2067 copy The Meteoritical Society 2007 Printed in USA

Meteoritics amp Planetary Science 42 Nr 12 2067ndash2077 (2007)Abstract available online at httpmeteoriticsorg

Mass-dependent fractionation of nickel isotopes in meteoritic metal

David L COOK1 2 3dagger Meenakshi WADHWA1 2 3daggerdagger Robert N CLAYTON1 2 4 Nicolas DAUPHAS1 2 4 Philip E JANNEY3daggerdagger and Andrew M DAVIS1 2 4

1Department of the Geophysical Sciences The University of Chicago 5734 South Ellis Avenue Chicago Illinois 60637 USA2Chicago Center for Cosmochemistry 5640 South Ellis Avenue Chicago Illinois 60637 USA

3Department of Geology The Field Museum 1400 South Lake Shore Drive Chicago Illinois 60605 USA4Enrico Fermi Institute 5640 South Ellis Avenue Chicago Illinois 60637 USA

daggerPresent address Department of Chemistry and Chemical Biology Rutgers University 610 Taylor Road Piscataway New Jersey 08854 USA

daggerdaggerPresent address Center for Meteorite Studies School of Earth and Space Exploration Arizona State University PO Box 871404 Tempe Arizona 85287ndash1404 USA

Corresponding author E-mail davecookrcirutgersedu

(Received 02 February 2007 revision accepted 08 June 2007)

AbstractndashWe measured nickel isotopes via multicollector inductively coupled plasma massspectrometry (MC-ICPMS) in the bulk metal from 36 meteorites including chondrites pallasites andirons (magmatic and non-magmatic) The Ni isotopes in these meteorites are mass fractionated thefractionation spans an overall range of asymp04permil amuminus1 The ranges of Ni isotopic compositions (relativeto the SRM 986 Ni isotopic standard) in metal from iron meteorites (asymp00 to asymp03permil amuminus1) andchondrites (asymp00 to asymp02permil amuminus1) are similar whereas the range in pallasite metal (asympminus01 to00permil amuminus1) appears distinct The fractionation of Ni isotopes within a suite of fourteen IIIAB irons(asymp00 to asymp03permil amuminus1) spans the entire range measured in all magmatic irons However the degreeof Ni isotopic fractionation in these samples does not correlate with their Ni content suggesting thatcore crystallization did not fractionate Ni isotopes in a systematic way We also measured the Ni andFe isotopes in adjacent kamacite and taenite from the Toluca IAB iron meteorite Nickel isotopesshow clearly resolvable fractionation between these two phases kamacite is heavier relative to taeniteby asymp04permil amuminus1 In contrast the Fe isotopes do not show a resolvable fractionation between kamaciteand taenite The observed isotopic compositions of kamacite and taenite can be understood in termsof kinetic fractionation due to diffusion of Ni during cooling of the Fe-Ni alloy and the developmentof the Widmanstaumltten pattern

INTRODUCTION

Studies of the mass-dependent fractionation of stableisotopes of light elements (eg H C N O) have been andstill are a useful tool for investigating early solar systemprocesses operating in the nebula and on parent bodies Theadvent of multi-collector ICPMS has greatly expanded therange of elements now available for such investigations withthe transition metals being of particular interest Studies of Feisotopes in meteorites (eg Zhu et al 2001 2002 Kehm et al2003 Poitrasson et al 2004 2005 Weyer et al 2005) as wellas of Cu and Zn isotopes (Luck et al 2005) show that theseelements underwent mass-dependent fractionation duringprocesses occurring in the early solar system These studieshave attributed mass-dependent fractionation of transitionmetals to both low- and high-temperature events that took

place in the solar nebula and on meteorite parent bodiesExcept for isotopic anomalies found in refractory inclusionsfrom carbonaceous chondrites (eg Voumllkening andPapanastassiou 1989 1990) the isotopes of Fe and Zn appearto have been fractionated from an originally homogeneousreservoir in the early solar system (Zhu et al 2001 Kehmet al 2003 Dauphas et al 2004 Luck et al 2005)

Analyses of metallic deep-sea spherules havedemonstrated that Ni isotopes can be mass fractionated due toevaporation (Davis and Brownlee 1993 Herzog et al 1994Engrand et al 2005) Furthermore Ni isotopes in metal fromvarious meteorite groups follow a mass-dependentfractionation trend (Moynier et al 2004 2005) Neverthelessthe extent and causes of mass-dependent fractionation of Niisotopes in natural samples remain largely unexplored Thuswe have analyzed the Ni isotopic composition of meteoritic

2068 D L Cook et al

metal from chondrites pallasites and iron meteorites with thefollowing goals 1) to constrain the range of mass-dependentfractionation of Ni isotopes in these objects 2) to examine thepossible effects of fractional crystallization of a liquid Fe-Nialloy during core formation on the isotopic composition of Niand 3) to investigate the fractionation of Ni isotopes betweenkamacite and taenite in iron meteorites

ANALYTICAL METHODS

Samples and Sample Preparation

Metal from 36 different meteorites was analyzed Thesemeteorites included 6 chondrites representing the LL H EHCR and CBa groups 3 main group pallasites the EagleStation pallasite and 26 iron meteorites from magmatic(IIAB IID IIIAB IVA IVB) and non-magmatic (IAB)groups All metal samples from chondrites pallasites non-magmatic irons and some magmatic irons are those for whichnon-mass-dependent effects in Ni isotopes were previouslyinvestigated (Cook et al 2006) In addition to these a suite offourteen IIIAB irons was also selected to investigate thepotential effects of core crystallization on Ni isotopes Forthis purpose we acquired six samples from the Field Museumin Chicago and eight from the Natural History Museum inLondon These latter eight samples were originally allocatedto E Mullane for an investigation of Fe isotopes (Mullane etal 2005) Digested but chemically unprocessed aliquots ofthese samples were obtained from E Mullane for this studyAliquots of these fourteen IIIAB irons were derived fromdissolved samples of relatively large pieces (03 to 14 g) tomake sure that they were representative of the bulk metal inthese meteorites For all other iron meteorite samples and forpallasite samples a small piece of fresh metal was cut using aslow speed saw equipped with a diamond wafering bladeChondrite samples were crushed and the metal was separatedwith a hand magnet Additional details of the samplepreparation techniques are described by Cook et al (2006)

The Toluca IAB iron meteorite was chosen in order toinvestigate the isotopic fractionation between kamacite andtaenite A polished and etched thick section of Toluca wasexamined via scanning electron microscopy (SEM) (JEOL5800 LV with Oxford Link Isis-300 X-ray microanalysissystem University of Chicago) to identify Ni-poor (kamacite)and Ni-rich (taenite) metal Elemental X-ray maps coveringan area of 28 cm2 were acquired for Fe Ni S and P Fivesamples were then collected along a traverse of adjacent Ni-poor and Ni-rich metal regions using a computer-controlledmicrodrill (MicroMill New Wave Research) equipped with atungsten carbide mill bit (Brasseler scriber H162111) Thedrilled pits were asymp400 μm in diameter and 700 μm in depthAlthough care was taken to obtain pure samples of kamaciteand taenite from Toluca it is possible that mixtures of bothphases were collected given the depths of the drilled spots

The potential consequences of such sampling issues areaddressed in the discussion of the observed isotopicfractionation between kamacite and taenite

Sample Digestion and Ni Separation

Details of sample digestion and Ni separation areprovided by Cook et al (2006) and are described only brieflyhere Metal samples separated from chondrites and twopallasites (Albin and Molong) that contained visible olivineinclusions were digested in 1 M HCl The solutions were thencentrifuged and the supernatant was decanted (thusundigested silicate grains were excluded) evaporated todryness and redigested in reverse aqua regia (21 concHNO3 conc HCl) All other samples were digested in reverseaqua regia only Nickel was separated using a combination ofanion and cation exchange chromatography Depending onwhether or not there was significant Ti in our samplesolutions we used either a two-column or three-columnprotocol for Ni separation Additionally the Fe fractions forthe microdrilled Toluca samples and the six IIIAB samplesfrom the Field Museum were recovered from the first anionexchange column as described by Dauphas et al (2004)

Isotopic Measurements

Nickel isotopic measurements were made on aMicromass IsoProbe MC-ICPMS housed in the IsotopeGeochemistry Laboratory of the Field Museum All fivestable Ni isotopes (58 60 61 62 64) were measuredsimultaneously and 57Fe and 66Zn were monitored and usedto correct for isobaric interferences on 58Ni from 58Fe and on64Ni from 64Zn Samples were corrected for instrumentalmass bias using the standard-sample bracketing techniqueThe NIST SRM 986 reference material was used as thebracketing standard and all Ni isotope data are reportedrelative to this standard Each reported Ni datum representsthe mean of a minimum of four (but typically five or more)repeat measurements of the sample bracketed bymeasurements of SRM 986 Iron isotopes were also measuredin the microdrilled samples from Toluca and six of the IIIABirons (see previous section) Iron isotopes were measured onthe same instrument using IRMM-014 as the bracketingstandard A more detailed discussion of the measurementprotocols for Ni and Fe isotopic analyses is provided by Cooket al (2006) and Dauphas et al (2004) respectively

Precision and Accuracy

All Ni isotope ratio data are reported relative to SRM 986in units of permil (permil) as given by

δij = [(Rsample minus Rstandard)Rstandard] times 103 (1)

where R is the iNi58Ni ratio (i = 60 61 62 or 64) In most

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2069

cases the data are discussed using the 61Ni58Ni ratio becausethis ratio is free from possible effects due to decay of theshort-lived isotope 60Fe (t12 = 149 My) and also frompotential nucleosynthetic effects on the most neutron-rich Niisotopes (62 and 64) It is convenient to discuss the massfractionation effects in terms of permil per amu (permil amuminus1)calculated using

FNi = δ61583 (2)

The internal precisions of the Ni isotopic compositions ofindividual samples based on a number of repeatmeasurements are represented by the standard error of themean (2SE) Repeated analyses of an Aesar Ni solutionbracketed with the SRM 986 standard over a 27-month periodwere used to determine the external precision of our Niisotope ratio measurements Figure 1 shows the FNi valuesfrom these measurements each datum is the mean valuecalculated from 5 repeat measurements during a singleanalytical session These data show that the Aesar Ni solutionis mass fractionated relative to SRM 986 by asympminus01permil amuminus1and the external precision (2SD) is plusmn0065permil amuminus1 Theexternal precision based on the other Ni isotope ratios issimilar (δ60582 plusmn0063permil amuminus1 δ62584 plusmn0061permil amuminus1δ64586 plusmn0066permil amuminus1)

Isotopic fractionation during chemical processing ofsamples can potentially affect the isotopic composition andlead to inaccurate measurements To demonstrate that our Niseparation procedure does not introduce any analyticalartifact within uncertainties three aliquots each of the SRM986 isotopic standard were processed through both the two-column and the three-column Ni separation protocols (Cook

et al 2006) Figure 2 shows the FNi values for the sixprocessed aliquots of SRM 986 Within the uncertaintiesnone of the measured aliquots of the processed SRM 986 havean isotopic composition that differs from the unprocessedSRM 986 used as a bracketing standard Additionally noresolvable effects were observed on the other three measuredNi isotope ratios (δ6058 δ6258 δ6458) Hence these datademonstrate that our Ni separation chemistry does notintroduce any resolvable mass-dependent isotopicfractionation effects

Effects of Fe Zn and Signal Intensity

Corrections for isobaric interferences on 58Ni from 58Feand on 64Ni from 64Zn must be applied if Fe andor Zn arepresent in the sample solutions Furthermore Fe is thedominant element in meteoritic metal and an inefficientseparation of Ni from Fe could introduce matrix effects thatmay alter the isotopic measurements Thus aliquots of SRM986 were doped with varying amounts of Fe and Zn(4 aliquots each) to test our ability to correct Fe and Zninterferences and to quantify the possible matrix effects dueto the presence of Fe Figure 3 shows the interference-corrected FNi values for the SRM 986 aliquots doped withFe as well as an undoped aliquot These data show that theFe interference correction is effective and that no resolvablematrix effects are present for FeNi le 1 Additionally noresolvable effects were observed on the other threemeasured Ni isotope ratios (δ6058 δ6258 δ6458) due to thepresence of Fe Cook et al (2006) showed that theinterference correction for 64Zn on 64Ni is not effective forZnNi ge 001 when measuring non-mass-dependent effects

Fig 1 FNi values (permil amuminus1) for repeated analyses of an Aesar Nisolution over the course of a 27-month period Each datum representsthe mean of five repeat measurements performed during a singleanalysis session The individual error bars (2SE) are based on the fiverepeat measurements The external precision is the standarddeviation (2SD) based on all of the plotted data and is shown by thetwo gray lines (plusmn0065permil amuminus1)

Fig 2 FNi values (permil amuminus1) for six aliquots of the SRM 986 isotopicstandard processed through our Ni separation procedure (squares2-column protocol circles 3-column protocol) Plotted errors are2SE the external precision (2SD) is shown by the two dashed lines(plusmn0065permil amuminus1)

2070 D L Cook et al

This is also true for measurements of mass-dependenteffects However no resolvable effects were observed onthe other three measured Ni isotope ratios (δ6058 δ6158δ6258) due to the presence of Zn thus FNi was unaffected inthe range of ZnNi ratios investigated (10minus5 to 10minus1) Similarto our previous work (Cook et al 2006) all measuredsample solutions had FeNi and ZnNi ratios lt50 times 10minus2

and lt45 times 10minus4 respectively The interference correctionfor FNi was typically le001permil amuminus1

We also investigated whether a mismatch between the Nisignal intensities for standard and sample solutions couldaffect our mass-dependent isotopic measurements Sixaliquots of SRM 986 with Ni concentrations ranging from055 to 10 ppm were analyzed by bracketing with a 10 ppmSRM 986 solution Figure 4 shows the FNi values from themeasurements of these six aliquots of SRM 986 No clearlyresolvable effects were observed except when the samplesignal differed from the standard signal by 45 Similarresults were obtained for the other three measured Ni isotoperatios (δ6058 δ6258 δ6458) All measured sample solutions ofmeteoritic metal had signal intensities within 10 of thestandard signal intensity Thus all sample and standard signalintensities were matched adequately to provide accuratemeasurements

RESULTS AND DISCUSSION

In our previous study of non-mass-dependent effects inNi isotopes in the metal of chondrites pallasites and ironmeteorites (Cook et al 2006) we demonstrated that thereare small deficits in the mass-bias-corrected 60Ni58Ni ratiorelative to the terrestrial standard (ie ε60) in the metalfrom some primitive chondrites and possibly some iron

meteorites and that these deficits are most likely due to thepresence of live 60Fe (t12 = 149 My) at the time of FeNifractionation in these metal samples However there are noother clearly resolvable non-mass-dependent effects in anyof the other measured Ni isotope ratios As such in Table 1we present the results of our measurements of the 61Ni58Ni62Ni58Ni and 64Ni58Ni ratios relative to the SRM 986standard in permil (ie 6158 6258 and 6458) For eachof the samples the Ni isotope fractionation in terms ofpermil amuminus1 is the same within uncertainties regardless ofwhich ratio (ie 61Ni58Ni 62Ni58Ni or 64Ni58Ni) isconsidered Thus the data are discussed primarily in termsof FNi (Equation 2)

We note that while we have performed analyses of asingle piece of bulk metal from most of the meteoritesstudied we analyzed two separate pieces (identified as 1and 2 in Table 1) of three magmatic IIIAB iron meteorites(Bella Roca Casas Grandes and Henbury) Sample 1 ineach of these three cases was a small piece (lt50 mg) thatwas previously analyzed primarily to investigate non-mass-dependent effects in Ni isotopes (Cook et al 2006) Weadditionally selected a second larger piece (gt300 mg) ofeach of these three samples with the goal of obtaining arepresentative sample of the bulk metal for ourinvestigation of the mass-dependent fractionation of Niisotopes in the IIIAB iron meteorites In the case ofHenbury the Ni isotope compositions of both pieces agreewithin error This is not the case for the two pieces of BellaRoca and Casas Grandes It is likely that the smaller piecesof these two meteorites do not provide representative Niisotopic compositions of the bulk metal due to thecoarseness of the Widmanstaumltten pattern Our results for

Fig 3 FNi values (permil amuminus1) versus the FeNi elemental ratio Dataare from analyses of four aliquots of SRM 986 doped with variousconcentrations of Fe and one undoped aliquot (FeNi asymp 10minus4) Plottederrors are 2SE the external precision (2SD) is shown by the twodashed lines (plusmn0065permil amuminus1)

Fig 4 FNi values (permil amuminus1) versus the ratio of the sample tostandard signal intensity Data are from analyses of six aliquots ofSRM 986 with Ni concentrations ranging from 055 to 10 ppmbracketed by a 10 ppm SRM 986 solution Plotted errors are 2SE theexternal precision (2SD) is shown by the two dashed lines(plusmn0065permil amuminus1)

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2071

individual kamacite and taenite in Toluca (discussed below)show that Ni isotopes are fractionated between these twophases and can lead to mass-dependent Ni isotopicvariations in iron meteorites at the millimeter scaleTherefore small metal samples of iron meteorites (andpallasites) characterized by a coarse Widmanstaumltten patternmay be skewed toward the isotopic composition ofkamacite or taenite As such in the discussion on the Niisotope fractionation in bulk samples of IIIAB ironmeteorites only the larger pieces (samples 2 in Table 1) ofeach of these three meteorites are considered

Meteoritic Metal Bulk Samples

The Ni isotopic compositions of the bulk metal from the36 meteorites analyzed (Table 1) are shown in a three-isotopeplot (Fig 5) Within error all samples in Fig 5 plot on a 11line which is consistent with mass-dependent isotopicfractionation The overall range of fractionation is041permil amuminus1 The observed mass-dependent fractionationand the overall range are both consistent with the resultspreviously reported by Moynier et al (2004 2005) for metalfrom iron meteorites pallasites chondrites and mesosiderites

Table 1 The Ni isotopic composition of meteoritic bulk metal (permil) relative to the SRM 986 standard Sample Group δ6158 plusmn 2SE δ6258 plusmn 2SE δ6458 plusmn 2SE n

Renazzoa CR 047 plusmn 012 058 plusmn 010 086 plusmn 014 14Gujbaa CBa 060 plusmn 012 084 plusmn 012 121 plusmn 019 13Semarkonaa LL 30 018 plusmn 012 009 plusmn 017 016 plusmn 029 10Bishunpura LL 31 043 plusmn 014 053 plusmn 018 071 plusmn 026 5Forest Valea H4 053 plusmn 026 068 plusmn 020 103 plusmn 045 5Indarcha EH4 026 plusmn 021 030 plusmn 024 049 plusmn 045 5

Eagle Stationa PES 015 plusmn 005 013 plusmn 007 039 plusmn 014 5Albina PMG 002 plusmn 010 013 plusmn 009 025 plusmn 013 9Brenhama PMG minus025 plusmn 016 minus034 plusmn 007 minus034 plusmn 036 5Molonga PMG minus019 plusmn 008 minus027 plusmn 011 minus041 plusmn 029 5

