Multiple melting in a four-layered barred-olivine chondrule with compositionally heterogeneous glass from LL3.0 Semarkona

Authors


E-mail: aerubin@ucla.edu

Abstract

Chondrule K7p from LL3.0 Semarkona consists of four nested barred-olivine (BO) chondrules. The innermost BO chondrule (chondrule 1) formed by complete melting of an olivine-rich dustball. After formation, the chondrule was incorporated into another olivine-rich dustball. A second heating event caused this second dustball to melt; the mesostasis and some of the olivine in chondrule 1 were probably also melted at this time, but the chondrule 1 structure remained largely intact. At this stage, the object was an enveloping compound BO chondrule. This two-step process of melting and dustball enshrouding repeated two more times. The different proportions of olivine and glass in chondrules 1–4 suggest that the individual precursor dustballs differed in the amounts of chondrule fragments they contained and the mineral proportions in those fragments. The final dustball (which ultimately formed chondrule 4) was somewhat more ferroan; after melting, crystallizing, and quenching, chondrule 4 contained olivine and glass with higher FeO and MnO contents than those of the earlier formed chondrules. Subsequent aqueous alteration on the LL parent body transformed the abundant metal blebs and stringers at the chondrule surface into carbide, iron oxide, and minor Ni-rich metal. Portions of the mesostasis underwent dissolution, producing holes and adjacent blades of more resistant material. Much of the glass in the chondrule remained isotropic, even after minor hydration and leaching. The sharp, moderately lobate boundary between the extensively altered mesostasis and the isotropic glass represents the reaction front beyond which there was little or no glass dissolution.

Introduction

There is broad consensus that the majority of chondrules were heated multiple times. Evidence includes the presence of relict grains within chondrules (e.g., Nagahara 1981; Rambaldi 1981; Jones 1996; Wasson and Rubin 2003); enveloping (i.e., nested) compound chondrules (Wasson et al. 1995); igneous rims around chondrules (Rubin 1984, 2010; Rubin and Wasson 1987; Krot and Wasson 1995); and the presence of nonspherical, multilobate chondrules (Rubin and Wasson 2005). (Had the latter objects been completely molten, they would have collapsed into spheres thousands of times faster than the time it would have taken their phenocrysts to nucleate and grow.)

Remelting temperatures appreciably above the liquidus totally melted chondrules and erased all petrographic evidence of their original textures. In cases where the peak remelting temperatures were significantly below the liquidus, only those components with the lowest melting temperatures would have been affected: mesostasis, metal-sulfide, pyroxene, and, perhaps, some olivine (e.g., Jones et al. 2005). There may have been many such low-energy remelting events (e.g., Wasson 1993). During remelting, chondrules would have lost volatiles, some of which would have condensed onto neighboring dust grains (e.g., Wasson 2008).

In between melting events, chondrules could have been chipped, fragmented, or shattered during collisions with other chondrules and inclusions in the nebula. Intact chondrules, CAIs, chondrule- and CAI-fragments, dislodged phenocrysts, and phenocryst fragments could have been incorporated in various amounts and proportions into porous dustballs (e.g., fig. 5 of Jones et al. 2005). The presence of such dustballs in the nebula is indicated by the occurrence of matrix lumps in ordinary and carbonaceous chondrites (Kurat 1970; Ikeda et al. 1981; Scott et al. 1984), coarse-grained igneous rims around chondrules (e.g., Rubin 1984; Krot and Wasson 1995), the secondary shells of enveloping compound chondrules (Wasson et al. 1995), and microchondrule-bearing fine-grained silicate-rich chondrule rims (Rubin et al. 1982; Krot and Rubin 1996; Krot et al. 1997a).

I report here the occurrence of a multilayered barred-olivine (BO) chondrule from the Semarkona LL3.0 ordinary chondrite that sheds light on multiple melting processes, the enshrouding of chondrules by porous dustballs, and subsequent parent-body aqueous alteration.

Analytical Procedures

The UCLA LEO 1430 scanning electron microscope was used to make backscattered electron (BSE) images of Semarkona thin section AMNH 4128-2, on loan from the American Museum of Natural History in New York. The separate images were combined into a mosaic using Adobe Photoshop software; a millimeter-sized grid was imposed on the mosaic and used to label individual chondrules (Fig. 1). Each millimeter-sized square was subdivided into a 5 × 5 subgrid designated by the letters a–y. Chondrule K7p occurs toward the lower right side of the mosaic image.

Figure 1.

Mosaic backscattered electron (BSE) image of Semarkona thin section AMNH 4128-2. A grid is superimposed on the image, facilitating the location of individual chondrules. Each square is 1 mm on a side and is further divided into a 5 × 5 subgrid composed of 25–200 μm-sized squares (not shown). These small squares are labeled with the letters a–y; the top row, starting at the left, is composed of the squares a–e; the following rows are f–j, k–o, p–t, and u–y. The chondrule that is the focus of the present study is K7p. (The letter p marks the approximate center of the object, located at the lower left of the K7 square.)

