Fractionated matrix composition in CV3 Vigarano and alteration processes on the CV parent asteroid

Authors

  • Sean M. HURT,

    1. Department of Earth and Space Sciences, University of California, Los Angeles, California 90095–1567, USA
    Search for more papers by this author
  • Alan E. RUBIN,

    1. Institute of Geophysics and Planetary Physics, University of California, Los Angeles, California 90095–1567, USA
    Search for more papers by this author
  • John T. WASSON

    Corresponding author
    1. Department of Earth and Space Sciences, University of California, Los Angeles, California 90095–1567, USA
    2. Institute of Geophysics and Planetary Physics, University of California, Los Angeles, California 90095–1567, USA
    3. Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095–1567, USA
    Search for more papers by this author

Corresponding author. E-mail: jtwasson@ucla.edu

Abstract

Abstract— Although CV3 Vigarano is one of the most primitive CV chondrites, it has lost most of the S from the matrix; matrix Na is also depleted relative to the concentration in bulk CV chondrites. We used a matrix-grid technique to study thirteen 50 × 50 μm regions in Vigarano; in each area, we used an electron microprobe to gather data with an electron beam 3 μm in width. We found two end-member types of matrix textures. One is coarse and porous, has lower Fe contents and lower analytical totals; it appears to be contaminated with comminuted chondrule debris. The other is finer grained and appears smooth; its mean composition has higher Fe, but lower S and Al contents, than the coarse matrix areas. Our tentative interpretation is that the larger grain size of the coarse areas resulted from the admixing of comminuted chondrule materials, and thus that the initial fraction of nebular fines was higher in the fine matrix regions. Aside from volatiles, the overall composition of Vigarano matrix is similar to that observed in matrix-grid studies of other carbonaceous chondrites: Al, Si, Fe, and Mn have high whole-chondrite-normalized abundance ratios; Ca concentrations are low and highly variable. Because asteroidal alteration effects are present in our sample, it is difficult to resolve nebular signatures in the compositions of the grid areas.

Introduction

The matrix material of primitive chondrites is important because it preserves the composition, heterogeneity, and chemical trends of fines (material smaller than approximately 1 μm) from the early solar nebula (e.g., McSween and Richardson 1977; Huss et al. 1981; Nagahara 1984; Scott et al. 1984, 1988; Zolensky et al. 1993; Brearley 1996). However, the record of presolar processes contained within the interchondrule matrix can be destroyed by secondary parent-body alteration or terrestrial weathering (e.g., Tomeoka and Tanimura 2000; Brearley 2006). Studies of matrix that make inferences about nebular fines must either demonstrate that samples have escaped secondary alteration or invoke mechanisms capable of separating altered and unaltered material.

Nebular materials that accreted to form chondrites mainly consisted of chondrules and submicrometer fines, metallic Fe-Ni and sulfide, and minor amounts of refractory and other inclusions. Fines are a major constituent of chondrite matrix. Modal matrix abundances vary significantly among largely anhydrous chondrite groups, ranging from 8% in EH chondrites to approximately 35% in anhydrous carbonaceous chondrites (McSween 1977; Rubin et al. 2009; Rubin 2010). McSween and Richardson (1977) concluded that fines mainly formed by condensation of materials evaporated during an early, high-temperature period in the solar nebula. More recent articles have recognized that material evaporated during chondrule formation contributed to a large fraction of the fines (Brearley 1993; Huss et al. 2005; Nuth et al. 2005). Wasson (2008) proposed that approximately 99% of chondritic fines are condensates of materials evaporated during chondrule formation and that the compositions of these materials gradually evolved as a result of changes in nebular composition and ambient temperature.

To constrain the origin of nebular fines, Wasson and Rubin (2009) investigated primitive matrix from the CR2 chondrite LAP 02342. They used electron microprobe (EMPA) methods to determine the concentrations of 10 elements in matrix-rich areas and observed (1) a direct relationship between an element’s volatility and its enrichment in the matrix relative to the whole rock and (2) compositional heterogeneities that were interpreted as reflecting variations preserved in fine-grained nebular carriers. To examine whether such variations are recognizable in other primitive chondrites, we analyzed matrix material in CV3 Vigarano following the same procedures.

Vigarano was thought to be a good choice for studies of primitive matrix because it is the only fall among reduced CV chondrites, and together with Efremovka and Leoville, is one of the least-metamorphosed members of this subgroup. The Allende and Bali subgroups of oxidized CV chondrites have experienced much more parent-body aqueous alteration than the reduced CV chondrites. Vigarano is brecciated (Scott et al. 1992) and contains some oxidized, Bali-like material (Krot et al. 2000; MacPherson and Krot 2002). Some late-recovered samples of Vigarano may have been exposed to rain (see below), but there are no textural or mineralogical features that have been specifically attributed to terrestrial weathering.

Samples and Techniques

The section that we mainly studied (USNM 6295-3) has some oxidized areas (including a few magnetite-bearing metal-sulfide nodules present within some chondrules and at some chondrule surfaces), but most areas (including the ones we analyzed) are generally reduced. Nevertheless, section USNM 6295-3 is not ideal: it shows much evidence of chondrule crushing. The second section we have access to, USNM 6295-1, has an even higher abundance of crushed chondrule debris.

