Characterization of matrix material in Northwest Africa 5343: Weathering and

thermal metamorphism of the least equilibrated CK chondrite

Tasha L. DUNN 1*, Oriana K. BATTIFARANO1, Juliane GROSS2,3, and Emma J. O’HARA1

1Department of Geology, Colby College, Waterville, Maine 04901, USA2Department of Earth and Planetary Sciences, Rutgers University, Piscataway, New Jersey 08854, USA

3Department of Earth and Planetary Sciences, American Museum of Natural History, New York, New York 10024, USA*Corresponding author. E-mail: tldunn@colby.edu

(Received 17 October 2017; revision accepted 24 April 2018)

Abstract–Based on the chemical heterogeneity of chondrule and matrix olivine, NorthwestAfrica (NWA) 5343 is the least metamorphosed CK chondrite reported so far. To betterconstrain the lower limit of metamorphism in the CK chondrites, we performed a detailedanalysis of matrix material in NWA 5343, including characterization of the texture and bulkcomposition and analyses of individual silicate minerals. Results suggest that NWA 5343 ispetrologic type 3.6 or 3.7. Although silicate minerals in the matrix seem to be equilibratedto roughly the same extent throughout the sample, there are recognizable differences ingrain size and shape. These textural differences may be the result of transient heating eventsduring impacts, which would be likely on the CK chondrite parent body. The differencebetween the extent of chemical equilibration and texture may also suggest that grain sizeand shape are still sensitive to metamorphism at petrologic subtypes where silicate mineralequilibration is nearly complete (e.g., >3.7). Carbonate material present in NWA 5343 is aproduct of terrestrial weathering; however, infiltration of a Ca-bearing fluid did notinfluence the composition of silicate minerals in the matrix. To evaluate the possibility of acontinuous metamorphic sequence between the CV and CK chondrites, the bulk matrixcomposition of NWA 5343 is compared to the CVred chondrite, Vigarano. Although thematrix composition of NWA 5343 could be derived by secondary processing of a Vigarano-like precursor, porosity and texture of matrix olivine in NWA 5343 are hard to reconcilewith a continuous metamorphic sequence.

INTRODUCTION

The Karoonda-like (CK) chondrites are a group ofhighly oxidized carbonaceous chondrites characterizedby the presence of nickel-rich matrix olivine (>0.2 wt%NiO) and Cr2O3-rich magnetite (>3 wt% Cr2O3)(Kallemeyn et al. 1991; Noguchi 1993; Geiger andBischoff 1995). The CK chondrites contain abundantmatrix material and are dominated by porphyritic, Fe-rich olivine chondrules (Kallemeyn et al. 1991; Rubin2010; Chaumard and Devouard 2016). Bulk lithophileelemental abundances and oxygen isotopes of the CKchondrites are similar to those of the Vigarano-like(CV) chondrites (Greenwood et al. 2010). Thesesimilarities led Greenwood et al. (2010) to suggest that

the CV and CK chondrites originated from a single,thermally stratified (i.e., onion-shell) parent asteroid.The single parent body model has since gained support,though alternatives to the onion-shell model have beenproposed (e.g., Chaumard et al. 2012, 2014; Wassonet al. 2013; Chaumard and Devouard 2016). However,Dunn et al. (2016a) contended that the single parentbody model is unlikely, and they showed thatcompositions of magnetite in the CVs and CKs couldnot be reconciled by a continuous metamorphicsequence (i.e., CV3 ? CK3 ? CK4-6). Most recently,Yin et al. (2017) also argued against a single parentbody model based on Cr isotopic compositions, whichare clearly distinguishable between the CV and CKchondrites.

Meteoritics & Planetary Science 1–16 (2018)

doi: 10.1111/maps.13118

1 © The Meteoritical Society, 2018.

One significant difference between the CV and CKchondrite groups is the degree of metamorphismexperienced by each. All CV chondrites are classified aspetrologic type 3 (McSween 1977; Bonal et al. 2006),while CK chondrites exhibit the full range of thermalmetamorphism, from type 3 to type 6 (Kallemeyn et al.1991; Geiger et al. 1993). Most CK chondrites areequilibrated (petrologic types 4–6). However, 35unequilibrated (type 3) CK chondrites have beenidentified since the first CK3 chondrite, Watson 002, wasdiscovered in 1991 (Geiger et al. 1993) (see theMeteoritical Bulletin for a current list of CK3s). Inmeteorites that have experienced a low degree ofmetamorphism (i.e., petrologic type 3), subtle differencesin the extent of thermal processing can be recognized.Resultantly, meteorites of this grade can be furtherdivided into petrologic subtypes (i.e., type 3.0–3.9) (Searset al. 1980, 1991; Chizmadia et al. 2002). It has beensuggested that the unequilibrated CKs are all petrologictype 3.7 or higher based on abundances andcompositions of refractory inclusions (Greenwood et al.2010), and six unequilibrated CK chondrites have beenassigned petrologic type 3.8 or 3.9 (Chaumard et al.2014; Dunn et al. 2016a, 2016b). In contrast, determiningthe petrologic subtypes of the CV chondrites iscomplicated by their complex alteration histories, whichinvolve fluid-assisted metamorphism and metasomatism(Brearley and Krot 2013). Although approaches based onpetrography are fairly consistent (suggesting that the CVsare petrologic subtype 3.3 or lower) (Guimon et al. 1995;Grossman and Brearley 2005), Raman spectroscopy oforganic material yields petrologic subtypes ≥3.6 for theCV chondrites Allende, Axtel, Bali, Mokoia, andGrosnaja (Bonal et al. 2006). It has been suggested thatthese inconsistencies are due to decoupling of organicmatter and silicates during metamorphism (Pearson et al.2007). Despite the uncertainty, the general consensus isthat CV chondrites have experienced lower degrees ofmetamorphism (≤ subtype 3.6) than the unequilibratedCK chondrites (≥ subtype 3.7).

In our efforts to determine the petrologic subtypesof the unequilibrated CK chondrites, we have analyzedchondrule and matrix olivine in several CK3 and CK4chondrites (Dunn et al. 2016b). Consistent withprevious observations (e.g., Chaumard and Devouard2016), the homogeneity and chemical composition ofchondrule olivine suggests that most of these CK3s havebeen metamorphosed to at least petrologic subtype 3.8conditions. However, heterogeneity of olivine inNorthwest Africa 5343 (NWA 5343) suggests that it isless metamorphosed than other CK chondrites(Chaumard and Devouard 2016; Dunn et al. 2016b).Accurate determination of petrologic subtypes is criticalwhen assessing the onion-shell model. Therefore, to

better constrain the lower limit of metamorphism in theCK chondrites, we performed a detailed analysis ofmatrix material in NWA 5343. Here we use “matrix” torefer to all materials that occupy the space betweenchondrite components (e.g., chondrules). Following theprocedures of Hurt et al. (2012), we characterized thetexture and bulk composition of matrix. We alsoanalyzed individual silicate minerals (olivine, pyroxene,and plagioclase), as these phases have been shown torecord the effects of metamorphism in unequilibratedordinary chondrites (e.g., Dodd and Van Schmus 1967;Huss et al. 1981). Finally, to assess possible links withthe CV chondrites, we have compared the bulk matrixcomposition of NWA 5343 to that of CVred chondriteVigarano.

METHODOLOGY

The polished thin section examined in this studywas prepared from a 36.7 g slice of NWA 5343, whichwas purchased from Mirko Graul in 2014. This thinsection (NWA 5343_01) and the remaining sample massare housed in the meteorite collection at Colby College.Backscattered electron (BSE) images were obtainedusing the JXA JEOL 8200 electron microprobe (EMP)at Rutgers University. These images were used todetermine the textural characteristics of the sample andto measure porosity. We measured grain sizes andcalculated porosity using the image processing softwareImageJ. More than 300 grains were measured (along thelong axis) throughout the sample. Porosity wasdetermined by thresholding an image to highlight thepore space, converting the images to black and white,and then calculating the thresholded area of the image.

