0.3	−7.5	0.2	−4.98	0.14	0.3	15
C30	I, PO	98.8	0.1/0.1	6,0,0	−3.2	0.3	−5.6	0.3	−3.90	0.26	0.2	15
	relict ol	98.7	n.a.	1,0,0	1.0	0.3	−1.3	0.4	−1.9	0.5	n.a.	15
	relict ol	98.4	n.a.	1,0,0	−4.3	0.3	−6.3	0.4	−4.1	0.5	n.a.	15
C18	I, POP	98.7	0.1/0.3	0,3,0	−1.2	0.5	−4.0	0.4	−3.31	0.29	0.4	15
	relict ol	99.3	n.a.	1,0,0	2.8	0.2	0.7	0.5	−0.8	0.4	n.a.	15
	relict ol	99.1	n.a.	1,0,0	2.4	0.2	0.0	0.5	−1.2	0.4	n.a.	15
	relict ol	99.1	n.a.	1,0,0	2.3	0.2	0.0	0.5	−1.2	0.4	n.a.	15
	relict ol	99.1	n.a.	1,0,0	2.0	0.2	−0.5	0.5	−1.5	0.4	n.a.	15
	relict ol	99.1	n.a.	1,0,0	2.4	0.2	−0.4	0.5	−1.6	0.4	n.a.	15
C11	I, PP	93.2	0.2/0.7	0,8,0	1.5	0.3	−1.7	0.3	−2.51	0.18	0.5	15
C12	I, PP	93.6	0.8/0.6	0,7,0	1.3	0.4	−1.7	0.2	−2.37	0.13	0.3	15
C16	I, BO, frag.	99.4	0.2/0.2	8,0,0	−7.9	0.4	−11.1	0.3	−6.99	0.24	0.4	10
C7	I, POP	99.0	0.1/0.2	6,2,0	−6.7	0.5	−9.4	0.4	−5.90	0.24	0.4	10
C8	I, POP	98.9	0.0/0.1	3,2,0	−5.3	0.4	−8.2	0.3	−5.42	0.22	0.4	10
	relict ol	98.8	n.a.	1,0,0	2.2	0.4	1.2	0.4	0.1	0.4	n.a.	10
	relict ol	98.7	n.a.	1,0,0	−2.1	0.4	−5.7	0.4	−4.6	0.4	n.a.	10
	relict ol	98.7	n.a.	1,0,0	−0.9	0.4	−3.8	0.4	−3.3	0.4	n.a.	10
C14	I, POP+BO core	99.0	0.2/0.3	7,4,0	−5.3	0.5	−8.0	0.3	−5.26	0.23	0.6	10
	relict ol	99.1	n.a.	1,0,0	−22.7	0.3	−24.8	0.4	−13.0	0.4	n.a.	10
C1	I, GOP	98.6	0.3/0.3	2,3,0	−3.7	0.5	−6.9	0.4	−4.99	0.33	0.6	10
	relict ol	99.0	n.a.	1,0,0	−14.6	0.6	−17.2	0.6	−9.5	0.5	n.a.	10
	relict lpx	98.8	n.a.	0,1,0	−8.4	0.6	−11.4	0.6	−7.0	0.5	n.a.	10
C25	I, PP	98.8	0.4/0.8	3,4,1	−3.2	0.4	−6.3	0.4	−4.66	0.26	0.6	10
C5	I, POP+BO core	98.6	0.2/0.3	4,4,0	−4.8	0.8	−7.3	0.4	−4.77	0.23	0.4	10
C31	I, PP	97.3	0.5/1.6	1,5,0	0.0	0.4	−3.3	0.4	−3.28	0.23	0.4	10
	relict ol	97.8	n.a.	1,0,0	−3.5	0.5	−7.1	0.6	−5.3	0.4	n.a.	10
	relict ol	97.6	n.a.	1,0,0	−5.2	0.5	−8.9	0.6	−6.2	0.4	n.a.	10
C22	II, PO	69.7	9.3/9.4	7,0,0	2.3	0.4	−1.0	0.3	−2.21	0.15	0.3	15
C17	II, PO	74.4	5.2/4.5	5,0,0	2.8	0.3	−0.6	0.3	−2.09	0.25	0.5	15
	relict ol	99	n.a.	1,0,0	−7.1	0.1	−10.2	0.3	−6.5	0.3	n.a.	15
	relict ol	76.6	n.a.	1,0,0	−3.7	0.1	−7.0	0.3	−5.1	0.3	n.a.	15
	relict ol	71.4	n.a.	1,0,0	0.5	0.1	−2.6	0.3	−2.9	0.3	n.a.	15
Heterogeneous
C9	I, GO	99.4	0.1/0.1	7,0,0	3.5	0.4	1.7	0.3	−0.15	0.20	0.4	10
		99.3	n.a.	0,1,0	−11.3	0.4	−13.7	0.4	−7.8	0.4	n.a.	10
C15	II, PO	98.5	n.a.	1,0,0	−6.6	0.1	−9.8	0.2	−6.3	0.2	n.a.	15
		98.3	n.a.	1,0,0	−3.4	0.1	−6.6	0.2	−4.8	0.2	n.a.	15
		97.7	n.a.	1,0,0	−2.6	0.1	−5.9	0.2	−4.5	0.2	n.a.	15
		91.3	n.a.	1,0,0	−2.4	0.1	−5.4	0.2	−4.1	0.2	n.a.	15
		77.3	n.a.	1,0,0	1.2	0.1	−2.7	0.2	−3.3	0.2	n.a.	15
		66.9	n.a.	1,0,0	1.6	0.1	−2.4	0.2	−3.3	0.2	n.a.	15
		63.4	n.a.	1,0,0	2.5	0.1	−1.0	0.2	−2.3	0.2	n.a.	15
		58.9	n.a.	1,0,0	1.9	0.1	−2.2	0.2	−3.2	0.2	n.a.	15

^aMg# = Mg/(Fe+Mg) molar % of olivine and/or pyroxene in each chondrule.