The chondrule was examined and photographed using an Olympus BX60 petrographic microscope. High-magnification BSE images were made with the JEOL electron microprobe at UCLA. Mineral compositions were determined with the JEOL microprobe using natural and synthetic standards, a sample current of 15 nA, an accelerating voltage of 15 keV, 20 s counting times per element, ZAF corrections, and a focused beam. Cobalt values in metallic Fe-Ni and carbide grains were corrected for the interference of the Fe-Kβ peak with the Co-Kα peak. To minimize devolatilization, mesostasis was analyzed with the JEOL probe using a sample current of 10 nA and a 3 μm-diameter beam. The modal abundances of phases in chondrule K7p were determined with Adobe Photoshop software from a BSE image of the chondrule (Fig. 2). Volume-percent was converted into weight-percent using the following densities (in g cm−3): forsterite (3.23), enstatite (3.209), glassy mesostasis (2.9), and metallic Fe-Ni (7.85). Grain size was determined from the BSE images using the scale bar.

Figure 2.

Chondrule K7p is a compound barred-olivine (BO) chondrule containing olivine bars, phenocrysts, and rims (medium gray) interspersed with mesostasis (light gray). Portions of the mesostasis at the bottom and left side of the chondrule have been aqueously altered and appear mottled in the image. Blebs of metallic Fe-Ni (white) occur in the chondrule interior. Blebs and stringers composed mainly of oxide and carbide (white) occur at the outer margin of the chondrule. The largest opaque bleb at the top of the image appears to be associated with the outer rim. Black areas are plucked regions, holes, and fractures. BSE image.

Results

Petrography

Chondrule K7p is 1090 × 1540 μm in size (Figs. 2 and 3); in plan view, it forms an ellipse with an eccentricity of approximately 0.7. Prior to parent-body alteration, the chondrule consisted of forsterite (58 vol%; 44 wt%), enstatite (2 vol%; 2 wt%), glassy mesostasis (17 vol%; 12 wt%), and metallic Fe-Ni (23 vol%; 42 wt%). In addition, small Ca-pyroxene crystallites were probably present in some mesostasis regions. The opaque phases (presumably metallic Fe-Ni prior to alteration) in the outer part of the chondrule now consist of carbide and oxide (cf. Krot et al. 1997b).

Figure 3.

Same image of chondrule K7p as Fig. 2, but divided into four enveloping (i.e., nested) BO chondrules. The lines mark the approximate inferred boundaries of the subchondrules. They are superimposed on the internal olivine shells and tend to be drawn through concentrations of opaque blebs. Chondrule 1 is the primary chondrule; it is surrounded by an olivine shell that is substantially structurally coherent. The overlying chondrules are divided into sections with the letters a on the left and b on the right. The olivine shells around chondrules 2 and 3 are thicker on the right sides than on the left. Patches of mesostasis are present below the olivine shells surrounding chondrules 1, 2, 3, and 4. BSE image.

The olivine grains in K7p range in size from 5 to 180 μm and in shape from quasi-equant to elongated (with aspect ratios ranging from 2 to 8). More of the olivine grains resemble detached bars from BO chondrules (e.g., fig. 1d of Weisberg 1987) than the quasi-equant phenocrysts common in Type-IA porphyritic-olivine chondrules (e.g., figs. 1 and 2 of Jones and Scott 1989).

When viewed microscopically in crossed-polarized transmitted light (Fig. 4), the chondrule appears to consist mainly of 13 discrete regions (labeled with lower-case Roman numerals). Most of the olivine in each individual region is in optical continuity and represents a single crystal.

Figure 4.

Transmitted light, crossed-polar image of chondrule K7p in approximately the same orientation as in Fig. 3. The main portions of the chondrule are divided into 13 separate regions (for the non-triskaidekaphobic); the olivine in each individual region is in optical continuity. A colored image of this chondrule appears as Fig. S1.

In BSE images (Figs. 2 and 3), there appear to be four sets of imperfectly defined internal spheroidal shells of olivine in the chondrule that resemble the outer shells of classic BO chondrules (e.g., fig. 1a of Nagahara 1983; fig. 1a of Weisberg 1987). These internal shells are similar to the rims around some primary BO chondrules that are surrounded by secondary chondrules in enveloping and adhering compound chondrules (fig. 1f of Weisberg 1987; Rubin 1992; figs. 1a, c, and 2b of Wasson et al. 1995). The internal olivine shells in K7p divide the chondrule into four discrete nested BO chondrules (Fig. 3). The shells are thicker and more structurally coherent on the right side of chondrule K7p than the left.

Each olivine shell contains more metallic-Fe-Ni blebs than are associated with mesostasis in the chondrule interior. These metal blebs consist of kamacite containing approximately 5 vol% 1–3 μm-sized, rounded-to-vermicular taenite grains. Larger metal blebs tend to be located farther from the chondrule center. The chondrule surface is marked by a near-continuous shell of blebby opaque stringers consisting mainly of carbide, oxide, and small amounts of Ni-rich metal. These stringers texturally resemble the metallic Fe-Ni rings surrounding many chondrules in CR carbonaceous chondrites (e.g., fig. 3 of Wasson and Rubin 2010). Patches of mesostasis occur throughout chondrule K7p; within chondrule 1; and on both sides of olivine shells 1, 2, and 3.

