Correlated isotopic and chemical evidence for condensation origins of olivine in comet 81P/Wild 2 and in AOAs from CV and CO chondrites

Abstract

Magnesium stable isotope ratios and minor element abundances of five olivine particles from comet 81P/Wild 2 were examined by secondary ion mass spectrometry (SIMS). Wild 2 olivine particles exhibit only small variations in δ²⁵Mg values from −1.0 ^+0.4/_−0.5 ‰ to 0.6 ^+0.5/_{− 0.6} ‰ (2σ). This variation can be simply explained by mass-dependent fractionation from Mg isotopic compositions of the Earth and bulk meteorites, suggesting that Wild 2 olivine particles formed in the chondritic reservoir with respect to Mg isotope compositions. We also determined minor element abundances, and O and Mg isotope ratios of olivine grains in amoeboid olivine aggregates (AOAs) from Kaba (CV3.1) and DOM 08006 (CO3.01) carbonaceous chondrites. Our new SIMS minor element data reveal uniform, low FeO contents of ~0.05 wt% among AOA olivines from DOM 08006, suggesting that AOAs formed at more reducing environments in the solar nebula than previously thought. Furthermore, the SIMS-derived FeO contents of the AOA olivines are consistently lower than those obtained by electron microprobe analyses (~1 wt% FeO), indicating possible fluorescence from surrounding matrix materials and/or Fe,Ni-metals in AOAs during electron microprobe analyses. For Mg isotopes, AOA olivines show more negative mass-dependent fractionation (−3.8 ± 0.5‰ ≤ δ²⁵Mg ≤ −0.2 ± 0.3‰; 2σ) relative to Wild 2 olivines. Further, these Mg isotope variations are correlated with their host AOA textures. Large negative Mg isotope fractionations in olivine are often observed in pore-rich AOAs, while those in compact AOAs tend to have near-chondritic Mg isotopic compositions. These observations indicate that pore-rich AOAs preserved their gas-solid condensation histories, while compact AOAs experienced thermal processing in the solar nebula after their condensation and aggregation. Importantly, one ¹⁶O-rich Wild 2 LIME olivine particle (T77/F50) shows negative Mg isotope fractionation (δ²⁵Mg = −0.8 ± 0.4‰, δ²⁶Mg = −1.4 ± 0.9‰; 2σ) relative to bulk chondrites. Minor element abundances of T77/F50 are in excellent agreement with those of olivines from pore-rich AOAs in DOM 08006. The observed similarity in O and Mg isotopes, and minor element abundances suggest that T77/F50 formed in an environment similar to AOAs, probably near the proto-Sun, and then was transported to the Kuiper belt, where comet 81P/Wild 2 likely accreted.

1. Introduction

Particles collected from the Jupiter family comet 81P/Wild 2 (NASA Stardust mission) have provided unique opportunities to study materials from the outer Solar System, likely within the Kuiper Belt (Brownlee et al., 2006; 2012; 2014). Wild 2 particles consist of a variety of materials, including presolar grains (McKeegan et al., 2006; Messenger et al., 2009; Leitner et al., 2010, 2012) and organic matter (Sandford et al., 2006; Sandford, 2008; Matrajt et al., 2008) as well as numerous high temperature materials such as Ca, Al-rich inclusions (CAIs; e.g., McKeegan et al., 2006; Zolensky et al., 2006; Chi et al., 2009; Simon et al., 2008; Joswiak et al., 2012; 2017) and ferromagnesian silicates (Nakamura et al., 2008; Bridges et al., 2012; Joswiak et al., 2012; Nakashima et al., 2012; Ogliore et al., 2012; Frank et al., 2014; Gainsforth et al., 2015; Defouilloy et al., 2017). Determining the origins of such particles can shed light on dynamic properties of the solar protoplanetary disk, including transport and processing of comet precursors. For example, the existence of high temperature minerals in Wild 2 suggests that materials originating from the hotter inner part of the solar protoplanetary disk were transported outward to the cooler Kuiper Belt region and eventually accreted to comets (e.g., McKeegan et al., 2006; Ciesla, 2007; Nakamura et al., 2008; Brownlee et al., 2012). In addition, Nakashima et al. (2012) suggested that at least a fraction of the ferromagnesian silicates in comet Wild 2 formed in the outer regions of the asteroid belt, based on the observed similarity of oxygen isotope systematics between Wild 2 particles and chondrules from CR chondrites. Furthermore, Bridges et al. (2012) hypothesize an outer Solar System origin for some Wild 2 silicate particles that have similar oxygen isotope and elemental characteristics when compared to Al-rich chondrules from carbonaceous chondrites. Therefore, at least some Wild 2 silicate particles may also hold information about the physicochemical conditions of the outer Solar System where comets accreted.

In the past decade, oxygen three-isotope ratios of forty-seven Wild 2 silicate particles larger than 2 μm have been acquired by secondary ion mass spectrometry (SIMS) (McKeegan et al., 2006; Nakamura et al., 2008; Nakamura-Messenger et al., 2011; Bridges et al., 2012; Nakashima et al., 2012; Ogliore et al., 2012, 2015; Joswiak et al., 2014; Gainsforth et al., 2015; Defouilloy et al., 2017; Chaumard et al., 2018), and they show a bimodal distribution on an oxygen three-isotope diagram (Fig. 1). The mass-independent fractionation of oxygen isotopes among Wild 2 particles are expressed as Δ¹⁷O (= δ¹⁷O − 0.52 × δ¹⁸O, where δ^17,18O = [(^17,18O/¹⁶O)_sample/(^17,18O/¹⁶O)_VSMOW − 1] × 1000; VSMOW = Vienna Standard Mean Ocean Water; Baertschi, 1976), and range from −24‰ to +3‰ for all data with uncertainties smaller than 5‰. Among them, the Δ¹⁷O values of thirty-six Wild 2 particles (≥ 2 μm) range from −7‰ to +3‰ (McKeegan et al., 2006; Nakamura et al., 2008; Nakashima et al., 2012; Ogliore et al., 2012, 2015; Joswiak et al., 2014; Gainsforth et al., 2015; Defouilloy et al., 2017; Chaumard et al., 2018), similar to those of chondrules found in chondrites (Tenner et al., 2018a). In contrast, seven Wild 2 particles show more ¹⁶O-rich oxygen isotopic characteristics (Δ¹⁷O = −24‰ to −20‰). These seven particles are nearly pure forsterite or enstatite, some of which are low-iron, manganese-enriched (LIME) olivines and pyroxenes (Nakashima et al., 2012; Defouilloy et al., 2017; Chaumard et al., 2018). LIME olivines and pyroxenes (hereafter LIME silicates) have MnO/FeO (wt%) ratios much higher than 0.1 and low FeO contents (typically ≤ 1 wt%). As Mn is not stable as metal at solar nebula conditions, it likely condensed at ~1100 K, as Mn₂SiO₄ in solid solution with forsterite, and as MnSiO₃ with enstatite (Larimer, 1967; Wai and Wasson, 1977; Lodders, 2003). In contrast, Fe co-condenses as metal with forsterite (if the oxidation state is below the Fe-FeO buffer), meaning that high MnO/FeO ratios observed in LIME silicates suggest a condensation origin from the solar nebula (Klöck et al., 1989; Weisberg et al., 1993; 2004; Ebel et al., 2012). LIME silicates have been found in chondritic interplanetary dust particles (IDPs) (Klöck et al., 1989) that are believed to have a cometary origin (e.g., Bradley, 2003 and references therein), implying that LIME silicates may have been abundant nebular condensates and they were present in comet accretion regions. Thus, Wild 2 LIME silicates contain important clues for understanding the formation and transportation of solids in the early Solar System.

Fig. 1.

Oxygen three-isotope ratios of Wild 2 crystalline silicate particles larger than 2 μm with uncertainties of Δ¹⁷O smaller than 5‰. TF, Y&R, CCAM, and PCM represent terrestrial fractionation (Clayton et al., 1991), the Young and Russell (Young and Russell, 1998), the carbonaceous chondrite anhydrous mineral (Clayton et al., 1977), and primitive chondrule mineral (Ushikubo et al., 2012; 2017) lines, respectively. Average values for each particle are shown. Data from [1] McKeegan et al. (2006), [2] Nakamura et al. (2008), [3] Ogliore et al. (2012, 2015), [4] Bridges et al. (2012), [5] Joswiak et al. (2014), [6] Gainsforth et al. (2015), [7] Nakashima et al. (2012), [8] Defouilloy et al. (2017), [9] Chaumard et al. (2018). Two particles, Torajiro and Gozen-sama, that show heterogeneous O isotope ratios within a single particle (Nakamura et al., 2008) are not shown.

LIME silicates are also observed in amoeboid olivine aggregates (AOAs; e.g., Weisberg et al., 2004; Komatsu et al., 2015) that are the most common type of refractory inclusions in carbonaceous chondrites (e.g., Krot et al., 2004a and references therein; Sugiura et al., 2009; Ruzicka et al., 2012; Krot et al., 2014). Based on their irregularly-shaped, porous, and fine-grained textures, AOAs are thought to be aggregates of solids condensed from the solar nebula and appear to have avoided significant melting after aggregation (e.g., Grossman and Steele, 1976; Komatsu et al., 2001, 2009; Krot et al., 2004a and references therein; Weisberg et al., 2004; Sugiura et al., 2009; Han and Brearley, 2015). In addition, Mg stable isotope ratios of AOAs are indicative of a condensation origin. δ²⁵Mg values of bulk AOAs (where δ²⁵Mg = [(²⁵Mg/²⁴Mg)_sample/(²⁵Mg/²⁴Mg)_DSM-3 − 1] × 1000; per-mil deviation from Mg reference material DSM-3; Galy et al., 2003) range from −2.47‰ to −0.03‰ (Larsen et al., 2011; Olsen et al., 2011), some of which are negatively fractionated from those of bulk chondrites (δ²⁵Mg = −0.15 ± 0.04‰; Teng et al., 2010). In situ SIMS Mg isotope analyses of AOA olivines also show similar variations in δ²⁵Mg values, as low as ~ −2.8‰ (MacPherson et al., 2012; Ushikubo et al., 2017; Marrocchi et al., 2019a). The light isotope enrichments observed in AOAs may be the result of disequilibrium condensation from solar nebula gas due to kinetic isotope fractionation (Richter, 2004). Note that AOAs from the least metamorphosed chondrites are ¹⁶O-rich (Δ¹⁷O < −20‰; Aléon et al., 2002; Fagan et al., 2004; Krot et al., 2004a; Itoh et al., 2007; Yurimoto et al., 2008 and references therein; Ushikubo et al., 2017; Komatsu et al., 2018). The similarities in oxygen isotope ratios and mineral chemistry (i.e., ¹⁶O-rich oxygen isotope signatures and the occurrence of LIME silicates) suggest a genetic link between Wild 2 LIME silicates and AOAs (Nakashima et al., 2012; Defouilloy et al., 2017). However, unlike AOAs, the petrological context of ¹⁶O-rich Wild 2 particles is not well understood because most particles have been found as single mineral grains. If these particles formed by disequilibrium condensation, their Mg isotope ratios could be enriched in lighter isotopes (Richter, 2004), which could be used to trace condensation processes. Due to small particle sizes (< 10 μm), Mg isotope data from Wild 2 particles at sub-‰ precision are currently lacking, but would nonetheless be valuable for testing condensation hypotheses for ¹⁶O-rich Wild 2 particles.

Here, we report Mg three-isotope ratios and minor element concentrations of eight Wild 2 olivine particles (four that are ¹⁶O-rich and four that are ¹⁶O-depleted) in order to better understand their formation processes. To obtain high spatial resolution ( ≤ 2 μm) SIMS Mg isotope analyses, we upgraded the WiscSIMS Cameca IMS 1280 with a radio frequency (RF) plasma oxygen ion source (Hertwig et al., 2019) and applied an instrumental bias correction method that combines forsterite endmember values (= Mg/[Mg + Fe] molar %) and secondary ion Mg⁺/Si⁺ ratios (Fukuda et al., 2020). We also conducted SIMS O and Mg isotope analyses, and SIMS minor element analyses of olivine grains in AOAs from Kaba (CV3.1; Bonal et al., 2006) and the least metamorphosed CO chondrite DOM 08006 (classified as CO3.00 from Davidson et al., 2019, while we adopt CO3.01 from Fe,Ni-metal analyses of Tenner et al., 2018b), in order to compare their formation histories with those of Wild 2 particles.

2. Wild 2 Samples

Eight Wild 2 olivine particles were analyzed from four Stardust tracks (Track 57, 77, 149, and 175) including one particle from track 57 (T57/F10), four particles from track 77 (T77/F4, T77/F6, T77/F7, and T77/F50), two particles from track 149 (T149/F1 and T149/F11a), and one particle from track 175 (T175/F1). Secondary electron images of the eight Wild 2 olivine particles are shown in Fig. 2. Mineral chemistry and oxygen isotope ratios of these particles have been previously reported (Joswiak et al., 2012; Nakashima et al., 2012; Defouilloy et al., 2017; Chaumard et al., 2018). Oxygen isotope ratios of the eight Wild 2 particles (Nakashima et al., 2012; Defouilloy et al., 2017; Chaumard et al., 2018) are shown in Fig. 1. Four of eight particles are ¹⁶O-rich (T57/F10, T77/F6, T77/F50, and T175/F1) while the other four particles are ¹⁶O-depleted (T77/F4, T77/F7, T149/F1, and T149/F11a). Fo contents (Fo = Mg/[Mg + Fe] molar %) of the ¹⁶O-rich particles range from 98.0 to 99.8, while ¹⁶O-depleted particles are more FeO-rich (Fo_60–85) (Joswiak et al., 2012; Defouilloy et al., 2017; Chaumard et al., 2018). Among the ¹⁶O-rich particles, T77/F6 and T77/F50 were found in the same bulbous track identified as T77 (type B track; Hörz et al., 2006; Burchell et al., 2008; Joswiak et al., 2012). It was suggested that the original projectile was a loose aggregate and that the two ¹⁶O-rich particles as well as the other ¹⁶O-rich particle in the same track (T77/F104) which are co-linearly aligned were derived from a single grain that fragmented during capture (Joswiak et al., 2012).

Fig. 2.

SE images of the eight comet 81P/Wild 2 olivine particles before (a-g) and after (h) SIMS Mg isotope analyses. (a-d) ¹⁶O-rich olivine particles (a) T57/F10, (b) T77/F6, (c) T77/F50, (d) T175/F1. (e-h) ¹⁶O-depleted olivine particles (e) T77/F4, (f) T77/F7, (g) T149/F11a (and T149/F11b; not used in this study), (h) T149/F1. For all images, SIMS pits of previous oxygen isotope analyses are visible (Nakashima et al., 2012; Defouilloy et al., 2017; Chaumard et al., 2018). FIB marks for Mg isotope analyses can be seen as squares at the surface of the grains (a-g). For T149/F1, SIMS pits for 6 oxygen isotope analyses (Defouilloy et al., 2017) and four Mg isotope analyses (this study; #1–4) are visible. Additional images can be found in Electronic Annex EA1.

