HETEROGENEOUS DISTRIBUTION OF ²⁶Al AT THE BIRTH OF THE SOLAR SYSTEM

Acid-resistant residues of ordinary and CCs used in this study were prepared by Huss & Lewis (1995) and Huss et al. (2003). The residues consist of SiC, spinel (MgAl₂O₄), hibonite (CaAl₁₂O₁₉), and corundum and range in size from 0.5 to 5 μm. The residues were diluted by a mixture of 90% isopropanol and 10% water and mixed in an ultrasonic bath. About 0.2 μl of the mixture was siphoned using a micro-aspirator and dispersed onto a gold substrate under a stereomicroscope. More than 100 grains were dispersed on each gold substrate. Prior to the dispersion, the gold substrate was examined for possible contamination using the UH JEOL JXA-8500F field emission electron microprobe (FE-EPMA) equipped with a cathodoluminescence detector (CL) and an energy dispersive spectrometer (EDS). Several relatively large (>5 μm) corundum grains embedded in the substrates were identified and avoided during subsequent oxygen- and magnesium-isotopic measurements. To locate the chondritic corundum grains comprising less than 1% of the dispersed grains dominated by spinel, we used CL and EDS of the FE-EPMA. Among the dispersed grains, only corundum and SiC show intense CL; the latter is very rare and can easily be distinguished by EDS. The corundum grains identified were imaged in secondary and backscattered electrons.

Oxygen-isotope compositions of individual corundum grains were measured with the UH Cameca ims-1280 (for details see Makide et al. 2009a).

Aluminum- and magnesium-isotope measurements with UH Cameca ims-1280 were conducted on the corundum grains which have been previously analyzed for oxygen isotopes. An ¹⁶O⁻ primary beam, either defocused (∼30 μm spotsize) or focused (∼3 μm), was used to generate secondary Mg⁺ and Al⁺ ions. A field aperture of 1000×1000 μm² corresponding to ∼7 μm on the sample was used to minimize the contribution of Mg signals from the substrate and other dispersed grains surrounding the grain of interest. The contribution of the Mg signal from the substrate was estimated to be <1% of Mg from the sample. The mass resolving power was set to ∼3500–3800, sufficient to separate interfering hydrides and ⁴⁸Ca⁺⁺. Secondary ions of ²⁴Mg⁺, ²⁵Mg⁺, and ²⁶Mg⁺ were measured simultaneously using the monocollector electron measure (EM) and two multicollector EMs; ²⁷Al⁺ was subsequently measured with the monocollector EM by peak-jumping. The primary beam current was adjusted so that the ²⁷Al⁺ count rate did not exceed 4.5×10⁵ cps. Acquisition times for Mg isotopes and ²⁷Al were 15 s and 1 s, respectively. Each measurement consisted of 20 cycles. Excess of radiogenic ²⁶Mg (δ²⁶Mg*) was calculated using the following equation: δ²⁶Mg* = Δ²⁶Mg − 2×Δ²⁵Mg, where Δ^25,26Mg = ((^25,26Mg/²⁴Mg)_sample)/(^25,26Mg/²⁴Mg)_ref−1) × 1000 and reference Mg-isotope ratios are from Catanzaro et al. (1966). We note that Δ^25,26Mg values wre calculated using the mean of ratios for all cycles. Since Mg count rates are low, it results in bias in ²⁵Mg/²⁴Mg and ²⁶Mg/²⁴Mg ratios (Ogliore et al. 2011). The typical reproducibility of standard measurements was ∼10‰ (2σ) in Δ²⁵Mg and Δ²⁶Mg. To correct for instrumental mass fractionation effects and estimate the relative sensitivity factor for the ²⁷Al/²⁴Mg ratio, we used a Montana sapphire (α-Al₂O₃) standard (NMNH 126321 from the Smithsonian Institution). The instrumental fractionation correction was made based on the measurements of micron-sized Montana sapphire grains dispersed on the gold substrate. To correct the ²⁷Al⁺/²⁴Mg⁺ ratios measured by SIMS in micron-sized corundum grains, we measured the Mg concentration in the Montana sapphire grain by the UH FE-EPMA; a piece of the same grain was also measured by multicollector inductively coupled plasma mass spectrometry (MC-ICPMS) by E. D. Young (UCLA). To measure trace amount of Mg in the sapphire by FE-EPMA, we used defocused (10 μm) electron beam at ∼300 nA current and 15 keV accelerating voltage. Madagascar hibonite was used as a standard. The ²⁷Al/²⁴Mg ratios in the sapphire obtained by FE-EPMA and LA MC-ICP-MS are 6084±498 (2σ, 45 spots) and 5830 ± 510 (2σ, 4 spots), respectively, and are the same within the uncertainties and show no evidence for heterogeneity.

