Evidence for accretion of fine-grained rims in a turbulent nebula for CM Murchison

Earth and Planetary Science Letters 481 (2018) 201–211

Contents lists available at ScienceDirect

Earth and Planetary Science Letters www.elsevier.com/locate/epsl

Evidence for accretion of fine-grained rims in a turbulent nebula for CM Murchison Romy D. Hanna ∗ , Richard A. Ketcham Jackson School of Geosciences, University of Texas, Austin, TX 78712, USA

a r t i c l e

i n f o

Article history: Received 15 May 2017 Received in revised form 11 October 2017 Accepted 14 October 2017 Available online xxxx Editor: W.B. McKinnon Keywords: X-ray CT chondrules fine-grained rims accretionary rims solar nebula nebular turbulence

a b s t r a c t We use X-ray computed tomography (XCT) to examine the 3D morphology and spatial relationship of fine-grained rims (FGRs) of Type I chondrules in the CM carbonaceous chondrite Murchison to investigate the formation setting (nebular vs. parent body) of the FGRs. We quantify the sizes, shapes, and orientations of the chondrules and FGRs and develop a new algorithm to examine the 3D variation of FGR thickness around each chondrule. We find that the average proportion of chondrule volume contained in the rim for Murchison chondrules is 35.9%. The FGR volume in relation to the interior chondrule radius is well described by a power law function as proposed for accretion of FGRs in a weakly turbulent nebula by Cuzzi (2004). The power law exponent indicates that the rimmed chondrules behaved as Stokes number St η > 1 nebular particles in Kolmogorov η scale turbulence. FGR composition as inferred from XCT number appears essentially uniform across interior chondrule types and compositions, making formation by chondrule alteration unlikely. We determine that the FGRs were compressed by the impact event(s) that deformed Murchison (Hanna et al., 2015), resulting in rims that are thicker in the plane of foliation but that still preserve their nebular morphological signature. Finally, we propose that the irregular shape of some chondrules in Murchison is a primary feature resulting from chondrule formation and that chondrules with a high degree of surface roughness accreted a relatively larger amount of nebular dust compared to smoother chondrules. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Fine-grained rims (FGRs) are dust-sized material that surrounds chondrules and refractory inclusions in chondritic meteorites and are usually distinguishable from the matrix in optical and scanning electron microscopy images on the basis of their differing texture and/or composition (Lauretta et al., 2006). Among the different chondrite groups FGR characteristics differ, most notably in that the carbonaceous chondrites have significantly thicker rims compared to the ordinary chondrites and contain hydrated phases (e.g., serpentines, smectites) (Metzler and Bischoff, 1996; Zolensky et al., 1993). Among the carbonaceous chondrites CMs contain the thickest, most visible FGRs and have been the most extensively studied (e.g., Metzler and Bischoff, 1996; Metzler et al., 1992; Trigo-Rodriguez et al., 2006). In some CM chondrites, FGRs surround all macroscopic components and dominate the proportion of fine-grained material in the rock with only minor interstitial matrix; chondrites with this texture have been termed “primary accretionary rock” (Metzler et al., 1992).

*

Corresponding author. E-mail address: [email protected] (R.D. Hanna).

https://doi.org/10.1016/j.epsl.2017.10.029 0012-821X/© 2017 Elsevier B.V. All rights reserved.

The origin of FGRs in CMs is debated and formation in both nebular (Brearley et al., 1999; Bunch and Chang, 1984; Greshake et al., 2005; Hua et al., 2002; Metzler et al., 1992; Zega and Buseck, 2003) and parent body (Sears et al., 1993; Takayama and Tomeoka, 2012; Trigo-Rodriguez et al., 2006) settings have been proposed (Fig. 1). Nebular FGR formation involves accretion of unequilibrated nebular dust to chondrule surfaces (e.g., Metzler et al., 1992). Mechanisms of FGR formation in a parent body setting include: 1) formation in the parent body regolith through attachment and/or compaction of fine-grained material onto the chondrule surface (Sears et al., 1993; Takayama and Tomeoka, 2012; Trigo-Rodriguez et al., 2006) and 2) formation via aqueous alteration of the chondrule (Sears et al., 1993; Takayama and Tomeoka, 2012). Evidence cited for nebular formation includes hydrated phases in the rim in direct contact with chondrule glass (Metzler et al., 1992) and sedimentary rim textures such as layering and grain-size coarsening (Brearley et al., 1999). Parent body formation evidence includes embayment textures along the chondrule/rim boundary indicating that areas of the rim formed through alteration of the chondrule (Takayama and Tomeoka, 2012) and low rim porosity compared to the matrix which may have arisen due

202

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

Fig. 1. Three proposed formation scenarios for FGRs in CM chondrites.

to preferential compaction of material next to the chondrules (Trigo-Rodriguez et al., 2006). Another observation taken as evidence for nebular formation of FGRs in CMs is a positive correlation between rim thickness and the size of the interior chondrule (Greshake et al., 2005; Metzler et al., 1992). Conversely, Trigo-Rodriguez et al. (2006) did not find strong evidence for a such a correlation and argued that this was evidence against nebular formation and instead proposed parent body formation. All of these studies used uncorrected 2D thin section measurements, which give only apparent chondrule size and rim thickness unless the appropriate measurement corrections are applied (e.g., Sahagian and Proussevitch, 1998). Another complicating factor is that many CMs show evidence of deformation that has altered the shape of the chondrules and their rims (Hanna et al., 2015; Lindgren et al., 2015; Rubin, 2012; Zolensky et al., 1997). Although astrophysical arguments have been made for the positive linear correlation between rim thickness and interior chondrule radius (Morfill et al., 1998), subsequent refinements hypothesize that dust accretion around chondrules results in a power law relationship between the FGR volume and interior chondrule radius that in turn provides information on nebular turbulence conditions (Cuzzi, 2004; Cuzzi and Hogan, 2003). Accurate measurement of FGR thickness and volume requires three-dimensional (3D) data of the chondrules and their rims. Xray computed tomography (XCT) allows non-destructive measurement of textures and components in situ and preserves their 3D spatial context (e.g., Hanna and Ketcham, 2017; Ketcham and Carlson, 2001). This is particularly important for CM chondrites that have undergone deformation resulting in a petrofabric (e.g., Rubin, 2012). Previous XCT studies have found a moderately strong foliation and a weak lineation in Murchison due to impact(s) (Hanna et al., 2015; Lindgren et al., 2015), which also resulted in aqueous alteration of the chondrules (Hanna et al., 2015). If CM FGRs are produced via aqueous alteration on the parent body, we may see evidence for it in the form of greater FGR volume with increased degree of chondrule alteration. This study examines the 3D morphology of FGRs in a sample of Murchison for which we have previously documented impactinduced deformation and aqueous alteration (Hanna et al., 2015). Using 3D measurement of the FGRs, we look for evidence of FGR deformation that may have occurred during the impact(s) such as compaction of the rims. In addition to documenting any postaccretional deformation, we determine if there is a positive correlation between the thickness or volume of the rim and the size of the interior chondrule as predicted for nebular FGR formation and whether the nebular environment was laminar or turbulent based on FGR/chondrule size relationships (Cuzzi, 2004; Morfill et al., 1998). We also look for evidence for parent body FGR formation via aqueous alteration that may have occurred after the chondrules accreted to the CM parent body.

