Lithium isotope compositions of Martian and lunar reservoirs

Hans-Michael Seitz a,b,⁎, Gerhard P. Brey a, Stefan Weyer a, Soodabeh Durali a,Ulrich Ott b, Carsten Münker c,d, Klaus Mezger d

a Institut für Mineralogie, Universität Frankfurt, Senckenberganlage 28, 60054 Frankfurt, Germanyb Max-Planck-Institut für Chemie, Postfach 30 60, 55020 Mainz, Germany

c Mineralogisch-Petrologisches Institut, Universität Bonn, Poppelsdorfer Schloss, 53115 Bonn, Germanyd Institut für Mineralogie, Universität Münster, Correnstrasse 24, 48149 Münster, Germany

Received 2 September 2005; received in revised form 3 March 2006; accepted 3 March 2006Available online 19 April 2006

Editor: V. Courtillot

Abstract

Lithium isotope compositions and concentrations of 12 lunar samples (including two high-Ti, three low-Ti mare basalts, fivehighland breccias, one orange and one green glass) and 7 Martian meteorites (three basaltic and one lherzolitic shergottite, twonakhlites, and the orthopyroxenite ALHA 84001 were measured using MC-ICP–MS. Most of the Martian samples have a narrowrange of δ7Li (+3.6 to +5.2‰). Only ALHA 84001 is isotopically lighter, with δ7Li=−0.6‰. The range in Li concentrations islimited and all shergottites have identical Li concentrations (1.8–2.1μg/g) and isotope compositions within the error. Despite alarger variation in Li concentrations (5–49μg/g), Li isotope variation of most lunar samples is also very limited (+3.5 to +6.6‰)with an average of +5.2‰ (±1.2, 2σ). The only exception is one KREEP-rich highland breccia (15445a), which has a δ7Li value of +18.6‰. Consequently, the majority of lunar and Martian samples have an isotopic signature similar to the Earth's mantle (MORBand OIB). These results imply a homogeneous Li isotope composition of the inner solar system with a δ7Li≈+4‰. The resultsfurther indicate that planetary silicate differentiation by partial melting on planets under either wet or dry conditions does notsignificantly fractionate Li isotope compositions. Lithium abundances of lunar basalts and glasses are similar to those of terrestrialbasalts. In contrast, Martian basalts have generally lower Li concentrations, more similar to BSE, although the concentrations inshergottitic clinopyroxenes and nakhlitic pyroxenites do not indicate a lower Li abundance for bulk Mars. These systematics implythat the Martian basalts were depleted in Li by a process that did not fractionate the Li isotope composition.© 2006 Elsevier B.V. All rights reserved.

Keywords: lithium isotopes; moon; Mars; SNC meteorites; mare basalts; KREEP

1. Introduction

The early accretion history, the evolution of theterrestrial planets and planetesimals and the bulkcomposition of the solar system is poorly understood

with respect to its Li isotope composition predominantlydue to a lack of data. Published Li isotope data forextraterrestrial material are limited to the C1 chondriteOrgueil (δ7Li=+3.9‰ [1]), 6 carbonaceous chondrites(δ7Li=+3.0‰ to −3.5‰ [2]), a few chondrules andcalcium–aluminium-rich inclusions (CAIs) [3,4], 2 bulknakhlite samples and their constituent olivines andclinopyroxenes (+5.1‰ and +3.7‰ [5]) and SIMSstudies on SNC meteorites [6,7] that show pronounced

Earth and Planetary Science Letters 245 (2006) 6–18www.elsevier.com/locate/epsl

⁎ Corresponding author. Tel.: +49 69 798 22547; fax: +49 69 79828066.

E-mail address: h.m.seitz@em.uni-frankfurt.de (H.-M. Seitz).

0012-821X/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.epsl.2006.03.007

variations between and within grains (e.g. core to rimvariation in shergottitic pyroxene from −17 to +10‰)[7]. Brooker et al. [8] suggested that due to themoderately high condensation temperature of Li (ca.950°C [9]) and the lack of isotope fractionation of themore volatile element K, isotope fractionation duringplanetary accretion did not occur for Li. Likewise,elements such as Fe and Mg with similarly highcondensation temperatures (∼1060°C [9]) did notfractionate during accretion [10,11]. It is thereforeexpected that terrestrial planets should display a δ7Lisignature close to carbonaceous chondrites. These showa range of +3.9 to −3.5‰ with an average of ca. 0‰.McDonough et al. [2] suggested that this valuecorresponds to that of the bulk solar system.

OIB and MORB e.g. [12] consistently give δ7Livalues around +4‰, the value reported for Orgueil [1].This observation, coupled with mineral separate com-positions (ol, opx, cpx) from relatively fertile spinelperidotites from rift settings led [13] to estimate the bulksilicate Earth (BSE) to have a δ7Li value around +4‰.This estimated value is in accord with [12] whosuggested a similar mantle value based on the MORBisotope compositions.

The differences in the history of planetary accretion,growth and subsequent differentiation with their distinctphysical, chemical and petrological characteristics mayhave led to distinct Li isotope compositions of the bulkplanetary bodies and their differentiated reservoirs. Forexample, Earth, Mars and the Moon are very differentwith respect to their water contents: the Moon has nohydrosphere, there is copious water on Earth, and sinceits formation, Mars has almost completely lost its water.Plate tectonic processes are active on Earth but absenton Mars and the Moon. Mixing and homogenisation ofthe Martian or lunar mantle is solely dependent onthermal convection and is therefore less effective thanon Earth, where oceanic crust and lithosphere is beingrecycled into the mantle. Radiogenic isotopes in Martianmeteorites are rather heterogeneous, suggesting that theMartian mantle has developed distinct geochemicalreservoirs that could not be subsequently homogenised[14]. Despite these significant differences, the Martianand lunar mantles are similar to the Earth's mantle inmany aspects. However, the Moon is more enriched inrefractory elements such as Mg, Ti and Al and depletedin siderophile, chalcophile and volatile elements, suchas Fe, Ni, Na, K, S and H2O (e.g. [15,16]). This is shownby the compositions of early lunar magmas, which arevirtually anhydrous. In SNC meteorites the ratios ofvolatile to refractory elements (e.g. K/La) are higherrelative to the Earth and oxygen isotope compositions

are different. Based on the model of [17], the Marsmantle has chondritic abundances of the refractorylithophile elements such as Al, Mg, Mn and Cr, whileNa and other volatile lithophile elements are depleted.Some stable isotopes indicate that the mantle reservoirsare distinctively different from the outgassed Martianatmosphere [14]. Martian meteorites are remarkably dryexcept for hydrous phases in rare melt inclusions inminerals. However, according to experimental studies atmedium pressures [18], crystallisation of clinopyrox-enes in shergottite magmas must have commenced atH2O-saturated conditions (∼1.8 wt.% H2O), withsubsequent degassing at low pressures. Such a modelwas used by Beck et al. [7] to explain the dramatic coreto rim zoning of Li isotopes in single pyroxenes from ashergottite from −17 to +10‰ at overall constant Liconcentrations. This model follows [19] and is taken upagain by Herd et al. [20] on the basis of no zoning orcore to rim depletion of light lithophile elements inshergottites. Bridges et al. [5] measured Li isotopes onbulk samples for the nakhlites Lafayette and Nakhla andobtained δ7Li values of +4.1‰ and +4.5‰, respec-tively. Pyroxene and olivine rich portions of Nakhlagave δ7Li values of +3.7‰ and +5.1‰, respectively.These values agree with the value for Orgueil and theestimated bulk silicate Earth.

