Atomic-scale Element and Isotopic Investigation of ²⁵Mg-rich Stardust from an H-burning Supernova

The ALH 77307 carbonaceous chondrite was collected by the ANtarctic Search for METeorites (ANSMET) Program from the Allan Hills of Antarctica and curated at NASA's Johnson Space Centre (JSC; Allen et al. 2011; Righter et al. 2014). This meteorite is considered among the most pristine and primitive meteorites ever discovered and has an average matrix-normalized presolar silicate grain abundance of 119–190 ppm. Its matrix is comprised of fine-grained (∼100–500 nm) primarily anhydrous phases, enriched in silicates, Fe-Ni sulfides, organics, and amorphous material (Brearley 1993). Consequently, presolar silicates have largely escaped the effects of thermal and aqueous alteration on their parent asteroid.

Characteristic O isotopic signatures consistent with external stellar systems were identified using NanoSIMS ion imaging with the Compagnie des Applications Mécaniques et Electroniques au Cinéma et á l'Atomistique (CAMECA) NanoSIMS 50L ion probe at the Centre for Microscopy, Characterisation, and Analysis (CMCA), University of Western Australia. Ion imaging throughout this project was conducted by rastering a 1.6–2 pA, <200 nm focused 16 keV Cs⁺ primary ion beam over a 20 μm square field of view. Initially negative secondary ions of ¹⁶O, ¹⁷O, ¹⁸O, ²⁸Si, ²⁹Si, ³⁰Si, and ³²S were measured using electron multipliers in multi-detection mode. Grains were considered presolar (formed in an external stellar system) when O isotopes exceeded 3σ from solar values (or "normal" grains in the same image; Zinner et al. 2003; Nguyen et al. 2007). The image run consisted of two frames over 1 hr. Isotopic delta values and O isotopic ratios across all techniques in this study were calculated using standard practices (Barnes et al. 1975; Lodders 2003). Chemical ion imaging was collected using NanoSIMS for improved accuracy when locating presolar grains within complex mineral matrices. Ion imaging focused on collection of ¹⁶O, ¹⁷O, ¹⁸O, ²⁴Mg ¹⁶O, ²⁷Al ¹⁶O, ⁵⁶Fe ¹⁶O, ⁴⁸Ti ¹⁶O, and SE imaging at 15 ms dwell times. The grain within this study was measured as part of a larger study representing the first measurement of O-rich presolar grains using APT (Nevill 2022). Consequently, to aid in locating each presolar grain within the matrix at higher precisions for focused ion beam (FIB) extraction, Mg, Fe, Al, and Ti elements were selected as they are representative of the most common major elements within presolar oxide and silicate grains. The data reduction used software developed by the team at NASA JSC. A detailed breakdown of this software and methods used for quasi-simultaneous arrival of secondary ions, electron multiplier dead times, and instrumental mass fractionation are discussed in Nguyen et al. (2022).

Initial chemical analysis and morphology of the target presolar grain and its mineralogical context with surrounding mineral matrices was determined using the Tescan Clara SEM at JdLC equipped with a high-sensitivity Ultim Max 170 SDD energy dispersive X-ray (EDX) detector. This EDX detector has a larger sensor than detectors used in previous studies of presolar grains, measuring at a higher sensitivity and accuracy when data are taken at the same beam energy as previous instruments. Point analysis and hyperspectral elemental maps were collected at 10 kV, with a beam current of 3 pA. Point analysis and hyperspectral elemental maps were processed using Aztec Oxford instruments software.

