PRESOLAR GRAINS FROM NOVAE: EVIDENCE FROM NEON AND HELIUM ISOTOPES IN COMET DUST COLLECTIONS

Samples were selected from the Grigg-Skjellerup (GS) and Tempel-Tuttle (TT) stratospheric collection surfaces L2054, L2055, U2121, and U2108 at the NASA Johnson Space Center Cosmic Dust Curatorial Facility (Zolensky 2004), and were packaged there in individual platinum foils for mounting in the Minnesota sample furnace. All samples are ∼nanogram (ng)-mass fragments of cluster IDPs—larger particles that shattered on collision with the collection surfaces. Two separate particle allocations were analyzed for He and Ne by static mass spectrometry, the primary sampling (Suite 1, Table 1) in 2004, and a later re-sampling of the GS collectors L2054, L2055, and U2121 (Suite 2, Table 3) as received over the period 2007–2010. Sample identifiers in Tables 1–3 were assigned by a convention in which an IDP, for example L2054 A1*1 in Suite 1, is uniquely designated first by collector (L2054), then by particle (A1), and last by cluster number (1). Thus, L2054 A1*1 and the Suite 2 sample L2054 L4*1 are two different particles (A1 and L4) from cluster 1 on collector plate L2054. Of the 15 different clusters sampled for Suite 1, 8 were re-sampled for Suite 2. The Suite 2 particles were mostly smaller in mass (0.04–1.2 ng) than the Suite 1 set (0.28–1.9 ng) and their Ne abundances were below the levels necessary for accurate isotope ratio measurements.

None of the Suite 1 samples selected for noble gas analysis were mineralogically or chemically characterized before they were destructively analyzed at Minnesota by high-temperature heating. Some fragments of the same cluster particles from which the Suite 1 and 2 samples were extracted, and other clusters from the Grigg-Skellerup collection (GSC), are composed primarily of silicates, aluminum oxides, GEMS (glass with embedded metal and sulfides), and carbonaceous matter in various proportions (Nakamura-Messenger et al. 2008). One of these contains the new mineral Brownleeite (Nakamura-Messenger et al. 2010). However, given the chemical and mineralogic heterogeneity observed in different fragments from a single cluster IDP (Thomas et al. 1995), the extent to which these compositional data are specifically applicable to the Suite 1 and 2 particles is uncertain.

After loading of sample and empty platinum foil packets into the multiple-sample furnace (the empty packets for blank analyses), the extraction system was baked at ∼150 °C for three days. Samples were then degassed in a series of computer-controlled 10 s heating pulses from ∼350 °C to a maximum temperature of ∼1400 °C. Helium and Ne released in successive temperature steps were accumulated in the mass spectrometer. Helium abundances were monitored during the pulse heating sequence, and heating was interrupted and evolved He and Ne abundances and isotopic compositions measured whenever enough He had accumulated for accurate isotopic measurement. Following this the spectrometer was emptied and gas accumulation resumed on initiation of the next step in the heating sequence. All sample and blank foils were heated and analyzed in the same way.

Since this procedure differs from the conventional analytic protocol in which noble gases are first purified and then released into a mass spectrometer at a single well-defined time, we describe it in more detail. The gas extraction, processing, and analysis system is composed of three sections in series: (a) a multi-sample furnace in which samples wrapped in Pt foil are connected across Pt leads and heated by stepped pulses of increasing electrical power; (b) a gas purification section equipped with Westinghouse^@ WL and NP10 getters, a SAES^@ metal appendage getter, and cryopumping on stainless steel (SS) and activated charcoal fingers at liquid N₂ temperature; and (c) a 375 double-focusing Mattauch–Herzog geometry mass spectrometer equipped with pulse-counting detectors (Mattauch & Herzog 1934; Nier & Schlutter 1985). Gases evolved during a heating step in (a) are released directly into section (b) where they are exposed to the getters and cryopumped for one minute on the SS finger to reduce abundances of chemically active and condensable species. After one minute the SS finger is isolated from the system and cryopumping is switched to the activated charcoal. The mass spectrometer inlet is then opened for a quick analysis of masses 4, 18, 20, 40, and 44. These masses are rapidly scanned three times in succession, and the amount of ⁴He used to determine if the estimated threshold for accurate helium isotopic analysis has been reached. If so, heating is halted and sections (b) and (c) isolated from (a). While the evolved gases continue to be exposed to the cold charcoal finger and getters in (b), the ion source electron voltage is reduced to 35 eV (Section 2.6) and the mass spectrometer electronics set to maximize sensitivity for Ne. Following the Ne analysis, electron voltage is increased to 75 eV, the spectrometer re-tuned for He, and the evolved He analyzed. The spectrometer is then pumped.

If the ⁴He released in a particular heating step is judged insufficient to provide an accurate isotopic analysis, sections (a) and (b) are again isolated from the mass spectrometer and stepped heating is resumed. Thus, the gas ultimately analyzed is the sum of gases released during the number of sample heating steps required to reach either the ⁴He threshold or the ∼1400 °C maximum temperature. In samples whose heating is uninterrupted, the total time taken to complete the 14 heating steps to ∼1400 °C is approximately 37 minutes, after which Ne and then He analyses are performed.

Neon analyses of samples, blanks, and calibrations comprise 10 individual blocks of computer-controlled peak-to-peak jump-scan measurements of masses 18, 19, 20, 22, 40, and 44 in each block. To determine the small mass 21 abundance with higher precision, it is measured 10 times individually in each of the 10 blocks, for a total of 100 measurements. Automatic re-centering on the mass 18 water signal in each block compensates for any magnetic field drift—usually negligible—that might have occurred during this time. This full Ne analysis requires approximately 60 minutes. The following ∼75 minute He analysis includes jump-scan measurements of masses 2, 3, and 4, and is performed in a similar fashion. ⁴He is measured 10 times and ³He 100 times, with re-centerings on mass 4 (for ⁴He) and on mass 2 (for ³He). Since count rates for both ²¹Ne and ³He are generally low, signals are corrected for small electron multiplier "dark currents" measured in parallel with ²¹Ne and ³He.

