Multi-technique investigation reveals new mineral, chemical, and textural heterogeneity in the Tagish Lake C2 chondrite

Planetary and Space Science 58 (2010) 1347–1364

Contents lists available at ScienceDirect

Planetary and Space Science

0032-06

doi:10.1

n Corr

E-m

journal homepage: www.elsevier.com/locate/pss

Multi-technique investigation reveals new mineral, chemical, and texturalheterogeneity in the Tagish Lake C2 chondrite

M.R.M. Izawa n, R.L. Flemming, P.J.A. McCausland, G. Southam, D.E Moser, I.R. Barker

Department of Earth Sciences, the University of Western Ontario, 1151 Richmond St. London, Ontario, Canada N6A 5B7

a r t i c l e i n f o

Article history:

Received 1 December 2009

Received in revised form

20 May 2010

Accepted 21 May 2010Available online 1 June 2010

Keywords:

Carbonaceous chondrite

Micro-XRD

Cathodoluminescence

Chondrules

CAI

Olivine

Sulfide

Magnetite

Carbonate

Parent body alteration

33/$ - see front matter & 2010 Elsevier Ltd. A

016/j.pss.2010.05.018

esponding author.

ail address: [email protected] (M.R.

a b s t r a c t

The Tagish Lake meteorite, an ungrouped C2 chondrite that is related to CI and CM chondrites, is a

heterogeneous accretionary breccia with several distinct lithologies that, in bulk, are thought to

represent the first known sample of a primitive carbonaceous D-type asteroid. Textural and chemical

zoning of clasts and matrix have been little studied and promise additional insight into early solar

system processes in both the solar nebula and on the Tagish Lake parent asteroid. We have examined an

intact 2.9 g fragment and two polished thin sections from the spring 2000 (non-pristine) Tagish Lake

collection to ascertain the major mineralogy and textures of notable features such as chondrules,

amoeboid olivine aggregates (AOAs), inclusions, clasts, matrix, and fusion crust. We designed three

stages of analysis for this friable meteorite: an initial, non-destructive in situ reconnaissance by mXRD to

document meteorite mineralogy and textures and to identify features of interest, followed by spatially

correlated mXRD, SEM-EDX and colour SEM-CL analysis of polished thin sections to fully understand

mineralogy and the record of texture development, and finally higher resolution SEM-BSE mapping to

document smaller scale relationships.

Our analyses reveal several previously unreported or poorly characterized features: (1) distinctive

colour cathodoluminescence (CL) zoning in relict CAI spinel, in chondrule and AOA forsterite, and in

calcite nodules occurring throughout the Tagish Lake matrix. Forsterite frequently shows CL colour and

intensity zonation that does not correspond with major or minor element differences resolvable with

EPMA, indicating a trace element and/or structural CL-activation mechanism for the zonation that is

likely of secondary origin; (2) an irregular inclusion dominated by magnesioaluminate spinel, dolomite,

and phyllosilicates with traces of a Ca, Ti oxide phase (likely perovskite) interpreted to be a relict CAI;

(3) variable preservation of mesostasis glass in porphyritic olivine chondrules. We anticipate that our

multi-technique methodology, particularly non-destructive mXRD, can be successfully applied to other

rare and friable materials such as the pristine Tagish Lake fragments.

& 2010 Elsevier Ltd. All rights reserved.

1. Introduction: the Tagish Lake C2 chondrite

Carbonaceous (C) chondrites provide the most primitive solarsystem material accessible on Earth. They carry textural, miner-alogical, and compositional clues to processes and environmentsin the earliest stages of solar system evolution, both in the solarnebula and on parent asteroids (Brearley and Jones, 1998; Scottand Krot, 2005). Carbonaceous chondrites are commonly friableand porous, and are prone to rapid contamination and alterationin the terrestrial environment. Carbonaceous chondrites are rarelyrecovered soon after they fall; therefore, the recovery of a freshlyfallen and relatively unaltered C chondrite is a rare andscientifically valuable event.

ll rights reserved.

M. Izawa).

The Tagish Lake carbonaceous chondrite fell in northern BritishColumbia, Canada, on 18 January 2000 (Brown et al., 2000).Fortuitously, a large quantity of fragments fell on the frozensurface of Tagish Lake. A small amount, �1 kg was recoveredwithin days of the fall event by local resident Jim Brook. In spring2000, a recovery operation headed by the University of WesternOntario and the University of Calgary recovered �10 kg ofadditional material. A comprehensive review of the fall andrecovery of the Tagish Lake meteorite was provided by Hildebrandet al. (2006). Tagish Lake is among the most primitive andphysically weak meteorites ever studied (Brown et al., 2000). It isalso among the few meteorite falls for which pre-entry orbitalelements have been determined through the use of a combinationof eyewitness, video, photograph, infrasound, and satellite data(Brown et al., 2000, 2002). The orbital data suggest an associationwith the primitive, organic-rich C, D or P-type asteroids of theouter main belt (Brown et al., 2000; Hildebrand et al., 2006).Infrared spectroscopy also suggests a relationship with outer belt

www.elsevier.com/locate/pss

dx.doi.org/10.1016/j.pss.2010.05.018

mailto:[email protected]

dx.doi.org/10.1016/j.pss.2010.05.018

M.R.M. Izawa et al. / Planetary and Space Science 58 (2010) 1347–13641348

D and P-type asteroids (Hiroi et al., 2001; Hiroi and Hasegawa,2003).

Tagish Lake is an ungrouped C2 chondrite with affinities toboth CI and CM chondrites. It is an accretionary breccia consistingof at least two major lithologies, identified as carbonate-poor andcarbonate-rich (Zolensky et al., 2002; Simon and Grossman, 2003;Bland et al., 2004). There may be some degree of gradationbetween these lithologies. The ratio of carbonate-poor tocarbonate-rich lithology has been estimated as �3:2 (Zolenskyet al., 2002), but given the heterogeneity of the meteorite and thesmall fraction of the total amount of recovered material that hasbeen examined thus far, such assessments may be non-represen-tative of the main mass: the amount of material investigated indetail between 2000 and 2006 was no more than 18 g of thepristine material (Herd and Herd, 2007). Both lithologies arecharacterized by a dark, fine-grained opaque matrix consisting ofcomplex intergrowths of the Fe-bearing phyllosilicate saponitewith magnetite, Fe, Ca, Mg, Mn carbonates and Fe, Ni sulfides(Zolensky et al., 2002; Bland et al., 2004). Matrix sulfides consistprimarily of pyrrhotite with rare pentlandite (Zolensky et al.,2002; Boctor et al., 2003). The phyllosilicates in the Tagish Lakematrix have previously been identified as saponite complexlyintergrown with serpentine based on HR-TEM lattice fringes(Keller and Flynn, 2001; Mikouchi et al., 2001; Noguchi et al.,2002; Zega et al., 2005) and XRD (Bland et al., 2004; Russell,2006). Matrix carbonate in the carbonate-poor lithology ispredominantly siderite whereas the carbonate-rich lithology isdominated by calcite (Nakamura et al., 2003). The carbonate-richlithology is also much lower in magnetite abundance. Izawa et al.(2010) reported two new possible lithologies: a magnetite- andsulfide-rich ‘dark, dusty’ material, and a carbonate-rich materialdominated by siderite (up to �25 wt%) rather than calcite (Izawa,2008; Izawa et al., 2010). Blinova and Herd (2010) and Blinovaet al. (2009) also reported new Tagish Lake lithologies including ahighly friable, clast-poor and phyllosilicate-rich ‘‘dark, dusty’’lithology.

Set within the matrix are a variety of clasts including sparsechondrules, amoeboid olivine aggregates (AOA), isolated grainsof olivine, uncommon coarse-grained phyllosilicates and rarepyroxene, rare calcium–aluminium rich inclusions (CAI), andirregular carbonate nodules (Zolensky et al., 2002; Bland et al.,2004). Geochemical (Mittlefehldt, 2002; Nakamura et al., 2003),oxygen isotopic (Clayton and Mayeda, 2001; Baker et al., 2002;Russell, 2006; Russell et al., 2010), organic chemical (Pizzarelloet al., 2001; Kminek et al., 2002; Binet et al., 2004), and petrologic(Zolensky et al., 2002) studies suggest a relationship betweenTagish Lake and the CI and CM chondrites.

2. Rationale

Despite the comprehensive overview provided by Zolenskyet al. (2002), much of the mass of the remarkably heterogeneousTagish Lake meteorite remains unstudied and unknown. Here wereport mineralogical and textural observations from new TagishLake fragments using a reconnaissance strategy that follows threesuccessive stages: an initial, non-destructive in situ reconnais-sance step using micro X-ray diffraction (mXRD) to documentmeteorite mineralogy and textures and to identify features ofinterest, followed by spatially correlated mXRD, scanning electronmicroscopy with energy-dispersive X-ray spectroscopy (SEM-EDX), and cathodoluminescence (CL) analysis of polished thinsections to fully understand their mineralogy and textures, andfinally higher resolution SEM-BSE mapping to establish spatialcontext for textural variation.