Backscattered electron images were compiled tomake a mosaic image of the entire thin section (Fig. 1).We superimposed a 1 mm by 1 mm grid, with rowslabeled 1–24 and columns labeled A–X, onto the mosaicBSE image (data not shown on Fig. 1) and selected fiveareas (L14, I21, K5, S6, and U12) for additional study.The selected areas are illustrated on Fig. 1, and BSEimages of each of the five areas are provided in thesupporting information Figs. S1 and S2. Wavelength-dispersive (WDS) and energy-dispersive (EDS)elemental Ka X-ray intensity maps for Fe, Al, Ca, S,and Mg (WDS) as well as Si, K, Na, and Cr (EDS)were obtained using the EPMA JEOL 8200 at RutgersUniversity. These maps were collected using 15 kVaccelerating voltage, 25 nA beam current, 1 lm focusedbeam, a pixel step size of 2 lm, and a dwell time of55 ms. X-ray maps were used to identify mineral phasesin regions S6, K5, U12, L14, and I21.

Quantitative mineral chemical compositions werealso determined using the JEOL 8200 EMP. Operating

2 T. L. Dunn et al.

conditions were: 15 kV accelerating voltage; 20 nAbeam current; focused electron beam (1 lm in size) forolivine, pyroxene, and magnetite; defocused electronbeam (3 lm) for plagioclase; and peak and backgroundcounting times of 10–40 s per element. Analyticalstandards used to determine both mineral compositionsand bulk matrix composition were well-characterizedsynthetic oxides and natural minerals including albite(Si, Na), olivine (Mg, Fe), Ni-olivine (Ni), orthoclase(K), plagioclase (Al), chromite (Cr), troilite (S),anorthite (Ca), rutile (Ti), and rhodonite (Mn). Dataquality was ensured by analyzing standard materials asunknowns.

Bulk composition of the matrix was determinedusing grid analyses similar to that implemented byWasson and Rubin (2009) and Hurt et al. (2012). Inthis technique, several square grids measuring 150 by150 lm were analyzed within the regions S6, K5, U12,L14, and I21. Grid areas are highlighted on the BSEmaps provided in the supporting information. Each gridarea was analyzed with a step size of 10 lm betweeneach point and background counting times of 10–40 sper element. We used standard, fully matrix-corrected,wavelength-dispersive analyses. For further discussion,see Wasson and Rubin (2009). We used 10 lm beam at15 nA to measure 11 elements: Na, Mg, Al, Si, S, K,Ca, Cr, Mn, Ni, and Fe. Detection limits were 0.01

wt% for Al, Mg, Si, Cr, S, K, Ti, and Ca; 0.02 wt% forNa, Ni; 0.05 wt% for Mn; and 0.08 wt% for Fe. Dataquality was ensured by analyzing standard materials asunknowns.

RESULTS

Petrography of the Matrix

Two distinct regions of matrix are visible in theBSE image of the thin section (Fig. 1). In the firstregion, located predominately in the southwest quadrantof the thin section, the matrix is light gray andcontinuous. Matrix material in the second region,located primarily in the northwest quadrant, is darkerand more porous. There is a sharp contact betweenthese two matrix regions (Fig. 1 insert). We will refer tothese regions henceforth as the BSE-light matrix and theBSE-dark matrix. Detailed inspection of areas withinthe BSE-light and BSE-dark matrix indicates that thereare noteworthy textural differences between the tworegions. BSE-dark matrix material (areas K5, S6, andU12) mainly consists of subhedral to anhedral grains ofolivine with an average grain size of 19.5 � 16.5 lm(Fig. 2). Minor, euhedral, high-Ca pyroxene(16.1 � 9.5 lm) and euhedral plagioclase grains(18.3 � 8.6 lm) are also present; however, olivine is the

Fig. 1. Mosaic backscattered electron (BSE) image of NWA 5343 showing the five areas selected for detailed textural andcompositional study. These areas represent boxes on a 1 9 1 mm grid, with rows labeled 1–24 and columns labeled A–X (datanot shown). The insert illustrates the sharp boundary (dashed line) between the dark-colored matrix material (top) and the light-colored matrix material (bottom).

Weathering and metamorphism of NWA 5343 3

dominant phase. Grains are separated by pore space,which is sometimes filled with submicron-sized silicatesand oxides (Fig. 2). Porosity of areas within the darkermatrix ranges from 10 to 14%. Much of the BSE-darkmatrix is typical of a breccia and is similar to texturesseen in the CV chondrite Vigarano (Abreu and Brearley2011). We also identified a 1.5 9 3 mm area within theBSE-dark region that is characterized by smaller(<10 lm), more angular olivine grains intermixed with afew large olivine clasts (90–100 lm) (Fig. 3). This areamost likely represents a crushed chondrule rather thanactual matrix material. Fragmental matrix textures arecommon in the CK chondrites and have been attributedto impact-induced crushing events prior tometamorphism (Noguchi 1993; Wasson et al. 2013).Some of the larger grains adjacent to the crushedchondrule exhibit characteristic 120° triple junctionsthat are attributed to thermally driven metamorphism(Fig. 3 insert).

The BSE-light matrix (areas I21 and L14) consistsmostly of anhedral olivine grains surrounded by aninterstitial Ca-rich material that is responsible for thelight color of this area in BSE images (Fig. 4). Thismaterial, identified as calcium carbonate using EDS, isa common terrestrial weathering product in desertregions. The source of this interstitial material is a largevein (150–300 lm wide) in area L14 (Fig. 5). Skeletalpyroxene, more variable in size than the olivine(18.3 � 13.1 lm), and large (55.7 � 21 lm), subhedralplagioclase grains are also present in CaCO3 vein(Fig. 5). We also observed a few small apatite crystalswithin the vein (not shown in Fig. 5) that most likelyrepresent an alteration product. Porosity in this regionranges from 3 to 7%. When carbonate is present in the

pore spaces, it often contains magnetite grains that areslightly larger (1–3 lm) than the magnetite in the BSE-dark matrix. Olivine in the BSE-light matrix is alsolarger (30.5 � 14.9 lm) and more anhedral than olivinein the BSE-dark matrix. Despite textural differences,olivine grains in both the BSE-dark and BSE-lightmatrix contain submicron voids and oxides similar tothose described by Brearley (2009) (Fig. 6).

Bulk Composition of the Matrix

Rules for Discarding PointsIn previous studies of extremely fine-grained

primitive nebular matrix material in CR and CVchondrites (Wasson and Rubin 2009; Hurt et al. 2012),analyses with extreme values were discarded to avoidthe contributions of coarse nonnebular mineral grains.However, here our aim was to study the coarse,recrystallized matrix material, and even the finestgrained material present in some interstices in the BSE-dark matrix appears to be recrystallized. Thus, we basedour rules for discarding points solely on oxide totalsand discarded all analyses with totals <85 wt% underthe assumption that totals below this thresholdrepresented voids rather than mineral phases. Ingeneral, areas in BSE-dark matrix had a higherproportion of discarded points (54–68%) than areas inthe BSE-light matrix (33–38%). The number ofdiscarded points is consistent with differences inporosity, which ranges from 10 to 14% in the BSE-darkmatrix and from 3 to 7% in the BSE-light matrix. Thelower porosity in the BSE-light matrix is an artifact ofthe interstitial calcium carbonate material. Thus, theporosity of the BSE-light matrix originally would have

Fig. 2. Representative areas of the BSE-dark matrix. This matrix is dominated by subhedral to anhedral olivine (olv), mostly< 30 lm in size. Some euhedral high-Ca pyroxene (cpx) and plagioclase (plag) is also present.

4 T. L. Dunn et al.

likely been closer to that of the BSE-dark matrix priorto terrestrial alteration.

Compositional VariabilityInsights into the compositional variability of the

bulk matrix in NWA 5343 can be examined usingscatter diagrams of major and minor oxides. Bulkcompositions of BSE-light and BSE-dark matrix ineight representative grids are plotted on Fig. 7. Because

the matrix of NWA 5343 is relatively coarse-grained,compositional trends are driven by mineralogy. Asillustrated Figs. 7a and 7b, olivine is the dominantphase present, followed by high-Ca pyroxene. The effectof the interstitial CaCO3 material on the bulkcomposition of the BSE-light matrix is illustrated inFig. 7b by the negative correlation between SiO2 andCaO. Data points with a higher CaO content reflect ahigher proportion of CaCO3 material in an individual

Fig. 3. Crushed chondrule (1.5 9 3 mm) in the BSE-dark matrix. Scale bar is equal to 100 lm. The insert illustrates the triplejunctions (TJ) in area K5-2 (see supporting information). Scale bar of insert is 10 lm.

Fig. 4. Representative regions of the BSE-light matrix. Matrix is dominated by anhedral olivine (olv), which is surrounded byinterstitial calcium carbonate (CaCO3). Skeletal high-Ca pyroxene (cpx) and micron-sized magnetite (mag) are also present. Someeuhedral plagioclase is also present in this matrix (not pictured above).