3. Analytical procedures

3.1. Electron microscopy

Two thin sections, Kaba USNM 1052–1 and DOM 08006, 50, were allocated from the Smithsonian Institution and NASA JSC, respectively. Backscattered electron (BSE) and secondary electron (SE) images of individual AOAs were obtained with a Hitachi S-3400 scanning electron microscope (SEM) at the University of Wisconsin-Madison (UW-Madison). Major and minor elements (Na₂O, MgO, Al₂O₃, SiO₂, K₂O, CaO, TiO₂, Cr₂O₃, MnO, and FeO) of minerals in AOAs were obtained with a Cameca SXFive FE electron-probe microanalyzer (EPMA) at UW-Madison. EPMA analyses were performed with an accelerating voltage of 15 kV and a beam current of 20 nA, with a 3 micron diameter beam. Off peak backgrounds were acquired, with 10 second counts on the background and 10 seconds on the peaks. The following reference materials (RMs) were used for olivine analyses: Burma jadeite (Na), synthetic forsterite (Mg, Si), Grass Valley anorthite (Al), NMNH 122142 Kakanui augite (Ca), microcline (K), TiO₂ oxide (Ti), synthetic Cr₂O₃ (Cr), synthetic Mn₂SiO₄ (Mn), and synthetic FeO (Fe). For high-Ca pyroxene analyses, the same suite of RMs were used except for Mg and Si, which were calibrated using NMNH 122142 Kakanui augite. Probe for EPMA^™ (PFE) software was used for data reduction. Calculated detection limits (99% confidence) for the measured oxides in wt% were Na₂O-0.02, MgO-0.02, Al₂O₃-0.03, SiO₂-0.03, K₂O-0.02, CaO-0.02, TiO₂-0.05, Cr₂O₃-0.07, MnO-0.05, and FeO-0.06 wt%.

Endmembers of olivine and pyroxene are expressed as Fo (= Mg/[Mg + Fe] molar %) and En (= Mg/[Mg + Fe + Ca] molar %), respectively. We also report Wo (= Ca/[Mg + Fe + Ca] molar %) as the Ca endmember for pyroxene. The Fo, En, and Wo contents were calculated by considering total iron as ferrous iron.

3.2. FIB marking

All Wild 2 particles studied here have been previously analyzed by SIMS for oxygen isotope ratios (Nakashima et al., 2012; Defouilloy et al., 2017; Chaumard et al., 2018) such that remaining areas for Mg isotope analyses were limited (typically 3 × 3 μm). In order to aim the primary ion beam precisely, we employed focused ion beam (FIB) marking to target areas prior to SIMS analyses (Nakashima et al., 2012; Defouilloy et al., 2017). FIB marking was performed with a field emission (FE) SEM equipped with a FIB system (Zeiss Auriga, UW-Madison), following procedures described in Defouilloy et al. (2017). After FIB marking, high-resolution SE images of all Wild 2 particles (Fig. 2 and Electronic Annex EA1) were obtained using the same instrument and were used for precise aiming of the primary ion beam.

3.3. Secondary ion mass spectrometry (SIMS)

Oxygen three-isotopes, Mg three-isotopes, and minor element analyses were performed with the WiscSIMS Cameca IMS 1280 equipped with the RF plasma ion source at UW-Madison. Five separate sessions (one for O isotopes, three for Mg isotopes, and one for minor elements) were conducted and analytical conditions were optimized for each session described below. Following analytical sessions, SIMS pits were investigated by SEM (S3400 at UW-Madison) or FE-SEM (LEO 1530 at UW-Madison), of which images are shown in Electronic Annex EA1–3.

3.3.1. SIMS oxygen isotope analyses

For oxygen three-isotope analyses of olivine from Kaba and DOM 08006 AOAs, the primary ¹³³Cs⁺ ion beam was focused to a ~8 μm diameter spot at 0.7 nA. Secondary oxygen ions (¹⁶O⁻, ¹⁷O⁻, and ¹⁸O⁻) were detected simultaneously using multi-collector Faraday cups (FCs) (Kita et al., 2010). To improve precision of δ¹⁷O analyses, the FC for ¹⁷O⁻ employed a high gain feedback resistor (10¹² ohm) that can reduce thermal noise (Electronic Annex EA4–1). The other FCs for ¹⁶O⁻ and ¹⁸O⁻ employed 10¹⁰ ohm and 10¹¹ ohm resistors, respectively. The mass resolving power (MRP) was set to ~2200 for ¹⁶O⁻ and ¹⁸O⁻, and 5000 for ¹⁷O⁻. A single analysis took ~7 min, including 100 s of presputtering and baseline measurement, ~80 s for automated centering of secondary ion deflectors (DTFA-X, DTFA-Y), 200 s of integration (10 s × 20 cycles) of the secondary ion signals. At the end of each analysis, ¹⁶O¹H⁻ was measured automatically to estimate the contribution of the tailing of ¹⁶O¹H⁻ to the ¹⁷O⁻ peak, which was always insignificant (<0.05‰ on δ¹⁷O; Electronic Annex EA4–1). A synthetic pure forsterite RM, HN-Ol (δ¹⁸O = 8.90‰: Kita et al., 2010), was used as the bracketing standard. The typical secondary ¹⁶O⁻ ion intensity when measuring the HN-Ol was 7.5 × 10⁸ counts per second (cps). Ten to fifteen unknown analyses were bracketed by 8 analyses of the HN-Ol standard (4 before and 4 after the unknowns). The external reproducibility of the HN-Ol standard was typically 0.4‰, 0.9‰, and 0.9‰ (2SD) for δ¹⁸O, δ¹⁷O, and Δ¹⁷O, respectively. The external reproducibilities (2SD) for each bracket are assigned as the uncertainties of bracketed unknown analyses. Instrumental biases in δ¹⁸O were calibrated based on the bracket HN-Ol analyses. Since all olivine grains in AOAs analyzed are near-pure forsterite (>Fo₉₉), no additional correction depending on Fo contents was applied.

3.3.2. SIMS magnesium isotope analyses

We conducted three sessions for Mg three-isotope analysis of olivine. For each session, the primary O₂⁻ ion beam was accelerated by 23 kV (−13 kV at the ion source and +10 kV at the sample surface) to the sample surface and secondary ²⁴Mg⁺, ²⁵Mg⁺, and ²⁶Mg⁺ ions were detected simultaneously using multi-collector FCs (L′2, C, and H1). Secondary ion optics are similar to those described in Kita et al. (2012) and Ushikubo et al. (2013). The MRP at 10% peak height was set to ~2500 and contributions of ⁴⁸Ca⁺ and ²⁴Mg¹H⁺ to ²⁴Mg⁺ and ²⁵Mg⁺ peaks, respectively, were negligibly small (EA2 in Fukuda et al., 2020). As described below, the primary beam intensity (and size) was optimized depending on the size of unknown samples, which led to different secondary Mg⁺ ion intensities. Therefore, the combinations of resistors for the multi-collector FCs were optimized for each session. Each analysis duration was ~8–12 min, including 100 s of presputtering, ~100 s for automated centering of the secondary ion deflectors, and ~300–500 s of integration (10 s × 30 cycles for AOA analyses and 10 s × 50 cycles for Wild 2 analyses) of secondary ion signals. The baselines of the three FC detectors were monitored during each presputtering step, and their values were averaged over eight analyses.

Data reduction procedures follow those from Ushikubo et al. (2017). Terrestrial reference ²⁵Mg/²⁴Mg and ²⁶Mg/²⁴Mg values (0.12663 and 0.13932, respectively; Catanzaro et al., 1966) were used to convert raw-measured, background-corrected ratios to delta notation (δ²⁵Mg_m and δ²⁶Mg_m). The δ²⁵Mg and δ²⁶Mg matrix effects were calibrated based on analyses of ~13 olivine reference materials (RMs), of which Mg isotope ratios in DSM-3 scale (Galy et al., 2003) have been determined by solution nebulization multi-collector inductively coupled plasma mass spectrometry (SN-MC-ICP-MS) or laser ablation MC-ICP-MS (Fukuda et al., 2020). Recently, we found complex matrix effects related to SIMS Mg isotope analyses of olivine, leading to instrumental biases on δ²⁵Mg that could not be fit by a simple regression to Fo contents. Instead, we apply a calibration that uses a combination of Mg⁺/Si⁺ signal ratios and Fo contents of olivine (Fukuda et al., 2020). Specifically, instrumental biases are estimated as a function of X = (²⁴Mg⁺/²⁸Si⁺)/Fo, where (²⁴Mg⁺/²⁸Si⁺) is the secondary ion ²⁴Mg⁺/²⁸Si⁺ ratio obtained from SIMS minor element analysis (see section 3.3.3) and where Fo contents of unknown samples were obtained with TEM-EDX (Wild 2 olivine; Joswiak et al., 2012; Nakashima et al., 2012; Defouilloy et al., 2017; Chaumard et al., 2018) or FE-EPMA (AOA olivine; this study).

Magnesium isotope ratios of unknowns are reported as δ²⁵Mg and δ²⁶Mg values in DSM-3 scale. The bias corrected δ²⁵Mg and δ²⁶Mg values are calculated as follows:

$${\delta }^i\text{Mg}=\left[\left(1+{\delta }^i{\text{Mg}}_{\text{m}}/1000\right)/\left(1+{\delta }^i{\text{Mg}}_{\text{bias }}/1000\right)-1\right]\times 1000$$

where i = 25 or 26 and δⁱMg_bias is the instrumental bias determined from the calibration curve (Electronic Annex EA5). Uncertainties in δ²⁵Mg and δ²⁶Mg are estimated by combining the analytical uncertainties and error of the fit to the instrumental bias correction. More details for the estimation of uncertainties are summarized in Electronic Annex EA5.

The mass-independent raw-measured Δ²⁶Mg_m is defined as Δ²⁶Mg_m = [(1 + δ²⁶Mg_m/1000)/(1 + δ²⁵Mg_m/1000)^{1/ β} − 1] × 1000. We assume that natural and instrumental mass-dependent fractionations follow a power law function, where an applied β value of 0.5128 was determined by vacuum evaporation experiments on melts of a Type B CAI bulk composition (Davis et al., 2015). Δ²⁶Mg_m values of olivine RMs obtained in the same analysis session were used to correct the non-mass dependent instrumental bias (see more details in Electronic Annex EA5). The mass-independent, bias-corrected Δ²⁶Mg of unknowns is defined as follows:

$${\text{Δ}}^{26}\text{Mg}=\left[\left(1+{\text{Δ}}^{26}{\text{Mg}}_{\text{m}\left(\text{unknown}\right)}/1000\right)/\left(1+{\text{Δ}}^{26}{\text{Mg}}_{\text{m}\left(\text{RM}\right)}/1000\right)-1\right]\times 1000$$

where Δ²⁶Mg_{m (unknown)} and Δ²⁶Mg_{m (RM)} are Δ²⁶Mg_m of unknowns and Δ²⁶Mg_m of olivine RMs, respectively. Finally, δ²⁶Mg* of unknowns are estimated as: δ²⁶Mg* = Δ²⁶Mg × (1 + δ²⁵Mg/1000)^{1/ β} (Ushikubo et al., 2017). We note that natural mass-dependent fractionations in Wild 2 and AOA olivines may not follow this power law β value, but that any potential δ²⁶Mg* inaccuracy this causes is likely to be minimal. For example, if assuming kinetic Mg isotope fractionation (β = 0.511; Young and Galy, 2004), the δ²⁶Mg* of Wild 2 and AOA olivines only change by a maximum of 0.03‰ within δ²⁵Mg values ranging from −3.8 to 0.6‰. This small offset is well within our analytical uncertainties (~0.05–0.47‰, 2SD).

The Wild 2 olivine particles were analyzed in session 1 (S1; 2018 Oct). The primary O₂⁻ beam was focused to 2 μm at 24 pA. All three FCs employed 10¹² ohm resistors for detection of ^{24, 25, 26}Mg⁺ ions. Prior to each analysis, secondary ²⁵Mg⁺ ion images of the Wild 2 olivine particles were obtained by rastering the primary ion beam (24pA) over a 10 μm × 10 μm area in order to identify the location of FIB marks. For this mode of operation, the secondary ²⁵Mg⁺ ions were detected with a multi-collector electron multiplier (EM; L1) by applying X-deflection after the secondary magnet (DSP2X), which is similar to previous techniques for oxygen and magnesium isotope analyses of Wild 2 particles (Nakashima et al., 2012, 2015; Defouilloy et al., 2017). One to four analyses were performed for each Wild 2 particle, bracketed by eight analyses of the San Carlos olivine (hereafter SC-Ol) grains mounted in the same disks. In addition, twelve other olivine RMs were measured for instrumental bias corrections (Electronic Annex EA4–2–1). The typical ²⁴Mg⁺ count rate during measurement of SC-Ol was 4.6 × 10⁶ cps. The standard deviations of FC baselines over eight analyses were typically 120 cps for L′2, 90 cps for C, and 110 cps for H1, corresponding to 3 × 10⁻⁵ relative to ²⁴Mg⁺ and 1.6 × 10⁻⁴ relative to ²⁵Mg⁺ and ²⁶Mg⁺ ion intensities. In this session, we estimated the time constant of each 10¹² ohm feedback resistor by monitoring the decay of secondary ^{24, 25, 26}Mg⁺ signals (i.e., after turning off the primary beam) from SC-Ol (Electronic Annex EA4–5). The time constants of the three 10¹² ohm feedback resistors range from 0.49 to 0.74 seconds, which are similar to each other. Thus we did not apply any tau corrections (e.g., Pettke et al., 2011). The external reproducibilities (2SD) of δ²⁵Mg_m, δ²⁶Mg_m, and Δ²⁶Mg_m for the SC-Ol bracket standard were 0.25‰, 0.26‰ and 0.47‰, respectively. The final uncertainties including error of the fit to the instrumental bias correction were typically 0.58‰ 1.05‰, and 0.54‰ for δ²⁵Mg, δ²⁶Mg, and δ²⁶Mg*, respectively.

Olivine grains from Kaba AOA (K27) were analyzed in session 2 (S2; 2018 Oct). The primary O₂⁻ beam was focused to 10 μm at 2 nA. The FC for ²⁴Mg⁺ (L′2) employed a 10¹⁰ ohm resistor, and the other two FCs for ²⁵Mg⁺ and ²⁶Mg⁺ (C and H1) employed 10¹¹ ohm resistors. SC-Ol grains were used as a bracketing standard. In addition, five Mg-rich olivine RMs (>Fo₉₄) listed in Electronic Annex EA4–2–2 were measured for instrumental bias corrections. The typical ²⁴Mg⁺ count rate for SC-Ol was 4.4 × 10⁸ cps. The external reproducibilities (2SD) of δ²⁵Mg_m, δ²⁶Mg_m, and Δ²⁶Mg_m for the SC-Ol bracket standard were 0.08‰, 0.15‰ and 0.05‰, respectively. The final uncertainties including error of the fit to the instrumental bias correction were typically 0.31‰, 0.60‰, and 0.06‰ for δ²⁵Mg, δ²⁶Mg, and δ²⁶Mg*, respectively.