Neon evolved in this stepped heating and data acquisition procedure may reside in the spectrometer for times as long as ∼37 minutes prior to analysis, and He up to ∼100 minutes, raising the possibility that isotopic distributions could have been altered during this storage period by spectrometer ion-pumping and memory contributions. Both ion-pumping and memory effects generate time trends in the measurements, and are customarily removed from sample data by extrapolating measured abundances and isotope ratios versus time back to the instant of gas inlet into the spectrometer ("time-0"). There is no definable time-0 when gases accumulate in the instrument during stepped heating, as they do here. However, time trends were not observed in Ne and He sample measurements; abundances were constant within statistical uncertainties over the duration of all step-heated sample and blank analyses. This was also the case in calibrations, where no time dependences of Ne signals were seen in the ∼60 minute analyses following known gas inlet times. We conclude that the absence of a time-0 marker for accumulated gases is therefore not a critical issue, and that significant perturbations of actual compositions during pre-analysis gas storage in the spectrometer are unlikely.

Sensitivities for He (2.58 × 10⁻¹³) and Ne (4.95 × 10⁻¹³) in units of cm³ STP per count s⁻¹ (1 cm³ STP = 2.69 × 10¹⁹ atoms) were determined from repeated analyses of air samples metered into the system from calibration cans designed to permit accurate volumetric splitting, with O₂, N₂, and other chemically active constituents continually purified by getters and cryopumping (Section 2.3). The helium isotopic standard for mass discrimination corrections was reagent grade helium spiked with ³He [³He/⁴He = (7.61 ±.08) × 10⁻⁴] for Suite 1 and most Suite 2 IDPs. A new standard [³He/⁴He = (5.00 ±.05) × 10⁻⁴] was used for the remainder of the Suite 2 samples. Air Ne isotope ratios ²⁰Ne/²²Ne = 9.80 ± 0.08, ²¹Ne/²²Ne = 0.0290 ± 0.0003 (Eberhardt et al. 1965) served as the Ne standard.

One set of procedural blanks was measured on empty Pt foil packets in two ways: cold (14 steps, no heating) and hot (14 heating steps to ∼1400 °C). Averages of these cold and hot blanks within a suite of measurements were indistinguishable. There were no statistically significant differences between these values and those for re-heatings of sample foils. These observations demonstrate that the blanks derive primarily from the room temperature gas purification system, not from degassing of the furnace hardware or Pt sample foils during stepped heating.

For Suite 1, averages of 36 He and 15 Ne blank abundances accumulated in 14 steps (in cm³ STP, errors ±1σ standard deviation of the mean) were ³He = (9.8 ± 0.3) × 10⁻¹⁵, ⁴He = (6.7 ± 0.4) × 10⁻¹², ²⁰Ne = (8.9 ± 0.4) × 10⁻¹³, ²¹Ne = (6.3 ± 0.4) × 10⁻¹⁵, and ²²Ne = (8.6 ± 0.4) × 10⁻¹⁴. Blank isotope ratios were ³H/⁴He ≃ (1.5 ± 0.1) × 10⁻³, ²⁰Ne/²²Ne ≃ 10.3 ± 0.7, and ²¹Ne/²²Ne ≃ (7.3 ± 0.6) × 10⁻². For most of Suite 2, ³He = 6.3 × 10⁻¹⁵, ⁴He = 8.6 × 10⁻¹², ²⁰Ne = 5.4 × 10⁻¹³ cm³ STP. Blanks were somewhat different for the last three Suite 2 samples: ³He = 2.6 × 10⁻¹⁵, ⁴He = 5.9 × 10⁻¹², ²⁰Ne = 7.9 × 10⁻¹³ cm³ STP for L2055 L1*7, and ³He = 4.9 × 10⁻¹⁵, ⁴He = 9.7 × 10⁻¹², ²⁰Ne = 3.1 × 10⁻¹² cm³ STP for L2054 L3*2 and U2121 E2*2.

A separate set of procedural blanks was applied to samples where stepped heating was interrupted and interim analyses performed. Here, the blank procedure duplicated that for a sample—i.e., gases accumulated through prior steps were measured at each point in the heating sequence where the sample gases had been analyzed. Again, there was no statistical difference between cold and hot blanks, or with sample re-heatings. Blank abundances for these interim measurements were very similar to those for the uninterrupted 14 step sequence (above), independent of the number of steps to the interruption temperature. No dependence on accumulation time was observed for any blanks in this study.

Blank corrections were most critical for ³He and ²¹Ne (and, for samples discussed in Section 2.6, for ²²Ne as well). A detailed analysis of the He blank sources in our extraction and analysis system was reported earlier (Pepin et al. 2000). In cases where the isotope was detectable above the blank, blank contributions to the measured gas in the cumulative sample heating ranged from 5% to 85% (average ∼40%) for ³He and from 40% to 90% (average ∼65%) for ²¹Ne. Uncertainties due to blank corrections are included in errors listed for all data in Tables 1–3.