2. Analytical methods 2.1. X-ray computed tomography We imaged six small chips (0.143–3.63 g; some embedded in epoxy) from the Murchison USNM 5487 sample that was previously measured by Hanna et al. (2015) with X-ray computed tomography (XCT) at the University of Texas High-Resolution X-ray Computed Tomography Facility (UTCT) (Table 1). XCT produces a 3D data set of the X-ray attenuation of an object; values are referred to as XCT numbers and are mapped to grayscale values for viewing. Because X-ray attenuation is a function of a material’s atomic number and density, the relative XCT numbers are a proxy for different compositions within the sample. In a typical XCT image, the lowest XCT numbers (darkest grayscales) represent the least X-ray attenuating materials while relatively higher XCT numbers (brighter grayscales) are materials that are relatively more attenuating. Samples were imaged on two XCT systems at varying data resolutions (5.5 to 9.9 μm) to provide a statistically significant number of measurable chondrules at a range of sizes (Table 1). Relatively low energies (70–90 kV) were used to enhance the contrast between iron-bearing phases and helped to more clearly distinguish the rims that are lower in iron content than the surrounding matrix (Section 3.1). The first sample is an unlabeled chip of the USNM 5487 main mass scanned twice on an Xradia (now Zeiss) microXCT 400 system using a 0.35 mm SiO2 filter to reduce beamhardening artifacts. For the first scan (Chip A) the chip was imaged in two parts and the image stacks combined. The final reconstructed voxel size was 5.50 μm. The second scan (Chip A_HR) targeted a subvolume of the sample at higher resolution (3.00 μm) to investigate the different object types and their accompanying FGRs in more detail. To measure additional larger chondrules (≥300 μm in radius), two more individual scans of five larger-mass samples were done on the North Star Imaging (NSI) scanner, using a FeinFocus FXE X-ray source with no filter and a Perkin Elmer 16” 2048 × 2048 detector. The first NSI scan imaged samples 5487-3-3 and 5487-3-6 with a voxel size of 9.49 μm (Chips 3-3 and 3-6) and the second imaged 5487-1, 5487-3-1, and 5487-3-5 at 9.99 μm resolution (Chips 1, 3-1, and 3-5). 2.2. XCT measurement of chondrules and FGRs Chondrules within the six XCT datasets were manually segmented using AvizoTM software. Each chondrule was segmented twice, with and without FGR. Specifically, we used the brush tool to outline the exterior of the chondrule and then used the fill command to complete the segmentation. This was done for every other XCT slice across the chondrule in one orthogonal plane (segmentation plane varied across chondrules to minimize systematic measurement bias) and then the interpolate tool was used

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

203

Table 1 Physical characteristics and scanning parameters of six chips of USNM 5487 examined for this study. Data name Chip Chip Chip Chip Chip Chip Chip a b c d

Ab A_HRd 1 3-1 3-3 3-5 3-6

USNM sample (#)

Weight (g)

Embedded in epoxy?

Location in original USNM 5487 main massa

UTCT system used

Energy (kV)

Scan time (hr)

Voxel size (μm)

5487c – 5487-1 5487-3-1 5487-3-3 5487-3-5 5487-3-6

0.143 – 3.63 0.87 1.22 1.32 0.77

No – Yes Yes Yes Yes Yes

Sub-Volume – Sub-Volume Sub-Volume Sub-Volume Sub-Volume Sub-Volume

Xradia Xradia NSI NSI NSI NSI NSI

70 70 90 90 70 90 70

6.08 11.55 2.50 2.50 3.03 2.50 3.03

5.50 3.00 9.99 9.99 9.49 9.99 9.49

#2 #1 #2 #2 #2 #2

See Hanna et al. (2015), Fig. 1. Scanned in two parts. Unlabeled portion of main mass. Higher-resolution subvolume scan of Chip A.

to fill in the remaining slices. After segmentation, smoothing was applied once in each of the other two orthogonal planes to minimize jagged edges that resulted from manual segmentation and interpolation in a single plane. Only intact (non-fragmented) chondrules with clearly defined rims were segmented and we avoided any chondrules that were on the sample edge, crossed by a visible fracture, or impinged by another chondrule. Segmentation data was exported into the Blob3D software (Ketcham, 2005a; Ketcham, 2005b) and chondrule size and orientation measured with (whole chondrule) and without (interior chondrule) the FGR. Orientations were derived from best-fit ellipsoids to the whole and interior chondrule exteriors and plotted on stereonets using Stereo32 (Roeller and Trepmann, 2010). This software also calculated the strength of fabrics defined by the chondrule orientations using the C parameter, with lower values representing weaker fabrics and relatively higher values indicating stronger fabrics. Further details on this fabric strength parameter can be found in Hanna et al. (2015). Reported aspect ratios are the ratios of the long axis to the short axis of the fitted ellipsoids. The reported whole and interior chondrule radii are spherical volume equivalent radii using the respective XCT measured chondrule volumes. The FGR volume was determined by subtraction of the volumes measured with and without the FGR. To determine the average FGR thickness for each chondrule the interior chondrule radius was subtracted from the whole chondrule radius. The measurement of the actual 3D FGR thickness around each chondrule is described in Section 2.4. A summary of our estimated measurement errors for all reported parameters are listed in Table S1 and described in detail in Supplementary Note 1. 2.3. Shape measurement To quantify chondrule shape irregularity (departure from an idealized sphere or ellipsoid) we used three indices (Fig. S1). The first is the Convolution Index (CVI), which is the ratio of the chondrule surface area to the surface area of an equivalent volume sphere (Hertz et al., 2003; Zanda et al., 2002).

CVI =

surface areachondrule surface areasphere

(1)

This is the inverse of the particle shape measurement sphericity (Blott and Pye, 2008). A CVI value of one indicates that the chondrule is a perfect sphere and higher values indicate higher degrees of departure in shape from a sphere. One ambiguity with this index is that it does not distinguish between a smooth ellipsoid and a spherical but rough surface, as both will increase the surface area of the chondrule relative to an equivalent volume sphere. Because our previous work had shown that Murchison chondrules are deformed into ellipsoids (Hanna et al., 2015), we deconvolved the

CVI into two new indices which quantify ellipticity and roughness separately. The ellipticity index is defined as

Ellipticity =

surface areabest-fit ellipsoid surface areasphere

(2)

where the best-fit ellipsoid is that found by Blob3D (Section 2.2) normalized in size to the volume of the chondrule. An ellipticity value of 1 indicates that the best-fit ellipsoid is a perfect sphere and higher values indicate a more ellipsoidal shape. The roughness index is defined as