Olivine grains from DOM 08006 AOAs were analyzed in session 3 (S3; 2018 Dec). The primary O₂⁻ beam was focused to ~5 μm at 160 pA. As for Wild 2 analyses, all three FCs employed 10¹² ohm resistors for detection of ^24,25,26Mg⁺ ions. SC-Ol grains were used as a bracketing standard. An additional 11 olivine RMs, listed in Electronic Annex EA4–2–3, were measured for instrumental bias corrections. The typical ²⁴Mg⁺ count rate when measuring SC-Ol was 3.0 × 10⁷ cps. External reproducibilities (2SD) of δ²⁵Mg_m, δ²⁶Mg_m, and Δ²⁶Mg_m for the SC-Ol bracket standard were 0.13‰, 0.22‰ and 0.19‰, respectively. The final uncertainties including error of the fit to the instrumental bias correction were typically 0.44‰ 0.88‰, and 0.23‰ for δ²⁵Mg, δ²⁶Mg, and δ²⁶Mg*, respectively.

3.3.3. SIMS Mg⁺/Si⁺ ratios and minor element analyses

Mg⁺/Si⁺ ratios and minor element (Ca, Cr, Mn, and Fe) concentrations were obtained using the same analytical conditions reported in Fukuda et al. (2020). Secondary ²⁴Mg⁺ ions were detected by an axial FC (FC2) with a 10¹¹ ohm resistor. The other secondary ions (²³Na⁺, ²⁷Al⁺, ²⁸Si⁺, ⁴⁰Ca⁺, ⁵²Cr⁺, ⁵⁵Mn⁺, ⁵⁶Fe⁺, and ⁶⁰Ni⁺), as well as mass 22.7 (used for stabilizing the magnetic field for ²³Na⁺), were detected by an axial EM that operated in magnetic peak jumping mode. Due to consumption of sample area from prior O and Mg isotope analyses of the Wild 2 olivine particles, we focused the primary O₂⁻ beam down to 1.5 μm in diameter (6 pA) and aimed it inside the previous SIMS pits (~2 μm in diameter) from Mg isotope analyses. For AOAs from DOM 08006, the primary ion beam was also aimed inside previous SIMS pits (~5 μm in diameter) from Mg isotope analyses. Prior to each analysis, a secondary ²⁴Mg⁺ ion image of the targeted SIMS pit was obtained by rastering the primary ion beam (6 pA) over a 10 × 10 μm area in order to aim the primary ion beam inside the SIMS pit. The AOA (K27) from Kaba had abundant areas for minor element analyses so that flat surface areas near the previous SIMS pits from Mg isotope analyses were analyzed. SC-Ol grains were used as a running standard. An additional 12 olivine standards (Electronic Annex EA4–3) were also measured to obtain ²⁴Mg⁺/²⁸Si⁺ ratios and to determine relative sensitivity factors (RSFs) of minor elements (Ca, Cr, Mn, and Fe). The typical ²⁴Mg⁺ count rate for SC-Ol was ~7.6 × 10⁵ cps. We also conducted multiple analyses of olivine RMs in order to compare the instrumental bias within previously generated SIMS pits and flat surface areas (see more details in Electronic Annex EA5). A single analysis was ~6 min in duration, including 100 s of presputtering, ~80 s for automated centering of secondary ion deflectors, and 160 s of integration (32 s × 5 cycles) of secondary ion signals. Detection limits were evaluated based on the measurement of synthetic pure forsterite RM (HN-Ol), which is expected not to contain Ca, Cr, Mn, and Fe. The detection limits of CaO, Cr₂O₃, MnO, and FeO were 0.002, 0.0001, 0.001, and 0.003 wt%, respectively.

4. RESULTS

4.1. AOA petrography and mineral chemistry

One AOA (K27) from Kaba and eight AOAs (501, 502, 503, 505, 511, 512, 513, and 514) from DOM 08006 were studied by SEM-EDS and FE-EPMA. All nine AOAs are composed of forsteritic olivine, opaque minerals (Fe,Ni-metal and/or weathering products), and Ca-, Al-rich portions (±spinel, ±anorthite, and Al-(Ti)-diopside). Low-Ca pyroxene is rare and was identified in two out of nine AOAs (502 and 503). The mineralogical characteristics of the nine AOAs are summarized in Table 1. The AOAs studied here are less than ~400 μm, except for K27 from Kaba (~1.7 mm). Average elemental compositions (wt%) and Fo, En, and Wo contents of AOA minerals are summarized in Table 2. Individual EPMA data and their analysis locations are summarized in Electronic Annex EA4–4 and EA6, respectively. Olivine grains from the nine AOAs have a narrow range of Fo content (Fo_99–100). CaO, Cr₂O₃, MnO, and FeO contents of olivine grains vary by ~0.4, ~0.5, ~0.5, and ~1.1 wt%, respectively. As we discuss later, FeO contents obtained with EPMA analyses might be overestimated due to secondary fluorescence effects (e.g., Fournelle et al., 2005; Buse et al., 2018) from surrounding FeO enhanced phases. En components of diopside in K27, 505, 511, and 513 range from 44.0 to 50.2. Al₂O₃ and TiO₂ contents of these diopside grains range from 3.4 to 11.9 wt% and 0.2 to 2.9 wt%, respectively.

Table 1.

Mineralogy of AOAs from Kaba (CV3.1) and DOM 08006 (CO3.01) carbonaceous chondrites

Chondrite	AOA name	AOA texture	fo	FeNi	Ca-, A1-rich portions				LIME o1b	1px
					A1-(Ti)-di	an	sp	haloa
Kaba	K27	compact	+	+	+	+	+	−	−	−
DOM 08006	501	pore-rich	+	+	+	+	−	+	+	−
DOM 08006	502	pore-rich	+	+	+	+	−	+	+	+
DOM 08006	503	pore-rich	+	+	+	+	−	+	+	+
DOM 08006	505	compact	+	+	+	−	+	−	+	−
DOM 08006	511	pore-rich	+	+	+	+	−	+	+	−
DOM 08006	512	pore-rich	+	+	+	+	−	+	+	−
DOM 08006	513	compact	+	+	+	+	+	−	−	−
DOM 08006	514	pore-rich	+	+	+	+	−	+	−	−

^aDiffuse area that appears to be brighter in BSE images than surrounding forsterite. See section 4.1.

^bOlivine grains with MnO/FeO (wt%) ≥ 1. The MnO/FeO ratios are calculated from SIMS data in Table 4, except for AOA 514.

The MnO/FeO ratios of olivine grains from AOA 514 are calculated from FE-EPMA data due to an absence of SIMS minor element data.

fo = forsterite, FeNi = Fe,Ni-metal and/or weathering products, Al-(Ti)-di = Al-(Ti) diopside, an = anorthite, sp = spinel, LIME ol = LIME olivine, lpx = low-Ca pyroxene

Table 2.

Fo, En, and Wo contents and average elemental compositions (wt%) of minerals in Kaba and DOM 08006 AOAs obtained with electron microprobe.

Chondrite	AOA name	Minerala	Fob	1SD	Enb	1SD	Wob	1SD	No. of analyses	Na2O	MgO	A12O3	SiO2	K2O	CaO	TiO2	Cr2O3	MnO	FeO	Total
Kaba	K27	fo	99.9	0.1					15	b.d	57.17	0.06	42.31	b.d	0.15	0.06	0.20	0.10	0.19	100.23
Kaba	K27	hpx			48.9	1.4	51.1	1.4	2	b.d	17.15	5.32	50.88	b.d	24.88	1.94	0.11	b.d	b.d	100.29
DOM 08006	501	fo	99.5	0.3					12	b.d	57.04	0.06	42.24	b.d	0.19	0.05	0.16	0.11	0.49	100.34
DOM 08006	502	fo	99.5	0.2					7	b.d	57.14	0.05	42.35	b.d	0.07	0.10	0.34	0.27	0.48	100.81
DOM 08006	503	fo	99.7	0.1					7	b.d	57.61	0.05	42.50	b.d	0.09	0.06	0.22	0.10	0.34	100.96
DOM 08006	505	fo	99.8	0.1					7	b.d	57.56	0.04	42.15	b.d	0.11	0.06	0.20	0.12	0.18	100.41
DOM 08006	505	hpx			50.1	0.1	49.6	0.2	3	b.d	18.12	3.87	53.00	b.d	24.92	0.28	0.09	b.d	0.20	100.48
DOM 08006	511	fo	99.5	0.1					7	b.d	57.36	0.04	42.17	b.d	0.12	0.05	0.13	0.16	0.54	100.56
DOM 08006	511	hpx			45.1	1.6	54.3	1.4	2	b.d	14.90	10.17	47.95	b.d	24.90	2.07	0.10	b.d	0.35	100.44
DOM 08006	511	an							1	0.03	0.55	35.94	42.55	b.d	19.84	0.07	0.08	b.d	0.37	99.43
DOM 08006	512	fo	99.7	0.1					7	b.d	57.35	0.08	42.32	b.d	0.09	0.08	0.27	0.15	0.27	100.61
DOM 08006	513	fo	99.6	0.1					7	b.d	57.32	0.04	42.44	b.d	0.33	0.05	0.10	b.d	0.40	100.69
DOM 08006	513	hpx			47.9	1.1	51.4	1.2	4	b.d	16.63	6.85	50.15	b.d	24.83	1.38	0.10	b.d	0.46	100.40
DOM 08006	514	fo	99.7	0.2					7	b.d	57.40	0.07	42.36	b.d	0.35	0.07	0.07	b.d	0.33	100.65

^afo = forsterite, hpx = high-Ca pyroxene, an = anorthite

^bFo = Mg/(Mg + Fe) molar %, En = Mg/(Mg + Fe + Ca) molar %, Wo = Ca/(Mg + Fe + Ca) molar %.

The AOAs investigated show variations in abundances of pores, shapes of external margins and ratios of minor minerals to olivine (Figs. 3–5). Three (K27, 505 and 513) show a compact texture with little porosity in the forsterite-rich regions (Fig. 3). The other AOAs (e.g., 502 and 503) are more porous and are irregularly shaped (Figs. 4–5). Detailed information for each inclusion is summarized below.

Fig. 3.

BSE images of compact AOAs from Kaba and DOM 08006. (a-c) K27 from Kaba, (d-e) 513 from DOM 08006, (f-g) 505 from DOM 08006. Mineral phases shown are olivine (Ol), anorthite (An), spinel (Sp) and Al diopside (Al-di).

Fig. 5.

BSE images of pore-rich AOAs from DOM 08006. (a) 501, (b) pore-rich lithology in 501, (c) compact lithology in 501, (d) 511, (e) 512, (f) 514. The locations of Mg isotope analyses for pore-rich and compact lithologies in AOA 501 are indicated in (b-c). Mineral phases shown are olivine (Ol), Al-diopside (Al-di) and anorthite (An).

Fig. 4.

BSE images of pore-rich, low-Ca pyroxene-bearing AOAs from DOM 08006. (a-d) 502, (e-g) 503. Mineral phases shown are olivine (Ol), low-Ca pyroxene (Lpx), Al diopside (Al-di), anorthite (An) and Fe,Ni-metal (FeNi). Pores, and haloes that represent diffuse area consisting of mixtures of tiny Al-diopside grains and forsterite, are indicated.

AOA K27 from Kaba is the largest aggregate (~1.7 mm; Fig. 3a) studied here. It consists of multiple nodules, and each nodule is composed of Ca-, Al-rich domains enclosed by sinuous forsterite (Fo_99.5–100) and opaque minerals (Fig. 3b). Forsterite regions display dense textures with little porosity between grain boundaries (Figs. 3b and c). Ca-, Al-rich domains have anorthite ± spinel cores rimmed by Al-diopside (Fig. 3c). Spinel grains are anhedral and appear corroded by the surrounding anorthite (Fig. 3c). Al-diopside occurs between anorthite and forsterite (Figs. 3b and c).

AOA 513 from DOM 08006 (Fig. 3d) consists of multiple Ca-, Al-rich portions enclosed by nearly pure forsterite grains (Fo_99.5-_99.7) and rare Fe,Ni-metal. This AOA has anorthite ± anhedral spinel cores rimmed by Al-diopside, which are enclosed by dense forsterite (Fig. 3e). Al-diopside occurs between anorthite and forsterite (Fig. 3e). The mineralogy and texture of this AOA are similar to those of Kaba AOA K27 (Figs. 3b and c).

AOA 505 from DOM 08006 (Fig. 3f) consists of spinel-diopside-rich nodules enclosed by dense forsterite (Fo_99.7–99.9) with minor Fe,Ni-metal. This inclusion may also be classified as a CAI with a thick forsterite rim. Spinel is separated from forsterite by a layer of Al-diopside (Fig. 3g).

AOA 502 from DOM 08006 is irregularly shaped and porous (Fig. 4a), consisting mostly of forsterite (Fo_99.3–99.7), with minor Fe,Ni-metal and numerous irregularly shaped diffuse areas (Figs. 4b–d). Rare low-Ca pyroxene is also observed (Figs. 4b and d), and forsterite regions of this AOA are porous (Figs. 4b and c). SEM-EDS analyses indicate that the diffuse areas are mixtures of tiny Al-diopside grains, and forsterite. Such diffuse areas are also observed in AOAs from ALHA77307 (CO3.03) as brighter areas in BSE images than surrounding forsterite, referred to as haloes by Sugiura et al. (2009). Forsterite grains within this AOA have the highest MnO abundances (~0.5 wt%) among the nine AOAs studied here.

AOA 503 from DOM 08006 is a fine-grained, porous object (Fig. 4e), consisting mostly of forsterite (Fo_99.4–99.8), with minor Fe,Ni-metal and numerous irregularly shaped Ca-, Al-rich portions (tiny Al-diopside grains and an anorthitic halo; Fig. 4f). Rare low-Ca pyroxene is also observed (Fig. 4g). The mineralogy and texture of this AOA is similar to those of AOA 502, but MnO contents of forsterite grains are lower (~0.1 wt%).

AOA 501 from DOM 08006 (Fig. 5a) consists of forsterite (Fo_99.0–99.9), Fe,Ni-metal, and numerous irregularly shaped, fine-grained Ca, Al-rich portions. Most of this AOA is composed of porous forsterite with numerous irregularly shaped, fine-grained Ca, Al-rich portions (Fig. 5b). However, a region shown in Fig. 5c is less porous, and has nodule-like (Ca-, Al-rich core and sinuous forsterite) textures, similar to characteristics observed in Kaba AOA K27 (e.g. Fig. 3b).