Coahuilaa IIAB 021 plusmn 024 032 plusmn 021 059 plusmn 061 5Santa Luziaa IIAB 059 plusmn 006 076 plusmn 004 126 plusmn 013 10Carboa IID 097 plusmn 041 130 plusmn 045 190 plusmn 086 5Augustinovkab IIIAB minus005 plusmn 022 minus012 plusmn 027 minus020 plusmn 061 5Avocab IIIAB 022 plusmn 017 028 plusmn 024 035 plusmn 026 5Bald Eagleb IIIAB 036 plusmn 026 054 plusmn 018 085 plusmn 051 5Bear Creek IIIAB 015 plusmn 022 027 plusmn 023 061 plusmn 039 5Bella Roca 1a IIIAB 008 plusmn 016 010 plusmn 012 005 plusmn 014 8Bella Roca 2b 092 plusmn 025 117 plusmn 023 177 plusmn 045 5Carthage IIIAB 055 plusmn 006 066 plusmn 007 108 plusmn 010 5Casas Grandes 1a IIIAB 062 plusmn 003 082 plusmn 003 135 plusmn 009 14Casas Grandes 2 029 plusmn 012 037 plusmn 004 061 plusmn 006 5Grant IIIAB 013 plusmn 024 010 plusmn 020 023 plusmn 039 5Henbury 1a IIIAB 033 plusmn 004 042 plusmn 004 061 plusmn 007 14Henbury 2b 031 plusmn 022 036 plusmn 016 045 plusmn 010 5Nova Petropolisb IIIAB 024 plusmn 013 026 plusmn 011 030 plusmn 008 5Orange River Ironb IIIAB 023 plusmn 031 025 plusmn 041 043 plusmn 094 4Seneca Falls IIIAB 037 plusmn 011 050 plusmn 014 081 plusmn 015 5Thunda IIIAB 042 plusmn 016 046 plusmn 005 073 plusmn 012 5Wellandb IIIAB 072 plusmn 058 089 plusmn 071 140 plusmn 119 5Gibeona IVA 028 plusmn 021 037 plusmn 009 078 plusmn 025 5Yanhuitlana IVA 064 plusmn 018 075 plusmn 016 112 plusmn 045 5Cape of Good Hopea IVB 036 plusmn 033 043 plusmn 014 085 plusmn 026 5Hobaa IVB 045 plusmn 011 064 plusmn 012 094 plusmn 018 13Tlacotepeca IVB 014 plusmn 035 022 plusmn 048 033 plusmn 069 9

Canyon Diabloa IAB 039 plusmn 006 051 plusmn 005 095 plusmn 012 19Tolucaa IAB 044 plusmn 013 062 plusmn 009 094 plusmn 010 5Daytona IAB minus015 plusmn 010 minus014 plusmn 005 001 plusmn 035 5Mundrabillaa IAB 026 plusmn 007 036 plusmn 006 055 plusmn 011 14

aSamples previously investigated for non-mass-dependent effects in Ni isotopes (Cook et al 2006)bSample solutions provided by E Mullane

n = Number of repeat measurements

2072 D L Cook et al

The FNi values for all 36 bulk metal samples are plottedin Fig 6 according to meteorite class Figure 6 shows that 1)the metal from chondrites and from iron meteorites(magmatic and non-magmatic) is typically similar to orheavier than SRM 986 and 2) metal from pallasites is similarto or lighter than SRM 986 The mean and two standarddeviations (2SD) of the FNi values for each class are asfollows chondrites (014 plusmn 011permil amuminus1) iron meteorites(012 plusmn 017permil amuminus1) and pallasites (minus002 plusmn 012permil amuminus1)The Ni isotopic compositions for each meteorite class werecompared to one another using t-tests These comparisonsshow that there is no significant difference between thechondrites and the iron meteorites (t = 0417) However theNi isotopic composition for pallasite metal is significantlydifferent from both chondrite metal (t = 4325) and from ironmeteorites (t = 3200) It is possible that the mean Ni isotopiccomposition of pallasitic metal is biased due to the smallnumber of samples measured However if this is not the casethere are several possibilities for explaining the apparentlydistinct Ni isotopic composition of pallasite metal Ifpallasites form at the core-mantle boundary of planetesimals(eg Scott 1977) the Ni isotopic composition of pallasitemetal may reflect isotopic fractionation between the metalliccore and the overlying mantle silicates Zhu et al (2002) andPoitrasson et al (2005) reported that Fe isotopes arefractionated between metal and olivine in pallasites and theseeffects range from 002 to 011permil amuminus1 Recent experimentsindicate that Fe isotopes can undergo fractionation up toasymp23permil amuminus1 between metallic and silicate melts due tokinetic effects (Roskosz et al 2006) and are fractionated by

asymp017permil amuminus1 between sulfide and silicate melts attemperatures between 840 and 1000 degC due to equilibriumeffects (Schuessler et al 2007) Currently a lack of Niisotopic data for non-metal phases from pallasites (egsilicates and oxides) precludes an assessment of planetarydifferentiation as a potential explanation for the apparentlydistinct Ni isotopic composition of pallasite metalAlternatively the Ni isotopic compositions of pallasite metalmay reflect differences at the parent-body scale that wereestablished prior to accretion rather than reflectingfractionation effects from differentiation and core formationon their parent bodies Finally it is also possible that in thecase of some of the smaller samples (ie lt50 mg) frommeteorites with metal that is compositionally heterogeneouson the millimeter-scale (eg the four pallasites) themeasured Ni isotope compositions may not be representativeof the bulk metal in these meteorites Analyses of additionaland larger samples would help to ascertain if pallasitic metaldoes indeed have a mean Ni isotopic composition that differssignificantly from metal in other meteorite classes

IIIAB Iron Meteorites

Nickel IsotopesWith over two hundred members the IIIABs constitute

the largest group of magmatic iron meteorites Thesemeteorites cover a wide compositional range and the originalIIIAB core is likely well-represented by the large number ofsamples (Chabot and Haack 2006) Mullane et al (2005)reported that Fe isotopes from a suite of 33 IIIAB irons weremass fractionated over a large range of 049permil amuminus1however the Fe isotopic compositions did not correlate with

Fig 5 A three-isotope plot of the data from the analyses of bulkmetal from 36 meteorites The data follow the trend expected formass-dependent fractionation a line with a slope equal to 1 is shownfor reference The overall range of fractionation is 041permil amuminus1Plotted errors are 2SE Filled triangles = chondrites squares =pallasites open triangles = magmatic irons circles = non-magmaticirons

Fig 6 FNi values (permil amuminus1) for the bulk metal from 36 meteoritesPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Filled triangles = chondritessquares = pallasites open triangles = magmatic irons circles = non-magmatic irons

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2073

the Ni contents which are a proxy for the degree ofcrystallization of the metallic core (eg Scott 1972) Hencefourteen IIIAB irons were selected to examine the degree ofmass-dependent fractionation of Ni isotopes possibly due tofractional crystallization of the core and to determine whetherthe fractionation effects on Fe and Ni isotopes werecorrelated Figure 7 shows the individual members of theIIIAB group in terms of their Ir and Ni contents including thefourteen samples chosen for Ni isotopic analyses Sampleswere chosen to represent the full compositional range of theIIIAB irons Figure 8 shows the FNi values for the fourteenIIIAB irons The Ni isotopes show fractionation over aconsiderable range from minus002 to 031permil amuminus1 The earliercrystallized low-Ni IIIA irons and the later crystallized high-Ni IIIB irons have similar mean compositions of01permil amuminus1 Furthermore the degree of mass fractionationof Ni isotopes does not correlate with the Ni content (r2 =0003) The data in Fig 8 suggest that core crystallization didnot mass fractionate Ni isotopes in a systematic way

The isotopic composition of Ni in different IIIAB ironmeteorites likely reflects a combination of events such asfractionation between 1) liquid metal and silicate during coresegregation 2) immiscible melts in the molten core and 3)various phases during core solidification The later stages ofcore crystallization of iron meteorites were likely affected byliquid immiscibility resulting in the formation ofcompositionally distinct liquids due to the presence of S and Pin the molten core (Ulff-Moslashller 1998 Chabot and Drake2000) Nickel isotopes may have been fractionated as a resultof the interaction of a metallic melt with a S andor P-richmelt during core crystallization Furthermore Fe isotopes arefractionated between metal phosphide (ie schreibersite)

and sulfide (ie troilite) in pallasites (Poitrasson et al 2005Weyer et al 2005) and between metal and sulfide in ironmeteorites (Williams et al 2006) Crystallization followed byslow subsolidus cooling of metallic and non-metallic phasesin IIIAB irons may have fractionated Ni isotopes in a similarmanner This last mechanism is particularly appealingbecause the Ni concentrations are quite different betweenmetal troilite and schreibersite in iron meteorites (Buchwald1977 Jochum et al 1980) Diffusion of Ni along theseconcentration gradients during crystallization and cooling ofthe various phases would likely be accompanied by kineticisotopic fractionation The proximity of non-metallic phasesrelative to the metal could produce metal in iron meteoriteswith a range of FNi values and potentially account for theobserved variation among the IIIAB irons shown in Fig 8Measurements of Ni isotopes in adjacent metallic and non-metallic phases in iron meteorites could be used to test thishypothesis

Iron IsotopesMullane et al (2005) reported Fe isotope fractionation

over a large range (049permil amuminus1) for IIIAB irons in whichsamples were both heavier and lighter than the IRMM-014standard These data are inconsistent with other studies of Feisotopes in IIIAB iron meteorite metal in which all sampleshad Fe isotopic compositions similar to or slightly heavierthan IRMM-014 and spanned a narrow range of007permil amuminus1 (Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006) We analyzedthe Fe isotopes in six of the IIIAB irons with low and high Niabundances in which we also measured Ni isotopes The56Fe54Fe ratios in permil amuminus1 (or FFe) for these six samples are

Fig 7 A plot of Ir versus Ni abundances for members of the IIIABiron meteorite group The fourteen samples selected for Ni isotopicanalyses are shown by the black circles Data are from Scott et al(1973) Scott and Wasson (1976) Kracher et al (1980) Malvin et al(1984) and Wasson et al (1989 1998)

Fig 8 FNi values (permil amuminus1) versus Ni abundances for IIIAB ironsPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Nickel abundance data are fromScott et al (1973) Scott and Wasson (1976) and Malvin et al(1984)

2074 D L Cook et al

shown in Fig 9 along with Fe isotope data for IIIABs fromother studies As shown in Fig 9 our measurements do notshow the large range reported by Mullane et al (2005) All sixsamples have Fe isotopic compositions similar to orsomewhat heavier than IRMM-014 and span an overall rangeof only 007permil amuminus1 Both of these features are consistentwith the measurements of Zhu et al (2001) Kehm et al(2003) Schoenberg et al (2006) and Williams et al (2006)Furthermore there is no correlation between the Ni and Feisotopic compositions of these six samples (r2 = 0105) Ourresults suggest that the processes that fractionated the Niisotopes did not affect the Fe isotopes to the same degree orthat Ni and Fe isotopes were fractionated by differentprocesses on the IIIAB parent body

Isotopic Fractionation Between Kamacite and Taenite

Poitrasson et al (2005) and Horn et al (2006) showedthat Fe isotopes were fractionated between kamacite andtaenite in IAB irons These authors reported that kamacite(Fe-rich phase) is lighter in terms of its Fe isotopiccomposition relative to taenite (Ni-rich phase) BecauseToluca has been studied for both cooling rates (eg Wood1964 Hopfe and Goldstein 2001) and Fe isotopes it wasselected for an investigation of Fe and Ni isotopicfractionation associated with growth of the Widmanstaumlttenpattern The region of Toluca chosen for analyses is shownin Fig 10 Samples were collected along a traverse ofadjacent kamacite (spots 1 3 and 5) and taenite (spots 2and 4) The Ni isotopic composition of these five samples isshown in Fig 11 and Table 2 Nickel isotopes show clearlyresolvable fractionation between these two phases kamaciteis heavier relative to taenite by 036permil amuminus1 Bourdon

et al (2006) reported a similar fractionation of Ni isotopesbetween kamacite and taenite in iron meteorites The Feisotopic compositions of the five samples shown in Fig 10were also measured with the results shown in Fig 12 andTable 2 Unlike the isotopes of Ni the Fe isotopes do notshow a resolvable fractionation between kamacite andtaenite These results are in contrast to those of Poitrassonet al (2005) and Horn et al (2006) As discussed below thereason for this discrepancy is likely due to differences insampling

Dauphas and Rouxel (2006) showed that phase growth ina diffusion-limited regime can create diffusive kinetic isotopefractionation between two phases and control the expressionof equilibrium isotope fractionation between the co-existingphases More recently Dauphas (2007) developed anumerical simulation to model the formation of theWidmanstaumltten pattern during cooling of Fe-Ni alloy and tocalculate the associated fractionation of Ni and Fe isotopesDuring formation of the Widmanstaumltten pattern nucleation ofkamacite occurs in the solid state and its growth is controlledby diffusion (Wood 1964 Goldstein and Ogilvie 1965) It islikely that kinetic isotopic fractionation will accompany thisprocess because light isotopes diffuse faster than heavyisotopes of the same element (eg Richter et al 2003Roskosz et al 2006) The model predicts enrichment of lightNi isotopes in taenite relative to kamacite caused by diffusionduring cooling and formation of the Widmanstaumltten pattern inFe-Ni alloy These results are consistent with the measuredfractionations shown in Fig 11

Model results show that diffusion-driven kineticisotopic fractionation between two phases depends both on

Fig 9 Fe isotopic compositions (permil amuminus1) for the metal from IIIABiron meteorites relative to the IRMM-014 standard FFe = δ56542Circles = this study squares = Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006 hatched area = rangereported by Mullane et al (2005) Fig 10 SEM X-ray Ni map for the region of the Toluca IAB iron

chosen for microdrilling The numbers show the five spots sampledwith the microdrill Bright areas are Ni-rich Scale bar is 1000 μm

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2075

elemental concentration gradients between the two phasesand on the absolute elemental concentrations in each phaseSmaller concentration gradients lead to less fractionationand large absolute elemental concentrations act to dilutefractionation effects The absolute Fe concentrations inkamacite and taenite are much greater than the Niconcentrations in these two phases This factor can accountfor the difference in the magnitude of the measuredfractionations for Ni and Fe isotopes shown in Figs 11 and12 Furthermore inadvertent sampling of kamacite-taenitemixtures in Toluca by drilling through one phase and intothe other could explain why only resolvable effects in FNiwere measured in this study Because the FeNi ratios areunequal between kamacite and taenite the effects on FFeand FNi from mixing these two phases during sampling isnot linear Depending on the proportions of kamacite andtaenite involved in mixing FNi may change rapidly as FFechanges slowly or vice-versa relative to the isotopiccompositions of the pure phases For example mixingtaenite and kamacite in a 31 ratio changes FFe by 28whereas FNi changes only by 76 relative to pure taeniteSampling of pure kamacite and taenite-kamacite mixturescould account for the observed differences in FNi and lackof observed differences in FFe between kamacite and taenitein Toluca as shown in Figs 11 and 12

Although the observed isotopic fractionations in Tolucaare consistent with kinetic effects (especially given the role ofdiffusion in formation of the Widmanstaumltten pattern)equilibrium fractionation cannot be ruled out Modelingresults show that equilibrium and kinetic fractionation effectscould potentially be resolved if precise measurements of Niand Fe isotopes could be made at a higher spatial resolution(ie asymp20 μm) than has been achieved thus far

CONCLUSIONS

The Ni isotopes in bulk metal samples from 36meteorites follow a mass-dependent fractionation trendThese results are consistent with previous work (Moynieret al 2004 2005) and imply that Ni isotopes werefractionated from an initially isotopically homogeneousreservoir The Ni isotopic composition of metal fromchondrites and iron meteorites (magmatic and non-magmaticgroups) is similar to or heavier than SRM 986 whereaspallasite metal is similar to or lighter than SRM 986Measurements of additional pallasites are needed todetermine whether this difference is real or the result ofsampling bias

A suite of fourteen IIIAB irons shows fractionationover a considerable range The overall degree of massfractionation of Ni isotopes does not correlate with the Nicontent suggesting that core crystallization did not massfractionate Ni isotopes in a systematic way The variabilityobserved in the IIIAB irons may be due to the interaction ofimmiscible melts during the later stages of corecrystallization or due to the crystallization of metallic andnon-metallic phases Measurements of Ni isotopes inadjacent metal phosphide and sulfide phases in ironmeteorites are needed to determine the cause of the Ni

Fig 11 FNi values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Kamacite is heavier relative totaenite by 036permil amuminus1 Plotted errors are 2SE the externalprecision (2SD) is shown by the two dashed lines (0065permil amuminus1)

Fig 12 FFe values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Within uncertainty Fe isotopesshow no fractionation between kamacite and taenite

Table 2 The FNi and FFe values (permil amuminus1) for kamacite and taenite from the Toluca IAB iron

Sample FNi plusmn 2SE n FFe plusmn 2SE n

Spot 1 (kamacite) 0084 plusmn 0028 5 0080 plusmn 0041 4Spot 2 (taenite) minus0302 plusmn 0082 5 0078 plusmn 0041 4Spot 3 (kamacite) 0169 plusmn 0073 5 0072 plusmn 0041 4Spot 4 (taenite) minus0178 plusmn 0045 5 0086 plusmn 0041 4Spot 5 (kamacite) 0103 plusmn 0034 5 0051 plusmn 0041 4

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2069

cases the data are discussed using the 61Ni58Ni ratio becausethis ratio is free from possible effects due to decay of theshort-lived isotope 60Fe (t12 = 149 My) and also frompotential nucleosynthetic effects on the most neutron-rich Niisotopes (62 and 64) It is convenient to discuss the massfractionation effects in terms of permil per amu (permil amuminus1)calculated using

FNi = δ61583 (2)

The internal precisions of the Ni isotopic compositions ofindividual samples based on a number of repeatmeasurements are represented by the standard error of themean (2SE) Repeated analyses of an Aesar Ni solutionbracketed with the SRM 986 standard over a 27-month periodwere used to determine the external precision of our Niisotope ratio measurements Figure 1 shows the FNi valuesfrom these measurements each datum is the mean valuecalculated from 5 repeat measurements during a singleanalytical session These data show that the Aesar Ni solutionis mass fractionated relative to SRM 986 by asympminus01permil amuminus1and the external precision (2SD) is plusmn0065permil amuminus1 Theexternal precision based on the other Ni isotope ratios issimilar (δ60582 plusmn0063permil amuminus1 δ62584 plusmn0061permil amuminus1δ64586 plusmn0066permil amuminus1)

Isotopic fractionation during chemical processing ofsamples can potentially affect the isotopic composition andlead to inaccurate measurements To demonstrate that our Niseparation procedure does not introduce any analyticalartifact within uncertainties three aliquots each of the SRM986 isotopic standard were processed through both the two-column and the three-column Ni separation protocols (Cook

et al 2006) Figure 2 shows the FNi values for the sixprocessed aliquots of SRM 986 Within the uncertaintiesnone of the measured aliquots of the processed SRM 986 havean isotopic composition that differs from the unprocessedSRM 986 used as a bracketing standard Additionally noresolvable effects were observed on the other three measuredNi isotope ratios (δ6058 δ6258 δ6458) Hence these datademonstrate that our Ni separation chemistry does notintroduce any resolvable mass-dependent isotopicfractionation effects

