The origin of volatile element depletion in early solar system material: Clues from Zn isotopes in chondrules

(1)

HAL Id: insu-02917495

https://hal-insu.archives-ouvertes.fr/insu-02917495

Submitted on 20 Aug 2020

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

Distributed under a Creative Commons Attribution - NoDerivatives| 4.0 International

License

The origin of volatile element depletion in early solar

system material: Clues from Zn isotopes in chondrules

Emily Pringle, Frédéric Moynier, Pierre Beck, Randal Paniello, Dominik Hezel

To cite this version:

Emily Pringle, Frédéric Moynier, Pierre Beck, Randal Paniello, Dominik Hezel. The origin of volatile

element depletion in early solar system material: Clues from Zn isotopes in chondrules. Earth and

Planetary Science Letters, Elsevier, 2017, 468, pp.62-71. �10.1016/j.epsl.2017.04.002�. �insu-02917495�

(2)

Contents lists available atScienceDirect

Earth

and

Planetary

Science

Letters

www.elsevier.com/locate/epsl

The

origin

of

volatile

element

depletion

in

early

solar

system

material:

Clues

from

Zn

isotopes

in

chondrules

Emily A. Pringle

a

,

Frédéric Moynier

a

,

b

,

∗

,

Pierre Beck

b

,

c

,

Randal Paniello

d

,

Dominik C. Hezel

e

,

f

a_Institut_de_Physique_du_Globe_de_Paris,_Université_Paris_Diderot,_Sorbonne_Paris_Cité,_CNRS_UMR_7154,₁_rue_Jussieu,₇₅₂₃₈_Paris,_France b_Institut_{Universitaire}_de_France,_Paris,_France

c_Institut_{d’Astrophysique}_et_de_{Planétologie}_de_Grenoble,_Université_Grenoble_Alpes,_France d_Department_of_Earth_and_Planetary_Sciences,_Washington_University_in_St._Louis,_USA

e_University_of_Cologne,_Department_of_Geology_and_Mineralogy,_Zülpicher_Str._49b,_{50674 Köln,}_Germany f_Department_of_Mineralogy,_Natural_History_Museum,_Cromwell_Road,_London,_SW7_5BD,_UK

a

r

t

i

c

l

e

i

n

f

o

a

b

s

t

r

a

c

t

Articlehistory:

Received14November2016

Receivedinrevisedform29March2017 Accepted1April2017

Availableonline14April2017 Editor: D.Vance Keywords: carbonaceouschondrites chondrules zincisotopes volatiles protoplanetarydisk

Volatilelithophileelementsaredepletedinthedifferentplanetarymaterialstovariousdegrees,butthe origin ofthese depletionsis stilldebated. Stable isotopesofmoderately volatile elements suchas Zn can beused tounderstandthe origin ofvolatileelementdepletions. Sampleswith signiﬁcantvolatile elementdepletions, includingtheMoonand terrestrialtektites,display heavyZnisotope compositions (i.e. enrichmentof66Znvs.64Zn), consistentwithkineticZnisotopefractionationduring evaporation. However, Luck etal. (2005) found anegativecorrelation between δ66_Zn_and _1/[Zn] _between _CI, _CM,

CO, and CVchondrites, oppositetowhat would beexpectedif evaporationcausedthe Znabundance variationsamongchondritegroups.

WehaveanalyzedtheZnisotopecompositionofmultiplesamplesofthemajorcarbonaceouschondrite classes: CI (1), CM(4), CV (2), CO (4), CB (2), CH (2), CK(4), and CK/CR (1). The bulk chondrites defineanegativecorrelationinaplotofδ66Znvs1/[Zn],confirmingearlierresultsthatZnabundance variationsamongcarbonaceouschondritescannotbeexplainedbyevaporation.ExceptionsareCBandCH chondrites,whichdisplayZnsystematicsconsistentwithacollisionalformationmechanismthatcreated enrichmentinheavyZnisotopesrelativetothetrenddefinedbyCI–CK.

WefurtherreportZnisotopeanalysesofchondritecomponents,includingchondrulesfromAllende(CV3) andMokoia(CV3),aswellasanaliquotofAllendematrix.AllchondrulesareenrichedinlightZnisotopes (∼500 ppmon66Zn/64Zn)relativetothebulk,contrarytowhatwouldbeexpectedifZnweredepleted duringevaporation,ontheotherhandthematrixhasacomplementaryheavyisotopecomposition.We reportsequentialleachingexperimentsinun-equilibratedordinarychondrites,whichshow sulﬁdesare isotopicallyheavycomparedtosilicatesandthebulkmeteoritebyca.+0.65 permilon66Zn/64Zn.We suggest isotopically heavysulﬁdes were removedfromeither chondrulesor theirprecursors, thereby producingthelightZnisotopeenrichmentsinchondrules.

1. Introduction

Carbonaceous chondrites (CC) are chemically primitive early Solar System materials that provide important clues toward un-derstanding the origin and evolution of the terrestrial planets. Carbonaceous chondrites consist of assemblages of four compo-nents in different proportion: high-temperature condensates in

*

Correspondingauthorat:InstitutdePhysiqueduGlobedeParis,UniversitéParis Diderot,SorbonneParisCité,CNRSUMR7154,1rueJussieu,75238Paris,France.

E-mailaddress:moynier@ipgp.fr(F. Moynier).

the form of Calcium Aluminum-rich Inclusions (CAIs), roughly spherical small igneous objects known as chondrules, metallic FeNi and sulfides, and a fine-grained matrix (Krot et al., 2009; Scott and Krot, 2014). Carbonaceous chondrites are undifferenti-ated meteorites that are relatively volatile-rich and did not ex-perience high-temperature processingsufficientto resultin melt-ing, but they do show evidence of variable degrees of aqueous alteration andthermalmetamorphism (Scott and Krot, 2014). Al-thoughCC areingeneralconsideredaschemicallyprimitive, indi-vidual carbonaceous chondrite classesexhibit distinct patterns in volatile element (here “volatile” refers to those elements with a http://dx.doi.org/10.1016/j.epsl.2017.04.002

(3)

50% condensation temperature, Tc, between 250 K and 1250 K,

Lodders, 2003) abundances relative to solar composition (Palme et al., 2014a). The CI-type CC are believed to be the closest in compositiontothesolarnebuladuetoelementalabundancesthat closelymatchthecomposition ofthesolarphotosphere,withthe exceptionoflithium, thehighly volatileelements (Tc

<

250K:H,

C,N,O),andthe noblegases (Palme et al., 2014a). The otherCC classesshow volatiledepletionincreasing intheorderCI-CM-(CO, CV)-CK(Palme et al., 2014a).The originofthischemical trendis asyet unresolved. Hypotheses for the volatileelement depletion inbulk chondrites includeincomplete condensationfromthe so-larnebula (Wasson and Chou, 1974), volatilelossby evaporation during accretion(Ringwood, 1966), ormixing ofdistinct primor-dialreservoirs consistingofa volatile-richCI-likecomponentand a volatile-poor refractory component(Larimer and Anders, 1967; Clayton and Mayeda, 1999; Luck et al., 2003). Chondrulesareone proposedcarrierofthevolatiledepletion forthelattercase, such thatvolatileelementdepletionsinbulkmeteoritesareasignature of distinct chondrule compositions and the different CC classes represent varying mixtures of a compositionally uniform CI-like matrix and compositionally variable chondrules (e.g. Alexander, 2005).

Themechanism ofchondrule formationis still debated andit hasnumerous implications forthe origin of volatile elements in the terrestrial planets and for dynamical processes in the early SolarSystem. Centralquestionsconcernthe relationship between chondrules and matrix, namely, whether these two chondritic componentsformedtogether inthesameregionofthesolar neb-ula or formed in separate reservoirs and were later combined (Zanda et al., 2006; Hezel and Palme, 2008, 2010; Palme et al., 2015andreferencestherein).Theformationoftheﬁrstsolid mate-rialsisthereforedirectlylinkedtoconditionsinthesolar environ-mentduringearlyplanetary formationandcansupportor contra-dictmodelsofSolarSystemformation.Thediﬃcultyliesin recon-cilingthecontrastingformationconditionsofCAIs(high tempera-tures,variableisotopicreservoirs,shortformationtimescaleseveral Myrbeforeincorporationintochondrites),chondrules(moderately hightemperatures,episodicformation likelyat highertotal pres-sures), and matrix (lesser thermal processing, volatile-rich) with theincorporation ofall three componentsinto a coherent aggre-gateataparticulartimeandplaceintheprotoplanetarydisk(Krot et al., 2009).