The specimen was extracted and milled into a needle-like shape for APT analysis (Figure A1) using the Tescan Lyra3 GM FIB-SEM at the JdLC, Curtin University. The specimen needle was milled 50 nm in diameter at the apex, with a ∼5° half-shank angle (Nevill 2022). Backscattered electron images (BSE), SE images, and EDX data were collected in the Tescan Lyra3 for targeting and correlating target submicrometer grains for extraction. Hyperspectral elemental maps in the Tescan Lyra3 were collected at 15–20 kV using a beam current of 1 nA as required across target regions and an Oxford X-max 50 EDX detector. APT needles were prepared using established methods with a few adaptations (Miller & Russell 2007; McKenzie et al. 2010; Miller & Forbes 2014; Lee & Ahn 2016; Nevill 2022). (1) A platinum (Pt) "button" fiducial was electron-beam deposited on the surface of the sample for improved centering at the apex of the needle over the desired grain or region per the requirements of the sample (Miller & Russell 2007). This button was visible beneath the subsequently deposited protective Pt layer (5 × 2 μm; Rickard et al. 2020). (2) The outer edges of the Pt layer were milled away to form a wedge-like shape (2.5 × 3.8 × 3.8 μm) using a 30 kV Ga⁺ ion beam (Miller & Forbes 2014; Lee & Ahn 2016). The shape, apex, and shank angles of the needle tips are designed to optimize controlled evaporation of specimen ions during acquisition. Final polishing of the apex of the specimen used an ion-beam accelerating voltage of 2 kV to remove the damaged/gallium implanted layer. During this process, unwanted sample coatings like carbon, deposited Pt and Cs⁺ ions (consistent with the NanoSIMS ion beam) were milled away. This stage is conducted visually under SEM conditions, permitting a degree of human error regarding the removal of surface coatings and remnants of previous techniques. Residual contaminants had no impact on APT data processing or interpretation and were readily identified from spectral and spatial distribution relationships. For example, naturally occurring Ga⁺ exhibits two isotopes (⁶⁹Ga⁺ and ⁷⁰Ga⁺) unlike the pure ⁶⁹Ga⁺ ion beam used within the FIB-SEM. Collection of BSE images throughout the preparation process enabled tracking of each stage of sample preparation and estimation of starting radii at the apex of each specimen as required for APT data processing.

Elemental abundances, subnanometer spatial distributions of each atom in 3D, and Mg isotopic compositions were measured using the CAMECA Local Electrode Atom Probe, LEAP 4000X HR, within the Geoscience Atom Probe facility at the JdLC, Curtin University. Isotopic compositions were calculated using custom methods detailed in Section 2.7.1. APT uses an electric field coupled with a high-frequency pulsed laser to desorb ions from the surface, resulting in ionization yields close to 100%. These ions are measured using a position-sensitive detector, with a detection efficiency of 35%. See Reddy et al. (2020) for a recent review of this technique and its application to the geosciences.

Data reconstruction and processing used CAMECA Atom Probe (AP) Suite v6.0 processing software and Integrated Visualization and Analysis Software (IVAS) v3.8. Experimental conditions, acquisition parameters, and data-reconstruction parameters are detailed per recommendations by Blum et al. (2017) in Table A1. Preselection of elements and standards are not required, with background noise, mass peak overlaps, and detector saturation assessed within distinct regions of individual data sets. Any residual coating contaminants were readily identified and eliminated from data sets.

Data quality is assessed using background noise, spatial resolution, mass resolving power (MRP), and multiple detector (multi-hit) events (Kinno et al. 2012; Kirchhofer et al. 2013; Thuvander et al. 2013; Meisenkothen et al. 2016, 2020; Blum et al. 2017; Reddy et al. 2020). N-AL6 has a high MPR (918.78) and spatial resolution, with elevated background levels and peak tails primarily associated with one hump in the lower end of the spectrum, which degraded into background noise and had minimal impact on measured isotopes.

2.7.1. Quantifying Atom Probe Tomography Isotopic Measurements

2.7.2. Corrections for Hydride Complexes

2.7.3. Isobaric Interferences

2.7.4. Uncertainty Calculations

The following equations detail the calculation of isotopic ratios using peaks without isobaric interferences and their calculated uncertainties. This equation considers background counts and the number of ions collected. All measurements were propagated to 1σ uncertainties.