Oxygen-isotope compositions of micron-sized corundum grains from acid-resistant residues of UOCs (Semarkona (LL3.0), Bishunpur (LL3.1), Roosevelt County (RC) 075 (H3.1)) and unmetamorphosed CCs (Orgueil (CI1), Murray (CM2), Renazzo (CR2), and Allan Hills (ALH) A77307 (CO3.0)) together with compositions of the mineralogically pristine CAIs from CR2 CCs (Makide et al. 2009b) and the solar wind returned by the Genesis spacecraft (McKeegan et al. 2010) are listed in online Table 1 and plotted in Figure 1. All but three corundum grains have ¹⁶O-rich compositions similar to those of CR CAIs and solar wind, consistent with a condensation origin from an ¹⁶O-rich gas of solar composition. Three corundum grains, not shown in Figure 1, have highly anomalous O-isotope compositions, and are probably presolar in origin.

Zoom In Zoom Out Reset image size

The solar corundum grains were subsequently measured for magnesium-isotope compositions and ²⁷Al/²⁴Mg ratio; the results are listed in Table 1 and plotted in Figure 2. The grains show large variations of the inferred initial ²⁶Al/²⁷Al ratio ((²⁶Al/²⁷Al)₀): 52% of grains have no resolvable ²⁶Mg*: an upper limit on (²⁶Al/²⁷Al)₀ is 2×10⁻⁶; 40% of grains have high (²⁶Al/²⁷Al)₀, (3.0–6.5)×10⁻⁵; 8% of grains have intermediate values of (²⁶Al/²⁷Al)₀, (1–2)×10⁻⁵.

Zoom In Zoom Out Reset image size

The coexistence of the ²⁶Al-rich and ²⁶Al-poor corundum grains in the same primitive meteorite precludes late-stage (after decay of ²⁶Al) resetting of the ²⁶Al–²⁶Mg systematics of the ²⁶Al-poor corundum grains during thermal metamorphism in the host chondrite parent body. Late-stage resetting in the solar nebula during thermal processing associated with chondrule formation is also excluded. Chondrule formation occurred in an ¹⁶O-poor (Δ¹⁷O > −10‰) nebular reservoir (e.g., Krot et al. 2006) when ²⁶Al was still alive (Kita et al. 2005; Kurahashi et al. 2008; Krot et al. 2009). We infer that ¹⁶O-rich corundum grains with low initial ²⁶Al/²⁷Al ratios never contained the canonical abundance of ²⁶Al, i.e., the lack of ²⁶Mg excess in 52% of ¹⁶O-rich corundum grains is a primary characteristic.

A high abundance (>50%) of ²⁶Al-poor, ¹⁶O-rich CAIs has been previously reported in CH CCs (Krot et al. 2006). Similar to corundum grains, the CH CAIs are among the smallest (<20 μm) and the most refractory (rich in hibonite and grossite, CaAl₄O₇) objects known (Krot et al. 2002). The ²⁶Al heterogeneity and high abundance of ²⁶Al-poor, ¹⁶O-rich refractory objects in a fine fraction of solar dust can be explained in the context of the formation setting of refractory objects (MacPherson et al. 1995; Krot et al. 2009) and theoretical modeling of their radial transport in the protoplanetary disk (Boss 2008; Ciesla 2009, 2010; Yang & Ciesla 2010). Refractory objects are thought to have formed in region(s) with high ambient temperature (at or above condensation temperature of forsterite (Mg₂SiO₄), ∼ 1350 K at total pressure of <10⁻⁴ bar), most likely within 1–2 astronomical units from the Sun (Krot et al. 2009, and references therein). They were subsequently transported outward to the accretion regions of chondrite parent asteroids and cometary nuclei (Gounelle & Meibom 2007; Boss 2008; Ciesla 2009, 2010). The oldest ²⁰⁷Pb–²⁰⁶Pb absolute ages of CAIs from CV (Vigarano-like) chondrites (Amelin et al. 2002; Bouvier & Wadhwa 2010), a narrow range of their initial ²⁶Al/²⁷Al ratios (Jacobsen et al. 2008), corresponding to a range of crystallization ages of <0.1 Ma, and the high ambient temperature in the CAI-forming region(s), are all consistent with an early and brief duration of CAI formation, during a period of high mass-accretion rate of dust and gas to the proto-Sun (∼10⁻⁶ M_☉ yr⁻¹; Krot et al. 2009; Ciesla 2009). We propose that the ²⁶Al-poor and ²⁶Al-rich ¹⁶O-rich corundum grains represent different generations of refractory objects formed during a brief epoch of CAI formation, and, therefore, reflect heterogeneous or variable ²⁶Al at the birth of the SS.

To understand the distribution of refractory objects with the different (²⁶Al/²⁷Al)₀ in the early SS, we modeled the evolution of a viscous disk and redistribution of refractory objects formed in the high-temperature (above 1400 K) region within it during the first million years of its evolution (Figure 3). We consider both the mass and angular momentum transport in the disk as well as the addition of mass to both the Sun and the disk due to infall from the parent molecular cloud (Yang & Ciesla 2010; Hueso & Guillot 2005). The parent cloud is assumed to be 1 M_☉, spherical, at a temperature of 15 K, rotating with a uniform angular velocity of Ω_c = 10⁻¹⁴ s⁻¹, and we adopt a mass accretion rate of ∼3×10⁻⁶ M_☉ yr⁻¹; the disk viscosity is parameterized with a turbulent viscosity coefficient of α = 10⁻³. The accreted mass falls onto the disk at locations where the specific angular momentum in the disk equals that in the parent cloud (Yang & Ciesla 2010; Hueso & Guillot 2005). We track populations of fine refractory objects as they are redistributed throughout the disk by viscous expansion, dividing them into batches depending on the timing of when they are last exposed to the high-temperature region (Figures 3(a) and (b)).