Proposedchondruleformationmodelsthat attemptto account fortheseconstraintsincludetheflashheatingofchondrule precur-sors bya shockwave (Connolly and Love, 1998; Ciesla and Hood, 2002; Desch and Connolly, 2002; Connolly and Desch, 2004) or formationinplumesgeneratedby impacts(Asphaug et al., 2011). Theflashheatingmodelreproducesthechemicalrelationships be-tween chondritic components, including the thermal history of chondrules,chondrule-matrixcomplementarity,andepisodic chon-drule formation (Asphaug et al., 2011). On the other hand, the impactoriginwouldpredict ahighgaspressureanddust density that would explain the retention of volatile elements in chon-drules.Alternatively,thex-windmodel(Shu et al., 1996) isbased onastronomicalobservationsofextra-solarsystemsandtheoretical modelingofstellarmagneticfields.ItsuggeststhatCAIsand chon-drulesformed closetothe youngsunandwere latertransported radiallyoutwardinthediskwheretheymixedwiththermally un-processedmatrixmaterial.Thismodelaccountsforthecontrasting thermalhistory betweenchondrules/CAIs andmatrix butfails to explaingeneticrelationshipsbetweenchondrulesandmatrix.

The study ofzinc (Zn) isotopes in chondrites and their com-ponents has the potential to investigatethe numerous questions outlinedabove.Zincisamoderatelyvolatileelement(Tc

=

726K)

withacondensationtemperaturesigniﬁcantlybelowthatofmany othermoderatelyvolatileelements,includingK(Tc

=

1006K),Na

(Tc

=

958 K), and Rb (Tc

=

800 K) (Lodders, 2003). Further, Zn

exhibitsbothchalcophileandlithophile behaviorand,hence, pro-videsinformationaboutthehistoryofsilicateaswell assulphide phases. Zinc displays signiﬁcant heavy isotope enrichments cou-pledwithvolatiledepletion,asobservedinlunarbasaltsand un-brecciatedeucrites(Paniello et al., 2012a, 2012b;Kato et al., 2015; Moynier et al., 2017).Impacteventshavebeenshowntovolatilize Zn and drive the composition of the residue toward isotopically heavier compositions (Moynier et al., 2009a), making Zn a pow-erfulprobeofvolatility-dependentprocessesduringchondriteand chondrule formation.The pioneer study ofLuck et al. (2005) re-ported a negative correlation between

δ

66Zn and 1/[Zn] in bulk CC,such thattheyareprogressivelyenrichedinthelightisotopes ofZnintheorderCI-CM-(CO,CV).Thisobservationisopposite to theeffectexpectedduring Znisotope fractionationdueto evapo-rationandsuggeststhattheabundanceofvolatileelementsinCC isinheritedfrompre-accretionaryprocesses,reﬂectingmixing be-tween atleasttwo isotopicallydistinct reservoirs.The Znisotope compositionofseveralCCgroups(CK,CB,CH),aswellasindividual componentssuch aschondrules andmatrix,whichare themajor reservoirsofZn,havenotyetbeenanalyzed.However,these sam-ples may providecritical clues toward furtherunderstanding the originofthevolatileelementvariationsamongCC.

HerewepresentZnisotopedataforacomprehensivesetofCC, includingsamplesfromtheCI,CM,CO,CV,CK,CB,andCHclasses, to furtherassess the originandevolution ofvolatile elements in theinnerSolarSystem.WealsopresentZnisotopemeasurements ofindividualchondrules fromAllende(CV3)andMokoia(CV3)as well asan aliquot of Allendematrix to investigatethe formation ofsolid materialsinthe earlySolarSystemandthecompositions oftheprimitivematerialsthatrepresentthebuildingblocksofthe terrestrialplanets.

2. Samplesandmethods

2.1. Samples

The20bulkchondritesstudiedincludeoneCI(Y-980115),four CM (Murchison, Murray, Cold Bokkeveld, LON 94101), two CV (Allende, GRA 06101), four CO (Felix, Ornans, Lancé, Isna), two CB (Gujba,MIL 05082), two CH (PCA 91467, A-881020), fourCK (ALH 85002, Karoonda, EET 92002, Maralinga), and one CK/CR (A-881595). Y-980115 is CI-like but has mineralogical character-isticsthat suggestitexperiencedthermalmetamorphism(King et al., 2015). In addition its highZn content (337 ppm, see below) issimilartootherCIchondritesandsupportitsclassiﬁcationasa CI chondrites. A-881595is classiﬁedasa CR2, butmore likelyis a CK basedon Oisotopes (Schrader et al., 2011). The bulk com-positions oftheselected meteoritesrepresenta rangeinZn con-centrationof morethan two ordersof magnitude, from

∼

3 ppm to

∼

300 ppm. Since Zn is highly mobile in aqueous ﬂuids, me-teorite fallswere selected over ﬁnds when possible. In addition, three unequilibrated ordinary chondrites (UOC), including Clovis (H3.6), GRA95208 (H3.7), andALH90411 (L3.7), were subjected toasequentialdissolutionproceduretoisolateandanalyzetheZn isotopecomposition ofvariousZn carrierphases.Finally,nine in-dividual chondrules (including eight from Allende and one from Mokoia)aswellasonematrix-richaliquotfromAllendewere ex-tractedandtheirZnisotopecompositionswereanalyzed.

2.2. Methods

For bulk chondrites,

>

500 mgof sample was crushed into a homogeneous powder. Approximately 20–50 mg of each sample powder was dissolved usinga mixture of concentrated HF/HNO3

(4)

HF/HNO3,6N HClwas added tothe residueand againheatedto

dissolveremaining ﬂuoridecomplexes.Sampleswere then evapo-ratedto dryness.Wholeindividual chondrules were separatedby hand,crushed,anddissolved usingthesameprocedureasforthe bulk chondrites exceptthat the volume of acids was reduced to limittheproceduralblank.

ThethreeUOCwerefurthersubjectedtophysicalphase separa-tion(toseparatemagneticandnon-magneticphases)and sequen-tialaciddissolutiononthenon-magnetic phasetoisolatespecific carriersofZnfollowingtheproceduredescribedbyMoynier et al. (2011). Aftercrushing, the magnetic and non-magnetic fractions where separated with a hand magnet. The magnetic phase was dissolved in aqua-regia. The non-magnetic fraction was first dis-solvedincold3NHClfor6 h;thisdissolutioncontainedthesulfide fraction (see Luck et al., 2005). The remaining residue was then dissolvedinHF/HNO3tocompletelydissolvethesilicatefraction.

Chemicalpuriﬁcation ofZnwas achievedfollowingthe proce-dure described by Moynier and Le Borgne (2015). Samples were loaded in1.5N HBron AG-1 X8 (200–400mesh) anion-exchange resininPTFEcolumns.Matrixelementswereremovedbyan addi-tionalwashof1.5NHBrandZnwaselutedusing0.5N HNO3.The

collected sample solutionswere then evaporated todryness.This procedurewasperformedforatotalofthreecolumnpassesto en-sure clean separation of Zn: samples were passed once through columnscontaining0.5mLresinandthentwice throughcolumns containing 0.1 mL resin.Procedural blank is

<

7 ng and insigniﬁ-cantrelativetotheamountofZninthesamplemassanalyzedfor bulkchondrites(

>

1 μgZn).Forchondrules,wehaveslightly mod-iﬁedthemethodwereweonlypassedthesamplestwicethrough the0.1 mLcolumns.Thislimitstheamountofacidsusedand re-ducedthe blankto

<

1 ngwhichrepresentsup to3%ofthetotal signal.

Zinc isotope compositions of bulk carbonaceous chondrites were measured using a Thermo Scientiﬁc Neptune Plus Multi-CollectorInductively-Coupled-PlasmaMass-Spectrometer (MC-ICP-MS) at either Washington University in Saint Louis (WUSTL) or the Institut de Physique du Globe in Paris (IPGP) follow-ing protocol recently described elsewhere (Chen et al., 2013; Moynier and Le Borgne, 2015). Carbonaceous chondrite compo-nentsweremeasuredatIPGP,whileUOCleachatesweremeasured atWUSTL.Externalreproducibilitywas assessedthroughfull pro-cedural replicateanalyses ofthe same sample to be 40ppm for

66_Zn/64_Zn _and_{50 ppm} _for68_Zn/64_Zn ₍_{Chen et}_al.,₂₀₁₃_). _The _Zn

isotopecompositionoftheUSGSbasaltstandardBHVO-2was mea-suredthroughoutthestudy;the dataare inagreementwith cur-rentlyacceptedliteraturevalues(seeTable 1;Moynier et al., 2017). Additionally, one sample (Murray) was independently processed through the entire procedure (dissolution, chemical puriﬁcation, massspectrometry) atboth Washington University inSaint Louis andattheInstitutdePhysiqueduGlobeinParis;thesereplicates yield identical results within analytical error, signifying that the methodisreproducible.