Roughness =

surface areachondrule surface areabest-fit ellipsoid

(3)

where the best-fit ellipsoid is the same as in Equation (2). A value of 1 indicates that the chondrule has a smooth surface and higher values indicate increasing degrees of surface roughness. 2.4. 3D FGR thickness measurement To determine the variable 3D thickness of the FGR around each chondrule we implemented a new algorithm in Blob3D. For each rimmed chondrule, the algorithm “looks outward” from the chondrule center along a set of 3D traverses evenly covering a unit sphere and measures the thickness of the rim along each vector. The chondrule center is defined as the center (average vertex) of the surface (set of vertices) that defines the outer extent of the rimmed chondrule. Because FGR thickness was found to vary systematically with chondrule size, FGR thickness measurements were normalized by the maximum diameter of the chondrule and are reported as FGR percent thickness. To provide a more random sampling of vectors for FGR measurement that sufficiently covers 3D space, several sets of uniform vectors were generated using the method of Ketcham and Ryan (2004); each set was randomly rotated and then collated to create a list of 2065 vectors. Each chondrule FGR shell was measured with this set of vectors and then the average percent FGR thickness among all chondrules was calculated for each vector. These data were inspected for spatial coherence using stereoplots. 3. Results 3.1. General observations The XCT data of all samples are similar in appearance and highlight differences in X-ray attenuation primarily due to the presence of iron. An XCT slice of the Chip A sample bears a striking resemblance to a low-resolution backscattered electron (BSE) image (Fig. 2), which is unsurprising as both XCT and BSE imaging are

204

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

Fig. 2. XCT slice 529 along XZ axis of Chip A. Grayscale (e.g., X-ray attenuation) variations are primarily due to differing amounts of iron. Dark objects are Type I chondrules dominated by Mg olivine and pyroxene and brighter objects are Type II chondrules (Fe-bearing olivine), Fe-bearing olivine crystal fragments, Fe serpentine, and Fe sulfides. FGRs appear as a moderate grayscale rim around most objects. The smaller, more dispersed bright phases are clumps of tochilinite, cronstedtite, and serpentine (generally absent in the rims). White and black arrows point out three examples of prominent FGRs.

sensitive to elemental composition (Hanna and Ketcham, 2017; Reed, 2005). We identify phases based on our previous work utilizing optical and electron microscopy to support XCT examination of Murchison (Hanna et al., 2015). The darkest, or least attenuating objects, are Type I chondrules dominated by pure Mg silicates (olivine and pyroxene). The largest bright, highly attenuating objects are Type II chondrules (dominated by Fe-bearing olivine), Fe olivine crystal fragments, and Fe sulfide. The smaller, more dispersed bright phases are clumps of tochilinite, cronstedtite, and serpentine [TCS, sometimes referred to as PCP, or poorly/partly characterized phases, Fuchs et al., 1973; Nakamura and Nakamuta, 1996]. FGRs surrounding chondrules and other objects appear as a medium grayscale and are primarily distinguished by the absence of the bright TCS clumps that are common in the surrounding matrix (Fig. 2; see also Figs. 3 and 5). The absence of TCS clumps in FGRs has been noted previously for other CMs (Trigo-Rodriguez et al., 2006). Many of the discrete objects within the XCT data have distinguishable FGRs (Fig. 3). The first (Fig. 3A) is an example of a Type II chondrule with an FGR located on the edge of Chip 3-3. The olivines are clearly zoned and a thick FGR is distinguishable around the entire portion of the chondrule that is present. A likely rimmed Ca–Al-rich refractory inclusion (CAI) contains pores and has a highly irregular interior boundary in contrast to the smoother exterior of its FGR (Fig. 3B). An object approximately 400 μm in diameter with minimal internal structure and an angular shape is likely a lithic clast (Fig. 3C). We searched for but did not find any other apparent lithic clasts. The object in Fig. 3D is an unknown object with a very diffuse and irregular boundary and large interior pores and cracks. It is unclear if the lighter-toned rim around the object is a true FGR or is part of the object itself (i.e., an alteration rind of some kind). We found two other smaller objects with a similar appearance. We also actively searched for clasts of FGR material with no interior object, as have been reported in other CMs (Trigo-Rodriguez et al., 2006). It is possible that the unknown objects like those

in Fig. 3C are clasts of this type, but we do not favor this interpretation because they clearly show a rim. When looking through the XCT slices it is very common to see what looks like an isolated clast of FGR material with no interior object but nearby slices confirm that this is merely the edge an FGR-enclosed object (Fig. 4E–F). We also searched in all datasets for groups of chondrules or other objects that are enclosed by a single FGR but did not find any. Although many objects were found to have FGRs only Type I chondrules were measured for this study. Testing formation theories based on chondrule and FGR sizes requires a statistically significant number of objects and Type I chondrules are abundant in Murchison. In addition, it is hypothesized that other object types, such as refractory inclusions, have different rimming properties in a nebular environment due to their larger size and/or different densities (Cuzzi, 2004). A total of 61 Type I chondrules among the 6 samples were segmented and measured. 3.2. Chondrule and FGR sizes and shapes Table 2 summarizes the XCT measurements of 61 chondrules including the range of chondrule sizes and FGR properties (full list of measurements can be found in the Supplementary Material Table S18). Measured chondrule sizes range from 176.2 to 1503.6 μm in equivalent spherical diameter. Average FGR thickness among the chondrules ranged from 27.8 to 102.5 μm. The best-fit line to the interior chondrule radius and the average FGR thickness (Fig. 4A), including uncertainty in FGR thickness (4.1%; Table S1), has a slope of 0.11 ± 0.004 (2-sigma uncertainty) and a moderately strong correlation (R 2 = 0.73). This is consistent with Metzler et al. (1992) who also reported a positive correlation between chondrule size and FGR thickness and which has been argued as evidence for nebular accretion of the rims (Metzler et al., 1992; Morfill et al., 1998). Modeling of nebular dust accumulation in magnetohydrodynamic (MHD) turbulence by Carballido (2011) also resulted in a linear relationship of FGR thickness to chondrule

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

205

the chondrule roughness than the chondrule ellipticity, to which it is only weakly correlated (Figs. 4E–F). Similar results are obtained when examining the relationships between the CVI, ellipticity, and roughness of the whole chondrules, although the CVI shows a slightly stronger correlation with ellipticity (R 2 = 0.39) and a relatively weaker correlation (compared to the interior chondrules) with roughness (R 2 = 0.69) (not shown). When comparing the shapes of the interior chondrule to that of the whole chondrule (including FGR), the interior chondrule usually has a larger aspect ratio (Fig. 4G), higher ellipticity (Fig. 4H), and greater roughness (Fig. 4I; see also Tables 2 and S18). The fact that the interior chondrule aspect ratio and ellipticity are consistently, slightly larger than and also highly correlated (R 2 = 0.90 for both) to that of the whole chondrule (Figs. 4G–H) is most likely a consequence of the interior chondrule plus outer rim structure (Supplementary Note 2). The surface roughness of the interior chondrule can be significantly larger than the whole chondrule surface roughness and in general the interior chondrules display a much wider range in roughness (1.13–1.51) compared to the whole chondrules (1.13–1.26) (Fig. 4I; Table 2). Finally, interior chondrule roughness does not show any correlation with chondrule size (interior chondrule volume or relative FGR volume; Figs. 4J–K). Examples of chondrules with relatively high and low surface roughness are shown in Fig. 5. The first chondrule has a relatively high roughness (1.47; Figs. 5A and 5B) that is evident in its segmented shape while the exterior (with FGR) is much smoother with a lower roughness (1.22). In contrast, the other interior chondrule is much smoother and its roughness index correspondingly lower (1.17; Figs. 5C and 5D). The whole chondrule roughness (1.19) of this chondrule is similar to the other chondrule (Fig. 5A), and within measurement uncertainty of its interior roughness (Table S1). 3.3. Chondrule and FGR thickness orientations Fig. 3. Various rimmed objects in the XCT data. (A) Portion of Type II chondrule in Chip 3-3. XZ slice 609. Note zoned olivines. (B) CAI with FGR in chip 3-6. XZ slice 894. The CAI has interior pores and a highly irregular boundary that is not reflected in the outer boundary of the FGR, which is much more smooth. (C) Presumed lithic clast in Chip A_HR. YZ slice 239. FGR is difficult to discern in places but completely encases object. (D) Unidentified object that appears to have a faint, but discernable FGR in Chip A_HR. YZ slice 443. (E) In YZ slice 630 of Chip A_HR, the edge of an FGR appears as if it is an isolated clast of FGR material with no interior object. (F) However, YZ slice 665 reveals that the FGR does indeed enclose a Type I chondrule.

