PRE-ACCRETIONAL SORTING OF GRAINS IN THE OUTER SOLAR NEBULA

Collected from the Earth's stratosphere, chondritic porous interplanetary dust particles (CP IDPs) are small (typically <50 μm), extremely fragile, fine-grained porous aggregates composed of crystalline mineral grains, amorphous silicates, and carbonaceous material (Figure 1). The most abundant crystalline mineral grains are anhydrous silicates, iron-rich sulfides, and iron–nickel metal grains, and the most abundant form of amorphous silicates are known as glass with embedded metal and sulfides (GEMS; Bradley 2013). The chemical and physical properties of CP IDPs are consistent with an origin from comets: their porous, fragile structures are consistent with the inferred properties of cometary meteors that readily disaggregate during atmospheric entry (Verniani 1969); their ∼10 μm infrared silicate features are similar to those observed in the astronomical spectra of comets (Sandford & Walker 1985; Bradley et al. 1999; Molster & Waters 2003); the high speeds at which some CP IDPs enter the atmosphere, derived from He release profiles, are consistent with capture from cometary orbits (Brownlee et al. 1993; Brownlee & Joswiak 1995); and their anhydrous mineralogy, high carbon content, and high abundance of isotopically anomalous presolar grains are consistent with an origin from small, frozen parent bodies in the outer solar system (Bradley 2013; Messenger 2000; Nguyen et al. 2007). Given that comets accreted in the outer regions of the solar nebula, beyond 5 AU, and have since remained in distant reservoirs (the Kuiper Belt and Oort Cloud), they are thought to have undergone minimal alteration since accretion and hence preserve the best examples of early outer solar system materials (Brownlee 2003). Indeed, the CP IDPs thought to sample them are extremely primitive with the highest abundance of measurable presolar grains (Floss et al. 2006) and no evidence of significant post-accretional parent body alteration (Bradley & Brownlee 1986; Bradley 2013). Therefore, despite their small sizes, CP IDPs are an extremely important class of extraterrestrial material, having the potential to provide a significant amount of information regarding the conditions and mechanisms operating in the outer, comet-forming regions of the solar nebula.

Zoom In Zoom Out Reset image size

Our recent comparative study of the sizes and size distributions of crystalline silicate and sulfide grains in the CP IDPs U220GCA, U211B6, and U212A34A and in comet 81P/Wild 2 samples collected by NASA's Stardust mission found that the grain size distributions of these components within each object were statistically significantly different from one another (Wozniakiewicz et al. 2012). From the calculated mean sizes, these components were found to vary from object to object, with sulfides being consistently smaller than their accompanying silicates. Comparisons between the two components in each object showed that they exhibit a size–density relationship indicative of a sorting mechanism that operated in the outer solar nebula, although the exact mechanism responsible remains unknown.

In an effort to explore and understand the sorting relationship observed, we aimed to study and compare the grain sizes and size distributions of other components in these CP IDPs to determine whether they too are sorted—and, if so, whether they are sorted in the same manner. Good statistics are imperative when performing such comparative studies. Therefore, we chose to study an additional component that is both common and abundant in CP IDPs—the amorphous silicates known as GEMS. Our previous study of crystalline silicates and sulfides included data for cometary particles collected by Stardust, and although comet 81P/Wild 2 may have contained GEMS, the method of collection for the Stardust samples (via hypervelocity impact) prevents positive identification of them (e.g., see Ishii et al. 2008 for examples of impact-generated "GEMS-like" objects). Consequently, we have also performed measurements of GEMS, silicates, and sulfides in an additional CP IDP (a Grigg-Skjellerup particle, GS1) to enable more reliable identification of trends and to determine whether the previously identified silicate and sulfide size–density relationship still holds.

In preparation for grain size analysis, CP IDPs were disaggregated and dispersed onto copper mesh transmission electron microscope (TEM) grids (Figure 1(b)) by using the method of Brownlee & Joswiak (2004). The constituent grains of CP IDPs are found to separate readily without significant fragmentation because of their characteristic fragile, loosely aggregated structures. The size data for silicate, sulfide, and GEMS component grains in the four CP IDPs of this study (U220GCA, U211B6, U212A34A, and GS1) were then obtained using scanning transmission electron microscopy (STEM). Analyses were performed using an 80–300 keV FEI TITAN³ G2 STEM equipped with a monochromator, dual spherical aberration (C_s) correctors for focused probe and imaging modes, an EDAX Genesis 4000 Si(Li) solid-state energy dispersive X-ray (EDX) spectrometer, and a Gatan Tridiem GIF (Gatan Imaging Filter). By using a combination of high-angle annular dark field (HAADF) imaging and EDX analyses, the sizes and compositions of grains were determined, allowing the compilation of size data for each component (crystalline silicates and sulfides and GEMS) in each CP IDP (Figure 2). Areas totaling approximately 3000 μm² on the U211B6 grid, 1000 μm² on the U220GCA grid, 3000 μm² on the U212A34A grid, and 2000 μm² on the GS1 grid were surveyed (different areas were analyzed because of different IDP sizes as well as different degrees of grain dispersion). Full details of the preparation and data collection methods can be found in Wozniakiewicz et al. (2012).