The 15 IDPs in Suite 1 display two distinctly different patterns of gas release with increasing temperature. Six of the samples (3, 4, 7, 8, 11, and 15 in Table 1) were comparatively rich in He; in five of the six, gases were measured in two or more heating steps to ∼1400 °C, with outgassing beginning at temperatures ranging from ∼350 °C to ∼800 °C. Blank-corrected data obtained in individual temperature steps of samples where the heating sequence was paused and gases released to that point were measured are given in Table 2. In contrast, so little He evolved from all but one of the other nine IDPs that only a single analysis, of total gases accumulated over all 14 heating steps to ∼1400 °C, was performed. The exception was IDP U2108 A1*1 (No. 6 in Table 2), which released solar-wind (SW)-like He together with a small Ne component, isotopically resembling an air–SW mixture, in three steps between 520 °C and 870 °C. Little additional He was evolved in the subsequent steps to 1430 °C. Analysis of this high-temperature fraction, which contained >70% of the total Ne, revealed an Ne component strikingly similar to that in the other eight low-He IDPs. U2108 A1*1 was therefore assigned to this group. The low-temperature releases could plausibly have derived from a small high-He particle adhering to this IDP, together with a minor air contaminant.

Except for U2108 A1*1 (No. 6) where only the 1430 °C data are given, abundances, concentrations, and isotope ratios reported in Tables 1 and 3 are integrations of measurements over all heating steps. It is important to emphasize that every sample, high-He and low-He, was subject to the same multi-step heating sequence for gas extraction whether or not it was interrupted for interim analyses. As was the case for most of the low-He IDPs, gases evolved from two of the high-He group were also measured only after termination of heating at maximum temperature (No. 15, U2108 A2*2), or close to it (No. 11, U2121 A3*3—see Table 2), and therefore had relatively long pre-analysis residence times in the mass spectrometer. As seen in Table 1, they are isotopically similar to the other four high-He IDPs where interim measurements were made at lower temperatures. Note also that the sample ordering in Tables 1 and 2 is by analysis date, and that appearances of individual high- and low-He IDPs are well interspersed throughout the duration of the experimental study.

The nine He-poor IDPs (Section 2.5) are characterized by He and Ne isotope ratios that differ markedly from solar system compositions, whereas those displayed by the six He-rich samples reflect implantation of their noble gas inventories by SW irradiation, in common with most IDPs (Table 1). For this and other reasons discussed in Section 3.1, the two groups are, respectively, designated as "anomalous" and "normal." Measurement of extraordinarily high ²⁰Ne/²²Ne ratios in the anomalous IDPs directed particular attention to blank and memory corrections to the low ²²Ne abundances and to the possibility of mass 20 augmentation by doubly ionized ⁴⁰Ar. The average ²²Ne blank correction was 67% with a range in individual samples of ∼54%–77%. Average measured ²²Ne was 50% (range ∼30%–85%) above the ²²Ne blank, about 10 times the blank uncertainty. Although measurement errors for ²²Ne in the samples were substantial (1σ average of ±37%, range ∼±18%–70%), minimum (−1σ) abundances were still above the blank level in five of the nine. Errors in Table 1 for ²⁰Ne/²²Ne ratios in the anomalous IDPs primarily reflect uncertainties in blank-corrected ²²Ne abundances.

Memory buildup during pre-analysis storage in the spectrometer is thought to be negligible (Section 2.3). But even if it were present, a memory Ne component with its known ²⁰Ne/²²Ne ratio of ∼11 ± 1—reflecting a residual ion-buried mixture of air–Ne and SW–Ne compositions previously analyzed in the spectrometer—would depress measured ²⁰Ne/²²Ne ratios in the low-He anomalous group below their actual values. An elevated and undetected blank contribution with ²⁰Ne/²²Ne = 10.3 ± 0.7 (Section 2.4) would have the same effect.

Measurements of ²⁰Ne generally had much smaller statistical fluctuations, and all were many standard deviations greater than the blank value. The question is whether mass 20 abundances could have been enhanced by ⁴⁰Ar⁺⁺. Care was taken to ensure against this contribution. An electron impact voltage of 45 eV was experimentally determined to be low enough to make doubly ionized ⁴⁰Ar and ¹²C¹⁶O₂ undetectable at masses 20 and 22; for added insurance that ⁴⁰Ar⁺⁺ and ¹²C¹⁶O₂⁺⁺ were completely suppressed it was further reduced to 35 eV for Ne analyses, without substantial loss of sensitivity. Moreover, the spectrometer was continually cryopumped with liquid N₂-chilled charcoal during the ∼1 hr duration of Ne analyses, reducing ⁴⁰Ar to very low levels. Plots of ⁴⁰Ar versus ²⁰Ne and CO₂ versus ²²Ne show no evidence of correlation.

³He was not detected in any of the anomalous samples. Two of them, No. 13 (U2108 A3*3) and No. 10 (L2055 A5*8) in Table 1, contained enough ⁴He that ³He would have been seen above the ³He blank, by factors of ∼6 and ∼3 times the blank uncertainty, if the ³He/⁴He ratio were as high as 10⁻⁴. The actual maximum is therefore lower. Although orders of magnitude higher than the theoretical—and unmeasurable—prediction of <10⁻⁸ in a neon nova outburst (Starrfield et al. 2009), the <10⁻⁴ upper limit is still below the ³He/⁴He range in solar system matter, from 1.66 × 10⁻⁴ in Jupiter's atmosphere (Mahaffy et al. 1998) to an average of 4.49 × 10⁻⁴ in the SW (Mabry et al. 2007; Heber et al. 2009; Table 4).

The ²¹Ne/²²Ne and ²⁰Ne/²¹Ne ratios assigned to the anomalous IDPs are from a single imprecise measurement of ²¹Ne in sample 2 (L2054 A4*4). Despite the large uncertainty, minimum (−1σ) ²⁰Ne/²¹Ne is still a factor of 1.7 above the SW ratio and within theoretical predictions for neon novae. Other anomalous IDPs have larger ²⁰Ne abundances than L2054 A4*4 but contained no detectable ²¹Ne above blank, indicating higher ²⁰Ne/²¹Ne ratios that would be fully consistent with a neon nova source (Section 4.2).