TheZnconcentrations wereobtainedby comparisonofthe in-tensityofthestandardandofthesamplesontheMC-ICP-MS for whichweestimate theerrorto be

∼

10%.ThismethodgivesaZn concentrationof103 ppminBHVO,whichcompareswellwiththe USGScertiﬁcation(103

±

6ppm).

3. Results

ZincisotopecompositionsarereportedinTable 1,Table 2,and

Table 3andFig. 1,Fig. 2,Fig. 3,Fig. 4,andFig. 5 aspermil devi-ationsfromtheJMC-Lyon (Moynier et al., 2017) Znisotope stan-dard,

Fig. 1. Threeisotopeplotofδ68_Zn_versus_δ66_Zn_for_bulk_chondrites._Samples_plot

alongthecalculatedequilibrium(solid;slope1.94)andkinetic(dashed;slope1.97) mass-dependentfractionationlines(±2 se).

δ

xZn

=

(

xZn

/

64Zn

)

sample

(

x_Zn

_/

64_Zn

₎

JMC-Lyon

−

1

×

1000 (1)

where x

=

66 or68.Errors are givenasthe2 standarddeviation (2sd) of replicate measurements. Where only one measurement waspossibleweusetheerrorobtainedonmultipleanalysesoffull procedural replicates (see Section 2.2) forthe bulk chondrites or theerrorobtainedforthesinglechondrulewithmultipleanalyses (see Table 2). All the data fall on a mass-dependent fractiona-tion line ina plot of

δ

68Zn versus

δ

66Zn(Fig. 1) consistent with

Moynier et al. (2009b), with the exception of three chondrules slightlyoff theline, whichis likelydueto anerror underestima-tionduetothelowZncontentofthesesamplesandthefactthat replicateanalysescouldnotbeperformed.

Bulk CC (Table 1 and Fig. 2) span a range of

δ

66Zn from

+

0.43

h

(Y-980115,CI)to

−

0.69

h

(A-881595,CK/CR).Thevalues for the single CI (

δ

66Zn

=

0

.

43

h ±

0

.

01) and the class averages for the CM (

δ

66Zn

=

0

.

38

h ±

0

.

04), CV (

δ

66Zn

=

0

.

24

h ±

0

.

12), andCO(

δ

66Zn

=

0

.

17

h ±

0

.

11)aresimilartopreviouslyreported data (Luck et al., 2005; Barrat et al., 2012, see Fig. 2). Notably, each CCclasshasacharacteristicZnisotopecomposition.As pre-viously observed,

δ

66_Zn _values _correlate _with _Zn _{concentration,}

[Zn] (Fig. 3); samples with the lowest [Zn] are the most en-riched in light Zn isotopes (Luck et al., 2005). The CK (on aver-age,

δ

66_Zn

₌

₀

_.

₁₉

_{h ±}

₀

_.

₃₁₎_are _more_variable _in_both_Zn _isotope

(5)

Table 1

Zincisotopiccompositionsandconcentrationdataforbulkcarbonaceouschondrites.

Sample Class Fall/ﬁnd δ66_Zn _2sda _δ68_Zn _2sda _nb _[Zn]

(ppm) BHVO-2 0.30 0.03 0.59 0.06 11 103 Y-980115c _CI1 _ﬁnd _0.43 _0.01 _0.79 _0.03 ₅ ₃₃₇ Murchisonc _CM2 _fall _0.40 _0.06 _0.76 _0.09 ₆ ₁₈₆ Murray (1)c _CM2 _fall _0.38 _0.02 _0.68 _0.05 ₅ ₁₈₇ Murray (2)d _CM2 _fall _0.36 _– _0.65 _– ₁ Average-Murray 0.37 0.02 0.67 0.04 2 187

Cold Bokkeveldc _CM2 _fall _0.46 _0.03 _0.89 _0.04 ₄ ₁₄₅

LON 94101c _CM2 _ﬁnd _0.37 _0.04 _0.66 _0.02 ₅ ₁₈₄ Average-CM 0.38 0.04 0.70 0.12 3 172 Allende (1)c _CV3 _fall _0.27 _0.04 _0.48 _0.08 ₂ ₁₁₂ Allende (2)c _CV3 _fall _0.30 _0.02 _0.54 _0.02 ₅ ₁₀₈ Average-Allende 0.29 0.04 0.51 0.08 2 110 GRA 06101d _CV3 _ﬁnd _0.20 _– _0.37 _– ₁ ₁₂₀ Average-CV 0.24 0.12 0.44 0.21 2 110

Felixd _CO3.3 _fall _0.14 _– _0.22 _– ₁ ₉₉

Ornansd _CO3.4 _fall _0.13 _– _0.19 _– ₁ ₁₀₅

Lancéd _CO3.5 _fall _0.16 _– _0.26 _– ₁ ₁₀₃

Isnad _CO3.8 _find _0.25 _– _0.41 _– ₁ ₁₀₀ Average-CO 0.17 0.11 0.27 0.19 4 102 Gujbad _CB _fall _0.20 _– _0.39 _– ₁ ₃ MIL 05082d _CB _find _0.05 _– _0.10 _– ₁ ₆ Average-CB 0.13 0.22 0.24 0.41 2 5 PCA 91467c _CH3 _find _0.32 _0.05 _0.58 _0.10 ₂ ₃₉ A-881020c _CH3 _find _0.29 _0.04 _0.53 _0.07 ₂ ₃₁ Average-CH 0.30 0.05 0.56 0.08 2 35 ALH 85002c _CK4 _find _0.22 _0.01 _0.43 _0.02 ₄ ₇₈ Karoondac _CK4 _fall _0.30 _0.05 _0.57 _0.05 ₄ ₈₃ EET 92002c _CK5 _find _0.28 _0.05 _0.54 _0.06 ₄ ₉₀ Maralingac _CK4 _find ₋_0.04 _0.03 ₋_0.01 _0.02 ₃ ₅₇ Average-CK 0.19 0.31 0.38 0.54 4 77 A-881595c _CK/CR2 _find ₋_0.69 _0.03 ₋_1.20 _0.11 ₂ ₂₂ a _2sd₌₂_×_{standard deviation.} b _n₌_{number of}_{measurements.} c _Measured_at_IPGP. d _Measured_at_WUSTL. Table 2

ZincisotopiccompositionsandconcentrationdataforindividualchondrulesfromAllende andMokoiaandan Allendematrix-richaliquot.

Sample δ66Zn 2sda δ68Zn 2sda _nb _Mass (mg) [Zn] ppm Allende CH4 −0.21 0.10 −0.39 0.15 2 1.7 60 CH5 −0.03 – 0.05 – 1 3.3 32 CH6 0.18 – 0.43 – 1 1.9 31 CH8 −0.45 – −0.63 – 1 0.7 42 CH10 0.06 – 0.27 – 1 0.7 71 A64 −0.08 – −0.07 – 1 1.6 94 A70 −0.24 – −0.32 – 1 2 52 A78 −0.16 – −0.08 – 1 2.2 47 Matrix-rich 0.35 150 Mokoia M53 −0.36 – −0.41 – 1 9 59 a _2sd₌₂_×_{standard deviation.} b _n₌_{number of}_{measurements.}

composition and[Zn], butrangefrom valuessimilar to thosefor CV-CO towards lighter Zn isotope compositions, further extend-ing thetrend relating

δ

66ZnandZn depletion deﬁnedinLuck et al. (2005). The chondrite with the lowest bulk

δ

66Zn, A-881595 (CK/CR;

δ

66Zn

= −

0

.

69

±

0

.

03

h

)is

>

10x moredepletedinZnand offsetby

>

1

h

onthe66_Zn/64_Zn_ratio_compared_to_CI._Chondrites

fromtheCBandCHclasses(onaverage,

δ

66Zn

=

0

.

13

±

0

.

22

h

and

δ

66Zn

=

0

.

30

±

0

.

05

h

, respectively) have low Zn concentrations (3–39 ppm) but fall off the trend deﬁned by the CI-CK classes, withhigher

δ

66Zn values fora given[Zn]. In addition,there is a cleartrend between

δ

66Zn and

ε

54_Cr _for_CI-CK _chondrites, _while

(6)

Table 3

Zincisotopecompositionsofequilibrated ordinarychondriteleachates.

