C, N, AND O ISOTOPIC HETEROGENEITIES IN LOW-DENSITY SUPERNOVA GRAPHITE GRAINS FROM ORGUEIL

Low-density graphite grains (ρ ∼ 1.7 g cm⁻³) have been isolated from the Murchison CM2 (Amari et al. 1994) and Orgueil CI (Jadhav et al. 2006) carbonaceous chondrites. Isotopic signatures of most low-density graphite grains indicate an origin in Type II supernovae, specifically excesses in ¹⁵N, ¹⁸O, ²⁸Si, and the inferred presence of ²⁶Al and ⁴⁴Ti from excesses in ²⁶Mg and ⁴⁴Ca, respectively, relative to solar composition (Amari et al. 1995; Travaglio et al. 1999; Stadermann et al. 2005; Jadhav et al. 2006). Low-density graphite grains also exhibit a wide range of ¹²C/¹³C ratios. Stadermann et al. (2005) obtained nano-scale secondary ion mass spectrometry (NanoSIMS) isotope images in ^{12, 13}C, ^{16, 17, 18}O, and ^{46, 47, 48, 49, 50}Ti of nine microtome sections of the graphite spherule KE3e10 (Croat et al. 2003) from the Murchison low-density KE3 fraction (ρ = 1.65–1.72 g cm⁻³). These microtomed graphite slices contained gradients in the ¹²C/¹³C and ¹⁶O/¹⁸O isotopic ratios, with their interiors being more anomalous than their exteriors. It was suggested that these gradients are due to partial isotopic equilibration with the grain's environment during its lifetime. Strong initial O⁻ secondary ion signals with respect to the "bulk" graphite were found in small regions that contain internal TiC subgrains. The TiC subgrains are more enriched in ¹⁸O than the parent grain, although their C isotopic compositions are indistinguishable from those of the surrounding graphite.

We report on the results of NanoSIMS isotope imaging of low-density supernova graphite grains from the Orgueil meteorite. We acquired isotope images of the surfaces of numerous grains as well as isotope images of microtomed sections of their interiors. Isotopic imaging is a useful tool for investigating the heterogeneous isotopic composition within presolar grains. This study is part of correlated isotopic, chemical, and crystallographic investigations undertaken to characterize individual presolar grains as fully as possible.

Seven size/density fractions were isolated from the Orgueil CI chondrite (OR1b-OR1h)¹ with densities increasing alphabetically (Jadhav et al. 2006). The OR1d fraction (ρ = 1.75–1.92 g cm⁻³) contains considerable insoluble organic matter of solar isotopic composition, which had not been removed during the chemical separation, and which surrounds the graphite grains. A few ∼0.5 μL drops of OR1d suspended in a 4:1 mixture of isopropanol/water were pipetted onto high-purity Au foil and briefly dessicated in a vacuum oven at 70℃ (Groopman et al. 2011). These graphite grains were identified in the scanning electron microscope (SEM) based upon surface morphology and Energy Dispersive X-Ray Spectroscopy (EDXS). Candidate grains were picked with a sharp W micromanipulator needle and were deposited on a clean Au-foil mount to reduce contamination from the insoluble organic matter; some of this material, however, remains stuck to the grains' surfaces. Twenty candidate grains with diameters >5 μm were selected, labeled OR1d6m.² Raman microprobe spectra with 532 nm excitation and ∼1 μm resolution were obtained for all grains (Wopenka et al. 2011).