Neon concentrations in the 15 Suite 1 IDPs from the GS and TT collections (Table 5) are shown in Figure 1. Plotted concentrations are numbered to identify the corresponding IDP data in Tables 1, 2, and 5. They clearly fall into two isotopically distinct groups, one "normal" and the other "anomalous." The normal population is characterized by the presence of SW-He, and by correlated ²⁰Ne and ²²Ne concentrations with ²⁰Ne/²²Ne close to the SW ratio of 13.9 (Table 4), identifying these as particles that experienced substantial implantation of solar ions during their residence in space. In the following discussion we define "normal" IDPs as those containing only Ne and He inventories attributable to space irradiation, by SW or by cosmic rays producing spallogenic components. None of these contain recognizable non-SW gases indigenous to the particles.

Zoom In Zoom Out Reset image size

The anomalous IDPs in Figure 1 are strikingly different. They contain no detectable SW-He, and contain low and approximately uniform ²²Ne concentrations together with highly variable ²⁰Ne, an unusual isotopic distribution noted in an earlier abstract (Palma et al. 2005). Measured ²⁰Ne/²²Ne ranges from ∼25 to ∼150. These high values greatly exceed the SW ratio and are not seen in meteorites, other IDPs, or planetary atmospheres. The anomalous IDPs are compared in Figure 2 to 64 normal IDPs with measured Ne compositions, utilizing an ⁴He/²⁰Ne versus ²²Ne/²⁰Ne representation that allows both groups and their measurement uncertainties to be shown at high resolution. The average offset between them is evident for both ²²Ne/²⁰Ne (factor of ∼6) and ⁴He/²⁰Ne (factor of ∼20). Within the normal IDPs themselves, the rightward scattering of ²²Ne/²⁰Ne in Figure 2 away from the incident SW ratio toward values deficient in ²⁰Ne suggests varying degrees of isotopic fractionation due to sputter erosion in space (Becker 1998; Pepin et al. 1999).

Zoom In Zoom Out Reset image size

Figure 2 data are re-plotted in Figure 3 with Ne isotopic compositions converted to the more conventional ²⁰Ne/²²Ne representation. Collapse of the SW and normal IDP data set in Figure 2 to a narrow band at the bottom of the diagram emphasizes the dramatically higher ²⁰Ne/²²Ne ratios in many of the anomalous IDPs.

Zoom In Zoom Out Reset image size

Examination of He alone confirms that the distinctly low ⁴He/²⁰Ne signature in the anomalous IDPs (Figure 2) is due to deficits in ⁴He content relative to normal IDPs. In the Figure 4 plot of ³He/⁴He versus ⁴He g⁻¹ the average ⁴He concentration in the anomalous particles is seen to be a factor of ∼3 lower than in any of the comparison groups that collectively include 83 normal IDPs. The unique nature of this He is even more evident in ³He contents, not normalized to particle masses. A logarithmic plot of measured ³He versus ⁴He abundances (Figure 5) shows that the groups of normal IDPs fall closely along a power-law distribution (from which an average ³He/⁴He ratio of ∼3.2 × 10⁻⁴ may be derived). Sample U2108 A3*3, with the highest ⁴He content of the anomalous IDPs (2.06 × 10⁻¹¹ cm³ STP: Table 1) but no detectable ³He, sets an ³He/⁴He upper limit of <10⁻⁴ (Section 2.7). U2108 A3*3 is clearly distinct from the normal samples, falling below their power-law distribution by a factor of ≳4 in ³He. One of the normal groups (yellow symbol) is lower than U2108 A3*3 in average measured ⁴He by about the same factor, yet SW-³He was readily detected in all 12 of its constituent particles. The failure to see ³He in U2108 A3*3 is therefore not due to an He content too low to measure an SW contribution at or well below the levels seen in normal IDPs.

Zoom In Zoom Out Reset image size

3.2.1. Space Residence Time of Anomalous IDPs in the Grigg-Skjellerup Collection

The absence of detectable ³He indicates that the anomalous IDPs have spent little time exposed to solar radiation as free particles in space. With some assumptions, an upper limit to their average space residence time in a comet dust stream orbit can be estimated. This requires, in addition to maximum ³He content [³He]_max, approximate measures of (a) the He loss factor f_h from drag heating during atmospheric entry, and (b) the factor f_s due to partial shielding of a cluster IDP fragment from SW irradiation by fractional burial in its larger parent particle. With the average SW-³He flux over the dust stream orbit represented as F₃, the upper time limit to SW exposure is T_max = [³He]_max f_s f_h/F₃.

The anomalous IDPs from the GS collectors (L2054 A4*4, L2055 A1*1, L2055 A5*8, U2121 A1*1, and U2121 A2*2 in Table 1, excluding the two with no detectable ⁴He) have an average cross-section area of ∼8.0 × 10⁻⁷ cm² and an ⁴He abundance of ∼5.9 × 10⁻¹² cm³ STP. With ³He/⁴He < 10⁻⁴, [³He]_max ≃ 2.0 × 10¹⁰ atoms cm⁻². An average value of F₃ for the GS dust stream may be calculated from the near-Earth SW-⁴He flux (F₄)_o = 1.24 × 10⁷ atoms cm⁻² s⁻¹ (Reisenfeld et al. 2007) and the perihelion (R_o ∼1 AU) and aphelion (R = 4.93 AU) of the GS orbit. The average ⁴He flux over the orbit from R_o to R is ∫(F₄)_odr/r²/∫dr = (F₄)_o/RR_o = 2.51 × 10⁶ atoms cm⁻² s⁻¹, and the corresponding ³He flux, with ³He/⁴He in the SW = 4.49 × 10⁻⁴ (Table 4), is F₃ = 1.13 × 10³ atoms cm⁻² s⁻¹.