We have analysed Li isotopes in most available SNCmeteorites and in a number of representative types oflunar samples to evaluate (1) whether the Martian andlunar silicate reservoirs are different from the Earth, (2)whether the different water contents in the differentbodies led to Li isotope fractionation and (3) to whatextent the planetary isotope signatures reflect the bulksolar system composition.

2. Methods

Samples were aliquots from powdered material(>100mg), which had been investigated in previousstudies on Nb/Ta [21] and Fe isotopes [10].

Rock digestion and column chemistry followed theprocedure of Arnold et al. [22] and Seitz et al. [13].Aliquots from dissolved samples from which Fe wasalready separated by ion exchange chromatographywere used. During this first chromatographic step for Feseparation, Li was collected together with the bulkmatrix of the sample. Lithium loss using this procedurewas virtually impossible, as the entire eluted fractionprior to the Fe cut (including the loaded sample matrix)was collected. To ensure that no Li resided on thecolumn, an olivine and a clinopyroxene matrix wasdoped with 10μg of Li. The Fe fraction as well as a

7H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

subsequent column wash fraction (of 5cv7M HCl+5cv0.5M HCl, cv=column volume) were analysedfor Li, but the combined amount of Li in all thesefractions was <0.01%. The same check (for Li in the Feeluent+column wash) was carried out for several lunarsamples (without Li doping). However, Li concentra-tions were below detection in all cases.

Rock digestion of replicate dissolutions (ALHA84001 and 15445) followed the procedure of [13].Powdered rock samples (10 and 5mg, respectively)were digested in a mixture of 1ml conc. HNO3 and 1mlconc. HF on a hot plate (140°C) for several days.Subsequently, samples were dissolved in 6M HCl andfinally reconstituted in 6M HNO3 prior to chromato-graphic Li purification (see [13] for more detail). For Lichromatography, clear sample solutions of 0.18ml 5MHNO3 and 0.72ml 100% methanol (analytical grade)were passed through a single cation exchange column.

Small 1.4-ml exchange columns with a resin heightof 6cm filled with BioRad AG50W-X8 (200–400 mesh)were used. Geometry and a small column size allow amaximum sample size of 50mg; however, 3–10mg wassufficient for the lunar samples and SNCmeteorites. Thecolumn was calibrated using a large variety of rockstandards as well as matrices that closely match thesamples. It was discovered, that even large composi-tional differences (e.g. 5 to 75wt/% MgO) cause onlyminor shifts in elution curves ([13] and Seitz unpub-lished data). Lithium peaks are always tight (ca. 6ml)and were well separated from the Na peak. With aneluted cut of 10ml elution, all lithium is recovered. Toasses the accuracy and consistency of the dissolutionand Li purification procedure, the JB-2 basalt standardand a blank were measured with each set of samples. Toensure 100% recovery, random checks were performedon fractions eluted prior to and after the Li cuts. ExcessLi was never detected and the concentrations alwaysremained within the analytical blank. For measurementsby MC-ICP–MS the samples were taken up in 2%HNO3 (10ng/g solution).

The ThermoFinnigan Multiple-Collector InductivelyCoupled Plasma Mass Spectrometer (MC-ICP–MS) inFrankfurt allows static measurements of both 6Li and 7Li.Measurements were performed with dry plasma condi-tions using a Cetac Aridus® nebuliser fitted with a PFAspray chamber and an ESI microconcentric nebuliserwith an uptake rate of 20μl/min. The sample gas is driedat 160°C before being introduced into the plasma. Withthe Finnigan standard cones (H-Cones) an intensity of40–50pA (=4–5V using 1011Ω resistor) for 7Li at a10ppb concentration level is achieved. The high ionyields guarantee high precision isotope measurements

on sample amounts as small as 2–3ng Li (e.g. 3mgsample weights with less than 1μg/g Li. The analyticalblank (chemistry blank minus background signal ondouble distilled 2%HNO3) was usually 10–15pg (≈4–6mV on 7Li). Sample analysis was carried outsequentially by ‘bracketing’ the sample with the L-SVEC standard [23]. Sample solutions were diluted tomatch the intensities of the 10ng/g L-SVEC standardsolution. All samples were measured at least 3times;each measurement consisted of 4blocks with 5cyclesper block. The total integration time for each Li isotopemeasurement was ≈4min, followed by an electronicbaseline measurement (at masses 5.9, 6.1 and 6.9, 7.1,respectively) of ≈1min. Isotope compositions areexpressed as ‰ deviations from the NIST L-SVECstandard: δ7Li=[(7Li/6Li)sample / (

7Li/6Li)L-SVEC standard−1]⁎1000. The internal precision is typically between ±0.2 and 0.6‰ (2S.E.). The best estimate for externalprecision is long term reproducibility, determined onreplicate dissolutions of the basalt standard JB-2, whichis about ±1.2‰ (2S.D.). The average external repro-ducibility on replicate analyses of meteorite samples isworse in some cases, as these samples (e.g. the KREEP-rich highland breccia 15445) can be very heterogeneous.The precision of replicate measurements of the samealiquot is <1‰ (2S.D.) in most cases. The accuracy ofthe Li isotope measurements is documented by theanalyses of rock standards (JB-2 and JGB-1) andmeteorite samples which have previously been mea-sured by other laboratories [e.g. our measurements ofOrgueil yield a δ7Li value of +3.4‰ ±0.7‰ (2S.D.)and 1.56μg/g Li ±0.25 (2S.D.), compared to +3.9‰ ±1.2‰ (2S.D.) and 1.49μg/g [1], also see Table 1a)]. Theδ7Li determined in this study agrees with literaturevalues to within 1.2‰.

Lithium concentrations of samples were determinedalong with the isotope measurements by comparing theion beam intensities with those of the 10μg/g NIST L-SVEC standard solution. The uncertainty of this methodwas estimated by measuring the international rockstandards JB-2 and JGB-1 several times during a dailymeasurement campaign. The daily precision of theseconcentration measurements was typically 10% (2S.D.).The long-term reproducibility of the JB-2 basaltstandard was 15% (2S.D.).

3. Samples

3.1. Mars

The SNC meteorites are generally accepted to bederived from the planet Mars [24–26]. Several studies

8 H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

Table 1Lithium concentration and isotope compositions of SNC (a) meteorites and (b) lunar samples

Sample Rock type Material Run Li (μg/g) 2σ δ7Li 2σ n Method Source

(a)Zagami Basaltic shergottite Bulk 34 2.1 0.04 4.4 0.5 3 MC-ICP–MS This study

Maskelynite 2.5 SIMS [6]Pigeonite 3.3 to −12.6 SIMS [6]Augite 6.5 SIMS [6]Bulk 3.8 ID [72]Bulk 2.9 2.60 Compilation [43]Augite 2.4 to 5.2 SIMS [20]Maskelynite 2.1 to 4.5 SIMS [20]Phosphate 2.6 to 2.8 SIMS [20]Oxides 4.3 SIMS [20]Augite 3.1 to 8.0 SIMS [19]

Shergotty Basaltic shergottite Bulk 35 2.0 0.4 5.2 0.5 3 MC-ICP–MS This studyMaskelynite 4.8 to −17.6 SIMS [6]Pigeonite 3.9 to −11.5 SIMS [6]Augite 2.2 SIMS [6]Bulk 4.1 ID [72]Bulk 3.3 ES [73]Bulk 4.5 1.8 Compilation [43]Bulk 5.6 INNA [70]Augite 1.9 to 4.8 SIMS [20]Maskelynite 1.4 to 2.3 SIMS [20]Phosphate 1.1 to 1.4 SIMS [20]Oxides 1.4 to 1.9 SIMS [20]Augite 1.7 to 8.1 SIMS [19]

EETA 79001 A Depleted basalt Bulk 33 1.8 0.05 3.6 0.8 3 MC-ICP–MS This studyEETA 79001 Maskelynite 2.5 SIMS [6]