Zoom In Zoom Out Reset image size

Once the data were collected and compiled, Kolmogorov–Smirnov (K-S) tests comparing the size distributions of the different components were performed. These indicated that the populations of silicate, sulfide, and GEMS grains within each CP IDP and between the four CP IDPs are statistically different from one another. In order to quantify this difference, the mean radii and standard deviations were determined for each data set. Specifically, the geometric mean radii (M_r) and standard deviations (+σ and −σ in nanometers) were determined since the size distribution profiles are skewed toward smaller sizes and best fit by log-normal distributions (as determined by performing K-S tests comparing the data sets against cumulative distributions of possible fit distributions). Because of data binning during compilation, values of M_r, +σ, and −σ could not be calculated directly via analytical expressions and were instead determined graphically using the modified geometric Folk & Ward (1957) equations of Blott & Pye (2001) as described in Wozniakiewicz et al. (2012, 2013). Results are shown in Figure 2 and Table 1.

The cumulative frequency distributions of grain sizes for the silicates, sulfides, and GEMS in each CP IDP are shown in Figure 2, with vertical lines marking the geometric mean radii for each component. With the addition of the CP IDP GS1 data set, we note that the sulfide distribution profiles remain consistently offset toward smaller sizes in comparison with their accompanying silicates. In Figure 3, the geometric mean radii of the silicates (M_r-silicate) are plotted against those of their accompanying sulfides (M_r-sulfide) for the four CP IDPs. Also included in this plot are the data for cometary samples returned by Stardust (derived by Wozniakiewicz et al. 2012 from data in Price et al. 2010). Here error bars represent the standard error of the mean, and an error-weighted trend line has been fitted (M_r-sulfide = 0.60M_r-silicate). This slope is almost identical to that reported previously in Wozniakiewicz et al. (2012)for CP IDPs U220GCA, U212A34A, and U211B6 and Stardust samples alone and, as reported therein, is approximately equal to the ratio of the average grain densities: Assuming densities of 3.22 g cm⁻³ for the silicates and 4.71 g cm⁻³ for the sulfides (the midpoints of forsterite and enstatite densities and the range of possible Ni-free pyrrhotite densities, respectively, Deer et al. 1992), the ratio of silicate to sulfide densities is approximately 0.68. Although only measured in CP IDP U220GCA, rare crystalline FeNi metal grains (not embedded in GEMS) also appear to demonstrate size–density equivalence with their accompanying silicates and sulfides, having cumulative frequency distributions shifted toward smaller radii as would be predicted for such dense grains (7.91 g cm⁻³; Henderson & Perry 1954). These grains are not, however, included in the present data set because of the large uncertainties associated with the small population of 26 grains measured to date.

Zoom In Zoom Out Reset image size

Unlike the silicates and sulfides, the geometric mean radii of the GEMS (M_r-GEMS) do not vary consistently with those of their accompanying silicates and sulfides: although the largest GEMS are found in the CP IDP with the largest silicates and sulfides (U220GCA) and the smallest GEMS are in the CP IDP with the smallest silicates and sulfides (U212A34A), GS1 has larger GEMS than U211B6 despite having smaller silicates and sulfides (Figure 2). In addition, the GEMS in U220GCA and U211B6 are larger than their accompanying sulfides and smaller than their silicates, while those in GS1 and U212A34A are larger than both their silicates and sulfides. These differences can be visualized by observing the position of the mean radii of GEMS in each CP IDP in Figure 2, indicated by a vertical black dashed line, relative to the mean radii of the silicates and sulfides, indicated by vertical green and blue lines, respectively. Figure 4 is a plot that compares the GEMS mean radii (M_r-GEMS) with their accompanying crystalline silicates (M_r-silicate) in gray and sulfides (M_r-sulfide) in black for all CP IDPs in this study. The poor fit of GEMS to the size–density equivalency trend observed between crystalline components is evident in Figure 4, where although the (origin-forced) error weighted fits are plausible (they point to a GEMS density that is less than sulfides and close to silicates), they are poor fits to the data points.