A several-nanometer-thick coating of Pt was deposited in a sputter coater on the grains to secure them to the Au foil. Whole-grain NanoSIMS measurements of negative secondary ions (^{12, 13}C, ^{14, 15}N, ^{16, 18}O, and ^{28, 29, 30}Si) were made with a Cs⁺ primary beam (∼100 nm), and measurements of positive secondary ions (^{24, 25, 26}Mg, ²⁷Al, ^{28, 29, 30}Si, ^{39, 41}K, ^{40, 42, 43, 44}Ca, ^{46, 47, 48, 49, 50}Ti, ⁵¹V, and ⁵²Cr) were made with an O⁻ primary beam. To obtain the N isotopic composition we measured ¹²C¹⁴N⁻ and ¹²C¹⁵N⁻, as N does not have a stable negative ion. We used sufficiently high mass resolution to separate the isobaric peaks of ¹²C¹⁵N and ¹³C¹⁴N; and ¹²C¹⁴N and ¹³C¹³C, and ¹²C¹³C H. Each grain's bulk isotopic composition was measured by rastering the primary beam over the entire grain's surface and summing the counts for each isotopic species. Additionally, we acquired 256 pixel by 256 pixel isotope images of the surfaces of grains G6, G13, G17, G21 in ¹²C, ^{14, 15}N, ^{16, 18}O and of G24 in ^{12, 13}C, ^{28, 29, 30}Si.

After NanoSIMS analysis, three grains (G17, G18, and G24) with sizes 18, 12, and 13 μm were picked from the Au foil mount, embedded in LR White hard acrylic resin, and ultramicrotomed into 70nm thick sections. Grains were selected based upon their isotopic compositions and relatively large sizes (Table 1). Slices of each grain were deposited on Si wafers and on Cu transmission electron microscope (TEM) grids, both SiO and holey-C coated. Due to the grains' large sizes, we were able to obtain hundreds of microtome slices of each grain and distribute these among the 3 different substrates. Each substrate has its own advantage: holey-C substrates are useful for determining Si content and performing electron energy loss spectroscopy (EELS) over holes; the SiO substrate reduces the C background necessary for C K-edge X-ray absorption near-edge structure (XANES) measurements (Nittler et al. 2011); and the Si wafers, while precluding TEM studies, are a superior heat sink for Raman spectroscopy (high signal-to-noise ratio), and provide a highly stable substrate for NanoSIMS measurements. Due to the large sizes of the grains, many of the slices contain tearing holes from microtoming, shown as gray in the figures. For the slices on Si wafers, isotope images in ¹²C, ^{14, 15}N, ^{16, 18}O or ^{12, 13}C, ^{28, 29, 30}Si were obtained with the resin and Si wafer used as isotopic standards. Only one image was obtained for each slice. Each isotope image consists of 20 layers of a 256 pixel by 256 pixel raster area, with a dwell time per pixel per layer of 10ms. Each raster area is chosen to be 1–2 μm wider than the grain slice. While isotope images require much longer measurement times than bulk analyses to gather meaningful statistics, they allow us to investigate heterogeneities in the grains' isotopic compositions.

3.1. Whole-grain Isotopic Measurements

3.2. Surface Images

Isotopic imaging on the graphite surfaces show heterogeneities in their C, O, and N isotopic ratios. While regions of roughly solar C, O, and N isotopic composition may be explained due to surface contamination, isotopically anomalous regions are found on the grains' surfaces, with extreme excesses in ¹⁸O and ¹⁵N relative to solar ratios. On G6, G13, and G21 these anomalous regions in have a good spatial correlation. If expressed as delta values (see Figure 1 for definition), δ¹⁵N and δ¹⁸O, respectively, range up to 6400‰ and 98,000‰ (¹⁵N/¹⁴N = 0.03 and ¹⁸O/¹⁶O = 0.20; ¹⁴N/¹⁵N = 37 and ¹⁶O/¹⁸O = 5) in G6; 2300‰ and 49,000‰ (0.012 and 0.1; 82 and 10) in G13; and 2300‰ and 98,000‰ (0.012 and 0.20; 82 and 5) in G21. In G17 the anomalous regions on the surface in δ¹⁵N and δ¹⁸O range up to 15,000‰ and 15,000‰ (0.06 and 0.03; 17 and 31), respectively, but are not spatially correlated. Hotspots on the grains' surfaces vary in size from 110 nm to 425 nm.