Pigeonite 3.8 to −18.9 SIMS [6]Bulk 4.5 Compilation [43]Bulk 1.5 ICP–MS [44]Bulk 1.7 ICP–MS [44]

ALHA 77005 Lherzolite Bulk 34 1.9 0.1 4.8 0.4 3 MC-ICP–MS This studyBulk 1.6 ID [74]Bulk 1.3 INNA [71]Bulk 1.5 0.4 Compilation [43]

Lafayette Clinopyroxenite Bulk 34 12.2 5.0 0.7 3 MC-ICP–MS This studyBulk 4.1 0.6 MC-ICP–MS [5]Bulk 3.9 ID [45]Augite 4.5 to 6.7 SIMS [19]

Nakhla Clinopyroxenite Bulk 35 6.3 0.6 4.1 1.0 3 MC-ICP–MS This studyBulk 4.5 0.4 MC-ICP–MS [5]Augite 3.7 0.6 MC-ICP–MS [5]Ol-rich 5.1 0.6 MC-ICP–MS [5]Bulk 3.9 Compilation [43]Bulk 3.8 INNA [70]Augite 3.8 to 6.0 SIMS [19]

ALHA 84001 Orthopyroxenite Bulk 35 2.9 0.9 −0.6 0.5 5 MC-ICP–MS This studyBulk 50 3.0 0.20 1.2 1.4 4 MC-ICP–MS This studyOrthopyroxene 50 1.8 0.09 1.2 1.3 4 MC-ICP–MS This studyCarbonate 50 0.1 0.02 3.2 1.2 3 MC-ICP–MS This studyInterstitial 50 1.0 0.07 −0.5 1.3 3 MC-ICP–MS This studyMaskelynite −4.8 SIMS [6]Orthopyroxene 5.8 to −9.7 SIMS [6]Phosphate −5.4 SIMS [6]

(continued on next page)

9H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

(e.g. [14,19,27–31]) report on their mineralogy, major,minor and trace elements and isotope composition. Inthis study three basaltic (Shergotty, Zagami and EETA79001A) and one lherzolitic (ALHA 77005) shergot-tite, two augite-olivine cumulates (Nakhla and Lafay-ette) and the orthopyroxenite ALHA 84001 wereinvestigated. All samples except ALHA 84001(which is believed to have crystallised at ca. 4.5Gaago [32,33]), have relatively young crystallisation ages(1.3 to 0.175Ga) and are inferred to have come fromvolcanic flows of the Tharsis or Elysium plume [30].

The basalts (Shergotty, Zagami and EETA 79001A)are partial melts derived from different parts of theMartian mantle that crystallised almost entirely insurface lava flows. EETA 79001 consists of twolithologies (A and B) that appear to represent twodistinct magmas (e.g. [29], and references therein). Inthis study, we investigated EETA 79001A which alsocontains lherzolitic xenocrysts and xenoliths.

The cumulate rocks Lafayette and Nakhla crystal-lised at depth. Beside the cumulus phases clinopyr-oxene and olivine, they contain interstitial glass,mesostasis with lathy plagioclase and re-crystallised‘melt’ inclusions in olivine [34], documenting thepresence of trapped melt. In fact, Fe-rich pyroxenerims, which are commonly observed in Nakhlites,suggest interaction with more evolved liquids [35].

The lherzolite ALHA 77005 is a cumulate rock withtwo texturally and mineralogically distinct lithologies([34], and references therein). Nonetheless, bulk

chemical compositions of both lithologies are similar[36].

The orthopyroxenite ALHA 84001 is a coarse-grained cumulate. Its primary mineralogy (97% ortho-pyroxene, 2% chromite, ∼1% maskelynite and minoraccessories, e.g. [34]) has been modified by secondaryalteration overprint including carbonates (e.g. [29,37],and references therein).

3.2. Moon

Altogether 12 lunar samples were analysed: 3 low-Timare basalts (15495, 15555, 15475), 2 high-Ti marebasalts (75035 sample split 72153a and sample split72153b), 3 KREEP-rich and 1 KREEP-poor highlandbreccias (15445 sample a and b, 65015, 62235, 78155,respectively), a polymict highland breccia (14310), A-15 green glass (15426) and A-17 orange glass (74220).

The selected samples encompass a large variety ofigneous rock types and therefore differ greatly in theirmineralogy, composition, crystallisation age and theirmagmatic and metamorphic histories.

Mare basalts are classified according to their TiO2

contents into high-Ti mare basalts with 6–15wt.%TiO2 (3.55–3.85Ga old) and younger (3.15–3.45Ga)very low- to medium-Ti mare basalts with <6wt.%TiO2. The chemical, isotopic, mineralogical and agedifferences of these rocks cannot be the result ofdifferent degrees of partial melting from a commonmantle source [38]. They are considered to have

Table 1 (continued)

Sample Rock type Material Run Li (μg/g) 2σ δ7Li 2σ n Method Source

(b)14310 Polymict highland breccia Bulk 35 48.8 3.8 6.6 1.1 5 MC-ICP–MS This study15445a KREEP-rich highland breccia Bulk 35 13.8 1.5 18.4 1.2 5 MC-ICP–MS This study15445b Bulk 52 3.5 0.3 1.9 0.8 3 MC-ICP–MS This study15495 Low-Ti mare basalt Bulk 34 6.4 0.7 5.6 0.2 3 MC-ICP–MS This study15555 Low-Ti mare basalt Bulk 34 5.8 0.2 3.6 0.4 4 MC-ICP–MS This study

Bulk 6.4 [46]65015 KREEP-rich highland breccia Bulk 33 21.1 0.5 3.9 0.3 3 MC-ICP–MS This study72153a High-Ti mare basalt Bulk 43 6.3 0.2 5.4 1.5 7 MC-ICP–MS This study72153b High-Ti mare basalt Bulk 34 7.8 0.1 6.2 0.4 4 MC-ICP–MS This study75035 High-Ti mare basalt Bulk 34 9.2 0.1 6.6 0.4 3 MC-ICP–MS This study15475 Low-Ti mare basalt Bulk 38 5.0 1.2 4.2 0.2 3 MC-ICP–MS This study

Bulk 15.3 [46]62235 KREEP-rich highland breccia Bulk 38 16.7 3.2 5.3 0.2 3 MC-ICP–MS This study78155 KREEP-poor highland breccia Bulk 38 6.6 1.5 5.7 0.6 3 MC-ICP–MS This study74220 A-17 orange glass Bulk 43 4.6 0.3 3.5 1.4 6 MC-ICP–MS This study15426 A-15 green glass Bulk 38 8.8 1.4 5.8 0.5 3 MC-ICP–MS This study

Corresponding to ‘Run’ numbers ‘n’ represents the number of measurements on a given solution. Uncertainty is the 2σ (standard deviation). SIMS:secondary ion mass spectrometry, ID: isotope dilution, INNA: instrumental neutron activation, ES: emission spectroscopy. Data are taken from:[5,6,19,20,43,44,46,70–74].

10 H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

formed as partial melts of a heterogeneous layeredlunar mantle. This mantle formed from cumulatesderived from an early lunar magma ocean ([39], andreferences therein). Recent support for this modelcomes from the Fe isotope compositions of high- andlow-Ti mare basalts which are clearly distinguishable[10]. The picritic green (15426) and orange (74220)glasses are interpreted to represent primitive liquidsfrom the lunar mantle (e.g. [40]).