Recent developments in micro X-ray diffraction (mXRD) haveexpanded the applicability of X-ray diffraction to in situ study ofEarth and planetary materials (Flemming, 2007). Micro XRDextends X-ray diffractometry to the scale of 10 s of microns,enabling rapid, non-destructive, in situ measurements of crystalstructure. This technique is valuable as a ‘no-touch’ first-passreconnaissance, which may identify areas for further study,helping to minimize sample manipulation and loss. Furthermore,we demonstrate that micro XRD also provides point-by-pointcorrelation of crystal structure data with other microscopic andmicroanalytical data, such as imaging and elemental mapping viaSEM-EDX and quantitative chemical analysis by electron probemicroanalysis (EPMA).

Micro XRD was applied in two ways in this study: to identifythe major minerals in an intact fragment of the Tagish Lake C2carbonaceous chondrite without sample preparation; and inTagish Lake polished thin sections, in conjunction withSEM-EDX mapping, SEM-CL imagery, and EPMA chemical analy-sis, providing crystal structure data to complement thesechemical analyses. Micro X-ray diffraction has also providedqualitative textural information using a two-dimensional detec-tor, and provides an in situ probe of long range order.

3. Experimental methods

3.1. Micro X-ray diffraction (mXRD)

All mXRD measurements were made using a Bruker-AXS D8Discover diffractometer using Cu Ka radiation (l¼1.5418 A) at40 kV accelerating voltage and 40 mA beam current. Gobel mirrorparallel beam optics removed Kb radiation, and maximizeddiffracted beam intensity for non-flat samples. A nominal beamdiameter of 500 mm or 50 mm was produced by an exchangeablepinhole collimator snout. Diffracted X-rays were detected with atwo-dimensional general area diffraction detector system(GADDS). Two-dimensional detection also provides a measure ofthe texture and crystallinity of the sample, by examination of thedistribution of diffracted ray intensity along the Debye rings (He,2003; Helming et al., 2003; Tissot, 2003; Flemming, 2007). Theintegrated GADDS images were analysed using Bruker-AXS DiffracPlus Evaluation software (BrukerAXS, 2005), and the InternationalCenter for Diffraction Data (ICDD) Powder Diffraction File (PDF-4)database (rev. 2006) for phase identification. Flemming (2007)presented a comprehensive description of mXRD instrumentationand techniques.

3.2. Secondary electron microscopy with multi-spectral

cathodoluminescence (SEM-CL)

Backscatter electron (BSE), secondary electron (SE), and colourcathodoluminescence (CL) images of polished thin sections werecollected with a tungsten-filament Hitachi S-2500C SEM at theZircon and Accessory Phase Laboratory, University of WesternOntario using Robinson Backscatter and Gatan ChromaCL detec-tors. The carbon coated thin sections were analysed at highvacuum with an accelerating voltage of 15–20 kV and a workingdistance of �10 mm. SEM-CL has the ability to reveal submicronscale variations in electron-stimulated light emissions frompolished surfaces; these variations are a proxy for trace elementand/or structural variations within several microns of the emittersurface. The electron beam was rastered using a Digiscan II beamcontroller and the dwell time per pixel was varied over the rangeof 80–500 ms, depending on the amount of long-lived infraredluminescence (commonly observed in carbonates). CL emission

M.R.M. Izawa et al. / Planetary and Space Science 58 (2010) 1347–1364 1349

was detected in the range 300–850 nm and passed through adiffraction grating that divides the CL signal into four channels:red (�600–850 nm, including near-infrared from �700–850 nm),green (�500–600 nm), blue (�400–500 nm), and ultraviolet(�300–400 nm). Digital images were assembled using GatanDigital Micrograph software.

Fig. 1. Photomosaic of images of the intact 2.9 g fragment from site MM-02,

indicating the 500 mm points examined by mXRD, as summarized in Table 1. Scale

bar (2 mm) is identical for both images. The two images are on approximately

opposite sides of the fragment. Note the well-developed irregular fractures in the

fragment, which has subsequently broken into two sub-fragments. Variation in

brightness is an artefact of autogain in the Bruker D8 videocamera/microscope

system.

3.3. Scanning electron microscopy with energy-dispersive X-ray

spectroscopy (SEM-EDX)

Backscattered electron images and EDX element maps of theTagish Lake sections were acquired with the Leo 440 SEM atSurface Science Western. This SEM is equipped with a Greshamlight element detector and a Quartz XOne EDX analysis system,capable of detecting all elements from C to U, with a detectionlimit of �0.5 wt% for most elements. Backscattered electron(BSE) imaging and elemental X-ray mapping provide graphicalrepresentations of elemental distribution. The Quartz XOnesystem uses full spectral imaging, recording all X-rays collectedfrom each pixel location, allowing live and post-collection dataanalysis.

High-resolution BSE imaging and EDX spot analysis werecarried out with the Leo 1540 FIB/SEM CrossBeam field emissionSEM at the Nanofabrication laboratory at UWO. This SEM isequipped with an Oxford Instruments INCA EDX system allowingfor elemental analysis. Additional high-resolution BSE imaging,EDX mapping, and spot analyses were carried out using a HitachiS-4300SE/N field emission SEM at the Imaging Center, Texas TechUniversity, and with a Hitachi SU6600 Schottky Field EmissionAnalytical SEM at the Earth Sciences Department, University ofWestern Ontario.

3.4. Electron probe microanalysis (EPMA)

Chemical compositions for selected targets on the Tagish Lakethin sections were analysed for major and minor elementchemistry by EPMA by Renaud Geological Consulting Ltd. Analysiswas carried out with a JEOL JXA-733 electron microprobeequipped with 5-wavelength-dispersive spectrometers (WDS)and an energy-dispersive spectrometer (EDX). Count times formajor elements were 20 s on peak and 10 s (on each side) forbackground measurements. For trace elements both peak andbackground count times were 50 s. Natural mineral microbeamstandards from the Smithsonian Institution were used forcalibration (Jarosewich, 2002). Data reduction was performedusing the F(rZ) oxide correction of Armstrong (1995). Analyticalaccuracy was verified using secondary standards: San Carlosolivine standards of Brey and Kohler (1990) and Smithsonianmicrobeam standards (Jarosewich, 2002). The instrument calibra-tion was deemed successful when the composition of secondarystandards was reproduced within the error margins defined bythe counting statistics.

3.5. Optical petrography

Polished thin sections were examined with a petrographicmicroscope in transmitted and reflected light. The method ofpetrographic shock classification developed by Stoffler et al.(1991) for ordinary chondrites and extended to carbonaceouschondrites by Scott et al. (1992) was used to assign a petrographicshock stage based on olivine in the Tagish Lake polished sectionsamples.

4. Tagish Lake samples

Samples for this study are from the spring 2000 (non-pristine)collection and include an intact 2.9 g fragment from the collectionsite MM-02 (Figs. 1 and 2) and six standard petrographic polishedthin sections prepared from a single fragment weighing 0.472 g(Fig. 3), subsampled from collection site MG-02 (Brown et al., 2000;Hildebrand et al., 2006). These sections were prepared by hand byGord Wood (UWO Earth Science) using isopropyl alcohol as the onlysolvent, and were set in Struers EpoFix epoxy following themanufacturer’s instructions (N. D. MacRae, personal communication2006; and G. Wood, personal communication 2010). Micro XRDanalyses of 34 points on the intact fragment were made using a500 mm nominal beam diameter (Fig. 1). An additional 132 points onthe thin sections were examined by mXRD for correlated study withSEM-EDX. Of these, 64 were 500 mm context analyses intended tocharacterize large matrix areas and grain assemblages, with theremaining 58 analyses being 50 mm nominal beam diameter spotstargeted on areas of particular interest. All six thin sections weremapped in BSE (Fig. 3) and features of interest were further examinedusing BSE at higher resolution.

5. Micro X-ray diffraction of the intact MM-02 fragment

The results of the 34 points examined on surface of the intactMM-02 fragment are mapped in Fig. 1 and summarized in Table 1.The interpretations were made on the basis of minerals identifiedby mXRD, the morphology of large inclusions (if present), theappearance of the target in the optical photomicrograph producedby the Bruker D8 camera, and the grain size and texture inferredfrom the two-dimensional GADDS images.

Fig. 2. Example of in situ mXRD of a small forsterite-bearing clast in matrix, spot 34 of Fig. 1. (A) Standard integrated X-ray diffractogram plotting peak intensity versus 2y,

with ICDD mineral pattern matches: forsterite—00-34-0189, magnetite—00-019-0629, pyrrhotite—00-029-0723, saponite—00-012-0160; (B) Camera image; white circle

is the nominal footprint of the 500 mm beam; (C) Two-dimensional GADDS images of the diffracted X-rays. Continuous diffracted rings correspond to fine-grained saponite,

pyrrhotite and magnetite, whereas the discrete spots are due to coarse-grained single crystals of forsterite. The forsterite spots are circular, consistent with an unstrained

mineral. Note that the coarse grain size of the forsterite can affect the relative intensities of the characteristic forsterite peaks in the diffraction pattern, due to the

likelihood of some diffracted spots lying outside the detector area. Sample surface topography can also cause missing peaks due to beam shadowing, particularly at low 2y.

In both instances, however, the 2y peak positions, if present, remain unaffected and may still be used for mineral identification.