Chondrule 1 is the primary chondrule, constituting 7% of K7p; in plan view, it forms an ellipse with an eccentricity of approximately 0.5. It is completely surrounded by an olivine shell that is 50–120 μm thick and approximately 1260 μm in circumference. The internal olivine bars are discontinuous; some are completely surrounded by mesostasis (in two dimensions), others are attached to the olivine shell. At their interface with the mesostasis, the bars display rounded to pyramidal terminations. A few rounded blebs of kamacite (7–20 μm in diameter) occur at the inside and outside boundaries of the olivine shell. The mesostasis throughout chondrule 1 consists of aqueously altered glass composed of elongated crystalline blades (10–40 × 0.8–1.5 μm) flanked by ellipsoidal (1–2 μm) to elongated (1 × 15 μm) holes (Fig. 5). Some small (5–10 μm-sized) Ca-pyroxene crystallites occur within portions of the altered mesostasis.

Figure 5.

High-magnification view of a portion near the right edge of chondrule 1 showing the altered mesostasis of chondrule 1 and some isotropic glass from chondrule 2 (region 2b). The altered mesostasis contains holes flanked by blades of material that might have been more crystalline and harder to dissolve by the aqueous fluid. BSE image.

Chondrule 2 is approximately 870 μm in maximum diameter. In plan view, it constitutes 13% of K7p. The olivine shell of chondrule 2 is much better defined on the right side (area 2b) than on the left (area 2a). The mesostasis at the bottom of the chondrule and within the lower halves of areas 2a and 2b has been altered and appears mottled (Fig. 3); mesostasis at the upper ends of areas 2a and 2b consists of isotropic glass. The boundary between the altered mesostasis and the isotropic glass is very sharp and somewhat lobate (Fig. 6). Between the olivine shells of chondrules 1 and 2, olivine occurs mainly as bars and quasi-equant grains surrounded by mesostasis; some short olivine bars in area 2a nucleated on the olivine shell surrounding chondrule 1. A few small kamacite blebs occur in the shell surrounding chondrule 2, particularly on the left side (i.e., between areas 2a and 3a).

Figure 6.

Boundary between altered mesostasis and isotropic glass in area 2b. The altered mesostasis consists of elongated holes and blades of presumably more resistant material. The olivine grain at top center is surrounded by a sheath of less-mottled mesostasis that is compositionally indistinguishable from the altered mesostasis adjacent to it. The boundary between the altered mesostasis and the isotropic glass is sharp and somewhat lobate; it probably marks the farthest extent of glass dissolution during parent-body alteration. BSE image.

Chondrule 3 is approximately 1290 μm in maximum diameter, constituting 20% of K7p in plan view. As in chondrule 2, the olivine shell of chondrule 3 is much better defined on the right side (area 3b) than on the left (area 3a). The mesostasis at the bottom of area 3a has been altered; mesostasis at the upper ends of areas 3a and 3b consists of isotropic glass. The mesostasis abundance in chondrule 3 is relatively low (approximately 20 vol%); it occurs in 100 μm-sized patches between coarse (100–200 μm) olivine grains in area 3b and surrounding smaller (15–70 μm) olivine grains in area 3a. The metal blebs in chondrule 3 tend to be appreciably coarser (25–100 μm) than in chondrules 1 or 2 (typically 5–15 μm). Some small metal blebs occur within the olivine shell surrounding chondrule 3, particularly on the bottom and right sides.

Chondrule 4 constitutes the outer portion of the entire chondrule. In plan view, it constitutes 60% of K7p. It has a thick olivine shell intergrown with abundant opaque blebs and stringers that presumably consisted of metallic Fe-Ni prior to parent-body aqueous alteration. Some of the stringers extend across the poorly defined boundary between chondrules 3 and 4. The largest opaque bleb (305 × 410 μm) occurs at the top of the chondrule and extends into the interior; it appears to be associated with the opaque stringers at the K7p-chondrule margin. The large bleb consists of iron-carbide with numerous micrometer-sized (and a few 10 μm-sized) patches of taenite and tetrataenite; iron-oxide and rare grains of troilite occur at the margin of the bleb.

There are a few small patches of mesostasis between the olivine shell of chondrule 3 and the olivine shell and opaque stringers at the margin of chondrule 4. The mesostasis at the bottom and left side (area 4a) of chondrule 4 has been altered; area 4b contains isotropic glassy mesostasis.