Highland breccias are extremely heterogeneous.Sample 15445, for example, contains a lithic clastassemblage of plutonic/metamorphic spinel troctolite,troctolite, norite and anorthosite in a matrix of melt-bounded small olivine, plagioclase and spinel clasts[41]. The melt, which is interpreted to be an impact meltproduced by the Imbrium impact event, has crystallisedolivine, plagioclase, pyroxene and opaque minerals[41]. The KREEP components in these impact meltrocks are thought to have formed early in the history ofthe Moon during the late solidification of the lunarmagma ocean. The polymict highland breccia 14310 isthought to have crystallised from a feldspathic impactmelt (e.g. [42]).

4. Results

4.1. Mars

The basaltic and lherzolitic shergottites depict anarrow range in their Li concentrations (1.8 to 2.1μg/g) as well as in their δ7Li (+3.6 to +5.2‰, Table 1aand Fig. 1a). The two clinopyroxenites Nakhla andLafayette have similar δ7Li values (+4.1 and +5.0‰)but much higher Li abundances (6.3 and 12.2μg/g,respectively). Our data agree well with the valuesdetermined by Bridges et al. [5] for nakhlites. Theseauthors reported δ7Li of +4.1‰ and +4.5‰ for bulkrock analyses of Lafayette and Nakhla and δ7Li of+3.7‰ and +5.1‰ for a pyroxene separate and anolivine rich portion of Nakhla. However, for twodifferent splits of the orthopyroxenite ALHA 84001,we obtained distinctly lighter isotopic signatures of−0.6 and +1.2‰ at somewhat higher Li contents (2.9and 2.5μg/g). An orthopyroxene separate of thissample gave a similar δ7Li of +0.5‰ at 1.8μg/g.Interstitial phases yielded a lighter value (δ7Li of−0.5‰ and 1μg/g Li), and a low Li (0.2μg/g)carbonate separate gave a heavier isotope signature(δ7Li of +3.2‰). The Li concentrations (2.1, 1.9 and1.8μg/g, respectively), determined in this study forZagami, ALHA 77005 and EETA 79001A agree wellwith the values given by Lodders [43] and Neal et al.

[44]. For Shergotty we obtained a lower (2.0μg/g) andfor Nakhla and Lafayette a higher (6.3 and 12.2μg/g)Li concentration compared to literature data [43,45].Published Li concentration data for some meteorites,however, display a considerable range, suggestingchemical heterogeneities between different splits ofthe same meteorite (see Table 1a).

4.2. Moon

The samples from the Moon show large variations inLi-contents (4.6–48.8μg/g) but a narrow range in δ7Lifrom +3.5 to +6.6‰, except for one KREEP-richhighland breccia (15445) which has a δ7Li value of +18.6‰. A second split of this inhomogeneous sample(both splits originate from a 130mg chip), however,gave a much lighter δ7Li value of +1.9‰ and aconsiderably lower Li abundance (3.5μg/g, Table 1band Fig. 1b). Low- and high-Ti mare basalts as well asthe orange and green glasses have Li concentrationsbetween 4.6 and 9.2μg/g. Highland breccias are morevariable and generally higher in their Li concentrationswhich may be related to the amount of assimilatedKREEP in the impact breccia. The KREEP-poor sample78155 contains only 6.6μg/g Li, whereas the KREEP-rich highland breccias, including the polymict breccia14310, always have greater than 10, up to 50μg/g Li.The Li concentration that we determined for the low-Timare basalt 15555 (5.8μg/g) agrees well with the 6.4μg/gvalue determined by G. Lofgren and E. Lofgren [46].However, our value for the low-Ti basalt 15475 of5.0μg/g is significantly lower than the value of 15.3μg/greported by G. Lofgren and E. Lofgren [46]. Thesubstantial differences of our two values for KREEP-rich highland breccia splits 15445a and b are most likelydue to microsampling of a heterogeneous sample.

5. Effects from cosmic ray spallation reactions

In terms of cosmic abundances Li is rare, with severalhigher-Z elements considerably more abundant that mayserve as target elements for spallation contributions bycosmic rays. Nevertheless, it is necessary to establish thesize of possible cosmic ray effects on the Li isotopiccomposition. The most critical effect is the productionof 6Li. This is because 7Li is ∼12times more abundantthan 6Li, so that with both isotopes being produced byspallation in equal proportions the net effect will be adecrease of the 7Li/6Li isotopic ratio. For a semi-quantitative treatment, the production of 6Li can becompared to that of 3He, which allows an estimate basedon measured noble gas isotope compositions.

11H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

Unlike for the noble gases and radioactive nuclides(10Be, 26Al, etc.) that are used in determining cosmic rayexposure ages (e.g. [47,48]), there is little hardexperimental data on production of the stable Liisotopes. Existing data is mostly confined to carbonand oxygen, due to the commonly accepted belief thatmuch of the Li, Be and B in the universe was created byspallation of the relatively abundant CNO elements inthe interstellar medium. Cross-sections for productionof 6Li on 12C ranging between 11 and 7mb wereobtained for protons with energies in the range 44 to550MeV by Bernas et al. [49]. Somewhat higher values

of 11.6 and 15.5mb were found for 150 and 600MeVprotons by Raisbeck et al. [50]. For production onoxygen similar values in the range 10 – 20mb werefound [49,51]. As Li isotopes are mostly produced as asmall fragment from much heavier nuclei, similar to Heisotopes, it seems reasonable to assume that theirproduction also depends only weakly on the targetelement. This can be compared to cross sections fordirect production of 3He that are on the order of 30mb[52]. There is production of similar amounts oftritium that contributes to the integrated production of3He [53] so that overall production of 6Li should be on

Fig. 1. δ7Li values of Martian (a) and lunar (b) rocks as a function of Li abundances. Data for bulk silicate Earth (BSE) and Orgueil C1 chondrite aretaken from [13] and [1], respectively.

12 H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

the order ∼1/4 that of 3He, at an uncertainty of a factorof 2.

Among the Martian meteorite sample, the cosmic rayeffect on Li isotopes must be largest for ALHA 84001which has the longest exposure age and relatively low Licontent. Applying the production rate estimate derivedabove, a concentration of cosmogenic 3He of∼25⁎10−8 cm3 STP/g e.g. [54], and a Li concentrationof 2.9μg/g (Table 1) implies an effect on the Li isotopicratio of ∼0.09‰, well within the analytical uncertainty.Effects for other Martian meteorites in Table 1a arelower by factors of ∼3 or more.

Some of the lunar samples have considerably longercosmic ray exposure; but these samples often also have asignificantly higher abundance of non-spallogenic Li.The polymict highland breccia 14310 has a multi-stageexposure history and a cosmogenic 3He content (aftercorrection for diffusion losses) of 294⁎10−8 cm3 STP/g[55]. Using the scaled production of cosmogenic 6Li andthe Li content of 48.8μg/g (Table 1a), the effect on δ7Liis on the order of 0.06‰. Because of its lower Licontent, the effect may be an order of magnitude higherfor low-Ti mare basalt 15555 with an exposure age of∼90Ma [56]. For the orange glass 74220 no cosmo-genic 3He was detected by Eugster et al. [57]. An upperlimit on its production of 20⁎10−8 cm3 STP/g can bederived from the measured concentration of cosmogenicNe, so the effect on Li should be <0.04‰. We are notaware of any relevant noble gas data for the other lunarsamples in Table 1a, but the close agreement in Liisotope composition between the individual samplessuggests that cosmogenic effects are also smaller thanthe analytical uncertainties for these samples as well.Regardless, the signature of the only outlier, highlandbreccia 15445a, cannot be explained by spallation, sinceit shows depletion rather than enrichment of 6Li.