AOAs 511 and 512 from DOM 08006 are irregularly shaped objects (Figs. 5d and 5e, respectively), consisting of forsterite (Fo_99.3–99.6), Ca-, Al-rich portions (Al-diopside and anorthite), and minor Fe,Ni-metal. Ca, Al-rich diffuse areas (haloes) are observed in each AOA.

AOA 514 from DOM 08006 is a sub-rounded object (Fig. 5f), consisting of forsterite (Fo_99.2–99.9) and irregularly shaped Ca-, Al-rich portions (Al-diopside and anorthite) that are concentrated in the inner part of the AOA. Forsterite grains of the inner part of the AOA, along with Ca, Al-rich portions, are more porous relative to the outer regions of this AOA.

4.2. Oxygen isotope ratios of AOA olivines

We obtained 32 olivine oxygen isotope ratio data from eight of the nine AOAs studied here (the exception being AOA 512, due to small olivine grains). δ¹⁸O, δ¹⁷O, and Δ¹⁷O values are reported in Table 3. Raw-measured O isotope ratios of AOA olivines are reported in Electronic Annex EA4–1. As shown in Fig. 6a, oxygen isotope ratios of all measured AOA olivines are ¹⁶O-rich (Δ¹⁷O ≤ −23‰) and plot either on the Primitive Chondrule Mineral (PCM; Ushikubo et al., 2012; 2017) line or between the PCM and CCAM (Carbonaceous Chondrite Anhydrous Mineral; Clayton et al., 1977) lines. Individual olivine analyses from the eight AOAs have limited variations in δ¹⁸O and Δ¹⁷O values, ranging from −44.0‰ to −47.4‰ and −23.0‰ to −24.8‰, respectively. Subtle variations of oxygen isotope ratios are observed within each AOA (e.g., K27 and 514; Fig. 6a), but averaged oxygen isotope ratios of individual AOAs form a tight cluster on the PCM line (Fig. 6b), and are identical within uncertainties.

Table 3.

Oxygen isotope ratios of AOA olivines from Kaba and DOM 08006

Sample/Location #a		δ18O (‰)	Unc. (2σ)	δ17O (‰)	Unc. (2σ)	Δ17O (‰)	Unc. (2σ)
Kaba(CV3.1)
K27
	#1	−44.6	0.5	−47.0	0.9	−23.8	1.0
	#2	−45.1	0.5	−47.2	0.9	−23.7	1.0
	#3	−44.0	0.5	−46.8	0.9	−24.0	1.0
	#4	−45.2	0.5	−47.2	0.9	−23.7	1.0
	#5	−45.7	0.5	−46.8	0.9	−23.0	1.0
	#6	−45.4	0.5	−47.1	0.9	−23.5	1.0
	#7	−45.6	0.4	−48.2	0.7	−24.5	0.7
	#8	−45.1	0.4	−48.1	0.7	−24.7	0.7
	#9	−45.6	0.4	−48.3	0.7	−24.6	0.7
	#10	−46.1	0.4	−47.7	0.7	−23.7	0.7
	Average & 2SD	−45.2	1.2	−47.5	1.2	−23.9	1.1
DOM 08006 (CO3.01)
501
	#1	−46.4	0.2	−47.4	0.9	−23.3	0.8
	#2	−45.7	0.2	−47.7	0.9	−23.9	0.8
	#3	−46.2	0.2	−47.9	0.9	−23.9	0.8
	#4	−45.9	0.2	−48.1	0.9	−24.2	0.8
	#5	−46.5	0.2	−47.7	0.9	−23.5	0.8
	Average & 2SD	−46.1	0.7	−47.8	0.5	−23.8	0.7
502
	#1	−46.2	0.2	−48.2	0.9	−24.2	0.8
503
	#1	−45.9	0.2	−47.5	0.9	−23.7	0.8
505
	#1	−46.5	0.4	−48.0	0.7	−23.8	0.7
	#2	−45.9	0.4	−47.3	0.7	−23.5	0.7
	#3	−46.4	0.4	−48.1	0.7	−24.0	0.7
	Average & 2SD	−46.3	0.7	−47.8	0.8	−23.8	0.5
511
	#1	−45.4	0.4	−47.7	0.7	−24.1	0.7
513
	#1	−45.5	0.2	−48.0	0.9	−24.3	0.8
	#2	−46.0	0.2	−48.3	0.9	−24.4	0.8
	#3	−45.9	0.2	−47.7	0.9	−23.9	0.8
	#4	−45.6	0.2	−47.5	0.9	−23.8	0.8
	Average & 2SD	−45.8	0.4	−47.9	0.7	−24.1	0.6
514
	#1	−45.5	0.4	−47.7	0.7	−24.0	0.7
	#2	−45.9	0.4	−47.7	0.7	−23.8	0.7
	#3	−45.8	0.4	−48.1	0.7	−24.3	0.7
	#4	−47.4	0.4	−49.1	0.7	−24.5	0.7
	#5	−45.5	0.4	−47.9	0.7	−24.2	0.7
	#6	−45.6	0.4	−48.6	0.7	−24.8	0.7
	#7	−45.8	0.4	−47.9	0.7	−24.1	0.7
	Average & 2SD	−45.9	1.3	−48.1	1.1	−24.3	0.7

^aAnalysis points are shown in Electronic Annex EA2.

Fig. 6.

Oxygen three-isotope ratios of the eight AOA olivines from Kaba (K27) and DOM 08006 (501, 502, 503, 505, 511, 513, 514). All measured data (= individual olivine analyses) are shown in (a). The insert in (a) shows detail of the analyses highlighted by the dotted box. Average δ^{17, 18}O values of each AOA are shown in (b), though for three AOAs (502, 503, 511) only a single analysis was collected. For comparison, data for AOAs from Acfer 094 (Ushikubo et al., 2017) are also shown in (b). Reference lines are the terrestrial fractionation (TF), the carbonaceous chondrite anhydrous mineral (CCAM), the Young and Russell (Y&R), and the primitive chondrule mineral (PCM) lines. Errors are 2SD for data obtained in this study. Errors for Acfer 094 data are propagated, as described in Ushikubo et al. (2017).

4.3. Magnesium isotope ratios of AOA and Wild 2 olivines

We obtained 36 olivine magnesium three-isotope ratio data from eight of the nine AOAs studied here. We did not acquire Mg isotope data from AOA 514 because large secondary deflector values (DTFA values > 50) had to be applied, and under such a condition δ²⁵Mg analyses are inaccurate due to additional instrument fractionation. Magnesium isotope ratios (δ²⁵Mg and δ²⁶Mg) are listed in Table 4 and plotted in Fig. 7. Individual SIMS data including raw-measured Mg isotope ratios are reported in Electronic Annex EA4–2–6 and EA4–2–7. δ²⁵Mg values of AOA olivines range from −3.8 ± 0.5‰ (2σ) to −0.2 ± 0.3‰ (2σ). All Mg isotope analyses obtained in AOA olivines show a linear correlation between δ²⁶Mg and δ²⁵Mg (Fig. 7a). We also calculated logarithmic values (δ²⁵Mg′ and δ²⁶Mg′) that are converted from δ²⁵Mg and ²⁶Mg values using the following formula: δⁱMg′ = 1000 × ln (1 + δⁱMg/1000) (Electronic Annex EA4–2–8). The logarithmic values are used to determine the slope and intercept in δ²⁶Mg′ versus δ²⁵Mg′ space (Fig. 8), with values of 0.516 ± 0.012 (2σ) and −0.06 ± 0.04 (2σ), respectively (using Isoplot 4.15 model 1; Ludwig, 2012). Variation in δ²⁵Mg and δ²⁶Mg within each AOA is identified, which follows a line with the same slope of 0.516 (Figs. 7b–i). Among them, δ²⁵Mg and δ²⁶Mg values of AOA 501 are bimodally distributed on the slope 0.516 line (Fig. 7c).

Table 4.

Magnesium isotope ratios and minor elemental abundances of AOA olivine from Kaba and DOM 08006 obtained with SIMS

		Sample/Location #a	δ25Mgdcm-3 (‰)	Unc. (+/−) (2σ)		δ26Mgdsm-3 (‰)	Unc. (+/−) (2σ)		δ26Mg* (‰)	Unc. (2σ)	CaO (wt%)	Unc. (2σ)	Cr2O3 (wt%)	Unc. (2σ)	FeO (wt%)	Unc. (2σ)	MnO (wt%)	Unc. (2σ)	MnO/FeO	Unc. (2σ)
Kaba (CV3.1)
K27
		#3b	−0.39	0.30	0.31	−0.68	0.59	0.59	0.04	0.06	0.128	0.004	0.110	0.007	0.324	0.017	0.027	0.002	0.08	0.01
		#4	−0.42	0.31	0.31	−0.78	0.60	0.60	−0.01	0.06	0.150	0.005	0.081	0.005	0.064	0.002	0.008	0.002	0.13	0.04
		#5	−0.25	0.30	0.31	−0.48	0.59	0.59	−0.03	0.06	0.158	0.005	0.083	0.005	0.065	0.003	0.008	0.001	0.12	0.02
		#6	−0.51	0.30	0.33	−0.95	0.59	0.62	0.00	0.06	0.230	0.022	0.057	0.004	0.116	0.003	0.009	0.001	0.08	0.01
		#7	−0.64	0.31	0.31	−1.21	0.60	0.60	−0.01	0.06	0.202	0.017	0.113	0.007	0.072	0.003	0.024	0.001	0.34	0.02
		#8	−0.69	0.31	0.31	−1.12	0.60	0.60	0.16	0.06	0.136	0.005	0.105	0.007	0.081	0.003	0.007	0.002	0.09	0.02
		#9b	−0.23	0.32	0.32	−0.38	0.61	0.61	0.01	0.06	0.089	0.006	0.138	0.012	0.184	0.006	0.098	0.004	0.53	0.03
		Average & 2SD	−0.45	0.35	0.35	−0.80	0.63	0.63	0.02	0.13	0.156	0.094	0.098	0.037	0.129	0.192	0.026	0.066	0.20	0.59
DOM 08006 (CO3.01)
501
		#1_outer left partc	−0.42	0.44	0.44	−0.55	0.87	0.87	0.17	0.24	0.119	0.003	0.188	0.006	0.072	0.003	0.088	0.001	1.21	0.06
		#2	−3.31	0.44	0.44	−6.27	0.86	0.86	0.09	0.24	0.082	0.002	0.177	0.010	0.073	0.002	0.108	0.004	1.49	0.07
		#3	−2.90	0.40	0.40	−5.50	0.80	0.80	0.10	0.24	0.165	0.005	0.100	0.005	0.058	0.002	0.032	0.003	0.55	0.05
		#4	−3.30	0.42	0.42	−6.21	0.84	0.84	0.15	0.24	0.189	0.004	0.072	0.003	0.052	0.001	0.028	0.002	0.54	0.04
		#5_outer left partc	−0.66	0.43	0.43	−1.20	0.85	0.85	0.01	0.24	0.151	0.005	0.117	0.005	0.071	0.003	0.078	0.007	1.10	0.11
		#6	−3.79	0.45	0.45	−7.27	0.92	0.92	0.05	0.21	0.128	0.003	0.093	0.004	0.059	0.002	0.051	0.003	0.87	0.06
		#7	−3.08	0.46	0.46	−5.85	0.93	0.93	0.05	0.21	0.142	0.003	0.163	0.005	0.073	0.002	0.093	0.003	1.28	0.05
		Average (outer left) & 2SD	−0.54	0.34	0.34	−0.87	0.93	0.93	0.09	0.24	0.135	0.046	0.153	0.100	0.072	0.003	0.083	0.014	1.16	0.21
		Average (#2–4 & #6–7) & 2SD	−3.28	0.67	0.67	−6.22	1.32	1.32	0.09	0.08	0.142	0.081	0.121	0.093	0.063	0.019	0.063	0.073	0.99	1.19
502
		#1	−0.62	0.44	0.44	−1.16	0.85	0.85	−0.05	0.24	0.033	0.001	0.320	0.011	0.063	0.002	0.074	0.003	1.18	0.06
		#2	−0.58	0.44	0.44	−0.87	0.85	0.85	0.14	0.24	0.045	0.002	0.316	0.010	0.078	0.004	0.084	0.004	1.08	0.08
		#3	−0.59	0.41	0.41	−0.92	0.77	0.77	0.13	0.24	0.037	0.002	0.321	0.011	0.075	0.002	0.314	0.015	4.20	0.23
		#4	−0.50	0.40	0.40	−0.90	0.77	0.77	−0.03	0.24	0.037	0.001	0.295	0.010	0.072	0.002	0.283	0.002	3.92	0.10
		Average & 2SD	−0.57	0.11	0.11	−0.96	0.26	0.26	0.05	0.21	0.038	0.010	0.313	0.017	0.072	0.013	0.189	0.255	2.62	3.57
503
		#1	−1.72	0.45	0.45	−3.34	0.93	0.93	−0.07	0.21	0.039	0.001	0.212	0.008	0.062	0.003	0.088	0.002	1.42	0.07
		#2	−1.85	0.44	0.44	−3.48	0.91	0.91	0.06	0.21	0.063	0.001	0.164	0.006	0.061	0.002	0.024	0.002	0.40	0.04
		#3	−2.06	0.46	0.46	−3.87	0.94	0.94	0.05	0.21	0.052	0.003	0.195	0.008	0.063	0.002	0.042	0.002	0.67	0.05
		#4	−2.26	0.47	0.47	−4.24	0.94	0.94	0.06	0.21	0.052	0.002	0.152	0.005	0.072	0.004	0.067	0.003	0.93	0.07
		Average & 2SD	−1.97	0.48	0.48	−3.74	0.81	0.81	0.03	0.13	0.051	0.020	0.181	0.039	0.065	0.010	0.055	0.056	0.86	0.87
505
		#1	−0.86	0.46	0.46	−1.51	0.94	0.94	0.08	0.21	0.060	0.001	0.259	0.010	0.080	0.004	0.135	0.006	1.70	0.11
		#2	−0.68	0.40	0.40	−1.18	0.84	0.84	0.13	0.21	0.079	0.003	0.128	0.004	0.058	0.003	0.026	0.002	0.45	0.05
		#3	−1.17	0.42	0.42	−2.30	0.88	0.88	−0.06	0.21	0.066	0.002	0.179	0.006	0.062	0.002	0.073	0.002	1.18	0.05
		#4	−1.06	0.41	0.41	−2.15	0.86	0.86	−0.10	0.21	0.075	0.003	0.129	0.004	0.057	0.002	0.019	0.002	0.33	0.03
		Average & 2SD	−0.94	0.44	0.44	−1.79	1.05	1.05	0.02	0.22	0.070	0.017	0.174	0.087	0.064	0.021	0.063	0.108	0.99	1.71
511
		#1	−2.36	0.44	0.44	−4.49	0.86	0.86	0.08	0.25	0.090	0.002	0.096	0.006	0.046	0.003	0.020	0.003	0.44	0.06
		#2	−2.91	0.48	0.48	−5.57	0.92	0.92	0.00	0.25	0.077	0.002	0.123	0.005	0.044	0.003	0.090	0.005	2.02	0.16
		#3	−2.52	0.43	0.43	−5.02	0.85	0.85	−0.15	0.25	0.101	0.002	0.128	0.007	0.044	0.001	0.095	0.002	2.14	0.08
		#4	−3.34	0.46	0.46	−6.41	0.89	0.89	0.03	0.25	0.109	0.005	0.113	0.007	0.056	0.003	0.005	0.001	0.09	0.01
		Average & 2SD	−2.78	0.87	0.87	−5.37	1.64	1.64	−0.01	0.20	0.094	0.027	0.115	0.020	0.048	0.011	0.052	0.093	1.10	1.97
512
		#1	−1.09	0.47	0.47	−1.96	0.94	0.94	0.05	0.21	0.040	0.001	0.273	0.009	0.068	0.002	0.215	0.003	3.18	0.10
		#2	−1.10	0.47	0.47	−1.97	0.94	0.94	0.06	0.21	0.037	0.001	0.233	0.008	0.068	0.002	0.088	0.002	1.28	0.05
		#3	−1.25	0.46	0.46	−2.30	0.92	0.92	0.03	0.21	0.032	0.001	0.245	0.008	0.066	0.002	0.048	0.002	0.73	0.04
		Average & 2SD	−1.15	0.18	0.18	−2.08	0.38	0.38	0.05	0.04	0.036	0.008	0.250	0.029	0.067	0.003	0.117	0.174	1.74	2.59
513
		#1	−0.47	0.42	0.42	−0.78	0.82	0.82	0.13	0.25	0.269	0.006	0.055	0.003	0.042	0.003	0.0039	0.0009	0.09	0.02
		#2	−0.83	0.45	0.45	−1.44	0.87	0.87	0.12	0.25	0.249	0.006	0.062	0.004	0.032	0.001	0.0046	0.0011	0.14	0.03
		#3	−0.57	0.44	0.44	−1.02	0.86	0.86	0.06	0.25	0.283	0.007	0.078	0.004	0.039	0.001	0.0040	0.0005	0.10	0.01
		Average & 2SD	−0.62	0.37	0.37	−1.08	0.67	0.67	0.10	0.08	0.267	0.035	0.065	0.017	0.038	0.010	0.004	0.001	0.11	0.04

