⁶⁰Fe AND ²⁶Al IN CHONDRULES FROM UNEQUILIBRATED CHONDRITES: IMPLICATIONS FOR EARLY SOLAR SYSTEM PROCESSES

Records of now-extinct short-lived radionuclides (SLRs) present at the time of formation of early solar system objects, such as Ca–Al-rich inclusions (CAIs), chondrules, and differentiated meteorites, provide high-resolution relative chronology of events leading to their formation (Russell et al. 2006; Halliday & Kleine 2006; Nyquist et al. 2009). The short-lived nuclide ⁶⁰Fe, which decays to ⁶⁰Ni with a half-life of 1.5 Myr, is a unique product of stellar nucleosynthesis (see, e.g., Meyer & Zinner 2006). A robust value of solar system initial (SSI) ⁶⁰Fe/⁵⁶Fe at the time of CAI formation would allow unambiguous identification of its stellar source and contribution from the same to the inventory of the other co-injected SLRs in the early solar system. A hint for the presence of ⁶⁰Fe in refractory CAIs, considered to be the first solid to form in the solar system, was reported more than two decades ago (Birck & Lugmair 1988). However, the lack of primary phases with high Fe/Ni ratio and possible presence of nucleogenic Ni isotope anomaly make CAIs unsuitable for inferring SSI ⁶⁰Fe. Differentiated meteorites, which provided direct evidence for the presence of live ⁶⁰Fe in early solar system (Shukolyukov & Lugmair 1993), are also prone to disturbance in Fe–Ni isotope systematics. Recent studies of Fe-rich phases (sulfides, oxides, and silicates) in matrix and chondrule from unequilibrated ordinary chondrites (UOCs) reported SSI ⁶⁰Fe/⁵⁶Fe values ranging from 2 × 10⁻⁷ to 1.6 × 10⁻⁶ (Tachibana & Huss 2003; Mostefaoui et al. 2005; Tachibana et al. 2006). However, the time of formation of the analyzed phases was not measured directly and plausible assumptions for the same were made to derive SSI ⁶⁰Fe/⁵⁶Fe. We report here results obtained from the first combined study of ²⁶Al and ⁶⁰Fe records in a set of UOC chondrules. The time of formation of the chondrules, relative to the CAIs, inferred from their Al–Mg isotope records, was used in conjunction with Fe–Ni isotope data to estimate SSI ⁶⁰Fe/⁵⁶Fe.

Chondrules from the UOCs, Semarkona, LEW86314, and Bishunpur were analyzed. Semarkona and Bishunpur are observed falls, while LEW86314 was collected in Antarctica. Studies of various indices of thermal metamorphism in bulk samples as well as individual chondrules (Huss et al. 2006, and references therein) suggest Semarkona (LL3.0) to be the least thermally affected sample followed by LEW86314 (L3.0) and Bishunpur (LL3.1). These UOCs have not experienced temperature exceeding 250–300°C during their residence in meteorite parent bodies (Huss et al. 2006) and chondrules in them should preserve pristine isotope records. Polished sections of these meteorites were studied using electron probe micro analyzer (EPMA) to identify chondrules (see Figure 1) hosting Fe-rich (olivine/pyroxene) and Al-rich (glassy mesostasis) phases. Al–Mg and Fe–Ni isotope studies of selected chondrules were carried out using a Cameca ims-4f ion microprobe at the Physical Research Laboratory (PRL), India, and a Cameca ims-1280 at the University of Hawaii (UH).