6. Discussion

Apart from differences in their chemical and isotopecomposition, the planetary bodies Earth, Mars and theMoon also differ in the duration of their volcanicactivity. While volcanism on Earth is an ongoingprocess, the major volcanic activity on the Moonoccurred between 3 and 4.5Ga (e.g. [39], and referencestherein), and on Mars possibly up to 180Ma ago (it isnot clear whether Mars still has active volcanism, e.g.[29]). The role of water during volcanism is fundamen-tally different on Earth, Mars and the Moon. Basalticmagmas on Earth are generally low in volatiles (0.2wt.% H2O) but are capable of holding several percent ofwater vapour before they become saturated. The

relatively low water content in basalts is a measure ofH2O in the source region. In comparison, basalticMartian meteorites have significantly lower waterconcentrations (e.g. Shergotty 280μg/g±120 andNakhla 570μg/g±120 [14]) and lunar mare basalts areabsolutely anhydrous. Another difference between theseplanetary bodies is the distinct Fe content of their silicatemantle. While BSE is relatively Fe-poor (Mg# 80–90),the mantles of Moon and Mars are more Fe-rich, withbulk Mg values of 75–85 and 65–78, respectively [58].If the age of volcanic activity, the bulk chemistry and thewater content of the source regions and the subsequentdegassing of basaltic magmas are crucial for thefractionation of Li isotopes, the isotope signature ofvarious mantle melts from Earth, Moon and Mars shouldbe systematically different.

Lithium isotope signatures of Martian meteorites areshown in Fig. 1a as a function of their Li-abundances.The values for bulk silicate Earth (BSE [13] and OrgueilC1 chondrite [1]) are plotted for comparison. Thebasaltic and lherzolitic shergottites have very similar Liisotope signatures to BSE and C1 (Orgueil) at slightlyhigher Li abundances (by 0.5μg/g). These abundances,however, are distinctively lower (by a factor of 3–4)when compared to their basaltic analogues from Earth(e.g. fresh MORB). Modelling Li abundances of thebasaltic shergottite sources, by using the available Liconcentrations for basaltic shergottites of this study andliterature values and applying partition coefficient ofDcpx/l =0.11–0.18 [59], gives concentrations that aremarkedly lower (260–810ppb) than the estimatedabundances of the Earth mantle [13]. These resultscould either mean that the Mars mantle sources aredepleted in Li relative to the terrestrial mantle [60,61] orthat the Martian basalts have lost a considerable amountof their Li. The latter model was evaluated by severalstudies. An experimental study by Dann et al. [18]suggested that the crystallisation of clinopyroxenes inshergotitte magmas began at H2O-saturated conditions(∼1.8wt.% H2O) with subsequent degassing at lowpressures. Based on mineral zoning patterns of lightlithophile elements and Li-depletion in pyroxene rimsthis model was also favoured by [7,19,20]. Beck et al.[7] used the magma degassing model to explain thedramatic core to rim zoning of Li isotopes (δ7Li from−17 to +10‰) in single pyroxenes from a shergottite atoverall constant Li-concentrations. The low Li concen-trations we measured in the basaltic lherzolites would, inprinciple, be consistent with such a model of Lidepletion by degassing. If so, the unfractionated Liisotope data would imply that the Li fractionationbetween a basaltic melt and a fluid phase causes no

13H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

detectable Li isotope fractionation under Martiansurface conditions. However, this model contradictsthe experimental results of Wunder et al. [62] who foundmineral-fluid Li isotope fractionation to occur and to behighly temperature dependent.

The clinopyroxenites (augite–olivine cumulates)Nakhla and Lafayette have Li isotope signatures thatare virtually identical with those of the basalticshergottites, but they have much higher Li concentra-tions. These high concentrations are consistent withthose measured in pyroxenes from Nakhla (3.5μg/g coreand 6.5μg/g rim) by Beck et al. [63]. Melts coexistingwith these cumulates must have had Li-abundances of atleast 17 to 33μg/g (applying Dcpx/l =0.11–0.18 [59]).Thus, if considering their present Li contents, Nakhlaand Lafayette could not have been in equilibrium withthe basaltic shergottites. It is notable, however, that bothcumulate rocks display trapped inter-cumulus melt, nowpresent as interstitial glass or a crystalline mesostasiswith lathy plagioclase. Re-crystallised ‘melt’ inclusionsalso occur in olivine [34]. Thus, the Li abundances inNakhla and Lafayette may be controlled by trappedmelts. The groundmass of Nakhla is enriched in Li(10μg/g [63]) which is too low to be in equilibrium withthe pyroxene rims. In a study by Szymanski et al. [35],Fe-rich pyroxene rims commonly observed in Nakhliteswere interpreted as being caused by late reactions withnon-parent Fe-rich liquids, directly before eruption.

Similar processes are likely to have occurred in thelherzolitic shergottite ALHA 77005. This sample is agabbroic cumulate with a Li abundance of 1.9μg/gsimilar to the concentrations in the basaltic shergottites.Lundberg et al. [64] suggested that the compositionaldifferences between single cumulus olivines (Fo75 toFo70 [65]) are the result of re-equilibration with theinter-cumulus liquid upon cooling.

Despite the differences in Li concentrations andcrystallisation histories, Li isotope variations of theMartian meteorites (except for ALHA 84001) aresurprisingly limited, with a mean δ7Li of 4.5‰ (±1.2,2σ). The similarity of δ7Li values for the basaltic andlherzolitic shergottites, as well as for the augite–olivinecumulates, suggests that neither different degrees ofpartial melting nor the formation of a clinopyroxenecumulate rock has led to isotope fractionation. This isin agreement with Tomascak et al. [66] who demon-strated the absence of Li isotopic fractionation inHawaiian basalts during magma differentiation. How-ever, distinctly lighter δ7Li values (−0.6‰ and 1.2‰)were obtained for two separate bulk samples of ALHA84001. In mineralogical terms, this ∼4.5Ga oldorthopyroxenite is very different from the other SNC

meteorites [29]. Interstitially it contains small patchesof olivine and clinopyroxene, maskelynite and phos-phate and accessory phases [6]. It also containssecondary carbonate minerals (∼3.9Ga old), whichprobably formed during impact metamorphism byreaction with hydrothermal or CO2-rich fluids (e.g.[33]). Shock metamorphism was followed by thermalmetamorphism, which led to textural annealing andmineral equilibration [29]. This sample is particularlyheterogeneous not only in its mineralogy but also in itsLi isotope composition. Interstitial phases, taken hereas a mineral mixture of what was left after orthopyr-oxene and carbonate were separated, yield a relativelylight isotope composition of δ7Li −0.5‰. Our resultsare in accord with the SIMS data of Chaussidon andRobert [6] who found that maskelynite and phosphateare isotopically light (δ7Li −4.8 and −5.4‰, respec-tively, see Table 1a for comparison). They also found alarge heterogeneity of δ7Li in orthopyroxene rangingfrom −9.7 to +5.8‰.

Seitz et al. [13] suggested that Li isotope fraction-ation between olivine, ortho- and clinopyroxene istemperature dependent (and decreases e.g. from 950°Cto 1150°C). The complex metamorphic and hydrother-mal history as well as the relatively low temperatureequilibration conditions of the orthopyroxenite ALHA84001 (two-pyroxene thermometry gives 875°C [67])has most likely resulted in a strong fractionation of Liisotopes.