Sample Component δ66Zn δ68Zn Mass fraction [Zn] ppm 66Znsulfide-silicatea Clovis (H3.6) Magnetic −1.64 −2.92 0.12 28 Silicate −2.57 −4.98 0.43 68 Sulfide −1.93 −3.73 0.44 62 Whole rock −2.42 −4.65 0.64 GRA 95208 (H3.7) Magnetic 0.14 41 Silicate −0.42 −0.79 0.46 58 Sulfide 0.37 0.75 0.40 44 Whole rock −0.04 0.07 0.79 ALH 90411 (L3.7) Magnetic −1.07 −1.94 0.08 20 Silicate −0.91 −1.74 0.47 72 Sulfide −0.41 −0.75 0.45 33 Whole rock −0.75 −1.41 0.50 a 66_Zn

sulﬁde-silicate= δ66Zn(sulﬁde)− δ66Zn(silicate).

Fig. 2. ZincisotopedataforbulkCCfromthisstudyaswellasliteraturedataforCI, CM,CV,andCO(Lucketal.,2005; Barratetal.,2012).Theδ66Znvaluesfromthis studyareingoodagreementwithliteraturedata.Errorbarsare2sd.

The

δ

66Zn valuesforindividual chondrules are highly variable andrangefrom

+

0.18

h

to

−

0.45

h

(Table 2and Fig. 4), witha mean

δ

66Zn of

−

0

.

12

±

0

.

39

h

(2sd).In addition the Zncontent ofthechondrulesisalsovariableaspreviouslyobservedbyRubin and Wasson (1987) and Palme et al. (2014b). All Allende chon-drulesareenrichedinlightZnisotopesrelativetothecomposition ofthebulk(Allendebulk

δ

66Zn

=

0

.

29

±

0

.

04

h

).Forthe individ-ualchondrulesthefullproceduralblankcanrepresentamaximum of3% ofthetotalbeamintensityduring MC-ICP-MSanalysis. Ifa terrestrialisotopiccompositionof

δ

66Zn

=

0

.

28 (Chen et al., 2013; notethat itisnotpossible toanalyzetheisotopiccompositionof the blank due to the extremely low total content of Zn) is as-sumed fortheblank, then thiscould shift themeasured isotopic composition of a chondrule towards heavier

δ

66Zn values by a maximumof only 0.03

h

.The matrix-rich aliquot of Allende has a slightly heavier Zn isotope composition (

δ

66Zn

=

0

.

35

h

) rela-tivetothebulk.TheZnconcentrationisaboutthreetimeshigher in Allende matrix (

∼

150 ppm) compared to Allende chondrules (

∼

50 ppm).TheZnisotopecompositionofthesingleMokoia chon-drule (

δ

66Zn

= −

0

.

36

h

) falls within the range of the Allende chondrules.

Fig. 3.δ66_Zn_versus_Zn_depletion_(represented_by_1/[Zn]_in_ppm)_for_bulk_chondrites.

MostCCclassesdefineatrendoflightZnisotopeenrichmentwithincreasing de-greeofZndepletion,asindicatedbythegreybest-fitline.Theexceptionsarethe CHandCBchondrites,whichplotaboveandtothe rightofthistrend;theirZn isotopecompositionsmayhavebeenmodifiedbyevaporationduringparent-body formation.

The

δ

66Zn values for the bulk UOC span a wide range, from

−

2

.

42

h

to

−

0

.

04

h

(Table 3andFig. 5). AlthoughthebulkUOC have varying Zn isotopecompositions, theleachates display con-sistently systematicbehavior. Notably, thesulfide andthesilicate phasesmakeasimilarcontributiontotheZnbudgetofthewhole rock: each contain between 40–50% of the bulk Zn. The sulfide phase is consistently enriched in the heavy isotopes of Zn com-paredtothebulk,whilethesilicatephaseisenrichedinthelighter isotopesofZn.ThisconfirmstheresultsfoundbyLuck et al. (2005)

fortheUOC Krymka(LL3.1). Thedifference between

δ

66Zninthe sulﬁdes and in the silicates (denoted

66Znsulﬁde-silicate) ranges

from0

.

5

h

to0

.

79

h

. 4. Discussion

4.1. Zincisotopevariationsinbulkcarbonaceouschondrites

Evaporative losswould create elementalabundance variations as a function of elemental volatility and would be expected to produce isotopicfractionation ifevaporationoccurredin anopen system. However, aspreviously noticed byLuck et al. (2005)and conﬁrmedby ournew datathereis astrong negativecorrelation between

δ

66Zn and1/[Zn].Carbonaceous chondrites are progres-sivelyenriched inthelightisotopesofZnintheorderCI-CM-(CV,

(7)

Fig. 4.δ66_Zn_versus _Zn _depletion_(represented _by_1/[Zn]_in_ppm)_for _CV

chon-dritecomponents,includingindividualchondrulesfromAllende andMokoiaand amatrix-richaliquotofAllende.AllanalyzedAllendechondruleshaveδ66_Zn_values

lowerthanthatofbulkAllende,whileAllendematrixhasaslightlyhigherδ66_Zn

value.ChondrulesshowatrendoflightZnisotopeenrichmentcoupledwith in-creasingZnloss,oppositetotheeffectexpectedduringevaporationofZn.

Fig. 5. ZincisotopedataforbulkUOCandUOCleachates.Zinc-bearingphasesofthe UOCClovis(H3.6),GRA95208(H3.7),andALH90411(L3.7)wereseparatedusing asequentialaciddissolution.Plottedaretheδ66_Zn_values_of_the_bulk_rock,_silicate,

sulﬁde,andmetalphases.Inallcasesthesulﬁdesareenrichedintheenrichedin theheavyisotopesofZncomparedtothesilicates(by∼0.65hforδ66Zn,alsosee

Table 3).

CO)-CK;this is infact opposite to the effectexpected dueto Zn isotope fractionation during evaporation. Therefore, as suggested byLuck et al. (2005),Znisotopefractionationduringfree evapora-tiononCCparentbodiesortheirprecursors doesnot explainthe ZnisotopevariabilityinCC.

Aqueousalterationonthechondriteparent bodiescould mod-iﬁed the Zn isotope composition of CC, as observed in labora-toryexperiments(seeMoynier et al., 2017andreferencestherein) AqueousalterationwasprevalentforCCandlikelyoccurredunder avarietyofconditions,includingvariablepressures,temperatures, water/rockratios,andoxygenfugacities(Brearley, 2003). Thehigh solubilityofZncouldresultinopen-systemlossofZnifthe water-rockinteractionon thesurface ofthe chondriteparent body was coupledwith subsequent ﬂuid loss.This is unlikely,as Zn abun-dancesare infact highestinCC that are therichest in –OH/H2O

(i.e.CI, CM).Since aqueous alteration was a parent-body speciﬁc processthatoccurredunderarangeofconditions,genetically un-relatedsamplesshouldnotshowsystematicrelationshipsof

δ

66Zn. ThelinearcorrelationofZnconcentrationandisotopiccomposition

Fig. 6.δ66_Zn_versus_ε54_Cr_(in_parts_per_ten_thousand)_for_bulk_carbonaceous

chon-drites. Shown aregroup averages for carbonaceous chondriteswith bulk anal-yses for CI, CM, CV, CO,and CK. Since CB δ66_Zn _and _ε54_Cr _bulk _rock

analy-ses are limited and no CB has been analyzed for both δ66_Zn _and _ε54_Cr, _the

δ66_Zn _data_from_this _work_are_plotted_against _ε54_Cr_data_for _the _CB_Bencubbin

(Trinquieret al., 2007). Thereis a clear relationshipbetween δ66Zn and ε54_Cr

for CI-CKchondrites.TheCBfallbelowthe trenddeﬁned byCI-CK(ε54_Cr_data

fromTrinquieret al.,2007;Qin et al.,2010.δ66_Zn _data_from_Luck_et _al.,_2005;

Barratetal.,2012;thiswork).

observed throughtherange ofmostCCclasses wouldnotbe ex-pectedifaqueousalterationwasresponsibleforgeneratingtheZn isotopevariationsinbulkCC.

Finally, the variable degrees of volatile element depletion among the differentCC classes mayreﬂect the mixing of chem-ically and isotopically distinct reservoirs during CC accretion (Clayton and Mayeda, 1999; Luck et al., 2003, 2005). Evidence for this may be observed in the relationships between

δ

66Zn and measures of stable isotope anomalies, such as

ε

54_Cr _(see

Fig. 6). Varying mixtures of volatile-rich material relatively en-riched in heavy Zn isotopes andvolatile-poor material relatively enriched in light Zn isotopes could explain the volatile contents and Zn isotope compositions of the chondrite classes on the CI-CK trend. Chondrules have been proposed as a carrier of the volatiledepletion inCCandapossibleexplanation forZn isotope variations observed previously in CC (Luck et al., 2005). In this case, a CI-like matrix material combined with varying amounts of chondrules (with a Zn-depleted, isotopically light Zn signa-ture)couldbe responsiblefortheCI-CKvariations.However,such a model is not compatible with what is currently known from chondrule-matrix complementarity which argue that the matrix of each chondrites group has a different composition which is complementary tothechondrule composition(Hezel et al., 2010; Palme et al., 2015).