Zoom In Zoom Out Reset image size

We find that the mass of the disk reaches its maximum by 300,000 years of its evolution (Figure 3(a)), and that the most dominant population of refractory objects in a 1 million year old disk consists of those that formed immediately before and after cessation of infall from the parent molecular cloud (Figure 3(b)). This prediction, combined with the observed heterogeneity of (²⁶Al/²⁷Al)_0, suggests that ²⁶Al was introduced into the SS contemporaneously with the collapse of the protosolar cloud core; injection into the disk (Ouellette et al. 2007) is much less likely. Alternatively, if ²⁶Al was heterogeneously distributed in the parent molecular cloud, mixing in the disk did not homogenize it until the infall and formation of refractory solids had ceased. Thus, refractory objects formed prior to the end of collapse would not record the average solar nebula ²⁶Al/²⁷Al ratio, but rather the ratio that existed in the hot nebular region at the time of their formation. This would allow refractory objects with different initial ²⁶Al/²⁷Al ratios (²⁶Al-poor and ²⁶Al-rich) to form and to be preserved in the disk. It also implies that the initial ²⁶Al/²⁷Al ratio of (5.23 ± 0.13)×10⁻⁵ recorded by CV CAIs (Jacobsen et al. 2008; Davis et al. 2010) may not necessarily represent the initial ²⁶Al/²⁷Al ratio in the SS.

Our modeling indicates that a significant population of refractory objects that formed prior to the cessation of infall survives in the disk (Figure 3(b)). The survival time of dust particles (up to ∼1 m in size) in the disk is inversely proportional to their sizes: fine (5 μm) dust is well-coupled to the nebular gas and less affected by head wind than the coarse (∼1 mm) dust (Ciesla 2010). These observations may explain the high proportion of ²⁶Al-poor refractory objects in a fine fraction of solar dust, e.g., as we have observed in corundum grains and CH CAIs, as well as the presence of an ²⁶Al-poor, ¹⁶O-rich CAI in comet Wild2 (Matzel et al. 2010).

The heterogeneous distribution of ²⁶Al at the birth of the SS may preclude inheritance of ²⁶Al from a molecular cloud enriched by a previous generation of stars (Gounelle et al. 2009), which is expected to have a uniform distribution of ²⁶Al. Instead, it indicates injection of ²⁶Al nearly contemporaneously with the collapse molecular cloud core that formed the SS. The similar oxygen-isotope compositions of the ²⁶Al-rich and ²⁶Al-poor corundum grains and the Sun suggest injection of ²⁶Al by a massive (>20 M_☉) Type II SN, by a wind from a WR star or by a low-mass (1.5 M_☉) AGB star (Sahijpal & Soni 2006; Krot et al. 2008; Ellinger et al. 2010). Injection of ²⁶Al by a less massive SN or by a 3–6 M_☉ AGB star is unlikely, because it is expected to produce significant change in oxygen-isotope composition of the SS that has not been observed (Sahijpal & Soni 2006; Gounelle & Meibom 2007; Ellinger et al. 2010).

To distinguish between an explosive SN and a WR wind of ²⁶Al, measurements of other isotopes (e.g., ⁶⁰Fe–⁶⁰Ni isotope systematics) of ²⁶Al-rich and ²⁶Al-poor ¹⁶O-rich CAIs or refractory grains are required: in contrast to SNe, the WR stars do not produce ⁶⁰Fe (Sahijpal & Soni 2006). Therefore, if ²⁶Al was injected by a WR star (Gaidos et al. 2009; Tatischeff et al. 2010) and ⁶⁰Fe was inherited from a molecular cloud of an earlier generation (Gounelle et al. 2009), the two isotopes would be decoupled, i.e., the ²⁶Al-rich and ²⁶Al-poor CAIs will have similar initial abundance of ⁶⁰Fe. In contrast, if ²⁶Al and ⁶⁰Fe were injected by the same SN, the ²⁶Al-poor CAIs would have lower initial abundance of ⁶⁰Fe than the ²⁶Al-rich CAIs.

We note that the initial abundance of ⁶⁰Fe in the SS remains controversial. The inferred initial ⁶⁰Fe/⁵⁶Fe ratios range from (2–5)×10⁻⁷ (Telus et al. 2011) to <1×10⁻⁸ (Tang & Dauphas 2011). Due to the presence of nucleosynthetic isotopic anomalies in Ni, the detection of radiogenic ⁶⁰Ni in bulk CAIs is difficult (Quitté et al. 2007) and may require high-precision Fe–Ni isotope measurements of CAI mineral separates to construct an internal Fe–Ni isochron.