^bUncertainties represent the range in measured Mg#’s of olivine and/or pyroxenes.

^cNumbers of mineral phases analyzed (olivine, low-Ca pyroxene, and high-Ca pyroxene).

Fig. 6.

Δ¹⁷O values of individual host chondrules and isolated olivine grains in Paris. Relict olivine (ol.) and low-Ca pyroxene (Lpx) grains are shown together. Data are sorted according to the host Δ¹⁷O values. Chondrule C9 (GO) contains homogeneously ¹⁶O-poor olivine, which is likely relict. A single analysis of pyroxene would represent melt oxygen isotope ratios that show lower Δ¹⁷O compared to other host chondrule values. For the heterogeneous chondrule C15 (II PO) and isolated Fe-poor olivine grain G24, individual analyses are shown. Error bars represents the propagated uncertainties for host chondrules and olivine in C9 and external reproducibility of individual analyses for relict grains, pyroxene in C9, and all data in C15 and G24.

In chondrule C18, five olivine grains display variable Δ¹⁷O value of −1.3±0.6‰ (2SD), which differ significantly from three pyroxene analyses (−3.3±0.3‰; Fig. 1a). We consider pyroxene data to represent the host chondrule value and olivine grains to be relict grains, as was discussed in previous studies (chondrule Y22 in Tenner et al., 2013; chondrule A6 in Chaumard et al., 2018). Similarly, chondrule C8 contains dusty olivine grains (SIMS spot #252, EA2 and EA5) with a Δ¹⁷O value of 0.1±0.4‰, which are significantly different from the host chondrule Δ¹⁷O value of −5.4±0.2‰ that are calculated from five olivine and pyroxene analyses (Table 1). In chondrule C9, seven olivine grains display homogeneous isotope ratios with the mean Δ¹⁷O value of −0.15±0.43‰ (2SD), which differ significantly from a single analysis of pyroxene (−7.8‰; Fig. 2e). All olivine grains in C9 contain numerous micron-sized inclusions of metal, which suggests they are dusty olivine grains and thus are likely to be relict. The pyroxene data are more likely to represent the host value. However, since there is only a single pyroxene analysis, we do not assign it to be host Δ¹⁷O value.

We also did not assign a host Δ¹⁷O value to the type II chondrule C15 (Fig. 4a) that show a large range of δ¹⁸0 and δ¹⁷O values between PCM and CCAM lines from −6.6‰ to +2.5‰ and from −9.8‰ to −1.0‰, respectively, among 8 olivine analyses. Chondrule C15 contains forsteritic olivine grains in its core (F0_91-99), indicating that these grains represent relict grains (based on FeO contents, Fe/Mn ratios; Berlin et al., 2011; Frank et al., 2014; Schrader and Davidson, 2017). This is consistent with the lower Δ¹⁷O values (−4.1‰ to −6.3‰) of forsteritic olivine grains compared to those of the more FeO-rich olivine grains with variable Δ¹⁷O values (−3.3‰ to −2.3‰).

Host Δ¹⁷O values from most of the type I chondrules vary continuously from −7.0±0.2‰ to −3.3±0.3‰ (Fig. 6). The two other type I chondrules (C11; Fig. 1d, and C12) and the two type II chondrules (C17 and C22) display a narrow range of host Δ¹⁷O values ranging from −2.5±0.2‰ to −2.1±0.2‰ (Fig. 6). Most of the relict olivine have Δ¹⁷O values between ~−2‰ and −8‰ (Fig. 6) overlapping those of the host values of other chondrules. Similar observation have been made in other carbonaceous chondrites such as Acfer 094, Murchison, Y-82094, CV, CO, and CR chondrites (Rudraswami et al., 2011; Ushikubo et al., 2012; Schrader et al., 2013, 2014, 2017; Tenner et al., 2013, 2015, 2017; Hertwig et al., 2018; 2019a; Chaumard et al., 2018). Only two relict grains have Δ¹⁷O values lower than −8‰; the relict grains in chondrule C1 (−9.5‰) and C14 (−13.0‰) (Fig. 6). Among the 29 relict grains analyzed, 13 (only olivine) display Δ¹⁷O values between −1.6‰ and 0.2‰ (Fig. 6). Most of these were found in the chondrule C9 that contains 7 relict olivine grains with Δ¹⁷O values ranging from −0.4‰ to 0.2‰ (Fig. 2e), and in chondrule C18 that contains 5 relict olivine grains with Δ¹⁷O values between −1.6‰ to −0.8‰ (Fig. 1a).

3.2.2. Oxygen isotope ratios of isolated grains and AOA

Table 2 shows the oxygen isotope ratios of three isolated grains and AOA. Four analyses of FeO-rich olivine grains G32 and G33 are homogeneous with Δ¹⁷O values of −2.7±0.5‰ and −1.9±0.3‰, respectively (Fig. 4b-c). These analyses are in agreement with type II chondrules C17 and C22. In FeO-poor olivine G24, 8 analysis of olivine are widely distributed close to PCM line with δ¹⁷O and δ¹⁷O values from −3.2‰ to +3.3‰ and from −5.5‰ to +1.7‰, respectively. The range of Δ¹⁷O values is from −4‰ to 0‰. As shown in Fig. 4d, the core of G24 is ¹⁶O-poor (Δ¹⁷O: −0.8‰ to 0.0‰) compared to the coarse rim (−2‰ to −4‰).