4.2. Theuniqueimpact-plumeformationofCHandCBchondrites

TwoCC classesin thecurrentdatasetdo not lie onthe trend deﬁned by CI-CK; speciﬁcally, the CH and CB chondrites have higher

δ

66Zn values than expected based on their low Zn con-centrations (on average, 35 ppm and 5 ppm, respectively). This is an indicator that CB and CH chondrites experienced a differ-ent formation history compared to the other CC. CB chondrites havehighmetal content,arehighlydepleted involatileelements, andthemetalcompositionssuggestdirectcondensationfromagas (Weisberg et al., 2001).TheCB chondriteshavechemicaland iso-topiccharacteristics that are similar to CH, andboth seemto be relatedtoCR(Weisberg et al., 2001). ChondrulesinCB arehighly depleted inmoderately volatileelements. It isbelievedthat they

(8)

condensedfromagas–meltplumeproducedduringahigh-energy impact eventin the early Solar System(Krot et al., 2005). Addi-tionally,Pb–PbdatingofindividualchondrulesfromtheCBGujba indicatesthat thechondrules haveuniformages andformed dur-ingasingleevent–incontrasttotherangeofindividualchondrule ages(spanning

∼

2Myr)determinedforchondritesfromotherCC classes(Bollard et al., 2015).

The unique impact-plume formation process proposed for CH andCB chondritesmayexplain theirenrichmentinheavy Zn iso-topesrelativetotheCI-CKtrend(i.e.theyhavehigher

δ

66Znvalues foragiven[Zn]andthusfallabovethetrendlinedeﬁnedbyCI-CK inFig. 3).Inthismodel,theZnreservoiroftheCBandCH precur-sorsmayhaveinitiallybeenontheCI-CKtrend,butsubsequently Znisotopecompositionsweremodiﬁedbyisotopefractionation as-sociated withZn lossbyevaporation. Loss ofisotopicallylightZn to thevapor phase, which was not fullyincorporated during the condensationofCBandCHparentbody-forming materials,ledto heavyZnisotopeenrichmentcoupledwithZndepletionontheCB andCHparent bodies.Therefore,the newZnisotope dataforCB andCHfurther supporttheoriginofthe CHandCB parentbody asaconsequenceofcollisionalaccretion.

4.3. MechanismscapableofgeneratingZnisotopefractionationin chondrules

Based on the igneous texture of chondrules, their formation occurred during a high-temperature process under conditions of rapidheating andcooling to retain volatileelements andto pre-serverelictgrainsfrompreviouschondrulegenerations(e.g.Jones, 2012).High-temperatureheatingatlowpressureinthesolar neb-ula should be recorded in the evaporative loss of volatile ele-ments from chondrules. However, the absence of volatileloss of Naorlargeisotopefractionationsindicateschondruleformationat asolid/gasratioseveralordersofmagnitudehigherthansolar(e.g.

Alexander et al., 2000; 2008;Hezel et al., 2010).

It has been shown that CC matrix has a different

com-position across the different classes, with varying degrees of volatileelementdepletionrelativetoCI(Rubin and Wasson, 1987; Bland et al., 2005). Furthermore, the degree of depletion is not strictly related to elemental volatility (i.e. elements with sim-ilar volatilities may show very different depletions relative to CI; Bland et al., 2005). Particularly, moderatelyvolatile lithophile elements show relatively small depletions, whereas moderately volatile chalcophile and siderophile elements are generally de-pletedtoagreaterdegreerelativetoCI(Bland et al., 2005).

Inallcases,AllendechondrulesaredepletedinZnandtheirZn isotopecompositionsarelightrelativetobulkAllende(seeFig. 4). TheMokoia chondrulehasaZn isotopecomposition and concen-tration within therange oftheAllende chondrules, showingthat thelight Zn isotopesignature in chondrules isnot limitedto Al-lende butmay be widespread. There are two general cases that couldresultinlightZn isotopeenrichmentsinchondrules: either thisisaneffectofparentbodyalterationoritisasignatureofthe chondruleformationprocess,namelycondensation,evaporation,or someotherprocess.

Chemical exchange betweenchondrule and matrix during al-teration on the parent body could lead to Zn isotope fractiona-tionifZnwas suﬃciently(anddirectionally)redistributed among chondritecomponents.Since thisstudyanalyzedbulk chondrules only, Zn isotope redistribution between phases within an indi-vidualchondrule wouldnotcauseanyobservable effect.Aqueous alterationinCVchondrites isvariable,withMokoiaexhibiting in-dicatorsofahigherdegreeofaqueousalteration comparedto Al-lende(Brearley, 2003).Whilewehavedataforonlyonechondrule fromMokoia,ithasasimilarZnabundanceandisotopic composi-tionsthanmostAllendechondrules;thisisunlikelytobethecase

if alteration was thecause ofthe light Zn isotopeenrichment in mostchondrules considering thediffering aqueousalteration his-toryofAllendeandMokoia.Itshouldbenotedthatincontrastto the oxidizedCV3analyzedhere, reduced-typeCV3show minimal evidence for aqueous alteration (Brearley, 2003); future study of the Znisotope composition ofchondrules fromreducedCV3 will be importanttoconﬁrmtheseresults.Finally,thelow permeabil-ityobservedinAllendewouldnot allowforﬂuidmovement over largescales;thiswouldrestrictaqueousalterationtolocaldomains only (Hezel et al., 2013). Forthesereasons,itappears that aque-ous alteration isunlikely to be the cause ofthe light Zn isotope enrichmentinchondrules; insteadthisisotopicallylightsignature isderivedfromthechondruleformingprocess. However,itisstill possiblethataqueousalterationmayaffectthemostZn-poor chon-drules and be theoriginof the deviationfromthe generaltrend ofthetwo chondrulesthatfall outofit(Fig. 4).Thesetwo chon-druleshavethelowestconcentrationsofZn(

∼

30ppm,seeTable 2,

Fig. 4); Therefore they would be the most susceptible to aque-ousalterationthatwouldpullthecompositionuptowardthebulk value.

Condensation can resultin isotopicallylight solids under spe-ciﬁcconditions(Richter, 2004).Thisraisesthequestionofwhether the Zn isotope composition of chondrules merely records the Zn isotope composition inherited during the condensationof the chondrule precursors. One possibility is that chondrule-forming materialcondensedfromagasenrichedinlightZnisotopes com-pared to CI, which resulted from the partial evaporation of Zn-bearing solids. This early partial evaporation would require that thelightZnisotope-enriched gaswas separatedfromthe remain-ingsolidsandthatsuchaseparation(e.g.byanineﬃcientdynamic couplingofthegasandsolids)happenedbeforeisotope equilibra-tion could occur. Evidence fora nebular gas enriched inlight Zn isotopes comparedtoCImayexistinCAIrims.AlthoughCAIsare not a major carrier ofZn, Allende CAIsexhibit Zn elemental en-richmentsatthesurface (Chou et al., 1976).Limiteddataindicate that Allende CAIs mayrecord very light Znisotope compositions (

δ

66_Zn

_{= −}

₂

_.

₆₅

_h

₎ _relative _to _bulk _Allende, _suggesting _that _CAIs

interacted with a gas enriched in light Zn isotopes (Luck et al., 2005);a CAI fromMurchison showssimilar light isotope enrich-ment (

δ

66_Zn

_{= −}

₁

_.

₂₈

_h

_; _Moynier_et_al.,₂₀₀₇_). _However, _an

iso-topically lightnebular gasalsoimpliesa complementary volatile-depleted reservoir enriched in the heavy Znisotopes, but such a reservoir has not been identiﬁed. Furthermore, the timing, dura-tion, andprevalenceconstraintssurrounding chondruleformation (e.g.Jones, 2012)suggeststhatZnisotopefractionationasaresult of chondrulecondensation isunlikely to be a signiﬁcant process, and that another mechanism is required to explain the light Zn isotopeenrichmentsinchondrules.

If chondrules experienced open system evaporation (under Rayleighconditions)thentheresidueshouldbeenrichedinheavy isotopes for volatile elements. The absence ofheavy isotope en-richment formoderately volatile elements (e.g. K,Humayun and Clayton, 1995;Zn,thisstudy)inchondrulesleadstooneof follow-ingconclusionsregardingchondruleevaporation:

(1)The isotopesignatureofchondruleevaporationwas erased by subsequentparent body processes, e.g. aqueous alteration. As previously discussed, this is unlikely for the chondrules in the presentstudy.