Zoom In Zoom Out Reset image size

3.3. Microtome Sections

Multiple ultramicrotome sections of G17 and G18 contain highly anomalous and spatially correlated hotspots of δ¹⁸O and δ¹⁵N (Figures 2 and 3). Five hotspots were found in five slices of G17 and G18 imaged in O and N isotopes (Groopman et al. 2012). δ¹⁵N and δ¹⁸O in the hotspots range up to 4400‰ and 6100‰ (¹⁵N/¹⁴N = 0.02 and ¹⁸O/¹⁶O = 0.01; ¹⁴N/¹⁵N = 50 and ¹⁶O/¹⁸O = 70), respectively. There is a strong spatial correlation between the observed excesses in ¹⁸O and ¹⁵N. It is likely that internal subgrains, particularly TiC, are the carriers of the isotopic anomalies in O and N within these hotspots. Preliminary TEM studies have confirmed the presence of internal TiC subgrains in other slices of G17, G18, and G24, although no δ¹⁸O, δ¹⁵N hotspots were found in the two slices of G24 imaged in O and N isotopes. Auger Nanoprobe scans of slice G18-C after NanoSIMS imaging show a strong Ti signal at the location of the δ¹⁵N, δ¹⁸O hotspots (Figure 3(b)). The four other slices with δ¹⁸O and δ¹⁵N hotspots were sputtered away completely during NanoSIMS imaging; we found no elemental Ti traces in Auger scans of these sections. We did not observe differences between the ¹⁶O⁻ and ¹²C¹⁴N⁻ signals between any of these hotspots and the surrounding material. Therefore, the anomalous δ-value compositions of these hotspots are characterized by excesses in ¹⁸O and ¹⁵N and not by depletions in the major isotopes, ¹⁶O and ¹⁴N, respectively. Si isotope images show uniform isotopic distributions within the grains and yield no evidence of internal SiC grains within these slices.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The interiors of the graphite grains contain C isotopic heterogeneities of two forms. Slices M and N of G24 exhibit fairly smooth gradients with ¹²C/¹³C increasing radially from 8 to 14 (Figures 4(a)–(c)). Based on the ¹²C/¹³C ratio measured in concentric regions of the slices, there exists a statistically significant gradient in the grain's C isotopic composition. In slices G17-A and G17-B (Figures 4(d) and (e)), ¹²C/¹³C ranges from 400 to 1800, whereas the bulk surface isotopic measurement was 340. The largest anomalies lie in pockets up to ∼2 μm in size close to the center of the slice. These pockets have the same Si isotopic composition and ²⁸Si⁻ count rate as the rest of the graphite grain slice, again yielding no evidence for the presence of internal SiC subgrains.

Zoom In Zoom Out Reset image size

We have found evidence for heterogeneities in the C, N, and O isotopic compositions of six low-density graphite grains. Several of these grains contain bulk excesses of ¹⁸O and ¹⁵N, evidence of a supernova origin; however, they also contain spatially correlated isotopic anomalies on their interiors and exteriors that are much more extreme than the bulk graphite measurements. TiC subgrains are a likely carrier of the observed δ¹⁸O and δ¹⁵N heterogeneities in the grains' interiors. Previous NanoSIMS studies of microtomed slices of a supernova graphite found that internal TiC subgrains have much larger ¹⁸O excesses than the parent grain (Stadermann et al. 2005). Initial ¹⁶O⁻ signals associated with these internal TiCs were also much higher than those in the parent grain. It is unknown whether the higher initial O⁻ secondary signal is due to intrinsically higher O content in the TiC subgrains or whether O ion yields from TiC are greater than those from graphite. However, recent EELS of supernova TiC subgrains has shown that they contain amorphous rims (Daulton et al. 2009) enriched in O up to 20 at.% relative to the interiors of the TiC subgrains (Daulton et al. 2012). These enrichments thus may be due to O-rich amorphous rims around the TiC grains. In the present study, we do not observe higher initial O⁻ secondary signals from the δ¹⁸O and δ¹⁵N hotspots. During our NanoSIMS measurements we presputtered the graphite sections until we achieved a constant ¹²C⁻ signal before we began imaging; this may have sputtered away the top surface of any O-rich rims on TiCs in contrast to Stadermann et al. (2005), who did not presputter their slices before imaging.