Effects of Fe Zn and Signal Intensity

Corrections for isobaric interferences on 58Ni from 58Feand on 64Ni from 64Zn must be applied if Fe andor Zn arepresent in the sample solutions Furthermore Fe is thedominant element in meteoritic metal and an inefficientseparation of Ni from Fe could introduce matrix effects thatmay alter the isotopic measurements Thus aliquots of SRM986 were doped with varying amounts of Fe and Zn(4 aliquots each) to test our ability to correct Fe and Zninterferences and to quantify the possible matrix effects dueto the presence of Fe Figure 3 shows the interference-corrected FNi values for the SRM 986 aliquots doped withFe as well as an undoped aliquot These data show that theFe interference correction is effective and that no resolvablematrix effects are present for FeNi le 1 Additionally noresolvable effects were observed on the other threemeasured Ni isotope ratios (δ6058 δ6258 δ6458) due to thepresence of Fe Cook et al (2006) showed that theinterference correction for 64Zn on 64Ni is not effective forZnNi ge 001 when measuring non-mass-dependent effects

Fig 1 FNi values (permil amuminus1) for repeated analyses of an Aesar Nisolution over the course of a 27-month period Each datum representsthe mean of five repeat measurements performed during a singleanalysis session The individual error bars (2SE) are based on the fiverepeat measurements The external precision is the standarddeviation (2SD) based on all of the plotted data and is shown by thetwo gray lines (plusmn0065permil amuminus1)

Fig 2 FNi values (permil amuminus1) for six aliquots of the SRM 986 isotopicstandard processed through our Ni separation procedure (squares2-column protocol circles 3-column protocol) Plotted errors are2SE the external precision (2SD) is shown by the two dashed lines(plusmn0065permil amuminus1)

2070 D L Cook et al

This is also true for measurements of mass-dependenteffects However no resolvable effects were observed onthe other three measured Ni isotope ratios (δ6058 δ6158δ6258) due to the presence of Zn thus FNi was unaffected inthe range of ZnNi ratios investigated (10minus5 to 10minus1) Similarto our previous work (Cook et al 2006) all measuredsample solutions had FeNi and ZnNi ratios lt50 times 10minus2

and lt45 times 10minus4 respectively The interference correctionfor FNi was typically le001permil amuminus1

We also investigated whether a mismatch between the Nisignal intensities for standard and sample solutions couldaffect our mass-dependent isotopic measurements Sixaliquots of SRM 986 with Ni concentrations ranging from055 to 10 ppm were analyzed by bracketing with a 10 ppmSRM 986 solution Figure 4 shows the FNi values from themeasurements of these six aliquots of SRM 986 No clearlyresolvable effects were observed except when the samplesignal differed from the standard signal by 45 Similarresults were obtained for the other three measured Ni isotoperatios (δ6058 δ6258 δ6458) All measured sample solutions ofmeteoritic metal had signal intensities within 10 of thestandard signal intensity Thus all sample and standard signalintensities were matched adequately to provide accuratemeasurements

RESULTS AND DISCUSSION

In our previous study of non-mass-dependent effects inNi isotopes in the metal of chondrites pallasites and ironmeteorites (Cook et al 2006) we demonstrated that thereare small deficits in the mass-bias-corrected 60Ni58Ni ratiorelative to the terrestrial standard (ie ε60) in the metalfrom some primitive chondrites and possibly some iron

meteorites and that these deficits are most likely due to thepresence of live 60Fe (t12 = 149 My) at the time of FeNifractionation in these metal samples However there are noother clearly resolvable non-mass-dependent effects in anyof the other measured Ni isotope ratios As such in Table 1we present the results of our measurements of the 61Ni58Ni62Ni58Ni and 64Ni58Ni ratios relative to the SRM 986standard in permil (ie 6158 6258 and 6458) For eachof the samples the Ni isotope fractionation in terms ofpermil amuminus1 is the same within uncertainties regardless ofwhich ratio (ie 61Ni58Ni 62Ni58Ni or 64Ni58Ni) isconsidered Thus the data are discussed primarily in termsof FNi (Equation 2)

We note that while we have performed analyses of asingle piece of bulk metal from most of the meteoritesstudied we analyzed two separate pieces (identified as 1and 2 in Table 1) of three magmatic IIIAB iron meteorites(Bella Roca Casas Grandes and Henbury) Sample 1 ineach of these three cases was a small piece (lt50 mg) thatwas previously analyzed primarily to investigate non-mass-dependent effects in Ni isotopes (Cook et al 2006) Weadditionally selected a second larger piece (gt300 mg) ofeach of these three samples with the goal of obtaining arepresentative sample of the bulk metal for ourinvestigation of the mass-dependent fractionation of Niisotopes in the IIIAB iron meteorites In the case ofHenbury the Ni isotope compositions of both pieces agreewithin error This is not the case for the two pieces of BellaRoca and Casas Grandes It is likely that the smaller piecesof these two meteorites do not provide representative Niisotopic compositions of the bulk metal due to thecoarseness of the Widmanstaumltten pattern Our results for

Fig 3 FNi values (permil amuminus1) versus the FeNi elemental ratio Dataare from analyses of four aliquots of SRM 986 doped with variousconcentrations of Fe and one undoped aliquot (FeNi asymp 10minus4) Plottederrors are 2SE the external precision (2SD) is shown by the twodashed lines (plusmn0065permil amuminus1)

Fig 4 FNi values (permil amuminus1) versus the ratio of the sample tostandard signal intensity Data are from analyses of six aliquots ofSRM 986 with Ni concentrations ranging from 055 to 10 ppmbracketed by a 10 ppm SRM 986 solution Plotted errors are 2SE theexternal precision (2SD) is shown by the two dashed lines(plusmn0065permil amuminus1)

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2071

individual kamacite and taenite in Toluca (discussed below)show that Ni isotopes are fractionated between these twophases and can lead to mass-dependent Ni isotopicvariations in iron meteorites at the millimeter scaleTherefore small metal samples of iron meteorites (andpallasites) characterized by a coarse Widmanstaumltten patternmay be skewed toward the isotopic composition ofkamacite or taenite As such in the discussion on the Niisotope fractionation in bulk samples of IIIAB ironmeteorites only the larger pieces (samples 2 in Table 1) ofeach of these three meteorites are considered

Meteoritic Metal Bulk Samples

The Ni isotopic compositions of the bulk metal from the36 meteorites analyzed (Table 1) are shown in a three-isotopeplot (Fig 5) Within error all samples in Fig 5 plot on a 11line which is consistent with mass-dependent isotopicfractionation The overall range of fractionation is041permil amuminus1 The observed mass-dependent fractionationand the overall range are both consistent with the resultspreviously reported by Moynier et al (2004 2005) for metalfrom iron meteorites pallasites chondrites and mesosiderites

Table 1 The Ni isotopic composition of meteoritic bulk metal (permil) relative to the SRM 986 standard Sample Group δ6158 plusmn 2SE δ6258 plusmn 2SE δ6458 plusmn 2SE n

Renazzoa CR 047 plusmn 012 058 plusmn 010 086 plusmn 014 14Gujbaa CBa 060 plusmn 012 084 plusmn 012 121 plusmn 019 13Semarkonaa LL 30 018 plusmn 012 009 plusmn 017 016 plusmn 029 10Bishunpura LL 31 043 plusmn 014 053 plusmn 018 071 plusmn 026 5Forest Valea H4 053 plusmn 026 068 plusmn 020 103 plusmn 045 5Indarcha EH4 026 plusmn 021 030 plusmn 024 049 plusmn 045 5

Eagle Stationa PES 015 plusmn 005 013 plusmn 007 039 plusmn 014 5Albina PMG 002 plusmn 010 013 plusmn 009 025 plusmn 013 9Brenhama PMG minus025 plusmn 016 minus034 plusmn 007 minus034 plusmn 036 5Molonga PMG minus019 plusmn 008 minus027 plusmn 011 minus041 plusmn 029 5

Coahuilaa IIAB 021 plusmn 024 032 plusmn 021 059 plusmn 061 5Santa Luziaa IIAB 059 plusmn 006 076 plusmn 004 126 plusmn 013 10Carboa IID 097 plusmn 041 130 plusmn 045 190 plusmn 086 5Augustinovkab IIIAB minus005 plusmn 022 minus012 plusmn 027 minus020 plusmn 061 5Avocab IIIAB 022 plusmn 017 028 plusmn 024 035 plusmn 026 5Bald Eagleb IIIAB 036 plusmn 026 054 plusmn 018 085 plusmn 051 5Bear Creek IIIAB 015 plusmn 022 027 plusmn 023 061 plusmn 039 5Bella Roca 1a IIIAB 008 plusmn 016 010 plusmn 012 005 plusmn 014 8Bella Roca 2b 092 plusmn 025 117 plusmn 023 177 plusmn 045 5Carthage IIIAB 055 plusmn 006 066 plusmn 007 108 plusmn 010 5Casas Grandes 1a IIIAB 062 plusmn 003 082 plusmn 003 135 plusmn 009 14Casas Grandes 2 029 plusmn 012 037 plusmn 004 061 plusmn 006 5Grant IIIAB 013 plusmn 024 010 plusmn 020 023 plusmn 039 5Henbury 1a IIIAB 033 plusmn 004 042 plusmn 004 061 plusmn 007 14Henbury 2b 031 plusmn 022 036 plusmn 016 045 plusmn 010 5Nova Petropolisb IIIAB 024 plusmn 013 026 plusmn 011 030 plusmn 008 5Orange River Ironb IIIAB 023 plusmn 031 025 plusmn 041 043 plusmn 094 4Seneca Falls IIIAB 037 plusmn 011 050 plusmn 014 081 plusmn 015 5Thunda IIIAB 042 plusmn 016 046 plusmn 005 073 plusmn 012 5Wellandb IIIAB 072 plusmn 058 089 plusmn 071 140 plusmn 119 5Gibeona IVA 028 plusmn 021 037 plusmn 009 078 plusmn 025 5Yanhuitlana IVA 064 plusmn 018 075 plusmn 016 112 plusmn 045 5Cape of Good Hopea IVB 036 plusmn 033 043 plusmn 014 085 plusmn 026 5Hobaa IVB 045 plusmn 011 064 plusmn 012 094 plusmn 018 13Tlacotepeca IVB 014 plusmn 035 022 plusmn 048 033 plusmn 069 9

Canyon Diabloa IAB 039 plusmn 006 051 plusmn 005 095 plusmn 012 19Tolucaa IAB 044 plusmn 013 062 plusmn 009 094 plusmn 010 5Daytona IAB minus015 plusmn 010 minus014 plusmn 005 001 plusmn 035 5Mundrabillaa IAB 026 plusmn 007 036 plusmn 006 055 plusmn 011 14

aSamples previously investigated for non-mass-dependent effects in Ni isotopes (Cook et al 2006)bSample solutions provided by E Mullane

n = Number of repeat measurements

2072 D L Cook et al

The FNi values for all 36 bulk metal samples are plottedin Fig 6 according to meteorite class Figure 6 shows that 1)the metal from chondrites and from iron meteorites(magmatic and non-magmatic) is typically similar to orheavier than SRM 986 and 2) metal from pallasites is similarto or lighter than SRM 986 The mean and two standarddeviations (2SD) of the FNi values for each class are asfollows chondrites (014 plusmn 011permil amuminus1) iron meteorites(012 plusmn 017permil amuminus1) and pallasites (minus002 plusmn 012permil amuminus1)The Ni isotopic compositions for each meteorite class werecompared to one another using t-tests These comparisonsshow that there is no significant difference between thechondrites and the iron meteorites (t = 0417) However theNi isotopic composition for pallasite metal is significantlydifferent from both chondrite metal (t = 4325) and from ironmeteorites (t = 3200) It is possible that the mean Ni isotopiccomposition of pallasitic metal is biased due to the smallnumber of samples measured However if this is not the casethere are several possibilities for explaining the apparentlydistinct Ni isotopic composition of pallasite metal Ifpallasites form at the core-mantle boundary of planetesimals(eg Scott 1977) the Ni isotopic composition of pallasitemetal may reflect isotopic fractionation between the metalliccore and the overlying mantle silicates Zhu et al (2002) andPoitrasson et al (2005) reported that Fe isotopes arefractionated between metal and olivine in pallasites and theseeffects range from 002 to 011permil amuminus1 Recent experimentsindicate that Fe isotopes can undergo fractionation up toasymp23permil amuminus1 between metallic and silicate melts due tokinetic effects (Roskosz et al 2006) and are fractionated by

asymp017permil amuminus1 between sulfide and silicate melts attemperatures between 840 and 1000 degC due to equilibriumeffects (Schuessler et al 2007) Currently a lack of Niisotopic data for non-metal phases from pallasites (egsilicates and oxides) precludes an assessment of planetarydifferentiation as a potential explanation for the apparentlydistinct Ni isotopic composition of pallasite metalAlternatively the Ni isotopic compositions of pallasite metalmay reflect differences at the parent-body scale that wereestablished prior to accretion rather than reflectingfractionation effects from differentiation and core formationon their parent bodies Finally it is also possible that in thecase of some of the smaller samples (ie lt50 mg) frommeteorites with metal that is compositionally heterogeneouson the millimeter-scale (eg the four pallasites) themeasured Ni isotope compositions may not be representativeof the bulk metal in these meteorites Analyses of additionaland larger samples would help to ascertain if pallasitic metaldoes indeed have a mean Ni isotopic composition that differssignificantly from metal in other meteorite classes

IIIAB Iron Meteorites

Nickel IsotopesWith over two hundred members the IIIABs constitute

the largest group of magmatic iron meteorites Thesemeteorites cover a wide compositional range and the originalIIIAB core is likely well-represented by the large number ofsamples (Chabot and Haack 2006) Mullane et al (2005)reported that Fe isotopes from a suite of 33 IIIAB irons weremass fractionated over a large range of 049permil amuminus1however the Fe isotopic compositions did not correlate with

Fig 5 A three-isotope plot of the data from the analyses of bulkmetal from 36 meteorites The data follow the trend expected formass-dependent fractionation a line with a slope equal to 1 is shownfor reference The overall range of fractionation is 041permil amuminus1Plotted errors are 2SE Filled triangles = chondrites squares =pallasites open triangles = magmatic irons circles = non-magmaticirons

Fig 6 FNi values (permil amuminus1) for the bulk metal from 36 meteoritesPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Filled triangles = chondritessquares = pallasites open triangles = magmatic irons circles = non-magmatic irons

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2073

the Ni contents which are a proxy for the degree ofcrystallization of the metallic core (eg Scott 1972) Hencefourteen IIIAB irons were selected to examine the degree ofmass-dependent fractionation of Ni isotopes possibly due tofractional crystallization of the core and to determine whetherthe fractionation effects on Fe and Ni isotopes werecorrelated Figure 7 shows the individual members of theIIIAB group in terms of their Ir and Ni contents including thefourteen samples chosen for Ni isotopic analyses Sampleswere chosen to represent the full compositional range of theIIIAB irons Figure 8 shows the FNi values for the fourteenIIIAB irons The Ni isotopes show fractionation over aconsiderable range from minus002 to 031permil amuminus1 The earliercrystallized low-Ni IIIA irons and the later crystallized high-Ni IIIB irons have similar mean compositions of01permil amuminus1 Furthermore the degree of mass fractionationof Ni isotopes does not correlate with the Ni content (r2 =0003) The data in Fig 8 suggest that core crystallization didnot mass fractionate Ni isotopes in a systematic way

The isotopic composition of Ni in different IIIAB ironmeteorites likely reflects a combination of events such asfractionation between 1) liquid metal and silicate during coresegregation 2) immiscible melts in the molten core and 3)various phases during core solidification The later stages ofcore crystallization of iron meteorites were likely affected byliquid immiscibility resulting in the formation ofcompositionally distinct liquids due to the presence of S and Pin the molten core (Ulff-Moslashller 1998 Chabot and Drake2000) Nickel isotopes may have been fractionated as a resultof the interaction of a metallic melt with a S andor P-richmelt during core crystallization Furthermore Fe isotopes arefractionated between metal phosphide (ie schreibersite)

and sulfide (ie troilite) in pallasites (Poitrasson et al 2005Weyer et al 2005) and between metal and sulfide in ironmeteorites (Williams et al 2006) Crystallization followed byslow subsolidus cooling of metallic and non-metallic phasesin IIIAB irons may have fractionated Ni isotopes in a similarmanner This last mechanism is particularly appealingbecause the Ni concentrations are quite different betweenmetal troilite and schreibersite in iron meteorites (Buchwald1977 Jochum et al 1980) Diffusion of Ni along theseconcentration gradients during crystallization and cooling ofthe various phases would likely be accompanied by kineticisotopic fractionation The proximity of non-metallic phasesrelative to the metal could produce metal in iron meteoriteswith a range of FNi values and potentially account for theobserved variation among the IIIAB irons shown in Fig 8Measurements of Ni isotopes in adjacent metallic and non-metallic phases in iron meteorites could be used to test thishypothesis

Iron IsotopesMullane et al (2005) reported Fe isotope fractionation

over a large range (049permil amuminus1) for IIIAB irons in whichsamples were both heavier and lighter than the IRMM-014standard These data are inconsistent with other studies of Feisotopes in IIIAB iron meteorite metal in which all sampleshad Fe isotopic compositions similar to or slightly heavierthan IRMM-014 and spanned a narrow range of007permil amuminus1 (Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006) We analyzedthe Fe isotopes in six of the IIIAB irons with low and high Niabundances in which we also measured Ni isotopes The56Fe54Fe ratios in permil amuminus1 (or FFe) for these six samples are

Fig 7 A plot of Ir versus Ni abundances for members of the IIIABiron meteorite group The fourteen samples selected for Ni isotopicanalyses are shown by the black circles Data are from Scott et al(1973) Scott and Wasson (1976) Kracher et al (1980) Malvin et al(1984) and Wasson et al (1989 1998)

Fig 8 FNi values (permil amuminus1) versus Ni abundances for IIIAB ironsPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Nickel abundance data are fromScott et al (1973) Scott and Wasson (1976) and Malvin et al(1984)

2074 D L Cook et al

shown in Fig 9 along with Fe isotope data for IIIABs fromother studies As shown in Fig 9 our measurements do notshow the large range reported by Mullane et al (2005) All sixsamples have Fe isotopic compositions similar to orsomewhat heavier than IRMM-014 and span an overall rangeof only 007permil amuminus1 Both of these features are consistentwith the measurements of Zhu et al (2001) Kehm et al(2003) Schoenberg et al (2006) and Williams et al (2006)Furthermore there is no correlation between the Ni and Feisotopic compositions of these six samples (r2 = 0105) Ourresults suggest that the processes that fractionated the Niisotopes did not affect the Fe isotopes to the same degree orthat Ni and Fe isotopes were fractionated by differentprocesses on the IIIAB parent body