Zoom In Zoom Out Reset image size

The experimental procedures adopted at PRL for Al–Mg isotope studies and results obtained for three chondrules were reported earlier (Rudraswami et al. 2008). The Fe–Ni isotope measurements were performed using the ims-4f at a mass resolution (M/ΔM) of ∼4000, sufficient to resolve major interferences at the masses of interest. Magnitude of unresolved hydride interference is at less than per mil level. A focused primary ¹⁶O⁻ beam with intensity ≥5 nA was used to sputter secondary ions from the sample surface that are accelerated to 4.5 kV and energy sorted using an electrostatic analyzer. Ions within a small energy window (∼25 eV) were mass analyzed using an electromagnet and ion counting was done using an electron multiplier. Each analysis consisted of 20 blocks of data, each of 5 cycles, through the mass sequence 56.7, ⁵⁷Fe, ⁶⁰Ni, and ⁶²Ni. Mass 56.7 was included to monitor dynamic background that ranged from 0.01 to 0.02 cps. Typical count times were 10 s (56.7), 1 s (⁵⁷Fe), 10 s (⁶⁰Ni), and 50 s (⁶²Ni) and an analysis lasted about 2 hr. Data for multiple analyses, conducted on the same spot or multiple spots in the same phase (olivine or pyroxene), were combined as long as the count rates and Fe/Ni ratio remained steady to improve precision of the measured isotope ratios. Instrument mass fractionation was monitored by analyzing terrestrial olivine standards (San Carlos and Vernadsky olivine) during and in between sample runs over the entire period of study; the values are close to one per mil per amu (Figure 2). This is also corroborated by the data for low Fe/Ni phases in the analyzed chondrules. The relative ion yield of Fe and Ni from olivine and pyroxene has been extensively studied earlier (Kita et al. 1998; Sugiura et al. 2006) and indicated a dependence on Fe content and the nature of primary ion beam (¹⁶O⁻and O₂⁻). Based on our data for the Vernadsky olivine standard with Fe content (16%) close to those for the analyzed chondrule phases (11%–18%), we adopt a value of unity for relative Ni/Fe ion yield.

Zoom In Zoom Out Reset image size

At UH, Fe–Ni isotope measurements were made using a Cameca ims-1280. A focused, 15–30 μm, 3–5 nA, primary O⁻ beam was rastered over a 20 × 20 μm square on samples and standards. The instrument was operated at 10 kV with a 50 eV energy window and a mass resolving power of ∼5000, sufficient to resolve all interferences except for hydrides. Secondary ions (⁵⁷Fe⁺, ⁶⁰Ni⁺, ⁶¹Ni⁺, and ⁶²Ni⁺) were counted on a monocollector electron multiplier for 0.3, 3, 15, and 5 s in each cycle. Each spot was measured for 2 hr. Data are corrected for electron multiplier background (0.02 cps) and deadtime (29.5 ns). Instrumental mass fractionation was corrected internally using the measured ⁶²Ni/⁶¹Ni ratio. Contributions of interferences of hydrides to count rates of nickel isotopes were less than 1‰ (see Tachibana et al. 2006, 2007 for additional details). The Mg isotope analyses of the Al-rich chondrule glass were conducted using a 0.1 nA focused O⁻ beam. Secondary ions (²⁴Mg⁺, ²⁵Mg⁺, and ²⁶Mg⁺) were collected on the monocollector electron multiplier and ²⁷Al⁺ was measured on the monocollector Faraday cup. Miyake-jima plagioclase and San Carlos olivine were used as standards.

Presence of both high Al/Mg and high Fe/Ni phases in individual chondrule is rare. In general, porphyritic chondrules hosting glassy phase (mesostasis) with high Al/Mg are devoid of high Fe/Ni phase, while reverse is the case for non-porphyritic chondrules. Based on EPMA data for more than a few hundred chondrules, we selected about two dozen chondrules from the three UOCs for this study. In 10 of them we could obtain data for either Al–Mg or Fe–Ni isotope system or both (see Table 1). Four chondrules (SEM3, SEM21, SMK3–6, and LEW36) yielded data for both; the Fe–Ni isotope data for these are shown in Figure 3. The initial ⁶⁰Fe/⁵⁶Fe at the time of formation of these chondrules, defined by the slope of the best-fit line through each data set, along with their inferred initial ²⁶Al/²⁷Al (this study and Rudraswami et al. 2008) are also shown in the figure. In three additional chondrules, we could obtain initial ⁶⁰Fe/⁵⁶Fe; however, they yielded only upper limits for initial ²⁶Al/²⁷Al.