As for the SNC meteorites, Li isotope composition ofthe lunar samples is fairly limited (Fig. 1b). Thereappears to be a weak trend in isotope compositionamong the lunar basalts: from orange glass (δ7Li+3.5‰) over low-Ti to high-Ti mare basalts (δ7Li+6.6‰), which may reflect different mantle sourcecompositions or isotopic fractionation during partialmelting or during magmatic differentiation. Overall,mare basalts have similar Li abundances (5.0–6.4μg/g)to MORB basalts, suggesting that Li concentrations ofthe lunar mantle are comparable to abundances in theEarth's mantle. The mean δ7Li of all lunar samples(except 15445a) is +4.9‰ (2.8, 2σ). This valueencompasses the estimated δ7Li value for BSE (+4‰[13] and with that of the Orgueil C1 chondrite +3.9‰[1]). The extremely heavy δ7Li value (+18.3‰) of theKREEP-rich highland breccia 15445a is somewhatsurprising as the Moon lacks a hydrosphere andanhydrous fractionation of Li isotopes to such heavyvalues has not been observed before. A second samplesplit of this non-homogeneous highland breccia gave amuch lighter isotopic signature (+1.9‰) and signifi-cantly lower Li concentrations (3.5μg/g). However,

14 H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

SIMS data by Barrat et al. [67] document the existenceof a heavy component in lunar meteorites. These authorsobtained values of up to +15‰ in melt inclusions inolivine from a lunar meteorite (NWA 479) andconcluded that the parental melt was enriched in 7Li.

Given the present database, we only can speculateon possible heterogeneities within samples, and as towhether the heavy isotopic signature reflects involvementof H2O or mantle source characteristics. There is scopeto look further into this problem.

Fig. 2. δ7Li values (a) and Li abundances (b) in Martian and lunar rocks are plotted against the Mg-value of the bulk rock [39,34, and referencestherein]. Fresh MORB data are taken from [68,12, and references therein]. Data for C1 mean are from [69]. Error bars are 2σ (mean).

15H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

In Fig. 2, δ7Li values of lunar samples and Martianmeteorites are plotted against the Mg number of theirbulk compositions. The latter reflects the major elementchemistry, magmatic differentiation processes and theoxidation state of a planet. With respect to their majorelement compositions, samples from the Moon andMars are similarly variable (Mg# 38 to 82) butvariations in δ7Li are very limited and unrelated toMg#, with the exception of one lunar and one Martiansample. The field for fresh MORB samples is shown forcomparison. With the exception of one lunar (15445)and one Martian sample (ALHA 84001), all samplesfrom both planetary bodies fall in the isotopic range offresh MORB δ7Li +1.5 to +6.8‰ ([12], and referencestherein). Most samples (Zagami, ALHA 77005, mostKREEP highland breccias and the green glass 15426)have Mg values similar to MORB. More distinctdifferences are apparent in the Li concentration versusMg# diagram of Fig. 2b. The basaltic shergottites,including the lherzolite (ALHA 77005) and theorthopyroxenite (ALHA 84001) have consistentlylower Li abundances than the lunar basalts andMORB, whereas high- and low-Ti mare basalts andorange and green glasses have MORB-like Liabundances.

7. Conclusions

Neither lunar samples nor SNC meteorites exhibit aclear correlation between their Mg values and Liabundances and Li isotopic signature, suggesting thatLi isotopes do not fractionate significantly duringmagmatic differentiation. A tendency of slightlyheavier Li in high-Ti-basalts compared to low-Ti-basalts is indicated suggesting a fractionation effect inthe formation of the different source rocks of the twolithologies. Taken together with the highland breccias(except for the KREEP-rich sample 15445a) the lunarsamples investigated here have an average δ7Li valuearound +4.9‰, similar to basaltic melts from theEarth's mantle, such as MORB and OIB. Since alllunar samples were generated from mantle sources thatwere formed in an early lunar magma ocean, thisimplies that magma ocean formation and crystallisationhad only a small effect on Li isotope composition.Lithium abundances and isotope signatures of thebasaltic and lherzolitic shergottites are also very similarto the values of C1 chondrites, fresh MORB and bulksilicate earth (BSE). The limited variations in Liisotopes for mantle melts and cumulates from Earth,Mars and the Moon, as well as their systematicdifferences in Li abundances, have implications for

their magmatic and planetary differentiation historiesand also for the behaviour of Li isotopes. The Liisotope composition appears to be largely independentof (1) planetary accretion and early differentiationhistory, (2) bulk chemistry of the terrestrial planets, and(3) water contents of the magmas and differentiationand degassing of basaltic melts.

This study implies that the terrestrial planets all havean unfractionated Li isotope signature of around δ7Li=+4‰. This value may furthermore reflect the Li isotopecomposition of the inner solar system.

Acknowledgements

This study was financially supported by the DeutscheForschungsgemeinschaft (Grant No. BR 1012/19-1).We profited much from discussions with Dmitri Ionov,Frank Brenker, Yann Lahaye, Jutta Zipfel, AndrewBerry and Margaret Hanrahan. For technical assistancewe thank Anna Karina Neumann, Thomas Kautz, andFranz Kneissl. Vincent Courtillot (editorial handling),Roberta Rudnick, Tim Elliott and an anonymousreviewer are thanked for their constructive reviews.

References

[1] R.H. James, M.R. Palmer, The lithium isotope composition ofinternational rock standards, Chemical Geology 166 (2000)319–326.

[2] W.F. McDonough, F.-Z. Teng, P.B. Tomascak, R.D. Ash, J.N.Grossman, R.L. Rudnick, Lithium isotopic composition ofchondritic meteorites, Lunar and Planetary Science XXXIV(2003) 1931.

[3] M. Chaussidon, F. Robert, K.D. McKeegan, Incorporation ofshort-lived 7Be in on CAI from the Allende meteorite, Lunar andPlanetary Science Conference, Houston, TX, XXXIII, 2002,Abstract #1563.

[4] M. Chaussidon, F. Robert, K.D. McKeegan, A.N. Krot, Lithiumand boron isotopic compositions of refractory inclusions fromprimitive chondrites: a record of irradiation in the early solarsystem, Lunar and Planetary Science XXXII (2001) (Abstract#1862).

[5] J.C. Bridges, R.H. James, V.K. Pearson, L. Baker, A.B.Verchovsky, I.P. Wright, Lithium and carbon isotopic fractiona-tions between the alteration assemblages of Nakhla andLafayette, Lunar and Planetary Science XXXVI (2005) 1758.

[6] M. Chaussidon, F. Robert, 7Li/6Li and 11B/10B ratios of SNCmeteorites, Lunar and Planetary Science Conference XXX(1999) 1592 (pdf).

[7] P. Beck, J.A. Barrat, M. Chaussidon, P. Gillet, M. Bohn, Liisotopic variations in single pyroxenes from the Northwest Africa480 shergottite (NWA 480): a record of degassing of Martianmagmas? Geochimica et Cosmochimica Acta 68 (2004)2925–2933.

[8] R.A. Brooker, R.H. James, J.D. Blundy, Li isotope systematics inZabargad peridotites: evidence of ancient subduction processesin the Red Sea mantle, Chemical Geology (2004) 179–204.

16 H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

[9] J.T. Wasson, Meteorites: Their Record of Early Solar SystemHistory, 1985. 267 pp.

[10] S. Weyer, A.D. Anbar, G.P. Brey, C. Müncker, K. Mezger, A.B.Woodland, Iron isotope fractionation during planetary differen-tiation, Earth and Planetary Science Letters 240 (2005)251–264.

[11] M. Bizzarro, J.A. Baker, H. Haack, M. Rosing, Mg isotope studyof terrestrial and extraterrestrial materials, Geochimica etCosmochimica Acta 67 (2003) A40–A40.

[12] P.B. Tomascak, Developments in the understanding andapplication of lithium isotopes in the Earth and Planetarysciences, in: C. Johnson, B. Beard, F. Albarède (Eds.),Geochemistry of Non-Traditional Stable Isotopes, MineralSociety of America, Washington, DC, 2004, pp. 153–195.

[13] H.-M. Seitz, G.P. Brey, Y. Lahaye, S. Durali, S. Weyer, Lithiumisotopic signatures of peridotite xenoliths and isotopic fraction-ation at high temperature between olivine and pyroxenes,Chemical Geology 212 (2004) 163–177.