radius. However, Cuzzi (2004) suggested that nebular formation of FGRs may manifest more precisely as a power law relationship between the interior chondrule radius and the FGR volume. To determine this relationship while including the uncertainty in FGR volume (4.1%; Table S1) we used the Levenberg–Marquardt method to perform a non-linear least squares fit to the data (Press et al., 2007) and found a much stronger fit (R 2 = 0.98; Fig. 4B) with a power law exponent of 2.50 ± 0.02 (2-sigma uncertainty). This power law dependence of rim volume on chondrule size results in a greater proportion of the chondrule volume contained in the rim for smaller chondrules (Fig. 4C). To avoid the bias of smaller chondrules we divided the sum of FGR volumes by the sum of whole chondrule volumes to find an average proportion of chondrule volume contained in the rim for Murchison chondrules of 35.9%. For chondrule shape, we first examined the relationships of the CVI, ellipticity, and roughness indices for the interior chondrules. We find that ellipticity and roughness are uncorrelated and thus well separated indices that are independently quantifying the ellipticity and surface roughness of the chondrules (Fig. 4D). The CVI measurement (which is a mixture of both the ellipticity and surface roughness indices) is more strongly influenced by

Chondrule orientations display a foliation fabric with primary (long) axes along a great circle girdle and tertiary (short) axes in a cluster (e.g., Turner and Weiss, 1963) (Fig. 6). This fabric, as well as its orientation, matches that of the foliation fabric previously found in this Murchison sample (Hanna et al., 2015). The chondrules also show evidence of the weak lineation (clustering of primary axes) found previously in Murchison (Hanna et al., 2015). The measured fabric in this study is weaker (C = 1.43–2.13) compared to the earlier study (C = 2.17–2.49) but is likely the result of fewer data points (n = 61 compared to n = 187). Additionally, the strength of the fabric defined by the interior chondrules (C = 1.43–1.48) is weaker than that of the whole chondrules when including the FGR (C = 1.69–2.13), probably due to the irregular shape of the interior chondrules (Figs. 4–5). FGR thickness around the chondrules also shows a consistent spatial relationship with the foliation plane orientation in Murchison (Figs. 6D–F). Figs. 6D–E display the average FGR percent thickness among all 61 chondrules in both the upper and lower projections which show that the variability of FGR thickness is nearly symmetrical around the chondrules and that the FGRs are thicker by up to 3.8% in a planar distribution (i.e., great circle girdle). Further, the orientation of this plane approximately matches that of the foliation fabric defined by the deformed chondrules (Figs. 6A–C). Because the variability of FGR thickness around the chondrules is nearly symmetric, we combined the upper and lower projection plots into a single lower projection plot to reduce noise and enable easier comparison to the lower projection plot of the foliated chondrule orientations (Fig. 6F). For each directional measurement, we averaged both the upper and lower projection FGR percent thickness for each chondrule. The average of these measurements among all chondrules for each direction were then plot-

206

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

Fig. 4. Size and shape relationships between interior chondrules and FGRs. (A) Average FGR thickness and interior chondrule radius are moderately positively linearly correlated. Interior chondrule radius measurement error (1.3%) is about the size of the symbols. (B) FGR volume and interior chondrule radius are very well fit using a power law relationship with exponent 2.50. Data is less well fit with a simple linear regression (R 2 = 0.89). Interior chondrule radius measurement error (1.3%) is about the size of the symbols. (C) A power law fit confirms a trend of greater proportion of chondrule volume contained in the rim for smaller chondrules. Whole chondrule volume measurement error (1.7%) is about the size of the symbols. (D) Interior chondrule roughness and ellipticity are uncorrelated, indicating that they are well-separated indices. Measurement error of ellipticity is about the size of the symbols. (E) There is a very weak linear correlation between the interior chondrule CVI and ellipticity, indicating that the CVI measurement is not strongly influenced by the ellipticity of the chondrule. (F) In contrast, there is a strong linear correlation between the interior chondrule CVI and roughness, indicating that the CVI is strongly influenced by the surface roughness of the particle. (G) Interior and whole chondrule aspect ratios are strongly linearly correlated and the interior chondrule aspect ratio is consistently, slightly larger than that of the whole chondrule. Measurement errors of both are about the size of the symbols. Dashed line is 1:1 line. (H) Interior chondrules tend to be slightly more elliptical in shape than whole chondrules. Measurement errors of both are about the size of the symbols. Dashed line is 1:1 line. (I) The surfaces of interior chondrules are usually rougher than the whole chondrule (rim) surface and display a much wider range in surface roughness (1.13–1.51) compared to the whole chondrules (1.13–1.26). Dashed line is 1:1 line. (J) There is no correlation between interior chondrule surface roughness and volume. Measurement error of interior chondrule volume is about the size of the symbols. (K) There is no correlation between surface roughness and the FGR volume, suggesting that a significant portion of FGR volume was not formed via alteration of the chondrule.

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

Fig. 5. Examples of chondrules with high (A–B) and low (C–D) roughness. (A) Segmented FGR (purple) and interior chondrule (blue) bisected by XCT YZ slice 181. The interior chondrule has a higher roughness (1.47) than the whole chondrule (1.22). (B) XCT YZ slice 181 only showing irregular 2D expression of interior chondrule compared to exterior FGR. (C) Segmented FGR (purple) and interior chondrule (blue) bisected by XCT XY slice 198. The interior chondrule has a slightly lower roughness (1.17) than the whole chondrule (1.19). This interior chondrule has a much smoother surface than the chondrule in Fig. 5A–B which is reflected in its lower roughness value. (D) XCT XY slice 198 only showing smoother 2D expression of interior chondrule and whole chondrule compared to chondrule in Fig. 5A–B. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

ted (Fig. 6F). This more clearly shows that the FGRs are consistently (up to 2.8% of maximum chondrule length) thicker in the plane of foliation defined by the deformed chondrules. While 2.8% represents a very small difference in actual FGR thickness and is below our estimated measurement error of 4.1% [Table S1; although the actual error is likely lower (Supplementary Note 1)], their spatial consistency and correspondence to the deformed chondrule foliation orientation give us confidence that the measured differences are real. Overall, this result indicates that the FGRs were already formed at the time of the impact-induced chondrule deformation and aqueous alteration of Murchison (Hanna et al., 2015). 4. Discussion 4.1. Nebular formation of FGRs These results strongly support a nebular formation environment for FGRs around Murchison chondrules. First, the FGR volume and interior chondrule radius data are well fit by a power law as suggested by Cuzzi (2004) (Fig. 4B). Neither of the parent body formation scenarios proposed to date (regolith compaction and aqueous alteration) predict any such relationship. Additionally, we find strong evidence that FGRs were deformed during the impact(s) that foliated the chondrules, resulting in thicker rims in the plane of the foliated chondrules (Fig. 6). This requires that the FGRs were already formed prior to the deformation (impact) event on the parent body. While this does not discount formation of the FGRs on the parent body (i.e., they could have formed prior to the deformation), it is more consistent with a pre-accretionary (nebular) origin because the deformational event itself resulted in both regolith gardening (via impact) and aqueous alteration (Hanna et al., 2015). If either of these two processes were responsible for creating FGRs, this latest, significant impact event would certainly have formed a substantial portion of the rims. However, we see no evidence for this as the power law relationship between FGR volume and interior chondrule radius (which can only be attributed