We used the UCLA LEO 1430 scanning-electron microscope to create a back-scattered electron (BSE) mosaic image of the entire section of USNM 6295-3 (Fig. 1) and superimposed a grid with intervals of 1 mm. Columns were assigned letters; rows were assigned numbers.

Figure 1.

 Mosaic BSE image of section USNM 6295-3; grid lines are spaced 1 mm apart. This section has a high density of large, relatively intact chondrules compared with some other Vigarano sections, but a detailed examination shows that much chondrule crushing has occurred.

We carried out two types of studies using the UCLA JEOL electron microprobe, both focusing, primarily, on two matrix-rich regions (G3 and D5). We initially created X-ray maps for five elements in the two regions and used these to select thirteen 50 × 50 μm areas for detailed study as 7 × 7 arrays; each step was approximately 7 μm.

X-Ray Maps

We created BSE maps and wavelength-dispersive X-ray maps for five elements: Al, Mg, Ca, Fe, and S. Mosaic X-ray maps were made for two 1.5 × 1.5 mm areas: D5 and G3 (Figs. 2 and 3). The maps were produced by moving the sample under a focused 1.5 μm electron beam. Readings were spaced 2 μm apart and dwell times were 50–55 ms.

Figure 2.

 Element (X-ray) and BSE maps for region D5 in Vigarano section USNM 6295-3. Matrix is recognizable as light gray regions on the Fe and BSE images and as uniformly medium gray areas on the Al images. Note the near absence of S in the matrix; higher S is always associated with chondrules or chondrule fragments.

Figure 3.

 Element (X-ray) and BSE maps for region G3 in Vigarano section USNM 6295-3. Matrix is recognizable as light gray regions on the Fe and BSE images and as uniformly medium gray areas on the Al images.

Electron Probe Studies of Matrix Grids

We studied the bulk matrix composition by analyzing 49 points, using an electron beam 3 μm in diameter in a square 50 × 50 μm grid. Both natural and synthetic standards were used: chromite for Cr, forsterite for Si and Mg, albite for Na, Mn-rich garnet for Mn, orthoclase for K, millerite for S, grossular for Al and Ca, and magnetite for Fe. The beam current was limited to 15 nA to minimize loss of alkalis. We used ZAF corrections. Because the matrix material is fine grained, it approaches the homogeneity of a single mineral phase; thus, errors introduced by applying this procedure should be minor.

The beam diameter of 3 μm (twice the beam diameter used in making the x-ray maps) was chosen to be larger than matrix grains, but small enough to recognize individual, single-phase grains that should be eliminated from matrix data sets. The locations of the 13 areas studied are shown on BSE maps in Figs. 4 and 6; detailed images are in Figs. 5 and 7. Magnetite-sulfide assemblages occur near each of these areas. We measured 10 elements (Na, Ca, K, Cr, Fe, S, Al, Mg, Si, Mn). With the exception of Ni and O, these represent all the elements with concentrations >1 mg g−1; O concentrations can be roughly inferred from concentrations of cations and stoichiometric parameters.

Figure 4.

 This BSE image of most of region D5 and a small portion of region F7 shows the locations of seven matrix-grid areas marked by squares. Note the bright area extending through region D5-6 to the NE and also surrounding the F7 grid area. Our data show systematic differences between such regions and the coarser grained areas such as that near grid area D5-2.

Figure 5.

 Detailed BSE views of seven grid areas: six in D5 and one in F7. Areas D5-6 and F7-1 have a fine texture, whereas D5-4 has a coarse and porous texture.

Results

Petrography of the Studied Regions: Variable Textures in the Interchondrule Regions of Vigarano

Generalizations about Texture, Grain Size, and Crushing

Although the studied section of Vigarano shows less brecciation than observed in some other sections, it shows strong evidence of crushing that has compromised the record in the interchondrule regions. These areas show highly variable textures, several of which we sampled in our studies. In detailed BSE images such as those in Figs. 5 and 7, we observed large variations in surface texture ranging from fine and smooth (e.g., area F7-1; Fig. 5f) to coarse and rough (e.g., area G3-3; Fig. 7c); the finer regions are brighter in BSE.

In the D5 region (Fig. 4) there is fine, smooth matrix in the NW portion of the figure, just SW of grid area 3 and the larger region trending NE from grid area 6 near the bottom of this image. The uniformly fine textures of these areas are reminiscent of aerial views of playas, the dry lakes that are common in southern California. In contrast, much plucking (or inherent porosity) can be seen in some of the matrix regions such as the one just SW of grid area 2, and the one just N of the smooth matrix region that extends NE of area 6. The coarser, more plucked grid areas tend to have lower analytical totals.

The same general textures are observed in the G3 region (Fig. 6). The area that includes G3-2 is relatively fine grained and smooth (Fig. 7b), whereas the region that includes G3-3 is coarser and has greater porosity (Fig. 7c). The coarse matrix appears to be associated with an admixture of comminuted chondrules. For example, there is a dark, coarse, elliptical region (probably a chondrule fragment) that includes area G3-3 and stretches NE from this location; its slightly darker appearance in BSE reflects a relatively low Fe concentration and greater porosity.

Figure 6.

 A BSE image showing most of region G3; locations of six matrix-grid areas are marked by squares. Grid area G3-2 is in a bright area; area G3-3 is near the SW end of a large (∼350 μm) oval dark matrix region oriented SW-NE.