(2) Open-systemchondrule evaporationoccurred, but the iso-topesignaturewaserasedduringsubsequentgas–meltinteraction (Friend et al., 2016) beforethechondruleswereaccretedtoa par-ent body.Back-reaction of chondrules andnebular gas has been previously suggestedtoexplainthelackofKisotopefractionation (Alexander et al., 2000) and the constant Na contents in chon-drules(Alexander et al., 2008).Chondruleformationcouldhave in-volvedevaporativevolatilelosswithoutisotopefractionationunder

(9)

speciﬁc conditions, if such a reaction between evaporated gases andtheresiduesuppressedisotopefractionation.Thisisconsistent withtheabsence ofheavy Zn isotope enrichmentsin chondrules butdoesnotexplaintheirlightZnisotopeenrichments.

(3) The loss of volatiles from chondrules by evaporationwas insigniﬁcantandchondruleisotopiccompositionsrecord an alter-nate process. The loss of volatileelements fromchondrules dur-ingevaporationcouldbe minimizedby rapidheatingandcooling ofand/or by highdust/gas ratios inthe chondrule-forming envi-ronment (as proposed to explain Na abundances in chondrules;

Alexander et al., 2008).There is some evidencethat the elemen-tal abundance of Zn inAllende chondrules is not strictly due to its lithophile or its chalcophile behavior; chondrule Zn could be hostedinsilicatesorinopaquephases.InAllendebulkchondrules, Znabundancesdonotcorrelatewithotherelements,incontrastto intercorrelationsamongotherelementssuchasthealkalielements (Grossman and Wasson, 1985; Rubin and Wasson, 1987). The ab-senceofsuchcorrelationsforZnsuggeststhatZnisindeedhosted inbothsilicateandsulﬁdeminerals(Grossman and Wasson, 1985). In addition, Zn is over-depleted relative to both the moderately volatilelithophilesandthemoderatelyvolatilechalcophiles(Rubin and Wasson, 1987).ThissuggeststhatZnabundanceinchondrules maybe dueto two separate contributions: a partial volatileloss viaevaporationfromchondrulesilicatesinwhichisotope fraction-ation was suppressed by gas–melt reaction andan additional Zn depletionrelatedtothechalcophilebehaviorofZn.

Forthesereasons,theZnisotopedataindicatethatchondrules mayhaveexperiencedevaporationunderopen systemconditions aslongastheisotopesignatureofevaporationwaserasedby back-reaction with the evaporated gas, or that evaporation may have occurredundernon-Rayleighconditions.However,neitherprocess explainsthelight Znisotope enrichmentin chondrules,so an al-ternativeexplanationisrequired.

4.4.TheoriginoftheisotopicallylightZninchondrulesviasulﬁde removal

Looking at sulﬁdes fromUOC might help to understand how opaquephasesproducedisotopicallylight CC chondrules.Sulﬁdes fromUOC arestronglyenrichedintheheavyisotopes ofZn com-pared to both silicates and the bulk chondrite (by

∼

0.65

h

for

δ

66Zn, see Table 3 and Fig. 5; Luck et al., 2005 and thisstudy). Thesegregationofan isotopicallyheavy sulﬁdephasefromeither chondrules ortheir precursors would thereforehave two effects: it would deplete the chondrules in Zn and remove a reservoir of isotopically heavy Zn, which would leave the chondrules en-riched in light Zn isotopes relative to the bulk chondrite. Such an effect would createthe general negative correlation observed between

δ

66Zn and 1/[Zn] (Fig. 4). A silicate-metal/sulﬁde phase separationhas similarly been used to explain complementary W and Mo nucleosynthetic anomalies between Allende chondrules andmatrix(Becker et al., 2015; Budde et al., 2016a, 2016b), and opaquephasesaremostprobablyalsoresponsibleforthevariable Fe isotope compositions ofAllende, Mokoia, and other CV chon-dritechondrules(Hezel et al., 2010). Thetwochondrulesthat fall outofthiscorrelationarethechondruleswiththelowest concen-trationsofZn(

∼

30 ppm,see Table 2,Fig. 4);they wouldbe the mostsusceptible to aqueous alteration that could partially over-printtheirinherentZnisotopecompositionandconcentration(see discussionoftheaqueousalterationabove).

The segregation of sulfides from chondrules or their precur-sors could occur through several different mechanisms, includ-ing immiscible liquid separation coupled with chondrule fission orquantitative evaporation of sulfides from chondrules (or their precursors). The ejection of metal and sulfide from the interior ofchondrules hasbeen modeled by Uesugi et al. (2008) and by

Wasson and Rubin (2010) to explain thesiderophile element de-pletions in chondrules. These models involve the separation of metal/sulfideand silicateasimmiscible liquids during the (brief) time that the chondrules were molten (Grossman and Wasson, 1985).Centrifugalforceswithinarotatingchondrulewouldfurther contributetodrivethedensermetal/sulfidephasetowardsthe sur-faceofthemeltedchondrule.Finally,themetal/sulfidephasemay beejectedfromthechondruleduetothefactthatsilicatemelthas alower surfaceenergythanFeandFeSmelt (Uesugi et al., 2008; Wasson and Rubin, 2010). In terms of Zn, the ejection of sul-fideviachondrulefissioncouldexplain theZndepletionandlight Zn isotopeenrichment observed inchondrules if thereis an iso-topicfractionation betweenZn dissolved in thesilicate melt and theZn-bearingsulfides,asobservedinUOC.However,thereare a numberof issueswiththe model.Onlya smallfractionof chon-drules show textural evidence for loss of immiscible liquids by fission(Grossman and Wasson, 1985).Palme et al. (2014b) demon-stratedthat chondriticNi/CoratiosofbulkAllendechondrules ar-gueagainstsignificantlossofopaques.Additionally,thefateofthe ejected metal/sulfideisunclear; a limit on theubiquity of chon-drulefissionmaybeplacedbythelackofcorrespondingmetalor sulfidespheroidaldropletsinthematrix.

Separationofsulfideandsilicatemayalsohaveoccurredduring quantitativeevaporationofsulfidefromchondrulesortheir precur-sors.Oneadvantageofthismechanismisthatsulfideevaporation couldoccurbelowthesilicateliquidus,soitwouldnotnecessarily needtooccurduring theshorttimethatchondrulesweremolten. Evaporationofsulfideswouldneedtobesignificantoroccurina highdust/gasenvironment(i.e.underahighpartialpressureofZn) toavoidkineticZnisotopefractionationduringevaporation.If sul-fidelossoccurredduringchondruleformation,thenevaporationof Znmayhaveprovided sufficientvaporpressure ofZntolimit ki-neticisotopefractionationduringanysubsequentevaporativeloss of Znfrom chondrules (e.g.from chondrulesilicates during flash heating), if it did occur. Finally, if the gas resulting from sulfide evaporationwas heterogeneously lostfromthe CC-forming reser-voirprior tothefinal assemblageofCC components,thisscenario couldalsoexplaintheZnisotopesystematicsobservedinbulkCC. Loss ofthegasphase produced bysulfideevaporationwould de-plete the CC reservoir in Zn and heavy Zn isotopes, driving the entire reservoir to lighter Zn isotope composition and lower Zn concentrations.This wouldproduce theCI-CK trendwhile allow-ingallcomponentstooriginatefromasinglereservoir.

5. Conclusions

There is an inverse correlation between

δ

66Zn and 1/[Zn] be-tweenbulkCCfortheCI,CM,CO,CO,andCKclasses,whichcannot beexplainedby volatilelossduring evaporationontheCCparent body. CB and CH fall off the trend deﬁned by CI-CK, consistent withevaporativelossofZninlinewiththeimpact-plume forma-tionhypothesis.

Individual chondrules from the CV chondrites Allende and MokoiaexhibitZndepletionsbutenrichmentsinlightZnisotopes. ThemostlikelyexplanationforthelightZnisotopeenrichmentin chondrulesisthesegregationofanisotopicallylight sulfidephase during chondrule formation.This most probablyoccurred during quantitative evaporationofsulfidesfromchondrules ortheir pre-cursors, or sulfides were preferentially incorporated into matrix ratherthan intochondruleprecursors. Loss ofthegas phasethat was produced during sulfide evaporation from chondrules from the CC-forming reservoir would create depletions in Zn and en-richmentsinlightZnisotopes.Ifthisgaswasheterogeneouslylost fromtheCC-formingreservoirprior tothefinalassemblage ofCC components, thisscenario could also explain the Zn isotope sys-tematicsobservedinbulkCC.