(1) Calculating the isotopic ratio:

$$\begin{eqnarray}&&\displaystyle \frac{{A}_{\mathrm{Isotope}}}{{B}_{\mathrm{Isotope}}}=\ \displaystyle \frac{{A}_{C}}{{B}_{c}},\end{eqnarray} \tag{ 1 }$$

where A_isotope is the numerator ion of the isotopic ratio, B_isotope is the denominator ion of the isotopic ratio, A_c is the total value given when the background of ion A or the corrected value given by the software is subtracted from the raw ion counts, and B_c is the total value given when the background of ion B or the corrected value given by the software is subtracted from the raw ion counts.

(2) Calculating the uncertainty of m/z peaks A and B:

$$\begin{eqnarray}&&{A}_{u}=\ \sqrt{{\left(\sqrt{{A}_{r}}\right)}^{2}\ +2\times \ {\left(\sqrt{{A}_{b}}\right)}^{2}},\end{eqnarray} \tag{ 2 }$$

where A_u is the uncertainty of the A m/z peak ion count, A_r is the raw ion count of the A peak, and A_b is the background count (take the corrected value from raw counts to isolate the background if the background is software corrected). The same formula should be applied to the B isotope.

(3) Calculating the uncertainty of the isotopic ratio:

$$\begin{eqnarray}&&(\pm )=\displaystyle \frac{A}{B}\times \ \sqrt{{\left(\displaystyle \frac{{B}_{u}}{{B}_{c}}\right)}^{2}+{\left(\displaystyle \frac{{A}_{u}}{{A}_{c}}\right)}^{2}},\end{eqnarray} \tag{ 3 }$$

where ± denotes the uncertainty associated with the isotopic ratio, A_u is the uncertainty of the A m/z peak ion count, A_c is the background-corrected A value (see Equation (1)), B_u is the uncertainty of the B m/z peak ion count, and B_c is the background-corrected B value.

Delta ratio uncertainties were calculated by adding the calculated uncertainty to the isotopic ratio, determining the new delta ratio using the isotopic delta formula associated with that element (Section 2.4), and extracting it from the original delta value. The difference represents the delta uncertainty. The same calculation must then be conducted for confirmation whereby the uncertainty is instead subtracted. If accurate, its difference will yield the same value as the first equation.

Only ⁶⁹Ga⁺ was detected in N-AL6, suggesting its Pt and carbon coating cap was sufficiently removed during sample preparation. Consequently, contaminants had no impact on data-processing procedures or isotopic or geochemical calculations.

Presolar grain N-AL6 is a 400 × 580 nm silicate grain within ALH 77307 that has a triangular sectional profile with round corners. It has a ¹⁷O/¹⁶O isotopic ratio 3.2× solar values, consistent with classification as a Group 1 presolar grain (Figure 1). The ¹⁷O-rich isotopic composition is distinct from its surrounding matrix, highlighting a presolar origin and formation external to the parent body (Table 1; Figure 1).

Zoom In Zoom Out Reset image size

The Si isotopic compositions plot within the mainstream correlation line and main presolar cluster (Table 1; Figure 1). Comparisons of Si isotopes in NanoSIMS and APT show similar values within analytical uncertainties, with both exhibiting large errors which are attributed to matrix-dilution effects and counting statistics from the low abundance of less common isotopes ²⁹Si and ³⁰Si and small analytical volume, respectively (Figure 1). Nevertheless, NanoSIMS-measured values are used for further comparisons.

The APT mass spectra from N-AL6 showed peaks ²⁴Mg, ²⁵Mg, ²⁶Mg, ²⁸Si, ²⁹Si, and ³⁰Si at (+) and (2+) charge states, Al²⁷O¹⁶, and a series of Mg and Si oxide complexes, e.g., MgO, SiO, and SiO₂ (Table 1; Figure 2). Mg, O, Al, and Si ions were homogeneously distributed throughout both the 3D tomographic reconstruction and along the 1D concentration profile of the presolar grain (Figure 2).