Fig. 3. Backscattered electron image mosaics of the six Tagish Lake thin sections

investigated in this study. Note the predominance of fine-grained matrix and the

wide range of inclusion size, morphology, and composition. The sections are six

members of a set of eight that were produced soon after the Tagish Lake collection

in spring 2000, that is why they are labelled A, B, C, D, G, and H.


Matrix sites investigated by mXRD were dominated by Fe-richsaponite, magnetite, siderite with accessory pyrrhotite andcalcite. Most matrix sites contained some forsterite, likelyoriginating from small fragments of chondrules and/or amoeboidolivine aggregates (AOA). Coarse-grained, forsterite-dominatedclasts are identified as chondrules and/or AOA. A white-coloured

millimetre-scale inclusion bearing spinel and calcite may be arelict CAI (point #1 in Table 1, Fig. 1).

Fig. 2 illustrates a typical in situ mXRD analysis point (#34)containing matrix and a forsterite-bearing aggregate (shown inFig. 2B). The coarse-grained nature of the forsterite aggregate maybe inferred from the pattern of discrete diffraction spots visible inthe GADDS image. Polycrystalline rings are due to the prominentfine-grained matrix minerals saponite, pyrrhotite, and magnetite.These analyses demonstrate the potential of the mXRD techniquefor rapid in situ mineral and textural/microstructural character-ization, as a reconnaissance investigative step prior to thin sectionpreparation, subsampling or crushing and digestion.

6. Multi-method examination of MG-02 polished thin sections

Mosaic SEM-BSE images for six polished thin sections from aTagish Lake MG-02 fragment are shown in Fig. 3. Fine-grainedmatrix dominates most of the sections. A wide variety ofinclusions can be seen, including chondrules, AOA, isolated olivinegrains, large assemblages of sulfide grains, and a carbonate nodulewhich we interpret as a CAI aqueously altered during parent bodyprocessing. Many of the large inclusions are surrounded byaccretionary rims. Fusion crust occurs as a thin band of brighter,sulfide-rich vesicular material containing euhedral, chemicallyzoned olivine grains along smoothly curved edges of sections A, B,C, D, and H. Coordinated crystal structural (mXRD) and chemical(SEM-EDX, SEM-CL, EPMA) analyses were made for many of thefeatures, and are discussed in more detail below.

6.1. Matrix

Fine-grained matrix is the most abundant component of theTagish Lake meteorite. The matrix consists of very fine-grainedFe-bearing saponite clay, together with fine-grained magnetite,pyrrhotite, and siderite. These observations are consistent withprevious studies of other Tagish Lake samples (Zolensky et al.,2002; Simon and Grossman, 2003; Greshake et al., 2005). Thematrix is quite uniform in both mineralogy (as observed by mXRD)and elemental composition (as observed by SEM-EDX and EPMA).The predominance of siderite as the matrix carbonate and the

Fig. 4. Chondrule with accretionary rim in thin section D (Fig. 3). The lower part of the chondrule contains bars of forsteritic olivine surrounded by Ca- and Al-rich material

that is likely a mixture of mesostasis glass and breakdown products thereof. Also visible are small spherules of Fe, Ni metal and Fe sulfides within chondrule forsterite. The

CL behaviour of this chondrule is shown in Fig. 8.

Table 1Results from in situ mXRD analyses of selected 500 mm spots on the intact Tagish Lake MM-02 fragment.

Spots Phases identified by XRD Interpretation

1 Forsterite, magnetite, spinel, dolomite, pyrrhotite Aqueously altered CAI

2, 6, 8, 11, 13, 19 Saponite, magnetite, siderite, pyrrhotite Matrix

3, 4, 18, 27–30, 32, 33 Forsterite, magnetite, saponite, pyrrhotite Chondrule/AOA+matrix

7, 9 Forsterite, magnetite Chondrule/AOA

10 Forsterite, magnetite, pyrrhotite Chondrule/AOA+matrix

12 Forsterite, magnetite, siderite Carbonate nodule+matrix

14 Magnetite, saponite, siderite, forsterite Chondrule/AOA+matrix

15, 17 Forsterite, magnetite, saponite Chondrule/AOA+matrix

16 Forsterite, magnetite, siderite, saponite Chondrule/AOA+matrix

20 Siderite, saponite, calcite Carbonate nodule

21 Forsterite, saponite, magnetite Chondrule/AOA+matrix

22, 25, 26, 31, 34 Magnetite, saponite, siderite, forsterite Matrix

23 Forsterite Chondrule

24 Siderite, saponite, calcite, pyrrhotite Carbonate nodule+matrix


high magnetite content indicate that these samples correspond tothe carbonate-poor lithology described by Zolensky et al. (2002).Calcite occurs scattered throughout the matrix as fine-grainedrounded nodules, typically several microns in diameter (e.g.,matrix areas in Fig. 4 and individual nodules in Fig. 5). Sideritedoes not appear to form discrete nodules but instead occurs asvery fine-grained material dispersed throughout the Tagish Lakematrix. No veins of aqueous alteration products have beenobserved in the samples studied, perhaps indicating that fluidmigration through the Tagish Lake matrix was primarily viaseepage through pore space. This model of fluid migration in

Tagish Lake is consistent with its very high porosity of �40%(Hildebrand et al., 2006).

6.2. Accretionary rims or mantles

Many inclusions in Tagish Lake are surrounded by a mantle offine-grained phyllosilicates, oxides, and sulfides, mineralogicallyand chemically very similar to the surrounding matrix but ofsomewhat lower porosity (Fig. 4). It has been suggested that thesemantles are the alteration product of primary fine-grained nebular

Fig. 5. Backscattered electron images of representative calcite nodules in the

Tagish Lake matrix. Subangular to rounded calcite nodules are scattered

throughout matrix areas, but are very rare in the fine-grained and less porous

rims surrounding most nebular clasts (chondrules, AOA, CAI). Calcite precipitation

was likely restricted to areas of higher porosity than the fine-grained rims. No

instances of veins of calcite or other carbonates have been observed, however, the

brecciated nature of Tagish Lake makes it impossible to be certain that such veins

did not exist at one time. (A) Calcite grain with pronounced zoning due to

variations in Mn content. Note the cluster of magnetite framboids, possibly a

pseudomorphic replacement of sulfide. (B) Unzoned calcite grain.


dust rims that formed by accretion of fine particles onto chondrules,CAI, AOA, and other nebular clasts before their incorporation intothe Tagish Lake parent body and subsequent aqueous alteration(e.g., Greshake et al., 2005). Our observations support thishypothesis, as we have observed accretionary rims on only ‘nebular’inclusions (Chondrules, AOA and CAI), and not surrounding grains ofsulfides, magnetite, and calcite that are interpreted as products ofparent body aqueous processing. The accretionary rims are notablysimilar in composition and mineralogy to the surrounding matrix,but are notably less porous. Calcite nodules are rare or absent in theaccretionary rims. Some rims are enriched in Fe, Ni sulfides(pyrrhotite and pentlandite), whereas others are enriched inmagnetite, and some display concentric zones of alternating sulfideand magnetite enrichment (in no consistent order). The variationsin sulfide and magnetite content in the accretionary rims may relateto variations in fS and fO2 in the nebular source regions of theinclusions that affected their parent body alteration history; or tofine scale variations in geochemical conditions within the TagishLake parent object during aqueous processing. In some cases,accretionary mantles appear to have protected the enclosedinclusion from aqueous alteration on the Tagish Lake parent body,possibly due to the lower porosity of the accretionary rimsinhibiting the diffusion of fluids.

6.3. Olivine-rich inclusions

Three populations of olivine have been observed in the TagishLake sections: chondrules (Figs. 4 and 6), Amoeboid OlivineAggregates (AOA—Fig. 7) and isolated olivine grains (Table 2).Olivines in chondrules and AOA are usually nearly pureend-member forsterite, whereas some of the isolated grains aresomewhat more Fe-rich, consistent with the olivine populationsreported in previous studies (Zolensky et al., 2002; Simon andGrossman, 2003). Chondrules and AOA are usually mantled by anaccretionary rim (e.g., Figs. 4, 6, and 7).

Fig. 6 shows a mostly unaltered porphyritic olivine chondrule.Chondrule olivine crystals often contain small (�1–10 mm)rounded blebs of Fe, Ni metal, sulfides, and/or magnetite. Somechondrules contain unaltered mesostasis glass enriched in Al andCa (Fig. 6B; Table 3). Porphyritic olivine chondrules are the mostcommon chondrule type in Tagish Lake, followed by barredolivine, consistent with previous reports (Zolensky et al., 2002;Simon and Grossman, 2003). We have not observed otherchondrule types in these samples.

Micro XRD of the AOA (Fig. 7B) shows the presence of enstatite,consistent with EPMA analysis of this AOA, which shows one grainwith an enstatite composition (Mg0.966Fe0.008Ca0.009Mn0.001-

Ti0.002Cr0.003Ni0.001Al0.004Si0.974O6). Olivine grains in many AOAare slightly Cr-enriched relative to chondrules and isolated olivinegrains (Table 2), indicating that they have not undergonesubstantial heating since their formation (Grossman, 2004). Thepresence of enstatite in AOA has been interpreted as the result ofthe reaction of forsterite with nebular SiO gas during rapid high-temperature thermal processing of AOA in the solar nebula, andmay support a linkage between chondrules and refractoryinclusions (e.g., Krot et al., 2004, 2005).