Weathering and metamorphism of NWA 5343 5

analysis (Fig. 7b). Some plagioclase is also present inthe matrix, though it is a minor phase compared toolivine and pyroxene (Fig. 7c). Sulfur and NiO are bothpresent at low abundances (<1 wt%), suggesting thatpentlandite is not a prominent matrix phase. Asillustrated in Fig. 7d, most bulk matrix compositions

plot parallel to, but below, the pentlandite trend line,suggesting that there is an additional source of Ni in thematrix, most likely olivine.

Mean CompositionsMean composition of each area reflect its dominant

mineralogy. Because mineralogy is relativelyhomogenous throughout the sample, averagecompositions of most grids cannot be resolved from oneanother when standard deviations are considered. GridsL14-1 and L14-2 are the only exceptions, as they havemineralogies dominated by high-Ca pyroxene andCaCO3 material rather than olivine. Mean compositionsof all grids, plotted as MgO versus FeO, Al2O3 versusCaO, MnO versus Cr2O3, and NiO versus S, areprovided in Figs. S1 and S2. Table 1 lists the meancomposition of each grid in areas U12, K5, S6, L14,and I21, and the mean composition of the BSE-lightand BSE-dark matrix. Because the pyroxene-dominatedmineralogy of grids L14-1 and L14-2 is notrepresentative of the overall matrix mineralogy, we haveexcluded these grids from the calculated mean of theBSE-light matrix. With the exception of CaO, Al2O3,and Cr2O3, the mean compositions of the BSE-dark andBSE-light matrix are nearly identical. The mostsignificant difference is CaO content, which is 2.29higher in the BSE-light matrix due to the presence of

Fig. 5. Calcium carbonate weathering vein in area L14. The insert is a close-up of the vein material, highlighting the skeletalpyroxene (cpx), euhedral plagioclase (plag), and anhedral olivine (olv).

Fig. 6. BSE image of matrix in the BSE-light regionhighlighting the “dusty” texture of the olivine. This texture isalso present in the BSE-dark matrix. Olivine containssubmicron voids and inclusions of magnetite and sulfides.

6 T. L. Dunn et al.

the CaCO3 weathering material in this region.Conversely, Al2O3 and Cr2O3 are higher in the BSE-dark matrix (1.29 and 0.669 higher, respectively), mostlikely due to compositional differences between high-Capyroxene in the BSE-dark matrix and the BSE-lightmatrix (see the Discussion section).

Individual Mineral Phases

Representative compositions of silicate minerals inthe BSE-dark and BSE-light matrix are reported in

Table 2. The Fa content of matrix olivine is plotted as ahistogram in Fig. 8, and compositions of high-Capyroxenes are plotted in Fig. 9.

OlivineMatrix olivine ranges in composition from Fa30 to

Fa38 in the BSE-dark matrix and from Fa35 to Fa40 in theBSE-light matrix. The average composition of olivine isthe same in both the matrix regions (Fa36.3 and Fa36.2,respectively). As is typical for CK chondrites, matrixolivine is nickel-rich (average of 0.31 wt% NiO in BSE-

0

5

10

15

20

25

30

35

0 10 20 30 40 50 60

MgO

(wt%

)

SiO2 (wt%)

olivine

cpx

(a)

0

5

10

15

20

25

30

35

40

0 10 20 30 40 50 60

CaO

(wt%

)

SiO2 (wt%)

I21-1I21-3L14-2L14-3K5-1K5-2S6-1U12-2

olv

cpx

plag

BSE-dark

BSE-light

(b)

0

5

10

15

20

25

30

0 2 4 6 8

Al2O

3(w

t%)

Na2O (wt%)

(c)

plagioclase

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.0 0.2 0.4 0.6 0.8

S (w

t%)

NiO (wt%)

Pentlandite(d)

Fig. 7. Bulk compositions (wt%) of eight representative areas of the matrix plotted as (a) SiO2 versus MgO, (b) SiO2 versusCaO, (c) Na2O versus Al2O3, and (d) NiO versus S. Analyses often plot along mineral trend lines, illustrating the dominantmineralogies of the areas. )

Weathering and metamorphism of NWA 5343 7

dark; 0.29 wt% NiO in BSE-light). NiO content ofolivine is more homogeneous in the BSE-light matrix(0.24–0.48 wt%) than in the BSE-dark matrix (0.23–0.67wt%). CaO is also slightly more homogenous in the BSE-light matrix. There is no discernible difference betweenaverage abundances and heterogeneity of MgO, SiO2,and MnO in the BSE-light and BSE-dark matrix.Compositions of matrix olivine in NWA 5343 areconsistent with other CK3 chondrites we have analyzedin our previous work (Dunn et al. 2016b).

PyroxeneAll pyroxene in the matrix of NWA 5343 is diopside

(En32–54Wo46–50) (Fig. 9). This is consistent with previous

studies of CK chondrites, in which diopside is thedominant, or only, pyroxene identified (e.g., Kallemeynet al. 1991). The average composition of diopside is thesame in both the matrix regions (En40Fs11Wo49) and isidentical to that of the CK3 chondrite NWA 1628(En41Fs10Wo49) (Brearley 2009). The Fs content of high-Ca pyroxene in NWA 5343 is also consistent with theCK4 chondrites Karoonda, Maralinga, and Dhofar 015(Fs10–11) (Noguchi 1993; Ivanova et al. 2000). Theheterogeneity of high-Ca pyroxene is the same inBSE-light and BSE-dark material (Fs6–18) and is similarto high-Ca pyroxene in the CK3 chondrite Watson 002(Fs7–20) (Geiger et al. 1993). Also, like Watson 002,pyroxene in NWA 5343 is more heterogeneous than

Table 1. Average bulk compositions (wt%) and standard deviations (in italics) of grids in all five areas of thematrix.

Points Na2O Al2O3 MgO SiO2 NiO Cr2O3 S K2O CaO FeO MnO Total

U12-1 77 0.05 1.49 21.33 31.10 0.40 0.30 0.16 0.02 2.78 30.62 0.13 88.37BSE-dark 0.07 0.89 3.54 6.31 0.10 0.21 0.11 0.03 6.12 7.88 0.06 2.99

U12-2 71 0.03 2.33 22.14 30.67 0.41 0.59 0.11 0.02 1.18 30.96 0.14 88.56BSE-dark 0.04 2.50 3.39 3.59 0.09 0.77 0.08 0.03 2.79 3.62 0.06 3.06K5-1 107 0.05 2.52 23.65 30.83 0.30 0.42 0.14 0.07 1.36 28.29 0.16 87.79BSE-dark 0.06 1.52 2.43 2.16 0.08 0.27 0.15 0.11 2.28 2.89 0.05 2.92

K5-2 116 0.05 2.27 22.78 31.78 0.31 0.40 0.10 0.05 1.73 27.56 0.19 87.21BSE-dark 0.05 1.17 2.18 2.30 0.05 0.22 0.11 0.07 2.45 2.82 0.11 2.54K5-3 159 0.22 1.90 21.97 34.89 0.32 0.27 0.06 0.04 3.87 30.52 0.16 94.38

BSE-dark 1.03 3.54 6.80 9.33 0.14 0.56 0.12 0.13 7.64 12.18 0.09 4.57S6-1 111 0.35 2.58 20.69 31.31 0.39 0.31 0.08 0.05 0.68 32.69 0.15 89.29BSE-dark 1.27 4.20 4.53 4.45 0.08 0.32 0.08 0.06 0.85 5.37 0.05 3.07

S6-2 115 0.03 1.22 23.29 31.95 0.40 0.18 0.10 0.03 0.63 32.78 0.17 90.78BSE-dark 0.03 0.89 2.46 2.72 0.07 0.15 0.08 0.04 1.28 2.14 0.05 3.34S6-3 169 0.03 1.08 25.57 32.66 0.41 0.25 0.09 0.03 0.39 34.67 0.17 95.54

BSE-dark 0.05 2.55 5.70 6.75 0.23 0.60 0.27 0.17 0.55 6.55 0.08 4.56L14-1 87 0.11 0.46 11.78 43.52 0.03 0.03 0.04 0.00 27.31 6.91 0.04 90.22BSE-light 0.10 0.47 2.06 4.89 0.05 0.04 0.05 0.01 3.27 2.72 0.04 3.71L14-2 104 0.34 1.37 11.88 44.08 0.02 0.05 0.04 0.02 26.73 6.31 0.05 90.87