Zoom In Zoom Out Reset image size

The APT data yield an elemental composition of Si (13.85 ± 0.15 at%), Mg (38.57 ± 0.15 at%), O (47.58 ± 0.19 at%), and Al (0.69 ± 0.04 at%). Comparisons of the APT-measured composition of N-AL6 with that of a terrestrial forsterite sample from the island of Holsnøy (SW Norway) analyzed under similar acquisition parameters, voltage ranges, and conditions (Tacchetto et al. 2021) demonstrates a similar stochiometric bias in the terrestrial olivine APT data, i.e., Si (14.94 at%), Mg (29.05 at%), Fe (9.04 at%), and O (46.98 at%). The quantification of oxygen anions via APT is known to be problematic due to the inability of APT to discriminate between the O⁺ and O₂ ⁺⁺ at a mass/charge of 16 Da (Reddy et al. 2020). Therefore, the APT results from N-AL6 are consistent with the analysis of stoichiometrically normal forsterite (Mg₂SiO₄). The theoretical value correcting for the over- and underestimation was calculated using total ion counts and correcting Mg and O for 10 at% each (Table 1).

The Mg isotopic compositions of N-AL6 calculated using APT data are δ²⁵Mg = 3025‰ ± 38‰ (2σ) and δ²⁶Mg = 746‰ ± 26‰ (2σ) (Table 1; Figure 3), which represents the highest value of δ²⁵Mg ever measured in a silicate grain (Figure 3). A comparison of δ²⁵Mg isotopic calculations from peaks Mg⁺⁺ (peak series represents ∼90% of Mg ions within spectra) and Mg⁺ (peak series with second highest number of counts) yielded δ²⁵Mg = 3025‰ ± 38‰ (Mg⁺⁺) and δ²⁵Mg = 3043‰ ± 383.9‰ (Mg⁺). The values derived from the Mg⁺⁺ and Mg⁺ peak series are similar within analytical uncertainties and demonstrate internal consistency of APT isotope results. To demonstrate the accuracy of these results, we calculated the Mg isotopic compositions of two terrestrial olivine grains from Holsnøy (SW Norway) measured under similar acquisition parameters (Tacchetto et al. 2021) to N-AL6 and similar distributions of ion counts within each Mg peak. Using the same method, isotopic compositions were calculated as δ²⁵Mg = −20.0‰ ± 11.9‰ and δ²⁶Mg = 14.5‰ ± 11.4‰ for grain 3521, and δ²⁵Mg = −28.9‰ ±20.1‰ and δ²⁶Mg = 17.9‰ ± 17.4‰ for grain 3442 (Tacchetto et al. 2021). A minor overestimation of δ²⁵Mg background correction during calculations resulted in slightly underestimated δ²⁵Mg values in both terrestrial grains and N-AL6, with terrestrial grains likely preserving isotopic signatures of δ²⁵Mg = 0‰. Results demonstrate that the measured anomalies in N-AL6 are accurate.

Zoom In Zoom Out Reset image size

Presolar grain N-AL6 is classified as a Group 1 presolar silicate due to its O isotopic signature and has an extreme δ²⁵Mg anomaly (δ²⁵Mg = 3025‰ ± 38‰ and δ²⁶Mg = 746‰ ± 26‰). Group 1 grains are characteristically ¹⁷O-rich with solar to slightly depleted ¹⁸O compositions (Nittler et al. 2008). Until recently, these Group 1 presolar grains were considered to have exclusively formed within the main-sequence RGB or low-mass AGB stars of close-to-solar metallicity (Boothroyd et al. 1994; Nittler et al. 2008). However, recent Mg isotopic studies using the Hyperion II RF plasma O primary ion source and 3D modeling of stellar atmospheres indicate that Group 1 and 2 presolar grains with δ²⁵Mg >600‰ originate from Type II SN of solar metallicity which undergo hydrogen ingestion into the helium (He) shell prior to the explosion (Leitner & Hoppe 2019; Hoppe et al. 2021). This is because comparisons between isotopic compositions of presolar grains and theoretical modeling found that ²⁵Mg excesses were challenging to reconcile with AGB nucleosynthesis irrespective of their metallicities. Nova models may account for some of the δ²⁵Mg-rich presolar grains considering they predict ¹⁷O-rich, slightly ¹⁸O-poor, ²⁵Mg-rich, and comparably low ²⁶Mg isotopic compositions. However, current nova models reveal serious discrepancies between measured grains and expected ratios (i.e., oxygen was 10× to 70× measured ¹⁷O/¹⁶O ratios) and dust production was inefficient to account for the percentage of grains attributed to this origin (Hoppe et al. 2021). This discrepancy may be due to the limited coverage of the range of conditions that could occur within these stellar environments (Iliadis et al. 2018), an issue which may be addressed in future 3D modeling of nova atmospheres to explore a wider range of environmental parameters, e.g., pressure and temperature, and achieve more realistic conditions, e.g., changes in mixing length which alters as a function of stellar surface gravity and temperature.