Olivines in chondrules and AOA are CL active, and commonlyshow distinct zoning with a bright interior and dark rim (Fig. 8).At least two types of dark red CL zoning are noted based onrelationship to grain surfaces. The dominant zoning, present inthe central region of the grain where olivine grain diameters arelargest (type 1), is parallel and concordant to grain boundaries,has width of tens of microns and a gradational, diffuse inner zoneedge. Haloes of equal CL intensity and colour are sometimes seenalong inner grain boundaries in contact with large inclusions(Fig. 8). The second type is distributed linearly, discordant to grainboundaries. In some cases these linear features are associatedwith, and symmetric about, fractures cross cutting olivine grains.It is noted, however, that not all fractures coincide with dark redCL lineaments. The type 2 discordant CL zones indicate that some,and perhaps all, of the dark red zoning formed after chondrulecrystallization, and reflects a secondary metamorphic processperhaps related to chemical diffusion and alteration. Somefracturing in the chondrule clearly postdates this event. None ofthe isolated olivine grains observed were CL active, which mayrelate to their generally higher Fe content, as Fe quenches CLactivity in many minerals (e.g., Coulson et al., 2007; Lee et al.,2007). No compositional gradients matching the CL zoning in thepure forsterite have been identified by EPMA or EDX.

6.4. Relict CAI and spinel-rich inclusions

At least one relict CAI has been identified in thin section (Figs. 9and 10). It consists of an assemblage of magnesioaluminate spinelcomplexly intergrown with phyllosilicates and dolomite, asidentified by mXRD. Spinel is the major remaining primary nebularphase. A few tiny (�1–3 mm in their longest dimension) crystals of aCa, Ti oxide phase are present, which we interpret to be perovskite(Fig. 11). The perovskite has not been quantitatively analysed by

Fig. 6. (A) Backscattered electron images and chemical maps for a porphyritic olivine chondrule in thin section A (Fig. 3). Large circles I–III are the nominal 500 mm

footprints of the mXRD beam, small circles 1–5 are the nominal footprints of the 50 mm mXRD beam targeted on areas of interest. The results of these analyses are

summarized in Table 2. Note as well the near-total absence of calcite nodules in the accretionary rim; (B) Micro XRD analysis of point I in the chondrule is shown in A and

plotted as an integrated X-ray diffractogram with inset raw two-dimensional GADDS images. The broad feature between 121 and 271 in 2y is characteristic of an amorphous

phase, probably mesostasis glass. Note the large, high-intensity diffraction spots due to forsterite. The intensity is plotted logarithmically to allow the weak amorphous

signal to be distinguished from the intense forsterite peaks. ICDD pattern matches siderite—00-029-0696, magnetite—00-019-0629, forsterite—00-34-0189,

pyrrhotite—00-029-0723, phyllosilicates (saponite-dominated)—00-012-0160.


EPMA due to its small crystal size. Spinel and perovskite are moreresistant to aqueous alteration than many other CAI phases (Davisand Richter, 2003). Hibonite and melilite, for example, readily alterto phyllosilicates and carbonates, respectively (Simon andGrossman, 2003). Melilite more typically alters to calcite (Simonand Grossman, 2003; Zolensky et al., 2008a); the presence ofdolomite may suggest processing in Mg-rich fluids. Alternatively theprimary CAI assemblage could have contained large amounts of Mg-rich phases such as akermanite-rich melilite. Similar objects in otherTagish Lake samples have been interpreted as relict CAI by Zolensky

et al. (2002). The compositions of the spinel, phyllosilicates, anddolomite in the relict CAI are presented in Table 4.Magnesioaluminate spinel in the relict CAI shows intense red CL,with possible zonation in intensity from bright edges to slightlydarker interiors (Fig. 10). No corresponding chemical zonation wasobserved in either SEM-BSE or SEM-EDX images, which indicatesthat a trace element and/or structural mechanism may beresponsible for the variability in spinel CL.

Rare occurrences of other magnesioaluminate spinel-rich inclu-sions were observed, having highly irregular shapes, typically �tens

Fig. 7. (A) An amoeboid olivine aggregate (AOA) from thin section A (Fig. 3) with a sulfide-rich core; note also the distinct accretionary rim and the Cr enrichment in the

AOA forsterite. Backscattered electron image and elemental maps obtained by SEM-EDX. The white circles I, II, III above are the nominal beam footprint of three mXRD

analysis points. The mXRD data also show evidence for; (B) high-resolution BSE image of the area in the small white box in A showing enstatite (En), forsterite (Fo), and

traces of sulfides. (C) Micro XRD data for the three circled areas in A. Background subtracted, count time 8 min. Patterns are vertically offset by 50 counts for clarity. ICDD

pattern matches: forsterite—00-34-018, pyrrhotite—00-029-0723, magnetite—00-019-0629, orthopyroxene (orthoenstatite)—01-083-0666. Points II and III have matches

with orthoenstatite, note especially the prominent (6 1 0) reflection at 31.051 2y (d¼2.8802 A), marked by the arrow.


by �hundreds of microns. Unlike the relict CAI described above,spinel-rich inclusions lack abundant dolomite and phyllosilicates,and consist almost entirely of angular spinel crystals.

6.5. Carbonates

The relict CAI described above is the largest carbonate-bearinginclusion we have observed in this study. Other occurrences ofcarbonate inclusions include small (typically a few tens ofmicrons) subangular to rounded nodules of calcite scatteredthroughout the matrix. We have also observed a few dolomitenodules similar in size and shape to the calcite nodules. The

carbonate content of the Tagish Lake matrix is dominated by veryfine-grained siderite, as discussed above. Fig. 12 displays theresults of EPMA analysis of carbonate in the relict CAI discussedabove, as well as in several calcite nodules, an isolated dolomitegrain and matrix siderite (analyses summarized in Table 5). Theisolated dolomite grain has a slightly more Ca-rich compositionthan the relict CAI dolomite.

6.6. Sulfides

Sulfides in the Tagish Lake meteorite display a wide range ofmorphologies (Fig. 13) and considerable variability in composition

Table 2Average olivine compositions by EPMA for the three populations of olivine

observed in the Tagish Lake MG-02 thin section. Oxygen contents were calculated

by difference.

Isolated chondrule AOA

Average olivine compositions (5 grains each type), wt% oxide

SiO2 36.803 42.135 42.176

TiO2 0.006 0.055 0.052

Al2O3 0.024 0.104 0.011

Cr2O3 0.320 0.296 0.623

FeO 30.866 0.560 0.806

MgO 31.394 56.376 55.702

MnO 0.273 0.028 0.126

K2O 0.019 0.017 0.020

CaO 0.279 0.515 0.263

Na2O 0.021 0.006 0.004

NiO 0.140 0.047 0.030

V2O5 0.030 0.006 0.022

Total 100.144 100.141 99.813

Cations per 4 oxygens

Si 0.999 0.991 0.996

Ti 0.000 0.001 0.001

Al 0.001 0.003 0.000

Cr 0.007 0.006 0.012

Fe 0.701 0.011 0.016

Mg 1.270 1.977 1.961

Mn 0.006 0.001 0.003

K 0.001 0.001 0.001

Ca 0.008 0.013 0.007

Na 0.001 0.000 0.000

Ni 0.003 0.001 0.001

V 0.001 0.000 0.000

O 4 4 4

Cation sum 2.997 3.004 2.997

Fo 64.25 99.45 99.19

Fa 35.75 0.55 0.81

Table 3Results of mXRD analyses for the locations marked in Fig. 6A. The presence of

glassy mesostasis at site I is inferred from the presence of a broad amorphous

feature in the mXRD measurement (Fig. 6B) and from the presence of optically

isotropic material.

Spots Phases identified Interpretation

I Forsterite, magnetite, phyllosilicates,

mesostasis

Chondrule olivine, possible

mesostasis

II Forsterite, pentlandite, saponite Chondrule olivine+sulfide

inclusions

III Forsterite, magnetite, phyllosilicates,

siderite, pyrrhotite

Chondrule+accretionary

rim

1, 3 Forsterite Chondrule olivine

2 Forsterite, magnetite Magnetite inclusion

4 Forsterite, pentlandite, pyrrhotite Sulfide inclusion

5 Saponite, pyrrhotite, magnetite Accretionary rim

Fig. 8. Chondrule olivine with distinct domainal variation in CL colour and

emission intensity which does not appear to correspond to chemical variation

resolvable with EPMA. Note the two types of dark CL zoning distribution; parallel

to outer grain boundaries (type 1) and linear zones sometimes associated with

fracturing (type 2). The latter suggests that zoning is not primary but reflects a

secondary metamorphic process. The small green CL active objects are likely an

artefact due to residual polishing compound. This chondrule is in polished section

D (Fig. 3), EDX chemical maps and BSE imagery for this chondrule are presented

in Fig. 4.