Isotopic Fractionation Between Kamacite and Taenite

Poitrasson et al (2005) and Horn et al (2006) showedthat Fe isotopes were fractionated between kamacite andtaenite in IAB irons These authors reported that kamacite(Fe-rich phase) is lighter in terms of its Fe isotopiccomposition relative to taenite (Ni-rich phase) BecauseToluca has been studied for both cooling rates (eg Wood1964 Hopfe and Goldstein 2001) and Fe isotopes it wasselected for an investigation of Fe and Ni isotopicfractionation associated with growth of the Widmanstaumlttenpattern The region of Toluca chosen for analyses is shownin Fig 10 Samples were collected along a traverse ofadjacent kamacite (spots 1 3 and 5) and taenite (spots 2and 4) The Ni isotopic composition of these five samples isshown in Fig 11 and Table 2 Nickel isotopes show clearlyresolvable fractionation between these two phases kamaciteis heavier relative to taenite by 036permil amuminus1 Bourdon

et al (2006) reported a similar fractionation of Ni isotopesbetween kamacite and taenite in iron meteorites The Feisotopic compositions of the five samples shown in Fig 10were also measured with the results shown in Fig 12 andTable 2 Unlike the isotopes of Ni the Fe isotopes do notshow a resolvable fractionation between kamacite andtaenite These results are in contrast to those of Poitrassonet al (2005) and Horn et al (2006) As discussed below thereason for this discrepancy is likely due to differences insampling

Dauphas and Rouxel (2006) showed that phase growth ina diffusion-limited regime can create diffusive kinetic isotopefractionation between two phases and control the expressionof equilibrium isotope fractionation between the co-existingphases More recently Dauphas (2007) developed anumerical simulation to model the formation of theWidmanstaumltten pattern during cooling of Fe-Ni alloy and tocalculate the associated fractionation of Ni and Fe isotopesDuring formation of the Widmanstaumltten pattern nucleation ofkamacite occurs in the solid state and its growth is controlledby diffusion (Wood 1964 Goldstein and Ogilvie 1965) It islikely that kinetic isotopic fractionation will accompany thisprocess because light isotopes diffuse faster than heavyisotopes of the same element (eg Richter et al 2003Roskosz et al 2006) The model predicts enrichment of lightNi isotopes in taenite relative to kamacite caused by diffusionduring cooling and formation of the Widmanstaumltten pattern inFe-Ni alloy These results are consistent with the measuredfractionations shown in Fig 11

Model results show that diffusion-driven kineticisotopic fractionation between two phases depends both on

Fig 9 Fe isotopic compositions (permil amuminus1) for the metal from IIIABiron meteorites relative to the IRMM-014 standard FFe = δ56542Circles = this study squares = Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006 hatched area = rangereported by Mullane et al (2005) Fig 10 SEM X-ray Ni map for the region of the Toluca IAB iron

chosen for microdrilling The numbers show the five spots sampledwith the microdrill Bright areas are Ni-rich Scale bar is 1000 μm

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2075

elemental concentration gradients between the two phasesand on the absolute elemental concentrations in each phaseSmaller concentration gradients lead to less fractionationand large absolute elemental concentrations act to dilutefractionation effects The absolute Fe concentrations inkamacite and taenite are much greater than the Niconcentrations in these two phases This factor can accountfor the difference in the magnitude of the measuredfractionations for Ni and Fe isotopes shown in Figs 11 and12 Furthermore inadvertent sampling of kamacite-taenitemixtures in Toluca by drilling through one phase and intothe other could explain why only resolvable effects in FNiwere measured in this study Because the FeNi ratios areunequal between kamacite and taenite the effects on FFeand FNi from mixing these two phases during sampling isnot linear Depending on the proportions of kamacite andtaenite involved in mixing FNi may change rapidly as FFechanges slowly or vice-versa relative to the isotopiccompositions of the pure phases For example mixingtaenite and kamacite in a 31 ratio changes FFe by 28whereas FNi changes only by 76 relative to pure taeniteSampling of pure kamacite and taenite-kamacite mixturescould account for the observed differences in FNi and lackof observed differences in FFe between kamacite and taenitein Toluca as shown in Figs 11 and 12

Although the observed isotopic fractionations in Tolucaare consistent with kinetic effects (especially given the role ofdiffusion in formation of the Widmanstaumltten pattern)equilibrium fractionation cannot be ruled out Modelingresults show that equilibrium and kinetic fractionation effectscould potentially be resolved if precise measurements of Niand Fe isotopes could be made at a higher spatial resolution(ie asymp20 μm) than has been achieved thus far

CONCLUSIONS

The Ni isotopes in bulk metal samples from 36meteorites follow a mass-dependent fractionation trendThese results are consistent with previous work (Moynieret al 2004 2005) and imply that Ni isotopes werefractionated from an initially isotopically homogeneousreservoir The Ni isotopic composition of metal fromchondrites and iron meteorites (magmatic and non-magmaticgroups) is similar to or heavier than SRM 986 whereaspallasite metal is similar to or lighter than SRM 986Measurements of additional pallasites are needed todetermine whether this difference is real or the result ofsampling bias

A suite of fourteen IIIAB irons shows fractionationover a considerable range The overall degree of massfractionation of Ni isotopes does not correlate with the Nicontent suggesting that core crystallization did not massfractionate Ni isotopes in a systematic way The variabilityobserved in the IIIAB irons may be due to the interaction ofimmiscible melts during the later stages of corecrystallization or due to the crystallization of metallic andnon-metallic phases Measurements of Ni isotopes inadjacent metal phosphide and sulfide phases in ironmeteorites are needed to determine the cause of the Ni

Fig 11 FNi values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Kamacite is heavier relative totaenite by 036permil amuminus1 Plotted errors are 2SE the externalprecision (2SD) is shown by the two dashed lines (0065permil amuminus1)

Fig 12 FFe values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Within uncertainty Fe isotopesshow no fractionation between kamacite and taenite

Table 2 The FNi and FFe values (permil amuminus1) for kamacite and taenite from the Toluca IAB iron

Sample FNi plusmn 2SE n FFe plusmn 2SE n

Spot 1 (kamacite) 0084 plusmn 0028 5 0080 plusmn 0041 4Spot 2 (taenite) minus0302 plusmn 0082 5 0078 plusmn 0041 4Spot 3 (kamacite) 0169 plusmn 0073 5 0072 plusmn 0041 4Spot 4 (taenite) minus0178 plusmn 0045 5 0086 plusmn 0041 4Spot 5 (kamacite) 0103 plusmn 0034 5 0051 plusmn 0041 4

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

2070 D L Cook et al

This is also true for measurements of mass-dependenteffects However no resolvable effects were observed onthe other three measured Ni isotope ratios (δ6058 δ6158δ6258) due to the presence of Zn thus FNi was unaffected inthe range of ZnNi ratios investigated (10minus5 to 10minus1) Similarto our previous work (Cook et al 2006) all measuredsample solutions had FeNi and ZnNi ratios lt50 times 10minus2

and lt45 times 10minus4 respectively The interference correctionfor FNi was typically le001permil amuminus1

We also investigated whether a mismatch between the Nisignal intensities for standard and sample solutions couldaffect our mass-dependent isotopic measurements Sixaliquots of SRM 986 with Ni concentrations ranging from055 to 10 ppm were analyzed by bracketing with a 10 ppmSRM 986 solution Figure 4 shows the FNi values from themeasurements of these six aliquots of SRM 986 No clearlyresolvable effects were observed except when the samplesignal differed from the standard signal by 45 Similarresults were obtained for the other three measured Ni isotoperatios (δ6058 δ6258 δ6458) All measured sample solutions ofmeteoritic metal had signal intensities within 10 of thestandard signal intensity Thus all sample and standard signalintensities were matched adequately to provide accuratemeasurements

RESULTS AND DISCUSSION

In our previous study of non-mass-dependent effects inNi isotopes in the metal of chondrites pallasites and ironmeteorites (Cook et al 2006) we demonstrated that thereare small deficits in the mass-bias-corrected 60Ni58Ni ratiorelative to the terrestrial standard (ie ε60) in the metalfrom some primitive chondrites and possibly some iron

meteorites and that these deficits are most likely due to thepresence of live 60Fe (t12 = 149 My) at the time of FeNifractionation in these metal samples However there are noother clearly resolvable non-mass-dependent effects in anyof the other measured Ni isotope ratios As such in Table 1we present the results of our measurements of the 61Ni58Ni62Ni58Ni and 64Ni58Ni ratios relative to the SRM 986standard in permil (ie 6158 6258 and 6458) For eachof the samples the Ni isotope fractionation in terms ofpermil amuminus1 is the same within uncertainties regardless ofwhich ratio (ie 61Ni58Ni 62Ni58Ni or 64Ni58Ni) isconsidered Thus the data are discussed primarily in termsof FNi (Equation 2)

We note that while we have performed analyses of asingle piece of bulk metal from most of the meteoritesstudied we analyzed two separate pieces (identified as 1and 2 in Table 1) of three magmatic IIIAB iron meteorites(Bella Roca Casas Grandes and Henbury) Sample 1 ineach of these three cases was a small piece (lt50 mg) thatwas previously analyzed primarily to investigate non-mass-dependent effects in Ni isotopes (Cook et al 2006) Weadditionally selected a second larger piece (gt300 mg) ofeach of these three samples with the goal of obtaining arepresentative sample of the bulk metal for ourinvestigation of the mass-dependent fractionation of Niisotopes in the IIIAB iron meteorites In the case ofHenbury the Ni isotope compositions of both pieces agreewithin error This is not the case for the two pieces of BellaRoca and Casas Grandes It is likely that the smaller piecesof these two meteorites do not provide representative Niisotopic compositions of the bulk metal due to thecoarseness of the Widmanstaumltten pattern Our results for

Fig 3 FNi values (permil amuminus1) versus the FeNi elemental ratio Dataare from analyses of four aliquots of SRM 986 doped with variousconcentrations of Fe and one undoped aliquot (FeNi asymp 10minus4) Plottederrors are 2SE the external precision (2SD) is shown by the twodashed lines (plusmn0065permil amuminus1)

Fig 4 FNi values (permil amuminus1) versus the ratio of the sample tostandard signal intensity Data are from analyses of six aliquots ofSRM 986 with Ni concentrations ranging from 055 to 10 ppmbracketed by a 10 ppm SRM 986 solution Plotted errors are 2SE theexternal precision (2SD) is shown by the two dashed lines(plusmn0065permil amuminus1)

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2071

individual kamacite and taenite in Toluca (discussed below)show that Ni isotopes are fractionated between these twophases and can lead to mass-dependent Ni isotopicvariations in iron meteorites at the millimeter scaleTherefore small metal samples of iron meteorites (andpallasites) characterized by a coarse Widmanstaumltten patternmay be skewed toward the isotopic composition ofkamacite or taenite As such in the discussion on the Niisotope fractionation in bulk samples of IIIAB ironmeteorites only the larger pieces (samples 2 in Table 1) ofeach of these three meteorites are considered

Meteoritic Metal Bulk Samples

The Ni isotopic compositions of the bulk metal from the36 meteorites analyzed (Table 1) are shown in a three-isotopeplot (Fig 5) Within error all samples in Fig 5 plot on a 11line which is consistent with mass-dependent isotopicfractionation The overall range of fractionation is041permil amuminus1 The observed mass-dependent fractionationand the overall range are both consistent with the resultspreviously reported by Moynier et al (2004 2005) for metalfrom iron meteorites pallasites chondrites and mesosiderites

Table 1 The Ni isotopic composition of meteoritic bulk metal (permil) relative to the SRM 986 standard Sample Group δ6158 plusmn 2SE δ6258 plusmn 2SE δ6458 plusmn 2SE n

Renazzoa CR 047 plusmn 012 058 plusmn 010 086 plusmn 014 14Gujbaa CBa 060 plusmn 012 084 plusmn 012 121 plusmn 019 13Semarkonaa LL 30 018 plusmn 012 009 plusmn 017 016 plusmn 029 10Bishunpura LL 31 043 plusmn 014 053 plusmn 018 071 plusmn 026 5Forest Valea H4 053 plusmn 026 068 plusmn 020 103 plusmn 045 5Indarcha EH4 026 plusmn 021 030 plusmn 024 049 plusmn 045 5

Eagle Stationa PES 015 plusmn 005 013 plusmn 007 039 plusmn 014 5Albina PMG 002 plusmn 010 013 plusmn 009 025 plusmn 013 9Brenhama PMG minus025 plusmn 016 minus034 plusmn 007 minus034 plusmn 036 5Molonga PMG minus019 plusmn 008 minus027 plusmn 011 minus041 plusmn 029 5

Coahuilaa IIAB 021 plusmn 024 032 plusmn 021 059 plusmn 061 5Santa Luziaa IIAB 059 plusmn 006 076 plusmn 004 126 plusmn 013 10Carboa IID 097 plusmn 041 130 plusmn 045 190 plusmn 086 5Augustinovkab IIIAB minus005 plusmn 022 minus012 plusmn 027 minus020 plusmn 061 5Avocab IIIAB 022 plusmn 017 028 plusmn 024 035 plusmn 026 5Bald Eagleb IIIAB 036 plusmn 026 054 plusmn 018 085 plusmn 051 5Bear Creek IIIAB 015 plusmn 022 027 plusmn 023 061 plusmn 039 5Bella Roca 1a IIIAB 008 plusmn 016 010 plusmn 012 005 plusmn 014 8Bella Roca 2b 092 plusmn 025 117 plusmn 023 177 plusmn 045 5Carthage IIIAB 055 plusmn 006 066 plusmn 007 108 plusmn 010 5Casas Grandes 1a IIIAB 062 plusmn 003 082 plusmn 003 135 plusmn 009 14Casas Grandes 2 029 plusmn 012 037 plusmn 004 061 plusmn 006 5Grant IIIAB 013 plusmn 024 010 plusmn 020 023 plusmn 039 5Henbury 1a IIIAB 033 plusmn 004 042 plusmn 004 061 plusmn 007 14Henbury 2b 031 plusmn 022 036 plusmn 016 045 plusmn 010 5Nova Petropolisb IIIAB 024 plusmn 013 026 plusmn 011 030 plusmn 008 5Orange River Ironb IIIAB 023 plusmn 031 025 plusmn 041 043 plusmn 094 4Seneca Falls IIIAB 037 plusmn 011 050 plusmn 014 081 plusmn 015 5Thunda IIIAB 042 plusmn 016 046 plusmn 005 073 plusmn 012 5Wellandb IIIAB 072 plusmn 058 089 plusmn 071 140 plusmn 119 5Gibeona IVA 028 plusmn 021 037 plusmn 009 078 plusmn 025 5Yanhuitlana IVA 064 plusmn 018 075 plusmn 016 112 plusmn 045 5Cape of Good Hopea IVB 036 plusmn 033 043 plusmn 014 085 plusmn 026 5Hobaa IVB 045 plusmn 011 064 plusmn 012 094 plusmn 018 13Tlacotepeca IVB 014 plusmn 035 022 plusmn 048 033 plusmn 069 9

Canyon Diabloa IAB 039 plusmn 006 051 plusmn 005 095 plusmn 012 19Tolucaa IAB 044 plusmn 013 062 plusmn 009 094 plusmn 010 5Daytona IAB minus015 plusmn 010 minus014 plusmn 005 001 plusmn 035 5Mundrabillaa IAB 026 plusmn 007 036 plusmn 006 055 plusmn 011 14

aSamples previously investigated for non-mass-dependent effects in Ni isotopes (Cook et al 2006)bSample solutions provided by E Mullane

n = Number of repeat measurements

2072 D L Cook et al

The FNi values for all 36 bulk metal samples are plottedin Fig 6 according to meteorite class Figure 6 shows that 1)the metal from chondrites and from iron meteorites(magmatic and non-magmatic) is typically similar to orheavier than SRM 986 and 2) metal from pallasites is similarto or lighter than SRM 986 The mean and two standarddeviations (2SD) of the FNi values for each class are asfollows chondrites (014 plusmn 011permil amuminus1) iron meteorites(012 plusmn 017permil amuminus1) and pallasites (minus002 plusmn 012permil amuminus1)The Ni isotopic compositions for each meteorite class werecompared to one another using t-tests These comparisonsshow that there is no significant difference between thechondrites and the iron meteorites (t = 0417) However theNi isotopic composition for pallasite metal is significantlydifferent from both chondrite metal (t = 4325) and from ironmeteorites (t = 3200) It is possible that the mean Ni isotopiccomposition of pallasitic metal is biased due to the smallnumber of samples measured However if this is not the casethere are several possibilities for explaining the apparentlydistinct Ni isotopic composition of pallasite metal Ifpallasites form at the core-mantle boundary of planetesimals(eg Scott 1977) the Ni isotopic composition of pallasitemetal may reflect isotopic fractionation between the metalliccore and the overlying mantle silicates Zhu et al (2002) andPoitrasson et al (2005) reported that Fe isotopes arefractionated between metal and olivine in pallasites and theseeffects range from 002 to 011permil amuminus1 Recent experimentsindicate that Fe isotopes can undergo fractionation up toasymp23permil amuminus1 between metallic and silicate melts due tokinetic effects (Roskosz et al 2006) and are fractionated by

asymp017permil amuminus1 between sulfide and silicate melts attemperatures between 840 and 1000 degC due to equilibriumeffects (Schuessler et al 2007) Currently a lack of Niisotopic data for non-metal phases from pallasites (egsilicates and oxides) precludes an assessment of planetarydifferentiation as a potential explanation for the apparentlydistinct Ni isotopic composition of pallasite metalAlternatively the Ni isotopic compositions of pallasite metalmay reflect differences at the parent-body scale that wereestablished prior to accretion rather than reflectingfractionation effects from differentiation and core formationon their parent bodies Finally it is also possible that in thecase of some of the smaller samples (ie lt50 mg) frommeteorites with metal that is compositionally heterogeneouson the millimeter-scale (eg the four pallasites) themeasured Ni isotope compositions may not be representativeof the bulk metal in these meteorites Analyses of additionaland larger samples would help to ascertain if pallasitic metaldoes indeed have a mean Ni isotopic composition that differssignificantly from metal in other meteorite classes

IIIAB Iron Meteorites

Nickel IsotopesWith over two hundred members the IIIABs constitute

the largest group of magmatic iron meteorites Thesemeteorites cover a wide compositional range and the originalIIIAB core is likely well-represented by the large number ofsamples (Chabot and Haack 2006) Mullane et al (2005)reported that Fe isotopes from a suite of 33 IIIAB irons weremass fractionated over a large range of 049permil amuminus1however the Fe isotopic compositions did not correlate with

Fig 5 A three-isotope plot of the data from the analyses of bulkmetal from 36 meteorites The data follow the trend expected formass-dependent fractionation a line with a slope equal to 1 is shownfor reference The overall range of fractionation is 041permil amuminus1Plotted errors are 2SE Filled triangles = chondrites squares =pallasites open triangles = magmatic irons circles = non-magmaticirons