Recently, the Mg, Si, and O isotopic compositions of 14 presolar grains were compared with theoretical modeling and astronomical observations of AGB, nova, and H-ingestion CCSN stellar atmospheres, to substantiate the proposed H-ingestion CCSN origin for ²⁵Mg-rich Group 1 and 2 grains (Leitner & Hoppe 2019; Hoppe et al. 2021). The latter was conducted by artificially increasing temperature and density to mimic the effects of explosive H-burning within a 15 M_⊙ SN model, which assumed 1.2% H ingestion into the He shell before the explosion (Pignatari et al. 2015; Leitner & Hoppe 2019; Hoppe et al. 2021). This same model was recently used to show that some presolar silicon carbide (SiC) grains originated from a H-ingestion CCSN rather than a nova (Liu et al. 2016).

Within CCSNe, once the core of the star reaches a critical mass and gravitational collapse occurs, a nascent neutron star is formed. During gravitational collapse, the infalling mass supersonically crashes onto the neutron star, launching a shock wave which passes through the outer layers of the star and ejects them. These shock waves cause instabilities among the layers of the stellar system, resulting in large-scale mixing within rapid timeframes (Harris & Lambert 1984). As a result, each zone alters in composition, becoming chemically distinct. These layers are therefore named according to their most abundant element and characteristics (Figure 4). As the shock wave passes through the layers they heat and compress, then expand and cool. Explosive H-burning occurs in the He layers as the material responds to the heating and compression resulting from the shock passage, leaving characteristic isotopic and elemental signatures not found within traditional CCSNe, e.g., an O/nova zone with nucleosynthesis characteristics of nova environments including strong ²⁵Mg enrichments (Figure 4). The He shell is comprised of a C/Si zone, an O/nova zone (mass range 6.82–7.16 M_⊙), and He/C zones (mass range 7.16–9.23 M_⊙). For grains with δ²⁵Mg exceeding values of 1000‰, mixing calculations required zone-selective mixing from the inner C/Si zone (mass range 6.8145–6.8247 M_⊙), and the outer region of the O/nova zone (mass range 7.05–7.2 M_⊙).

Zoom In Zoom Out Reset image size

As measured Mg and Si isotopic fractionation were in good agreement with H-burning CCSN models but O isotopic fractionation in moderate agreement (Boothroyd et al. 1994; Leitner & Hoppe 2019; Hoppe et al. 2021), Leitner & Hoppe (2019) and Hoppe et al. (2021) took two approaches regarding O isotopes for their 14 presolar silicates. The first used O isotopes from a 15 M_⊙ SN, and the second enhanced ¹⁷O isotopes within a pre-SN wind taken from infrared observations of pre-SN supergiant Betelgeuse, as this star had similar He shell physics assumptions as a model, accounting for the expected difference in solar metallicity (Harris & Lambert 1984; Pignatari et al. 2015; Leitner & Hoppe 2019; Hoppe et al. 2021). Modeling predictions in Hoppe et al. (2021) were best matched when SN envelopes were mixed with outer SN zones and matter ejected during the pre-SN phase. Thus, isotopic measurements of δ²⁵Mg within studied presolar grains were modeled with mixtures of 63%–96% pre-SN wind material (Pignatari et al. 2016; Leitner & Hoppe 2019; Hoppe et al. 2021). Considering the limited parameters of SN models explored and assumptions used in modeling, it has been argued results are favorable with a H-burning CCSN stellar environment (Hoppe et al. 2021).