(Table 6). Observed sulfide occurrences include:

(1)
Small (up to a few tens of microns), irregular to roundedaggregates of subhedral sulfide crystals, in which individualcrystals range from a few microns (Fig. 13A and B) up toseveral tens of microns each in size (Fig. 13C). Some of theseaggregates appear ‘hollow’, in that they enclose areas ofmatrix (Fig. 13A and C) while others are solid masses(Fig. 13B). It is unclear as to whether this is simply a resultof imaging two-dimensional slices through one type of objectin different orientations. It is also unclear whether there is arelationship between these and the submicron matrix sulfidegrains described below. The irregular and rounded sulfide
aggregates are more Ni-rich than other Tagish Lake sulfides inthis study (Table 6).

(2)
Spheroidal inclusions of sulfide in chondrule forsterite grains(Fig. 13D). These bear some similarity in size and shape to Fe,Ni metal spherules, which also occur in chondrule forsteritegrains sometimes occurring within the same crystal(Fig. 13D). These spheroidal inclusions may have formed fromdroplets of sulfide melt within chondrule forsterite.
(3)
A ‘flaky’ sulfide morphology forming large (up to �100 mm intheir longest dimension) irregular flaky masses, composed of�2�10 mm tabular to platy crystals (Fig. 13F and G). Thesesulfides are Ni-poor and therefore likely pyrrhotite-rich. Theflaky sulfides are texturally and chemically similar to sulfidesobserved in the core of a large AOA.
(4)
Large subangular to rounded aggregates of sulfide crystals upto several hundred microns in their longest dimension,composed of relatively large subhedral crystals of pyrrhotitewith subsidiary pentlandite (Fig. 13 E).
(5)
The core of a large AOA located in section A consists largely ofFe, Ni sulfide (Fig. 7). Micro XRD of this AOA confirms thepresence of forsteritic olivine and pyrrhotite (Fig. 7B). TheAOA core has a flaky texture very similar to that of the ‘flakysulfide’ grains observed as isolated grains within the matrix(cf. Fig. 13F, G, and H). Additionally, the composition of theflaky sulfides is very similar to that of the AOA core (Table 6).The ‘flaky sulfides’ may therefore be the result of disaggrega-tion of AOA, probably during regolith processing. Sulfide-richAOA cores are probably the result of in situ sulfidation of Fe, Nimetal after AOA formation, either in the nebula or duringparent body metamorphism.
(6)
Minute rounded sulfide grains, less than 1 mm in extent, occurscattered throughout the matrix, and are highly concentratedin some accretionary rims (e.g., Figs. 4 and 6). These minutesulfides are too small for reliable EPMA analysis.
Pyrrhotite is the most common sulfide in the Tagish Lakemeteorite; most other sulfide grains are pentlandite. This isconsistent with the mXRD results for the intact fragment as well as

Fig. 9. This assemblage of dolomite, spinel, and Fe-poor (relative to matrix) phyllosilicates from thin section B (Fig. 3) is interpreted as an aqueously altered CAI.

(A) Backscattered electron image (central panel) and chemical maps obtained by SEM-EDX. The white circle in BSE image is the nominal footprint of the 50 mm mXRD beam

(Fig. 11). Note the correlation of Al and Mg (spinel), and the anti-correlation of Ca with Al and Si (dolomite). This grain is surrounded by a sulfide-rich accretionary rim. Note

also that the small nodules of calcite in the matrix do not occur in the accretionary rim. (B) Micro XRD data showing matches to magnesioaluminate spinel, dolomite, and

phyllosilicates. Integrated mXRD pattern for the region of the altered CAI circled in A, with phases identified using the ICDD database. C: GADDS image corresponding to the

region circled in A and integrated to produce the diffractogram in B. The continuous Debye rings in the GADDS image are indicative of fine-grained, well-crystallized

minerals. ICDD pattern matches: spinel—00-021-1152, dolomite—00-036-0426, phyllosilicate, best ICDD match to Al-rich saponite 00-010-0426.


analyses of the thin sections, and agrees well with previousstudies (Zolensky et al., 2002; Boctor et al., 2003; Bland et al.,2004; Zolensky et al., 2008b; Izawa et al., 2010). Table 6summarizes the results of EPMA analysis of the various popula-tions of Tagish Lake sulfides.

Some, but not all large sulfide grain assemblages appear to beenriched in P as shown by the SEM-EDX chemical maps (Fig. 14).Electron microprobe analysis of the ‘flaky’ sulfides shows traces of P(Table 6). Areas with a high concentration of fine-grained sulfides,

such as the S-rich areas in accretionary rims, do not appear to becorrespondingly P-enriched (Fig. 6). This argues against spectral lineoverlap as the cause of this signal. Phosphorous-enriched sulfidegrains have been reported in numerous meteorites including CMand CI chondrites, and are proposed to have formed by sulfidation ofFe–Ni metal in the presolar nebula or by direct condensation ofP-enriched pyrrhotite or troilite in a reducing, high fS region of thenebula (Zolensky and Thomas, 1995; Lauretta et al., 1997, 1998;Boctor et al., 2002; Goreva and Lauretta, 2006).

Fig. 10. Cathodoluminescence in magnesioaluminate spinel contained in the altered CAI (Fig. 9). Boxed areas A and B in the top series BSE and CL images are enlarged

below. The magnesioaluminate spinel is CL active, mainly in the red channel. Enlargements in C and D show possible zonation in CL activity from brighter edges to dimmer

cores; SEM-BSE imaging and SEM-EDX mapping (Fig. 9) do not show chemical corresponding variations pointing to a possible structural and/or trace element mechanism

for CL activation. The faint streaking in the CL images is an artefact due to long-lived excitation (phosphorescence) in the spinel.


6.7. Magnetite, other oxides and metal

Magnetite is by far the dominant oxide phase in Tagish Lake. Itoccurs throughout the matrix as small clusters of micron-scaleframboids (Fig. 15A), as much larger (several 10 s of microns)euhedral to subhedral grains which occur isolated or as irregularclumps, and as euhedral crystals within an isolated grain of dolomitethat suggest that these two minerals formed in equilibrium(Fig. 15B). Table 7 summarizes the EPMA results for Tagish Lakemagnetite. There are no clear compositional trends between thevarious occurrences of magnetite. In many instances, magnetiteappears to pseudomorphically replace sulfides (Fig. 15C).

Minute (submicron) crystals of a Fe, Cr oxide phase have beenobserved associated with aqueously altered chondrule olivine. Thesegrains have not been analysed by EPMA due to their size, however,EDX spot analysis indicates that they contain more Cr than Fe,suggesting that this phase is chromite (Fig. 16). Chromite may be aproduct of parent body aqueous alteration of primary Cr-bearingforsterite. The loss of Cr from chondrule olivine has been shown tocorrelate with alteration in many chondrite types (Grossman, 2004).

Free Fe, Ni metal spherules up to a few tens of microns in diameterhave been observed in several chondrule forsterite grains (Figs. 6Aand 13C). Some free metal grains contain inclusions of an Fe, Niphosphide, probably schriebersite. The Fe, Ni metal spherules and theenclosing forsterite show no evidence of reaction with each other,

indicating that the metal and forsterite were near equilibrium andhave not subsequently experienced significant element mobility. Inparticular, SEM-BSE, EDX, and CL imaging show no evidence ofchemical exchange (Fe diffusion for instance) between metal andforsterite (Figs. 6A and 13C).

6.8. Fusion crust

A thin (�hundreds of microns) layer of vesicular, sulfide- andmagnetite-rich material was observed at the edges of all but one (G)thin section (Fig. 17). This material contains many small (�10 s ofmicrons) euhedral grains of olivine, many of which are elongated, aswould be expected from rapid crystallization. Most of these grainsshow strong chemical zonation with Fe-rich rims. In one locationobserved, the fusion crust interacted with a forsterite-rich chondrule(Fig. 17B). Here, rounded grains of olivine appear to have partiallymelted and reacted with the surrounding material producing Fe-richreaction rims around very Fe-poor cores. The fusion crust is also richin very fine-grained magnetite with dendritic or snowflakemorphology. Tagish Lake fusion crust is texturally, mineralogically,and compositionally heterogeneous.

As has recently been pointed out by Thaisen and Taylor (2009)in a study of lunar and eucrite meteorites, the composition ofmeteorite fusion crust is not necessarily representative of the bulk


rock, or even of the bulk refractory elements. Furthermore, thepresence of euhedral olivine crystals in the Tagish Lake fusioncrust should be borne in mind during studies using grainsseparated from their original context: such crystals could bemistaken for a pre-terrestrial feature, and would likely showchemical and isotopic evidence for a very different environmentof formation than other Tagish Lake olivine.

6.9. Shock stage assessment using optical petrography

Petrographic examination of Tagish Lake polished sections intransmitted light has enabled us to assign a shock stageto these samples following the methods developed by Scott,

Fig. 11. Backscattered electron image showing a small Ca, Ti oxide grain hosted in

magnesioaluminate spinel interpreted as a perovskite grain, within magnesioalu-

minate spinel in the relict CAI (Figs. 9 and 10).

Table 4Compositions of the main constituents of the relict CAI by EPMA.