Table 2.

Mg#’s and O-isotope ratios of isolated olivine grains and aggregate olivine inclusion.

Object	Type, texture	Mg#	+/−	n (ol)	δ18O	unc.	δ17O	unc.	Δ17O	unc.	Δ17O 2SD	Beam (μm)
G32	Isolated FeO-rich olivine	72.2	2.3/5.7	4	2.2	0.7	−1.6	0.5	−2.7	0.5	0.8	15
G33	Isolated FeO-rich olivine	37.3	11.4/7.7	4	1.8	0.9	−0.9	0.5	−1.9	0.3	0.3	15
G24	Isolated FeO-poor olivine	99.4	n.c.	1	3.3	0.3	1.7	0.5	0.0	0.5	n.a.	10
		99.4	n.c.	1	2.0	0.3	0.4	0.5	−0.6	0.5	n.a.	10
		99.4	n.c.	1	−1.8	0.3	−4.0	0.5	−3.1	0.5	n.a.	10
		99.4	n.c.	1	−3.2	0.3	−5.5	0.5	−3.8	0.5	n.a.	10
		99.3	n.c.	1	2.6	0.3	0.6	0.5	−0.8	0.5	n.a.	10
		99.3	n.c.	1	0.2	0.3	−2.2	0.5	−2.3	0.5	n.a.	10
		99.3	n.c.	1	−3.1	0.3	−5.4	0.5	−3.9	0.5	n.a.	10
I3	AOI	99.5	0.1/0.0	2	−45.5	0.5	−47.3	0.4	−23.7	0.2	0.4	10

We analyzed four spots in olivine from AOA I3. While we rejected two analyses that show numerous large cavities in their SIMS spots (EA2), the mean of δ¹⁷O, δ¹⁷O, and Δ¹⁷O values of the two remaining analyses of olivine in AOA I3 are −45.5±0.5‰, −47.3±0.4‰, and −23.7±0.2‰, respectively.

3.3. Chondrule Mg#

Following the method of Ushikubo et al. (2012) and Tenner et al. (2013, 2015, 2017), we calculated the “host chondrule Mg#” in taking a mean value of olivine and/or low-Ca pyroxene excluding relict grains of each chondrule investigated. These values are shown in Table 1. Uncertainties of host chondrule Mg# are defined so that they represent the range of Mg#s in each chondrule, by taking differences between maxima, or minima, and the mean Mg# of olivine and low-Ca pyroxene, respectively. The host chondrules Mg#s of type I chondrules show a narrow range between 98.6 and 99.6, except for the three type I PP chondrules C31, C11 and C12, with Mg# of 97.3, 93.3 and 93.6, respectively. The host chondrule Mg#s for the type II chondrules are calculated to be ~70 by excluding obvious relict forsteritic olivine.

4. DISCUSSION

4.1. A single or two separate isotope reservoir(s)?

The Δ¹⁷O values of host chondrules and isolated olivine grains (n=30) are shown in Fig. 6 in the sequence of decreasing Δ¹⁷O values from −1.9 ±0.3‰ to −7.0 ±0.2‰. The majority of type I chondrules show a narrow range of Δ¹⁷O between −6‰ and −4‰, while all FeO-rich chondrules and isolated olivine grains have Δ¹⁷O between −3‰ and −2‰. This is very similar to the results from Murchison, where the Δ¹⁷O values of high Mg# (>98) chondrules range from −6‰ to −4‰ and those of lower Mg# (96-65) chondrules are between −3‰ and −2‰ (Chaumard et al., 2018). Compared to Murchison data, Paris chondrule data are more continuous, without a gap between −4‰ and −3‰ and extend to lower Δ¹⁷O ~ −7‰.

In Fig. 7, the host chondrule Δ¹⁷O values in Paris are shown against their Mg#s along with data from Murchison (Chaumard et al., 2018). The majority of type I chondrules in Paris plot at highest Mg#s ~99 with Δ¹⁷O from −6‰ to −4‰, while the rest of data plot on a trend where Δ¹⁷O values increase with decreasing Mg#s. In Murchison, chondrules with high Mg#s (>98.5) display host Δ¹⁷O values ranging from −6.0‰ to −4.1‰ while chondrules with lower Mg# (~96–65) have host Δ¹⁷O values of ~−2.5‰. Thus, the chondrule Mg#-Δ¹⁷O relationship in Paris is nearly the same as that observed for Murchison. Indeed, the two isolated FeO-rich olivine grains analyzed in Paris plot along and extend the trend defined by type II chondrules (in both Paris and Murchison), towards lower Mg#s (~37 for the grain G33) than was previously measured in Murchison. One PP (C31) and one POP (C18) chondrule with host Δ¹⁷O values of −3.3‰ also plot between Mg# ~96–99 and two other PP chondrules (C11, C12) with Δ¹⁷O values of ~−2.5‰ have Mg#s ~93-94, a range of Mg#s not found in Murchison (Chaumard et al., 2018). Thus, by combining data from two CM chondrites, CM chondrule data define a single continuous trend of increasing Δ¹⁷O values with decreasing Mg#s. Marrocchi et al. (2018) also reported similar host chondrule Δ¹⁷O values and Mg# data for three chondrules in NWA 5958, which is a CM-related ungrouped carbonaceous chondrite.

Fig. 7.

Δ¹⁷O values of individual host chondrules in Paris vs. Mg#s (open diamonds). Data from the Murchison CM2 chondrite (black and white squares; Chaumard et al., 2018) are shown for comparison. Chondrule Mg# uncertainties correspond to the range of measured values, while uncertainties in Δ¹⁷O are the propagated 2SE.