(10)

Acknowledgements

We thank the Editor, Derek Vance, as well as 3 anonymous reviewers fortheir constructivecomments that greatly improved the quality of the manuscript. Julien Moureau, Pascale Louvat, and Jessica Dallas for maintaining the MC-ICP-MS at IPGP. FM acknowledges funding from the European Research Council un-dertheH2020frameworkprogram/ERCgrantagreement#637503 (Pristine) the ANR through the Cradle project, the UnivEarthS Labex program at Sorbonne Paris Cité (ANR-10-LABX-0023 and ANR-11-IDEX-0005-02). Parts of this work were supported by IPGPmultidisciplinary programPARI,andby RegionÎle-de-France SESAMEGrantno.12015908.WealsoareindebtedtoJoseph Boe-senberg andDentonEbel(American Museum ofNationalHistory, NewYork), TimothyMcCoy(US NationalMuseumofNatural His-tory,SmithsonianInstitution,WashingtonDC),CarolineSmith(The NaturalHistoryMuseum,London),AlexBevan(WesternAustralian Museum,Perth),JimKarnerandCarlAgee(UniversityofNew Mex-ico,Albuquerque),LudovicFerriere(NaturhistorischesMuseum, Vi-enna),PhilipHeck(TheFieldMuseum,Chicago),Meenakshi Wad-hwa (Arizona State University, Tempe),Cecilia Satterwhite (NASA Johnson SpaceCenter, Houston)andthe Comité de Gestion (Mu-seumNationale d’Histoire Naturelle,Paris) fortheir generous do-nationsofmeteoritesamplesforthisworkandtheirconﬁdencein ouranalyticalandscientiﬁccapabilities.

References

Alexander, C.M.O., 2005. Re-examining the role of chondrulesin producingthe elemental fractionations in chondrites. Meteorit. Planet. Sci. 40, 943–965.

http://doi.org/10.1111/j.1945-5100.2005.tb00166.

Alexander, C.M.O., Grossman, J.N., Wang, J., Zanda, B., Bourot-Denise, M., Hewins, R.H., 2000. The lack of potassium-isotopicfractionation in Bishun-purchondrules.Meteorit.Planet.Sci. 35,859–868. http://doi.org/10.1111/j.1945-5100.2000.tb01469.

Alexander, C., Grossman, J.N., Ebel, D.S., 2008. The formation conditions of chondrules and chondrites. Science 320, 1617–1619. http://doi.org/10.1126/ science.1157846.

Asphaug,E.,Jutzi,M.,Movshovitz,N.,2011.Hcondruleformationduring planetesi-malaccretion.EarthPlanet.Sci.Lett. 308,369–379.

Barrat,J.A.,Zanda,B.,Moynier,F.,Bollinger,C.,Liorzou,C.,Bayon,G.,2012. Geo-chemistryofCIchondrites:majorandtraceelements,andCuandZnisotopes. Geochim.Cosmochim.Acta 83,79–92.http://doi.org/10.1016/j.gca.2011.12.011. Becker,M.,Hezel,D.C.,Schulz,T.,Elfers,B.-M.,Münker,C.,2015.TheageofCV

chon-dritesfromcomponentspeciﬁcHf–Wsystematics.EarthPlanet.Sci.Lett. 432, 472–482.http://doi.org/10.1016/j.epsl.2015.09.049.

Bland,P.A.,Alard,O.,Benedix,G.K.,2005.Volatilefractionationintheearlysolar systemandchondrule/matrixcomplementarity.Proc.Natl.Acad.Sci.USA 102, 13755–13760.http://doi.org/10.1073/pnas.0501885102.

Bollard,J.,Connelly,J.N.,Bizzarro,M.,2015.Pb–Pbdatingofindividualchondrules from the CBa chondrite Gujba: assessment of the impact plume formation model.Meteorit.Planet.Sci. 50,1197–1216.http://doi.org/10.1111/maps.12461.

Brearley,A.J., 2003.Nebularversus parent-bodyprocessing. In:Treatiseon Geo-chemistry,vol. 1,pp. 247–268.

Budde,G.,Burkhardt,C.,Brennecka,G.A.,Fischer-Gödde,M.,Kruijer,T.,Kleine,T., 2016a.Molybdenumisotopicevidencefortheoriginofchondrulesanda dis-tinctgeneticheritageofcarbonaceousandnon-carbonaceousmeteorites.Earth Planet.Sci.Lett. 454,293–303.http://doi.org/10.1016/j.epsl.2016.09.020. Budde,G.,Kleine,T.,Kruijer,T.,Burkhardt,C.,Metzler,K.,2016b.Tungstenisotopic

constraintsontheageandoriginofchondrules.Proc.Natl.Acad.Sci.USA 113 (11),2886–2891.http://doi.org/10.1073/pnas.1524980113.

Chen,H.,Savage,P.S., Teng,F.-Z.,Helz, R.T.,Moynier,F.,2013.Zincisotope frac-tionation during magmatic differentiation and the isotopic composition of thebulkEarth.EarthPlanet.Sci.Lett. 369–370,34–42.http://doi.org/10.1016/ j.epsl.2013.02.037.

Chou, C.-L., Baedecker, P.A., Wasson, J.T., 1976. Allende inclusions: volatile-element distribution and evidence for incomplete volatilization of preso-larsolids.Geochim. Cosmochim.Acta 40,85–94. http://doi.org/10.1016/0016-7037(76)90196-4.

Ciesla,F.J.,Hood,L.L.,2002.Thenebularshockwavemodelforchondruleformation: shockprocessinginaparticle–gassuspension.Icarus 158,281–293.http://doi. org/10.1006/icar.2002.6895.

Clayton, R.N.,Mayeda,T.K.,1999. Oxygenisotopestudies ofcarbonaceous chon-drites. Geochim. Cosmochim. Acta 63, 2089–2104. http://doi.org/10.1016/ S0016-7037(99)00090-3.

Connolly, H.C.J., Love, S.G., 1998. The formation of chondrules: petrologic tests ofthe shockwavemodel. Science 280,62–67.http://doi.org/10.1126/science. 280.5360.62.

Connolly Jr., H.C., Desch, S.J., 2004. On the origin of the “kleine Kügelchen” called Chondrules. Chem. Erde 64, 95–125. http://doi.org/10.1016/j.chemer. 2003.12.001.

Desch,S.J.,Connolly,H.C.J.,2002.Amodelofthethermalprocessingofparticles insolarnebulashocks:applicationtothecoolingratesofchondrules.Meteorit. Planet.Sci. 37,183–207.http://doi.org/10.1111/j.1945-5100.2002.tb01104.x. Friend, P., Hezel,D.C., Mucerschi, D., 2016. The conditionsof chondrule

forma-tion,partII:opensystem.Geochim.Cosmochim.Acta 173,198–209.http://doi. org/10.1016/j.gca.2015.10.026.

Grossman, J.N., Wasson, J.T., 1985. The origin and history of the metal and sulﬁde components ofchondrules. Geochim. Cosmochim. Acta 49, 925–939.

http://doi.org/10.1016/0016-7037(85)90308-4.

Hezel,D.C.,Palme,H.,2008.Constraintsfor chondruleformationfromCa–Al dis-tribution incarbonaceous chondrites. EarthPlanet. Sci. Lett. 265, 716–725.

http://doi.org/10.1016/j.epsl.2007.11.003.

Hezel, D.C.,Palme, H.,2010. Thechemical relationshipbetweenchondrules and matrixandthechondrulematrixcomplementarity.EarthPlanet.Sci.Lett. 294, 85–93.

Hezel,D.C.,Needham,A.W.,Armytage,R.,Georg,B.,Abel,R.,Kurahashi,E.,Coles, B.J.,Rehkämper,M.,Russell,S.S.,2010.Anebulasettingastheoriginforbulk chondruleFeisotopevariations inCVchondrites.EarthPlanet.Sci.Lett. 296, 423–433.http://doi.org/10.1016/j.epsl.2010.05.029.

Hezel, D.C., Elangovan, P., Viehmann, S., Howard, L., Abel, R.L., Armstrong, R., 2013.VisualisationandquantiﬁcationofCVchondritepetrographyusing micro-tomography. Geochim. Cosmochim. Acta 116, 33–40. http://doi.org/10.1016/ j.gca.2012.03.015.

Humayun, M., Clayton, R.N., 1995. Potassium isotope cosmochemistry: genetic implications of volatile element depletion. Geochim. Cosmochim. Acta 59, 2131–2148.http://doi.org/10.1016/0016-7037(95)00132-8.