An estimate of f_h ∼ 10 for He comes from Figure 2 where the average ⁴He/²⁰Ne ratio in 64 normal IDPs is seen to be approximately an order of magnitude below the SW implantation ratio of ∼650 (Table 4); the assumption here is that drag heating primarily releases He while less labile Ne is largely retained. About two-thirds of the ratios in Figure 2 fall between ⁴He/²⁰Ne ∼20 and ∼200, corresponding to f_h values in the range ∼3–30. With this as an uncertainty estimate, f_h is ∼10 within a factor of ∼3.

The SW shielding factor f_s is more difficult to evaluate. Our approach was to compare ²⁰Ne contents (atoms cm⁻² cross section) in ∼60 individual IDPs and ∼40 cluster IDP fragments, utilizing data from Nier & Schlutter (1990), Pepin et al. (2000, 2001), Kehm et al. (2006), this work (Tables 1 and 3), and unpublished IDP data from the Minnesota laboratory. The comparison indicates an ∼four-fold-lower Ne loading in the average cluster fragment compared to the average individual particle, with a conservative range of ∼1–7 depending on how the various data sets are weighted. This factor of 4 ± 3 can be taken as an approximate measure of f_s if two implicit assumptions are roughly valid: that the individual IDPs experienced SW implantation over their entire surfaces—i.e., f_s = 1 (no shielding)—and the relatively large numbers of IDPs in the individual and cluster groups have about the same average SW exposure age. An f_s value of 4 means that, on average, a cluster fragment was exposed over 25% of its surface area. We note that virtually all cluster IDP fragments analyzed for noble gases, except for the anomalous ones reported here, contain ³He attributable to SW implantation. This argues against high values for f_s and implies relatively open in-space structures for the parent cluster particles, as opposed to geometries which would shield significant fractions of their volumes from short-range (∼10–30 nm) SW radiation.

Table 3. Helium and Neon Abundances and Isotope Ratios in Individual Suite 2 IDPs

Sample	4He	20Ne	4He/20Ne	3He/4He	20Ne/22Ne	21Ne/22Ne	20Ne/21Ne
Mass†	(10−11 cm3 STP)	(10−12 cm3 STP)		(×10−4)		(×10−2)
L2054 L4*1	2.73	0.082	332	2.97		nd: 21Ne, 22Ne
0.04 ng	±0.03	±0.006	±25	±0.35
L2054 L3*2	3.29	nd: 20Ne		4.27		nd: 21Ne, 22Ne
0.91 ng	±0.09			±0.33
L2054 L1*3	0.998	nd: 20Ne		2.9		nd: 21Ne, 22Ne
0.08 ng	±0.024			±1.2
L2054 L2*4	0.903	0.299	30.2	4.9	nd: 22Ne		88
0.16 ng	±0.021	±0.016	±1.8	±1.3			±55
L2055 I1*1‡	0.117	0.035	33.4	nd: 3He		nd: 21Ne, 22Ne
0.20 ng	±0.003	±0.004	±3.9
L2055 I2*6‡	0.107	0.097	11.0	nd: 3He		nd: 21Ne, 22Ne
0.16 ng	±0.003	±0.005	±0.6
L2055 I3*7	1.85	0.195	94.9	1.60		nd: 21Ne, 22Ne
0.08 ng	±0.04	±0.018	±9.0	±0.59
L2055 L1*7	25.6	0.927	276	2.99	18	0.12	152
0.40 ng	±0.3	±0.072	±22	±0.13	±11	±0.10	±83
U2121 E1*1	6.22	0.478	130.1	2.54	nd: 22Ne		290
0.29 ng	±0.05	±0.021	±5.8	±0.28			±260
U2121 E2*2	3.94	nd: 20Ne		3.24		nd: 21Ne, 22Ne
1.2 ng	±0.13			±0.22

Notes. IDP: interplanetary dust particle, nd: not detected. Uncertainties are ±1σ analytic errors. ^†Masses estimated from particle dimensions and assumed asteroidal IDP density of 2.5 g cm⁻³; uncertainty ≈±70%. ^‡Possible low-He anomalous IDP.

Download table as:  ASCIITypeset image

Table 4. Solar Wind Isotope Ratios in Genesis Collector Materials

Source	4He/20Ne	3He/4He	20Ne/22Ne	21Ne/22Ne	20Ne/21Ne
		(×10−4)		(×10−2)
Solar	656	4.64	13.78	3.29	418.8
winda	±5	±0.09	±0.03	±0.01	±1.6
Solar	643	4.34	13.97	3.46	403.8
windb		±0.02	±0.03	±0.03	±2.4
Average	650	4.49	13.88	3.38	411.3
solar wind	±7	±0.15	±0.10	±0.09	±7.5

Notes. ^aHeber et al. (2009). ^bMabry et al. (2007).

Download table as:  ASCIITypeset image

With these estimates for f_h and f_s, the nominal upper limit for SW exposure of an average anomalous IDP originating from the GS dust stream is T_max ≃ 20 yr. Uncertainties in f_h and f_s allow ranges for T_max of ∼7–65 yr and ∼6–40 yr, respectively. The actual maximum exposure age of a dust stream IDP released after perturbation of the comet into an Earth-crossing orbit in 1964 (Messenger 2002) and collected in 2003 (Zolensky 2004) is ∼40 yr. Many such particles, shed by the comet in apparitions closer in time to 2003, will have shorter exposures. The estimated average space residence time of at most a few decades for the anomalous IDPs after their assumed release from GS is therefore entirely compatible with the chronology of the GS Earth-crossing dust stream.