Zoom In Zoom Out Reset image size

Chondrules are products of melting of precursor microscopic solids followed by rapid cooling in the solar nebula (Scott & Krot 2005). The melting events homogenized the isotopic compositions of the chondrules, resetting the isotopic clocks, and the inferred initial ²⁶Al/²⁷Al and ⁶⁰Fe/⁵⁶Fe represent values at the time of their formation. The observed spread in initial ²⁶Al/²⁷Al values (0.55–1.6) × 10⁻⁵, combined with the canonical ²⁶Al/²⁷Al value of 5 × 10⁻⁵ for CAIs, suggest that these chondrules formed between 1.2 and 2.1 Ma after CAIs. If we consider the initial ⁶⁰Fe/⁵⁶Fe values for the four chondrules (Figure 3) and their mean time of formation, they yield SSI ⁶⁰Fe/⁵⁶Fe values (±2σ) of (8.3 ± 4.7) × 10⁻⁷ (SEM3), (4.3 ± 2.8) × 10⁻⁷ (SMK3–6), (7.2 ± 4.8) × 10⁻⁷ (LEW36), and (9.9 ± 5.3) × 10⁻⁷ (SEM21) and define a weighted average SSI ⁶⁰Fe/⁵⁶Fe value of (6.3 ± 2) × 10⁻⁷. This value is within the range inferred in an earlier study, assuming a time interval of 1.5–2 Ma between the formation of CAIs and the analyzed UOC chondrules (Tachibana et al. 2006). Our results do not support the SSI value of ∼1.6 × 10⁻⁶ inferred from the analysis of an Allende CAI (Birck & Lugmair 1988). It is also at variance with values inferred from studies of UOC sulfides of unknown formation time (Tachibana & Huss 2003; Mostefaoui et al. 2005). We note that during the course of this study a revised half-life of 2.6 Ma for ⁶⁰Fe has been proposed (Rugel et al. 2009). If true, this will lower the SSI ⁶⁰Fe/⁵⁶Fe value obtained in this study by ∼30% and does not alter the above inferences.

The use of various SLRs as relative chronometers of early solar system events is based on the assumptions that they were homogeneously distributed in the early solar system and characterized by unique SSI abundances. The above estimate of SSI ⁶⁰Fe/⁵⁶Fe, using ²⁶Al as a time marker, thus requires that the SLRs ⁶⁰Fe and ²⁶Al satisfy these assumptions. Self-consistent chronology obtained for solar system events based on Pb–Pb, ¹⁸²Hf–¹⁸²W, ⁵³Mn–⁵³Cr, and ²⁶Al–²⁶Mg isotope systems (Russell et al. 2006; Halliday & Kleine 2006; Nyquist et al. 2009) suggest that the SLRs, ²⁶Al, ⁵³Mn, and ¹⁸²Hf, were homogeneously distributed in the early solar system and have well-defined SSI abundances. However, based on the observed apparent deficits in ⁶⁰Ni from meteorites that formed very early in solar system history, Bizzarro et al. (2007) suggested a late injection of ⁶⁰Fe into the solar system, relative to ²⁶Al, from a massive star. It was proposed that ²⁶Al was injected via stellar wind during the early evolution of the star, and ⁶⁰Fe, about a million year later, when it ended its life as a Type Ib/Ic supernova. If true, the SLR ¹⁸²Hf, which is co-produced with ⁶⁰Fe in stellar nucleosynthesis (Meyer 2005; Meyer & Zinner 2006), would also be injected late into the solar system. A very systematic study of Fe–Ni isotope records of various meteorites (Dauphas et al. 2008) could not reproduce the results of Bizzarro et al. (2007) and negates the late injection hypothesis for ⁶⁰Fe. Further, records of ¹⁸²Hf and ²⁶Al in CAIs and later formed angrites of known formation ages are also consistent with their homogeneous distribution in the early solar system (Burkhardt et al. 2008; Nyquist et al. 2009). These results argue for co-injection and homogeneous distribution of ⁶⁰Fe, ²⁶Al, and ¹⁸²Hf in the early solar system and validate our approach for inferring SSI ⁶⁰Fe/⁵⁶Fe based on combined Fe–Ni and Al–Mg isotope data for chondrules.