Phase Compositions

Mafic silicates in chondrule K7p include major forsterite and rare enstatite and Ca-pyroxene. All three phases are very reduced and contain significant concentrations of minor elements (Table 1). Olivine in chondrules 1, 2, and 3 is rather homogeneous, but olivine in chondrule 4 is richer in FeO (0.75 versus 0.52–0.60 wt%) and MnO (0.07 versus 0.04 wt%). Two-sample, two-tailed t-tests show that chondrules 4 and 1 differ significantly in FeO (t = 3.10, 2α = 0.01, confidence level = 99%) and MnO (t = 135, 2α < 0.00001, confidence level > 99.999%). It is clear that olivine in the outermost portion of chondrule K7p is richer in FeO and MnO than in the interior portions.

Table 1. Compositions of olivine and pyroxene (wt%) in different areas of chondrule K7p
Area12343,44
MineralForsteriteForsteriteForsteriteForsteriteEnstatiteCa-pyx
# grains71217621
 MeanSDMeanSDMeanSDMeanSDMeanMean
SiO242.50.142.40.242.30.342.30.358.451.1
TiO20.040.030.040.030.040.030.040.030.161.3
Al2O30.100.070.110.100.200.060.070.060.937.3
Cr2O30.210.070.230.100.160.080.280.040.360.32
FeO0.520.050.600.140.530.140.750.190.490.33
MnO0.040.020.040.04<0.040.040.070.020.060.18
MgO56.90.256.70.456.80.256.90.238.920.6
CaO0.280.030.320.080.400.070.260.080.5418.9
Na2O<0.040.01<0.040.01<0.040.01<0.040.02<0.040.07
K2O<0.040.00<0.040.01<0.040.01<0.040.00<0.04<0.04
Total100.6 100.4 100.4 100.7 99.8100.1
mol% Fa0.510.050.590.140.520.140.740.18  
mol% Fs        0.700.54
mol% Wo        0.9839.5

Isotropic glass occurs in areas 2a, 2b, 3a, 3b, and 4b (Table 2). A CIPW-norm calculation of its mean composition indicates that the glass is dominantly quartzofeldspathic (74 wt%) with a significant pyroxene component (24 wt%). Except for FeO, which is slightly low, isotropic glass in K7p is within the compositional range of previously analyzed glassy mesostases in Type-IA Semarkona chondrules (table 3 of Jones and Scott 1989).

Table 2. Composition of mesostasis (wt%) in chondrule K7p
Area2a2b3a3b4bAverage glassAverage mesostasis
MesostasisIsotropicIsotropicIsotropicIsotropicIsotropicIsotropicAltered
# points12211616178232
 MeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSD
SiO260.00.258.30.460.70.359.10.357.50.459.11.053.01.4
TiO20.680.040.780.040.610.030.720.040.660.070.700.080.860.37
Al2O317.80.219.20.317.10.218.40.219.40.518.51.020.61.4
Cr2O30.380.050.410.070.220.070.270.050.190.060.300.110.380.15
FeO0.190.030.150.030.280.120.190.050.430.110.250.130.590.42
MnO0.040.020.060.020.080.020.070.020.110.030.070.030.050.03
MgO5.00.35.10.44.90.35.00.44.70.64.90.45.01.7
CaO15.60.215.40.215.50.315.50.215.00.315.40.316.41.0
Na2O0.320.420.410.400.630.520.460.441.20.40.610.520.930.42
K2O<0.040.01<0.040.01<0.040.01<0.040.01<0.040.01<0.040.010.120.08
Total100.0 99.8 100.0 99.7 99.2 99.8 97.9 

In individual regions, the glass in K7p is compositionally rather homogeneous, but there are small but significant differences among regions. For example, isotropic glass in area 4b differs from that in 2b in TiO2 (t = 6.64), Cr2O3 (t = 10.3), FeO (t = 11.2), MnO (t = 6.14), and Na2O (t = 6.14); from area 2a and 2b in CaO (t = 6.03 and 4.92, respectively); and from areas 3a and 3b in Al2O3 (t = 17.1 and 7.45, respectively). Isotropic glass in area 2a differs from that in 3a in SiO2 (t = 6.99), Al2O3 (t = 45.8), and Cr2O3 (t = 6.72). Isotropic glass in 2b differs from that in 3b in SiO2 (t = 6.69). All of these t-tests have 2α values < 0.0001 and confidence levels > 99.99%.

Altered mesostasis (Table 2) occurs at the bottom and left side of the chondrule (in areas 1, 2a, 2b, 3a, and 4a). A CIPW-norm calculation of its mean composition indicates that it is mainly quartzofeldspathic (71 wt%) with a significant pyroxene component (24 wt%), i.e., not that different from isotropic glass. Nevertheless, compared with isotropic glass, altered mesostasis has lower SiO2; more-heterogeneous SiO2; higher K2O; more-heterogeneous K2O; and more-heterogeneous TiO2, Al2O3, FeO, MgO, and CaO (Table 2). (Altered mesostasis and isotropic glass differ in their concentrations of SiO2 and K2O [with t values of 26.0 and 11.2, respectively; 2α < 0.0001].) The low mean analytical total (97.9 wt%) of altered mesostasis compared with that of mean isotropic glass (99.8 wt%) mainly reflects the presence of pores in the altered material; it may also reflect a greater degree of hydration. (The totals in isotropic glass and altered mesostasis are significantly different [t = 15.5; 2α < 0.0001].)