Zoom In Zoom Out Reset image size

Densities have been measured for IDPs classified as asteroidal (<14 km s⁻¹ entry velocity) and cometary (>18 km s⁻¹) in origin (Joswiak et al. 2000). Classification was based on release profiles of He in laboratory stepped heating and a model of particle heating during atmospheric entry (Love & Brownlee 1994). Reported averages and ranges are 2.5 (1–5.4) g cm⁻³ for the 12 IDPs in the asteroid group and 1.1 (0.6–1.9) g cm⁻³ for the 13 in the comet group. Particle masses in Tables 1 and 3, used to calculate Ne and He concentrations in Figures 1 and 4, are estimates based on assumed densities of 2.5 g cm⁻³ for the normal IDPs and 1.1 g cm⁻³ for the anomalous IDPs. The wide ranges in possible densities are primarily responsible for the large uncertainties, estimated overall to be ∼±70%, in the masses of individual samples listed in Tables 1 and 3.

Errors in Ne concentrations (cm³ STP/g-IDP) listed in Table 5 and plotted in Figure 1 do not include these large mass uncertainties; they indicate only the errors in measured ²⁰Ne and ²²Ne abundances (cm³ STP) given in Table 1. If mass uncertainties were included, error bars would span half the x-axis of Figure 1 for the high ²⁰Ne anomalous data. They are omitted because they have no impact on any of the paper's conclusions; the discussions and supporting figures, except for the anomalous ⁴He concentration plotted in Figure 4, involve either isotope ratios, where they cancel, or abundances, where they do not enter. Consideration of mass uncertainties, however, is germane to comparisons of the concentration distributions for the normal and anomalous IDPs in Figure 1. Use of actual masses—if we knew what they were—in calculating concentrations could generate quite different distributions.

The normal IDPs in Figure 1 are assumed to be asteroidal. They probably belong to the background population of IDPs present on most stratospheric collectors (Zolensky 2004). Their He and Ne inventories appear to be due solely to SW irradiation and indicate substantial time in space. All of the normal IDPs in Figure 2 do not necessarily derive from asteroids; some could be cometary, with solar components masking any indigenous noble gas signatures that might be present. However, their classification as "normal" is independent of assumed density since their ²⁰Ne and ²²Ne abundances are correlated. Different densities in the Suite 1 particles, for example, shift their positions in Figure 1 but they remain within error on the 13.9 SW line.

There are strong arguments for the assignment of cometary density to the anomalous IDPs. They are characterized by undetectable ³He, indicating minimal exposure to the SW (Section 3.2.1) as would be expected for recent release into cometary dust streams (Messenger 2002). They were taken from collectors unusually rich in presolar grains, consistent with expectations for primitive cometary matter (Busemann et al. 2009; Nakamura-Messenger et al. 2008, 2010), and they contain strikingly non-solar Ne and He compositions quite unlike those seen in meteorites derived from asteroid parent bodies. It seems most likely that they originated in the dust streams the collection flights were targeted to sample.

The absence of detectable SW components indicates that He and Ne inventories in the anomalous IDPs represent essentially unmodified indigenous gases. Of the combination of theoretical hydrodynamic calculations and nuclear reaction networks describing element synthesis in stellar environments, only those for total ejecta compositions in neon nova outbursts specifically predict the combination of negligible ³He, low ⁴He/²⁰Ne, high ²⁰Ne/²¹Ne, and ²⁰Ne/²²Ne ratios in and above the range of ∼25 to ∼150 (Figure 3) measured in these particles (Starrfield et al. 2009; José et al. 2004).

A classical nova explosion is caused by a thermonuclear runaway (TNR) on the surface of a WD that is accreting matter through the inner Lagrangian point in a close binary star system (Bode & Evans 2008). Radiation pressure from the TNR ejects into space a large fraction of the accreted gas as well as core material dredged up from the WD itself during the TNR. Each nova outburst ejects material with unique abundances because of the large range in parameters affecting the properties of the explosion. These include WD mass and composition, initial WD luminosity, the mass accretion rate, and the nuclear reactions in the TNR itself (José et al. 2004; Starrfield et al. 2008, 2009). Once propelled from the WD, the expanding shell of ejected gas can be observed spectroscopically to determine its chemical composition (Gehrz et al. 1998). The discovery of a strong [Ne ii] 12.8 μm emission line in nova QU Vul 1984#2 provided firm observational evidence for nova eruptions driven by TNRs on oxygen–neon (ONe) WD stars (Gehrz et al. 1985). Velocity broadening prevents assignment of observed Ne emission lines to a particular isotope, but the composition of the dredged-up ONe WD core material predicts that Ne initially in the ejecta is essentially all ²⁰Ne. Detailed calculations (José et al. 2004; Starrfield et al. 2008, 2009) show that repeated proton captures on ²⁰Ne will fuse some of the neon to ²²Na. Pre-existing ³He and ²²Ne are virtually annihilated in the TNR by fusion to heavier species; ³He/⁴He is driven to <10⁻⁸ and ²⁰Ne/²²Ne to >10⁶ (Starrfield et al. 2009). Total ²²Ne inventories (Σ²²Ne) in the neon nova models are thus dominated by beta decay of ²²Na (half-life = 2.60 yr), and ³He is essentially absent.