BSE-light 0.98 3.93 3.04 5.52 0.06 0.17 0.06 0.04 5.57 4.81 0.04 4.24L14-3 162 0.02 0.80 24.00 30.48 0.39 0.24 0.21 0.04 6.24 29.14 0.17 91.73BSE-light 0.03 0.78 4.04 4.26 0.12 0.28 0.13 0.08 5.25 3.98 0.06 4.10

L14-4 149 0.02 0.91 23.97 30.42 0.43 0.24 0.15 0.04 2.78 32.38 0.18 91.51BSE-light 0.02 1.23 3.33 3.89 0.09 0.32 0.13 0.08 3.34 3.80 0.05 3.87I21-1 161 0.04 0.62 22.77 33.67 0.35 0.17 0.14 0.02 7.52 27.23 0.16 92.69

BSE-light 0.06 0.63 5.32 6.36 0.14 0.19 0.14 0.03 8.48 8.73 0.07 4.09I21-2 152 0.04 0.77 22.49 33.51 0.30 0.18 0.11 0.02 8.97 26.25 0.16 92.79BSE-light 0.06 1.11 5.78 6.60 0.13 0.28 0.11 0.05 9.02 8.93 0.07 4.56

I21-3 169 0.02 1.23 24.14 30.98 0.38 0.26 0.11 0.03 2.47 32.86 0.20 92.67BSE-light 0.04 1.49 3.58 4.38 0.07 0.36 0.11 0.06 3.57 3.75 0.05 4.26I21-4 163 0.01 0.74 25.02 32.02 0.36 0.18 0.08 0.02 3.54 30.75 0.17 92.88BSE-light 0.02 0.83 3.51 3.99 0.08 0.20 0.06 0.03 4.85 3.69 0.05 4.27

Avg. BSE-dark 0.09 1.92 22.57 31.79 0.37 0.35 0.11 0.04 1.62 30.99 0.16 90.040.29 2.11 3.83 4.73 0.10 0.40 0.12 0.07 3.16 5.47 0.07 3.34

Avg. BSE-lighta 0.03 0.85 23.73 31.84 0.37 0.21 0.13 0.03 5.25 29.77 0.17 92.38

0.04 1.01 4.26 4.91 0.11 0.27 0.12 0.06 5.75 5.48 0.06 4.19aCompositions of L14-1 and L14-2 are excluded from the average bulk composition of the BSE-light matrix, as both areas are unrepresentative

of the overall matrix.

8 T. L. Dunn et al.

olivine (Geiger et al. 1993). However, this is notunexpected, as olivine equilibrates faster than pyroxeneduring metamorphism (e.g., Huss et al. 1981).

Abundance and heterogeneity of minor elements inpyroxene is also similar in both regions of the matrix.Cr2O3 and NiO are typically below detectable limits,though a few grains contain measurable amounts (up to0.13 wt% Cr2O3 and 0.26 wt% NiO). Na2O is present atabundances from 0.03 to 0.59 wt%, with an average of0.20 wt% in the BSE-dark and 0.11 wt% in the BSE-light. Published analyses of diopside in the CK3chondrites are lacking. However, abundances of minorelements in high-Ca pyroxene are lower in NWA 5343than in CK4 chondrites Karoonda and Maralinga(Noguchi 1993). The most remarkable difference is in

Al2O3, which is present at average abundances of 1.66and 1.83 wt% in Karoonda and Maralinga, respectively.The average abundance of Al2O3 in NWA 5343 is 0.58wt% in the BSE-dark matrix and 0.27 wt% in the BSE-light matrix.

PlagioclasePlagioclase in NWA 5343 is Na-rich, with

compositions of An6–45Ab54–91Or1–4 in the BSE-darkand An14–40Ab55–84Or1–2 in the BSE-light. This isconsistent with matrix plagioclase in CK3 chondritesWatson 002 (An7–50) (Geiger et al. 1993) and NWA1628 (An17–37) (Brearley 2009). Consistent with previousstudies, plagioclase in NWA 5343 is enriched in FeO.

Table 2. Representative compositions of olivine, high-Ca pyroxene, and plagioclase in the matrix.

Olivine High-Ca pyroxene Plagioclase

K5, 1 U12, 5 L14, 66 S6, 35 U12, 42 L14, 7 K5, 25 S6, 43 U12, 20

MgO 31.5 30.6 31.1 11.8 15.4 15.0 nm nm nm

Al2O3 0.02 0.26 <0.01 0.09 0.90 0.76 23.6 22.8 25.7SiO2 36.6 35.9 36.5 52.6 52.7 53.0 61.1 62.6 58.7Na2O nm nm nm 0.07 0.21 0.21 8.39 9.48 7.27CaO 0.08 0.06 0.10 24.4 24.6 24.9 4.9 3.9 7.0

K2O nm nm nm nm nm nm 0.31 0.32 0.19TiO2 0.01 0.01 0.03 <0.01 0.31 0.57 nm nm nmCr2O3 0.03 0.05 0.10 <0.01 <0.01 <0.01 nm nm nm

MnO 0.11 0.25 0.30 0.78 0.06 0.09 0.13 <0.05 <0.05FeO 31.7 31.3 31.9 10.5 4.5 4.9 0.43 0.15 0.36NiO 0.29 0.31 0.28 <0.02 0.04 <0.02 nm nm nm

Total 100.3 98.7 100.3 100.3 98.7 99.4 98.9 99.4 99.3

nm = not measured.

The number following the area name (i.e., K5, 1) indicates the analysis number.

Fig. 8. Histogram of olivine compositions in the BSE-lightand BSE-dark matrix. The Fa content of matrix olivine rangesfrom 30.5 to 39.8 mol% Fa and peaks at 36 mol% Fa.

)

Fig. 9. Compositions of high-Ca pyroxene in the BSE-lightand BSE-dark matrix of NWA 5343. All pyroxene in thematrix of NWA 5343 is diopside (En32–46Wo46–50).

Weathering and metamorphism of NWA 5343 9

Average abundances of FeO in plagioclase are 0.23wt% in the BSE-dark matrix and 0.13 wt% in the BSE-light matrix; abundances range from 0.04 to 0.64 wt%in all matrix material. These abundances are not as highas those reported from the CK4 chondrite Maralinga(0.75–1.9 wt% FeO) (Keller et al. 1992). Abundances ofMgO, NiO, TiO2, Cr2O3, and MnO are belowdetectable limits. The heterogeneity of plagioclase inNWA 5343 is consistent with other CK chondrites, andis thought to be the result of fractionation during shockmelting (Kallemeyn et al. 1991; Rubin 1992).

Oxides and SulfidesSulfides and oxides are present as accessory phases

in the matrix of NWA 5343. Magnetite is thepredominant oxide, though we also observed very minorchromite and hercynitic spinel. Sulfides are rare,reflecting the low sulfur content of the matrix (0.11wt%). Although compositional trends indicate thatminor pentlandite may be present, we did not encounterany sulfides of sufficient grain size to be analyzed in thematrix of NWA 5343. The sulfur that is present appearsto be dispersed, along with oxides, as nm-sizedinclusions in matrix olivine. Like other CK3 chondrites,magnetite in NWA 5343 is Cr-rich (2.3–4.1 wt%Cr2O3), Al-rich (2.8–5.0 wt% Al2O3), and contains highabundances of TiO2 (0.9–1.4 wt%) and NiO (0.16–0.22wt%) (Dunn et al. 2016a). Magnetite in NWA 5343plots in the CK3 chondrite field of our compositionaldiagrams for CV and CK chondrites (see figs. 4 and 5in Dunn et al. 2016a).

DISCUSSION

Comparison of Matrix and Whole-Rock Compositions

In Fig. 10, we plot Mg- and bulk-CK-normalizedabundance ratios of eight elements in the BSE-dark andBSE-light matrix. Grids L14-1 and L14-2 are excludedfrom the average composition of the BSE-light region.The mean CK bulk composition was determined usingbulk elemental abundances in two CK4 falls (Kobe andKaroonda) and five relatively unweathered CK4chondrites (Oura et al. 2002; Huber et al. 2006;Greenwood et al. 2010). In NWA 5343, abundanceratios of Al, Cr, Ni, and Na are depleted in the matrixrelative to the whole-rock, while abundance ratios of Feand Mn are near unity. Relative to whole-rock,abundance ratios of Ca are significantly enriched in theBSE-light matrix and slightly depleted in the BSE-darkmatrix. Conversely, abundance ratios of K are slightlydepleted in the BSE-light matrix and slightly enriched inthe BSE-dark matrix. However, if we averageabundance ratios of K in the BSE-light and BSE-dark

matrix regions, K plots near unity. These patterns ofdepletion in the matrix of NWA 5343 differ frompatterns in other anhydrous carbonaceous chondrites, asmatrix material is typically enriched in all elementsexcept Ca and Cr relative to the whole-rockcomposition (e.g., Wasson and Rubin 2009).