Kamacite is restricted to the interior of chondrule K7p and is similar in Co and Ni content (Table 3) to the mean kamacite composition in Semarkona determined by Rubin (1990). The moderately high Cr content (0.62 wt%) of K7p kamacite is similar to that of spheroidal kamacite blebs within chondrule interiors from LL3 Bishunpur (Table 1 of Rambaldi and Wasson 1981). The tiny taenite grains within the kamacite blebs in K7p are richer in Cr (1.5 wt%) than the kamacite.

Table 3. Composition of opaque phases (wt%) in chondrule K7p
MineralKamaciteTaeniteTetrataeniteSulfideCarbide
# points2413126
  1. Sulfide normalized to 100.0 wt%; prenormalization total is 94.4 wt%. C in carbide determined by difference.

  2. nd = not determined.

Fe94.677.241.464.486.7
Ni4.321.358.80.413.9
Co0.310.061<0.040.24
S<0.04<0.040.1133.8<0.04
Cr0.621.50.181.40.69
Cndndndnd8.5
Total99.8100.1101.5100.0100.0

Tetrataenite and carbide in the altered regions at the chondrule margin (Table 3) are within the established compositional ranges for these phases in Semarkona (fig. 9 of Krot et al. 1997b). Troilite (which was found only in the altered portion of K7p) has moderate amounts of Cr and Ni, but does not differ appreciably in composition from other occurrences of troilite in Semarkona.

Bulk Composition

The bulk composition of chondrule K7p and the bulk composition of its silicate portion can be calculated from the modal abundances and compositions of the individual phases (Table 4). A CIPW-norm calculation of the bulk silicate portion of K7p yields (in wt%): 11.4% plagioclase, 5.5% diopside, 18.3% hypersthene, 64.0% olivine, 0.34% ilmenite, and 0.25% magnetite.

Table 4. Inferred initial bulk composition (wt%) of chondrule K7p
 Silicate portionEntire chondrule Entire chondrulea
  1. All opaque phases are assumed to be kamacite.

  2. a

    O concentration determined by difference.

SiO246.527.0Si12.6
TiO20.180.10Ti0.06
Al2O34.02.3Al1.2
Cr2O30.230.13Cr0.35
FeO0.500.29Fe40.0
MnO0.05<0.04Mn<0.04
MgO44.926.0Mg15.7
CaO3.52.0Ca1.4
Na2O0.130.08Na0.06
K2O<0.04<0.04K<0.04
Fe<0.0439.7Ni1.8
Ni<0.041.8Co0.13
Co<0.040.13O26.7
Cr<0.040.26  
Total100.099.8Total100.0

The prealteration composition of chondrule K7p is inferred to have had much more metallic Fe-Ni (41.5 wt%) than typical Type-IA chondrules in Semarkona (0.43–8.0 wt%; table 5 of Jones and Scott 1989). Similarly, BO chondrules in ordinary chondrites (OC) generally have only negligible to accessory amounts of metallic Fe-Ni (Weisberg 1987). Although the silicate portion of chondrule K7p (Table 4) is within the compositional range of bulk Semarkona Type-IA chondrules (table 5 of Jones and Scott 1989), relative to their mean composition, K7p is depleted in volatiles (0.13 versus 0.52 ± 0.59 wt% Na2O; <0.04 versus 0.07 ± 0.06 wt% K2O; <0.04 versus 0.07 ± 0.05 wt% S) and FeO (0.50 versus 1.18 ± 0.89 wt%).

Discussion

Multiple Melting and Chondrule Crystallization

Dustballs and Chondrule Melting Episodes

The presence of porous dustballs in the nebula is indicated by matrix lumps in ordinary and CV chondrites (Kurat 1970; Ikeda et al. 1981; Scott et al. 1984), dark inclusions in different carbonaceous chondrite groups (e.g., Johnson et al. 1990; Krot et al. 1995; Kojima and Tomeoka 1996; Zolensky et al. 1997), microchondrule-bearing rims around normal-sized chondrules in ordinary chondrites (Rubin et al. 1982; Rubin 1989; Krot and Rubin 1996; Rubin and Krot 1996; Krot et al. 1997a), igneous rims around chondrules in every chondrite group except CI (Rubin 1984; Rubin and Wasson 1987; Rubin et al. 1990; Krot and Wasson 1995), and as the secondary shells of enveloping compound chondrules (Wasson et al. 1995).

Several of the thick ellipsoidal bands of olivine within chondrule K7p, particularly those between area 1 and 2a, 1 and 2b, and 2b and 3b, closely resemble the rims on individual BO chondrules and the internal olivine rims around primary BO chondrules in compound-chondrule objects. By analogy, it seems likely that the sets of internal olivine bands in K7p mark the outer boundaries of chondrule K7p during different stages of its development.