[14] H.Y.J. McSween, Mars, in: Andrew M. Davis (Ed.), Treatiseon Geochemistry, Meteorites, Comets, and Planets, vol. 1,2003, pp. 601–621.

[15] R. Wolf, E. Anders, Moon and Earth—compositional differencesinferred from siderophiles, volatiles, and alkalis in basalts,Geochimica et Cosmochimica Acta 44 (1980) 2111–2124.

[16] S.R. Taylor, Planetary Science: A Lunar Perspective, 1982.512 pp.

[17] H. Wänke, G. Dreibus, Chemical composition and accretionhistory of terrestrial planets, Philosophical Transactions of theRoyal Society of London. A 325 (1988) 545–557.

[18] J.C. Dann, A.H. Holzheid, T.L. Grove, H.Y. McSween, Phaseequilibria of the Shergotty meteorite: constraints on pre-eruptivewater contents of Martian magmas and fractional crystallizationunder hydrous conditions, Meteoritics and Planetary Science 36(2001) 793–806.

[19] R.C.F. Lentz, H.Y.J. McSween, J. Ryan, L.R. Riciputi, Water inthe Martian magmas: clues from light lithophile elements inshergottite and nakhlite pyroxenes, Geochimica et Cosmochi-mica Acta 65 (2001).

[20] C.D.K. Herd, A.H. Treiman, G.A. McKay, C.K. Shearer, Lightlithophile elements in Martian basalts: evaluating the evidencefor magmatic water degassing, Geochimica et CosmochimicaActa 699 (2005) 2431–2440.

[21] C. Münker, J. Pfänder, S. Weyer, A. Büchl, T. Kleine, K. Mezger,Evolution of planetary cores and the Earth–Moon system fromNb/Ta systematics, Science 301 (2003) 84–87.

[22] G.L. Arnold, S. Weyer, A.D. Anbar, Fe isotope variation innatural materials measured using high mass resolution MC–ICPMS, Analytical Chemistry 76 (2004) 322–327.

[23] G. Flesch, A.R. Anderson, H.J. Svec, A secondary isotopicstandard for 6Li/7Li determinations, International Journal of MassSpectrometry and Ion Physics 12 (1973) 265–272.

[24] J.T. Wasson, G.W. Wetherill, Dynamical, chemical, and isotopicevidence regarding the formation locations of asteroids andmeteorites, in: T. Gehrels (Ed.), Asteroids, Univ. Arizona Press,Tuscon, 1979, pp. 926–974.

[25] D.D. Bogard, P. Johnson, Martian gases in an Antarcticmeteorite, Science 221 (1983) 651–654.

[26] H.Y.J. McSween, E.M. Taylor, E.M. Stolper, Allan Hills 77005: anew meteorite type found in Antarctica, Science 204 (1979)1201–1203.

[27] L.E. Borg, L.E. Nyquist, H. Wiesmann, Y. Reese, Constraints onthe petrogenesis of Martian meteorites from the Rb–Sr and Sm–

Nd isotopic systematics of the lherzolitic shergottites ALH77005and LEW88516, Geochimica et Cosmochimica Acta 66 (2002)2037–2053.

[28] T. Kleine, K. Mezger, C. Münker, H. Palme, A. Bischoff,182Hf–182W isotope systematics of chondrites, eucrites, andMartian meteorites: chronology of core formation and earlymantle differentiation in Vesta and Mars, Geochimica etCosmochimica Acta 68 (2004) 2935–2946.

[29] H.Y. McSween, A.H. Treiman, Martian meteorites, in: J.J. Papike(Ed.), Planetary Materials, Reviews in Mineralogy, vol. 36,Mineralogical Society America, 1998, pp. 6-1–6-53.

[30] H.Y.J. McSween, The rocks of Mars, from far and near,Meteoritics and Planetary Science 37 (2002) 7–25.

[31] D.W. Mittlefehldt, M.M. Lindstrom, Petrology and geochemistryof Martian meteorites EETA79001 and ALHA77005, Lunar andPlanetary Science XXX (1999) (abs) #1817.

[32] L.E. Nyquist, B. Bansal, H. Weissmann, C.-Y. Shih, “Martians”young and old: Zagami and ALHA84001 (abstr), Lunar andPlanetary Science XXVI (1995) 106–1066.

[33] L.E. Nyquist, D.D. Bogard, C.-Y. Shih, A. Greshake, D. Stöffler,O. Eugster, Ages and geological histories of Martian meteorites,Chronology and Evolution of Mars 96 (2001) 105–164.

[34] C. Meyer, Mars Meteorite Compendium, 2003.[35] A. Szymanski, F. Brenker, H. Palme, A. El Goresy, Thermo-

barometry of Martian nakhlites and a lherzolite—a comparison,Meteoritics and Planetary Science 39 (2004) A102–A102.

[36] E. Jarosewich, Homogenized powders of Antarctic meteorites,Antarctic Meteorite Newsletter 13 (1990) 134.

[37] D.W. Mittlefehldt, ALH84001, a cumulate orthopyroxenitemember of the Martian meteorite clan, Meteoritics and PlanetaryScience (1994) 214–221.

[38] C. Meyer, The Lunar Petrographic Thin Section Set, 2003, p. 66.[39] J.J. Papike, G. Ryder, C.K. Shearer, Lunar samples, in: J.J.

Papike (Ed.), Planetary Materials, Reviews in Mineralogy,Mineralogical Society America, 1998, pp. 5-1–5-234.

[40] R.J. Walker, M.F. Horan, C.K. Shearer, J.J. Papike, Lowabundances of highly siderophile elements in the lunar mantle:evidence for prolonged late accretion, Earth and PlanetaryScience Letters 224 (2004) 399–413.

[41] G. Ryder, J.F. Bower, Petrology of Apollo 15 black-and-whiterocks 15445 and 15455—fragments of the Imbrium impact meltsheet? Proc. Lunar Planet. Sci. Conf. 8th, 1977, pp. 1895–1923.

[42] M.D. Norman, G. Ryder, A summary of the petrology andgeochemistry of pristine highlands rocks, Proc. Lunar Planet. Sci.Conf. 10th, 1979, pp. 531–559.

[43] K. Lodders, A survey of shergottite, nakhlite and Chassignymeteorites whole-rock compositions, Meteoritics and PlanetaryScience 33 (1998) A183–A190 (suppl.).

[44] C.R. Neal, L.A. Taylor, J.C. Ely, J.C. Jain, M.A. Nazarov,Detailed geochemistry of new shergottite, Dhofar 019, Lunar andPlanetary Science 32 (2001) (Abstract #1671).

[45] G. Dreibus, W. Huisl, B. Spettel, R. Haubold, Comparison of thechemistry of Y-000749 with other nakhlites, Lunar and PlanetaryScience XXXIV (2003) (Abstract #1586).

[46] G.E. Lofgren, E.M. Lofgren, Catalog of lunar mare basaltsgreater than 40grams: Part 1. Major and Trace Elements, LPIContribution, vol. 438, Lunar and Planetary Institute, Houston,1981.

[47] R. Michel, R. Bodemann, H. Busemann, R. Daunke, M. Gloris,H.-J. Lange, B. Klug, A. Krins, I. Leya, M. Lüpke, S. Neumann,H. Reinhard, M. Schnatz-Büttgen, U. Herpers, T. Schiekel, F.Sudbrock, B. Holmqvist, H. Condé, P. Malmborg, M. Suter, B.

17H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18

Dittrich-Hannen, P.-W. Kubik, H.-A. Synal, D. Filges, Crosssections from the production of residual nuclides by low- andmedium-energy protons from the target elements C, N, O, Mg,Al, Si, Ca, Ti, V, Mn, Fe, Co, Ni, Cu, Sr, Y, Zr, Nb, Ba and Au,Nuclear Instruments and Methods in Physics Research, B 129(1997) 153–193.