Zoom In Zoom Out Reset image size

Values of geometric mean radii, the standard deviations, the standard error of the mean (defined as the standard deviation over the square route of the number of grains), and the number of grains measured are summarized in Table 1.

Our data demonstrate that CP IDPs sample different populations of crystalline silicates and sulfides and GEMS. The crystalline silicates and sulfides (and the metal grains measured to date) exhibit a size–density relationship that suggests they have been sorted together prior to accretion (as reported in Wozniakiewicz et al. 2012). A similar relationship is observed between the silicate chondrules and iron sulfide and metal grains in primitive chondritic meteorites, although these objects are orders of magnitude larger and sample asteroidal bodies that accreted much closer to the Sun: these meteoritic components have narrow size distributions of comparable width that are shifted toward smaller sizes for the denser sulfides and metal grains (e.g., Hughes 1978; King & King 1978; Rubin & Keil 1984; Rubin & Grossman 1987; Dodd 1976; Skinner & Leenhouts 1993; Kuebler et al. 1999). It has been proposed that the meteoritic components have undergone aerodynamic sorting, whereby grains of equivalent size–density product are sorted together (e.g., Kuebler et al. 1999). Several mechanisms of aerodynamic sorting have been put forward to account for the meteoritic observations, such as the turbulent concentration of grains in low-vorticity zones in conjunction with parent body accretion (e.g., Cuzzi et al. 2008), sorting as a result of gas dynamics (Clayton 1980), or sorting during transport (e.g., Liffman 2005). Although similar sorting mechanisms may have operated in the outer, comet-forming regions of the solar nebula, their ability to sort grains at the sizescales applicable to CP IDP components is problematic given current models: this may be due to our poor understanding of the processes operating, or the conditions existing, within the outer nebula. Consequently, at present, it is unknown what physical mechanism or mechanisms are responsible for the observed sorting of the crystalline components in these cometary samples (Wozniakiewicz et al. 2012).

Although the CP IDPs sample different populations of GEMS, the GEMS do not appear to share similar size–density relationships with their accompanying silicates and sulfides. It could be argued that the GEMS density is poorly constrained, varying between GEMS grains as their porosity and silicate/sulfide/metal abundances vary. This could obscure a size–density relationship with the silicates and sulfides, despite sorting via the same mechanism. However, if so, it would be difficult to explain the different, yet tightly constrained, size ranges for GEMS in different CP IDPs: A variation in grain density might be expected to result in less efficient sorting and broader distributions, yet the size distributions of GEMS are, in most cases, better constrained than those of the crystalline silicates and sulfides.⁵ Instead, it could be argued that the GEMS densities may vary from one CP IDP to another, so that the single values of density used in this work adequately describe the crystalline silicates and sulfides but would not adequately describe the GEMS. In order to maintain a size–density relationship with their accompanying silicates and sulfides, this would require that GEMS densities vary from ∼2 g cm⁻³ in CP IDPs U212A34A and GS1to ∼3.5 g cm⁻³ in U220GCA and U211B6. This, combined with the moderately well-sorted nature of these grains, means that we would expect clear differences in the abundance of denser phases in GEMS (the embedded metals and sulfides) between samples if this were the case, yet no such difference was evident during their examination by TEM.

Our data may indicate that the reason behind the disparate sorting trends we observe between the crystalline and amorphous components may lie in their origins. It is generally accepted that the majority of crystalline silicate and sulfide grains in CP IDPs were likely formed within the early solar nebula (Bradley et al. 1983; Bradley 2013). Astronomical observations of disks around other stars at mid-infrared wavelengths show that abundant crystalline silicates are often present in the cold outer disk regions (e.g., van Boekel et al. 2004). Since the dominant dust grain components of the interstellar medium (ISM) are amorphous silicates (>97%, Kemper et al. 2004), these crystalline silicates must have formed within the disks by complete reprocessing of the ISM amorphous silicates via annealing or melting, followed by crystallization or vaporization, followed by recondensation. Indeed, enstatite crystals with whisker and platelet morphologies and enstatite and forsterite crystals with compositions relatively enriched in Mn are commonly observed in CP IDPs and are both characteristics thought to be indicative of vapor phase condensation (Bradley et al. 1983; Klöck et al. 1989). Experimental data have shown it likely that CP IDP sulfides were also produced in the solar nebula by condensation from nebula gases (Nakazawa et al. 1973) or via sulfidization of preexisting Fe–Ni metal grains (e.g., Lauretta et al. 1997; Zolensky & Thomas 1995), with the former being able to reproduce characteristic crystallographic structures observed in CP IDPs (experiments produced nickel-free pentlandite structures similar to the cubic "spinel-like" sulfides identified in CP IDPs by Dai & Bradley 2001).