There are two possible explanations for the depletionof Na in the matrix of NWA 5343; the first of whichbeing volatile loss during impact heating (Wlotzka 1993;Wasson et al. 2013). There is substantial evidence thatthe CK chondrites have been fragmented by impacts(e.g., crushed chondrules), so Na loss would be expected.An alternative, and more likely, explanation is that thisdepletion is due to the absence of Na-bearing phases inthe matrix. In the CK chondrites, whole-rockcompositions are influenced by contributions from thematrix, chondrules, and CAIs. Because plagioclase inCAIs is the main contributor of Na to the bulkcomposition, we would expect the matrix, which containsvery low abundances of plagioclase, to be depleted in Narelative to the whole-rock composition. The matrixdepletion of Al can also be attributed to the lowabundance of plagioclase in the matrix. CAIs alsocontain Cr-bearing phases, such as high-Ca pyroxene andspinel (Chaumard et al. 2014). Thus, the depletion of Crin the matrix is also likely an artifact of the contributionof CAIs to the whole-rock composition.

Fig. 10. Comparison of mean whole-rock and Mg-normalizedabundance ratios in the BSE-dark (diamond) and BSE-light(triangle) matrix regions of NWA 5343. The bulk CKcomposition is the mean of two CK4 falls and five unfractionatedCK chondrites (Oura et al. 2002; Huber et al. 2006; Greenwoodet al. 2010). Abundances of Al, Na, Cr, and Ni are depleted in thematrix relative to the bulk rock, while Ca is enriched in the BSE-light matrix due to the presence of carbonate.

)

10 T. L. Dunn et al.

The primary Ni-bearing phase in the CK chondritesis olivine, with magnetite being a secondary contributor.Because both phases are present in the matrix,chondrules, and to a lesser extent CAIs, the depletion ofNi in the matrix of NWA 5343 cannot be an artifact ofother components’ contribution to the bulk rockcomposition. However, depletion of Ni is characteristicof CK chondrite finds from hot deserts and has beenattributed to breakdown of pentlandite during terrestrialweathering (Rubin and Huber 2005; Huber et al. 2006).Abundances of Ni in the matrix of NWA 5343 areconsistent with Ni abundances in weathered hot desertfinds (Huber et al. 2006) and correspond to aweathering index (wi) of wi-3 (significantly weathered)(Rubin and Huber 2005). Although we must use cautionwhen comparing Ni abundances in NWA 5343, whichrepresent only the matrix contribution, to whole-rockabundances measured by Huber et al. (2006), a visualestimate of the proportion of silicate minerals withbrown staining confirms a weathering index of wi-3.

Terrestrial Weathering

Depletions of Ni and the brown staining of silicates,which forms when oxidized iron is mobilized (Rubinand Huber 2005), confirm that NWA 5343 experiencedterrestrial weathering. Because abundances of Ni are thesame in the BSE-dark and BSE-light regions (0.35wt%), it is likely that the breakdown of pentlandite washomogenous throughout the sample. The similaritybetween abundances of Fe and Mn in the BSE-darkand BSE-light matrix also support this, as bothelements are largely unaffected by weathering in the CKchondrites (Huber et al. 2006). However, in addition tothe characteristic weathering observed in CKchondrites, the BSE-light region of the matrix has alsobeen affected by a calcium carbonate weatheringproduct, which fills the pore space in this region. Thiscarbonate weathering product is responsible for the Cacontent of the BSE-light matrix, which is 2.29 higherthan the BSE-dark matrix. Huber et al. (2006) alsoobserved elevated Ca in hot desert finds, which theyattributed to terrestrial weathering.

The source of the interstitial CaCO3 material in theBSE-light matrix is a vein in region L14 (Fig. 5). This isnot the first occurrence in carbonate veins in a CKchondrite, as Keller et al. (1992) also observed fine-grained carbonate veins in Maralinga, an anomalousCK4 chondrite found in the desert of Australia. BecauseMaralinga was extensively degraded and covered bypatches of carbonate, Keller et al. (1992) attributed theseveins to terrestrial weathering. Carbonate veins havebeen observed in CV chondrites Vigarano, Leoville, andNWA 1456 (Abreu and Brearley 2005; Kereszturi et al.

2015). Although Kereszturi et al. (2015) argued thatveins in NWA 1456, which were filled with interlayeredcarbonate and Fe-Ni oxyhydroxides, formed undervarying oxidizing conditions on the parent body, Abreuand Brearley (2005) suggested that Fe-oxides associatedwith carbonate were produced during terrestrialweathering of metal and sulfides. Abreu and Brearley(2005) also argued that carbonate veins in Vigarano wereterrestrial in origin, due in large part to their associationwith a carbonate layer on the surface of the Vigaranospecimen. Additional support for a terrestrial origin ofthe carbonate in NWA 5343 is the large extent ofcarbonate material. The vein itself is 150–300 lm wide,and carbonate material infiltrates into ~170 cm2 of thetotal sample area. This extent of infiltration wouldrequire a significant amount of fluid, which would beunlikely on the CK chondrite parent body, as much ofthe fluid present would be driven off duringmetamorphism.

The presence of a well-developed CaCO3 vein in thematrix of NWA 5343 provides us with an opportunity toexamine the potential effects of terrestrial weathering onsilicate mineral compositions. Although diffusion offluids into silicate minerals at surface temperature andpressures is unlikely, if diffusion did occur, we wouldexpect silicate minerals in the BSE-light matrix to havehigher CaO contents than their counterparts in the BSE-dark matrix. Although we did observe a higher meanCaO content of olivine in the BSE-light matrix than inthe BSE-dark matrix (0.12 wt% versus 0.09 wt%), valuesfor both are consistent with matrix olivine in other CK3chondrite that we have analyzed (0.10–0.19 wt% CaO)(Dunn et al. 2016b). Therefore, the fluid does not appearto have influenced the composition of olivine. This alsoappears to be true for high-Ca pyroxene. Although theWo content of high-Ca pyroxene in NWA 5343(Wo46–50) is higher than some CK4 chondrites (Wo42–47)(Noguchi 1993; Ivanova et al. 2000), there is nodiscernible difference between the composition of high-Ca pyroxene in BSE-dark and BSE-light material (22.2–25.4 wt% CaO). Calcium content of plagioclase is alsothe same in both regions and is consistent withpreviously measured CK chondrites. The onlymineralogical changes that can be attributed toterrestrial weathering are the precipitation of carbonateand (possibly) apatite from the Ca-bearing fluid.

Thermal Metamorphism

Petrologic SubtypeThe most easily recognizable effects of thermal

metamorphism are textural changes, such asrecrystallization of matrix material (i.e., blurring ofchondrules) (Dodd 1969; Huss et al. 1978, 1981).

Weathering and metamorphism of NWA 5343 11

During metamorphism, opaque matrix materialrecrystallizes, becoming more coarse-grained, lessporous, and less friable and translucent (Dodd 1969;Huss et al. 1978, 1981). Although Huss et al. (1978,1981) suggested that matrix material begins torecrystallize at petrologic subtype 3.6 conditions, recenttransmission electron microscopy (TEM)-based studiessuggest that recrystallization begins much earlier (e.g.,Dobrică and Brearley 2011). Thus, it is difficult toestimate the lower limit of metamorphism in NWA5343 based on the extent of matrix recrystallization.However, because recrystallization is mostly completeby petrologic subtype 3.8 or 3.9 conditions (Huss et al.1978, 1981), the extent of recrystallization in NWA5343 suggests that it is less than petrologic subtype 3.8.