Jones,R.H.,2012.Petrographicconstraintsonthediversityofchondrulereservoirsin theprotoplanetarydisk.Meteorit.Planet.Sci. 47,1176–1190.http://doi.org/10. 1111/j.1945-5100.2011.01327.x.

Kato,C.,Valdes,M.C.,Dhaliwal,J.K.,Day,J.M.D.,Moynier,F.,2015.Extensivevolatile lossduringformationanddifferentiationoftheMoon.Nat.Commun. 6,1–4.

http://doi.org/10.1038/ncomms8617.

King,A.J.,Schoﬁeld,P.F.,Howard,K.T.,Russell,S.S.,2015.ModalmineralogyofCIand CI-likechondritesbyX-raydiffraction.Geochim.Cosmochim.Acta 165,148–160.

Krot,A.N.,Amelin,Y.,Cassen,P.,Meibom,A.,2005.YoungchondrulesinCB chon-drites froma giantimpact inthe earlySolar System.Nature 436,989–992.

http://doi.org/10.1038/nature03830.

Krot,A.N.,Amelin,Y.,Bland,P.,Ciesla,F.J.,Connelly,J.,Davis,A.M.,2009.Originand chronologyofchondriticcomponents:areview.Geochim.Cosmochim.Acta 73, 4963–4997.http://doi.org/10.1016/j.gca.2008.09.039.

Larimer,J.W.,Anders,E.,1967.Chemicalfractionationsinmeteorites—II.Abundance patternsand theirinterpretation.Geochim. Cosmochim.Acta 31,1239–1270.

http://doi.org/10.1016/S0016-7037(67)80014-0.

Lodders,K.,2003.Solarsystemabundancesandcondensationtemperaturesofthe elements.Astrophys.J. 591,1220–1247.http://doi.org/10.1086/375492. Luck,J.M.,Othman,D.B.,Barrat,J.A.,Albarède,F.,2003.Coupled63Cuand16O

ex-cessesinchondrites.Geochim.Cosmochim.Acta 67,143–151.http://doi.org/10. 1016/S0016-7037(02)01038-4.

Luck, J.-M., Othman, D.B., Albarède, F., 2005. Zn and Cu isotopic variations in chondrites and iron meteorites: early solar nebula reservoirs and parent-bodyprocesses. Geochim. Cosmochim.Acta 69,5351–5363.http://doi.org/10. 1016/j.gca.2005.06.018.

Moynier,F.,LeBorgne,M.,2015.Highprecisionzincisotopicmeasurementsapplied tomouseorgans.J.Vis.Exp. 99,1.http://doi.org/10.3791/52479.

Moynier,F.,Blichert-Toft,J.,Telouk,P.,Luck,J.-M.,Albarède,F.,2007.Comparative stableisotopegeochemistryofNi,Cu,Zn,andFeinchondritesandiron mete-orites.Geochim.Cosmochim.Acta 71,4365–4379.

Moynier,F., Beck,P.,Jourdan,F., Yin,Q.-Z.,Reimold,U., Koeberl,C., 2009a. Iso-topic fractionation ofzinc intektites. Earth Planet.Sci. Lett. 277,482–489.

http://doi.org/10.1016/j.epsl.2008.11.020.

Moynier,F.,Dauphas,N.,Podosek,F.,2009b.Asearchfor70Znanomaliesin mete-orites.Astrophys.J.Lett. 700,L92–L95.

Moynier,F.,Paniello,R.C.,Gounelle,M.,Albarède,F.,Beck,P.,Podosek,F.,Zanda,B., 2011.Natureofvolatiledepletionandgeneticrelationshipsinenstatite chon-dritesandaubritesinferredfromZnisotopes.Geochim.Cosmochim.Acta 75, 297–307.

Moynier,F.,Vance,D.,Fujii,T.,Savage,P.S.,2017.Theisotopegeochemistryofzinc andcopper.In:Teng,F.Z.,Watkins,J.,Dauphas,N.(Eds.),Non-traditionalStable Isotopes.In:ReviewsinMineralogyandGeochemistry,vol. 82,pp. 543–600. Palme,H.,Lodders,K.,Jones,A.,2014a.Solarsystemabundancesoftheelements.

(11)

Palme,H.,Spettel,B.,Hezel,D.C.,2014b.Siderophileelementsinchondrulesof CV-chondrites.Chem.Erde 74,507–516.

Palme,H.,Hezel,D.C.,Ebel,D.S.,2015.Theoriginofchondrules:constraintsfrom matrixcompositionand matrix-chondrulecomplementarity.EarthPlanet.Sci. Lett. 411,11–19.http://doi.org/10.1016/j.epsl.2014.11.033.

Paniello,R.C.,Day,J.M.D.,Moynier,F.,2012a.Zincisotopicevidencefortheoriginof theMoon.Nature 490,376–379.http://doi.org/10.1038/nature11507.

Paniello,R.C., Moynier,F.,Beck, P.,Barrat, J.-A., Podosek,F.A., Pichat, S., 2012b. ZincisotopesinHEDs:cluestotheformationof4-vEsta,andtheunique com-positionofPecora Escarpment82502.Geochim.Cosmochim.Acta 86,76–87.

http://doi.org/10.1016/j.gca.2012.01.045.

Qin,L.,Alexander,C.M.O.D.,Carlson,R.W.,Horan,M.F.,Yokoyama,T.,2010. Contribu-torstochromiumisotopevariationofmeteorites.Geochim.Cosmochim.Acta 74 (3),1122–1145.

Richter,F.,2004.Timescalesdeterminingthedegreeofkineticisotopefractionation byevaporationandcondensation.Geochim.Cosmochim.Acta 68,4971–4992.

Ringwood,A.E.,1966.Chemicalevolutionoftheterrestrialplanets.Geochim. Cos-mochim.Acta 30,41–104.http://doi.org/10.1016/0016-7037(66)90090-1. Rubin,A.E.,Wasson,J.T., 1987. Chondrules, matrixand coarse-grained chondrule

rims in the Allende meteorite: origin, interrelationships and possible pre-cursorcomponents.Geochim. Cosmochim.Acta 51, 1923–1937.http://doi.org/ 10.1016/0016-7037(87)90182-7.

Schrader,D.L., Franchi, I.A., Connolly,H.C.J.,Greenwood, R.C.,Lauretta, D.S., Gib-son,J.M.,2011.TheformationandalterationoftheRenazzo-likecarbonaceous

chondrites I:implicationsofbulk-oxygenisotopiccomposition.Geochim. Cos-mochim.Acta 75,308–325.http://doi.org/10.1016/j.gca.2010.09.028.

Scott, E.R.D.,Krot,A.N., 2014.Chondritesand theircomponents. In:Holland,H., Turekian,K.(Eds.),TreatiseonGeochemistry,vol. 1,pp. 65–137.http://doi.org/ 10.1016/B978-0-08-095975-7.00104-2.

Shu,F.H.,Shang,H.,Lee,T., 1996.Toward anastrophysicaltheoryofchondrites. Science 271,1545–1552.http://doi.org/10.1126/science.271.5255.1545. Trinquier,A.,Birck,J.-L.,Allègre,C.J.,2007.Widespread54Crheterogeneityinthe

innersolarsystem.Astrophys.J. 655,1179–1185.http://doi.org/10.1086/510360. Uesugi,M.,Sekiya,M.,Nakamura,T.,2008.Kineticstabilityofameltediron glob-uleduringchondruleformation.I.Non-rotatingmodel.Meteorit.Planet.Sci. 43, 717–730.http://doi.org/10.1111/j.1945-5100.2008.tb00680.x.

Wasson,J.T.,Chou,C.-L.,1974.Fractionationofmoderatelyvolatileelementsin or-dinarychondrites.Meteoritics 9,69–84.

Wasson, J.T., Rubin, A.E., 2010. Metal in CR chondrites. Geochim. Cosmochim. Acta 74,2212–2230.http://doi.org/10.1016/j.gca.2010.01.014.

Weisberg, M.K.,Prinz,M.,Clayton, R.N.,Mayeda,T.K.,Sugiura, N.,Zashu,S., Ebi-hara,M.,2001.Anewmetal-richchondritegrouplet.Meteorit.Planet.Sci. 36, 401–418.http://doi.org/10.1111/j.1945-5100.2001.tb01882.x.

Zanda,B.,Hewins,R.H.,Bourot-Denise,M.,Bland,P.A.,Albarède,F.,2006.Formation ofsolarnebulareservoirsbymixingchondriticcomponents.EarthPlanet.Sci. Lett. 248,650–660.http://doi.org/10.1016/j.epsl.2006.05.016.