207

to a nebular process) has been well preserved. This suggests that regolith gardening and aqueous alteration are not significant contributors to the formation of FGRs. There is additional specific evidence against a parent body aqueous alteration origin for FGRs. If they are formed through alteration of chondrules, it is expected that we would see examples of completely altered chondrules that result in FGR ‘clasts’. We searched for such objects but did not find them. In addition, aqueous alteration of chondrules would likely result in an FGR composition that is dependent on the initial composition of the chondrule. However, when we compared average FGR XCT numbers of various object types (Type I, Type II, and the lithic clast in Fig. 3C) we found no significant differences correlated with object type. We also found no correlation between the shape irregularity of the interior chondrule and the relative FGR volume (Fig. 4K) which might be expected if increasing aqueous alteration led to both a more irregular chondrule shape and increased FGR volume. While it is possible that aqueous alteration can proceed equally along the entire chondrule surface and therefore not produce an irregular inner chondrule shape, we do not favor this interpretation because 1) Type I chondrules in Murchison are heterogeneous in composition and texture (e.g., Fuchs et al., 1973; Hanna et al., 2015) which would be expected to result in uneven alteration and 2) in other CMs where aqueous alteration has taken place along the chondrule exterior, irregular degrees of alteration and embayments are seen (Takayama and Tomeoka, 2012). We do not preclude the possibility that some aqueous alteration might have increased the FGR volume, but the proportion of FGR volume created this way must have been limited enough that it did not significantly affect the FGR volume/chondrule radius power law relationship caused by nebular formation (Fig. 4B). 4.2. Nebular conditions during FGR formation The power law relationship between rim volume and interior chondrule radius proposed by Cuzzi (2004) is for the formation of FGRs around chondrule to CAI-sized particles in a laminar to weakly turbulent nebula. He showed that this nonlinear relationship is a direct consequence of the particle-size dependent gasrelative velocity of the particle as it moves through the gas (and the dust that is coupled to it) (Cuzzi, 2004; Cuzzi and Hogan, 2003). This size-dependent velocity leads to a size-dependent accumulation of dust around the chondrule as larger chondrules generally encounter a relatively higher amount of dust in the same amount of time. In a laminar flow the relative velocity of chondrules to gas is linearly proportional to particle radius, whereas in the presence of turbulence this relationship is modified such that the relative velocity approaches proportionality to the square root of the radius (Cuzzi, 2004). In other words, larger particles move slower and therefore encounter relatively less dust than in a laminar flow. Cuzzi (2004) derived an equation (Eq. (11) in that work) to express the volume of the rim produced as a function of the interior chondrule radius, the degree of turbulence, and the Stokes number St of the particles. The latter two variables are encapsulated in the variable p. The Stokes number St describes the aerodynamic behavior of a particle and is the ratio of a particle’s stopping time to the overturn time of some characteristic eddy. Stopping time refers to the timescale it takes a particle to equilibrate with changes in nebular gas flow – smaller particles are better coupled to the gas and equilibrate more quickly (e.g., Cuzzi and Weidenschilling, 2006). Nebular turbulence is characterized by an inertial range of length scales, from the largest (commonly regarded as the local orbital period) to the smallest, or Kolmogorov η, scale (Cuzzi and Hogan, 2003). Therefore particles with a stopping time similar to the overturn time of the smallest length scale of turbulence have a Kolmogorov stopping time and thus St η = 1.

208

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

Table 2 Summary of XCT measurements of 61 chondrules with FGRs among 6 samples.

Minimum Maximum Mean (n = 61) a b

ICa volume (mm3 )

FGR volume (mm3 )

IC radius (μm)

Average FGR thickness (μm)

IC aspect ratio

WCb aspect ratio

IC CVI

WC CVI

IC ellipticity

WC ellipticity

IC roughness

WC roughness

0.003 1.780 0.192

0.004 0.625 0.108

88.1 751.8 279.8

27.8 102.5 54.2

1.12 2.29 1.53

1.11 2.14 1.46

1.13 1.54 1.29

1.14 1.31 1.20

1.00 1.10 1.03

1.00 1.08 1.02

1.13 1.51 1.25

1.13 1.26 1.18

IC = interior chondrule. WC = whole chondrule.

Fig. 6. (A–C) Chondrule orientations using primary (long) and tertiary (short) axes of best-fit ellipsoids to chondrules with (A) and without (B) the FGR. (C) Orientation of chondrules from the whole Murchison sample from Hanna et al. (2015). Chondrules measured in this study display the same foliation fabric orientation as the whole Murchison sample from Hanna et al. (2015). Interior chondrules show weaker fabric (C parameter) that is most likely a result of their more irregular shape compared to the FGR exterior (Table 2 and Figs. 4–5. (D–F) Average FGR percent thickness of maximum chondrule length among the 61 measured chondrules in directions indicated. In these plots each dot represents a directional measurement, not a discrete chondrule. Lower (D) and upper (E) projection plots indicate that the FGRs are consistently thicker (orange) approximately in a plane. (F) Average symmetric (lower projection) FGR percent thickness of maximum chondrule length among the 61 chondrules in directions indicated. FGRs are consistently thicker (orange) in the plane of foliation. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The rim volume predictions using Equation (11) of Cuzzi (2004) for a range of p values are plotted in Fig. 7. Note that different p values represent different nebular turbulence conditions; the sizes of the particles (chondrules) themselves are not changing. As the degree (scale) of the turbulence changes so does the radiusdependence on chondrule velocity and thus the amount of dust accreted as a function of chondrule size. A p value of 1 indicates that the nebula was laminar or, if turbulent, that the Kolmogorov eddy was sized such that the nebular Kolmogorov stopping time was longer than that of the chondrules’ actual stopping time, leaving the size-dependent velocity gradient unaffected. In other

words, for the p = 1 model all chondrules were sized significantly smaller relative to Kolmogorov stopping time particles, indicating that the chondrules were St η  1 in this turbulence (Cuzzi, 2004). Lower values of p apply only to a turbulent nebula, and various p values correspond to particles ‘sized’ differently relative to the Kolmogorov stopping time particles due to different degrees of nebular turbulence. Relatively lower p values correspond to scales of turbulence that increasingly affect the size-dependent velocity of chondrules so that larger chondrules move more slowly and encounter less dust. A p value of 0.75 is a special case predicted to correspond to particles that are sized such that their stopping time

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

209

Fig. 7. Rim volume predictions for p = 0.0, 0.5, 0.7, 0.9, and 1.0 using Equation (11) from Cuzzi (2004) (thin gray lines) and the data for the Murchison samples (circles). Measurement uncertainty in rim volume (4.1%) and interior chondrule radius measurement uncertainty (1.3%) is about the size of the symbol. The thick solid line is the best fit through the Murchison data (same fit is shown Fig. 4B) and the dotted lines represent the 2-sigma uncertainties in the best fit.