Helium and Ne distributions from a recent modeling of total ejecta composition in a neon nova outburst (Starrfield et al. 2009) are given in Tables 6 and 7 and plotted as isotope ratios in Figure 6. Their non-solar character is evident for ratios involving the enhanced ²⁰Ne and severely depleted ³He. Since modeling results are sensitive to parameter choices, Figure 6 also shows ratio ranges derived from 10 ONe WD outburst models based on differing selections for input parameters (José et al. 2004). Ratio ranges in total ejecta from supernovae, calculated from the Type II supernova (SN II) nucleosynthesis models of Rauscher et al. (2002), are plotted for comparison.

Zoom In Zoom Out Reset image size

He and Ne compositions in the normal and anomalous IDPs (Table 7) are shown in Figure 7. SW ratios and calculated ratios and ranges for neon nova outbursts from Figure 6 are plotted for comparison. Normal IDP averages fall close to SW composition, consistent with prolonged space irradiation, except for a low ⁴He/²⁰Ne ratio suggesting loss of labile implanted He by drag heating during atmospheric entry. In contrast, as seen in Figure 7, measured isotopic averages and ranges in the anomalous IDPs, particularly the non-solar ⁴He/²⁰Ne and ²⁰Ne/Σ²²Ne ratios reflecting high ²⁰Ne abundances, are in good agreement with theoretical predictions for neon nova ejecta. This is also true for ratios involving ³He and ²¹Ne, but these are less definitive since they rest on a calculated ³He upper limit and one imprecise measurement of ²¹Ne (Section 2.7).

Zoom In Zoom Out Reset image size

This direct comparison of an elemental ratio, in this case ⁴He/²⁰Ne, with nucleosynthetic theory is similar to that carried out by Heck et al. (2007) for ⁴He/²²Ne ratios measured in their study of He and Ne in presolar SiC grains. Elemental composition is revealed here because the anomalous IDP gases apparently contain little or no interference from normal solar system components. In this respect they differ from the first detections of presolar noble gases in meteorites (Reynolds & Turner 1964; Black & Pepin 1969; Black 1972). There the isotopic signatures were diluted by mixing with other meteoritic gases, and the pure compositions were determined only when their carrier grains—nanodiamonds for Xe-HL, graphite and silicon carbide for Ne-E—were chemically or physically separated from the bulk meteorites (Anders & Zinner 1993; Nittler 2003; Zinner 2004). This is not the case for the nova-linked IDPs, even though the much larger sizes of the ∼8–15 μm fragments analyzed in this study compared to the ∼0.4–1.4 μm grains observed in nova outflows (Gehrz 2008) suggest that minerals from other sources are dominant constituents of the anomalous IDPs. Agreements of measured He and Ne isotope ratios with theoretically derived ejecta compositions indicate that extraneous non-nova materials in the samples contribute little if at all to the gas inventories.

A striking feature of Figure 1 is the non-correlation of the ²⁰Ne and ²²Ne concentrations in the anomalous IDPs. With one exception, ²²Ne is approximately uniform in the nine samples, but ²⁰Ne varies by a factor of ∼7 and ²⁰Ne/²²Ne ratios by a factor of ∼6. Noble gases incorporated into minerals by mechanisms such as irradiation, solution, or physical trapping closely reflect the isotopic compositions of their sources. An example is the normal IDP group in Figure 1 where the correlated ²⁰Ne and ²²Ne concentrations yield ²⁰Ne/²²Ne ratios consistent with that in the irradiating SW. Uncorrelated ²⁰Ne and ²²Ne in the anomalous group requires separate sources for the two isotopes. We propose that most of the ²²Ne derives from decay of ²²Na condensed in grains during their growth in the expanding nova outflow, and the non-condensable ²⁰Ne and ⁴He reflect shock or stellar wind emplacement into the grains from an ambient reservoir strongly depleted in ²²Ne and ³He (Section 4.4).

Other sources for the anomalous He and Ne appear unlikely. Nucleosynthesis in AGB stars (Heck et al. 2007) and CO novae (José et al. 2004; Heck et al. 2009) generates ²⁰Ne/Σ²²Ne ratios that are far too low. Comparison of the total yield ranges for SNe II (Rauscher et al. 2002) in Figure 6 with the anomalous IDP values in Figure 7 shows that the ⁴He/²⁰Ne, ²⁰Ne/Σ²²Ne, and ²⁰Ne/²¹Ne SN II ratios are significantly offset from the IDP data, in directions also indicating underproduction of ²⁰Ne. Total yields, however, reflect complete mixing of nucleosynthesis products generated in separate supernova shells. Infrared spectroscopic observations of a supernova remnant (Cas A) suggest that shell products may be largely unmixed in the ejecta (Rho et al. 2008), so grains condensing within individual shells could have acquired locally synthesized gases. However, for the case of ejecta from a 15 solar mass supernova (Rauscher et al. 2002, data accessed at http://www.nucleosynthesis.org), Figure 8 indicates that He and Ne isotopes produced in any of the separate SN II shells cannot reproduce the full set of anomalous IDP ratios, nor can those generated by mixing adjacent shells. Integrated Ne yields in the Ni shell come closest in Figure 8 to matching the average IDP Ne ratios, and ³He/⁴He (not shown) is ≪10⁻⁴. However, ⁴He/²⁰Ne in the shell is too high by a factor of ∼10⁴. For the mixing alternative, the best possibility appears to be a combination of products synthesized in the Ni through O/C shells, collectively ∼10% of the total ejecta mass, which yields ³He/⁴He ≪10⁻⁴, ⁴He/²⁰Ne reasonably close to the average IDP ratio, and ²⁰Ne/Σ²²Ne (∼35) near the lower end of the measured range. But this mixing scenario cannot account for ²⁰Ne/Σ²²Ne ratios >35 in seven of the nine anomalous IDPs (Figure 3), nor for a calculated ²⁰Ne/²¹Ne lower than the measured ratio by ∼6σ.