The Mg#-Δ¹⁷O relationship among chondrules observed in CM chondrites is similar to those obtained from Acfer 094 (Ushikubo et al., 2012) and the Y-81020 CO chondrite (Tenner et al., 2013), as shown in Fig. 8, and to a lesser extent for chondrules in the Y-82094 ungrouped carbonaceous chondrite and CV chondrites (Tenner et al., 2017; Hertwig et al., 2018; 2019a). However, chondrules in Acfer 094 and Y-81020 display a well-defined bimodal distribution of Δ¹⁷O at ~ −5‰ and −2.5‰ (Fig. 8). Ushikubo et al. (2012) and Tenner et al. (2013) argued that these chondrules formed in two distinct isotope reservoirs; (1) a reducing environment at lower dust enrichments (100 × solar composition gas) in which chondrules with Mg#>97 and Δ¹⁷O ~ −5‰ formed and (2) an oxidizing environment showing higher dust enrichments (1000× solar) in which chondrules with Mg# <97 chondrules and Δ¹⁷O ~ −2.5‰ formed. They interpreted the difference in oxygen isotope ratios between the two reservoirs as a result of varying amounts of ¹⁶O-poor H₂O ice mixed in with other chondrule precursor solids.

Fig. 8.

Δ¹⁷O values vs. Mg#s of individual host chondrules in CM (open diamonds; Chaumard et al., 2018; this work), CO (grey circles; Tenner et al., 2013), and Acfer 094 (black circles; Ushikubo et al., 2012) chondrites. Chondrule Mg# uncertainties correspond to the range of measured values, while uncertainties in Δ¹⁷O are the propagated 2SE.

Chondrule data from Acfer 094 and Y-810202 overlap almost exactly with the CM chondrule trend (Fig. 8), so that the two separate isotope reservoirs indicated from earlier studies could represent subsets of a single chondrule-forming region, which could have been shared by many CC chondrules. Chondrules in the ungrouped carbonaceous chondrite Y-82094 also show a similar Mg#-Δ¹⁷O relationship that overlaps with that of CM chondrites. In CV chondrites, chondrules generally have high Mg#s (>98) with Δ¹⁷O values predominantly ranging from −6‰ to −4‰ (Hertwig et al., 2018; 2019a; Marrocchi et al., 2019). These values overlap exactly with the CM chondrule data for Mg#>98. In all these CCs, type I chondrules with Mg#≤97 and Δ¹⁷O values between −4‰ and −3‰ are scarce. The Paris chondrule C31 (PP) with Mg# ~97.3 (+0.5/−1.6) shows intermediate Δ¹⁷O = −3.3±0.2‰ and is located in the gap between the two groups of chondrules (Δ¹⁷O ~−5‰ and ~−2.5‰ on the Δ¹⁷O versus Mg# plot, Fig. 8), so is one of the three chondrules studied by Marrocchi et al. (2018) in NWA 5958 with Mg# ~95 and Δ¹⁷O of −3‰. Identifying more chondrules with intermediate Mg#s and Δ¹⁷O values would further test if CC chondrules derived from a single continuous region, or if they were derived from two separate reservoirs in terms of time and locations, as argued earlier from Acfer 094 and CO chondrite chondrule data (Ushikubo et al., 2012; Tenner et al., 2013).

4.2. Nature of the oxygen isotope reservoir(s)

Tenner et al. (2015) presented a mass balance model to describe the observed negative correlation between Δ¹⁷O values and Mg#s among CR chondrules. As discussed by Tenner et al. (2015), Mg#s of olivine and pyroxene in chondrules mainly depend on the redox state of iron that in turn was controlled by the oxygen fugacity of the chondrule-forming environment. Equilibrium thermodynamic calculations (e.g., Ebel and Grossman 2000) indicate that type I chondrules with high Mg#s formed under reducing conditions, e.g., fO₂ ~3.5 to ~2.5 log units below the IW buffer for Mg# 99 and 96, respectively (Tenner et al., 2015). In the protoplanetary disk, H₂O ice would reside in fine-grained dust particles enriched in organic matter and amorphous silicates, similar to chondritic porous interplanetary dust particles and the matrix in primitive meteorites (e.g., Bradley and Brownlee, 1986; Abreu et al., 2010). Therefore, the model supposes that the chondrule-forming region consisted of varying relative proportions of these components. In the scope of the model, H₂O ice, anhydrous dust, organics, and solar-composition gas possessed Δ¹⁷O values of 5.1‰, −5.9‰, 11.3‰, and −28.4‰, respectively (Tenner et al., 2015).

The model assumes that chondrules formed in an environment whose effective oxygen isotopic composition and redox state is mostly controlled by the mixture of these four components. During transient heating, this environment comprised the solar-composition gas enriched in the evaporated portions of the precursors (= ambient gas), a melt component, and unmelted solids. Being a mixing model, no predictions are made about the actual physical processes operating during chondrule formation. It is implied by the model, however, that after evaporation, recondensation into the melt (e.g., Libourel et al., 2006; Nagahara et al., 2008; Libourel and Portail, 2018) lead to an effective oxygen exchange between melt and ambient gas (e.g., Kita et al., 2010; Ushikubo et al., 2012; Tenner et al., 2015; Marrocchi and Chaussidon, 2015; Marrocchi et al., 2019), so that melt and ambient gas display the same effective Δ¹⁷O value at the time of olivine crystallization. The Mg#s of olivine and pyroxene in chondrules would be determined by the oxygen fugacity of the ambient gas, which can be estimated from the relative abundances of H, C, and O of the mixture of ambient gas and the melt component. Further details of the mass balance model are described in Tenner et al. (2015; 2018).