Relict CAI spinel composition—average of 10 analyses, values in wt% oxide

Al2O3 MgO FeO Cr2O3 MnO ZnO

72.72 26.505 0.233 0.097 0.018 0.023

Relict CAI phyllosilicate composition—average of 8 analyses, values in wt% oxide

SiO2 TiO2 Al2O3 Cr2O3 FeO MgO

43.20 0.27 4.82 0.08 9.39 25.99

Relict CAI dolomite composition—average of 14 analyses, values in wt% carbonate

FeCO3 MgCO3 MnCO3 CaCO3 BaCO3 SrCO3

5.85 38.21 1.41 52.49 0.03 0.02

Stoffler and Keil (Scott et al., 1992; Stoffler et al., 1992). Olivinecrystals observed in Tagish Lake display sharp optical extinction,with occasional irregular fracturing, placing our Tagish Lakepolished sections in shock stage S1 (very weakly shocked orunshocked). It is notable that mXRD GADDS images containingolivine reflections also show no evidence of strain-relatedasterism or ‘streakiness’ (Flemming, 2007), consistent with thelow degree of shock-induced lattice damage inferred from opticalmicroscopy.

7. Discussion and conclusions

7.1. Micro XRD and other coordinated reconnaissance techniques

Micro X-ray diffraction has rapidly and reliably identified themajor mineral phases in situ in the Tagish Lake carbonaceouschondrite and provided texture and crystallinity information via2-D GADDS images. Phase identification enables the preliminarycharacterization of chondrules, CAI, AOA, carbonate nodules andother objects of interest. We have demonstrated the capability ofsurveying carbonaceous chondrites without sample preparation,of great potential value for the in situ, non-destructive study of thepristine Tagish Lake material, and for other precious samples,such as planetary sample returns.

NiO SiO2 TiO2 Nb2O5 V2O5 Total

0.037 0.08 0.065 0.019 0.031 99.828

MnO K2O CaO Na2O Total

0.05 0.16 0.45 0.34 84.74

CrCO3 Total

0.03 98.03

Fig. 12. Ternary composition diagram for matrix calcite nodules, relict CAI

dolomite, isolated dolomite, and matrix siderite, as determined by EPMA.

Fig. 13. Variations in sulfide morphology observed in the Tagish Lake thin sections. (A) large ‘hollow’ aggregate of sulfide crystals. (B) ‘solid’ aggregate of sulfide crystals.

(C) ‘flaky’ sulfide. (D) Heart-shaped, large irregular aggregate of sulfide crystals. (E) ‘flaky’ sulfide. (F) Rounded, possibly ‘hollow’ (i.e., enclosing an area of matrix material)

sulfide aggregate. (G) Chondrule forsterite crystal containing sulfide and metal spherules. (H) Rounded ‘solid’ sulfide aggregate. It is possible that the ‘hollow’ and ‘solid’

variations are different two-dimensional slices through similar assemblages.

Table 5Average Tagish Lake carbonate analyses by EPMA, values in wt%.

FeCO3 MgCO3 MnCO3 CaCO3 BaCO3 SrCO3 CrCO3 Total

Matrix calcite 1.04 0.44 0.02 94.92 0.05 0.04 0.04 96.55

Relict CAI dolomite 5.85 38.21 1.41 52.49 0.03 0.02 0.03 98.03

Isolated dolomite 5.24 34.74 2.58 59.33 0.00 0.01 0.00 101.91

Matrix siderite 68.00 19.70 0.00 4.03 0.00 0.00 0.13 91.87


The mXRD technique is an excellent complement to othermicroanalytical techniques, providing mineralogical context andpoint-correlated crystal structure information, enabling more

confident mineral identification. Non-destructive, in situ recon-naissance can identify objects such as CAI, chondrules, AOA, andcarbonate nodules for more in-depth analyses using other,

Table 6Tagish Lake sulfide compositions by EPMA for the various modes of occurrence observed, in wt% atomic. Sulfur content was measured, rather than calculated by difference.

Sulfide occurrence

Sulfide aggregates Sulfide inclusions in

chondrule forsterite

Large sulfide grain

assemblages

AOA cores Flaky sulfides

Cr 0.03 0.11 0.09 0.05 0.03

Cu 0.01 0.02 0.02 0.00 0.04

Zn 0.01 0.00 0.02 0.00 0.01

Fe 22.73 32.99 28.58 33.12 32.44

Co 1.03 0.08 0.70 0.01 0.01

Ni 13.96 1.81 7.40 1.92 2.79

S 62.10 64.75 63.26 64.77 64.29

Ti 0.04 0.02 0.02 0.00 0.03

Mn 0.05 0.00 0.01 0.01 0.02

P 0.00 0.00 0.00 0.01 0.01

V 0.01 0.05 0.07 0.00 0.00

Total 99.94 99.83 100.18 99.89 99.67

Fig. 14. An assemblage of phyllosilicates, rounded forsterite grains, and calcite surrounded by an accretionary mantle dominated by Fe-bearing saponite, magnetite, and

pyrrhotite. The cracking along the accretionary rim may have occurred during sectioning, and probably reflects the competency contrast between the accretionary rim and

the more porous matrix. Backscattered electron image and elemental maps obtained by SEM-EDX. This is probably a chondrule in which the primary phases have largely

been replaced by phyllosilicates during aqueous alteration. At the upper right is a large Fe–Ni–S grain. Calcite occurs throughout the matrix as small Ca-rich, Si- and Al-poor

nodules; note that the accretionary rim contains few calcite nodules compared to the surrounding matrix.


potentially more invasive methods. Such preliminary reconnais-sance will assist in minimizing sample losses, for example duringthin section preparation or subsampling.

7.2. Aqueous processing on the Tagish Lake parent body

Our combined crystal structural and chemical reconnaissanceof the Tagish Lake carbonaceous chondrite has produced bothnew insights and new questions regarding the fluid alteration

history on the Tagish Lake parent body. Hanowski and Brearley(2001) developed a scheme of alteration in CM chondriteswherein the fluid composition evolves from highly Fe-rich tohighly Mg-rich, as different mineral phases are progressivelyaltered. These authors suggest alteration in four stages, whichmay grade into one another, as follows: (1) reaction of primaryfine-grained troilite and metal to form tochilinite, pyrrhotite andpentlandite, and alteration of most chondrule mesostasis tophyllosilicates, with a very Fe-rich fluid composition; (2) alterationof silicate grains beginning with fayalitic olivine and Ca-poor

Fig. 15. Backscattered electron images of Tagish Lake magnetite grains. (A) Framboidal magnetite; the cracks likely formed during thin section preparation. (B) Euhedral

magnetite crystal within an isolated dolomite crystal, and a cluster of rounded magnetite crystals. (C) Clusters of large, rounded magnetite crystals surrounded by matrix.

(D) Irregular magnetite aggregate.

Table 7Tagish Lake magnetite compositions by EPMA, Fe2O3 calculated based on ideal stoichiometry.

wt% oxide

FeO Fe2O3 TiO2 MgO Cr2O3 MnO SiO2 Al2O3 ZnO NiO Nb2O5 V2O5 Total

Rounded magnetite grains 31.17 69.27 0.02 0.00 0.02 0.00 0.07 0.00 0.02 0.03 0.00 0.04 100.65

Framboidal magnetite 30.78 68.41 0.01 0.05 0.01 0.02 0.18 0.01 0.03 0.06 0.02 0.01 99.59

Magnetite in isolated dolomite 30.78 68.41 0.00 0.00 0.00 0.00 0.05 0.00 0.03 0.06 0.00 0.00 99.32

Cations per 4 oxygens

Fe2+ Fe3 + Ti Mg Cr Mn Si Al Zn Ni Nb V Cation sum

Rounded magnetite grains 1.00 1.99 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.00

Framboidal magnetite 0.99 1.99 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 3.00

Magnetite in isolated dolomite 1.00 2.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.00

Fig. 16. Backscattered electron image of minute Cr, Fe oxide grains, probably

chromite, near the interface of chondrule forsterite and their phyllosilicate-

dominated aqueous alteration products. These Cr, Fe oxide grains are only

observed associated with altered olivine, and are probably the result of parent

body aqueous alteration of Cr-bearing olivine.


clinopyroxene, concentrated at grain boundaries and alongfractures and/or cleavage planes and beginning with a fluidcomposition similar to that of phase 1 with progressively lower Feconcentrations as alteration proceeds; (3) extensive alteration offayalitic olivine and Ca-poor clinopyroxene with incipient altera-tion of forsterite and orthopyroxene and a Mg-rich fluid withsome compositional equilibration between the phyllosilicatealteration products of phases 1 and 2 and the new Mg-rich fluidcomposition; and (4) almost complete alteration of fayalite andCa-poor clinopyroxene, and alteration of forsterite and enstatitewith greater alteration of highly fractured and/or fine-grainedMg-rich silicates, a very Mg-rich fluid composition and extensiveequilibration of previous phyllosilicate alteration products of theprevious stages with the fluid phase.

The paragenetic sequence inferred from our observations forthe Tagish Lake materials in this study is similar to that proposedfor CM2 chondrites and is consistent with that of Zolensky et al.(2002) for the carbonate-poor Tagish Lake lithology. Earlyalteration events in Tagish Lake included the alteration of primaryphases probably including troilite, kamacite, and chondrulemesostasis to form pyrrhotite, pentlandite and/or Fe, Ni

Fig. 17. Vesicular, sulfide-rich material near the edge of thin section H, interpreted as fusion crust. (A) Note the rounded olivine grains reacting with the fusion crust melt,

probably originating from the rounded inclusion below. (B) Abundant euhedral olivine crystals, often with elongated morphologies consistent with rapid crystallization.