The most pronounced change that occurs duringthermal metamorphism is the chemical equilibration ofsilicate minerals due to diffusive exchange of cations (e.g.,Fe2+, Mg2+, Cr3+, Mn2+, and Ca2+) between chondritecomponents (e.g., chondrules and CAIs) and matrixmaterial (Huss et al. 1978, 1981). Because pyroxeneequilibrates more slowly during metamorphism, thecomposition and homogeneity of olivine is the mostcommonly used indicator of metamorphic grade (Doddand Van Schmus 1967; Huss et al. 1978, 1981). We canuse the ordinary chondrites, which were metamorphosedon onion-shell asteroids (Treiloff et al. 2003), to help usquantify the extent of metamorphism in the CKchondrites. Because both groups have different precursormaterials and reached equilibration at differentcompositions, we cannot directly compare olivinecompositions. However, we can compare the heterogeneityof olivine, as we would expect olivine to equilibrate atroughly the same rate during thermal metamorphism. Asillustrated in a histogram of matrix olivine compositions(Fig. 8), the Fa content of matrix olivine in NWA 5343ranges from 30.5 to 39.8 with a mean at 36 mol%. If wecompare this histogram to histograms of matrix olivine inhigher grade ordinary chondrites (subtypes 3.6–3.8), itmost closely resembles ordinary chondrites Mezo-Madras(L3.7) and Parnallee (LL3.6) in terms of distribution(shape) and compositional range (Huss et al. 1981;Matsunami et al. 1990). Comparison of matrix olivine inNWA 5343 to olivine in CO chondrites yields similarresults, as matrix olivine in NWA 5343 most closelyresembles matrix olivine in Warrenton (CO 3.6) (Brearleyand Jones 1998). Although a detailed examination ofchondrule olivine in the CK chondrites is beyond thescope of this study, the composition and homogeneity ofchondrule olivine in NWA 5343 (Fa15.9�8.5) is alsoconsistent with CO chondrites of petrologic type 3.6 orhigher (Scott and Jones 1990).

In addition to silicate minerals, the bulkcomposition of the matrix is also sensitive to the

effects of metamorphism (Huss et al. 1981). TheFeO/(FeO+MgO) ratio of the matrix decreases duringmetamorphism, approaching the FeO/(FeO + MgO)ratio of the bulk sample when equilibration is reached.Thus, the F/FMmatrix to F/FMbulk ratio for anequilibrated chondrite is 1.0. The FeO/(FeO+MgO)ratio of NWA 5343 matrix is 0.578. This value fallsbetween the CO chondrites ALHA 77003 (3.6) and Isna(3.8) (Scott and Jones 1990). Since we have notmeasured the bulk composition of NWA 5343, wecannot directly address how close to equilibrium thematrix is to the bulk. However, using the bulk FeO/(FeO+MgO) ratios of the CK4 finds Kobe andKaroonda (Oura et al. 2002; Greenwood et al. 2010) asa proxy for the bulk composition, the F/FMmatrix toF/FMbulk ratio of NWA 5343 is 1.05–1.08. Thesevalues are closest to ordinary chondrites of petrologicsubtypes 3.4–3.9, and are most consistent withBremerv€orde (H3.9) (Huss et al. 1981). Overall, thehomogeneity of the bulk matrix and silicate mineralssuggests that NWA 5343 is petrologic subtype 3.6–3.7.

Potential Differences in Metamorphism Between theBSE-Dark and BSE-Light Matrix

Because grain size of silicate minerals increasesduring metamorphism, the observation that olivinegrains appear to be larger and more uniform in size inthe BSE-light matrix than in the BSE-dark matrix maysuggest that the BSE-light matrix is moremetamorphosed. Chaumard et al. (2009) observed thattabular olivine grains found at lower petrologic typesbecame more globular at higher petrologic types. So,the more anhedral nature of grains in the BSE-lightmatrix may also indicate a higher degree ofmetamorphism. However, Chaumard et al. (2009)suggested this change occurred between petrologic types3.8 and 3.9. Because NWA 5343 is no higher thanpetrologic type 3.7, it may be that this change is moregradational, beginning at lower petrologic types andbecoming more evident at higher petrologic types. Analternative explanation may be that the apparenttextural differences between the two regions are theresult of transient heating during impacts rather thanthermal metamorphism. As there is substantial evidencethat the CK chondrites have been shocked andfragmented by impacts, such heating events would beexpected.

If the BSE-light matrix is more metamorphosedthan the BSE-dark matrix, we would also expect olivineand pyroxene in the BSE-light matrix to be morechemically homogeneous and Fe-rich than in the BSE-dark matrix. Although the average composition ofolivine is the same in both the regions (Fa36), olivine inthe BSE-light matrix is more homogeneous, spanning a

12 T. L. Dunn et al.

smaller range in Fa content (Fa35–40) than olivine in theBSE-dark matrix (Fa30–38). The average composition ofdiopside is the same in the both the BSE-dark and BSE-light matrix (En40Fs11Wo49). However, unlike olivine,the heterogeneity of high-Ca pyroxene is the same inboth regions of matrix. This may not be of significancethough, as pyroxene equilibrates slower than olivine.Heterogeneity of plagioclase, which is attributed tofractionation during shock rather than thermalmetamorphism, cannot be used as an indicator ofpetrologic type. Thus, there is minor mineralogicalevidence suggesting that the BSE-light matrix is moremetamorphosed than the BSE-light matrix, though thiswould be difficult to confirm.

Relationship Between Silicate Mineral Equilibrationand Texture

Though silicate minerals seem to be chemicallyequilibrated to the same extent throughout the sample,there are recognizable differences in grain size andshape. The relationship between grain size andpetrologic type has already been established for theequilibrated CK chondrites (Kallemeyn et al. 1991), andboth Chaumard et al. (2009) and Wasson et al. (2013)observed an increase in the average grain size frompetrologic subtype 3.8 to type 4. Wasson et al. (2013)measured an average grain size of 11 lm in NWA 1559(CK3) and 35 lm in NWA 5035 (CK4), whileChaumard et al. (2009) measured an average grain sizeof 16 lm in NWA 1559 and 65 lm in TNZ 057 (CK4).This may suggest that grain size and shape are stillsensitive to metamorphism at petrologic subtypes wheresilicate mineral equilibration is nearly complete (e.g.,>3.7). However, a more detailed examination ofmatrices in CK3 chondrites would be necessary in orderto fully quantify the extent of this textural changeduring the first stages of metamorphism.

Possible CV–CK Metamorphic Sequence

If the CV and CK chondrites are derived from thesame parent body, then the CV chondrites, which areless than petrologic subtype 3.6 (Guimon et al. 1995;Bonal et al. 2006), must be the precursors to the moremetamorphosed CK chondrites. The CV chondrites area highly complex group, with three subgroups (CVred,CVOXA, and CVOXB) characterized by differentoxidation states and types of secondary alteration. TheBali-type CVs (CVOXB) are an unlikely precursor, asthey have heavier oxygen isotope compositions than theCK chondrites (Greenwood et al. 2010) and often haveshock-induced petrofabrics (Rubin 2012) that are absentfrom the CKs. Greenwood et al. (2010) suggested thatthe reduced CV chondrites are a good isotopic match

for CK chondrite precursor material. If this is the case,two possible metamorphic sequences exist, i.e., (1)CVred ? CVOXA ? CK or (2) CVred ? CK. Abreuand Brearley (2011) argued that the CVOXA chondritescould not have formed from thermal processing ofCVred material because of the complex, fluid-assistedgrowth required to form matrix olivine in the CVOXA

chondrites. Thus, the most likely scenario would be thatCVred chondrites are a direct precursor to the CKchondrites.

To examine CVred chondrites as a possibleprecursor material for the CKs, we compared the meancomposition of the BSE-dark matrix in NWA 5343 tothe average matrix composition in Vigarano (CVred)(Hurt et al. 2012). We elected to use the BSE-darkmatrix in this comparison because the bulk compositionof the BSE-light reflects the contribution of Ca from thecarbonate weathering product. Magnesium- andVigarano-normalized abundance ratios of 10 elements inthe matrix of NWA 5343 are plotted in Fig. 11. If theCV and CK chondrites are genetically related, thenmatrix material in both groups should have nearlyidentical starting compositions, and any differencesbetween the two matrices must be attributable tosecondary alteration process (i.e., impact heating,aqueous alteration, or thermal metamorphism).Compared to Vigarano, the matrix of NWA 5343 issignificantly depleted in Fe, Mn, Na, Al, and S, andslightly less depleted in Cr, Si, and K (Fig. 11). Theonly element that is enriched in the matrix of NWA5343 is Ca, which is 2.89 higher than matrix inVigarano. Because we are using the BSE-dark in ourcomparison with Vigarano, the Ca enrichment in NWA5343 cannot be attributed to CaCO3 weatheringmaterial.