These spatially correlated δ¹⁸O and δ¹⁵N hotspots suggest that the internal TiC grains in the supernova outflow formed from inner He/C zone material where ¹⁸O/¹⁶O and ¹⁵N/¹⁴N are very high (Rauscher et al. 2002), while the bulk of the host graphite grain formed from a mixture that included material further out in the He/C zone or from other zones, resulting in more modest ¹⁸O and ¹⁵N excesses. The high ¹²C/¹³C ratios of G17 and G18 also indicate contributions from the He/C zone, which has ¹²C/¹³C ≫ solar. While it has been proposed that carbonaceous grains form in a supernova environment where the elemental ratio C/O < 1 (Clayton 2011), all of our observations point to major He/C zone contributions, where C/O > 1. G24 has a very low ¹²C/¹³C ratio and relatively smaller excesses of ¹⁸O and ¹⁵N, indicating that it may represent a mixture of material with a significant contribution from the He/N zone, where C/O is still >1. We did not observe any spatially correlated hotspots in δ¹⁸O and δ¹⁵N in G24; however, TEM investigations have found a large 300 nm TiC grain in another slice on a TEM grid. Since this slice is part of an ongoing TEM study, we will measure the O and N isotopic composition of this TiC in the future.

Isotope images of grain surfaces yield some of the largest excesses in ¹⁸O and ¹⁵N ever measured in presolar graphite grains. Although TiC abundance has been shown to decrease with radial position in several supernova graphites (Croat et al. 2003), TiC subgrains are not precluded from being the source of the surface δ¹⁸O, δ¹⁵N hotspots.

We observed two types of heterogeneities in the C isotopic compositions of microtomed graphites: smooth radial gradients in G18 and G24, and extremely anomalous pockets near the center of grain G17. Both heterogeneities are statistically significant. Radial gradients in C and O isotope composition have been previously observed (Stadermann et al. 2005). While partial isotopic equilibration with grains' environments may explain the O isotope compositions, gradients in C isotope composition require a different explanation. Since the graphite grains are overwhelmingly composed of carbon, partial isotopic equilibration would require the exchange of a significant portion of the grain's mass over its lifetime in order to reproduce the gradients we observe, which is unlikely. The equilibration of C isotopes would simultaneously have to preserve the N and O isotopic anomalies that we observe. Additionally, Zinner & Jadhav (2012) found that the radiogenic components of ²⁶Mg, and ⁴¹K within graphite grains from the same Orgueil size/density fraction are spatially correlated with the major isotopes of the parent species in each system, implying that radiogenic daughter isotopes are quantitatively retained and there is no redistribution within the grains. We therefore find it unlikely that significant redistribution or exchange of C has occurred. Instead, we suggest that the grains condensed in an environment whose isotopic composition was changing during grain formation. This would likely require small-scale mixing of material in the grains' formation environments on formation timescales and/or movement of the growing grain through regions of changing isotopic composition. Evidence of extinct ⁴⁹V (t_1/2 = 330 days) in SiC-X grains implies that the timescale of SiC formation in supernova ejecta is on the order of several months (Hoppe & Besmehn 2002). However, excluding very high pressure environments, graphite is likely to form before SiC from a gas with C/O > 1 and otherwise approximately solar composition (Bernatowicz et al. 1996). Such conditions indicate that formation of the graphite grains occurred within the first few months after the supernova explosion.

The heterogeneous ¹²C/¹³C anomalies in G17 (Figures 2(d) and (e)) range up to ¹²C/¹³C = 1800 and are located in distinct regions within the slices. These anomalous regions are unlikely to be due to the presence of either internal TiC or SiC X grains; their sizes range up to 2 μm, whereas the largest internal TiC found to date was ∼500 nm; they lack significant Si and Ti elemental signals with respect to the rest of the graphite slice as measured in the Auger Nanoprobe; and they show no significant enrichments in ²⁸Si with respect to the rest of the graphite in NanoSIMS images. We therefore believe that the pockets are graphite regions that formed from more ¹²C-enriched material before being incorporated by the parent graphite.

This work was funded, in part, by NASA Earth and Space Sciences Fellowship (NESSF) NNX11AN60H, and NASA grants NNX10AI45G and NNX11AH14G. We thank Frank Gyngard and Kevin Croat for their helpful discussions. We would like to thank the reviewer for revisions and suggestions, which have improved this paper.