Figure 7.

 Detailed BSE views of six grid areas in G3. G3-2 has a fine texture and G3-3 has a uniformly coarse, porous texture, and resides within a large dark oval of matrix.

Because the compositions are related to texture, we have assigned the grid areas to three textural classes: “coarse,”“smooth,” and “intermediate, ambiguous, or anomalous.” These categories are not fully independent of the compositional data; we were considering both when making the categorization, but the BSE images shown in Figs. 5 and 7 were the main factors.

A Survey of the Textures of the 13 Grid Areas

Thirteen grid areas were chosen (six from region D5, one from F7, and six from G3) that appeared to consist largely of matrix material (with few coarse components). The different areas exhibited a diversity of textures reflecting variations in grain size, porosity, and composition (as indicated by their brightness in BSE images).

Detailed views of the six grid areas in D5 and the one in F7 are shown in Fig. 5. The regions range from fine and smooth (D5-6, F7-1) to coarse and porous/ plucked (D5-4, D5-2) to intermediate, ambiguous, or anomalous (D5-1, D5-3, and D5-5). There is a striking contrast between the smooth texture of F7-1 and D5-6 and the coarse, plucked area of D5-4. Area D5-1, which has anomalously high Na and Ca contents, includes a dark-shaded clast and some short (approximately 4 μm) linear grooves, probably the result of plucking. Some grid areas seem to include patches of both fine and coarse materials.

The six grid areas in G3 are shown in Fig. 7. We group the observed textures as fine and smooth (G3-2, G3-6), coarse (G3-3) and intermediate, ambiguous or anomalous (G3-1, G3-4, G3-5). The coarse texture of G3-3 is remarkably uniform. The fine textures of G3-2 and G3-6 are interrupted in a few places by moderately round holes 3 to 5 μm in size, presumably formed by plucking. A 20 μm clast extends into grid area G3-5 on the upper right.

Rules for Discarding Anomalous Points

Because the initial goal of this study was to determine pristine matrix compositions uncontaminated by additional components that were incorporated during parent-body processes, it was necessary to discard anomalous data. There were sufficient numbers of analyses (49 in each grid) to allow us to develop criteria for rejecting anomalous points. We discarded points that are dominated by a single constituent such as a chondrule fragment or coarse mineral grain because these are not fine-grained matrix components. We also rejected aberrant data (e.g., >>2 σ away from the mean) even though there may be scant petrographic evidence indicating why particular points are compositionally deviant. Although some of the rejected points may represent nebular phases (and should not have been omitted), it seems likely that most compositionally anomalous points were produced by brecciation, minor aqueous alteration, or terrestrial weathering. Although Vigarano is an observed fall (22 January 1910 in Ferrara, Italy) and an 11.5 kg stone was recovered quickly, another stone (4.5 kg) was collected about a month after the fall (Grady 2000). Most research projects seem to have involved the second stone. Weather records for 1910 in Ferrara, Italy are unavailable, but the website, weatherspark.com, indicates that the probability for precipitation on any given day in January or February is approximately 25%. The probability that the second stone was rained on is approximately 1–(0.75)30 or 99.98%. Thus, Vigarano samples may have experienced some terrestrial weathering.

Although there is a degree of arbitrariness involved in rejecting points, we err on the side of making the matrix composition more uniform. This seems justified because CV-chondrite matrices appear to be rather homogeneous on 100 μm size scales (Table 2 of McSween and Richardson 1977).

Because Vigarano contains different matrix lithologies and because some elements showed very large (factors >3) ranges in concentration even in the “core” regions of individual grid areas, it was much more difficult to establish general limits for excluding points than it was in the Wasson and Rubin (2009) study of CR2 LAP 02342. Many of the excluded points in Vigarano are anomalous for more than a single element. We list the chosen elemental cut-off limits in Table 1.

Table 1.   Criteria employed to reject points from the grid areas.
ElementHigh (mg g−1)Low (mg g−1)
  1. For Mg, Al, Si, and Fe the points exceeding these limits were excluded from the means. The variance in Ca and K values was too large to allow practical limits to be set; the high Ca limit was chosen to avoid possible dilution of other phases by Ca-rich alteration products.

Na1.800.17
Mg13378
Al22.25.8
Si158100
S0.6150.080
Ca18.5
Cr4.40.62
Mn2.61.3
Fe393277

We found relatively few points dominated by a single phase. These few include two grains of troilite, several grains that have compositions close to that of Fa40 olivine, and a few grains of hedenbergite (indicating the presence of some oxidized CV material in our sample; cf. Krot et al. 1995). These points were discarded.

We also deleted high Ca values (those over 18.5 mg g−1) to decrease the likelihood of analyzing Ca-rich secondary minerals known to occur in Vigarano (andradite, Ca3Fe2Si3O12; hedenbergite, CaFeSi2O6; kirschsteinite, CaFeSiO4; MacPherson and Krot 2002) and to avoid the possible dilution of other elements by one of these Ca-rich phases.

Because the ranges for K were so extreme, we had to use different limits in different grid areas. In all cases, however, we discarded points that were >2 σ away from the mean of that grid.

Of the 637 analyzed points in the study, we eliminated about 130, a bit more than 20%. The largest number of discarded points is 20 in grid areas D5-3 and G3-1; the smallest is 3 in area F7-1.