Fig 6 FNi values (permil amuminus1) for the bulk metal from 36 meteoritesPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Filled triangles = chondritessquares = pallasites open triangles = magmatic irons circles = non-magmatic irons

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2073

the Ni contents which are a proxy for the degree ofcrystallization of the metallic core (eg Scott 1972) Hencefourteen IIIAB irons were selected to examine the degree ofmass-dependent fractionation of Ni isotopes possibly due tofractional crystallization of the core and to determine whetherthe fractionation effects on Fe and Ni isotopes werecorrelated Figure 7 shows the individual members of theIIIAB group in terms of their Ir and Ni contents including thefourteen samples chosen for Ni isotopic analyses Sampleswere chosen to represent the full compositional range of theIIIAB irons Figure 8 shows the FNi values for the fourteenIIIAB irons The Ni isotopes show fractionation over aconsiderable range from minus002 to 031permil amuminus1 The earliercrystallized low-Ni IIIA irons and the later crystallized high-Ni IIIB irons have similar mean compositions of01permil amuminus1 Furthermore the degree of mass fractionationof Ni isotopes does not correlate with the Ni content (r2 =0003) The data in Fig 8 suggest that core crystallization didnot mass fractionate Ni isotopes in a systematic way

The isotopic composition of Ni in different IIIAB ironmeteorites likely reflects a combination of events such asfractionation between 1) liquid metal and silicate during coresegregation 2) immiscible melts in the molten core and 3)various phases during core solidification The later stages ofcore crystallization of iron meteorites were likely affected byliquid immiscibility resulting in the formation ofcompositionally distinct liquids due to the presence of S and Pin the molten core (Ulff-Moslashller 1998 Chabot and Drake2000) Nickel isotopes may have been fractionated as a resultof the interaction of a metallic melt with a S andor P-richmelt during core crystallization Furthermore Fe isotopes arefractionated between metal phosphide (ie schreibersite)

and sulfide (ie troilite) in pallasites (Poitrasson et al 2005Weyer et al 2005) and between metal and sulfide in ironmeteorites (Williams et al 2006) Crystallization followed byslow subsolidus cooling of metallic and non-metallic phasesin IIIAB irons may have fractionated Ni isotopes in a similarmanner This last mechanism is particularly appealingbecause the Ni concentrations are quite different betweenmetal troilite and schreibersite in iron meteorites (Buchwald1977 Jochum et al 1980) Diffusion of Ni along theseconcentration gradients during crystallization and cooling ofthe various phases would likely be accompanied by kineticisotopic fractionation The proximity of non-metallic phasesrelative to the metal could produce metal in iron meteoriteswith a range of FNi values and potentially account for theobserved variation among the IIIAB irons shown in Fig 8Measurements of Ni isotopes in adjacent metallic and non-metallic phases in iron meteorites could be used to test thishypothesis

Iron IsotopesMullane et al (2005) reported Fe isotope fractionation

over a large range (049permil amuminus1) for IIIAB irons in whichsamples were both heavier and lighter than the IRMM-014standard These data are inconsistent with other studies of Feisotopes in IIIAB iron meteorite metal in which all sampleshad Fe isotopic compositions similar to or slightly heavierthan IRMM-014 and spanned a narrow range of007permil amuminus1 (Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006) We analyzedthe Fe isotopes in six of the IIIAB irons with low and high Niabundances in which we also measured Ni isotopes The56Fe54Fe ratios in permil amuminus1 (or FFe) for these six samples are

Fig 7 A plot of Ir versus Ni abundances for members of the IIIABiron meteorite group The fourteen samples selected for Ni isotopicanalyses are shown by the black circles Data are from Scott et al(1973) Scott and Wasson (1976) Kracher et al (1980) Malvin et al(1984) and Wasson et al (1989 1998)

Fig 8 FNi values (permil amuminus1) versus Ni abundances for IIIAB ironsPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Nickel abundance data are fromScott et al (1973) Scott and Wasson (1976) and Malvin et al(1984)

2074 D L Cook et al

shown in Fig 9 along with Fe isotope data for IIIABs fromother studies As shown in Fig 9 our measurements do notshow the large range reported by Mullane et al (2005) All sixsamples have Fe isotopic compositions similar to orsomewhat heavier than IRMM-014 and span an overall rangeof only 007permil amuminus1 Both of these features are consistentwith the measurements of Zhu et al (2001) Kehm et al(2003) Schoenberg et al (2006) and Williams et al (2006)Furthermore there is no correlation between the Ni and Feisotopic compositions of these six samples (r2 = 0105) Ourresults suggest that the processes that fractionated the Niisotopes did not affect the Fe isotopes to the same degree orthat Ni and Fe isotopes were fractionated by differentprocesses on the IIIAB parent body

Isotopic Fractionation Between Kamacite and Taenite

Poitrasson et al (2005) and Horn et al (2006) showedthat Fe isotopes were fractionated between kamacite andtaenite in IAB irons These authors reported that kamacite(Fe-rich phase) is lighter in terms of its Fe isotopiccomposition relative to taenite (Ni-rich phase) BecauseToluca has been studied for both cooling rates (eg Wood1964 Hopfe and Goldstein 2001) and Fe isotopes it wasselected for an investigation of Fe and Ni isotopicfractionation associated with growth of the Widmanstaumlttenpattern The region of Toluca chosen for analyses is shownin Fig 10 Samples were collected along a traverse ofadjacent kamacite (spots 1 3 and 5) and taenite (spots 2and 4) The Ni isotopic composition of these five samples isshown in Fig 11 and Table 2 Nickel isotopes show clearlyresolvable fractionation between these two phases kamaciteis heavier relative to taenite by 036permil amuminus1 Bourdon

et al (2006) reported a similar fractionation of Ni isotopesbetween kamacite and taenite in iron meteorites The Feisotopic compositions of the five samples shown in Fig 10were also measured with the results shown in Fig 12 andTable 2 Unlike the isotopes of Ni the Fe isotopes do notshow a resolvable fractionation between kamacite andtaenite These results are in contrast to those of Poitrassonet al (2005) and Horn et al (2006) As discussed below thereason for this discrepancy is likely due to differences insampling

Dauphas and Rouxel (2006) showed that phase growth ina diffusion-limited regime can create diffusive kinetic isotopefractionation between two phases and control the expressionof equilibrium isotope fractionation between the co-existingphases More recently Dauphas (2007) developed anumerical simulation to model the formation of theWidmanstaumltten pattern during cooling of Fe-Ni alloy and tocalculate the associated fractionation of Ni and Fe isotopesDuring formation of the Widmanstaumltten pattern nucleation ofkamacite occurs in the solid state and its growth is controlledby diffusion (Wood 1964 Goldstein and Ogilvie 1965) It islikely that kinetic isotopic fractionation will accompany thisprocess because light isotopes diffuse faster than heavyisotopes of the same element (eg Richter et al 2003Roskosz et al 2006) The model predicts enrichment of lightNi isotopes in taenite relative to kamacite caused by diffusionduring cooling and formation of the Widmanstaumltten pattern inFe-Ni alloy These results are consistent with the measuredfractionations shown in Fig 11

Model results show that diffusion-driven kineticisotopic fractionation between two phases depends both on

Fig 9 Fe isotopic compositions (permil amuminus1) for the metal from IIIABiron meteorites relative to the IRMM-014 standard FFe = δ56542Circles = this study squares = Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006 hatched area = rangereported by Mullane et al (2005) Fig 10 SEM X-ray Ni map for the region of the Toluca IAB iron

chosen for microdrilling The numbers show the five spots sampledwith the microdrill Bright areas are Ni-rich Scale bar is 1000 μm

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2075

elemental concentration gradients between the two phasesand on the absolute elemental concentrations in each phaseSmaller concentration gradients lead to less fractionationand large absolute elemental concentrations act to dilutefractionation effects The absolute Fe concentrations inkamacite and taenite are much greater than the Niconcentrations in these two phases This factor can accountfor the difference in the magnitude of the measuredfractionations for Ni and Fe isotopes shown in Figs 11 and12 Furthermore inadvertent sampling of kamacite-taenitemixtures in Toluca by drilling through one phase and intothe other could explain why only resolvable effects in FNiwere measured in this study Because the FeNi ratios areunequal between kamacite and taenite the effects on FFeand FNi from mixing these two phases during sampling isnot linear Depending on the proportions of kamacite andtaenite involved in mixing FNi may change rapidly as FFechanges slowly or vice-versa relative to the isotopiccompositions of the pure phases For example mixingtaenite and kamacite in a 31 ratio changes FFe by 28whereas FNi changes only by 76 relative to pure taeniteSampling of pure kamacite and taenite-kamacite mixturescould account for the observed differences in FNi and lackof observed differences in FFe between kamacite and taenitein Toluca as shown in Figs 11 and 12

Although the observed isotopic fractionations in Tolucaare consistent with kinetic effects (especially given the role ofdiffusion in formation of the Widmanstaumltten pattern)equilibrium fractionation cannot be ruled out Modelingresults show that equilibrium and kinetic fractionation effectscould potentially be resolved if precise measurements of Niand Fe isotopes could be made at a higher spatial resolution(ie asymp20 μm) than has been achieved thus far

CONCLUSIONS

The Ni isotopes in bulk metal samples from 36meteorites follow a mass-dependent fractionation trendThese results are consistent with previous work (Moynieret al 2004 2005) and imply that Ni isotopes werefractionated from an initially isotopically homogeneousreservoir The Ni isotopic composition of metal fromchondrites and iron meteorites (magmatic and non-magmaticgroups) is similar to or heavier than SRM 986 whereaspallasite metal is similar to or lighter than SRM 986Measurements of additional pallasites are needed todetermine whether this difference is real or the result ofsampling bias

A suite of fourteen IIIAB irons shows fractionationover a considerable range The overall degree of massfractionation of Ni isotopes does not correlate with the Nicontent suggesting that core crystallization did not massfractionate Ni isotopes in a systematic way The variabilityobserved in the IIIAB irons may be due to the interaction ofimmiscible melts during the later stages of corecrystallization or due to the crystallization of metallic andnon-metallic phases Measurements of Ni isotopes inadjacent metal phosphide and sulfide phases in ironmeteorites are needed to determine the cause of the Ni

Fig 11 FNi values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Kamacite is heavier relative totaenite by 036permil amuminus1 Plotted errors are 2SE the externalprecision (2SD) is shown by the two dashed lines (0065permil amuminus1)

Fig 12 FFe values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Within uncertainty Fe isotopesshow no fractionation between kamacite and taenite

Table 2 The FNi and FFe values (permil amuminus1) for kamacite and taenite from the Toluca IAB iron

Sample FNi plusmn 2SE n FFe plusmn 2SE n

Spot 1 (kamacite) 0084 plusmn 0028 5 0080 plusmn 0041 4Spot 2 (taenite) minus0302 plusmn 0082 5 0078 plusmn 0041 4Spot 3 (kamacite) 0169 plusmn 0073 5 0072 plusmn 0041 4Spot 4 (taenite) minus0178 plusmn 0045 5 0086 plusmn 0041 4Spot 5 (kamacite) 0103 plusmn 0034 5 0051 plusmn 0041 4

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2071

individual kamacite and taenite in Toluca (discussed below)show that Ni isotopes are fractionated between these twophases and can lead to mass-dependent Ni isotopicvariations in iron meteorites at the millimeter scaleTherefore small metal samples of iron meteorites (andpallasites) characterized by a coarse Widmanstaumltten patternmay be skewed toward the isotopic composition ofkamacite or taenite As such in the discussion on the Niisotope fractionation in bulk samples of IIIAB ironmeteorites only the larger pieces (samples 2 in Table 1) ofeach of these three meteorites are considered

Meteoritic Metal Bulk Samples

The Ni isotopic compositions of the bulk metal from the36 meteorites analyzed (Table 1) are shown in a three-isotopeplot (Fig 5) Within error all samples in Fig 5 plot on a 11line which is consistent with mass-dependent isotopicfractionation The overall range of fractionation is041permil amuminus1 The observed mass-dependent fractionationand the overall range are both consistent with the resultspreviously reported by Moynier et al (2004 2005) for metalfrom iron meteorites pallasites chondrites and mesosiderites

Table 1 The Ni isotopic composition of meteoritic bulk metal (permil) relative to the SRM 986 standard Sample Group δ6158 plusmn 2SE δ6258 plusmn 2SE δ6458 plusmn 2SE n

Renazzoa CR 047 plusmn 012 058 plusmn 010 086 plusmn 014 14Gujbaa CBa 060 plusmn 012 084 plusmn 012 121 plusmn 019 13Semarkonaa LL 30 018 plusmn 012 009 plusmn 017 016 plusmn 029 10Bishunpura LL 31 043 plusmn 014 053 plusmn 018 071 plusmn 026 5Forest Valea H4 053 plusmn 026 068 plusmn 020 103 plusmn 045 5Indarcha EH4 026 plusmn 021 030 plusmn 024 049 plusmn 045 5

Eagle Stationa PES 015 plusmn 005 013 plusmn 007 039 plusmn 014 5Albina PMG 002 plusmn 010 013 plusmn 009 025 plusmn 013 9Brenhama PMG minus025 plusmn 016 minus034 plusmn 007 minus034 plusmn 036 5Molonga PMG minus019 plusmn 008 minus027 plusmn 011 minus041 plusmn 029 5

Coahuilaa IIAB 021 plusmn 024 032 plusmn 021 059 plusmn 061 5Santa Luziaa IIAB 059 plusmn 006 076 plusmn 004 126 plusmn 013 10Carboa IID 097 plusmn 041 130 plusmn 045 190 plusmn 086 5Augustinovkab IIIAB minus005 plusmn 022 minus012 plusmn 027 minus020 plusmn 061 5Avocab IIIAB 022 plusmn 017 028 plusmn 024 035 plusmn 026 5Bald Eagleb IIIAB 036 plusmn 026 054 plusmn 018 085 plusmn 051 5Bear Creek IIIAB 015 plusmn 022 027 plusmn 023 061 plusmn 039 5Bella Roca 1a IIIAB 008 plusmn 016 010 plusmn 012 005 plusmn 014 8Bella Roca 2b 092 plusmn 025 117 plusmn 023 177 plusmn 045 5Carthage IIIAB 055 plusmn 006 066 plusmn 007 108 plusmn 010 5Casas Grandes 1a IIIAB 062 plusmn 003 082 plusmn 003 135 plusmn 009 14Casas Grandes 2 029 plusmn 012 037 plusmn 004 061 plusmn 006 5Grant IIIAB 013 plusmn 024 010 plusmn 020 023 plusmn 039 5Henbury 1a IIIAB 033 plusmn 004 042 plusmn 004 061 plusmn 007 14Henbury 2b 031 plusmn 022 036 plusmn 016 045 plusmn 010 5Nova Petropolisb IIIAB 024 plusmn 013 026 plusmn 011 030 plusmn 008 5Orange River Ironb IIIAB 023 plusmn 031 025 plusmn 041 043 plusmn 094 4Seneca Falls IIIAB 037 plusmn 011 050 plusmn 014 081 plusmn 015 5Thunda IIIAB 042 plusmn 016 046 plusmn 005 073 plusmn 012 5Wellandb IIIAB 072 plusmn 058 089 plusmn 071 140 plusmn 119 5Gibeona IVA 028 plusmn 021 037 plusmn 009 078 plusmn 025 5Yanhuitlana IVA 064 plusmn 018 075 plusmn 016 112 plusmn 045 5Cape of Good Hopea IVB 036 plusmn 033 043 plusmn 014 085 plusmn 026 5Hobaa IVB 045 plusmn 011 064 plusmn 012 094 plusmn 018 13Tlacotepeca IVB 014 plusmn 035 022 plusmn 048 033 plusmn 069 9

Canyon Diabloa IAB 039 plusmn 006 051 plusmn 005 095 plusmn 012 19Tolucaa IAB 044 plusmn 013 062 plusmn 009 094 plusmn 010 5Daytona IAB minus015 plusmn 010 minus014 plusmn 005 001 plusmn 035 5Mundrabillaa IAB 026 plusmn 007 036 plusmn 006 055 plusmn 011 14

aSamples previously investigated for non-mass-dependent effects in Ni isotopes (Cook et al 2006)bSample solutions provided by E Mullane

n = Number of repeat measurements

2072 D L Cook et al

The FNi values for all 36 bulk metal samples are plottedin Fig 6 according to meteorite class Figure 6 shows that 1)the metal from chondrites and from iron meteorites(magmatic and non-magmatic) is typically similar to orheavier than SRM 986 and 2) metal from pallasites is similarto or lighter than SRM 986 The mean and two standarddeviations (2SD) of the FNi values for each class are asfollows chondrites (014 plusmn 011permil amuminus1) iron meteorites(012 plusmn 017permil amuminus1) and pallasites (minus002 plusmn 012permil amuminus1)The Ni isotopic compositions for each meteorite class werecompared to one another using t-tests These comparisonsshow that there is no significant difference between thechondrites and the iron meteorites (t = 0417) However theNi isotopic composition for pallasite metal is significantlydifferent from both chondrite metal (t = 4325) and from ironmeteorites (t = 3200) It is possible that the mean Ni isotopiccomposition of pallasitic metal is biased due to the smallnumber of samples measured However if this is not the casethere are several possibilities for explaining the apparentlydistinct Ni isotopic composition of pallasite metal Ifpallasites form at the core-mantle boundary of planetesimals(eg Scott 1977) the Ni isotopic composition of pallasitemetal may reflect isotopic fractionation between the metalliccore and the overlying mantle silicates Zhu et al (2002) andPoitrasson et al (2005) reported that Fe isotopes arefractionated between metal and olivine in pallasites and theseeffects range from 002 to 011permil amuminus1 Recent experimentsindicate that Fe isotopes can undergo fractionation up toasymp23permil amuminus1 between metallic and silicate melts due tokinetic effects (Roskosz et al 2006) and are fractionated by

asymp017permil amuminus1 between sulfide and silicate melts attemperatures between 840 and 1000 degC due to equilibriumeffects (Schuessler et al 2007) Currently a lack of Niisotopic data for non-metal phases from pallasites (egsilicates and oxides) precludes an assessment of planetarydifferentiation as a potential explanation for the apparentlydistinct Ni isotopic composition of pallasite metalAlternatively the Ni isotopic compositions of pallasite metalmay reflect differences at the parent-body scale that wereestablished prior to accretion rather than reflectingfractionation effects from differentiation and core formationon their parent bodies Finally it is also possible that in thecase of some of the smaller samples (ie lt50 mg) frommeteorites with metal that is compositionally heterogeneouson the millimeter-scale (eg the four pallasites) themeasured Ni isotope compositions may not be representativeof the bulk metal in these meteorites Analyses of additionaland larger samples would help to ascertain if pallasitic metaldoes indeed have a mean Ni isotopic composition that differssignificantly from metal in other meteorite classes

IIIAB Iron Meteorites

Nickel IsotopesWith over two hundred members the IIIABs constitute

the largest group of magmatic iron meteorites Thesemeteorites cover a wide compositional range and the originalIIIAB core is likely well-represented by the large number ofsamples (Chabot and Haack 2006) Mullane et al (2005)reported that Fe isotopes from a suite of 33 IIIAB irons weremass fractionated over a large range of 049permil amuminus1however the Fe isotopic compositions did not correlate with