Using the model calculation, Tenner et al. (2015) suggested that the majority of type I chondrules in CR chondrites formed at a dust enrichment factor of 100-200 relative to solar-composition gas and with a H₂O-ice enhancement of 0-0.8 times relative to CI-composition dust. Subsequently, Hertwig et al. (2018) modified the model parameters by applying Δ¹⁷O values of the anhydrous dust and H₂O ice of −8‰ and +2‰, respectively, in order to explain the distribution of Mg#s and Δ¹⁷O values of chondrules in CV chondrites. These parameters were estimated based on the assumptions that (1) the lowest host-chondrule Δ¹⁷O value is representative of the anhydrous silicate dust component and (2) the average Δ¹⁷O value of type II chondrules constrain the Δ¹⁷O values of H₂O ice based on the oxygen isotope mass balance at high dust enrichments (Tenner et al., 2015). In Paris, the lowest host Δ¹⁷O value measured is −7.0‰, while the type II chondrules and FeO-rich olivine G33 reach Δ¹⁷O values as high as ~ −2‰. Consequently, the Δ¹⁷O values for anhydrous silicate dust and H₂O ice used by Hertwig et al. (2018) are useful to estimate both the dust and H₂O enrichments of the precursors during the formation of CM chondrules. Applying the model of Tenner et al. (2015) and the calculations of Hertwig et al. (2018), Mg# ~99 chondrules in Paris correspond to dust enrichment factors of 50-100× relative to solar-composition gas and to amounts of H₂O in the dust ranging from 0 (anhydrous dust) to 0.8× relative to the nominal abundance of H₂O ice in the CI dust (Fig. 9). For chondrules with Mg#s lower than 98 and down to mostly 60-70, the ice enrichment factor is roughly constant (between 0.8 and 1× the nominal abundance of H₂O ice, relative to the CI dust) while the dust enrichment factor increases from ~100× to ≥1,000× (Fig. 9). We note that Mg# should be controlled not only by oxygen fugacity, but also by iron abundance of the system relative to Mg and Si, which is assumed to be CI chondritic in the model. Chondrules with lower Mg# (<60 down to ~35) might have been formed from precursors that were significantly enriched in iron and may not represent chondrule formation under extremely high dust-enrichments (>3,000×).

Fig. 9.

Δ¹⁷O values of individual host chondrules in the CM chondrites Paris (this work) and Murchison (Chaumard et al., 2018) and Mg#s superimposed by oxygen isotope mixing curves of constant dust enrichment and ice enhancement from Tenner at el. (2015) and Hertwig et al. (2018). In this model, the anhydrous silicate dust, Solar gas, water ice, and organics in the dust are considered to have Δ¹⁷O values of −8.0‰, −28.4‰, +2.0‰, and +11.3‰, respectively (Hertwig et al., 2018). Chondrule Mg# uncertainties correspond to the range of measured values, while uncertainties in Δ¹⁷O are the propagated 2SE.

The observed relationship between Δ¹⁷O values and Mg#s from chondrules in CM chondrites might have resulted from their formation in a single large region in the protoplanetary disk that was radially zoned with respect to the effective Δ¹⁷O values, possibly due to variable abundances of ¹⁶O-poor H₂O ice. The region would likely have existed near to the snow line, the condensation front of H₂O ice. Inside the snow line, the relative abundance of ¹⁶O-poor H₂O ice among the solid chondrule precursors was low in contrast to outside of the snow line (e.g., Morbidelli et al., 2016). The most reduced chondrules (Mg#s >99) formed inside the snow line, within this large region, where dust enrichment factors were up to ~100×. More oxidized chondrules with higher Δ¹⁷O values and lower Mg#s would have formed towards the external part of this single chondrule-forming region where there was addition of ¹⁶O-poor H₂O ice to the nearly anhydrous chondrule precursors and/or an increase of the dust enrichment factor from ~ 100× to ~1,000×, as proposed for CR chondrites by Tenner et al. (2015).

Alternatively, systematic changes in Mg#-Δ¹⁷O could have developed over time. Most type I chondrules with Mg#>97 and lower Δ¹⁷O values ~ −5‰ formed in regions mainly inside of the snow line with dust enrichments lower than ~ 200×. The local disk temperatures would decrease with time and, consequently, the snow line could migrate through the chondrule forming regions. At the same time, the dust layer became thicker prior to the formation of asteroidal bodies (e.g., Alexander et al., 2008). Later-forming chondrules would form under oxidizing environments and with higher Δ¹⁷O values ~ −2‰ due to addition of ¹⁶O-poor H₂O ice to chondrule precursors.

Hartlep and Cuzzi (2020) estimated dust enrichment factors of the turbulent disk midplane to be typically −100× where cm-sized pebbles formed by streaming instability. They further discussed that chondrules and their precursors existed in the form of cm-sized pebbles and argued that the estimated particle concentrations are consistent with those of high Mg# chondrules, which are abundant in carbonaceous chondrites. Their model also predicts that particle concentrations may reach −1,000×, but probability is low. Therefore, in principle, formation of type II chondrules under higher dust-enrichment factors of −1,000× might have occurred in the protoplanetary disk.