Points that were above the cut-off limit in Mn, Cr or Na were omitted only from the means for these particular elements; the points were not omitted in diagrams or in means involving other elements.

Overview of Mean Compositions of the Matrix Grid Areas

Figure 8 shows means for each of the ten elements with 95% confidence limits on seven scatter diagrams. It is obvious that the 13 grid areas do not share a common, mean composition. In the Wasson and Rubin (2009) study of CR2 LAP 02342, it was noted that nearly all grid areas could be resolved in terms of two or more elements. The authors inferred that these differences were associated with the presence of fragile chondrule-like objects that had experienced low degrees (approximately 20%) of melting and had been crushed and mixed with nebular fines during the compaction of the parent asteroid.

Figure 8.

 Mean concentrations of the seven elements in the 13 grid areas are shown on seven element–element diagrams. Error bars give 95% confidence limits. Although the grid areas cluster on most diagrams, they are all resolvable in terms of one or more elements. The two areas with fine textures (F7-1 and G3-2) show high Fe and low Al concentrations. See text for more details.

The most variable elements in our set are K (which has mean values differing by a factor of six; Fig. 8a, b) and Ca (with means differing by a factor of three; Fig. 8d, g). We suspect that this indicates that in these regions, K and Ca experienced significant transport during asteroidal alteration; the K and Ca values might also reflect contamination by chondrule debris. Grid D5-3 with the highest K is one of the areas in which we discarded the highest number of points (including some with very high K contents). It is possible that some of these high-K points resulted from impact-induced volatilization of alkalis on the asteroid, followed by transport and preferential condensation of K. This would be consistent with the occurrence of a large number of impact-melt-rock clasts in ordinary chondrite (OC) breccias that are enriched in K (e.g., Kempe and Müller 1969; Fodor et al. 1974; Ikeda and Takeda 1979; Wlotzka et al. 1983). (It is also possible that some K is associated with small grains of aqueously produced halite or sylvite as in some OC breccias; Zolensky et al. 1999; Rubin et al. 2002.) Grid D5-3 also stands out because of its low Fe (Fig. 8e) and low Mn (Fig. 8f). The highest Ca is in area D5-1 (which also has one of the highest Na contents).

Aluminum is the most uniform element in terms of grid means. The two regions that are about 3 mg g−1 lower than the remainder are F7-1 and G3-2, areas that are singled out in the following section because of their fine textures; these two areas also share the distinction of having the highest Fe contents. At the other Al extreme is G3-3, discussed below as having a coarse texture.

There is no correlation among grid averages of Mg and Si in Fig. 8c, although there are positive correlations among Mg and Si points within each of the individual 13 grid areas. Grid averages for Mg and Si show the most consistency among all elements.

Relationship Between Composition and Texture

In Fig. 9, we show examples of the relationship between texture and composition. Most grid areas with fine textures (triangles) have high Fe (≥ 330 mg g−1) and high totals (≥910 mg g−1), whereas areas with coarse textures (diamonds) have low Fe (≤330 mg g−1) and low totals ≤930 mg g−1). In the coarse areas, there is a correlation between analytical totals and Fe content, whereas in the fine areas, there is no correlation.

Figure 9.

 Plots of analytical totals versus Fe contents in seven matrix grid areas. Diamonds and circles correspond to coarse matrix, triangles to fine matrix, and the squares are designated as fine, but show more scatter than the areas represented by the triangles. (a) Fine region G3-2, coarse region G3-3, and largely fine region G3-6. (b) Fine region F7-1, coarse region D5-4, largely fine region D5-6, and largely coarse region D5-2. Fine areas have higher Fe and higher analytical totals than coarse areas. The totals on the y-axis reflect that of the common oxides except for elemental S.

In the case of the areas represented by squares, we see that area D5-6 (Fig. 9b) differs from the other fine areas chiefly in terms of analytical totals that average 20 mg g−1 lower. (Totals are calculated on the basis of oxides except for S.) Fine area G3-6 (Fig. 9a) differs from the other fine areas in that it shows a greater spread in both Fe content and analytical totals. The data for area D5-2 (round points in Fig. 9b) show that it is similar in Fe content to coarse area D5-4, but has analytical totals that average approximately 3% higher.

Discussion

Correlation Between Analytical Totals and Fe Content

It appears that the correlation between totals and Fe contents could have largely resulted from a variation in porosity or water content; because the current fraction of hydrated phases in Vigarano is quite low, porosity seems the more likely contributing factor because only the coarse-grained (presumably more porous) areas exhibit a significant correlation between the Fe content and analytical total. This is essentially a dilution effect that should affect other major elements. There are moderate positive correlations between Mg (and Si) with totals (with similar relative variations), consistent with this conclusion. However, there is a possible correlation between the Mg/Si ratio and totals in nearly all the grid areas (with the exception of D5-3), not just within the coarse-grained ones.

Loss of Al, Na, K, and S from Fine Matrix

Although differences in porosity are probably responsible for the correlations between analytical totals and the contents of Fe, Mg, and Si, there are other factors that influence compositional variations among grid areas. Our working scenario is that the fine and coarse areas differ principally in terms of their relative fractions of crushed chondrule debris; coarse areas have appreciably larger fractions of this debris.