NWA 5343 is clearly more metamorphosed thanVigarano, as evidenced by the chemical composition,heterogeneity, and grain size of olivine in Vigarano(Abreu and Brearley 2011). This is confirmed by theabundance of Fe in the matrix of NWA 5343, which isdepleted relative to Vigarano. We would expect thematrix of NWA 5343 to have a lower Fe content thanVigarano, as iron diffuses from the matrix intochondrules during metamorphism (e.g., Huss et al.1981; Scott and Jones 1990). Calcium behaves in theopposite manner (diffusing from chondrules to matrix),so the enrichment of Ca in the matrix of NWA 5343 islikely a result of the destabilization of chondrulemesostasis during metamorphism. However, themagnitude of the enrichment seems to argue againstmetamorphism as the sole source. The depletion of S inthe matrix of NWA 5343 could be an effect of thermalmetamorphism or terrestrial weathering. Both processeshave been shown to result in a decrease in sulfur

Weathering and metamorphism of NWA 5343 13

content (Grossman and Brearley 2005; Huber et al.2006). Because Vigarano is a fall, we would expect it tohave higher S than a hot desert find such as NWA5343. The depletion of Na in the matrix of NWA 5343is probably due to impact heating and shock, which isindicated by the presence of fragmented matrix. It is notclear what secondary process could be responsible forthe depletion of Mn and Al, but the remainingdifferences between the matrices of Vigarano and NWA5343 can be explained by secondary effects, suggestingthat it is at least possible that NWA 5343 and Vigaranoare genetically linked.

Although it is certainly possible that the matrixcomposition of NWA 5343 could be derived bysecondary processing of a Vigarano-like precursor, thisalone does not provide conclusive evidence for a geneticrelationship between the CVred and CK chondrites. Inaddition to the arguments outlined by Dunn et al.(2016a) and Yin et al. (2017), porosity and texture ofmatrix olivine are also hard to reconcile with acontinuous metamorphic sequence. Because porositydecreases during metamorphism, a continuousmetamorphic sequence would imply that the CKchondrites should have lower porosities than the CVred

chondrites. Actual porosity measurements of the CVred

and CK chondrites are contrary to this, as porosity ofthe CVred chondrites is much lower than porosity of theCK chondrites (average of 3.6% and 17.8%,respectively) (Macke et al. 2011). However, the lowporosity of the CVred chondrites could be a result ofshock flattening. In addition, matrix olivine in theCVOX chondrites, Allende and ALH 84028 (CV3),

exhibits a “dusty” texture similar to matrix olivine inthe CK chondrites; however, this texture is absent fromCVred chondrite, Vigarano (Brearley 2009; Abreu andBrearley 2011). Abreu and Brearley (2011) argued thatthe absence of this texture required a differentformation mechanism for olivine in the CVred

chondrites, thus invalidating the argument that CVOX

chondrites could be derived from thermal processing ofthe CVred chondrites. Because matrix olivine in the CKchondrites appears to have formed via the same processas olivine in the CVOX chondrites, this argument thenapplies to the CK chondrites as well.

SUMMARY

Based on the chemical heterogeneity of chondruleand matrix olivine, NWA 5343 is the leastmetamorphosed CK chondrite reported so far. In aneffort to better constrain the lower limit ofmetamorphism in the CK chondrites, we performed adetailed analysis of matrix material in NWA 5343,including characterization of the texture and bulkcomposition and analyses of individual silicate minerals.To evaluate the possibility of a metamorphic sequencewithin the CV and CK chondrites, we also comparedthe bulk matrix composition of NWA 5343 to that ofthe reduced CV chondrite Vigarano. Our study ofmatrix in NWA 5343 yields the following conclusions.1. Compared to bulk compositions of CK4 chondrites,

the matrix of NWA 5343 is depleted in Al, Cr, Ni,and Na. The depletion of Na, Al, and Cr is likelyan artifact of the contribution of CAIs to the bulkcomposition, while the depletion of Ni is the resultof terrestrial weathering. The enrichment of Ca inthe BSE-light matrix is due to the presence ofinterstitial carbonate.

2. The CaCO3 material filling the pore space in theBSE-light matrix is a product of terrestrialweathering. Infiltration of a Ca-bearing, terrestrialfluid does not appear to have directly influenced thecomposition of silicate minerals in the matrix. Theonly mineralogical changes that can be attributed toterrestrial weathering are the precipitation ofcarbonate and (possibly) apatite from the Ca-bearing fluid.

3. The extent of matrix recrystallization, heterogeneityof matrix olivine, and bulk composition of thematrix suggest that NWA 5343 is petrologic subtype3.6 or 3.7.

4. Though silicate minerals seem to be equilibrated toroughly the same extent throughout the sample,there are recognizable differences in grain size andshape. These textural differences may be the resultof transient heating events during impacts, which

Fig. 11. Mean abundances of BSE-dark matrix in NWA 5343normalized to Mg and Vigarano (CVred) matrix (Hurt et al.2012). All elements except Ca have abundance ratios <1.

14 T. L. Dunn et al.

would be likely on the CK chondrite parent body.The difference between compositional equilibrationand texture may also suggest that grain size andshape is still sensitive to metamorphism atpetrologic subtypes where silicate mineralequilibration is nearly complete (e.g., >3.7).

5. Although the matrix composition of NWA 5343could be derived by secondary processing of aVigarano-like precursor, porosity and texture ofmatrix olivine are hard to reconcile with acontinuous metamorphic sequence. Porosity is lowerin the CVred chondrites than in the CK chondrites,and matrix olivine in the CKs appears to require adifferent formation mechanism than matrix olivinein the CVred chondrites.

Acknowledgments—This work was funded by NASACosmochemistry grant NNX14AG61G to TLD.Summer research support for undergraduate students O.Battifarano and E. O’Hara was provided by the ClareBoothe Luce program of the Henry Luce Foundation.We thank reviewers N. Chaumard and K. Howard andAssociate Editor A. Brearley for their thoughtfulcomments and suggestions. Our manuscript has beengreatly improved by their suggestions.

Editorial Handling—Dr. Adrian Brearley

REFERENCES

Abreu N. M. and Brearley A. J. 2005. Carbonate in theVigarano CV3 carbonaceous chondrite: Terrestrial,preterrestrial or both? Meteoritics & Planetary Science40:609–625.

Abreu N. and Brearley A. 2011. Deciphering the nebular andasteroidal record of silicates and organic material in thematrix of the reduced CV chondrite Vigarano. Meteoritics& Planetary Science 46:252–274.

Bonal L., Quirco E., Bourot-Denise M., and Montagnac G.2006. Determination of the petrologic type of C3chondrites by Raman spectroscopy of induced organicmatter. Geochimica et Cosmochimica Acta 70:1849–1863.

Brearley A. J. 2009. Matrix olivine in the metamorphosed CKchondrite NWA 1628: Possible affinities to olivines in thematrices of oxidized CV3 chondrites and dark inclusions(abstract #1791). 40th Lunar and Planetary ScienceConference. CD-ROM.

Brearley A. J. and Jones R. H. 1998. Chondritic meteorites. InPlanetary materials, edited by Papike J. J. Washington,D.C.: Mineralogical Society of America. pp. 3-1–3-398.

Brearley A. J. and Krot A. N. 2013. Metasomatism inchondritic meteorites. In Metasomatism and the chemicaltransformation of rock: The role of fluids in terrestrial andextraterrestrial processes, edited by Harlov D. andAustrheim H. Lecture Notes in Earth Sciences Series.Heidelberg, Germany: Springer Verlag. pp. 653–782.

Chaumard N. and Devouard B. 2016. Chondrules in CKcarbonaceous chondrites and thermal history of the CV-CKparent body.Meteoritics & Planetary Science 51:547–573.

Chaumard N., Devouard B., Zanda B., and Ferri�ere L. 2009.The link between CV and CK carbonaceous chondritesbased on parent body processes (abstract #5206).Meteoritics & Planetary Science 44.

Chaumard N., Devouard B., Delbo M., Provost A., andZanda B. 2012. Radiative heating of carbonaceous near-Earth objects as a cause of thermal metamorphism for CKchondrites. Icarus 220:65–73.

Chaumard N., Devouard B., Bouvier A., and Wadhwa M.2014. Metamorphosed calcium-aluminum inclusions in CKcarbonaceous chondrites. Meteoritics & Planetary Science49:419–452.