^aAnalysis points are shown in Electronic Annex EA3.

^bAverage values are listed for minor element data. Individual data are reported in Electronic Annex EA4–3–3.

^cThese data are obtained from an outer left part of AOA 501, which shows compact textures relative to the most part of this AOA.

Fig. 7.

Magnesium-three isotope ratios of the eight AOA olivines from Kaba and DOM 08006. All measured data are shown in (a). (b-i) Individual AOA diagrams from Kaba (b) and DOM 08006 (c-i). Square symbols represent the Mg isotope ratios of bulk chondrites ([1] Teng et al., 2010). Solid lines represent a best-fit line for all the measured data determined by a least squares fitting (Ludwig, 2012). Errors are 2σ.

Fig. 8.

Relationships between δ²⁵Mg′_DSM-3 and δ²⁶Mg′_DSM-3 values of individual AOA olivine analyses. The square symbol represents the Mg isotope ratios of bulk chondrites ([1] Teng et al., 2010). The solid line represents a linear regression line (Isoplot model 1; Ludwig, 2012) for all AOA data (N = 36) obtained in this study. The dotted curves are error envelopes that show 95% confidence limits. Note that uncertainties of δ²⁵Mg′_DSM-3 and δ²⁶Mg′_DSM-3 values were not propagated from those related to instrumental mass fractionation. See more details in Electronic Annex EA5.

Magnesium isotope ratios (δ²⁵Mg and δ²⁶Mg) of Wild 2 olivine particles T57/F10, T77/F4, T77/F50, T149/F1, and T175/F1 are listed in Table 5 and plotted in Fig. 9. Difficulties with the analyses of three Wild 2 particles (T77/F6, T77/F7, and T149/F11a) or data reduction problems prevented the reporting of bias-corrected δ²⁵Mg and δ²⁶Mg values. For analyses of T77/F6 and T77/F7, (Mg⁺/Si⁺/Fo) values of these particles are outside the range of the instrumental fractionation calibration curve, meaning we could not properly correct data for instrumental mass fractionation. For T149/F11a, we observed that the Mg ion yields and δ²⁵Mg_m values changed significantly during the analysis. These changes were not observed among olivine RM analyses, meaning we could not properly correct data for instrumental mass fractionation. More detailed explanations for these three particles can be found in Electronic Annex EA7. All individual data, including raw-measured Mg isotope ratios of the eight Wild 2 particles, are reported in Electronic Annex EA4–2–5. Regarding the five Wild 2 particles for which Mg isotope ratios were successfully determined, they show slight variations in δ²⁵Mg, ranging from −1.0 ^+0.4/_−0.5‰ (2σ) to 0.6 ^+0.5/_−0.6‰ (2σ). The Mg isotopic data from these five Wild 2 particles fall on the same slope 0.516 line that is defined by individual AOA olivines (Fig. 9). Particles T77/F50 and T149/F1 were analyzed multiple times (N = 3 and N = 4, respectively) and do not show intra-grain variations in δ²⁵Mg and δ²⁶Mg values beyond uncertainties (Table 5 and Fig. 9).

Table 5.

Magnesium isotope ratios and minor elemental abundances of Wild 2 olivine particles obtained with SIMS

Sample/Location #a		δ25Mgdsm-3 (‰)	Unc. (+/−) (2σ)		δ26Mgdsm-3 (‰)	Unc. (+/−) (2σ)		δ26Mg* (‰)	Unc. (2σ)	CaO (wt%)	Unc. (2σ)	Cr2O3 (wt%)	Unc. (2σ)	FeO (wt%)	Unc. (2σ)	MnO (wt%)	Unc. (2σ)	MnO/FeO	Unc. (2σ)
16O-depleted particles
T77/F4
	#1	−0.17	0.77	0.51	−0.46	1.17	0.87	−0.07	0.43	0.186	0.015	0.07	0.01	29.72	0.99	0.481	0.005	0.016	0.001
T149/F1
	#1	−0.53	0.70	0.70	−0.74	1.05	1.05	−0.06	0.76	0.115	0.009	0.51	0.05	12.24	0.41	0.332	0.008	0.027	0.001
	#2	−0.40	0.70	0.70	−0.78	1.05	1.05	−0.33	0.76	0.116	0.009	0.53	0.06	12.84	0.43	0.342	0.008	0.027	0.001
	#3	−0.36	0.70	0.70	−0.77	1.05	1.05	−0.41	0.76	---	---	---	---	---	---	---	---	---	---
	#4	−0.70	0.70	0.70	−0.87	1.05	1.05	0.16	0.76	---	---	---	---	---	---	---	---	---	---
	Average & 2SD	−0.50	0.30	0.30	−0.79	0.11	0.11	−0.16	0.52	0.116	0.001	0.52	0.03	12.54	0.85	0.337	0.014	0.027	0.002
16O-rich particles
T57/F10
	#1	0.56	0.47	0.63	1.40	0.97	1.30	0.15	0.45	0.043	0.003	0.42	0.05	1.75	0.06	0.664	0.016	0.38	0.02
T77/F50
	#1	−0.97	0.41	0.49	−1.68	0.76	1.01	0.08	0.71	---	---	---	---	---	---	---	---	---	---
	#2	−0.93	0.41	0.49	−1.69	0.76	1.01	−0.01	0.70	0.031	0.002	0.31	0.03	0.065	0.003	0.247	0.004	3.80	0.16
	#3	−0.60	0.41	0.49	−0.94	0.76	1.01	0.11	0.70	---	---	---	---	---	---	---	---	---	---
	Average &2SD	−0.83	0.40	0.40	−1.44	0.87	0.87	0.06	0.13
T175/F1
	#1	−0.48	0.51	0.62	−0.47	1.04	1.26	0.31	0.37	0.0038	0.0003	0.15	0.02	0.151	0.015	0.055	0.004	0.37	0.05

^aAnalysis points are shown in Electronic Annex EA1.

Fig. 9.

Magnesium-three isotope ratios of the five Wild 2 olivine particles studied. The red square symbol represents the Mg isotope ratios of bulk chondrites ([1] Teng et al., 2010). All measured AOA data (Fig. 7a) are shown for comparison. Errors for Wild 2 olivine data are 2σ. For clarity, errors for AOA data are not shown.

δ²⁶Mg* values of AOA olivines and Wild 2 olivines are reported in Tables 4 and 5, and are plotted against their δ²⁵Mg values in Fig. 10. δ²⁶Mg* values of AOA olivines from Kaba and DOM 08006 range from −0.15 ± 0.25‰ (2σ) to 0.17 ± 0.24‰ (2σ) and do not show resolvable excesses in ²⁶Mg when compared to inferred Solar System initial δ²⁶Mg* values (−0.040‰; Jacobsen et al., 2008 or −0.016‰; Larsen et al., 2011). δ²⁶Mg* values of Wild 2 olivines range from −0.41 ± 0.76‰ (2σ) to 0.31 ± 0.37‰ (2σ) and do not show resolvable excesses in ²⁶Mg.

Fig. 10.

δ²⁶Mg*-δ²⁵Mg_DSM-3 plot for AOA olivines (Kaba and DOM 08006) and the five Wild 2 olivine particles studied.

4.4. Minor element abundances of AOA and Wild 2 olivines obtained with SIMS

CaO, Cr₂O₃, MnO and FeO contents in AOA and Wild 2 olivines were determined with a spatial resolution of ~1.5 μm by SIMS (Tables 4 and 5). Olivines from Kaba and DOM 08006 AOAs show one or two orders of magnitude variation in CaO, Cr₂O₃, and MnO contents ranging from 0.03 to 0.28 wt%, 0.06 to 0.32 wt%, and 0.004 to 0.31 wt%, respectively. FeO contents in olivines from DOM 08006 AOAs do not show significant variation among the seven examples investigated (0.03–0.08 wt%; Fig. 11). In contrast, olivines from Kaba AOA K27 are more variable (0.06–0.58 wt%) than DOM 08006 AOAs (Fig. 11). MnO/FeO (wt%) ratios of 38 olivine analyses show a large variation (0.1–4.2) even within a single inclusion (e.g., 0.1–2.1 in AOA511; see Table 4).

Fig. 11.

FeO contents (wt%) of AOA olivines from Kaba and DOM 08006. Data obtained with FE-EPMA (open symbols) and SIMS (closed symbols) are shown. Data obtained with FE-EPMA might be overestimated due to secondary fluorescence effects (see section 5.3.2). The dotted line represents the detection limit (0.05 wt%) for FE-EPMA analyses.

CaO, Cr₂O₃, and MnO contents in five Wild 2 olivine particles (T57/F10, T77/F4, T77/F50, T149/F1, and T175/F1) range from 0.004 to 0.19 wt%, 0.07 to 0.53 wt%, and 0.06 to 0.66 wt%, respectively. The ¹⁶O-depleted, FeO-rich Wild 2 olivines (T77/F4 and T149/F1) have higher CaO contents (0.12–0.19 wt%) than the ¹⁶O-rich, FeO-poor Wild 2 olivines (0.004–0.04 wt%). The MnO/FeO (wt%) ratio of ¹⁶O-rich olivine from T77/F50 is 3.8, is consistent with LIME olivine (Klöck et al., 1989). The MnO/FeO (wt%) ratios of the other ¹⁶O-rich olivines (T57/F10 and T175/F1) are 0.38 and 0.37, respectively. ¹⁶O-depleted Wild 2 olivines T77/F4 and T149/F1 have very low MnO/FeO ratios (≤0.03) due to their abundant FeO contents.

5. DISCUSSION

5.1. Distribution of O isotopes at the AOA-forming regions

Ushikubo et al. (2017) conducted O isotope analyses of individual AOA minerals from the Acfer 094 (C-ungrouped 3.00) carbonaceous chondrite and observed subtle variations in O isotope ratios along slope ~1 lines (CCAM and PCM). Some AOA olivines studied here show similar detectable variability (Fig. 6a), indicating slight oxygen isotope variation of early solar nebular gas (Ushikubo et al., 2017). However, these variations are very small compared to the differences between O isotope ratios of AOAs and more ¹⁶O-depleted objects (e.g., chondrules). In fact, the averaged O isotope ratios of individual AOAs from Kaba and DOM 08006 cannot be distinguished from each other (Fig. 6b). Those of AOAs from Acfer 094 (Ushikubo et al., 2017) also plot within the range of Kaba and DOM 08006 AOAs (Fig. 6b). These observations indicate that AOAs from Kaba, DOM 08006, and Acfer 094 formed in a single ¹⁶O-rich gaseous reservoir (average Δ¹⁷O = −23.9 ± 0.5‰; 2SD; N = 12; Ushikubo et al., 2017 and this study), probably near the proto-Sun (Krot et al., 2004) and were transported to accretion regions of each chondrite parent body.

In addition to AOAs, O isotope ratios of fine-grained CAIs from type 3.00–3.05 chondrites as well as spinel-hibonite inclusions (SHIBs) from CM chondrites are distributed between the CCAM and PCM lines with limited variations in Δ¹⁷O from −24‰ to −22‰ (Kööp et al., 2016; Ushikubo et al., 2017). Regarding Δ¹⁷O values, CAIs from the least metamorphosed CO and CR chondrites also exhibit consistently ¹⁶O-rich isotope characteristics (Δ¹⁷O ≤ −20‰; Makide et al., 2009; Bodénan et al., 2014; Simon et al., 2019; Zhang et al., 2020). The O isotope similarity among refractory inclusions from thermally unmetamorphosed chondrites suggests that they formed in a homogeneously ¹⁶O-rich reservoir.

5.2. Distribution of Mg isotopes at the AOA- and Wild 2 olivine-forming regions

We observe large variation in Mg isotopes of AOA olivine (Fig. 7). Other types of refractory inclusions (i.e., CAIs) also exhibit isotopic variations in multiple elements, including Mg (e.g., Liu et al., 2009; Wasserburg et al., 2012; MacPherson et al., 2017; Park et al., 2017), which are sometimes attributed to nucleosynthetic anomalies. However, a nucleosynthetic origin for the negative δ²⁵Mg and δ²⁶Mg values observed in AOA olivine seems unlikely because these values are linearly correlated with each other (Fig. 7), with a slope of ~0.5. Instead, this correlation is most likely the result of mass-dependent isotope fractionation. Based on theoretical studies, equilibrium and kinetic mass-dependent processes have different slopes in δ²⁶Mg′ versus δ²⁵Mg′ space (0.521 and 0.511, respectively; Young and Galy, 2004; Davis et al., 2015). The slope of the mass-dependent fractionation observed in the AOA olivines studied here (0.516 ± 0.012, 2σ; Fig. 8) is in between these expected slopes, and therefore cannot be distinguished as being the result of an equilibrium or kinetic process.