To examine the differences between the two end-member types of matrix, we compared averages of grid areas belonging to each type. At the bottom of Table 2, we list one coarse and two fine compositions; our preferred fine composition (A) is based on data for area F7-1 and G3-2 (the two fine areas with the highest analytical totals); the other composition (B) includes data for F7-1, G3-2, D5-6, and G3-6. Our mean composition for coarse matrix is based on D5-4, G3-3, and D5-2. To compensate for the dilution effects of porosity, Table 2 also includes the composition for each grid normalized to 1000 mg g−1 (under the assumption that porosity manifests itself primarily by low totals). Comparison of the un-normalized compositions of “fine A” and “coarse” shows that the mean Si and Mg contents of the fine and coarse areas are nearly the same, but that the Fe concentrations are statistically distinguishable. Forcing the data to total 1000 mg g−1 does not change the essential fact that the coarse-grained areas contain less Fe than the smooth areas, but both types of areas still have approximately the same Mg and Si contents. The most interesting elements are those having significantly higher concentrations in the coarse fraction despite the lower totals: Al is higher by a factor of 2.5 and S by a factor of 1.9. Na is higher by a factor of 1.2, which is marginally significant. Concentrations of most of the other elements (K, Ca, Mn) are about the same in the fine and coarse fractions considering their relatively large variances.

Table 2.   Mean elemental compositions of the matrix grids and their overall texture. The final three rows compare fine and coarse regions. Compositions normalized to totals of 1000 mg g−1 are in italics. All concentrations are in mg g−1.
AreaPointsNaMgAlSiSKCaCrMnFeTotal§
  1. *Fine A is the mean of grid areas F7-1 and G3-2; fine B is the mean of F7-1, D5-6, G3-2, and G3-6; coarse is the mean of areas D5-4, G3-3, and D5-2. Areas are listed with their overall texture as fine and smooth; coarse; or intermediate, ambiguous or anomalous.

  2. §The total assumes that all elements except S occur in their common oxide components.

D5-1
intermediate
351.03
1.13
101
110
14.6
16.0
127
139
3.1
3.4
0.28
0.31
6.9
7.6
2.05
2.24
1.85
2.02
332
362
914
1000
D5-2
coarse
371.07
1.20
107
119
15.1
16.8
129
144
2.5
2.8
0.34
0.38
4.5
5.0
2.17
2.43
1.93
2.16
308
344
894
1000
D5-3
intermediate
291.97
2.17
104
115
19.2
21.1
134
148
3.9
4.3
1.14
1.26
4.9
5.4
1.66
1.83
1.46
1.61
304
335
906
1000
D5-4
coarse
451.07
1.21
108
121
15.8
17.7
127
142
4.8
5.44
0.20
0.23
2.3
2.6
1.49
1.67
1.83
2.05
308
346
889
1000
D5-5
intermediate
420.95
1.05
101
112
15.9
17.6
124
137
2.7
2.95
0.21
0.23
3.7
4.1
1.20
1.33
2.08
2.31
340
366
899
1000
D5-6
fine
420.83
0.90
100
109
15.0
16.4
121
133
2.4
2.62
0.13
0.14
3.1
3.4
1.60
1.75
1.99
2.18
347
380
913
1000
G3-1
intermediate
240.75
0.76
106
113
13.8
14.7
127
136
2.9
3.0
0.24
0.24
4.8
6.9
1.64
1.70
1.81
1.88
347
368
933
1000
G3-2
fine
430.68
0.72
110
118
10.9
11.6
122
130
2.0
2.1
0.21
0.22
3.2
3.4
2.52
2.69
1.83
1.95
358
381
937
1000
G3-3
coarse
451.04
1.16
98
110
17.3
19.3
128
144
2.5
2.8
0.29
0.33
2.4
2.6
1.38
1.55
2.04
2.29
320
358
893
1000
G3-4
intermediate
400.91
0.98
108
116
14.0
15.0
135
145
2.6
2.8
0.26
0.28
4.7
5.0
2.35
2.52
1.91
2.05
327
351
932
1000
G3-5
intermediate
360.88
0.95
102
110
14.2
15.4
138
149
2.8
3.1
0.45
0.49
3.7
4.0
1.85
2.00
1.97
2.12
326
352
926
1000
G3-6
fine
440.63
0.68
102
110
15.8
17.1
128
139
3.0
3.2
0.22
0.24
4.4
4.8
2.15
2.31
1.94
2.09
339
366
925
1000
F7-1
Fine
451.07
1.14
99
105
10.6
11.3
131
139
1.5
1.6
0.32
0.35
2.5
2.7
1.73
1.84
2.11
2.25
359
382
939
1000
Fine A* 0.87
0.93
104.5
111
10.7
11.4
126
134
1.75
1.87
0.27
0.29
2.9
3.1
2.12
2.26
1.97
2.1
359
383
938
1000
Fine B* 0.80
0.86
102.6
111
13.1
14.1
126
136
2.22
2.39
0.22
0.24
3.3
3.6
2.00
2.16
1.97
2.12
351
378
928
1000
Coarse* 1.06
1.19
104.0
117
16.0
17.9
128
143
3.27
3.67
0.28
0.31
3.1
3.48
1.68
1.88
1.93
2.16
312
350
892
1000

In Fig. 10, the fine and coarse fractions are plotted normalized to Mg and to the mean CV composition given by Wasson and Kallemeyn (1988) with elements increasing in volatility from left to right. The first thing to note is that the difference between the fine and the coarse materials is small (10–20% higher in coarse) compared with the differences between matrix-grid areas and the bulk CV composition; this fractionation will be discussed below. Relevant to the present discussion are the fractionations discussed above; the depletion of Al, Na, K, and S in the fine materials relative to the coarse.