Zoom In Zoom Out Reset image size

Irrespective of WD class, ONe or CO, grains begin to nucleate and grow to maximum size in cooling nova outflows within ∼80 days or less after the eruptions (Gehrz 1988). While dominant mineralogies depend in part on nova class, individual novae are capable of producing a suite of grain types including C, SiC, silicates, and hydrocarbons (José et al. 2004; Gehrz 2008). Infrared emission from grains evolves to maximum luminosity at an average of roughly 90 days when temperatures are ∼900–1000 K (Gehrz 2008, Table 8.1). Most rock-forming elements will have condensed by this time (Lodders 2003).

Of primary interest, in the context of the ²⁰Ne and ²²Ne measurements, is the timescale for the condensation of Na-bearing minerals. In the modeled ONe ejecta composition chosen here to compare with the noble gas results (Tables 6 and 7), the ratio of the total Na to the major silicate-forming elements (O+Mg+Al+Si) is ∼10% by mass (Starrfield et al. 2009). A possible Na mineral could be albite (NaAlSi₃O₈), which condenses at ∼960 K from a solar composition gas (Lodders 2003). Grain temperatures near this are reached approximately 90 days after the explosion. At this time only 6.4% of the 2.6 yr ²²Na generated in the outburst will have decayed to ²²Ne during its pre-condensation residence in the nova outflow. The resulting ²⁰Ne/²²Ne ratio in the ejecta gas phase at the time of Na condensation is ∼800 from data in Table 6. Subsequent implantations of this ambient Ne into grains, in amounts required by measured ²⁰Ne in the anomalous IDPs (Figure 1), contribute <10% of their average ²²Ne concentration. The ²²Ne inventories remain dominated by in situ decay of ²²Na in grains even if Na precipitation occurs substantially later. For example, at 240 days when grain temperatures are estimated to be ∼200–400 K (Gehrz et al. 1986), ambient ²⁰Ne/²²Ne would be ∼320 and the average implanted ²²Ne contribution <25%. This sequestering of ²²Na in grains as dust condenses and cools in the nova outflow, together with destruction of preexisting ³He and ²²Ne in the TNR, results in a gas-phase He and Ne composition strongly depleted in ³He and ²²Ne.

Laboratory experiments have demonstrated that shocks are effective in implanting ambient gases into solids (e.g., Wiens & Pepin 1988). Shocks, albeit of a different type, should be common in nova outburst environments. It is well known that Rayleigh–Taylor instability causes outflow fragmentation in stellar winds (Shore 2007, Section 9.6.1). The manifestation of this fragmentation is seen in the clumpiness evident in optical images of the shells of old novae such as HR Del (Harman & O'Brien 2003). Dust grains in nova ejecta are most likely to grow within these clumps where they are shielded from ionizing radiation (Gehrz 2008), an effect graphically illustrated by separation of the dust formation zone and the ionization front in the wind of RY Scuti (Smith et al. 2001). The clumps will be subjected to strong radiative shock heating (Shore 2007, Section 4.5.4) by the hard radiation field of the WD central engine, which has a measured surface temperature as high as 8 × 10⁵ K by the time the grains have matured (Petz et al. 2005). Such shocks could lead to stochastic implantation of ambient Ne and He into grains embedded in the outflow clumps, in amounts governed by local shock conditions.

The physical, chemical, and mineralogic character of the anomalous IDP gas carriers has not been determined. The expectation, however, is that ONe novae with low C/O ratios will condense silicates, as observed in novae QU Vul 1984#2 (Gehrz et al. 1986), V2362 Cyg (Helton et al. 2011), and V1065 Cen (Helton et al. 2010). Size estimates of grains emitting silicate spectral features are in the range ∼0.4–1.4 μm, possibly eroded to smaller sizes before ejection into the interstellar medium (Gehrz 2008). An extensive SIMS–SEM search for presolar grains in IDPs from the GSC (Busemann et al. 2009) revealed presolar silicates with high abundances, up to >1% by weight (wt.%), and sizes (∼0.34 μm) notably similar to the spectroscopically identified nova silicates. With abundance of 1 wt.%, roughly 100 of the GSC presolar grains would be present in the average anomalous IDP. If they were the carriers of the nova-linked noble gases, the required He loading would be on the order of 1 cm³ STP g⁻¹ (∼1 He atom per 700 silicate atoms), comparable to levels seen in heavily SW-irradiated IDPs (e.g., Figure 4 and Pepin et al. 2000).

However, the role of the GSC presolar silicates as the actual nova gas carriers is questionable. Isotopic ratio ranges for C, N, O, Mg, and Si calculated from nuclear modeling of neon nova ejecta compositions are shown in Figure 9 (Starrfield et al. 2009; José et al. 2004). The GSC silicate grains exhibit the expected ¹⁷O enrichments, but their Si isotopic compositions differ from the Figure 9 predicted pattern (Busemann et al. 2009). It may be that the carriers have a distinctly different mineralogy. Carbonaceous matter, inferred from spectral observations of the Ne-rich novae V2361 Cyg and V2362 Cyg (Helton et al. 2011), has long been known to be a principal host of noble gases in meteorites (e.g., Frick & Pepin 1981). It is relatively abundant in the GSC IDPs (Busemann et al. 2009; Nakamura-Messenger et al. 2008, 2010).

Zoom In Zoom Out Reset image size

The actual nova gas carriers should display signatures of the non-solar isotopic distributions in C, N, O, Mg, and Si predicted by nuclear modeling (Figure 9). Dilution with normal solar system matter would mute these signatures, as is the case for the Amari et al. (2001) analyses of SiC grains of putative origin in ONe novae, but element-to-element patterns could be preserved. If such anomalies were found, their hosts would help to establish the nature of the gas-carrying grains.