These sets of internal olivine shells in chondrule K7p indicate that in the nebular region where LL chondrules formed, a moderate amount of dust was present. One dustball was melted to form chondrule 1. Sometime after solidification, this chondrule was enshrouded in a second dustball; this second dustball melted to form chondrule 2 (which occurs as an igneous BO shell around chondrule 1). This two-step process repeated two more times to form chondrules 3 and 4. It is unclear how much time elapsed between the cyclical processes involving repeated episodes of chondrule melting interspersed with incorporation of solidified chondrules into dustballs.

The similarities in olivine and mesostasis composition in each of the nested chondrules in K7p indicate that, in general, the dustball components were rather homogeneous in composition and underwent little compositional change during the period when LL chondrules underwent multiple episodes of melting. This suggests that the interval between one melting episode and the subsequent incorporation of the entire object into a new dustball may have been relatively short.

Moderate differences in the proportions of mesostasis and olivine among the nested chondrules in K7p suggest that individual dustballs may have had variable proportions of olivine and feldspathic material. Even if the dust itself was homogeneous, the dustballs may have contained different abundances of pre-existing chondrule fragments; the fragments themselves may not have been compositionally representative of the whole chondrules from which they were derived. The more ferroan compositions of olivine and glass in chondrule 4 (the outermost chondrule in K7p) indicate that, with time, the dust in the LL-chondrule formation region became richer in FeO, presumably as a result of increasing oxidation of metallic Fe.

Chondrule melting experiments indicate that objects with BO textures formed by essentially complete melting of olivine-rich precursors (Hewins 1988; Hewins and Fox 2004; Hewins et al. 2005); temperatures may have approached 1850 °C (Hewins et al. 2005). The occurrence of BO-like textures in areas 2a, 2b, and 3a suggest that the melts in these regions contained few nuclei. Areas 3b, 4a, and 4b texturally resemble porphyritic olivine (PO) as well as BO chondrules; their precursor melts might have contained some relict olivine grains.

Details of the formation of the different subchondrules within K7p are discussed below.

Chondrule 1

The first stage in the development of chondrule K7p was the total melting of a dustball to form chondrule 1.

At this point, the chondrule would have had dimensions similar to that of chondrule 1. However, the olivine rim could have been thinner and the olivine bars may have had a somewhat different texture.

Sulfide-rich liquids readily wet silicates (Barnes et al. 2008), but metallic Fe liquids do not (Minarik et al. 1996). It thus seems plausible that the nonsilicate melts in the interior of chondrule 1 were initially sulfide-bearing and migrated via porous flow (Gaetania and Grove 1999) to the chondrule surface through conduits in the silicate melt (Wasson and Rubin 2010). If S in the melt evaporated at the chondrule surface, this would have decreased the surface energy of the melts, reducing the contact area between the metal and silicate (Wasson and Rubin 2010). Individual metal beads would have formed at the chondrule surface (e.g., Yu et al. 1996) as it commonly occurs in type-I CR chondrules (e.g., fig. 3 of Wasson and Rubin 2010).

Chondrule 2

At some point after it solidified, chondrule 1 was incorporated into a porous olivine-rich dustball very similar in composition to the one from which chondrule 1 was itself derived. A second chondrule-heating event totally melted the dustball, but not most of the interior of the underlying chondrule. (Fine-grained porous dustballs have very large surface areas and are more readily melted than intact chondrules, particularly under conditions of minimal superheating [Wasson 1993].) This particular heating event did not extensively melt chondrule 1, but some of the olivine and much or all of the mesostasis in chondrule 1 were melted at this time. The experiments of Fox and Hewins (2005) suggest that peak temperatures for this melting event were probably approximately 1600 ± 50 °C. The right side of the rim around chondrule 1 and some of the underlying olivine phenocrysts were melted; the olivine in region vii (Fig. 4) then crystallized in optical continuity with the olivine grains in area 2b. (Compound olivine chondrules commonly exhibit optical continuity between the primary and secondary at their interface [Wasson et al. 1995], presumably due to epitaxial growth of the olivine crystals on the pre-existing substrate.) Chondrule 2 formed as an enveloping BO shell around chondrule 1 (but portions of area 2a were later melted). An olivine ring formed around chondrule 2; olivine bars nucleated from the melt, and metal or metal-sulfide migrated toward the chondrule-2 surface.

At this stage, the chondrule was a compound BO object, comparable in size to chondrule 2 (approximately 850–900 μm in diameter) and grossly resembling the nested BO chondrules illustrated in fig. 1f of Weisberg (1987).

Chondrule 3

The process repeated itself. Chondrule 2 was incorporated into an olivine-rich dustball similar in composition to the precursors that were melted to form chondrules 1 and 2. A third heating event totally melted this dustball (but not the underlying chondrules) to form a BO shell around chondrule 2. During this event, part of the (now-internal) olivine shell around chondrule 2 may have been melted, particularly in area 2a/3a and at the bottom of the chondrule. The mesostases in chondrules 1 and 2 may also have been melted during this event. Olivine in the chondrule-3 melt nucleated on the chondrule-2 shell and crystallized within the melt. Area 3b corresponds for the most part to region xi (Fig. 4); much of area 3a was later remelted.