Fig 5 A three-isotope plot of the data from the analyses of bulkmetal from 36 meteorites The data follow the trend expected formass-dependent fractionation a line with a slope equal to 1 is shownfor reference The overall range of fractionation is 041permil amuminus1Plotted errors are 2SE Filled triangles = chondrites squares =pallasites open triangles = magmatic irons circles = non-magmaticirons

Fig 6 FNi values (permil amuminus1) for the bulk metal from 36 meteoritesPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Filled triangles = chondritessquares = pallasites open triangles = magmatic irons circles = non-magmatic irons

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2073

the Ni contents which are a proxy for the degree ofcrystallization of the metallic core (eg Scott 1972) Hencefourteen IIIAB irons were selected to examine the degree ofmass-dependent fractionation of Ni isotopes possibly due tofractional crystallization of the core and to determine whetherthe fractionation effects on Fe and Ni isotopes werecorrelated Figure 7 shows the individual members of theIIIAB group in terms of their Ir and Ni contents including thefourteen samples chosen for Ni isotopic analyses Sampleswere chosen to represent the full compositional range of theIIIAB irons Figure 8 shows the FNi values for the fourteenIIIAB irons The Ni isotopes show fractionation over aconsiderable range from minus002 to 031permil amuminus1 The earliercrystallized low-Ni IIIA irons and the later crystallized high-Ni IIIB irons have similar mean compositions of01permil amuminus1 Furthermore the degree of mass fractionationof Ni isotopes does not correlate with the Ni content (r2 =0003) The data in Fig 8 suggest that core crystallization didnot mass fractionate Ni isotopes in a systematic way

The isotopic composition of Ni in different IIIAB ironmeteorites likely reflects a combination of events such asfractionation between 1) liquid metal and silicate during coresegregation 2) immiscible melts in the molten core and 3)various phases during core solidification The later stages ofcore crystallization of iron meteorites were likely affected byliquid immiscibility resulting in the formation ofcompositionally distinct liquids due to the presence of S and Pin the molten core (Ulff-Moslashller 1998 Chabot and Drake2000) Nickel isotopes may have been fractionated as a resultof the interaction of a metallic melt with a S andor P-richmelt during core crystallization Furthermore Fe isotopes arefractionated between metal phosphide (ie schreibersite)

and sulfide (ie troilite) in pallasites (Poitrasson et al 2005Weyer et al 2005) and between metal and sulfide in ironmeteorites (Williams et al 2006) Crystallization followed byslow subsolidus cooling of metallic and non-metallic phasesin IIIAB irons may have fractionated Ni isotopes in a similarmanner This last mechanism is particularly appealingbecause the Ni concentrations are quite different betweenmetal troilite and schreibersite in iron meteorites (Buchwald1977 Jochum et al 1980) Diffusion of Ni along theseconcentration gradients during crystallization and cooling ofthe various phases would likely be accompanied by kineticisotopic fractionation The proximity of non-metallic phasesrelative to the metal could produce metal in iron meteoriteswith a range of FNi values and potentially account for theobserved variation among the IIIAB irons shown in Fig 8Measurements of Ni isotopes in adjacent metallic and non-metallic phases in iron meteorites could be used to test thishypothesis

Iron IsotopesMullane et al (2005) reported Fe isotope fractionation

over a large range (049permil amuminus1) for IIIAB irons in whichsamples were both heavier and lighter than the IRMM-014standard These data are inconsistent with other studies of Feisotopes in IIIAB iron meteorite metal in which all sampleshad Fe isotopic compositions similar to or slightly heavierthan IRMM-014 and spanned a narrow range of007permil amuminus1 (Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006) We analyzedthe Fe isotopes in six of the IIIAB irons with low and high Niabundances in which we also measured Ni isotopes The56Fe54Fe ratios in permil amuminus1 (or FFe) for these six samples are

Fig 7 A plot of Ir versus Ni abundances for members of the IIIABiron meteorite group The fourteen samples selected for Ni isotopicanalyses are shown by the black circles Data are from Scott et al(1973) Scott and Wasson (1976) Kracher et al (1980) Malvin et al(1984) and Wasson et al (1989 1998)

Fig 8 FNi values (permil amuminus1) versus Ni abundances for IIIAB ironsPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Nickel abundance data are fromScott et al (1973) Scott and Wasson (1976) and Malvin et al(1984)

2074 D L Cook et al

shown in Fig 9 along with Fe isotope data for IIIABs fromother studies As shown in Fig 9 our measurements do notshow the large range reported by Mullane et al (2005) All sixsamples have Fe isotopic compositions similar to orsomewhat heavier than IRMM-014 and span an overall rangeof only 007permil amuminus1 Both of these features are consistentwith the measurements of Zhu et al (2001) Kehm et al(2003) Schoenberg et al (2006) and Williams et al (2006)Furthermore there is no correlation between the Ni and Feisotopic compositions of these six samples (r2 = 0105) Ourresults suggest that the processes that fractionated the Niisotopes did not affect the Fe isotopes to the same degree orthat Ni and Fe isotopes were fractionated by differentprocesses on the IIIAB parent body

Isotopic Fractionation Between Kamacite and Taenite

Poitrasson et al (2005) and Horn et al (2006) showedthat Fe isotopes were fractionated between kamacite andtaenite in IAB irons These authors reported that kamacite(Fe-rich phase) is lighter in terms of its Fe isotopiccomposition relative to taenite (Ni-rich phase) BecauseToluca has been studied for both cooling rates (eg Wood1964 Hopfe and Goldstein 2001) and Fe isotopes it wasselected for an investigation of Fe and Ni isotopicfractionation associated with growth of the Widmanstaumlttenpattern The region of Toluca chosen for analyses is shownin Fig 10 Samples were collected along a traverse ofadjacent kamacite (spots 1 3 and 5) and taenite (spots 2and 4) The Ni isotopic composition of these five samples isshown in Fig 11 and Table 2 Nickel isotopes show clearlyresolvable fractionation between these two phases kamaciteis heavier relative to taenite by 036permil amuminus1 Bourdon

et al (2006) reported a similar fractionation of Ni isotopesbetween kamacite and taenite in iron meteorites The Feisotopic compositions of the five samples shown in Fig 10were also measured with the results shown in Fig 12 andTable 2 Unlike the isotopes of Ni the Fe isotopes do notshow a resolvable fractionation between kamacite andtaenite These results are in contrast to those of Poitrassonet al (2005) and Horn et al (2006) As discussed below thereason for this discrepancy is likely due to differences insampling

Dauphas and Rouxel (2006) showed that phase growth ina diffusion-limited regime can create diffusive kinetic isotopefractionation between two phases and control the expressionof equilibrium isotope fractionation between the co-existingphases More recently Dauphas (2007) developed anumerical simulation to model the formation of theWidmanstaumltten pattern during cooling of Fe-Ni alloy and tocalculate the associated fractionation of Ni and Fe isotopesDuring formation of the Widmanstaumltten pattern nucleation ofkamacite occurs in the solid state and its growth is controlledby diffusion (Wood 1964 Goldstein and Ogilvie 1965) It islikely that kinetic isotopic fractionation will accompany thisprocess because light isotopes diffuse faster than heavyisotopes of the same element (eg Richter et al 2003Roskosz et al 2006) The model predicts enrichment of lightNi isotopes in taenite relative to kamacite caused by diffusionduring cooling and formation of the Widmanstaumltten pattern inFe-Ni alloy These results are consistent with the measuredfractionations shown in Fig 11

Model results show that diffusion-driven kineticisotopic fractionation between two phases depends both on

Fig 9 Fe isotopic compositions (permil amuminus1) for the metal from IIIABiron meteorites relative to the IRMM-014 standard FFe = δ56542Circles = this study squares = Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006 hatched area = rangereported by Mullane et al (2005) Fig 10 SEM X-ray Ni map for the region of the Toluca IAB iron

chosen for microdrilling The numbers show the five spots sampledwith the microdrill Bright areas are Ni-rich Scale bar is 1000 μm

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2075

elemental concentration gradients between the two phasesand on the absolute elemental concentrations in each phaseSmaller concentration gradients lead to less fractionationand large absolute elemental concentrations act to dilutefractionation effects The absolute Fe concentrations inkamacite and taenite are much greater than the Niconcentrations in these two phases This factor can accountfor the difference in the magnitude of the measuredfractionations for Ni and Fe isotopes shown in Figs 11 and12 Furthermore inadvertent sampling of kamacite-taenitemixtures in Toluca by drilling through one phase and intothe other could explain why only resolvable effects in FNiwere measured in this study Because the FeNi ratios areunequal between kamacite and taenite the effects on FFeand FNi from mixing these two phases during sampling isnot linear Depending on the proportions of kamacite andtaenite involved in mixing FNi may change rapidly as FFechanges slowly or vice-versa relative to the isotopiccompositions of the pure phases For example mixingtaenite and kamacite in a 31 ratio changes FFe by 28whereas FNi changes only by 76 relative to pure taeniteSampling of pure kamacite and taenite-kamacite mixturescould account for the observed differences in FNi and lackof observed differences in FFe between kamacite and taenitein Toluca as shown in Figs 11 and 12

Although the observed isotopic fractionations in Tolucaare consistent with kinetic effects (especially given the role ofdiffusion in formation of the Widmanstaumltten pattern)equilibrium fractionation cannot be ruled out Modelingresults show that equilibrium and kinetic fractionation effectscould potentially be resolved if precise measurements of Niand Fe isotopes could be made at a higher spatial resolution(ie asymp20 μm) than has been achieved thus far

CONCLUSIONS

The Ni isotopes in bulk metal samples from 36meteorites follow a mass-dependent fractionation trendThese results are consistent with previous work (Moynieret al 2004 2005) and imply that Ni isotopes werefractionated from an initially isotopically homogeneousreservoir The Ni isotopic composition of metal fromchondrites and iron meteorites (magmatic and non-magmaticgroups) is similar to or heavier than SRM 986 whereaspallasite metal is similar to or lighter than SRM 986Measurements of additional pallasites are needed todetermine whether this difference is real or the result ofsampling bias

A suite of fourteen IIIAB irons shows fractionationover a considerable range The overall degree of massfractionation of Ni isotopes does not correlate with the Nicontent suggesting that core crystallization did not massfractionate Ni isotopes in a systematic way The variabilityobserved in the IIIAB irons may be due to the interaction ofimmiscible melts during the later stages of corecrystallization or due to the crystallization of metallic andnon-metallic phases Measurements of Ni isotopes inadjacent metal phosphide and sulfide phases in ironmeteorites are needed to determine the cause of the Ni

Fig 11 FNi values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Kamacite is heavier relative totaenite by 036permil amuminus1 Plotted errors are 2SE the externalprecision (2SD) is shown by the two dashed lines (0065permil amuminus1)

Fig 12 FFe values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Within uncertainty Fe isotopesshow no fractionation between kamacite and taenite

Table 2 The FNi and FFe values (permil amuminus1) for kamacite and taenite from the Toluca IAB iron

Sample FNi plusmn 2SE n FFe plusmn 2SE n

Spot 1 (kamacite) 0084 plusmn 0028 5 0080 plusmn 0041 4Spot 2 (taenite) minus0302 plusmn 0082 5 0078 plusmn 0041 4Spot 3 (kamacite) 0169 plusmn 0073 5 0072 plusmn 0041 4Spot 4 (taenite) minus0178 plusmn 0045 5 0086 plusmn 0041 4Spot 5 (kamacite) 0103 plusmn 0034 5 0051 plusmn 0041 4

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

2072 D L Cook et al

The FNi values for all 36 bulk metal samples are plottedin Fig 6 according to meteorite class Figure 6 shows that 1)the metal from chondrites and from iron meteorites(magmatic and non-magmatic) is typically similar to orheavier than SRM 986 and 2) metal from pallasites is similarto or lighter than SRM 986 The mean and two standarddeviations (2SD) of the FNi values for each class are asfollows chondrites (014 plusmn 011permil amuminus1) iron meteorites(012 plusmn 017permil amuminus1) and pallasites (minus002 plusmn 012permil amuminus1)The Ni isotopic compositions for each meteorite class werecompared to one another using t-tests These comparisonsshow that there is no significant difference between thechondrites and the iron meteorites (t = 0417) However theNi isotopic composition for pallasite metal is significantlydifferent from both chondrite metal (t = 4325) and from ironmeteorites (t = 3200) It is possible that the mean Ni isotopiccomposition of pallasitic metal is biased due to the smallnumber of samples measured However if this is not the casethere are several possibilities for explaining the apparentlydistinct Ni isotopic composition of pallasite metal Ifpallasites form at the core-mantle boundary of planetesimals(eg Scott 1977) the Ni isotopic composition of pallasitemetal may reflect isotopic fractionation between the metalliccore and the overlying mantle silicates Zhu et al (2002) andPoitrasson et al (2005) reported that Fe isotopes arefractionated between metal and olivine in pallasites and theseeffects range from 002 to 011permil amuminus1 Recent experimentsindicate that Fe isotopes can undergo fractionation up toasymp23permil amuminus1 between metallic and silicate melts due tokinetic effects (Roskosz et al 2006) and are fractionated by

asymp017permil amuminus1 between sulfide and silicate melts attemperatures between 840 and 1000 degC due to equilibriumeffects (Schuessler et al 2007) Currently a lack of Niisotopic data for non-metal phases from pallasites (egsilicates and oxides) precludes an assessment of planetarydifferentiation as a potential explanation for the apparentlydistinct Ni isotopic composition of pallasite metalAlternatively the Ni isotopic compositions of pallasite metalmay reflect differences at the parent-body scale that wereestablished prior to accretion rather than reflectingfractionation effects from differentiation and core formationon their parent bodies Finally it is also possible that in thecase of some of the smaller samples (ie lt50 mg) frommeteorites with metal that is compositionally heterogeneouson the millimeter-scale (eg the four pallasites) themeasured Ni isotope compositions may not be representativeof the bulk metal in these meteorites Analyses of additionaland larger samples would help to ascertain if pallasitic metaldoes indeed have a mean Ni isotopic composition that differssignificantly from metal in other meteorite classes

IIIAB Iron Meteorites

Nickel IsotopesWith over two hundred members the IIIABs constitute

the largest group of magmatic iron meteorites Thesemeteorites cover a wide compositional range and the originalIIIAB core is likely well-represented by the large number ofsamples (Chabot and Haack 2006) Mullane et al (2005)reported that Fe isotopes from a suite of 33 IIIAB irons weremass fractionated over a large range of 049permil amuminus1however the Fe isotopic compositions did not correlate with

Fig 5 A three-isotope plot of the data from the analyses of bulkmetal from 36 meteorites The data follow the trend expected formass-dependent fractionation a line with a slope equal to 1 is shownfor reference The overall range of fractionation is 041permil amuminus1Plotted errors are 2SE Filled triangles = chondrites squares =pallasites open triangles = magmatic irons circles = non-magmaticirons

Fig 6 FNi values (permil amuminus1) for the bulk metal from 36 meteoritesPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Filled triangles = chondritessquares = pallasites open triangles = magmatic irons circles = non-magmatic irons

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2073

the Ni contents which are a proxy for the degree ofcrystallization of the metallic core (eg Scott 1972) Hencefourteen IIIAB irons were selected to examine the degree ofmass-dependent fractionation of Ni isotopes possibly due tofractional crystallization of the core and to determine whetherthe fractionation effects on Fe and Ni isotopes werecorrelated Figure 7 shows the individual members of theIIIAB group in terms of their Ir and Ni contents including thefourteen samples chosen for Ni isotopic analyses Sampleswere chosen to represent the full compositional range of theIIIAB irons Figure 8 shows the FNi values for the fourteenIIIAB irons The Ni isotopes show fractionation over aconsiderable range from minus002 to 031permil amuminus1 The earliercrystallized low-Ni IIIA irons and the later crystallized high-Ni IIIB irons have similar mean compositions of01permil amuminus1 Furthermore the degree of mass fractionationof Ni isotopes does not correlate with the Ni content (r2 =0003) The data in Fig 8 suggest that core crystallization didnot mass fractionate Ni isotopes in a systematic way

The isotopic composition of Ni in different IIIAB ironmeteorites likely reflects a combination of events such asfractionation between 1) liquid metal and silicate during coresegregation 2) immiscible melts in the molten core and 3)various phases during core solidification The later stages ofcore crystallization of iron meteorites were likely affected byliquid immiscibility resulting in the formation ofcompositionally distinct liquids due to the presence of S and Pin the molten core (Ulff-Moslashller 1998 Chabot and Drake2000) Nickel isotopes may have been fractionated as a resultof the interaction of a metallic melt with a S andor P-richmelt during core crystallization Furthermore Fe isotopes arefractionated between metal phosphide (ie schreibersite)

and sulfide (ie troilite) in pallasites (Poitrasson et al 2005Weyer et al 2005) and between metal and sulfide in ironmeteorites (Williams et al 2006) Crystallization followed byslow subsolidus cooling of metallic and non-metallic phasesin IIIAB irons may have fractionated Ni isotopes in a similarmanner This last mechanism is particularly appealingbecause the Ni concentrations are quite different betweenmetal troilite and schreibersite in iron meteorites (Buchwald1977 Jochum et al 1980) Diffusion of Ni along theseconcentration gradients during crystallization and cooling ofthe various phases would likely be accompanied by kineticisotopic fractionation The proximity of non-metallic phasesrelative to the metal could produce metal in iron meteoriteswith a range of FNi values and potentially account for theobserved variation among the IIIAB irons shown in Fig 8Measurements of Ni isotopes in adjacent metallic and non-metallic phases in iron meteorites could be used to test thishypothesis

Iron IsotopesMullane et al (2005) reported Fe isotope fractionation

over a large range (049permil amuminus1) for IIIAB irons in whichsamples were both heavier and lighter than the IRMM-014standard These data are inconsistent with other studies of Feisotopes in IIIAB iron meteorite metal in which all sampleshad Fe isotopic compositions similar to or slightly heavierthan IRMM-014 and spanned a narrow range of007permil amuminus1 (Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006) We analyzedthe Fe isotopes in six of the IIIAB irons with low and high Niabundances in which we also measured Ni isotopes The56Fe54Fe ratios in permil amuminus1 (or FFe) for these six samples are

Fig 7 A plot of Ir versus Ni abundances for members of the IIIABiron meteorite group The fourteen samples selected for Ni isotopicanalyses are shown by the black circles Data are from Scott et al(1973) Scott and Wasson (1976) Kracher et al (1980) Malvin et al(1984) and Wasson et al (1989 1998)

Fig 8 FNi values (permil amuminus1) versus Ni abundances for IIIAB ironsPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Nickel abundance data are fromScott et al (1973) Scott and Wasson (1976) and Malvin et al(1984)