Chemical zonation is apparent in most fusion crust olivine crystals.


monosulfide solid solution and Fe-rich phyllosilicates. This wasfollowed by progressive alteration of more susceptible crystallinesilicates probably including fayalitic olivine and clinoenstatiteforming additional phyllosilicates. Carbonate minerals includingcalcite, dolomite, and siderite were formed in pore spaces and aretherefore much rarer in the fine-grained accretionary rims. Calciteand dolomite are commonly zoned in their Mn contents, possiblyindicating an evolution towards more Mn-rich fluid compositionlater in the alteration sequence. Magnetite, including pseudo-morphs after earlier pyrrhotite, formed contemporaneously withcarbonates as indicated by equilibrium textures between magne-tite, dolomite, and calcite. Very fine-grained sulfides, probablydominated by pyrrhotite, are also formed later in the alterationsequence as evidenced by their common enrichment in fine-grained accretionary rims. Highly resistant phases includingcoarse-grained forsterite and enstatite were incipiently alteredtowards the end of Tagish Lake aqueous processing. The productsof each successive step in the alteration sequence were also atleast partially equilibrated with the evolved fluid; this wasprobably particularly true for the fine-grained phyllosilicates thatdominate the Tagish Lake matrix.

Tagish Lake is a brecciated meteorite, which complicates anyinterpretation of relative degrees of aqueous alteration betweenvarious clast populations; however, our observations indicate thatthe majority of minimally altered olivine-rich clasts includingchondrules and AOA are dominated by coarse-grained forsteriticolivine. Pyroxenes are relatively rare, and where present consist ofMg-rich orthopyroxene and show abundant evidence of altera-tion. Based on the petrographic/textural examination and mineralchemistry reported in this study, Tagish Lake is a C2 chondritewith affinities to both CI and CM chondrites that has experienced

aqueous alteration conditions similar to those of the CM2chondrites in phase 4 of the scheme of Hanowski and Brearley(2001) outlined above. If this is indeed the case, then it is probablethat the composition of the altering fluid evolved from Fe-rich toMg-rich.

The presence of small amounts of material corresponding tolower degrees of alteration is likely the result of mixing betweenmaterials with different alteration histories during regolithprocessing on the parent asteroid. These rare less-alteredmaterials, including coarse forsterite-dominated chondrules withrelict mesostasis and isolated fayalitic olivine grains can all fitwithin the general scheme of alteration presented here.

Magnesium sulfates produced by terrestrial alteration havebeen documented in many carbonaceous chondrites (Gounelleand Zolensky, 2001). Searches for Mg-bearing salts and otherhighly soluble phases have thus far been inconclusive; they mayhave been lost due to leaching by melt water prior to recovery. In

situ mXRD and other suitable investigations of a wider range ofsamples, especially the pristine Tagish Lake material, prior to thinsection preparation may be more successful at locating solublephases.

7.3. Some new Tagish Lake research directions arising from this

reconnaissance

Three populations of CL active grains have been identified:relict CAI spinel, calcite, and forsterite in chondrules and AOA. Itremains to be seen whether the bright CL domains are related toprimary growth or are relict, partially overprinted by fluidprocessing/metamorphism. Although the process and material


change responsible for the linear dark red CL zones in forsteriticolivine remains cryptic, the discordance of these zones relative tograin boundaries and their spatial association with fracturesindicates post-chondrule alteration perhaps due to chemicalmetasomatic change at the trace element level. The mechanismscausing CL activity in these phases may be elucidated by high-sensitivity compositional measurements by techniques such asToF-SIMS, SHRIMP or nano-SIMS, and likely hold clues forunderstanding the history of these grains. Tagish Lake containsrare large grains of Fe, Ni sulfide consisting of pyrrhotite andminor pentlandite. Some of these sulfides show evidence forenrichment in P, which may be an artefact due to the presence ofphosphides such as schriebersite below the spatial resolution ofthe current data set. Alternatively, these may actually be P-enriched sulfide grains, possibly originating from sulfidation of Fe,Ni metal and/or phosphides in the nebula or on the Tagish Lakeparent body.

The identification of distinct populations of carbonates, oxides,and sulfides opens up the possibility of defining distinct isotopicand geochronological constituents in the Tagish Lake meteorite, tohelp unravel its parent body history and its record of precursorevents and processes at the beginning of the Solar System.

Acknowledgements

Reviews of this manuscript by Chris Herd and an anonymousreviewer led to substantial improvements and are gratefullyacknowledged. We thank Peter Brown, Alan Hildebrand, NeilMacRae, Sam Russell, Richard Herd, Ron Peterson Chris Herd,Martin Beech, and Penny King for fruitful discussions on TagishLake and primitive solar system bodies. This study formed part ofthe M.Sc. Thesis of MRMI, who acknowledges research supportfrom NSERC (USRA and CGS programs), the Walker MineralogicalClub, the Mineralogical Association of Canada, the Lunar andPlanetary Institute Career Development Award, the WesternGraduate Thesis Research Award, the Allan D. Edgar PetrologyAward, and the Robert & Ruth Lumsden Award. RLF, GS, and DEMwere supported by NSERC discovery grants. Additional fundingwas supplied by the Canada Foundation for Innovation (DEM, RLF)and the Western Academic Development Fund (RLF, PJAM). MRMIthanks Surface Science Western and Ross Davidson; the WesternNanofabrication Laboratory and Todd Simpson; Jim Renaud ofRenaud Geological Consulting; the Imaging Center at Texas TechUniversity (NSF Grant MRI 04-511), Callum Hetherington andMark Grimson for training, advice, and access to SEM instruments.We also thank Margriet ten Napel for editorial handling.

References

Armstrong, J.T., 1995. CITZAF: a package of correction programs for thequantitative electron microbeam X-ray analysis of thick polished materials,thin films, and particles. Microbeam Analysis 4, 177–200.

Baker, L., Franchi, I.A., Wright, I.P., Pillinger, C.T., 2002. The oxygen isotopiccomposition of water from Tagish Lake: its relationship to low-temperaturephases and to other carbonaceous chondrites. Meteoritics & Planetary Science37, 977–985.

Binet, L., Gourier, D., Derenne, S., Pizzarello, S., Becker, L., 2004. Diradicaloidsin the insoluble organic matter from the Tagish Lake meteorite: comparisonwith the orgueil and murchison meteorites. Meteoritics & Planetary Science39, 1649–1654.

Bland, P.A., Cressy, G., Menzies, O.N., 2004. Modal mineralogy of carbonaceouschondrites by X-ray diffraction and Mossbauer spectroscopy. Meteoritics &Planetary Science 39, 3–16.

Blinova, A., Herd, C.D.K., 2010 (abstract). Insights into the mineralogy of the TagishLake meteorite through EPMA, XRD and CL. 41st Lunar and Planetary ScienceConference.

Blinova, A., Herd, C.D.K., Zega, T., De Gregorio, B., Stroud, R., 2009 (abstract).Preliminary SEM and TEM study of pristine samples of Tagish Lake meteorite.40th Lunar and Planetary Science Conference.

Boctor, N.Z., Kurat, G., Alexander, C.M.O.’D., 2003 (abstract). Sulfide-oxideassemblage in Tagish Lake carbonaceous chondrite. 34th Lunar and PlanetaryScience Conference.

Boctor, N.Z., Kurat, G., Alexander, C.M.O.’D., Prewitt, C.T., 2002 (abstract). Sulfidemineral assemblages in Boriskino CM chondrite. 33rd Lunar and PlanetaryScience Conference.

Brearley, A.J., Jones, R.H., 1998. Chondritic meteorites. In: Papike, J.J. (Ed.), PlanetaryMaterials. Mineralogical Society of America, Washington, DC, pp. 398.

Brey, G.P., Kohler, T., 1990. Geothermobarometry in four-phase Lherzolites II. Newthermobarometers, and practical assessment of existing thermobarometers.Journal of Petrology 31, 1353–1378.

Brown, P.G., Hildebrand, A.R., Zolensky, M.E., Grady, M.M., Clayton, R.N., Mayeda,T.K., Tagliaferri, E., Spalding, R., Macrae, N.D., Hoffman, E.L., Mittlefehldt, D.W.,Wacker, J.F., Bird, J.A., Campbell, M.D., Carpenter, R., Gingerich, H., Glatiotis, M.,Greiner, E., Mazur, M.J., Mccausland, P.J.A., Plotkin, H., Mazur, T.R., 2000. Thefall, recovery, orbit and composition of the Tagish Lake meteorite: a new typeof carbonaceous chondrite. Science 290, 320–325.

Brown, P.G., Revelle, D.O., Tagliaferri, E., Hildebrand, A.R., 2002. An entry model forthe Tagish Lake fireball using seismic, satellite and infrasound records.Meteoritics & Planetary Science 37, 661–675.

BrukerAXS, 2005. DIFFRACplus Evaluation Package Release 2005. Bruker AXS,Karlsruhe, Germany.

Clayton, R.N., Mayeda, T.K., 2001 (abstract). Oxygen isotopic composition of theTagish Lake carbonaceous chondrite. 32nd Lunar and Planetary ScienceConference.