Figure 10.

 Comparison of bulk-CV- and Mg-normalized ratios of smooth-fine (mean of areas F7-1 and G3-2) and coarse (G3-3, D5-4, and D5-2) grid areas in Vigarano. Aluminum, Na, K, and S are 20–50% lower in the fine area relative to the coarse. The mean CV-chondrite bulk composition has an abundance ratio of unity on the diagram.

There are three possible end-member explanations for the fractionations of labile elements: (1) These are nebular effects (e.g., Scott et al. 1984), and the coarse materials (which we infer to be enriched in comminuted chondrules) had higher concentrations of these elements. (2) These differences are caused by asteroidal aqueous alteration effects (e.g., Housley and Cirlin 1983). (3) Impact heating affected the concentrations of these elements (e.g., Wlotzka et al. 1983).

For the three volatiles (Na, K, S), there is abundant evidence in the literature that they were more enriched in nebular fines than in average chondrules or bulk chondrites (which should have higher concentrations of volatiles than chondrules). For example, the Na/Mg, K/Mg, and S/Mg ratios in CR matrix (Table 1 of Wasson and Rubin 2009) are approximately 3.5–5 times higher than the corresponding ratios in CR whole rocks (Table 3 of Kallemeyn et al. 1994; Table 1 of Mason and Wiik 1962). We therefore conclude that the second explanation is correct, i.e., that matrix Na, K, and S was mobilized during asteroidal alteration. In the case of Vigarano, as in other CV chondrites, both aqueous and thermal processes seem to have been involved. Hydration and dehydration of CV matrix (e.g., Krot et al. 1995) could have caused volatile elements in the finest-grained component to be lost. Although this hydration/ dehydration process was inferred for oxidized CV chondrites, it is plausible to assume that it also occurred (to a lesser extent) in reduced CV chondrites. These processes were likely to be more effective in fine matrix because of enhanced surface/volume ratios compared with coarse matrix.

The next question is where the sinks were for these mobilized volatile elements. Their volatility suggests that they could have been evaporated during a heating event and lost from (this part of) the asteroid. However, it is difficult in practice to lose volatiles from below the surface of an asteroid because there is no carrier phase and thus no appreciable pressure causing upward flow. Vigarano is a breccia (e.g., fig. 1 of Krot et al. 2000; p. 91 of Bevan and De Laeter 2002) containing solar-wind-implanted rare gases (Mazor et al. 1970; Matsuda et al. 1980); it probably formed in the near-surface environment. Impact processes can cause volatile loss from surface materials; here, it is easier to lose volatile elements entrained by H2O to space because the velocity of light gases can exceed escape velocities during impact events.

Although Al is not a volatile, it is commonly enriched in chondritic fines compared with bulk chondrites. Wasson and Rubin (2009) suggested that this enrichment could reflect loss of Al-rich mesostasis microdroplets during chondrule formation and incorporation of this material into nebular fines. Alternatively, chondrule Al could have been dissolved in aqueous fluids and precipitated as a hydroxide, perhaps in the voids; if these voids were outside compact matrix regions, we would not have analyzed them.

Although Al is not normally considered a leachable element, it has in fact been leached from the glassy mesostasis of radial pyroxene and cryptocrystalline chondrules in the Semarkona LL3.0 chondrite (Grossman et al. 2000). Semarkona shows only a few features associated with aqueous alteration; thus, the fraction of initial water must have been very low, perhaps 10 mg g−1. In nebular fines, Al is probably in an amorphous phase, and it is reasonable to infer that it is labile. Even though Vigarano has experienced far less alteration than oxidized CV chondrites, it appears to have experienced minor thermal alteration (e.g., Bonal et al. 2006).

Fe in the Matrix and High Analytical Totals in Vigarano

Iron is the dominant element in the matrix by mass fraction and is the best tracer of matrix among the elements we studied. However, like the other elements, Fe is heterogeneously distributed. The Fe concentration values range from lows near 270 mg g−1 to highs of 390 mg g−1 (more extreme points were attributed to contamination by single phases and discarded).

Analytical totals in Vigarano are positively correlated with Fe contents in the coarse areas. Some totals are quite high, particularly in the fine areas; as shown in Fig. 9, the totals in areas with fine textures are mostly >92 wt% (two are slightly above 100 wt%). Because our data-reduction protocol assumes that all Fe is bound to O, the presence of appreciable metallic Fe could produce high totals. However, in the two “classic” fine-texture areas (F7-1 and G3-2), there is no positive correlation between analytical totals and Fe content. Because Fe is calculated as FeO, if any magnetite were present (due to the occurrence of some oxidized CV material), we would obtain low analytical totals. Although some Fe is present in troilite, if we had analyzed coarse troilite grains, this would have been revealed by enhanced S concentrations.