Chizmadia L. J., Rubin A. E., and Wasson J. T. 2002.Mineralogy and petrology of amoeboid olivine inclusionsin CO3 chondrites: Relationship to parent-body aqueousalteration. Geochimica et Cosmochimica Acta 37:1781–1796.

Dobrică E. and Brearley A. J. 2011. Earliest stages ofmetamorphism and aqueous alteration observed in thefine-grained materials of two unequilibrated ordinarychondrites (abstract #2092). 42nd Lunar and PlanetaryScience Conference. CD-ROM.

Dodd R. T. 1969. Metamorphism of the ordinary chondrites:A review. Geochimica et Cosmochimica Acta 33:161–203.

Dodd R. T. and Van Schmus W. R. 1967. A survey of theunequilibrated ordinary chondrites. Geochimica etCosmochimica Acta 31:921–951.

Dunn T. L., Gross J., Ivanova M. A., Runyon S. E., andBruck A. M. 2016a. Magnetite in the unequilibrated CKchondrites: Implications for metamorphism and newinsights into the relationship between the CV and CKchondrites. Meteoritics & Planetary Science 51:1701–1720.

Dunn T. L., Gross J., and Ivanova M. 2016b. Homogeneity ofmatrix and chondrule olivine in the unequilibrated CKchondrites (abstract #1921). Meteoritics & PlanetaryScience 51.

Geiger T. and Bischoff A. 1995. Formation of opaque in CKchondrites. Planetary and Space Science 43:485–498.

Geiger T., Spettel B., Clayton R. N., Mayeda T. K., andBischoff A. 1993. Watson 002—The first CK/type 3chondrite. Meteoritics 28:352.

Greenwood R. C., Franchi I. A., Kearsley A. T., and AlardO. 2010. The relationship between the CK and CVchondrites. Geochimica et Cosmochimica Acta 74:1684–1705.

Grossman J. N. and Brearley A. J. 2005. The onset ofmetamorphism in ordinary and carbonaceous chondrites.Meteoritics & Planetary Science 40:87–122.

Guimon R. K., Symes S. J. K., Sears D. W. G., and Benoit P.H. 1995. Chemical and physical studies of chondrites. XII:The metamorphic history of CV chondrites and theircomponents. Meteoritics 30:704–714.

Huber H., Rubin A. E., Kallemeyn G. W., and Wasson J. T.2006. Siderophile-element anomalies in CK carbonaceouschondrites: Implications for parent-body aqeous alterationand terrestrial weathering of sulfides. Geochimica etCosmochimica Acta 70:4019–4037.

Hurt S. M., Rubin A. E., and Wasson J. T. 2012.Fractionated matrix in CV3 Vigarano and alterationprocesses on the CV parent asteroid. Meteoritics &Planetary Science 47:1035–1048.

Huss G. R., Keil K., and Taylor G. J. 1978. Composition andrecrystallization of the matrix of unequilibrated (type 3)ordinary chondrites. Meteoritics 13:495.

Weathering and metamorphism of NWA 5343 15

Huss G. R., Keil K., and Taylor G. J. 1981. The matrices ofunequilibrated ordinary chondrites: Implications for theorigin and history of chondrites. Geochimica etCosmochimica Acta 45:33–51.

Ivanova M. A., Nazarov M. A., Kononkova N. N., Taylor L.A., Patchen A., Clayton R. N., and Mayeda T. K. 2000.Dhofar 015, a new CK chondrite: A record of nebularprocesses (abstract #5046). 31st Lunar and PlanetaryScience Conference. CD-ROM.

Kallemeyn G. W., Rubin A. E., and Wasson J. T. 1991. Thecompositional classification of chondrites: V. TheKaroonda (CK) group of carbonaceous chondrites.Geochimica et Cosmochimica Acta 55:881–892.

Keller L. P., Clark J. C., Lewis C. F., and Moore C. B. 1992.Maralinga, a metamorphosed carbonaceous chondritefound in Australia. Meteoritics 27:87–91.

Kereszturi A., Ormandi S., and Joza S. 2015. Processing atransitional environment of CV and CK chondrites’ parentbodies in the light of mineralogical and petrologicalanalysis of NWA 1465 CV3 meteorite . Planetary andSpace Science 109–110:175–186.

Macke R. J., Consolmagno G. J., and Britt D. T. 2011.Density, porosity, and magnetic susceptibility ofcarbonaceous chondrites. Meteoritics & Planetary Science46:1842–1862.

Matsunami S., Nishimura H., and Takeshi H. 1990. Thechemical composition and textures of matrices andchondrule rims of unequilibrated unequilibrated ordinarychondrites-II. Their constituents and implications for theformation of matrix olivine. Proceedings of the NIPRSymposium on Antarctic Meteorites 35:147–180.

McSween H. Y. Jr. 1977. Petrographic variations amongcarbonaceous chondrites of the Vigarano type. Geochimicaet Cosmochimica Acta 41:1777–1790.

Noguchi T. 1993. Petrology and mineralogy of CK chondrites:Implications for the metamorphism of the CK chondriteparent body. Proceedings, 17th NIPR Symposium ofAntarctic Meteorites. pp. 2014–2233.

Oura Y., Ebihara M., Yoneda S., and Nakamura N. 2002.Chemical composition of the Kobe meteorite: Neutron-induced prompt gamma ray analysis study. GeochemicalJournal 36:295–307.

Pearson V. K., Greenwood R. C., Morgan G. H., Turner D.,Raade G, Roaldset E., and Gilmour I. 2007. Organicconstitution of the CO3 chondrites and implications forasteroidal processes (abstract #1846). 38th Lunar andPlanetary Science Conference. CD-ROM.

Rubin A. E. 1992. A shock-metamorphic model for silicatedarkening and compositionally variable plagioclase in CKand ordinary chondrites. Geochimica et Cosmochimica Acta56:1705–1714.

Rubin A. E. 2010. Physical properties of chondrules indifferent chondrite groups: Implications for multiplemelting events in dusty environments. Geochimica etCosmochimica Acta 74:4807–4828.

Rubin A. E. 2012. Collisional facilitation of aqueousalteration of CM and CV carbonaceous chondrites.Geochimica et Cosmochimica Acta 90:181–194.

Rubin A. E. and Huber H. 2005. A weathering index for CKand R chondrites. Meteoritics & Planetary Science40:1123–1130.

Scott E. R. D. and Jones R. H. 1990. Disentangling nebularand asteroidal features of CO3 carbonaceous chondritemeteorites. Geochimica et Cosmochimica Acta 54:2485–2502.

Sears D. W., Grossman J. N., Melcher C. L., Ross L. M., andMills A. A. 1980. Measuring metamorphic history ofunequilibrated ordinary chondrites. Nature 287:791–795.

Sears D. W., Hasan F. A., Batchelor J. D., and Lu J. 1991.Chemical and physical studies of type 3 chondrites. XI—Metamorphism, pairing, and brecciation of ordinarychondrites. Proceedings, 21st Lunar and Planetary ScienceConference. pp. 493–512.

Treiloff M., Jessberger E. K., Herrwerth I., Hopp J., Fi�eniC., Gh�elis M., Bourot-Denis M., and Pellas P. 2003.Structure and thermal history of the H-chondrite parentasteroid revealed by thermochronometry. Nature 422:502–506.

Wasson J. T. and Rubin A. E. 2009. Composition of matrix inthe CR chondrite LAP 02342. Geochimica etCosmochimica Acta 73:1436–1460.

Wasson J. T., Isa J., and Rubin A. E. 2013. Compositionaland petrographic similarities of the CV and CKchondrites: A single group with variations in textures andvolatile concentrations attributed to impact heating,crushing, and oxidation. Geochimica et Cosmochimica Acta108:45–62.

Wlotzka F. 1993. A weathering scale for ordinary chondrites.Meteoritics 28:460.

Yin Q.-Z., Sanborn M. E., and Ziegler K. 2017. Testing thecommon source hypothesis for the CV and CK chondriteparent body using D17O-e54Cr isotope systematics (abstract#1771). 48th Lunar and Planetary Science Conference.CD-ROM.

SUPPORTING INFORMATION

Additional supporting information may be found inthe online version of this article:

Fig S1. BSE images of the five regions examined inthis study. Areas that were analyzed for bulk

composition are indicated on each area by150 9 150 µm boxes. Areas K5, S6, and U12 areporous matrix, while I21 and L14 are weathered matrix.Scale bar is 500 µm.

Fig S2. Average bulk compositions of each of thegrid areas.

16 T. L. Dunn et al.