Coulson, I.M., Edwards, P.R., Lee, M.R., 2007. Recent developments in microbeamcathodoluminescence with applications to mineralogy. American Mineralogist92, 233–234.

Davis, A.M., Richter, F.M., 2003. Condensation and evaporation of solar systemmaterials. In: (Eds.), Davis, A. M., Holland, H., Turekian, K.K., Treatise onGeochemsitry.

Flemming, R.L., 2007. Micro X-ray diffraction (mXRD): a versatile technique forcharacterization of Earth and planetary materials. Canadian Journal of EarthSciences 44, 1333–1346.

Goreva, J.S., Lauretta, D.S., 2006 (abstract). Early oxidation of phosphorusassociated with sulfides in CM chondrites. 37th Lunar and Planetary ScienceConference.

Gounelle, M., Zolensky, M.E., 2001. A terrestrial origin for sulfate veins in CI1chondrites. Meteoritics & Planetary Science 36, 1321–1329.

Greshake, A., Krot, A.N., Flynn, G.J., Keil, K., 2005. Fine-grained dust rims in theTagish Lake carbonaceous chondrite: evidence for parent body alteration.Meteoritics & Planetary Science 40, 1413–1431.

Grossman, J.N., 2004 (abstract). Loss of chromium from olivine during themetamorphism of chondrites. 35th Lunar and Planetary Science Conference.

Hanowski, N.P., Brearley, A.J., 2001. Aqueous alteration of chondrules in the CMcarbonaceous chondrite, Allan Hills 81002: implications for parent bodyalteration. Geochimica et Cosmochimica Acta 65, 495–518.

He, B.B., 2003. Introduction to two-dimensional X-ray diffraction. PowderDiffraction 18, 71–85.

Helming, K., Lyubchenko, M., He, B.B., Preckwinkel, U., 2003. A new method fortexture measurements using a general area detector diffraction system.Powder Diffraction 18, 99–102.

Herd, R.K., Herd, C.D.K., 2007 (abstract). Towards systematic study of the TagishLake meteorite. 38th Lunar and Planetary Science Conference.

Hildebrand, A.R., Mccausland, P.J.A., Brown, P.G., Longstaffe, F.J., Russell, S.D.,Tagliaferri, E., Wacker, J.F., Mazur, M.J., 2006. The fall and recoveryof the Tagish Lake meteorite. Meteoritics & Planetary Science 41,407–431.

Hiroi, T., Hasegawa, S., 2003. Revisiting the search for the parent body of the TagishLake meteorite: case of a T/D asteroid 308 Polyxo. Antarctic MeteoriteResearch 16, 176–184.

Hiroi, T., Zolensky, M.E., Pieters, C.M., 2001. The Tagish Lake meteorite: a possiblesample from a D-type asteroid. Science 293, 2234–2236.

Izawa, M.R.M., 2008. Mineralogy and IR spectroscopy of the Tagish Lake C2chondrite and enstatite chondrites. M.Sc. thesis. The University of WesternOntario, London.

Izawa, M.R.M., Flemming, R.L., King, P.L., Peterson, R.C., Mccausland, P.J.A., 2010.Mineralogical and spectroscopic investigation of the Tagish Lake carbonaceouschondrite by X-ray diffraction and infrared reflectance spectroscopy. Meteori-tics & Planetary Science, in press, doi:10.1111/j.1945-5100.2010.01043.x.

Jarosewich, E., 2002. Smithsonian microbeam standards. Journal of Research of theNational Institute of Standards and Technology 107, 681–685.

Keller, L.P., Flynn, G., 2001 (abstract). Matrix Mineralogy of the Tagish Lakecarbonaceous chondrite: TEM and FTIR studies. 32nd Lunar and PlanetaryScience Conference.

Kminek, G., Botta, O., Glavin, D.P., Bada, J.L., 2002. Amino acids in the Tagish Lakemeteorite. Meteoritics & Planetary Science 37, 697–701.

Krot, A.N., Fagan, T.J., Nagashima, K., Petaev, M.I., Yurimoto, H., 2005. Origin of low-Ca pyroxene in amoeboid olivine aggregates: evidence from oxygen isotopiccompositions. Geochimica et Cosmochimica Acta 69, 1873–1881.

Krot, A.N., Petaev, M.I., Yurimoto, H., 2004. Amoeboid olivine aggregates with low-Ca pyroxenes: a genetic link between refractory inclusions and chondrules?Geochimica et Cosmochimica Acta 68 1923–1941.

Lauretta, D.S., Lodders, K., Fegley Jr., B., 1997. Experimental simulations of sulfideformation in the Solar Nebula. Science, 277.

http://10.1111/j.1945-5100.2010.01043.x


Lauretta, D.S., Lodders, K., Fegley Jr., B., 1998. Kamacite sulfurization in the SolarNebula. Meteoritics & Planetary Science 33, 821–833.

Lee, M.R., Parsons, I., Edwards, P.R., Martin, R.W., 2007. Identification ofcathodoluminescence activators in zoned alkali feldspars by hyperspectralimaging and electron-probe microanalysis. American Mineralogist 92,243–253.

Mikouchi, T., Kasama, T., Zolensky, M.E., Tachikawa, O., 2001 (abstract).Transmission electron microscopy of the matrix minerals in the Tagish Lakecarbonaceous chondrite. 32nd Lunar and Planetary Science Conference.

Mittlefehldt, D.W., 2002. Geochemistry of the ungrouped carbonaceous chondriteTagish Lake, the anomalous CM chondrite Bells, and comparison with CI andCM chondrites. Meteoritics & Planetary Science 37, 703–712.

Nakamura, T., Noguchi, T., Zolensky, M.E., Tanaka, M., 2003. Mineralogy and noble-gas signatures of the carbonate-rich lithology of the Tagish Lake carbonaceouschondrite: evidence for an accretionary breccia. Earth and Planetary ScienceLetters 207, 83–101.

Noguchi, T., Nakamura, T., Nozaki, W., 2002. Mineralogy of phyllosilicate-richmicrometeorites and comparison with Tagish Lake and Sayama meteorites.Earth and Planetary Science Letters 202, 229–246.

Pizzarello, S., Huang, Y., Becker, L., Poreda, R.J., Nieman, R.A., Cooper, G., Williams,M., 2001. The organic content of the Tagish Lake meteorite. Science 293,2236–2239.

Russell, S.D., 2006. The O and H isotope composition of the Tagish Lake meteorite.Master of Science thesis. The University of Western Ontario, London, Ontario,Canada.

Russell, S.D.J., Longstaffe, F.J., King, P.L., Larson, T.E., 2010. The oxygen-isotopecomposition of isolated olivine and chondrules from the Tagish Lakecarbonaceous chondrite. Geochimica et Cosmochimica Acta 74 (8), 2484–2499.

Scott, E.R.D., Keil, K., Stoffler, D., 1992. Shock metamorphism of carbonaceouschondrites. Geochimica et Cosmochimica Acta 56, 4281–4293.

Scott, E.R.D., Krot, A.N., 2005. Thermal processing of silicate dust in the solarnebula: clues from primitive chondrite matrices. The Astrophysical Journal623, 571–578.

Simon, S.B., Grossman, L., 2003. Petrography and mineral chemistry of theanhydrous component of the Tagish Lake carbonaceous chondrite. Meteoritics& Planetary Science 38, 813–825.

Stoffler, D., Keil, K., Scott, E.R.D., 1991. Shock metamorphism of ordinarychondrites. Geochimica et Cosmochimica Acta 55, 3845–3867.

Stoffler, D., Keil, K., Scott, E.R.D., 1992. Shock classification of ordinarychondrites—new data and interpretations. Meteoritics 27, 292–293.

Thaisen, K.G., Taylor, L.A., 2009. Meteorite fusion crust variability. Meteoritics &Planetary Science 44, 871–878.

Tissot, R.G., 2003. Microdiffraction applications utilizing a two-dimensionalproportional detector. Powder Diffraction 18, 86–90.

Zega, T. J., Stroud, R. M., Nittler, L. R., Alexander, C. M. O’D., Young, A. F., 2005. (abstract).Transmission electron microscopy of deuterium-rich matrix from Tagish Lake. 68thAnnual Meteoritical Society Meeting, # 5318.

Zolensky, M.E., Krot, A.N., Benedix, G.K., 2008a. Record of low-temperaturealteration in asteroids. Reviews in Mineralogy & Geochemistry 68, 429–462.

Zolensky, M.E., Nakamura, K., Gounelle, M., Mikouchi, T., Kasama, T., Tachikawa, O.,Tonui, E., 2002. Mineralogy of Tagish Lake: an ungrouped type 2 carbonaceouschondrite. Meteoritics & Planetary Science 37, 737–761.

Zolensky, M.E., Ohsumi, K., Mikouchi, T., Hagiya, K., Le, L., 2008b (abstract).Crystallinity of Fe–Ni sulfides in carbonaceous chondrites. 39th Lunar andPlanetary Science Conference.

Zolensky, M.E., Thomas, K.L., 1995. Iron and iron–nickel sulfides in chondriticinterplanetary dust particles. Geochimica et Cosmochimica Acta 59, 4707–4712.

Documents

Multi-technique investigation reveals new mineral, chemical, and textural heterogeneity in the Tagish Lake C2 chondrite