Mean totals are high, with an average of 91.5 wt%; this contrasts with the much lower totals of approximately 82 wt% in CR LAP 02342 (Wasson and Rubin 2009) and approximately 84 wt% in the ungrouped C3 chondrite Acfer 094 (Wasson and Rubin 2010). The low analytical totals in LAP 02342 appear to result largely from the presence of hydrous phases, whereas low totals in Acfer 094 are probably attributable to fine cracks and surface roughness. Although Vigarano and Acfer 094 are both dry, the higher analytical totals in the Vigarano matrix may have resulted from a late impact that led to additional compaction (i.e., reduced porosity). This may have been particularly effective in the “playa-like” regions that have the finest grain sizes.

Within the grid areas, the average molar (Fe + Mg)/ Si ratio is approximately 2.24 ± 0.14; it is roughly constant over all grid areas and 12% higher than that in olivine. Because the S content of the fines is only approximately 3 mg g−1, FeS can account only for approximately 8 mg g−1 of the matrix Fe (about 2.5% of the total). Our electron-probe beam did not sample any single phase magnetite or metal grains in our grid areas indicating that the mean grain size of such phases in the matrix is ≤2 μm.

Comparison of Matrix and Whole-Rock Compositions

In Fig. 11, we plot the matrix-grid elemental means normalized to the mean CV composition of Wasson and Kallemeyn (1988). To simplify the discussion, for each element, we have plotted the areas with fine textures on the left (with light fills); areas with coarse textures in the middle (with dark fills); and intermediate, ambiguous or anomalous areas on the right (with stick symbols ×, +, *). These diagrams offer an overview of the variation in abundances of the individual elements.

Figure 11.

 Whole-rock- and Mg-normalized abundance ratios of 10 elements in the 13 grid areas of Vigarano matrix. (a) Areas D and F. (b) Area G. The general pattern is similar to that observed in CR2 LAP 02342 with the major exception of the volatile elements Na, K, and S, which have lower abundance ratios than in LAP 02342. In addition, the concentrations of Al and Ca are lower, and Fe and Mn are higher in the Vigarano matrix than in the matrix of LAP 02342. Fine regions of Vigarano include D5-6, F7-1, G3-2, and G3-6; coarse regions include D5-2, D5-4, and G3-3; intermediate, ambiguous, or anomalous regions include D5-1, D5-3, D5-5, G3-1, G3-4, and G3-5.

In Fig. 11a, we plot abundance ratios of the six matrix-grid areas in D5 and of matrix area F7-1. The volatile elements Na and S are low; K is near CV whole-rock levels. Nevertheless, all three volatile elements have much lower whole-rock-normalized abundance ratios than in CR LAP 02342 (fig. 17 of Wasson and Rubin 2009). If we disregard Na, K, and S, we find that the overall pattern in Fig. 11 is typical of that observed in matrix regions of anhydrous carbonaceous chondrites (e.g., Al and Si ratios are slightly above unity, Fe and Mn ratios are further above unity, and Ca ratios are low). As also discussed above in connection with the concentration data, volatiles are generally lower in fine-textured areas compared with the coarse area; however, Na and K abundance ratios in F7-1 are about the same as those in coarse-area D5-2. Abundance ratios in the fine-textured areas for Fe and Mn are higher and those for Al are lower or the same as those in the coarse areas.

In Fig. 11b, abundance ratios for the six grid areas in G3 are depicted. With a couple of exceptions, the same generalizations hold as just discussed for Fig. 10. The differences are (1) that, in fine area G3-2, the Al ratio is below unity and that of Si is unity, (2) Mn abundances in fine areas are slightly below that in the coarse area, and (3) Na abundances in G3-2 and G3-6 are considerably lower than those in F7-1 and D5-6.

In summary, the patterns observed in Fig. 11 are typical of matrix in primitive chondrites with the exception of the volatile depletion. The differences between fine- and coarse-textured materials (the lower concentrations of Al, Na, K, and S in the fine materials) are important. The best explanation seems to be that the coarse materials are diluted by small amounts of comminuted chondrules. If this is correct, the chondrules seem to have had slightly enhanced Al values relative to nebular fines. This could suggest an enrichment in chondrule mesostasis, which would require a separate explanation. These compositional anomalies warrant further study.

Limited Information about Heterogeneity in Nebular Fines

Wasson and Rubin (2009) interpreted the variations among matrix-grid areas in terms of heterogeneities within the nebula; they suggested that weakly bonded clusters of fine materials were formed during chondrule-melting events in which the degree of melting was low, perhaps 20%. They proposed that the carriers resembled chondrules in having a moderate range in composition. An initial goal of this study was to search for evidence of nebular heterogeneity in Vigarano matrix.

Unfortunately, as discussed above, several features of the Vigarano matrix seem best explained by secondary alteration on the parent asteroid, alteration that involves both aqueous and thermal processes. Of special importance in the LAP 02342 CR chondrite was the enhanced concentrations of volatiles in matrix regions; these are not observed in Vigarano. Because of the pervasiveness of the alteration, it does not seem worthwhile to try to infer details of the nebular signatures in the different grid areas.