2074 D L Cook et al

shown in Fig 9 along with Fe isotope data for IIIABs fromother studies As shown in Fig 9 our measurements do notshow the large range reported by Mullane et al (2005) All sixsamples have Fe isotopic compositions similar to orsomewhat heavier than IRMM-014 and span an overall rangeof only 007permil amuminus1 Both of these features are consistentwith the measurements of Zhu et al (2001) Kehm et al(2003) Schoenberg et al (2006) and Williams et al (2006)Furthermore there is no correlation between the Ni and Feisotopic compositions of these six samples (r2 = 0105) Ourresults suggest that the processes that fractionated the Niisotopes did not affect the Fe isotopes to the same degree orthat Ni and Fe isotopes were fractionated by differentprocesses on the IIIAB parent body

Isotopic Fractionation Between Kamacite and Taenite

Poitrasson et al (2005) and Horn et al (2006) showedthat Fe isotopes were fractionated between kamacite andtaenite in IAB irons These authors reported that kamacite(Fe-rich phase) is lighter in terms of its Fe isotopiccomposition relative to taenite (Ni-rich phase) BecauseToluca has been studied for both cooling rates (eg Wood1964 Hopfe and Goldstein 2001) and Fe isotopes it wasselected for an investigation of Fe and Ni isotopicfractionation associated with growth of the Widmanstaumlttenpattern The region of Toluca chosen for analyses is shownin Fig 10 Samples were collected along a traverse ofadjacent kamacite (spots 1 3 and 5) and taenite (spots 2and 4) The Ni isotopic composition of these five samples isshown in Fig 11 and Table 2 Nickel isotopes show clearlyresolvable fractionation between these two phases kamaciteis heavier relative to taenite by 036permil amuminus1 Bourdon

et al (2006) reported a similar fractionation of Ni isotopesbetween kamacite and taenite in iron meteorites The Feisotopic compositions of the five samples shown in Fig 10were also measured with the results shown in Fig 12 andTable 2 Unlike the isotopes of Ni the Fe isotopes do notshow a resolvable fractionation between kamacite andtaenite These results are in contrast to those of Poitrassonet al (2005) and Horn et al (2006) As discussed below thereason for this discrepancy is likely due to differences insampling

Dauphas and Rouxel (2006) showed that phase growth ina diffusion-limited regime can create diffusive kinetic isotopefractionation between two phases and control the expressionof equilibrium isotope fractionation between the co-existingphases More recently Dauphas (2007) developed anumerical simulation to model the formation of theWidmanstaumltten pattern during cooling of Fe-Ni alloy and tocalculate the associated fractionation of Ni and Fe isotopesDuring formation of the Widmanstaumltten pattern nucleation ofkamacite occurs in the solid state and its growth is controlledby diffusion (Wood 1964 Goldstein and Ogilvie 1965) It islikely that kinetic isotopic fractionation will accompany thisprocess because light isotopes diffuse faster than heavyisotopes of the same element (eg Richter et al 2003Roskosz et al 2006) The model predicts enrichment of lightNi isotopes in taenite relative to kamacite caused by diffusionduring cooling and formation of the Widmanstaumltten pattern inFe-Ni alloy These results are consistent with the measuredfractionations shown in Fig 11

Model results show that diffusion-driven kineticisotopic fractionation between two phases depends both on

Fig 9 Fe isotopic compositions (permil amuminus1) for the metal from IIIABiron meteorites relative to the IRMM-014 standard FFe = δ56542Circles = this study squares = Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006 hatched area = rangereported by Mullane et al (2005) Fig 10 SEM X-ray Ni map for the region of the Toluca IAB iron

chosen for microdrilling The numbers show the five spots sampledwith the microdrill Bright areas are Ni-rich Scale bar is 1000 μm

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2075

elemental concentration gradients between the two phasesand on the absolute elemental concentrations in each phaseSmaller concentration gradients lead to less fractionationand large absolute elemental concentrations act to dilutefractionation effects The absolute Fe concentrations inkamacite and taenite are much greater than the Niconcentrations in these two phases This factor can accountfor the difference in the magnitude of the measuredfractionations for Ni and Fe isotopes shown in Figs 11 and12 Furthermore inadvertent sampling of kamacite-taenitemixtures in Toluca by drilling through one phase and intothe other could explain why only resolvable effects in FNiwere measured in this study Because the FeNi ratios areunequal between kamacite and taenite the effects on FFeand FNi from mixing these two phases during sampling isnot linear Depending on the proportions of kamacite andtaenite involved in mixing FNi may change rapidly as FFechanges slowly or vice-versa relative to the isotopiccompositions of the pure phases For example mixingtaenite and kamacite in a 31 ratio changes FFe by 28whereas FNi changes only by 76 relative to pure taeniteSampling of pure kamacite and taenite-kamacite mixturescould account for the observed differences in FNi and lackof observed differences in FFe between kamacite and taenitein Toluca as shown in Figs 11 and 12

Although the observed isotopic fractionations in Tolucaare consistent with kinetic effects (especially given the role ofdiffusion in formation of the Widmanstaumltten pattern)equilibrium fractionation cannot be ruled out Modelingresults show that equilibrium and kinetic fractionation effectscould potentially be resolved if precise measurements of Niand Fe isotopes could be made at a higher spatial resolution(ie asymp20 μm) than has been achieved thus far

CONCLUSIONS

The Ni isotopes in bulk metal samples from 36meteorites follow a mass-dependent fractionation trendThese results are consistent with previous work (Moynieret al 2004 2005) and imply that Ni isotopes werefractionated from an initially isotopically homogeneousreservoir The Ni isotopic composition of metal fromchondrites and iron meteorites (magmatic and non-magmaticgroups) is similar to or heavier than SRM 986 whereaspallasite metal is similar to or lighter than SRM 986Measurements of additional pallasites are needed todetermine whether this difference is real or the result ofsampling bias

A suite of fourteen IIIAB irons shows fractionationover a considerable range The overall degree of massfractionation of Ni isotopes does not correlate with the Nicontent suggesting that core crystallization did not massfractionate Ni isotopes in a systematic way The variabilityobserved in the IIIAB irons may be due to the interaction ofimmiscible melts during the later stages of corecrystallization or due to the crystallization of metallic andnon-metallic phases Measurements of Ni isotopes inadjacent metal phosphide and sulfide phases in ironmeteorites are needed to determine the cause of the Ni

Fig 11 FNi values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Kamacite is heavier relative totaenite by 036permil amuminus1 Plotted errors are 2SE the externalprecision (2SD) is shown by the two dashed lines (0065permil amuminus1)

Fig 12 FFe values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Within uncertainty Fe isotopesshow no fractionation between kamacite and taenite

Table 2 The FNi and FFe values (permil amuminus1) for kamacite and taenite from the Toluca IAB iron

Sample FNi plusmn 2SE n FFe plusmn 2SE n

Spot 1 (kamacite) 0084 plusmn 0028 5 0080 plusmn 0041 4Spot 2 (taenite) minus0302 plusmn 0082 5 0078 plusmn 0041 4Spot 3 (kamacite) 0169 plusmn 0073 5 0072 plusmn 0041 4Spot 4 (taenite) minus0178 plusmn 0045 5 0086 plusmn 0041 4Spot 5 (kamacite) 0103 plusmn 0034 5 0051 plusmn 0041 4

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2073

the Ni contents which are a proxy for the degree ofcrystallization of the metallic core (eg Scott 1972) Hencefourteen IIIAB irons were selected to examine the degree ofmass-dependent fractionation of Ni isotopes possibly due tofractional crystallization of the core and to determine whetherthe fractionation effects on Fe and Ni isotopes werecorrelated Figure 7 shows the individual members of theIIIAB group in terms of their Ir and Ni contents including thefourteen samples chosen for Ni isotopic analyses Sampleswere chosen to represent the full compositional range of theIIIAB irons Figure 8 shows the FNi values for the fourteenIIIAB irons The Ni isotopes show fractionation over aconsiderable range from minus002 to 031permil amuminus1 The earliercrystallized low-Ni IIIA irons and the later crystallized high-Ni IIIB irons have similar mean compositions of01permil amuminus1 Furthermore the degree of mass fractionationof Ni isotopes does not correlate with the Ni content (r2 =0003) The data in Fig 8 suggest that core crystallization didnot mass fractionate Ni isotopes in a systematic way

The isotopic composition of Ni in different IIIAB ironmeteorites likely reflects a combination of events such asfractionation between 1) liquid metal and silicate during coresegregation 2) immiscible melts in the molten core and 3)various phases during core solidification The later stages ofcore crystallization of iron meteorites were likely affected byliquid immiscibility resulting in the formation ofcompositionally distinct liquids due to the presence of S and Pin the molten core (Ulff-Moslashller 1998 Chabot and Drake2000) Nickel isotopes may have been fractionated as a resultof the interaction of a metallic melt with a S andor P-richmelt during core crystallization Furthermore Fe isotopes arefractionated between metal phosphide (ie schreibersite)

and sulfide (ie troilite) in pallasites (Poitrasson et al 2005Weyer et al 2005) and between metal and sulfide in ironmeteorites (Williams et al 2006) Crystallization followed byslow subsolidus cooling of metallic and non-metallic phasesin IIIAB irons may have fractionated Ni isotopes in a similarmanner This last mechanism is particularly appealingbecause the Ni concentrations are quite different betweenmetal troilite and schreibersite in iron meteorites (Buchwald1977 Jochum et al 1980) Diffusion of Ni along theseconcentration gradients during crystallization and cooling ofthe various phases would likely be accompanied by kineticisotopic fractionation The proximity of non-metallic phasesrelative to the metal could produce metal in iron meteoriteswith a range of FNi values and potentially account for theobserved variation among the IIIAB irons shown in Fig 8Measurements of Ni isotopes in adjacent metallic and non-metallic phases in iron meteorites could be used to test thishypothesis

Iron IsotopesMullane et al (2005) reported Fe isotope fractionation

over a large range (049permil amuminus1) for IIIAB irons in whichsamples were both heavier and lighter than the IRMM-014standard These data are inconsistent with other studies of Feisotopes in IIIAB iron meteorite metal in which all sampleshad Fe isotopic compositions similar to or slightly heavierthan IRMM-014 and spanned a narrow range of007permil amuminus1 (Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006) We analyzedthe Fe isotopes in six of the IIIAB irons with low and high Niabundances in which we also measured Ni isotopes The56Fe54Fe ratios in permil amuminus1 (or FFe) for these six samples are

Fig 7 A plot of Ir versus Ni abundances for members of the IIIABiron meteorite group The fourteen samples selected for Ni isotopicanalyses are shown by the black circles Data are from Scott et al(1973) Scott and Wasson (1976) Kracher et al (1980) Malvin et al(1984) and Wasson et al (1989 1998)

Fig 8 FNi values (permil amuminus1) versus Ni abundances for IIIAB ironsPlotted errors are 2SE the external precision (2SD) is shown by thetwo dashed lines (plusmn0065permil amuminus1) Nickel abundance data are fromScott et al (1973) Scott and Wasson (1976) and Malvin et al(1984)

2074 D L Cook et al

shown in Fig 9 along with Fe isotope data for IIIABs fromother studies As shown in Fig 9 our measurements do notshow the large range reported by Mullane et al (2005) All sixsamples have Fe isotopic compositions similar to orsomewhat heavier than IRMM-014 and span an overall rangeof only 007permil amuminus1 Both of these features are consistentwith the measurements of Zhu et al (2001) Kehm et al(2003) Schoenberg et al (2006) and Williams et al (2006)Furthermore there is no correlation between the Ni and Feisotopic compositions of these six samples (r2 = 0105) Ourresults suggest that the processes that fractionated the Niisotopes did not affect the Fe isotopes to the same degree orthat Ni and Fe isotopes were fractionated by differentprocesses on the IIIAB parent body

Isotopic Fractionation Between Kamacite and Taenite

Poitrasson et al (2005) and Horn et al (2006) showedthat Fe isotopes were fractionated between kamacite andtaenite in IAB irons These authors reported that kamacite(Fe-rich phase) is lighter in terms of its Fe isotopiccomposition relative to taenite (Ni-rich phase) BecauseToluca has been studied for both cooling rates (eg Wood1964 Hopfe and Goldstein 2001) and Fe isotopes it wasselected for an investigation of Fe and Ni isotopicfractionation associated with growth of the Widmanstaumlttenpattern The region of Toluca chosen for analyses is shownin Fig 10 Samples were collected along a traverse ofadjacent kamacite (spots 1 3 and 5) and taenite (spots 2and 4) The Ni isotopic composition of these five samples isshown in Fig 11 and Table 2 Nickel isotopes show clearlyresolvable fractionation between these two phases kamaciteis heavier relative to taenite by 036permil amuminus1 Bourdon

et al (2006) reported a similar fractionation of Ni isotopesbetween kamacite and taenite in iron meteorites The Feisotopic compositions of the five samples shown in Fig 10were also measured with the results shown in Fig 12 andTable 2 Unlike the isotopes of Ni the Fe isotopes do notshow a resolvable fractionation between kamacite andtaenite These results are in contrast to those of Poitrassonet al (2005) and Horn et al (2006) As discussed below thereason for this discrepancy is likely due to differences insampling

Dauphas and Rouxel (2006) showed that phase growth ina diffusion-limited regime can create diffusive kinetic isotopefractionation between two phases and control the expressionof equilibrium isotope fractionation between the co-existingphases More recently Dauphas (2007) developed anumerical simulation to model the formation of theWidmanstaumltten pattern during cooling of Fe-Ni alloy and tocalculate the associated fractionation of Ni and Fe isotopesDuring formation of the Widmanstaumltten pattern nucleation ofkamacite occurs in the solid state and its growth is controlledby diffusion (Wood 1964 Goldstein and Ogilvie 1965) It islikely that kinetic isotopic fractionation will accompany thisprocess because light isotopes diffuse faster than heavyisotopes of the same element (eg Richter et al 2003Roskosz et al 2006) The model predicts enrichment of lightNi isotopes in taenite relative to kamacite caused by diffusionduring cooling and formation of the Widmanstaumltten pattern inFe-Ni alloy These results are consistent with the measuredfractionations shown in Fig 11

Model results show that diffusion-driven kineticisotopic fractionation between two phases depends both on

Fig 9 Fe isotopic compositions (permil amuminus1) for the metal from IIIABiron meteorites relative to the IRMM-014 standard FFe = δ56542Circles = this study squares = Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006 hatched area = rangereported by Mullane et al (2005) Fig 10 SEM X-ray Ni map for the region of the Toluca IAB iron

chosen for microdrilling The numbers show the five spots sampledwith the microdrill Bright areas are Ni-rich Scale bar is 1000 μm

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2075

elemental concentration gradients between the two phasesand on the absolute elemental concentrations in each phaseSmaller concentration gradients lead to less fractionationand large absolute elemental concentrations act to dilutefractionation effects The absolute Fe concentrations inkamacite and taenite are much greater than the Niconcentrations in these two phases This factor can accountfor the difference in the magnitude of the measuredfractionations for Ni and Fe isotopes shown in Figs 11 and12 Furthermore inadvertent sampling of kamacite-taenitemixtures in Toluca by drilling through one phase and intothe other could explain why only resolvable effects in FNiwere measured in this study Because the FeNi ratios areunequal between kamacite and taenite the effects on FFeand FNi from mixing these two phases during sampling isnot linear Depending on the proportions of kamacite andtaenite involved in mixing FNi may change rapidly as FFechanges slowly or vice-versa relative to the isotopiccompositions of the pure phases For example mixingtaenite and kamacite in a 31 ratio changes FFe by 28whereas FNi changes only by 76 relative to pure taeniteSampling of pure kamacite and taenite-kamacite mixturescould account for the observed differences in FNi and lackof observed differences in FFe between kamacite and taenitein Toluca as shown in Figs 11 and 12

Although the observed isotopic fractionations in Tolucaare consistent with kinetic effects (especially given the role ofdiffusion in formation of the Widmanstaumltten pattern)equilibrium fractionation cannot be ruled out Modelingresults show that equilibrium and kinetic fractionation effectscould potentially be resolved if precise measurements of Niand Fe isotopes could be made at a higher spatial resolution(ie asymp20 μm) than has been achieved thus far

CONCLUSIONS

The Ni isotopes in bulk metal samples from 36meteorites follow a mass-dependent fractionation trendThese results are consistent with previous work (Moynieret al 2004 2005) and imply that Ni isotopes werefractionated from an initially isotopically homogeneousreservoir The Ni isotopic composition of metal fromchondrites and iron meteorites (magmatic and non-magmaticgroups) is similar to or heavier than SRM 986 whereaspallasite metal is similar to or lighter than SRM 986Measurements of additional pallasites are needed todetermine whether this difference is real or the result ofsampling bias

A suite of fourteen IIIAB irons shows fractionationover a considerable range The overall degree of massfractionation of Ni isotopes does not correlate with the Nicontent suggesting that core crystallization did not massfractionate Ni isotopes in a systematic way The variabilityobserved in the IIIAB irons may be due to the interaction ofimmiscible melts during the later stages of corecrystallization or due to the crystallization of metallic andnon-metallic phases Measurements of Ni isotopes inadjacent metal phosphide and sulfide phases in ironmeteorites are needed to determine the cause of the Ni

Fig 11 FNi values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Kamacite is heavier relative totaenite by 036permil amuminus1 Plotted errors are 2SE the externalprecision (2SD) is shown by the two dashed lines (0065permil amuminus1)

Fig 12 FFe values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Within uncertainty Fe isotopesshow no fractionation between kamacite and taenite

Table 2 The FNi and FFe values (permil amuminus1) for kamacite and taenite from the Toluca IAB iron

Sample FNi plusmn 2SE n FFe plusmn 2SE n

Spot 1 (kamacite) 0084 plusmn 0028 5 0080 plusmn 0041 4Spot 2 (taenite) minus0302 plusmn 0082 5 0078 plusmn 0041 4Spot 3 (kamacite) 0169 plusmn 0073 5 0072 plusmn 0041 4Spot 4 (taenite) minus0178 plusmn 0045 5 0086 plusmn 0041 4Spot 5 (kamacite) 0103 plusmn 0034 5 0051 plusmn 0041 4

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

2074 D L Cook et al

shown in Fig 9 along with Fe isotope data for IIIABs fromother studies As shown in Fig 9 our measurements do notshow the large range reported by Mullane et al (2005) All sixsamples have Fe isotopic compositions similar to orsomewhat heavier than IRMM-014 and span an overall rangeof only 007permil amuminus1 Both of these features are consistentwith the measurements of Zhu et al (2001) Kehm et al(2003) Schoenberg et al (2006) and Williams et al (2006)Furthermore there is no correlation between the Ni and Feisotopic compositions of these six samples (r2 = 0105) Ourresults suggest that the processes that fractionated the Niisotopes did not affect the Fe isotopes to the same degree orthat Ni and Fe isotopes were fractionated by differentprocesses on the IIIAB parent body