Chondrule 3 has a lower proportion of mesostasis than chondrules 1 or 2 and thus probably formed from a dustball that was richer in olivine. Metal or metal-sulfide migrated to the chondrule surface and formed a few approximately 100 μm-sized blebs. After solidification, the entire object consisted of three nested BO chondrules comparable in diameter to chondrule 3, i.e., approximately 1250–1300 μm.

Chondrule 4

Chondrule 3 became enmeshed in another olivine-rich dustball. The more-ferroan olivine and more-ferroan glass in chondrule 4 (Tables 1 and 2) suggest that this dustball had a higher FeO content than the dustball precursors of chondrules 1–3. This is consistent with various condensation models of a cooling solar nebula that show metallic Fe becoming more oxidized with time, resulting in dust with higher FeO (e.g., table 5 of Grossman 1972), particularly under conditions of enrichment in CI-composition dust (e.g., Petaev and Wood 2005). The relatively low amount of mesostasis in chondrule 4 and the large amount of metal at the rim of K7p indicate that the dustball precursor of chondrule 4 was very rich in olivine and metal, but relatively depleted in feldspathic material. However, it is possible that some of the metal at the margin of chondrule 4 was derived from the chondrule interior (e.g., from a pre-existing metal-rich rim around chondrule 3) and migrated to the new chondrule surface.

A fourth episode of chondrule heating melted the chondrule-4 precursor dustball and probably remelted mesostases in the underlying chondrule portions; olivine bars and internal olivine shells of the interior chondrules may have been partly melted as well. Metal or metal-sulfide migrated to the chondrule-4 surface; the present lack of sulfide indicates that if any sulfide had indeed been present, the S must have completely evaporated. Region xiii corresponds more or less to area 4b. Extensive melting of the left side of the chondrule resulted, upon cooling, in olivine grains crystallizing in optical continuity with those of chondrule 1, forming the upper left side of region iii.

Crystallization of Mafic Phases

A modified and normalized modal composition of the silicate portion of chondrule K7p would approximate a rock containing (in wt%) 76% forsterite, 2% enstatite, and 21% anorthite. This would plot deep within the forsterite field of the forsterite-anorthite-silica pseudoternary diagram. Crystallization of forsterite from a melt of this composition would move the residual liquid directly away from the forsterite corner of the diagram toward the forsterite–anorthite subtraction curve. The absence of anorthite crystals in the chondrule indicates that the melt must have quenched before reaching the curve. Because pyroxene would have crystallized (in small quantities) from this melt only after anorthite had crystallized, little pyroxene is expected to have formed. The overall compositional similarities among chondrules 1, 2, 3, and 4 indicate that after each melting event, crystallization would have proceeded in approximately the same manner.

Formation of Compositionally Heterogeneous Glass

The altered mesostasis occurs mainly at the bottom and left side of chondrule K7p. The holes in the mesostasis (Fig. 5) formed by dissolution. This is the same parent-body alteration process responsible for forming the bleached zones within nonporphyritic chondrules in OC (e.g., fig. 2 of Grossman et al. 2000). The blades adjacent to the holes in the altered mesostasis (Fig. 5) probably represent more-crystalline portions of the glass that remained after dissolution.

Grossman et al. (2002) found that 35% of Type-I (low-FeO) porphyritic chondrules in Semarkona exhibit compositional zoning in their mesostases: glass in the outer parts of the chondrules tends to be enriched in Na, K, Rb, F, Cl, and H and depleted in Ca, Ti, and Cr. They found that Al and Si may be either enriched or depleted in the outer parts of the chondrules, that Mn correlates with Na in some of the chondrules (but not in others), and that Fe and Mg are not zoned. Grossman et al. concluded that light aqueous alteration on the parent body caused glass in the outer parts of the chondrules to hydrate and undergo structural modification; this in turn aided elemental exchange between the glass and the chondrule rim and surrounding chondrite matrix.

Isotropic glass in chondrule K7p follows many of the same trends identified by Grossman et al. (2002). As shown in Table 2, glass in portions of chondrule 4 (the outermost part of chondrule K7p) has more Al2O3 than glass in portions of chondrule 3 and less CaO, TiO2, and Cr2O3 and more FeO, MnO, and Na2O than glass in portions of chondrule 2. Except for the enrichment of Fe in chondrule-4 glass, these trends are generally consistent with parent-body aqueous alteration (Grossman et al. 2002).

Although the high analytical totals (99.2–100.0 wt%) in isotropic glass from different chondrule regions (Table 2) indicate that the glass was not extensively hydrated, it is well established that igneous glass of various compositions can undergo moderate hydration and chemical alteration without devitrifying (e.g., Lipman et al. 1969; Tsong et al. 1978, 1980; Berger et al. 1988; Mungall and Martin 1994).