Aqueous and Thermal Alteration

Although Weisberg and Prinz (1998) doubted that parent-body hydration was important in any of the CV3 chondrites, most workers have concluded that the oxidized CV chondrites experienced appreciable aqueous alteration (e.g., Krot et al. 1998). Reduced CV chondrites largely escaped this alteration and previous studies of Vigarano matrix showed that the reduced lithology of Vigarano experienced little aqueous alteration (Zolensky et al. 1993). Nevertheless, Tomeoka and Tanimura (2000) studied chondrule rims in Vigarano and showed that there was intense parent-body hydration and subsequent aqueous alteration that was highly localized. MacPherson and Krot (2002) reported that portions of Vigarano contain such secondary phases as hedenbergite, andradite, and kirschsteinite in the matrix and at the surfaces of Wark-Lovering rims, accretionary rims, and CAIs. Because Vigarano’s mean porosity appears to be higher than those of other reduced CV chondrites (Macke et al. 2011), it seems likely that Vigarano is more altered than other reduced CV chondrites. Its alteration was caused by mobilized water that seeped into voids (MacPherson and Krot 2002; Rubin forthcoming).

Elements sensitive to aqueous alteration (by association with soluble phases such as CaSO4, KCl, and NaCl) generally show little correlation in Vigarano. Grid area F7-1 is the exception; it shows very significant, positive correlations among Ca and S, Na, and K, and Na and Ca (e.g., Fig. 8). These four elements appear to be correlated, suggesting that they were all affected by the same parent-body process. This is plausibly mobilization by aqueous fluids.

Magnetite, an important indicator of aqueous effects in oxidized CV chondrites (Housley and Cirlin 1983; Keller et al. 1994), is nevertheless also present in reduced CV chondrites. Most researchers infer that chondritic magnetite was formed during asteroidal alteration rather than in the solar nebula (e.g., Ramdohr 1963; McSween 1977; Krot et al. 1995; Choi et al. 1998; Greenwood et al. 2000). Magnetite is a minor constituent of Vigarano matrix (Krot et al. 1995). If alteration were advanced, one would expect to see coarse magnetite grains in the matrix, but none were found in our grid analyses.

We suggest that aqueous alteration played an important role in the formation of Vigarano matrix. In particular, the textures of the fine-matrix “playa” regions seem consistent with formation in an aqueous environment. The fact that Vigarano is now largely anhydrous implies that either there was relatively little water or that the mild thermal metamorphism experienced by Vigarano occurred after aqueous alteration, driving the water off (e.g., Krot et al. 1995). The high totals that we find in the fine regions suggest that the compaction event occurred after (or possibly during) mild metamorphic reheating.

It is unclear what fraction of the matrix heterogeneity originated in the nebula and what fraction was produced during secondary alteration. Because S is particularly sensitive to alteration effects (e.g., Grossman and Brearley 2005) and can be redistributed by shock (e.g., Nakamura et al. 2000), its pronounced absence from Vigarano matrix is a strong indicator that the study areas have not preserved their original compositions. It appears impossible to separate the relative contributions of these two sources.

Conclusions

We used the matrix-grid technique to analyze matrix in CV3 Vigarano; we found that the ca. 50 × 50 μm regions show sizable differences in composition and texture. The regions with coarser texture (and higher porosity) showed lower Fe contents and lower analytical totals. We tentatively attribute this correlation to a larger dilution by comminuted chondrules (which mostly have lower Fe contents) in the coarser region, with the low totals largely reflecting higher porosity.

The overall compositional trend in Vigarano matrix is similar to that observed in other carbonaceous chondrites; relative to whole rock, Al, Si, Fe, and Mn abundances are high. The key difference compared with primitive chondrites such as CR2 LAP 02342 (Wasson and Rubin 2009) and type-3.0 C-ungrouped Acfer 094 (Wasson and Rubin 2010) is that the volatile elements Na and S have very low abundance ratios in the Vigarano matrix. The K abundance ratio is also lower than in these more primitive chondrites, but shows large variations among grid areas. Another difference is that the mean analytical totals are much higher in Vigarano, approximately 91.5 wt% versus approximately 82 wt% in LAP 02342 and approximately 84 wt% in Acfer 094.

In the study of matrix in LAP 02342, Wasson and Rubin (2009) observed that each matrix-grid region had its own compositional signature; with one exception, all were resolvable from other regions. These differences were attributed to dilution by fragile nebular fine-grained objects during impact compaction on the asteroid. Although we observed significant heterogeneity in the matrix of Vigarano, many of the differences appear to be produced mainly by secondary alteration. In the most primitive chondrites, S is the best tracer of matrix (e.g., Wasson 2008). However, in Vigarano, S is not mostly in matrix, but instead occurs in irregularly distributed grains typically associated with chondrules or chondrule fragments.

We suggest that aqueous processes played an important role in the mobilization of Na and S and in the transportation of these elements out of the matrix. The present “dry” condition of Vigarano would then reflect water loss during metamorphic reheating, plausibly the result of impact heating. Although the studied section of Vigarano shows a number of largely intact chondrules, there is strong evidence of impact effects. The interchondrule region is littered with fragments 10–30 μm in size. Other samples of Vigarano show brecciation on a larger scale.

Acknowledgments

Acknowledgment— We are grateful to Frank Kyte for teaching S. M. Hurt how to use the electron microprobe. We thank the Smithsonian Institution for the loan of the Vigarano sections. The paper benefited greatly from the reviews of M. E. Zolensky, an anonymous referee, and Editor A. J. T. Jull. This research was mainly supported by NASA grants NNG06GG35G and NNX10AG98G (JTW) with additional support from NASA grant NNG06GF95G (AER).

Editorial Handling— Dr. A. J. Timothy Jull

Ancillary