Modeling indicates H-burning CCSNe produce large ²⁵Mg excesses due to the ²⁴Mg(p, γ)²⁵Al(β⁺)²⁵Mg reaction, with ²⁶Mg enriched to a lesser degree in contrast to standard CCSN models (Leitner & Hoppe 2019; Hoppe et al. 2021), consistent with Mg isotopic measurements in N-AL6 (Table 1; Figure 3). N-AL6 Mg, O, and Si isotopic values are comparable to initial modeling discussions (Leitner & Hoppe 2019; Hoppe et al. 2021). For example, N-AL6 O and Si N-AL6 O and Si isotopic values were ¹⁷O/¹⁶O = 11.78 ± 0.24, ¹⁸O/¹⁶O = 1.49 ± 0.03, δ²⁹Si = 5.09 ± 29.05, and δ³⁰Si = −8.81 ± 35.67, falling well within the range of Group 1 ²⁵Mg-rich grains measured within Leitner & Hoppe (2019) and Hoppe et al. (2021).

In calculations without zone-selective mixing, where ²⁵Mg isotopes in N-AL6 could be matched, Si isotopes could not. When using the zone-selective mixing taken by Leitner & Hoppe (2019), ²⁵Mg ²⁶Mg, ²⁹Si, and ³⁰Si could all be well matched. In comparison to Hoppe et al. (2021) and Leitner & Hoppe (2019), N-AL6 was modeled with a 5% pre-SN wind mixture. For the pre-SN wind, we used the O isotopic ratios in a 15 M_⊙ star predicted by Pignatari et al. (2016) and O isotopic values measured in Betelgeuse (Harris & Lambert 1984). Within the bounds of the 15 M_⊙ star, the ¹⁷O/¹⁶O ratios were too low, deviating by ∼60%, and the ¹⁸O/¹⁶O ratios deviated by a factor of 10, as seen in Hoppe et al. (2021). When using Betelgeuse isotopic compositions, the ¹⁷O/¹⁶O ratios were within ∼30% and the ¹⁸O/¹⁶O ratios were off by a factor of 8. The O isotopic values showed better matches with ¹⁷O/¹⁶O ratios but poorer matches to ¹⁸O/¹⁶O ratios compared with Hoppe et al. (2021). As O isotopic compositions were a closer match to modeling of a H-burning CCSN than a nova, Hoppe et al. (2021) proposed that the deviations could be attributed to the limited number of tested metallicities and lack of 3D modeling of CCSNe, which is required to account for peak density, temperature, and hydrodynamic simulations of H ingestion (Pignatari et al. 2015). The results within this study are therefore in good agreement with a H-burning CCSN. Together, these calculations suggest CCSNe were significant suppliers of dust to the molecular cloud from which our solar system formed. However, we argue that novae should not be ruled out as a potential stellar origin for ²⁵Mg-rich presolar grains given the extreme ²⁵Mg values within nova atmospheres and lack of extensive modeling of nova atmospheres. Based on discussions of H-burning CCSNe and modeling calculations, it is clear that the isotopic signatures in N-AL6 challenge current stellar models, implying there may be processes occurring within stellar environments which we do not yet fully understand.

N-AL6 was inferred to be a stoichiometric forsterite based on the similarity of APT results from a stochiometric terrestrial olivine grain measured under similar analytical conditions. The stoichiometry of N-AL6 implies that it formed during thermodynamic equilibrium condensation within a circumstellar environment. Thermodynamic equilibrium calculations of SN, nova, and AGB/RGB stellar systems have each identified forsterite as a primary condensing phase and the first Mg-rich silicate to condense within stellar atmospheres under equilibrium conditions (Lodders 2003; José et al. 2004; Fedkin et al. 2010; Pignatari et al. 2015; Agúndez et al. 2020). This supports the formation of N-AL6 within an H-burning CCSN atmosphere. The triangular sectional profile and round corners of N-AL6 (Table 1) indicate a degree of post-processing during ejection or migration throughout the ISM from processes that did not affect the overall chemistry, e.g., if ionizing processes in the ISM were sufficient to chemically alter N-AL6 we would expect depletion of Si relative to Mg, which was not observed (Jones 2000; Lodders 2003; Takigawa et al. 2014). Further constraints on N-AL6 evolution requires studying its crystallinity (Zinner 2014).