is exactly the Kolmogorov stopping time (St η = 1) (Cuzzi, 2004; Cuzzi and Hogan, 2003). Particles that are sized smaller relative to the Kolmogorov-stopping-time size, and thus stopping times shorter than the Kolmogorov overturn time (St η < 1) are indicated by p values between 0.75 and 1.0, and particles sized larger (St η > 1) will have p values below 0.75 (Cuzzi, 2004; Cuzzi and Hogan, 2003). Therefore, examination of where the Murchison data fall within these predicted rim volume/chondrule radius relationships could shed light on whether the nebula was turbulent or laminar and the size of the rimmed chondrules relative to the Kolmogorov scale (i.e., the scale of turbulence in the nebula and thus the St η of the chondrules) (Fig. 7). The exact value of p that the Murchison data mostly closely corresponds to is difficult to determine visually from the graph as the best fit line and its 2-sigma uncertainties cross all lines (p = 0.0 to 1.0) at some point. Among the various p cases (Fig. 7), the slope (power law exponent) of the line is what most uniquely determines each case as the vertical axis intercept represents the absolute rim volume that depends on other parameters such as the local dust density and rimming timescale (Cuzzi, https://doi.org/10.1016/j.chemer.2017.01.006, personal communication, August 2016). Therefore, we determined a best-fit linear relationship using linear regression (R 2 = 0.998) to the p parameter versus its power law exponent from Equation (11) of Cuzzi (2004) (all other parameters set at values determined in that work)

y = 1.08p + 1.90

(4)

where y is the power law exponent. Using this relationship, the best-fit line to the Murchison data with 2-sigma uncertainty corresponds to a p value of 0.56 ± 0.02. This indicates that the nebula was turbulent, and also that the rimmed particles were larger than Kolmogorov-stopping-time particles (St η > 1). This is a different result than that found for CV Allende which had best fit of p = 0.7–0.9 (Cuzzi, 2004). However, in that work the small number of data points (n = 14) and their large scatter made it impossible to completely rule out other values of p (from 0.5 to 1.0) (Cuzzi, 2004). Therefore, the different values of p found for Murchison could represent either a real difference compared to CV Allende (which could vary among the different chondrite classes) or simply reflect the more statistically significant data set. More data will be collected in the future for other chondrite classes to determine whether similar values of p are indicated. Regardless, this result indicates that in the area where Murchison chondrules were acquiring

Fig. 8. Murchison data split into 2 groups based on their interior chondrule roughness indices. (A) Best-fit power law (solid line) with 2-sigma uncertainties (dotted lines) to group (N = 49) with smoothest inner chondrule roughness values (below mid-range value of 1.32). Vertical error bars are 4.1% measurement uncertainty on the rim volume. (B) Best-fit power law (solid line) with 2-sigma uncertainties (dotted lines) to group (N = 12) with roughness values (above mid-range value of 1.32). Vertical error bars are 4.1% measurement uncertainty on the rim volume. (C) Graph of P values of each group shown in Figs. 8A–B calculated with Equation (4) with 2-sigma uncertainties. Groups are very well separated in P space, with rougher chondrules corresponding to a higher P value. The gray bar indicates the calculated P values using the full Murchison dataset (0.54–0.58; see text).

their rims the disk was weakly turbulent and the chondrules (∼100–1500 μm in diameter) were behaving as St η > 1 particles. This provides an important physical constraint for accretion models that incorporate nebular turbulence (e.g., Armitage, 2011; Hopkins, 2016; Johansen et al., 2015; Morbidelli and Raymond, 2016). One simplifying assumption in the Cuzzi (2004) model is that the particles are spherical. However, our analyses show that the interior chondrules can be quite irregular in shape (Figs. 4–5). We quantify shape irregularity separately as ellipticity and roughness, where the ellipticity is thought to have originated during impact deformation on the CM parent body (Hanna et al., 2015). Therefore, only the surface roughness of the chondrules is relevant when considering their dust accumulation to form FGRs in the nebula. To determine whether this surface roughness influenced dust accumulation and therefore the rim volume to chondrule radius relationship (and thus the estimate of p), we split the data into 2 groups based on their interior chondrule roughness value (Fig. 8). Here we show the data split based on the mid-range roughness value of interior chondrules (1.32) although using the mean rough-

210

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

ness of the interior chondrules (1.25) instead produces the same results (Supplementary Note 3). The best-fit power law relationship of the smoother chondrule group (n = 49) has a power law exponent of 2.47 (Fig. 8A) and converting this to p using Equation (4) with the 2-sigma uncertainties results in p = 0.52 ± 0.02 (Fig. 8C). The rougher particle group (n = 12) has a significantly higher best-fit power law exponent of 2.61 (Fig. 8B), and converting this to p results in 0.66 ± 0.05 which is significantly higher compared to the smoother chondrules (Fig. 8C). According to the Cuzzi (2004) model this indicates that the relatively rougher particles are behaving like they are slightly smaller relative to Kolmogorov-stopping-time particles than the smoother particle group (both are relatively larger than Kolmogorov-stopping-time particles in general, however, as p for both are <0.75). This requires either that the two groups acquired rims in different turbulent conditions (with different Kolmogorovscale turbulence) or that the surface roughness is influencing the volume of rim that is accreted. It is possible that the FGRs formed in separate areas with different degrees of nebular turbulence and were then mixed together to accrete on the CM parent body. Some compositional and chemical variation is seen among FGRs in CMs, including Murchison (Brearley and Geiger, 1993, 1991; Metzler et al., 1992) suggesting different nebular reservoirs, but it is unknown if these chemical differences are correlated with surface roughness. We compared the average FGR XCT grayscales (which are a function of composition) to interior chondrule roughness and did not find any systematic correlations. We therefore think it more likely that the chondrule surface roughness is influencing the accretion behavior of dust in the nebula (Section 4.3). Regardless, this is further support for the formation of FGRs by accretion in a nebular setting, as such a particle-roughness dependence of the rim volume to interior radius relationship would not be expected in a parent-body setting. 4.3. Irregular chondrule shape It is often assumed that chondrules were originally spherical because they represent crystallized melt that was freely floating in the nebula. Despite the deformed shape of chondrules in Murchison, ellipticity analysis has shown that they were not perfectly spherical prior to the deformation (Hanna et al., 2015). Numerous workers have observed that CM chondrules have irregular outlines that are effectively smoothed out by the FGRs which infill embayed areas (Cuzzi, 2004; Hua et al., 2002; Kinnunen and Saikkonen, 1983; Lauretta et al., 2000; Metzler et al., 1992; Tomeoka et al., 1991; Zega and Buseck, 2003). Our shape analysis confirms these observations, indicating that the whole chondrule exteriors are much smoother (lower roughness) than their interiors (Figs. 4I and 5). The origin of irregular shapes of interior chondrules is not known, but the 2D expression of highly irregular chondrules (e.g., Fig. 5B) is reminiscent of embayment such as might occur through alteration of chondrule edges along favorable pathways. However, if irregular shape is the result of alteration (and therefore forming the FGR in the process) we would expect a positive correlation between surface roughness and FGR volume, which is not observed (Fig. 4G and Section 4.1). In addition, there is evidence that the surface roughness of some chondrules may have influenced their dust accumulation behavior relative to their size (Fig. 8 and Section 4.2). Taken together, this suggests that the irregular shape of the chondrules is a primary feature, and predates both FGR accretion in the nebula and chondrule accretion to the CM parent body. This irregular shape could be the result of multiple melting events (Rubin and Wasson, 2005) or collisions between partially molten chondrules (Alexander and Ebel, 2012).