A plot of initial ⁶⁰Fe/⁵⁶Fe versus initial ²⁶Al/²⁷Al for the analyzed chondrules is shown in Figure 4. The time of formation of the chondrules, relative to CAIs, is also indicated. If ²⁶Al and ⁶⁰Fe were co-injected into the nascent solar system and distributed homogeneously within the solar system, the data should fall along a linear trend in this log–log plot. In spite of the relatively large error and upper limit values for ²⁶Al/²⁷Al in several cases, the data display such a trend and support co-injection of these two SLRs. This makes ⁶⁰Fe a potential heat source, along with ²⁶Al, for early thermal metamorphism of planetesimals (see, e.g., Yoshino et al. 2003). Our results suggest that with improved analytical precision in measurements of Fe–Ni isotope systematics, ⁶⁰Fe could be a viable chronometer for early solar system events.

Zoom In Zoom Out Reset image size

⁶⁰Fe is a product of s-process nucleosynthesis via neutron capture and both a core-collapse supernova (SN-II) and a high-mass (≥5 M_☉) thermally pulsing (TP)-asymptotic giant branch (AGB) star could be a source of this nuclide. Low-mass AGB stars do not attain the temperature required for activation of the high-density ²²Ne neutron source needed for effective production of ⁶⁰Fe (Wasserburg et al. 1995, 2006). Contemporaneous injection of ⁴¹Ca and ²⁶Al from a stellar source into the protosolar cloud had been proposed earlier (Sahijpal et al. 1998, 2000) and the present study suggests co-injection of ²⁶Al and ⁶⁰Fe. Thus, any proposed stellar source should be able to account for the inferred SSI ⁶⁰Fe/⁵⁶Fe as well as those for ²⁶Al/²⁷Al and ⁴¹Ca/⁴⁰Ca reported previously (see, e.g., Goswami et al. 2005). The same source should ideally account for the SSI abundances of other SLRs (e.g., ¹⁸²Hf and ⁵³Mn) that are also homogeneously distributed in the early solar system along with ²⁶Al and ⁶⁰Fe (Dauphas et al. 2008; Nyquist et al. 2009). Identification of a plausible stellar source is based on its ability to concurrently match the SSI abundance of multiple SLRs by considering injection of a specific amount of freshly synthesized stellar material, containing the SLRs, into the protosolar cloud and assuming a time interval between their synthesis and eventual incorporation into the first solar system solids. Studies of nucleosynthesis yields from a massive (≥25 M_☉) star that ends its life as a Type-II supernova suggest that it could be a plausible source for multiple SLRs. However, it is necessary to invoke mixing and fall back of material in the inner region of such stars as well as an injection mass cut for ejected stellar material containing freshly synthesized SLRs (Meyer 2005; Meyer & Adams 2006; Takigawa et al. 2008). The possibility of a high-mass TP-AGB star as a source of SLRs present in the early solar system has also been considered by Wasserburg et al. (2006) and Trigo-Rodrigues et al. (2009). Reasonable match with SSI abundances (within a factor of 2) could be obtained for ²⁶Al, ⁴¹Ca, and ⁶⁰Fe. However, both the amount of injected stellar material and the time interval between the production of these SLRs and their incorporation into solar system solids differ significantly in these two approaches. Further, unlike the case of massive stars, whose association with star-forming regions is well documented, the probability of association of a high-mass AGB star with star-forming regions is extremely small, ≤10⁻⁵ (see, e.g., Huss et al. 2009), making such an association very unlikely. It is also necessary to invoke additional contribution from a massive star to explain the presence of the SLRs ¹⁸²Hf and ⁵³Mn that are not products of nucleosynthesis in AGB stars. Injection of freshly synthesized material from an SN associated with a massive star remains the most plausible source for ⁶⁰Fe and several other SLRs present in the early solar system.