The minor amounts of aluminum detected within N-AL6 could be explained by direct condensation in the stellar environment. Comparative studies of condensation in AGB/RGB, novae, and CCSNe show similar relationships between condensation and environmental conditions. For example, during condensation of presolar grains, compatible elements are incorporated in minor and trace abundances at the tail end of the range of pressure and temperature conditions each phase condenses at, and increasing pressure reduces the condensing temperature range for each phase (Lodders 2003). Environmental changes within circumstellar envelopes can affect the timing and pressure/temperature ranges of each phase, which influences the abundances of each element incorporated into the lattice at the time of condensation and the compatibility of elements within a condensing phase. Pressure and temperature variations have minimal influence on gas composition, which also plays a role in the uptake of minor and trace elements. Consequently, pressure and temperature are unlikely to have solely contributed to a lack of additional minor and trace elements in N-AL6 regardless of whether it condensed under equilibrium or disequilibrium conditions.

We propose the following. (1) The collapse of the CCSN could have been so rapid that mixing and homogenization could not be achieved on smaller scales, resulting in pockets of different gas compositions which could occur at localized scales within the condensing envelope. (2) N-AL6 could have condensed just before collapse of the SN, where mixing of SN layers is minimal, consistent with mixing calculations and measured elemental abundances. Because Al is also the first condensate, this element would likely have been present within the condensing region in both scenarios when forsterite, the first Mg-rich silicate, formed, accounting for the elemental composition of N-AL6. (3) The elemental composition of N-AL6 could also be explained if the gas was enriched in Mg, Si, and O as well as additional heavy elements that are incompatible within a forsterite lattice under the pressure and temperature conditions within the circumstellar envelope at the time of condensation. A more extensive study of presolar nucleosynthesis and condensing phases from H-burning CCSNe and thermodynamic modeling of pressure conditions and equilibrium condensation is required to further refine geochemical relationships within H-burning CCSN environments.

The isotopic and elemental compositions of N-AL6 represent new constraints on mixing and processing in stellar atmospheres. Particles from external stellar systems such as this are among the original building blocks of our solar system, providing important records of its evolution. Ultimately, the study of all isotopic and chemical signatures of presolar silicates, at nanoscale spatial resolution, may therefore help reconcile the differences between (i) thermodynamic modeling and measured isotopic and chemical compositions, and (ii) the physical and chemical processes responsible for reproducing the signatures measured at micrometric to atomic scales in presolar grains from processes and environmental conditions in our solar system and external stellar systems.

Peak 43 m/z is consistent with ²⁷Al¹⁶O⁺. However, there were no other Al peaks present. If all the Al ions were bonded to O within the presolar grain before acquisition and the energy within the APT was not significant enough to break the Al-O bonds, all ions would accumulate at the single oxide charge state, peak 43 m/z, explaining the lack of additional peaks. Alternatively, peak 27 m/z could be missing if the Al concentration was low and the field too high (giving a high Al⁺⁺/Al⁺ ratio). Given the proportional heights of each isotope within the other series of Mg peaks, peak 43 m/z was proportionally inconsistent with ²⁶MgOH⁺, and was therefore assigned ²⁷Al¹⁶O⁺.

The following figures and table support the outcomes of applied methods within this study. Figure A1 shows the prepared needle of presolar grain N-AL6, using site-specific focus ion beam. The BSE image of the APT needle is accompanied by the atomic reconstruction acquired during the APT run. Figures A2 and A3 show supporting information for the custom method applied within this study which enables us to constrain the isotopic compositions of phases in APT quantitatively. Figure A2 focuses on the peak quantification and background approach used and Figure A3 focuses on demonstrating isotopic consistency across different ranges within the N-AL6 APT data set. Table A1 details the experimental conditions, acquisition parameters, and reconstruction parameters used to re-create the N-AL6 APT reconstruction.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size