In contrast to the crystalline components, the origin of most GEMS remains uncertain. Because of GEMS' small sizes (typically 50–500 nm in diameter) and amorphous structure, their chemistries are significantly more susceptible to subtle alteration processes, and those alterations are more difficult to diagnose relative to their associated crystalline components. Alteration may have occurred prior to accretion on the parent body or even after CP IDPs release from the parent body since these porous aggregates provide little isolation of GEMS from changes in their surrounding environment. As a result, it remains problematic to determine with any certainty the origin of most individual GEMS grains, although some researchers have made attempts to establish classifications by origin (Keller & Messenger 2011). Nonetheless, some GEMS large enough for statistically significant isotope measurements have been unambiguously identified as presolar, demonstrating that they are remnant dust from the ISM incorporated into the presolar molecular cloud (Messenger et al. 2003; Floss et al. 2006). It has been proposed that such GEMS grains originally formed in the outflows of post-main-sequence asymptotic giant branch stars and were subsequently amorphized by extensive radiation processing in the ISM and further processed in the molecular cloud prior to incorporation into CP IDP parent bodies. Laboratory irradiation experiments produce silicate glasses with embedded nanocrystalline metal from mineral standards (Bradley et al. 1996). GEMS share physical characteristics in common with interstellar amorphous silicates (as deduced by astronomical observations): they exhibit infrared spectral signatures that closely match that of interstellar amorphous silicates (Bradley et al. 1999) and their sizes typically fall within the range inferred for interstellar amorphous silicate grains from studies of interstellar extinction curves (100–500 nm in diameter, e.g., Kim et al. 1994). Furthermore, it has been proposed that the nanometer-sized inclusions of superparamagnetic FeNi metal in GEMS would explain the observed polarization of starlight caused by the magnetic alignment of interstellar silicates to the galactic magnetic field (Jones & Spitzer 1967; Martin 1995). GEMS are also reported to exhibit average bulk compositions that are approximately chondritic (within 3σ of solar), as might be expected for a sample of the (approximately homogenous) presolar molecular cloud (e.g., Bradley et al. 1989). These observations suggest that GEMS grains in CP IDPs may originate from a different source environment than the crystalline components in CP IDPs that likely condensed in the solar nebula.

Irrespective of their presolar or solar nebula origins, the amorphous structure of GEMS and the crystalline structures of the silicates and sulfides in CP IDPs clearly demonstrate that GEMS did not form in the same location, under the same conditions, and at the same time as their accompanying crystalline silicates and sulfides. Consequently, it is possible that GEMS were transported to the region of cometary accretion by a different mechanism or at a different time than the crystalline silicates and sulfides, thereby accounting for the different sorting behavior. Alternatively, they may have formed under conditions that produced specific grain sizes in different locations, with those distributions then being somehow transferred to different locations of comet accretion and again resulting in a different sorting behavior than their accompanying silicates and sulfides. For example, different comets may have accreted different populations of GEMS from different regions of the collapsing molecular cloud, while their silicates and sulfides were transported from inner nebula regions and sorted together, either prior to, during, or after transport.

Our data from cometary IDPs and comet 81P/Wild2 dust indicate that aerodynamic-type sorting mechanism(s) operated in the outer nebula on silicate, sulfide, and likely metal grains, resulting in size–density equivalence between these grains. The amorphous silicates (GEMS) in CP IDPs show a different sorting behavior than their accompanying crystalline silicates and sulfides. This may be due to a different sorting mechanism having operated on them or, alternatively, could reflect an origin from different source environments. Our understanding of the evolution and transport of materials in the early nebula is derived primarily from computational modeling, yet current models are unable to explain these observations, suggesting our fundamental knowledge of the processes operating and the conditions existing in the outer nebula is lacking. The ISM and early nebula were dominated by micron and submicron grains like those studied here, yet the computational expense of incorporating such small size grains in solar system scale models is too high to be feasible. This has led to their neglect in many models and highlights the potential value to this field of the physical study of CP IDPs that sample and preserve a record of the solar system's earliest history. Further studies of these samples, together with spatially resolved astronomical observations of disks around other stars that reveal details such as gas density throughout the disks, should provide a greater understanding of the observed sorting in cometary samples and the relation to the origins and transport of small grains around young stars.

This work was funded in part by a grant from NASA's Cosmochemistry program (J.P.B.) and LDRD grant 09-ERI-004 (J.P.B.). H. A. Ishii was supported by NASA's Laboratory Analysis of Returned Samples program. Portions of this work were performed under the auspices of the U.S. Department of Energy by LLNL under contract DE-AC52–07NA27344.