Teng et al. (2010) conducted Mg isotope analyses of 38 chondrites which show indistinguishable ratios (δ²⁵Mg = −0.15 ± 0.04‰, δ²⁶Mg = −0.28 ± 0.07‰, 2SD). In this study, the Mg isotope regression for AOA olivines is δ²⁵Mg′ = (−0.06 ± 0.04) + (0.516 ± 0.012) × δ²⁶Mg′ (Fig. 8). If using the δ²⁶Mg′ value of bulk chondrites (Teng et al., 2010) in this formula, the calculated δ²⁵Mg′ is −0.20 ± 0.04‰ (95% confidence limit), which is indistinguishable from that of the bulk chondrites within their uncertainties. This suggests that AOA olivines originated from a reservoir with a chondritic Mg isotope ratio.

Five Wild 2 particles show resolvable variability in Mg isotope ratios (Fig. 9). All Wild 2 olivine analyses fall on a mass-dependent fractionation line defined by individual AOA analyses (Fig. 9), suggesting Wild 2 olivine particles also derived from a chondritic Mg isotope reservoir. While uncertainties from comet particles are relatively larger, our data demonstrate for the first time that Mg isotopes are in agreement between Wild 2 comet olivine particles and other known Solar System solids.

5.3. Origin of AOAs: Condensation and thermal processing in the solar protoplanetary disk

5.3.1. Variations in texture among AOAs

Earlier studies have hypothesized a condensation origin for AOAs based on textural and mineralogical signatures (Krot et al., 2004a and references therein). The porous and irregular shapes of some AOAs indicate that they are aggregates of olivine grains and were never completely molten (e.g., Grossman and Steele, 1976; Krot et al., 2004a and references therein). Also, the presence of silica in an AOA from the Yamato-793261 (CR) is indicative of gas-solid condensation (Komatsu et al., 2018). The AOAs studied here show textural variations (Figs. 3–5), allowing for interpreting their origins. Three of them (K27, 505 and 513) have a compact texture with little porosity in their forsterite portions (Fig. 3), similar to “compact AOAs” from Yamato-81020 (CO3.05) (Sugiura et al., 2009). In contrast, AOAs 502 and 503 are porous, irregularly shaped objects (Fig. 4), similar to “porous AOAs” reported in Sugiura et al. (2009). Compact and porous AOAs are found in other types of carbonaceous chondrites (e.g., Krot et al., 2004a; Weisberg et al., 2004; Ruzicka et al., 2012; Rubin, 2013; Krot et al., 2014; Han and Brearley, 2015; Komatsu et al., 2015), suggesting this textural variation is a general characteristic among AOAs. From our observations, we categorize the nine AOAs investigated here into two types of AOAs; (1) three AOAs, K27, 505, and 513 are categorized as compact AOAs, as they have almost no pores and show a concentric texture with spinel and/or anorthite at the core → Al-diopside → forsterite (Figs. 3c, e, g); and (2) those with pore-rich, irregular-shaped, ±fine-grained textures are categorized as pore-rich AOAs (501, 502, 503, 511, 512, 514; Figs. 4–5). Spinel grains are observed in compact AOAs, but are not present in pore-rich AOAs of this study. Conversely, diffuse Ca-Al-rich haloes are observed in pore-rich AOAs, but are not a feature of compact AOAs. Moreover, the pores within pore-rich AOAs have polygonal outlines (e.g., Figs. 4b–d and 4f–g), suggesting these AOAs avoided melting (Han and Brearley, 2015).

Many AOAs do not appear to be the result of simple condensation (e.g., Komatsu et al., 2009; 2015). As shown in Fig. 3, compact AOA olivines have low porosity textures, indicating that they experienced high-temperature annealing after their condensation and aggregation (e.g., Krot et al., 2004a; Sugiura et al., 2009; Komatsu et al., 2009; Han and Brearley, 2015). In addition, compact AOAs K27 and 513 contain 5–10 μm size anorthite that is rimmed by 5–10 μm thick Al-diopside, which is in turn rimmed by olivine grains (Figs. 3c and 3e). This texture is not consistent with a sequence that is predicted by equilibrium condensation calculations (Petaev and Wood, 2005). Komatsu et al. (2009) performed heating experiments on mineral assemblages similar to those of AOAs and suggested that Al-diopside in AOAs can be produced from a small degree of melting of forsterite and anorthite. Based on the compact texture and the occurrence of Al-diopside between forsterite and anorthite (Figs. 3c and 3e), it appears that some AOAs experienced high-temperature annealing.

Similar occurrences of Al-diopside between forsterite and anorthite are also found among pore-rich AOAs (e.g., Figs. 5b–c), but at much finer scales. In addition, minor low-Ca pyroxenes are found in two pore-rich AOAs (502 and 503), which occur with Fe,Ni-metal (Figs. 4b, 4d, 4g). Similar textural occurrences of low-Ca pyroxene are present in AOAs from CR chondrites (Krot et al., 2004b). The existence of low-Ca pyroxene suggests interaction with gaseous SiO or direct condensation (Krot et al., 2004b). These observations indicate that some pore-rich AOAs have also experienced high-temperature annealing, although to a smaller degree relative to compact AOAs.

5.3.2. Variations in minor element abundances among AOAs

Among the AOAs investigated there is an inconsistency in FeO contents determined by EPMA versus SIMS (Fig. 11). Previous studies employing EPMA found that AOA olivines from the least metamorphosed chondrites have detectable FeO (~1 wt%; e.g., Krot et al., 2004a; Weisberg et al., 2004; Itoh et al., 2007; Sugiura et al., 2009; Komatsu et al., 2015; 2018), most of which are more enriched than concentrations predicted by equilibrium condensation calculations (e.g., ~0.14 wt% at a total pressure of 10⁻⁴ bar and with a dust enrichment factor of 1 (nominal solar composition); Sugiura et al., 2009). The enrichment in FeO suggests condensation of olivine at more oxidizing conditions (= high dust/gas ratios) (e.g., Komatsu et al., 2015; 2018). Our FE-EPMA-derived FeO contents from AOA olivines are variable (~0.05–1 wt%; open symbols in Fig. 11), with some values that are also much higher than that predicted by Sugiura et al. (2009). In contrast, our SIMS data reveal uniform olivine FeO contents of ~0.05 wt% among the AOAs studied (with a few exceptions for Kaba AOA K27; Fig. 11), which is consistent with equilibrium condensation calculations with a nominal solar composition gas from Sugiura et al. (2009). The uniform olivine FeO contents suggest that AOA olivine formed at more reducing conditions than previously thought. As higher FeO is mostly observed in porous AOAs, Sugiura et al. (2009) pointed out the possibility of contributions from FeO-bearing matrix materials during olivine EPMA analyses. Our present results also indicate that FeO contents of AOA olivine determined by EPMA analyses tend to be influenced by fluorescence from surrounding matrix materials and/or Fe,Ni-metals in AOAs, and that this may result in an overestimation of FeO contents and in an underestimation of MnO/FeO ratios in AOA olivine. The similar problem on EPMA analyses has also been pointed out by Wasson et al. (1994), who demonstrated that FeO contents in enstatite grains separated from EL6 meteorite chips are an order of magnitude lower than those in meteorite thin sections, suggesting a secondary fluorescence and/or double backscattering of electrons from surrounding opaque minerals. Therefore, we only refer to minor element abundances, including FeO contents, obtained by SIMS in the following discussion.

In Figure 12 CrO and MnO contents from SIMS olivine analyses are plotted as functions of CaO (Figs. 12a–b) and FeO (Figs. 12c–d), along with predicted values from Sugiura et al. (2009). Here we use CrO contents instead of Cr₂O₃ to compare SIMS data and predicted values from Sugiura et al. (2009). Generally, CrO (and possibly MnO) contents of AOA olivine are anti-correlated with CaO contents (Figs. 12a–b). Some olivines in pore-rich AOAs are enriched in Cr and Mn and are depleted in Ca relative to those in compact AOAs, but ~70% plot within a narrow range (CaO ~0.05–0.2 wt%; Figs. 12a–b). The correlations between textures and volatility controlled elemental abundances of AOAs are consistent with observations from Sugiura et al. (2009) and Komatsu et al. (2015), and indicate that compact AOAs are more refractory than pore-rich AOAs. As shown in Figs. 12c–d, plots of MnO and CrO versus FeO contents in AOA olivine are similar to those predicted by condensation models with a dust enrichment factor of 1 (nominal solar composition), except for some Kaba K27 AOA data. While all DOM 08006 data show a limited range in FeO contents regardless of their textural variations (compact versus pore-rich), it is possible that some Kaba AOA olivines were influenced by parent body alteration, leading to subtle Fe addition. Using data from DOM 08006 alone, we find CrO, MnO, and FeO compositions of AOA olivine are slightly offset, but broadly consistent with the condensation paths calculated by Sugiura et al. (2009) (Figs. 12c–d). In contrast, when CrO and MnO are plotted against CaO (Figs. 12a–b), they are clearly offset from the condensation paths, which is similar to AOA data reported in Sugiura et al. (2009). These relationships could be the result of two possibilities: (1) equilibrium condensation with lower dust/gas ratios (dust enrichment factor < 1), or (2) disequilibrium condensation of AOA olivine. In the case of (1), the condensation paths in Figs. 12c–d could be shifted to a lower side with respect to FeO contents (X-axis) because disk environments with very low dust/gas ratios could have resulted in lower FeO contents of condensed olivine without significant changes in MnO and CrO contents (Sugiura et al., 2009; Komatsu et al., 2015). However, even in environments with lower dust/gas ratios, CaO contents in AOA olivine may not be significantly reduced when compared to those predicted from condensation calculations (Figs. 12a–b). In the case of (2), disequilibrium condensation could explain low concentrations of CaO relative to those predicted from a model condensation path. In the temperature range of equilibrium olivine condensation (e.g., 1300–1370 K at 10⁻⁴ bar with a dust enrichment factor of 1), most Ca is predicted to be consumed by the formation of Ca-Al-rich minerals that do not remain in the gas phase (Sugiura et al., 2009). Thus, if AOA olivine condensed from the gas phase and was never equilibrated with pre-existing Ca-Al-rich minerals, CaO contents in olivine would be lower than those expected by equilibrium condensation. In contrast, Mn and Cr would have been largely in the gas phase, such that MnO and CrO abundances versus CaO contents in AOA olivine would plot parallel to the equilibrium condensation trajectories (Figs. 12a–d). Therefore, the observed offsets from condensation paths indicate that AOAs formed by disequilibrium condensation.

Fig. 12.

Minor element systematics of ¹⁶O-rich Wild 2 olivine particles (T57/F10, T77/F50, and T175/F1) and AOA olivines from Kaba and DOM 08006 obtained by SIMS. (a and c) CrO contents as functions of (a) CaO and (c) FeO. (b and d) MnO contents as functions of (b) CaO and (d) FeO. Equilibrium condensation trajectories at 10⁻⁴ bar with dust enrichment factor = 1 (nominal solar composition shown by solid curve), 2 (dashed curve), and 5 (dotted curve) calculated by Sugiura et al. (2009) are shown for comparison. Data for T57/F10 are not shown in (c) and (d) because of its high FeO content (1.75 wt%).

5.3.3. Kinetic isotope fractionation during condensation

During condensation, mineral phases can become enriched in light isotopes due to disequilibrium effects. For example, Richter (2004) modeled Mg isotope fractionation during condensation and demonstrated that Mg isotopes of the condensate are enriched in lighter isotopes when the ambient temperature changes faster than the condensation timescale. In this condition, the pressure of Mg gas is greater than its saturation vapor pressure, which facilitates disequilibrium condensation. This supersaturated condition could produce condensates enriched in lighter isotopes.

Fig. 13a shows δ²⁵Mg values for individual olivines from AOAs with different textures (compact versus pore-rich). Compact AOAs are more limited in δ²⁵Mg values (−1.2‰ to chondritic), while pore-rich AOAs exhibit greater variability (−3.8‰ to chondritic). As pore-rich AOAs may have experienced limited post-formation annealing, their larger and more negative δ²⁵Mg variations may have been caused by disequilibrium condensation processing. As discussed earlier, variations of trace element abundances in AOA olivine suggest they are controlled by volatility even though concentrations do not exactly follow equilibrium condensation trajectories. The δ²⁵Mg values in pore-rich AOA olivine correlate positively with CrO and negatively with CaO contents, respectively (Figs. 13b and c), which could represent a condensation trajectory in a cooling nebula gas.

Fig. 13.

Relationships among Mg isotope ratios and minor element abundances of ¹⁶O-rich Wild 2 olivine particles T57/F10, T77/F50, and T175/F1, and AOA olivines from Kaba and DOM 08006 obtained by SIMS. Open circle symbols represent AOA olivines with compact textures (compact AOA), while closed circle, diamond, and half-closed square symbols represent AOA olivines with pore-rich textures (pore-rich AOA; see section 5.3.1 for details). The dotted lines represent the Mg isotope ratio of bulk chondrites (Teng et al., 2010). (a) δ²⁵Mg_DSM-3 values of three ¹⁶O-rich Wild 2 particles and AOA olivines with different textures (compact versus pore-rich). (b-e) δ²⁵Mg_DSM-3 values as functions of (b) CrO, (c) CaO, (d) MnO contents (wt%) and (e) MnO/FeO (wt%) ratio. Errors are 2σ.

A similar positive correlation between δ²⁵Mg and MnO concentration is expected, but is not clearly seen in Fig. 13d. Most AOAs have MnO contents < 0.1wt% and a large range of δ²⁵Mg. Equilibrium condensation calculations by Sugiura et al. (2009) predict that condensation of Mn into olivine should occur at a lower temperature than Cr, and that the increase of MnO with decreasing temperature is steeper than that of CrO (see Fig. 2 in Sugiura et al., 2009). If the cooling rate is faster than the equilibrium condensation timescale (i.e., disequilibrium condensation), and if the temperature is within the range of Mn condensation, Mn will not sufficiently condense into olivine. Thus, relatively lower MnO contents are consistent with kinetic mass-dependent isotope fractionation of Mg, due to supersaturated conditions during condensation of olivine.