[48] I. Leya, H.-J. Lange, M. Lüpke, U. Neupert, R. Daunke, O.Fanenbruck, R. Michel, R. Rösel, B. Meltzow, T. Schiekel, F.Sudbrock, U. Herpers, D. Filges, G. Bonani, B. Dittrich-Hannen,M. Suter, P.W. Kubik, H.-A. Synal, Simulation of the interactionof galactic cosmic-ray protons with meteoroids: on theproduction of radionuclides in thick gabbro and iron targetsirradiated isotropically with 1.6GeV protons, Meteoritics andPlanetary Science 35 (2000) 287–318.

[49] R. Bernas, M. Epherre, E. Gradsztajn, R. Klapisch, F. Yiou,Mass-spectrometric determination of Li6 and Li7 production inC12 and O16 by high energy protons, Physics Letters 15 (1965)147–149.

[50] G.M. Raisbeck, J. Lestringuez, F. Yiuo, Cross sections for 6Liand 7Li production from the bombardment of 12C by 150- and600-MeV protons and 880-MeValpha particles, Physical ReviewC 6 (1972) 685–689.

[51] F. Yiou, M. Baril, J. Dufaure de Citres, P. Fontes, E. Gradstajn, R.Bernas, Mass-spectrometric measurement of lithium, beryllium,and boron isotopes produced in 16O by high-energy protons, andsome astrophysical implications, Physical Review 166 (1968)968–974.

[52] R. Michel, F. Pfeiffer, S. Theis, F. Begemann, H. Weber, P.Signer, R. Wieler, P. Cloth, P. Dragovitsch, D. Filges, P. Englert,Production of stable and radioactive nuclides in thick stonytargets (R=15 and 25cm) isototropically irradiated with600MeV protons and simulation of the production of cosmo-genic nuclides in meteorites, Nuclear Instruments andMethods inPhysics Research, B 42 (1989) 76–100.

[53] I. Leya, F. Begemann, H.W. Weber, R. Wieler, R. Michel,Simulation of the interaction of galactic cosmic ray protons withmeteoroids: on the production of 3H and light noble gas isotopesin isotropically irradiated thick gabbro and iron targets,Meteoritics and Planetary Science 39 (2004) 367–366.

[54] Y.N. Miura, K. Nagao, N. Sugiura, H. Sagawa, K. Matsubara,Orthopyroxenite ALH84001 and shergottite ALH77005: addi-tional evidence for a Martian origin from noble gases,Geochimica et Cosmochimica Acta 59 (1995) 2105–2113.

[55] O. Eugster, P. Eberhardt, J. Geiss, N. Grögler, M. Jungck, M.Mörgeli, F. Niederer, Cosmic ray exposure histories of Apollo 14,Apollo 15, and Apollo 16 rocks, Proc. 14th Lunar Planet. Sci.Conf. Journal Geophys. Res. 89 (1984) B498–B512.

[56] F.A. Podosek, J.C. Huneke, G.J. Wasserburg, Gas-retention andcosmic-ray exposure ages of lunar rock 15555, Science 175(1972) 423–425.

[57] O. Eugster, P. Eberhardt, J. Geiss, N. Grögler, M. Jungck, M.Mörgeli, The cosmic-ray exposure history of shorty cratersamples; the age of shorty crater, Proc. Lunar Sci. Conf. 8th,1977, pp. 3059–3082.

[58] BVSP, Basaltic Volcanism on the Terrestrial Planets, 1981.[59] J.M. Brenan, E. Neroda, C.C. Lindstrom, H.F. Shaw, F.J.

Ryerson, D.L. Phinney, Behaviour of boron, beryllium andlithium during melting and crystallization: constraints from

mineral–melt partitioning experiments, Geochimica et Cosmo-chimica Acta 62 (1998) 2129–2141.

[60] W.F. McDonough, S.-S. Sun, The composition of the Earth,Chemical Geology 120 (1995) 223–253.

[61] H.-M. Seitz, A.B. Woodland, The distribution of lithium inperodotite and pyroxenitic mantle lithologies—an indicator ofmagmatic and metasomatic processes, Chemical Geology 166(2000) 47–64.

[62] B. Wunder, A. Meixner, R.L. Romer, W. Heinrich, Temperature-dependent isotopic fractionation of lithium between clinopyrox-ene and high-pressure hydrous fluids, Contributions to Mineral-ogy and Petrology 151 (2006) 112–120.

[63] P. Beck, M. Chaussidon, J.-A. Barrat, P. Gillet, M. Bohn, An ion-microprobe study of lithium isotopes behaviour in nakhlites, 68thAnnual Meteoritical Society Meeting, 2005, Abs. 5044.

[64] L.L. Lundberg, G. Crozaz, H.Y.J. McSween, Rare earth elementsin minerals of the ALHA77005 shergottite and implications forits parent magma and crystallization history, Geochimica etCosmochimica Acta 54 (1990) 2535–2547.

[65] Y. Ikeda, Petrography and petrology of the ALHA77005shergottite, Proc. NIPR Symp. Antarctic Meteorites 7th, Nat.Inst. Polar Res., Tokyo, 1994, pp. 9–29.

[66] P.B. Tomascak, R.W. Carlson, S.B. Shirey, Accurate and precisedetermination of Li isotopic compositions by multi-collectorsector ICP–MS, Chemical Geology 158 (1999) 145–154.

[67] J.A. Barrat, M. Chaussidon, M. Bohn, P. Gillet, C. Göpel, M.Lesourd, Lithium behaviour during cooling of a dry basalt: Anion-microprobe study of the lunar meteorite Northwest Africa479 (NWA 479), Geochimica et Cosmochimica Acta (2005)5597–5609.

[68] L.H. Chan, J.M. Edmond, G. Thompson, K. Gillis, Lithiumisotopic composition of submarine basalts: implications for thelithium cycle in the oceans, Earth and Planetary Science Letters108 (1992) 151–160.

[69] H. Palme, A. Jones, Solar system abundances of the elements,in: Andrew M. Davis (Ed.), Treatise on Geochemistry,Meteorites, Comets, and Planets, vol. 1, 2003, pp. 41–61.

[70] G. Dreibus, P.H.W. Rammensee, B. Spettel, G. Weckwerth, H.Wänke, Composition of the Shergotty parent body: furtherevidence of a two component model for planet formation, Lunarand Planetary Science XIII (1982) 186–187 (abstracts).

[71] G. Dreibus, K.H. Jochum, H. Palme, B. Spettel, F. Wlotzka, H.Wänke, LEW88516: a meteorite compositionally close to the“Martian mantle”, Meteoritics 27 (1992) 216–217 (abstract).

[72] C.-Y. Shih, L.E. Nyquist, D.D. Bogard, G.A. McKay, J.L.Wooden, B.M. Bansal, H. Wiesmann, Chronology and petro-genesis of young achondrites, Shergotty, Zagami, andALHA77005: late magmatism on a geologically active planet,Geochimica et Cosmochimica Acta 46 (1982) 2323–2344.

[73] J.C. Laul, The Shergotty consortium and SNC meteorites: anoverview, Geochimica et Cosmochimica Acta 50 (1986)875–888.

[74] C.-Y. Shih, L.E. Nyquist, B.M. Bansal, H. Wiesmann, J.L.Wooden, G.A. McKay, REE, Sr and Nd isotopic studies onshocked achondrites—Shergotty, Zagami and ALHA77005,Lunar and Planetary Science XII (1981) 973–975 (abstracts).

18 H.-M. Seitz et al. / Earth and Planetary Science Letters 245 (2006) 6–18