Many type I POP chondrules show a mineralogical zoning with olivine grains being located in the core and pyroxene in periphery of individual chondrules (e.g., Krot et al., 2004; Hezel et al., 2006; Friend et al., 2016). Further, Villeneuve et al. (2020) observed a large mass-dependent Si isotope fractionation in type I chondrules, especially in those with high Mg#s. Both observations provide evidence for SiO molecules condensing from the ambient gas to the melt. Based on detailed CL-mapping of chondrule minerals and correlated oxygen isotope systematics, Marrocchi et al. (2019) proposed that SiO (along with Mg atoms) condensing from an ¹⁶O-poor nebula gas (Δ¹⁷O ~0‰) interacted with ¹⁶O-rich AOA-like chondrule precursors (Δ¹⁷O ~ −20‰) to form type I chondrules with Δ¹⁷O values ranging from −6‰ to −3‰. However, 2-3 Ma after CAIs, ambient temperatures of the protoplanetary disk should have been significantly lower than the condensation temperature of Si (e.g., Desch et al., 2018). Hence, the partial pressure of SiO in the nebula gas is expected to be low, prior to the heating events causing chondrule formation. Instead, heating and evaporation of solid precursors, such as silicates, during chondrule formation would produce the majority of gaseous SiO, residing in the ambient gas (e.g., Nagahara et al., 2008) and being available for exchange and interaction with the chondrule melt and solids. Thus, the observed correlation of Δ¹⁷O values and Mg#s is better explained by the contribution of ¹⁶O-poor H₂O ice (or icy fine-grained dust) rather than addition of ¹⁶O-poor SiO gas from nebula gas.

4.3. Origin of relict grains in CM chondrules

Some chondrules contain grains predating the host minerals that crystallized from the final chondrule melts. These grains are defined as relict grains and can be identified chemically and/or isotopically (e.g., Jones et al., 2004; Kunihiro et al., 2004; Krot et al., 2006; Rudraswami et al., 2011; Ushikubo et al., 2012; Schrader et al., 2013; Tenner et al., 2013; Schrader and Davidson, 2017; Hertwig et al., 2018; 2019a; Chaumard et al., 2018; Marrocchi et al., 2018; 2019). Ten of the 27 chondrules in this study contain relict olivine and/or low-Ca pyroxene grains (Table 1, Fig. 6). The abundance of relict-grain bearing chondrules (~37%) in Paris is similar to Murchison (CM2) (~38%; Chaumard et al., 2018), Acfer 094 (~ 45%; Ushikubo et al., 2012), and the Y-81020 CO chondrite (~ 42%; Tenner et al., 2013). We note that the abundance of relict-grain bearing chondrules in these chondrites is probably underestimated due to the limited number of analyses per chondrules (≤8), in comparison to the 30-50 analyses per chondrules performed in other studies (Marrocchi et al., 2018; 2019). While eight out of the nine relict grains in Murchison chondrules were ¹⁶O-rich compared to their host chondrules (Chaumard et al., 2018), four out of 10 chondrules in Paris contain relict olivine grains with Δ¹⁷O values higher than their host chondrules, which reach as high as 0‰. This difference might be due to selection bias for chondrules in Paris in this work (or in Murchison in previous work) and may not be statistically significant. Similar ¹⁶O-poor relict olivine grains have also been reported from other CCs but seem to be not as common (Ushikubo et al., 2012; Tenner et al., 2013; Hertwig et al., 2018; 2019a; Marrocchi et al., 2018; Schrader et al., 2020).

In type II chondrules, forsteritic cores in FeO-rich olivine phenocrysts are easily recognized as relict grains (e.g., Jones, 1990; Wasson and Rubin, 2003; Ruzicka et al., 2007, 2008). These observations indicate that at least a part of the type II chondrule precursors formed in more reducing conditions, similar to those during formation of type I chondrules (e.g., Ruzicka et al., 2008). Based on re-heating experiments of type I precursor materials at 1450–1500°C under oxidizing conditions (between the IW and NNO buffers), Villeneuve et al. (2015) proposed that type I chondrules (or fragments) could have been the main precursor material of type II chondrules. These results and the recognition of ¹⁶O-rich relict grains in the type II chondrule C17 with Δ¹⁷O values similar to those of host type I chondrules (Fig. 6) support this precursor origin of type II chondrules. There are many other examples where type II chondrules enclose FeO-poor olivine grains that are similar to those observed in type I chondrules (Kunihiro et al. 2004; 2005; Ushikubo et al., 2012; Tenner et al. 2013; 2017; Krot and Nagashima, 2017; Krot et al., 2018). Villeneuve et al. (2020) reported negative δ³⁰Si in relict olivine from type II chondrules in carbonaceous chondrites, indicating they could have originated from forsteritic olivine of type I chondrules.

Thirteen relict grains from 7 chondrules have Δ¹⁷O values within the range of the host values calculated for chondrules (from −7.0‰ to −2.1‰; Fig. 6). Including the ¹⁶O-rich relict olivine in type II chondrules, it had been suggested that most relict grains were result of mixing between two distinct major isotope reservoirs (e.g., Jones et al., 2005; Hewins and Zanda, 2012; Ushikubo et al., 2012; Tenner et al., 2013, 2015). The new dataset for Paris chondrules instead may point to mixing within a single chondrule-forming region. Within this region, the radial transport of solids would have occurred (Cuzzi et al., 2010) between an inner area, enriched in reduced and ¹⁶O-rich chondrules, and an outer area, enriched in more oxidized and ¹⁶O-poor chondrules.