Isotope fractionation of condensing forsterite has been modeled as functions of the condensation temperature and timescale parameter ε by Richter (2004),

$$\epsilon =\frac{{\tau }_{\mathit{\text{cond }}}}{{\tau }_T}=\frac{{\lambda }_T{V}_c{\rho }_i{\eta }^{-\frac{1}{3}}}{A{J}_{i,0}}$$ (1)

where τ_cond and τ_T are timescales for condensation and for ambient temperature change, respectively, V_c is the volume of the condensed phase, ρ_i is the molar density of i in the condensed phase, A is the surface area, and J_i,0 is the free evaporation rate at a specified temperature and pressure. The timescale of τ_T is an inverse of the cooling rate λ_T = −(1/T)dT/dt. As expected by this relationship, when the cooling rate is much faster than the timescale for condensation (i.e., larger ε value), it induces disequilibrium condensation and large isotopic fractionation by partial condensation. As a first approximation, the condensation temperature of AOA olivine is estimated by comparing the CrO content and the equilibrium condensation calculations at a total pressure of 10⁻⁴ bar and a dust enrichment factor of 1 (Sugiura et al., 2009). Fig. 14 shows δ²⁵Mg values of AOA olivines as a function of the estimated condensation temperatures, which are compared to the δ²⁵Mg trajectories during disequilibrium condensation modeled by Richter (2004). Since olivine crystallization would occur during disequilibrium condensation rather than during equilibrium condensation, the predicted temperatures are systematically higher than the actual temperatures depending on the extent of disequilibrium condensation. While there are uncertainties in our temperature estimates, pore-rich AOA olivine data are distributed between the two model disequilibrium condensation trajectories with ε =0.001 and 0.01, respectively (Fig. 14). The negative correlation among most pore-rich AOA data (except for two AOA 501 analyses) are in agreement with a small range of ε, suggesting that variations in δ²⁵Mg values result from differences in the condensation temperature at a fairly constant cooling rate. Note that the isotopic fractionation calculations by Richter (2004) were done at a total pressure of 10⁻³ bar, which is one order of magnitude higher than that of the condensation calculation by Sugiura et al. (2009) that we use to estimate condensation temperatures. The total pressure influences the parameter J_i,0 in Eq. (1) (see Richter, 2004 for more details), and decreases it by a factor of 3 when the total pressure is lowered by 10⁻¹ bar; in turn, this leads to increase ε by a factor of 3. Thus, it does not significantly affect our conclusion.

Fig. 14.

δ²⁵Mg_DSM-3 values as a function of estimated condensation temperatures for ¹⁶O-rich Wild 2 olivine particles T57/F10, T77/F50, and T175/F1, and AOA olivines from Kaba and DOM 08006. Condensation trajectories of δ²⁵Mg as a function of condensation temperature and the timescale parameter ε modeled by Richter (2004) are shown for comparison (solid line with ε = 0.01 and dotted line with ε = 0.001). Fractions of magnesium condensed as forsterite with ε = 0.01 are also shown at the top of the plot. Condensation temperatures (X axis for each data point) are estimated by comparing CrO contents obtained by SIMS and an equilibrium condensation calculation at 10⁻⁴ bar with a dust enrichment factor of 1 (nominal solar composition) from Sugiura et al. (2009). Although negative Mg isotope fractionations suggest formation by disequilibrium condensation, we assume the CrO contents of olivine reflect their condensation temperatures (see section 5.3.3 for more details). This assumption may result in additional uncertainties of estimated temperatures.

5.3.4. Effect of high-temperature annealing

If AOAs experienced high-temperature annealing in the solar nebula, Mg isotope signatures and correlated minor element concentrations that are seen in pore-rich AOA could have been modified from their primary values. At high temperatures, Cr could be lost and Mg isotope ratios could have exchanged with ambient gas (e.g., Solar δ²⁵Mg ~0). Also, as a result of the presence of Ca-Al-rich phases in AOAs, Ca would likely have diffused into olivine (Sugiura et al., 2009). Such processes could have modified δ²⁵Mg values of AOA olivine, making them negatively correlated with CrO and positively correlated with CaO. However, this is not observed for the AOA data in this study (Figs. 13b–c). In particular, the diffusivity of Cr in forsterite varies with Cr concentration, oxygen fugacity, and silica activity (a_SiO2) (Jollands et al., 2018). At the onset of olivine condensation (1370K at a total pressure of 10⁻⁴ bar and a dust enrichment factor of 1; Sugiura et al., 2009), the Cr diffusion rate in forsterite (along the c axis) ranges from 1 × 10⁻¹⁸ to 2 × 10⁻¹⁶ m²/s (Jollands et al., 2018), which is faster than the Mg diffusion rate in forsterite (5 × 10⁻¹⁹ m²/s; Chakraborty, 2010). Hence, high-temperature annealing would likely have resulted in the loss of Cr from olivine before Mg isotopes exchanged with ambient gas. Thus, the positive correlation between δ²⁵Mg and CrO contents in pore-rich AOAs (Fig. 13b) suggest that high-temperature annealing did not significantly modify the Mg isotope and chemical signatures of pore-rich AOAs, a finding which is consistent with textural observations, such as irregular shapes of these AOAs. Additionally, Ca diffusion in forsterite is comparable with that of Mg (1 × 10⁻¹⁹ m²/s at 1370 K along the c axis; Morioka, 1981; Coogan et al., 2005) meaning that the inverse relationship between δ²⁵Mg and CaO contents (Fig. 13c) also supports a conclusion that Mg isotope characteristics of pore-rich AOAs have not been modified by annealing.

As previously mentioned, a positive correlation between δ²⁵Mg and MnO contents (or MnO/FeO ratios) is not apparent (Figs. 13d–e), which is likely the result of disequilibrium condensation. Alternatively, high-temperature annealing could lead to a loss of Mn from olivine grains. If this is the case, MnO contents as well as MnO/FeO ratios of some AOAs might have been modified during the annealing. However, the Mn diffusion rate in forsterite, which is influenced by oxygen fugacity and silica activity, ranges from 3 × 10⁻²⁰ to 1 × 10⁻¹⁷ m²/s at 1370 K along the c axis (Jollands et al., 2016). This is similar to or slightly lower than the Mn-Mg inter-diffusion rate, which was determined by Morioka (1981) to be 4.2 × 10⁻¹⁷ m²/s at 1370 K along the c axis. Therefore, low MnO contents of pore-rich AOAs might not be caused by loss of Mn during high-temperature annealing.

5.3.5. Formation of compact AOAs

As mentioned above, δ²⁵Mg values of compact AOAs are nearly uniform (Fig. 13a), indicating partial isotope exchange with ambient gas during high-temperature annealing. Sugiura et al. (2009) proposed that compact AOA textures resulted from either annealing of pre-existing AOAs in the solar nebula or by slow cooling during condensation. Both hypotheses are consistent with the δ²⁵Mg uniformity in compact AOAs. If the former is the case, the duration of annealing would likely have been much longer than that for pore-rich AOAs, and may have induced Mg isotope exchange between AOA olivine and ambient gas. If the latter is the case, the cooling rate of compact AOAs would have been sufficiently slower than that of pore-rich AOAs (i.e., smaller ε value). Such slow cooling could achieve near-equilibrium condensation, which would suppress isotopic fractionation during condensation (e.g., dotted line, Fig. 14). This scenario is also consistent with the observed uniformity in compact AOA olivine δ²⁵Mg (Fig. 14). However, as discussed above, the mineralogical sequence observed in compact AOAs is inconsistent with an equilibrium condensation sequence (Petaev and Wood, 2005; Komatsu et al., 2009). In addition, minor element systematics in AOA olivines cannot be simply explained by equilibrium condensation (e.g., Fig. 12a). Therefore, long-term, high-temperature annealing of pre-existing AOAs appear to be the most likely scenario to explain the chemical, textural, and Mg isotopic signatures among compact AOAs, consistent with the conclusion of previous studies (e.g., Komatsu et al., 2001; Krot et al., 2004a; Han and Brearley, 2015; Marrocchi et al., 2019a). Regarding O isotope ratios, we do not find a resolvable difference between compact and pore-rich AOAs (Fig. 6b), suggesting that (1) the annealing process did not modify the O isotope ratios because of its slower diffusion rate than Mg (Ando et al., 1981), and/or (2) the annealing process occurred at ¹⁶O-rich environments similar to where pore-rich AOAs condensed.

One pore-rich AOA, 501 from DOM 08006, exhibits a bimodal distribution on a Mg three-isotope plot (Fig. 7c). The irregularly shaped Ca-Al-rich portions and abundant pores (Fig. 5b) indicate that the majority of this AOA has not experienced extensive annealing like compact AOAs. However, two Mg isotope analyses for AOA 501 do not follow the condensation trend observed in pore-rich AOAs (Figs. 13b–c), making it difficult to explain the δ²⁵Mg variation in AOA 501 by a simple condensation process. The near-chondritic δ²⁵Mg values are found only in the outer left portion of the AOA, which has a compact texture (SIMS pits #1 and #5 in Fig. 5c) compared to rest of the AOA (Fig. 5b). Thus, it is possible that Mg isotope ratios of the compact textured region exchanged with ambient gas during annealing. Porous and compact textures have also been found in a single AOA from Adelaide (ungrouped C), which does not show evidence for melting (AOA #34b; Krot et al., 2004b). This AOA mainly consists of porous forsterite regions with very fine-grained anorthite and Al-diopside, and these domains are surrounded by compact forsterite rims (see Fig. 4d in Krot et al., 2004b). As shown in Fig. 5c, Ca-Al-rich phases in the outer left part of the AOA 501 are concentrated centrally in this region, similar to nodules in compact AOA K27 (Fig. 3b). Note that, however, an analysis of the outer right portion of the AOA (location #7 in Electronic Annex EA3) reveals a negative Mg isotope ratio (δ²⁵Mg = −3.1 ± 0.4‰; 2σ), meaning there is no systematic relationship between analysis locations and δ²⁵Mg values in AOA 501. Furthermore, CrO contents of inner and outer regions are similar to each other. Since Cr diffusion rate in forsterite is faster than the Mg diffusion rate in forsterite, the similar CrO contents in both inner and outer regions may not support the Mg isotope exchange with ambient gas during annealing. Similarly, we did not observe notable differences in minor element abundances and oxygen isotope ratios among AOA 501 (Table 3 and 4).

Recently, Marrocchi et al. (2019a) conducted SIMS oxygen, magnesium, and silicon isotope analyses of AOA olivine from Kaba (CV3.1), NWA 5958 (ungrouped C2, CM-like), and MIL 07342 (CO3.0–3.2) and found large negative Si isotope fractionations in AOA olivine (down to −5‰/amu). In contrast, Mg isotope ratios of these AOA olivines are near-chondritic to positively fractionated (~2‰ in δ²⁵Mg). Based on a model involving the condensation of olivine, they concluded that AOA olivines formed by rapid condensation (~days to weeks) followed by Mg isotope homogenization during thermal annealing that did not affect primary Si isotope ratios, due to much lower diffusion rates of Si than Mg (Chakraborty, 2010). The textures of these AOAs indicate they are similar to compact AOAs in our study and their near-chondritic Mg isotope data are also generally consistent with this study. However, most δ²⁵Mg values reported in Marrocchi et al. (2019a) are slightly positively fractionated from the chondritic Mg isotopic compositions, which is not observed in this work, indicating that some AOAs experienced partial evaporative loss of Mg during annealing. Alternatively, this difference between datasets could be the result of complex matrix effects for SIMS Mg isotope analyses of olivine (Fukuda et al., 2020). Marrocchi et al. (2019a) also mentioned “complex” matrix effects on Mg isotope analyses, though they did not use the correction scheme employed here. Our preliminary results, that are calibrated using only Fo contents, also show slight positive δ²⁵Mg values of some AOA olivines (Fukuda et al., 2019a), similar to the results from Marrocchi et al. (2019a). Therefore, a component of the observed positive fractionations in AOA olivine from Marrocchi et al. (2019a) could be due to inaccurate corrections of SIMS matrix effects. For example, when we apply our new calibration method, that combines Fo contents and Mg⁺/Si⁺ ratios, δ²⁵Mg values do not show positive values, but instead range from −4‰ to bulk chondritic Mg isotope ratios (Fig. 7). These δ²⁵Mg values are in good agreement with δ²⁹Si values of AOA olivine (−5‰ < δ²⁹Si < 0‰) determined by Marrocchi et al. (2019a). This comparison is significant because similar δ²⁵Mg and δ²⁹Si values are predicted from theoretical calculations of isotope fractionation during disequilibrium condensation of olivine (Marrocchi et al., 2019a), supporting our conclusion that the pore-rich AOAs studied here preserved their primary condensation Mg isotopic fractionation characteristics.

5.4. Origin of ¹⁶O-rich Wild 2 olivine particles

Three Wild 2 particles studied here, T57/F10, T77/F50, and T175/F1, are forsteritic olivines (> Fo₉₈) with ¹⁶O-rich signatures (Nakashima et al., 2012; Defouilloy et al., 2017; Chaumard et al., 2018). TEM-EDX analyses revealed that one of the particles, T77/F50 is LIME olivine (Fo_99.8; Joswiak et al., 2012), which was confirmed by our SIMS-derived MnO/FeO (wt%) ratio of this particle (3.8 ± 0.2; 2σ). Magnesium isotope data from T77/F50 are negatively fractionated (δ²⁵Mg = −0.8 ± 0.4‰, δ²⁶Mg = −1.4 ± 0.9‰, 2SD; average values) relative to a chondritic composition. These values are similar to negative Mg isotope fractionations observed in some AOAs (Fig. 9), suggesting that the T77/F50 particle is linked to AOA olivines. In addition, minor element systematics of T77/F50 and their relationship with δ²⁵Mg are similar to those of pore-rich AOAs (Figs. 12a–d and Figs. 13b–e, respectively). In particular, the δ²⁵Mg and minor element values of T77/F50 cannot be distinguished from those of AOA 502, which is the most porous example studied here (Figs. 4a–d). Furthermore, the Δ¹⁷O value of T77/F50 (−23.9 ± 1.6‰, 2SD; Nakashima et al., 2012), is identical to that of AOA 502 (Δ¹⁷O = −24.2 ± 0.8‰, 2SD) within uncertainties. Thus, the chemical and isotopic agreement between T77/F50 and at least one pore-rich AOA indicates that T77/F50 likely formed by disequilibrium condensation from the solar nebula, within an AOA-forming region providing strong evidence for some materials in comet Wild 2 and chondrites to be sourced from similar nebular regions.

T175/F1 is also a nearly pure forsterite particle (Fo_99.8) with a Δ¹⁷O value of −23.3 ± 1.4‰ (2SD; Chaumard et al., 2018). However, unlike T77/F50, T175/F1 is not negatively fractionated in its Mg isotopes, beyond analytical uncertainties. On plots comparing δ²⁵Mg-CrO-MnO, T175/F1 falls within the range of compact AOAs (Figs. 13b and d–e), suggesting they are related. Note that, however, the CaO content of T175/F1 (38 ± 3 ppm; 2σ) is remarkably lower than those of AOAs (Fig. 13c). The FeO content of T175/F1 is slightly higher than AOAs from DOM 08006 and is more consistent with Kaba compact AOA K27 (Figs. 12c–d). The slight enrichment in FeO suggests that either (1) minor Fe addition occurred after initial formation, like Kaba AOA K27, or (2) T175/F1 condensed at a higher dust/gas ratio (dust enrichment factor > 1). In either case, it is difficult to explain the depletion of CaO compared to AOA olivines, meaning T175/F1 may have a different origin than AOA olivines and T77/F50. ¹⁶O-rich, FeO-poor olivines are also found in chondrules as relict materials (e.g., Ushikubo et al., 2012; Marrocchi et al., 2019b), but their CaO contents are similar to those of AOAs (Marrocchi et al., 2019b), meaning a link between T175/F1 and relict olivine is unlikely. Matrix materials of chondrites also include ¹⁶O-rich, FeO-poor olivines (e.g., Nagashima et al., 2015). The average Ca content of matrix olivine grains from the Kakangari meteorite is 0.13 ± 0.56 wt% (Nagashima et al., 2015), which is higher than that of T175/F1 by a factor of ~30, indicative of an origin that is different than that of T175/F1 olivine.