As shown in Fig. 6, the boundary between the regions of altered mesostasis and isotropic glass in chondrule K7p is sharp and moderately lobate. Although all of the glass in the chondrule has been subject to at least some aqueous alteration, the sharp lobate boundary between the obviously altered mesostasis and the still-isotropic glass marks a reaction front beyond which glass dissolution was not a significant process. Nevertheless, it seems likely that minor amounts of water diffused through the isotropic glass, affecting the glass structure and facilitating elemental diffusion and exchange with porous materials outside the chondrule (Grossman et al. 2002). Sodium was introduced into the isotropic glass and minor amounts of Ca were leached out.

Because Grossman et al. (2002) found that Fe in the mesostasis of aqueously altered Type-I chondrules in Semarkona was unzoned, the higher FeO content of the isotropic glass in chondrule 4 relative to that in chondrules 2 and 3 (0.43 versus 0.15–0.28 wt%; Table 2) probably represents an inherent difference in glass composition unaffected by parent-body alteration. This is supported by the higher FeO content of olivine in chondrule 4 relative to that in underlying chondrules 1–3 (0.75 versus 0.52–0.60 wt%; Table 1). It thus seems probable that the dustball precursor of chondrule 4 was more ferroan than those that gave rise to the earlier formed chondrules.

Aqueous Alteration of Metallic Fe-Ni

The opaque blebs and stringers at the outer margin of chondrule 4 texturally resemble the kamacite rings around many CR chondrules and were probably composed of kamacite ± troilite at the time the chondrule formed. However, at present, kamacite is absent from these blebs and stringers; they consist instead of carbide, iron oxide, Ni-rich metal and accessory troilite. Such phases are common in Semarkona: Taylor et al. (1981) described assemblages of carbide, iron oxide, kamacite, Ni-rich metal, troilite, and Ni-rich sulfide associated with the fine-grained porous matrices of Semarkona and other LL3 chondrites. Krot et al. (1997b) identified these same assemblages in unequilibrated members of all three OC groups and concluded that the assemblages formed from metal-sulfide nodules during parent-body hydrothermal alteration by a C-O-H-bearing fluid. This scenario is consistent with the reported presence of rare grains of phyllosilicate in Semarkona and LL3 Bishunpur (Hutchison et al. 1987; Alexander et al. 1989).

It seems likely that similar C-O-H-bearing fluids altered the kamacite at the margin of chondrule 4 and altered the glassy mesostasis to different extents throughout the interior of K7p. The O and H may have been derived from water in hydrated silicates (Wasson and Krot 1994) or ice (e.g., McSween and Labotka 1993), the C from hydrocarbons or graphite (Krot et al. 1997b).

Conclusions

Chondrule K7p is composed of four nested barred-olivine (BO) chondrules. The first olivine-rich dustball was completely melted to form chondrule 1 (the innermost BO chondrule). Sometime after solidification, chondrule 1 was incorporated into another olivine-rich dustball. A second episode of heating caused this dustball to melt, but much of chondrule 1 (the primary chondrule) remained intact even though its mesostasis and some of its olivine probably melted. At this stage, the object was an enveloping compound BO chondrule. The two-step process of chondrule incorporation into a dustball and the melting of the dustball (without extensive melting of the underlying chondrule) was repeated two more times.

The proportions of olivine and glass differ among chondrules 1–4, suggesting that the precursor dustballs were variable in bulk composition. This may have been due to the presence in individual dustballs of different amounts of chondrule fragments and different mineral proportions in the fragments. The final dustball was somewhat richer in FeO than the preceding dustballs; melting, crystallization, and quenching of this dustball produced the more-ferroan olivine and more-ferroan glass of chondrule 4. This indicates that during the period of chondrule formation in the LL nebular region, the composition of dust underwent a secular change, presumably due to increasing oxidation of metallic Fe. It is possible that more time elapsed between the formation of chondrules 3 and 4 than between the formation of chondrules 1, 2, and 3.

The entire chondrule later underwent parent-body aqueous alteration. Metal blebs and stringers at the chondrule surface were transformed into carbide and iron oxide with minor amounts of Ni-rich metal. Mesostasis in the chondrule was altered—in contiguous portions of the chondrule, the glass underwent dissolution, producing holes and adjacent blades of more resistant (perhaps slightly more crystalline) material. Although much of the glass remained isotropic, compositional zoning trends are consistent with minor hydration and leaching of the glass without causing devitrification. (These compositional trends show that the outer zones of the chondrule contain glass enriched in Na2O, MnO, and Al2O3 and depleted in CaO, TiO2, and Cr2O3 relative to interior zones.) The sharp and moderately lobate boundary between the extensively altered mesostasis and the isotropic glass represents the reaction front beyond which there was little or no glass dissolution.

Acknowledgments

I thank J. N. Grossman and two anonymous reviewers for helpful comments. This work was supported by NASA Cosmochemistry grant NNG06GF95G.

Editorial Handling

Dr. Randy Korotev

Ancillary