Isotopic Fractionation Between Kamacite and Taenite

Poitrasson et al (2005) and Horn et al (2006) showedthat Fe isotopes were fractionated between kamacite andtaenite in IAB irons These authors reported that kamacite(Fe-rich phase) is lighter in terms of its Fe isotopiccomposition relative to taenite (Ni-rich phase) BecauseToluca has been studied for both cooling rates (eg Wood1964 Hopfe and Goldstein 2001) and Fe isotopes it wasselected for an investigation of Fe and Ni isotopicfractionation associated with growth of the Widmanstaumlttenpattern The region of Toluca chosen for analyses is shownin Fig 10 Samples were collected along a traverse ofadjacent kamacite (spots 1 3 and 5) and taenite (spots 2and 4) The Ni isotopic composition of these five samples isshown in Fig 11 and Table 2 Nickel isotopes show clearlyresolvable fractionation between these two phases kamaciteis heavier relative to taenite by 036permil amuminus1 Bourdon

et al (2006) reported a similar fractionation of Ni isotopesbetween kamacite and taenite in iron meteorites The Feisotopic compositions of the five samples shown in Fig 10were also measured with the results shown in Fig 12 andTable 2 Unlike the isotopes of Ni the Fe isotopes do notshow a resolvable fractionation between kamacite andtaenite These results are in contrast to those of Poitrassonet al (2005) and Horn et al (2006) As discussed below thereason for this discrepancy is likely due to differences insampling

Dauphas and Rouxel (2006) showed that phase growth ina diffusion-limited regime can create diffusive kinetic isotopefractionation between two phases and control the expressionof equilibrium isotope fractionation between the co-existingphases More recently Dauphas (2007) developed anumerical simulation to model the formation of theWidmanstaumltten pattern during cooling of Fe-Ni alloy and tocalculate the associated fractionation of Ni and Fe isotopesDuring formation of the Widmanstaumltten pattern nucleation ofkamacite occurs in the solid state and its growth is controlledby diffusion (Wood 1964 Goldstein and Ogilvie 1965) It islikely that kinetic isotopic fractionation will accompany thisprocess because light isotopes diffuse faster than heavyisotopes of the same element (eg Richter et al 2003Roskosz et al 2006) The model predicts enrichment of lightNi isotopes in taenite relative to kamacite caused by diffusionduring cooling and formation of the Widmanstaumltten pattern inFe-Ni alloy These results are consistent with the measuredfractionations shown in Fig 11

Model results show that diffusion-driven kineticisotopic fractionation between two phases depends both on

Fig 9 Fe isotopic compositions (permil amuminus1) for the metal from IIIABiron meteorites relative to the IRMM-014 standard FFe = δ56542Circles = this study squares = Zhu et al 2001 Kehm et al 2003Schoenberg et al 2006 Williams et al 2006 hatched area = rangereported by Mullane et al (2005) Fig 10 SEM X-ray Ni map for the region of the Toluca IAB iron

chosen for microdrilling The numbers show the five spots sampledwith the microdrill Bright areas are Ni-rich Scale bar is 1000 μm

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2075

elemental concentration gradients between the two phasesand on the absolute elemental concentrations in each phaseSmaller concentration gradients lead to less fractionationand large absolute elemental concentrations act to dilutefractionation effects The absolute Fe concentrations inkamacite and taenite are much greater than the Niconcentrations in these two phases This factor can accountfor the difference in the magnitude of the measuredfractionations for Ni and Fe isotopes shown in Figs 11 and12 Furthermore inadvertent sampling of kamacite-taenitemixtures in Toluca by drilling through one phase and intothe other could explain why only resolvable effects in FNiwere measured in this study Because the FeNi ratios areunequal between kamacite and taenite the effects on FFeand FNi from mixing these two phases during sampling isnot linear Depending on the proportions of kamacite andtaenite involved in mixing FNi may change rapidly as FFechanges slowly or vice-versa relative to the isotopiccompositions of the pure phases For example mixingtaenite and kamacite in a 31 ratio changes FFe by 28whereas FNi changes only by 76 relative to pure taeniteSampling of pure kamacite and taenite-kamacite mixturescould account for the observed differences in FNi and lackof observed differences in FFe between kamacite and taenitein Toluca as shown in Figs 11 and 12

Although the observed isotopic fractionations in Tolucaare consistent with kinetic effects (especially given the role ofdiffusion in formation of the Widmanstaumltten pattern)equilibrium fractionation cannot be ruled out Modelingresults show that equilibrium and kinetic fractionation effectscould potentially be resolved if precise measurements of Niand Fe isotopes could be made at a higher spatial resolution(ie asymp20 μm) than has been achieved thus far

CONCLUSIONS

The Ni isotopes in bulk metal samples from 36meteorites follow a mass-dependent fractionation trendThese results are consistent with previous work (Moynieret al 2004 2005) and imply that Ni isotopes werefractionated from an initially isotopically homogeneousreservoir The Ni isotopic composition of metal fromchondrites and iron meteorites (magmatic and non-magmaticgroups) is similar to or heavier than SRM 986 whereaspallasite metal is similar to or lighter than SRM 986Measurements of additional pallasites are needed todetermine whether this difference is real or the result ofsampling bias

A suite of fourteen IIIAB irons shows fractionationover a considerable range The overall degree of massfractionation of Ni isotopes does not correlate with the Nicontent suggesting that core crystallization did not massfractionate Ni isotopes in a systematic way The variabilityobserved in the IIIAB irons may be due to the interaction ofimmiscible melts during the later stages of corecrystallization or due to the crystallization of metallic andnon-metallic phases Measurements of Ni isotopes inadjacent metal phosphide and sulfide phases in ironmeteorites are needed to determine the cause of the Ni

Fig 11 FNi values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Kamacite is heavier relative totaenite by 036permil amuminus1 Plotted errors are 2SE the externalprecision (2SD) is shown by the two dashed lines (0065permil amuminus1)

Fig 12 FFe values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Within uncertainty Fe isotopesshow no fractionation between kamacite and taenite

Table 2 The FNi and FFe values (permil amuminus1) for kamacite and taenite from the Toluca IAB iron

Sample FNi plusmn 2SE n FFe plusmn 2SE n

Spot 1 (kamacite) 0084 plusmn 0028 5 0080 plusmn 0041 4Spot 2 (taenite) minus0302 plusmn 0082 5 0078 plusmn 0041 4Spot 3 (kamacite) 0169 plusmn 0073 5 0072 plusmn 0041 4Spot 4 (taenite) minus0178 plusmn 0045 5 0086 plusmn 0041 4Spot 5 (kamacite) 0103 plusmn 0034 5 0051 plusmn 0041 4

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2075

elemental concentration gradients between the two phasesand on the absolute elemental concentrations in each phaseSmaller concentration gradients lead to less fractionationand large absolute elemental concentrations act to dilutefractionation effects The absolute Fe concentrations inkamacite and taenite are much greater than the Niconcentrations in these two phases This factor can accountfor the difference in the magnitude of the measuredfractionations for Ni and Fe isotopes shown in Figs 11 and12 Furthermore inadvertent sampling of kamacite-taenitemixtures in Toluca by drilling through one phase and intothe other could explain why only resolvable effects in FNiwere measured in this study Because the FeNi ratios areunequal between kamacite and taenite the effects on FFeand FNi from mixing these two phases during sampling isnot linear Depending on the proportions of kamacite andtaenite involved in mixing FNi may change rapidly as FFechanges slowly or vice-versa relative to the isotopiccompositions of the pure phases For example mixingtaenite and kamacite in a 31 ratio changes FFe by 28whereas FNi changes only by 76 relative to pure taeniteSampling of pure kamacite and taenite-kamacite mixturescould account for the observed differences in FNi and lackof observed differences in FFe between kamacite and taenitein Toluca as shown in Figs 11 and 12

Although the observed isotopic fractionations in Tolucaare consistent with kinetic effects (especially given the role ofdiffusion in formation of the Widmanstaumltten pattern)equilibrium fractionation cannot be ruled out Modelingresults show that equilibrium and kinetic fractionation effectscould potentially be resolved if precise measurements of Niand Fe isotopes could be made at a higher spatial resolution(ie asymp20 μm) than has been achieved thus far

CONCLUSIONS

The Ni isotopes in bulk metal samples from 36meteorites follow a mass-dependent fractionation trendThese results are consistent with previous work (Moynieret al 2004 2005) and imply that Ni isotopes werefractionated from an initially isotopically homogeneousreservoir The Ni isotopic composition of metal fromchondrites and iron meteorites (magmatic and non-magmaticgroups) is similar to or heavier than SRM 986 whereaspallasite metal is similar to or lighter than SRM 986Measurements of additional pallasites are needed todetermine whether this difference is real or the result ofsampling bias

A suite of fourteen IIIAB irons shows fractionationover a considerable range The overall degree of massfractionation of Ni isotopes does not correlate with the Nicontent suggesting that core crystallization did not massfractionate Ni isotopes in a systematic way The variabilityobserved in the IIIAB irons may be due to the interaction ofimmiscible melts during the later stages of corecrystallization or due to the crystallization of metallic andnon-metallic phases Measurements of Ni isotopes inadjacent metal phosphide and sulfide phases in ironmeteorites are needed to determine the cause of the Ni

Fig 11 FNi values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Kamacite is heavier relative totaenite by 036permil amuminus1 Plotted errors are 2SE the externalprecision (2SD) is shown by the two dashed lines (0065permil amuminus1)

Fig 12 FFe values (permil amuminus1) for the five sample spots shown inFig 10 for the Toluca IAB iron Within uncertainty Fe isotopesshow no fractionation between kamacite and taenite

Table 2 The FNi and FFe values (permil amuminus1) for kamacite and taenite from the Toluca IAB iron

Sample FNi plusmn 2SE n FFe plusmn 2SE n

Spot 1 (kamacite) 0084 plusmn 0028 5 0080 plusmn 0041 4Spot 2 (taenite) minus0302 plusmn 0082 5 0078 plusmn 0041 4Spot 3 (kamacite) 0169 plusmn 0073 5 0072 plusmn 0041 4Spot 4 (taenite) minus0178 plusmn 0045 5 0086 plusmn 0041 4Spot 5 (kamacite) 0103 plusmn 0034 5 0051 plusmn 0041 4

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

2076 D L Cook et al

isotope variation recorded in the IIIAB irons Iron isotopeswere also measured in six of the IIIAB samples Withinuncertainty all samples have the same Fe isotopiccomposition and we find no evidence of the large rangereported by Mullane et al (2005)

The Ni and Fe isotopes in adjacent kamacite and taenitefrom the Toluca IAB iron were measured Nickel isotopesshow clearly resolvable fractionation between these twophases nickel in kamacite is heavier relative to taenite Incontrast the Fe isotopes do not show a resolvablefractionation between kamacite and taenite The observedisotopic compositions can be understood in terms of kineticfractionation due to diffusion during cooling of the Fe-Nialloy and the development of the Widmanstaumltten patternHowever equilibrium fractionation at the interface betweenkamacite and taenite cannot be ruled out as the cause of themeasured isotopic compositions and future measurements atbetter spatial resolution could potentially discriminatebetween equilibrium and kinetic effects

AcknowledgmentsndashWe thank E Mullane for providingaliquots from her solutions of IIIAB irons the Smithsonianfor providing the sample of Semarkona (USNM 1805) andGreg Herzog and an anonymous reviewer for helpfulcomments and suggestions This work was supported in partby the National Aeronautics and Space Administrationthrough grants to M W R N C N D and A M D

Editorial HandlingmdashDr Edward Scott

REFERENCES

Bourdon B Quitteacute G and Halliday A N 2006 Kineticfractionation of nickel and iron between kamacite and taeniteInsights into cooling rates of iron meteorites (abstract)Meteoritics amp Planetary Science 41A26

Buchwald V F 1977 The mineralogy of iron meteoritesPhilosophical Transactions of the Royal Society of London A286453ndash491

Chabot N L and Drake M J 2000 Crystallization of magmatic ironmeteorites The effects of phosphorus and liquid immiscibilityMeteoritics amp Planetary Science 35807ndash816

Chabot N L and Haack H 2006 Evolution of asteroidal cores InMeteorites and the early solar system II edited by Lauretta D Sand McSween H Y Jr Tucson Arizona The University ofArizona Press pp 747ndash771

Cook D L Wadhwa M Janney P E Dauphas N Clayton R Nand Davis A M 2006 High-precision measurements of nickelisotopes in metallic samples via multi-collector ICPMSAnalytical Chemistry 788477ndash8484

Dauphas N 2007 Diffusion-driven kinetic isotope effect of Fe andNi during formation of Widmanstaumltten pattern Meteoritics ampPlanetary Science 421597ndash1613

Dauphas N Janney P E Mendybaev R A Wadhwa M RichterF M Davis A M Hines R and Foley C N 2004Chromatographic separation and multicollection-ICPMSanalysis of iron Investigating mass-dependent and -independentisotope effects Analytical Chemistry 765855ndash5863

Dauphas N and Rouxel O 2006 Mass spectrometry and natural

variations of iron isotopes Mass Spectrometry Reviews 25515ndash550

Davis A M and Brownlee D E 1993 Iron and nickel isotopic massfractionation in deep-sea spherules (abstract) 24th Lunar andPlanetary Science Conference pp 373ndash374

Engrand C McKeegan K D Leshin L A Herzog G FSchnabel C Nyquist L E and Brownlee D E 2005 Isotopiccompositions of oxygen iron chromium and nickel in cosmicspherules Toward a better comprehension of atmospheric entryheating effects Geochimica et Cosmochimica Acta 695365ndash5385

Goldstein J I and Ogilvie R E 1965 The growth of theWidmanstaumltten pattern in metallic meteorites Geochimica etCosmochimica Acta 29893ndash920

Herzog G F Hall G S and Brownlee D E 1994 Massfractionation of nickel isotopes in metallic spheres Geochimicaet Cosmochimica Acta 585319ndash5323

Hopfe W D and Goldstein J I 2001 The metallographic coolingrate method revised Application to iron meteorites andmesosiderites Meteoritics amp Planetary Science 36135ndash154

Horn I Blanckenburg F von Schoenberg R Steinhoefel G andMarkl G 2006 In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal oreformation processes Geochimica et Cosmochimica Acta 703677ndash3688

Jochum K P Seufert M and Begemann F 1980 On the distributionof major and trace elements between metal and phosphide phasesof some iron meteorites Zeitschrift fuumlr Naturforschung 35a57ndash63

Kehm K Hauri E H Alexander C M OrsquoD and Carlson R W2003 High-precision iron isotope measurements of meteoriticmaterial by cold plasma ICP-MS Geochimica et CosmochimicaActa 672879ndash2891

Kracher A Willis J and Wasson J T 1980 Chemical classificationof iron meteoritesmdashIX A new group (IIF) revision of IAB andIIICD and data on 57 additional irons Geochimica etCosmochimica Acta 44773ndash787

Malvin D J Wang D and Wasson J T 1984 Chemicalclassification of iron meteoritesmdashX Multielement studies of 43irons resolution of group IIIE from IIIAB and evaluation of Cuas a taxonomic parameter Geochimica et Cosmochimica Acta48785ndash804

Moynier F Teacutelouk P Blichert-Toft J and Albaregravede F 2004 Theisotope geochemistry of nickel in chondrites and iron meteorites(abstract 1286) 35th Lunar and Planetary Science ConferenceCD-ROM

Moynier F Blichert-Toft J Teacutelouk P and Albaregravede F 2005Excesses of 60Ni in chondrites and iron meteorites(abstract 1593) 36th Lunar and Planetary Science ConferenceCD-ROM

Mullane E Russell S S and Gounelle M 2005 Fractionation ofiron isotopes during magmatic processing on the IIIAB parentbody (abstract) Meteoritics amp Planetary Science 40A108

Poitrasson F Halliday A N Lee D-C Levasseur S andTeutsch N 2004 Iron isotope differences between Earth MoonMars and Vesta as possible records of contrasted accretionmechanisms Earth and Planetary Science Letters 223253ndash266

Poitrasson F Levasseur S and Teutsch N 2005 Significance ofiron isotope mineral fractionation in pallasites and ironmeteorites for the core-mantle differentiation of terrestrialplanets Earth and Planetary Science Letters 234151ndash164

Richter F M Davis A M DePaolo D J and Watson B E 2003Isotope fractionation by chemical diffusion between moltenbasalt and rhyolite Geochimica et Cosmochimica Acta 673905ndash3923

Mass-dependent fractionation of nickel isotopes in meteoritic metal 2077

Roskosz M Luais B Watson H C Toplis M J AlexanderC M Orsquo D and Mysen B O 2006 Experimental quantificationof the fractionation of Fe isotopes during metal segregation froma silicate melt Earth and Planetary Science Letters 248851ndash867

Schoenberg R and Blanckenburg F von 2006 Modes of planetary-scale Fe isotope fractionation Earth and Planetary ScienceLetters 252342ndash359

Schuessler J A Schoenberg R Behrens H and BlanckenburgF von 2007 The experimental calibration of the iron isotopefractionation factor between pyrrhotite and peralkaline rhyoliticmelt Geochimica et Cosmochimica Acta 71417ndash433

Scott E R D and Wasson J T 1976 Chemical classification of ironmeteoritesmdashVIII Groups IC IIE IIIF and 97 other ironsGeochimica et Cosmochimica Acta 40103ndash115

Scott E R D Wasson J T and Buchwald V F 1973 Chemicalclassification of iron meteoritesmdashVII A re-investigation of ironswith Ge concentrations between 25 and 80 ppm Geochimica etCosmochimica Acta 371957ndash1983

Ulff-Moslashller F 1998 Effects of liquid immiscibility on trace elementfractionation in magmatic iron meteorites A case study of groupIIIAB Meteoritics amp Planetary Science 33207ndash220

Voumllkening J and Papanastassiou D A 1989 Iron isotope anomalyThe Astrophysical Journal 347L43ndashL46

Voumllkening J and Papanastassiou D A 1990 Zinc isotope anomalyThe Astrophysical Journal 358L29ndashL32

Wasson J T Xinwei O Wang J and Jerde E 1989 Chemical

classification of iron meteorites XI Multi-element studies of 38new irons and the high abundance of ungrouped irons fromAntarctica Geochimica et Cosmochimica Acta 53735ndash744

Wasson J T Choi B-G Jerde E A and Ulff-Moslashller F 1998Chemical classification of iron meteorites XII New members ofthe magmatic groups Geochimica et Cosmochimica Acta 62715ndash724

Weyer S Anbar A D Brey G P Muumlnker C Mezger K andWoodland A B 2005 Iron isotope fractionation duringplanetary differentiation Earth and Planetary Science Letters240151ndash164

Williams H M Markowski A Quitteacute G Halliday A N TeutschN and Levasseur S 2006 Fe isotope fractionation in ironmeteorites New insights into metal-sulphide segregation andplanetary accretion Earth and Planetary Science Letters 250486ndash500

Wood J A 1964 The cooling rates and parent planets of several ironmeteorites Icarus 3429ndash459

Zhu X K Guo Y OrsquoNions R K Young E D and Ash R D 2001Isotopic homogeneity of iron in the early solar nebula Nature412311ndash313

Zhu X K Guo Y Williams R J P OrsquoNions R K Matthews ABelshaw N S Canters G W de Waal E C Weser U BurgessB K and Salvato B 2002 Mass fractionation processes oftransition metal isotopes Earth and Planetary Science Letters20047ndash62