Frank et al. (2014) conducted minor element analyses of more than forty Wild 2 olivine particles by TEM-EDX. Of six aerogel tracks (T57, T61, T111, T112, T130, and T154) with forsteritic olivine (> Fo₉₅), four (T57, T61, T111, and T112) have CaO-poor olivine (< 0.09 wt%). Among them, T61 terminal particles have low FeO (< 0.10 wt%), MnO and Cr₂O₃ (< 0.07 wt%), and CaO (< 0.09 wt%), which are similar to concentrations observed in T175/F1. Frank et al. (2014) also conducted minor element analyses of olivines from various types of chondrites. Since forsteritic olivines with very low CaO are not found in chondrites, Frank et al. (2014) suggested a link between T61 terminal olivine particles and olivines from aubrites. If this is also the case for T175/F1, it implies highly reducing, ¹⁶O-rich environments existed in the solar protoplanetary disk at the time of T175/F1 formation.

T57/F10 olivine also has ¹⁶O-rich oxygen isotopic characteristics (Δ¹⁷O = −22.0 ± 1.3‰, 2SD; Nakashima et al., 2012) with a slight enrichment in FeO (1.75 ± 0.06 wt%; 2σ) compared to the other two ¹⁶O-rich Wild 2 and AOA olivines (~ 0.58 wt%) discussed above. This particle is not LIME olivine (MnO/FeO wt% = 0.38 ± 0.02; 2σ), though the MnO content is distinctively higher (MnO = 0.66 ± 0.02 wt%; 2σ) than those of the other two ¹⁶O-rich Wild 2 and AOA olivines (Fig. 12b). The Cr₂O₃ content is also slightly higher (CrO = 0.42 ± 0.05 wt%; 2σ) than those of the other two ¹⁶O-rich Wild 2 and AOA olivines. Hence, T57/F10 olivine is relatively enriched in the moderately to slightly refractory elements Mn, Cr, and Fe. We note that T57/F10 exhibits slightly positive δ²⁵Mg and ²⁶Mg values (δ²⁵Mg = 0.6 ^+0.5/_−0.6‰, δ²⁶Mg = 1.4 ^+1.0/_−1.3‰; 2σ) that are not observed in AOA olivines (Fig. 9). Therefore, if T57/F10 olivine formed by condensation, it likely occurred under different conditions relative to those of AOA olivines studied here. For example, Komatsu et al. (2015) modeled equilibrium condensation under various dust/gas ratios (dust enrichment factor = ~ 0.1–100) and predict that FeO and MnO contents, as well as the MnO/FeO ratio of T57/F10 are explained by equilibrium condensation at a higher dust/gas ratio (dust enrichment factor = ~20), and at relatively lower temperatures (< 1200 K). If we assume T57/F10 olivine formed by equilibrium condensation with a dust enrichment factor of ~20, most olivine (~90%) is predicted to have condensed at temperatures near 1500K (Komatsu et al., 2015), meaning that the fraction of olivines condensed at lower temperatures (< 1200 K) like T57/F10 would be very minor. In contrast, olivine grains with similar chemical compositions are commonly reported from Wild 2 particles, chondrite matrices (Frank et al., 2014), and IDPs (Klöck et al., 1989), indicative of their ubiquitous presence in the outer Solar System, which is inconsistent with the minor fraction predicted by equilibrium condensation calculation (Komatsu et al., 2015). Alternatively, a volatile-rich olivine like T57/F10 could have experienced a complex thermal history, involving re-equilibration with surrounding gas (Komatsu et al., 2015). Although oxygen isotope ratios of primary matrix minerals in chondrites are not well investigated due to their extremely small sizes (≤ 1 μm), Nagashima et al. (2015) reported ¹⁶O-rich matrix olivine grains from the Kakangari meteorite. This suggests a link between T57/F10 and matrix olivine grains. Further studies that combined oxygen and magnesium isotope analyses, and minor element abundance measurements could provide important constraints on relationships between cometary and interplanetary dust particles, and fine-grained materials in chondrite matrices.

5.5. Origin of ¹⁶O-depleted Wild 2 olivine particles

Olivine particles T77/F4 and T149/F1 are FeO-rich (Fo₆₀ for T77/F4; Joswiak et al., 2012 and Fo₈₅ for T149/F1; Defouilloy et al., 2017) and have ¹⁶O-depleted oxygen isotopic signatures (Δ¹⁷O = −1.5 ± 1.2‰ (2SD) for T77/F4; Nakashima et al., 2012 and Δ¹⁷O = −1.7 ± 1.5‰ (2σ) for T149/F1; Defouilloy et al., 2017). δ²⁵Mg values of T77/F4 and T149/F1 are −0.2 ^+0.8/_−0.5‰ (2σ) and −0.5 ± 0.3 ‰ (2SD; average value), respectively, both of which are near-chondritic values. As discussed earlier for ¹⁶O-rich Wild 2 olivines, near-chondritic Mg isotope ratios can be explained by equilibrium condensation. However, FeO contents of T77/F4 and T149/F1 are too high compared to those expected by equilibrium condensation (e.g., Komatsu et al., 2015), meaning such an origin is unlikely. Alternatively, FeO-rich and ¹⁶O-depleted oxygen isotopic signatures are well explained by formation at high dust/gas ratios (dust enrichment factor = ~2500) in the protoplanetary disk (Tenner et al., 2015). In addition, the Fo contents and oxygen isotope ratios of T77/F4 and T149/F1 olivines are similar to those of type II chondrule olivines from CR and CH chondrites, indicative of a possible genetic link (Nakashima et al., 2012; Defouilloy et al., 2017). The δ²⁵Mg values of T77/F4 and T149/F1 are within the range of those of bulk chondrules (e.g., Bouvier et al., 2013; Van Kooten et al., 2016; Olsen et al., 2016; Chen et al., 2018), supporting a link between Wild 2 ¹⁶O-depleted particles and type II chondrules in chondrites. Note that these high precision Mg stable isotope data are from bulk chondrule analyses, not from chondrule olivine themselves. Future high precision Mg isotope analyses of chondrule olivines will give an opportunity to discuss their relations with ¹⁶O-depleted Wild 2 olivines.

5.6. Implications for dust transportation in the solar protoplanetary disk

Magnesium isotope and minor element analyses of Wild 2 olivine particles suggest they underwent a variety of nebular processing. For example, oxygen and magnesium isotopes, and chemical compositions of Wild 2 LIME olivine T77/F50 are in excellent agreement with those of pore-rich AOA olivines from the DOM 08006 (CO3.01) carbonaceous chondrite. The Mg isotope ratios of T77/F50 are negatively fractionated, indicating this particle formed by disequilibrium condensation from the solar nebula. Since characteristic nebular temperatures beyond 5 astronomical units (AU) should not exceed 200 K (e.g., Desch et al., 2018), the condensation of T77/F50 from comet-forming regions (~ tens of AU) seems unlikely. Instead, it is more likely that T77/F50 condensed within the inner protoplanetary disk (< 1 AU), as ambient temperatures would have exceeded 1000 K during early Solar System evolution (Desch et al., 2018). This would also explain the ¹⁶O-rich signatures of some Wild 2 olivine particles, as measurements of solar wind infer the proto-Sun was ¹⁶O-rich (McKeegan et al., 2011). Therefore, the presence of ¹⁶O-rich “condensed” olivine in comet 81P/Wild 2 provides evidence for material transport from the inner part of the Solar System outward, to comet-forming regions.

Mechanisms for such transportation have been hypothesized by several studies (Ciesla., 2007; Nakamura et al., 2008; Hughes and Armitage, 2010; Nakashima et al., 2012; Joswiak et al., 2017), in order to explain the presence of ¹⁶O-rich CAI-like materials in comet 81P/Wild 2 (McKeegan et al., 2006; Simon et al., 2008; Matzel et al., 2010). Al-Mg isotope analyses of two ¹⁶O-rich CAI-like materials in comet 81P/Wild 2, Coki and Inti, indicate the absence of live ²⁶Al at the time of their formation (Ishii et al., 2010; Matzel et al., 2010), in which the ²⁶Al/²⁷Al upper limit of Coki was determined to be < 1 × 10⁻⁵. The absence of live ²⁶Al at the time of ¹⁶O-rich Wild 2 materials indicates that they formed either (1) very early (i.e., before injection and homogenization of ²⁶Al in the Solar System), or (2) late formation (> 1.7 Myr after CAIs: Matzel et al., 2010) after the decay of ²⁶Al if using a canonical (²⁶Al/²⁷Al)₀ of 5.2 × 10⁻⁵ (Jacobsen et al., 2008; Larsen et al., 2011), and assuming a homogeneous distribution of ²⁶Al in the early Solar System. In the case of (1), transport was efficient enough to deliver the earliest formed materials to the outer Solar System, which is consistent with earliest formed CAIs to be transported into the outer part of the solar protoplanetary disk (Yang and Ciesla, 2012; Fukuda et al., 2019b; Larsen et al., 2020). In the case of (2), no evidence of live ²⁶Al within ¹⁶O-rich Wild 2 materials indicates, conversely, that transport was not efficient enough to deliver the earliest formed materials with the canonical (²⁶Al/²⁷Al)₀ values from inner (< 1 AU) to the outer Solar System (~30 AU). In either case, the presence of the AOA-like olivine (T77/F50) could place important constraints on the efficiency of such transportation processes as well as residence time of dust in the solar protoplanetary disk. Due to the absence of Al-rich phases (e.g., anorthite) associated with the ¹⁶O-rich olivine particles studied here, we cannot constrain the initial abundances of ²⁶Al when T77/F50 formed. However, Al-Mg isotope analyses of AOAs from primitive chondrites suggest they formed around a time similar to the formation of CV3 CAIs (²⁶Al/²⁷Al ~5 × 10⁻⁵; Itoh et al., 2007; Sugiura and Krot, 2007; MacPherson et al., 2012; Ushikubo et al., 2017), except for some AOAs in CH chondrites, which could have recorded admixing of ²⁶Al in the earliest stage of the Solar System evolution and/or might have experienced thermal annealing by an impact-generated plume (Krot et al., 2014). On the basis of chemical and isotopic similarities between T77/F50 and AOA olivines from DOM 08006, T77/F50 may also have formed at the same time as CAIs and AOAs, suggesting that the efficiency of the transportation processes was sufficient to transport solid particles from the inner (< 1 AU) to the outer Solar System (~30 AU) at the earliest stage of the Solar System evolution. Recently, Neveu and Vernazza (2019) conducted thermal evolution simulations for IDP-like bodies including D-type asteroids that may have formed within comet-forming regions (Fujiya et al., 2019), and predicted accretion of IDP-like bodies at least 5–6 Myr after the formation of CAIs in CV3 chondrites. If true, the presence of AOA-like olivine in the comet 81P/Wild 2 would indicate that the residence time of dust particles exceeded 5–6 Myr during the earliest stages of the Solar System evolution.

6. CONCLUSIONS

Magnesium isotope ratios and minor element abundances of ¹⁶O-rich and ¹⁶O-depleted Wild 2 olivine particles were investigated by SIMS. Oxygen and Mg isotope ratios and minor element abundances of AOA olivines from Kaba and DOM 08006 carbonaceous chondrites were also determined in order to compare with those of ¹⁶O-rich Wild 2 olivine particles.

Oxygen isotope data from Kaba and DOM 08006 AOA olivine are identical within uncertainties; they are also in agreement with oxygen isotope data from Acfer 094 AOAs. These results indicate AOAs originated from a single homogeneous reservoir.

SIMS minor element analyses reveal uniform FeO contents of olivine in AOAs from DOM 08006 (~0.05 wt%), suggesting AOAs formed at more reducing environments in the solar nebula than previously thought. In contrast, EPMA-derived FeO contents of the same AOA olivine are systematically higher than those obtained by SIMS analyses, indicating that EPMA-derived FeO contents of AOA olivine tend to be influenced by fluorescence from surrounding matrix materials and/or Fe,Ni-metals in AOAs.

Magnesium isotopes of Kaba and DOM 08006 AOA olivines are negatively mass-dependently fractionated down to −3.8‰ from the chondritic value, indicative of a disequilibrium condensation origin for some AOAs. Olivine grains from compact AOAs show limited variations in δ²⁵Mg ranging from −1.2‰ to −0.2‰. In contrast, larger variations in δ²⁵Mg, ranging from −3.8‰ to −0.4‰, are observed in olivine grains from pore-rich AOAs. The latter data are positively correlated with their CrO contents, suggesting that the variation in Mg isotope ratios was related to condensation temperature during formation. In contrast, the limited variation in δ²⁵Mg of compact AOAs likely resulted from partial isotope exchange with ambient gas through high-temperature annealing.

Magnesium isotope ratios of five Wild 2 olivine particles are consistent with mass-dependent fractionation from a chondritic Mg isotopic composition, indicating that Wild 2 olivine particles formed in the chondritic reservoir with respect to Mg isotope compositions. Three ¹⁶O-rich Wild 2 olivines have small variations in δ²⁵Mg, ranging from −1.0‰ to 0.6‰. Considering their minor element abundances, these particles may have experienced complex thermal histories including condensation followed by annealing. Two ¹⁶O-depleted Wild 2 olivines have near-chondritic Mg isotope ratios, supporting an origin that is similar to meteoritic chondrules.

One ¹⁶O-rich Wild 2 LIME olivine particle, T77/F50, shows agreement in terms of chemical and isotopic compositions with olivines in pore-rich AOAs from DOM 08006, suggesting that T77/F50 formed by disequilibrium condensation near the proto-Sun. If this is the case, this olivine particle likely traveled to the outer solar protoplanetary disk and then was accreted during the formation of comet 81P/Wild 2.

Supplementary Material

EA1

Electronic Annex EA1: SE images of Wild 2 olivine particles after (1) FIB marking, (2) Mg isotope analysis, and (3) Minor element analysis

EA2

Electronic Annex EA2: Position of SIMS measurements on AOAs (Oxygen isotopes)

EA3

Electronic Annex EA3: Position of SIMS measurements on AOAs (Magnesium isotopes and minor elements)

EA4

Electronic Annex EA4: Research data (EA4–1. SIMS O isotope data; EA4–2. SIMS Mg isotope data; EA4–3. SIMS major and minor element analyses data; EA4–4. EPMA data; EA4–5. Time-constant values for three Ω¹² ohm feedback resisters)

EA5-1

Electronic Annex EA5: SIMS data reduction procedure

EA6

Electronic Annex EA6: Position of EPMA measurements on AOAs

EA7

Electronic Annex EA7: Descriptions for SIMS analyses of T77/F6, T77/F7, and T149/F11a