The remaining relict grains are significantly ¹⁶O-rich (Δ¹⁷O: −13‰ and −9.5‰) or ¹⁶O-depleted (Δ¹⁷O: from −1.9‰ to 0.2‰; Fig. 6). Similar ¹⁶O-rich relict grains were also reported in Murchison chondrules, reaching values of down to −18‰ (Chaumard et al., 2018). The ¹⁶O-rich relict grains might be related to CAI and AOA-like refractory precursors (Ushikubo et al., 2012; Marrocchi et al., 2019), which formed during the earliest stage of the Solar System evolution (≤0.2 Ma; Kita et al., 2013). However, relict olivine grains in some chondrules (C1 and C14) do neither show any special textural nor major-element compositional difference compared to other coexisting (non-relict) olivine in the same chondrules. Identification of small differences in minor elements would require detailed elemental mapping of olivine at a high electron-beam intensity, as demonstrated by Marrocchi et al. (2018; 2019) who found that relict grains tend to have lower Ca, Ti, and Al concentrations. In contrast, chondrules C8 and C9 and the isolated olivine grain G24 all show dusty olivine textures and Δ¹⁷O values ~ 0‰, which are significantly higher than other olivine analyses in the same objects (Table 1). Schrader et al. (2020) reported ¹⁶O-poor dusty olivine with Δ¹⁷O ~ 0‰ in four type I chondrules from Murchison and Murray CM chondrites and suggested that precursors of dusty olivine in CM include those unrelated to type I and II chondrules, but originated instead from an unequilibrated ordinary chondrite source. Dusty olivine grains are thought to form by the reduction during partial melting of more FeO-rich olivine, derived from a previous generation of chondrules (Nagahara, 1981; Rambaldi, 1981; Jones and Danielson, 1997; Leroux et al., 2003). Consequently, their oxygen isotope ratios likely reflect those of the FeO-rich precursors. Our results from C8, C9 and G24 are consistent with those of Schrader et al. (2020) and imply an origin of these relict grains from a different reservoir than the one in which type I (Δ¹⁷O: ~−5‰) and type II (Δ¹⁷O: ~−2.5‰) chondrules in CM and other groups of CCs formed.

The ¹⁶O-poor relict grains may be related to OC-like chondrules in CCs that show homogeneous oxygen isotope ratios with Δ¹⁷O ~0‰ (e.g., Tenner et al., 2017). Tenner et al. (2017) argued that three type II chondrules in Y-82094 (ungroup C) are characterized by intermediate Mg#s (80-90), higher MnO contents in olivine, and oxygen isotope ratios on the terrestrial fractionation line but to the left side of the PCM line, which are similar to chondrules in ordinary chondrites. Recently, Williams et al. (2020) and Schneider et al. (2020) obtained coordinated ε⁵⁴Cr, ε⁵⁰Ti, and SIMS oxygen isotope analyses of individual chondrules from multiple chondrite groups and found that chondrules generally show isotope signatures similar to those of their host bulk meteorites. However, Williams et al. (2020) also found three BO chondrules in Allende (CV) with Mg#s of 80-90 that show negative ε⁵⁴Cr and ε⁵⁰Ti values and Δ¹⁷O =0‰, which are very different from those of the bulk Allende meteorite. These chondrules do not exactly match ordinary chondrite chondrules, but show similar isotope ratios to achondrites.

The Al-Mg ages of two OC-like chondrules in Acfer 094 are older by 0.4-0.8 Ma than other chondrules in the same meteorite (Ushikubo et al., 2013; Hertwig et al., 2019b). The age difference could represent the transit time for these chondrules to travel from the outside of the CC chondrule forming regions to the Acfer 094 accretion region. Schrader et al. (2020) and Williams et al. (2020) suggested that dusty olivine in CM chondrite chondrules and the FeO-rich BO chondrules in Allende, respectively, had migrated outward from inner to outer disk across the Jupiter’s gap. Collectively, both refractory ¹⁶O-rich precursors and OC-like ¹⁶O-poor precursors represent inner disk solids that would have been added to the solid precursors of CM chondrules.

4.4. Oxygen isotope ratios of AOA in CM

Here we compare oxygen isotope ratios of AOA I3 in Paris obtained during this work to AOAs in other CCs and refractory inclusions in CM chondrites, which were previously obtained in the WiscSIMS laboratory. The mean of 2 analyses of AOA I3 gives δ¹⁷O = −45.5±0.5‰, δ¹⁷O = −47.3±0.4‰, and Δ¹⁷O = −23.7±0.2‰ (Table 1). This is in agreement with AOA analyses in Acfer 094 (mean of four AOA and 2SD; δ¹⁷O = −45.8±0.7‰, δ¹⁷O = −47.7±0.7‰, and Δ¹⁷O = −23.8±0.6‰; Ushikubo et al., 2017) and those in DOM 08006 (CO3.0; δ¹⁷O = −45.8±0.7‰ and Δ¹⁷O = −24.0±0.4‰; Fukuda et al., 2021). Data from I3 are also consistent with spinel-hibonite inclusions (SHIBs) in Murchison studied by Kööp et al. (2016) with Δ¹⁷O = −23.4±1.1‰. Other studies reporting oxygen isotope ratios of AOA from multiple CCs (Fagan et al., 2004; Krot et al., 2005; 2014; Komatsu et al., 2017), Kakangari (Nagashima et al., 2015), ordinary chondrites (Itoh et al., 2007; Ebert et al., 2020), and enstatite chondrite (Guan et al., 2000) also show similar ranges. Thus, AOA I3 formed in a homogeneous ¹⁶O-rich isotope reservoir that formed a wide variety of CAIs (e.g., Ushikubo et al., 2017).