Finally, we consider how this primary irregular chondrule shape is causing relatively rougher particles to develop an FGR volume/particle radius relationship consistent with higher values of p (Fig. 8). Higher p values indicate steeper curves in the FGR volume/particle size plot (Fig. 7); in other words, the rate of FGR volume increase per chondrule size is higher than for lower p curves. Therefore, the rougher chondrules have a larger rate of FGR volume change per chondrule size compared to smoother particles. This indicates that the process causing rougher particles to accrete more nebular dust operates more efficiently at relatively larger particle sizes (i.e., it is a size-dependent process). It might seem unexpected that rougher particles would accrete a larger amount of dust as irregular particle shape causes an increase in drag force (decrease in velocity) in turbulent flow that would in turn decrease the amount of dust encountered and accreted (Dellino et al., 2005). But the magnitude of this effect is independent of particle size (Dellino et al., 2005) and therefore would not change the power law relationship (p) of FGR volume to chondrule size. However, irregularly-shaped particles also have higher surface to volume ratios, surface roughness, and curvature which can increase their adsorption and geometrical interlocking capabilities (e.g., Vonlanthen et al., 2015). Nebular dust aggregation models also show an increase in aggregate growth with irregular particle (dust) shapes due to an increase in sticking probability (Poppe et al., 2000). Because surface area increases proportionally to the particle radius as r 2 , larger particles provide exponentially more surface area for these effects to operate. 5. Conclusions We have found strong morphological evidence that FGRs in Murchison were formed in a weakly turbulent nebula, as the relationship between FGR volume and interior chondrule interior radius is well fit by a power law relationship consistent with the presence of turbulence (Cuzzi, 2004). The power law exponent of the best-fit power law including its 2-sigma uncertainty suggests that the rimmed chondrules were behaving as St η > 1 nebular particles. In addition, we rule out parent-body aqueous alteration as an origin for significant FGR formation based on several lines of evidence: the lack of correlation between interior chondrule shape irregularity (a proxy for degree of alteration) and FGR volume, the presence of a positive correlation between rim thickness and chondrule size, and evidence that FGRs are a similar composition despite their interior chondrule types (composition). We also find that the FGRs were deformed along with the interior chondrules during the impact(s) that foliated and lineated the chondrules on the CM parent body (Hanna et al., 2015) but that despite this deformation the morphological signature of FGR nebular formation was preserved. Finally, we propose that the highly irregular shape of some Murchison chondrules is a primary feature resulting from chondrule formation and that rougher particles accreted a relatively larger amount of nebular dust. Acknowledgements We would like to thank Jeff Cuzzi for extremely helpful discussions of his work related to nebular formation of FGRs that were critical for several interpretations made in this work. We also thank Travis Clow, Megan Hoffman, and Zoe Yin (University of Texas at Austin) for assistance in AvizoTM segmentation and error determination. We thank the Smithsonian National Museum of Natural History and curator Tim McCoy for the loan of Murchison USNM 5487. Thanks to Mike Zolensky for his help in identifying unknown objects in the XCT data. We are grateful to editor William McKinnon, Josep Trigo-Rodriguez, and an anonymous reviewer for comments that improved the quality of the manuscript. R.D.H. was

R.D. Hanna, R.A. Ketcham / Earth and Planetary Science Letters 481 (2018) 201–211

supported by the NASA Earth and Space Sciences Fellowship Program – grant NNX13AO64H during this work and portions of this work were funded through an Analytical Fees grant from the Jackson School of Geosciences to R.D.H. The UTCT Facility and R.A.K. are supported by the National Science Foundation Instrument and Facilities program in the EAR directorate, grant EAR-1561622. Appendix A. Supplementary material Supplementary material related to this article can be found online at https://doi.org/10.1016/j.epsl.2017.10.029. References Alexander, C.M.O.D., Ebel, D.S., 2012. Questions, questions: can the contradictions between the petrologic, isotopic, thermodynamic, and astrophysical constraints on chondrule formation be resolved? Meteorit. Planet. Sci. 47, 1157–1175. Armitage, P.J., 2011. Dynamics of protoplanetary disks. Annu. Rev. Astron. Astrophys. 49, 195–236. Blott, S.J., Pye, K., 2008. Particle shape: a review and new methods of characterization and classification. Sedimentology 55, 31–63. Brearley, A., Geiger, T., 1993. Fine-grained chondrule rims in the Murchison CM2 chondrite: compositional and mineralogical systematics. Meteoritics 28, 328–329. Brearley, A.J., Geiger, T., 1991. Mineralogical and chemical studies bearing on the origin of accretionary rims in the Murchison CM2 carbonaceous chondrite. Meteoritics 26, 33. Brearley, A.J., Hanowski, N.P., Whalen, J.F., 1999. Fine-grained rims in CM carbonaceous chondrites: a comparison of rims in Murchison and ALH 81002. Proc. Lunar Planet. Sci. Conf., 1460. Bunch, T., Chang, S., 1984. CAI rims and CM2 dustballs: products of gas-grain interactions, mass transport, grain aggregation and accretion in the nebula? Proc. Lunar Planet. Sci. Conf., 100–101. Carballido, A., 2011. Accretion of dust by chondrules in a MHD-turbulent solar nebula. Icarus 211, 876–884. Cuzzi, J.N., 2004. Blowing in the wind: III. Accretion of dust rims by chondrule-sized particles in a turbulent protoplanetary nebula. Icarus 168, 484–497. Cuzzi, J.N., Hogan, R.C., 2003. Blowing in the wind: I. Velocities of chondrule-sized particles in a turbulent protoplanetary nebula. Icarus 164, 127–138. Cuzzi, J.N., Weidenschilling, S.J., 2006. Particle-gas dynamics and primary accretion. In: Lauretta, D., McSween, H.Y.J. (Eds.), Meteorites and the Early Solar System II. University of Arizona Press, Tuscon, pp. 353–381. Dellino, P., Mele, D., Bonasia, R., Braia, G., La Volpe, L., Sulpizio, R., 2005. The analysis of the influence of pumice shape on its terminal velocity. Geophys. Res. Lett. 32, L21306. Fuchs, L.H., Olsen, E., Jensen, K.J., 1973. Mineralogy, mineral-chemistry, and composition of the Murchison (C2) meteorite. Smithson. Contrib. Earth Sci. 10, 39. Greshake, A., Krot, A.N., Flynn, G.J., Keil, K., 2005. Fine-grained dust rims in the Tagish Lake carbonaceous chondrite: evidence for parent body alteration. Meteorit. Planet. Sci. 40, 1413–1431. Hanna, R.D., Ketcham, R.A., 2017. X-ray computed tomography of planetary materials: a primer and review of recent studies. Chem. Erde. https://doi.org/10.1016/ j.chemer.2017.01.006. Hanna, R.D., Ketcham, R.A., Zolensky, M., Behr, W., 2015. Impact-induced brittle deformation, porosity loss, and aqueous alteration in the Murchison CM chondrite. Geochim. Cosmochim. Acta 171, 256–282. Hertz, J., Ebel, D.S., Weisberg, M.K., 2003. Tomographic study of shapes and metal abundances of Renazzo chondrules. In: XXXIV Lunar and Planetary Science Conference. Houston, TX, 1959. Hopkins, P.F., 2016. Jumping the gap: the formation conditions and mass function of ‘pebble-pile’ planetesimals. Mon. Not. R. Astron. Soc. 456, 2383–2405. Hua, X., Wang, J., Buseck, P.R., 2002. Fine-grained rims in the Allan Hills 81002 and Lewis Cliff 90500 CM2 meteorites: their origin and modification. Meteorit. Planet. Sci. 37, 229–244. Johansen, A., Low, M.-M.M., Lacerda, P., Bizzarro, M., 2015. Growth of asteroids, planetary embryos, and Kuiper belt objects by chondrule accretion. Sci. Adv. 1. Ketcham, R.A., 2005a. Computational methods for quantitative analysis of threedimensional features in geological specimens. Geosphere 1, 32–41. Ketcham, R.A., 2005b. Three-dimensional grain fabric measurements using highresolution X-ray computed tomography. J. Struct. Geol. 27, 1217–1228. Ketcham, R.A., Carlson, W.D., 2001. Acquisition, optimization and interpretation of X-ray computed tomographic imagery; Applications to the geosciences. Comput. Geosci. 27, 381–400.