4.5. Implications for the chondrule formation region

The large chondrule formation region discussed in 4.2. would likely be located at around 3 AU, where the snow line would be located at the time of chondrule formation (a few Ma after CAIs; Morbidelli et al., 2016). This region would have been outside of the OC chondrite accretion region, so that a minor fraction of inner disk materials might have been transported from the inner disk to the outer disk (Hertwig et al., 2018; Schrader et al., 2020; Williams et al., 2020). Several recent studies suggest that CM and other CCs form outside of proto-Jupiter in order to explain the isotope dichotomy between carbonaceous meteorites and non-carbonaceous meteorites (e.g., Kruijer et al., 2017; Desch et al., 2018; Nanne et al., 2019; Van Kooten et al., 2020), a hypothesis first advocated based on bulk meteorite nucleosynthetic anomalies of ⁵⁴Cr and ⁵⁰Ti by Warren (2011). Kruijer et al. (2017) further suggested that the isolation of carbonaceous and non-carbonaceous isotope reservoirs occurred early, within 1 Ma after CAI formation. This predates the events forming CC chondrules which are estimated to have occurred 2-3 Ma after CAI formation (e.g., Ushikubo et al., 2013; Nagashima et al., 2017; Hertwig et al., 2019b). Desch et al. (2018) proposed a comprehensive evolutionary model involving the proto-Jupiter’s gap that predicts a location and formation time for each meteorite parent body in order to explain the chemical diversity among meteorites. In their model, Jupiter originally formed ≥4 AU and migrated inward to 3 AU to open a gap that separated the two isotope reservoirs. Within this framework, chondrules in CCs should form beyond the snow line, which is in contrast to the dry environments expected for the abundant CM chondrules with high Mg# (>98) and Δ¹⁷O ~−5‰. Furthermore, complete isolation of carbonaceous and non-carbonaceous meteorite forming regions can not explain chondrules and relict grains with OC-like oxygen isotope ratios in CM chondrites. Schrader et al. (2020) argued that small fragments of UOC chondrules (<300 μm) could migrate outward beyond the Jupiter’s gap and so were incorporated to chondrule precursors for CM chondrites. A few chondrules in Paris contain dusty olivine grains with Δ¹⁷O ~0‰ (C8, C9, C24), which are also ~300 μm or smaller and could have migrated from UOC regions across the Jupiter’s gap. Chondrules with Δ¹⁷O ~0‰ in Acfer 094 and Y-82094 (ungrouped C) are also ≤300 μm in size, though those in CV are from ~ 500 μm (Hertwig et al., 2018; 2019) to ≥ 2 mm (Williams et al., 2020).

Alternatively, proto-Jupiter could have been located close to the current Jupiter orbit that is outside of the major CC chondrule-forming regions. Recent numerical simulation by Tanaka et al. (2020) indicated that Jupiter-sized giant planets would not experience a significant radial migration as opposed to those indicated in previous studies (e.g., Lin and Papaloizou 1986). If there were radial transport of solids in the protoplanetary disk (e.g., Cuzzi et al., 2010), a small amount of OC-like chondrules and their fragments could have migrated outward to CM and other major CC forming regions. The main drawback in this scenario is that inward drift of significant amounts of dust from the CC forming regions to the OC forming regions might have occurred. This would be in disagreement with the observed distinct isotope signatures between bulk carbonaceous and non-carbonaceous meteorites (e.g., Budde et al., 2016; Kruijer et al., 2017; Kleine et al. 2020).

CONCLUSIONS

In situ SIMS oxygen 3-isotope analyses of 29 chondrules and 3 isolated olivine grains in the least-altered CM chondrite Paris were performed. The results were used to estimate host chondrule Δ¹⁷O values. By combining this data set with that from Murchison chondrules (Chaumard et al., 2018), the Mg#- Δ¹⁷O relationship of CM chondrite chondrules was evaluated.

1. Host chondrule Δ¹⁷O values in CM chondrites increase continuously from −7‰ to −2‰ with decreasing Mg# from 99 to 37. The majority of types I and II chondrules in CM show high Mg#s ~99 with Δ¹⁷O from −6‰ to −4‰ and lower Mg#s of 60-70 with Δ¹⁷O of −2.5‰, respectively, while a few type I chondrules with lower Mg# (97-93) show Δ¹⁷O between −3.3‰ and −2.5‰, We suggest that bimodal distribution of chondrule Mg# and Δ¹⁷O from Acfer 094 and Y-81020 (CO3) previously reported by Ushikubo et al. (2012) and Tenner et al. (2013), respectively, are parts of the CM chondrule trend.

2. Relict grains in CM chondrite chondrules are either ¹⁶O-rich or ¹⁶O-poor relative to their host chondrules, with the total range from −18‰ to 0‰. The majority of relict grain Δ¹⁷O values overlap the range of host chondrule Δ¹⁷O values, suggesting they were derived from a precursor in the same CM chondrite chondrule forming region. Relict grains with Δ¹⁷O values <−8‰ and ~0‰ are likely from refractory precursors and OC-like precursors, respectively.

3. By adapting the oxygen isotope mass balance model of Tenner et al. (2015) and Hertwig et al. (2018), we argue that the majority of type I chondrules formed under relatively low dust enrichment factors (50-100× Solar) but with a range of ¹⁶O-poor H₂O ice (0-0.8× CI) in the precursors. Other chondrules formed under oxidizing environments with higher dust enrichments (100-1,000× Solar) and abundant ¹⁶O-poor H₂O ice (0.8-1× CI).

4. CM chondrite chondrule formation could have occurred in a large single region that spanned both sides of the snow line, facilitating a wide range of dust-enrichments and ¹⁶O-poor H₂O-ice enhancements. It is possible that the conditions in this region evolved with time, from reducing to more oxidizing, as the local disk became denser and colder. This chondrule forming region would have likely exist inside the proto-Jupiter orbit if Jupiter was large enough to produce a gap in the accretion disk.

5. In addition to chondrules, we also analyzed one AOA in Paris, and the data from this AOA are in good agreement with AOAs and other refractory inclusions from Acfer 094, DOM-08006 (CO3), and Murchison (CM2).