211

Ketcham, R.A., Ryan, T.M., 2004. Quantification and visualization of anisotropy in trabecular bone. J. Microsc. 213, 158–171. Kinnunen, K.A., Saikkonen, R., 1983. Kivesvaara C2 chondrite: Silicate petrography and chemical composition. Bull. - Geol. Surv. Finl. 55, 35–49. Lauretta, D.S., Hua, X., Buseck, P.R., 2000. Mineralogy of fine-grained rims in the ALH 81002 CM chondrite. Geochim. Cosmochim. Acta 64, 3263–3273. Lauretta, D.S., Nagahara, H., Alexander, C., 2006. Petrology and origin of ferromagnesian silicate chondrules. In: Lauretta, D., McSween Jr., H.Y. (Eds.), Meteorites and the Early Solar System II. University of Arizona Press, Tucson, pp. 431–459. Lindgren, P., Hanna, R.D., Dobson, K.J., Tomkinson, T., Lee, M.R., 2015. The paradox between low shock-stage and evidence for compaction in CM carbonaceous chondrites explained by multiple low-intensity impacts. Geochim. Cosmochim. Acta 148, 159–178. Metzler, K., Bischoff, A., 1996. Constraints on chondrite agglomeration from finegrained chondrule rims. In: Hewins, R.H., Jones, R.H., Scott, E.R.D. (Eds.), Chondrules and the Protoplanetary Disk. Cambridge University Press, Cambridge, pp. 153–162. Metzler, K., Bischoff, A., Stoeffler, D., 1992. Accretionary dust mantles in CM chondrites – evidence for solar nebula processes. Geochim. Cosmochim. Acta 56, 2873–2897. Morbidelli, A., Raymond, S.N., 2016. Challenges in planet formation. J. Geophys. Res., Planets 121, 1962–1980. Morfill, G.E., Durisen, R.H., Turner, G.W., 1998. An accretion rim constraint on chondrule formation theories. Icarus 134, 180–184. Nakamura, T., Nakamuta, Y., 1996. X-ray study of PCP from the Murchison CM carbonaceous chondrite. Antarct. Meteor. Res. 9, 37. Poppe, T., Blum, J., Henning, T., 2000. Analogous experiments on the stickiness of micron-sized preplanetary dust. Astrophys. J. 533, 454. Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P., 2007. Numerical recipes. In: The art of Scientific Computing, 3rd ed.. Cambridge University Press, New York, New York. Reed, S.J.B., 2005. Electron Microprobe Analysis and Scanning Electron Microscopy in Geology, 2nd ed.. Cambridge University Press, Cambridge. Roeller, K., Trepmann, C., 2010. Stereo32. Institut für Geologie, Ruhr University Bochum, Bochum, Germany. Rubin, A.E., 2012. Collisional facilitation of aqueous alteration of CM and CV carbonaceous chondrites. Geochim. Cosmochim. Acta 90, 181–194. Rubin, A.E., Wasson, J.T., 2005. Non-spherical lobate chondrules in CO3.0 Y-81020: general implications for the formation of low-FeO porphyritic chondrules in CO chondrites. Geochim. Cosmochim. Acta 69, 211–220. Sahagian, D.L., Proussevitch, A.A., 1998. 3D particle size distributions from 2D observations: stereology for natural applications. J. Volcanol. Geotherm. Res. 84, 173–196. Sears, D.W.G., Benoit, P.H., Jie, L., 1993. Two chondrule groups each with distinctive rims in Murchison recognized by cathodoluminescence. Meteoritics 28, 669–675. Takayama, A., Tomeoka, K., 2012. Fine-grained rims surrounding chondrules in the Tagish Lake carbonaceous chondrite: verification of their formation through parent-body processes. Geochim. Cosmochim. Acta 98, 1–18. Tomeoka, K., Hatakeyama, K., Nakamura, T., Takeda, H., 1991. Evidence for pre-accretionary aqueous alteration in the Yamato-793321 CM carbonaceous chondrite. In: 16th Symposium on Antarctic Meteorites. NIPR, Tokyo, Japan, pp. 37–39. Trigo-Rodriguez, J.M., Rubin, A.E., Wasson, J.T., 2006. Non-nebular origin of dark mantles around chondrules and inclusions in CM chondrites. Geochim. Cosmochim. Acta 70, 1271–1290. Turner, F.J., Weiss, L.E., 1963. Structural Analysis of Metamorphic Tectonites. McGraw-Hill. Vonlanthen, P., Rausch, J., Ketcham, R.A., Putlitz, B., Baumgartner, L.P., Grobéty, B., 2015. High-resolution 3D analyses of the shape and internal constituents of small volcanic ash particles: the contribution of SEM micro-computed tomography (SEM micro-CT). J. Volcanol. Geotherm. Res. 293, 1–12. Zanda, B., Bourot-Denise, M., Hewins, R.H., Cohen, B.A., Delaney, J.S., Humayun, M., Campbell, A.J., 2002. Accretion textures, iron evaporation, and re-condensation in Renazzo chondrules. In: XXXIII Lunary and Planetary Science Conference. Houston, TX, 1852. Zega, T.J., Buseck, P.R., 2003. Fine-grained-rim mineralogy of the Cold Bokkeveld CM chondrite. Geochim. Cosmochim. Acta 67, 1711–1721. Zolensky, M., Barrett, R., Browning, L., 1993. Mineralogy and composition of matrix and chondrule rims in carbonaceous chondrites. Geochim. Cosmochim. Acta 57, 3123–3148. Zolensky, M.E., Mittlefehldt, D.W., Lipschutz, M.E., Wang, M.-S., Clayton, R.N., Mayeda, T.K., Grady, M.M., Pillinger, C., Barber, D., 1997. CM chondrites exhibit the complete petrologic range from type 2 to 1. Geochim. Cosmochim. Acta 61, 5099.