sign in sign up
<<
Home , Isotopes of neodymium

1.03 SolarSystemAbundances ofthe Elements H.Palme Universita¨tzuKo¨ln,Germany and A.Jones Universite´ ParisSud, France

1.03.1 ABUNDANCESOFTHE ELEMENTS IN THE SOLAR NEBULA 41 1.03.1.1 HistoricalRemarks 41 1.03.1.2 SolarSystemAbundancesofthe Elements 42 1.03.1.2.1 Isthe elementalandisotopic composition ofthe solarnebulauniform? 42 1.03.1.2.2 The composition ofthe solarphotosphere 43 1.03.1.3 AbundancesofElements inMeteorites 45 1.03.1.3.1 Undifferentiated anddifferentiated meteorites 45 1.03.1.3.2 Cosmochemicalclassification ofelements 45 1.03.1.4CIChondritesasStandardfor SolarAbundances 47 1.03.1.4.1 Chemicalvariations amongchondritic meteorites 47 1.03.1.4.2 CI chondrites 51 1.03.1.4.3 The CI abundance table 51 1.03.1.4.4Comparison withAnders andGrevesseabundance table 53 1.03.1.5SolarSystemAbundancesofthe Elements 53 1.03.1.5.1 Comparison ofmeteoriteandsolarabundances 53 1.03.1.5.2 Solarsystemabundancesversus mass number 54 1.03.1.5.3 Othersourcesfor solarsystemabundances 55 1.03.2 THE ABUNDANCESOFTHE ELEMENTS IN THE ISM 55 1.03.2.1 Introduction 55 1.03.2.2 The Natureofthe ISM 56 1.03.2.3 The ChemicalComposition ofthe ISM 57 1.03.2.3.1 The composition ofthe interstellargasandelementaldepletions 57 1.03.2.3.2 The composition ofinterstellardust 57 1.03.2.3.3 Did the solarsysteminheritthe depletion ofvolatileelements from the ISM? 59 1.03.2.3.4The ISM oxygenproblem 59 1.03.3 SUMMARY 60 REFERENCES 60

1.03.1 ABUNDANCES OF THE ELEMENTS composition ofcosmic matter,utilizingcompo- IN THE SOLAR NEBULA sitionaldataon the Earth’s crust andmeteorites. Thisled to the discovery byHarkins (1917) that 1.03.1.1 HistoricalRemarks elements withevenatomic numbers aremore abundant thanthosewithodd atomic numbers,the Atthe beginningofthe twentiethcentury so-called Oddo–Harkins rule, best exemplified attempts weremadeto definethe average for the rareearthelements (REEs). Duringthe

41 42 SolarSystemAbundancesofthe Elements 1920s and1930s VictorMoritz Goldschmidtand years therehasbeenacontinuous convergence of hiscolleaguesinGo¨ttingen,andlaterinOslo, abundancesderived from meteoritesandthose measured andcompiled awealthofchemicaldata obtained from solarabsorption linespectroscopy. on terrestrialrocks,meteorites,andindividual Theagreement isnow betterthan ^ 10% for most phasesofmeteorites. Onthe basisofthesedata elements,asdescribed below. Goldschmidt(1938)setupacosmic abundance tablewhich he publishedin1938inthe ninth 1.03.1.2 SolarSystemAbundancesofthe Elements volumeofhis Geochemische Verteilungsgesetze derElemente ( The GeochemicalLaws ofthe 1.03.1.2.1 Isthe elementalandisotopic Distribution ofthe Elements )entitled Die Men- composition ofthe solarnebula genverha¨ltnissederElementeundderAtom-Arten uniform? ( The Proportions ofthe Elements andthe Various Inthe past itwasassumed thatthe Sun,the KindsofAtoms). Goldschmidtbelieved that planets andall otherobjects ofthe solarsystem meteoriteswouldprovide the average composition formed from agaseousnebulawithwell-defined ofcosmic matter. He used the word“cosmic” chemicalandisotopic composition.Thediscovery because, incitingcontemporaneous astronomers, ofcomparatively large andwidespread variations he thoughtthatmeteoritesrepresent interstellar inoxygenisotopic compositions hascast doubt materialfrom outside the solarsystem. Inhisbook upon thisassumption (see Begemann,1980and he mentioned asecondreasonfor usingmeteorite referencestherein).Inaddition,evidence for data.Most meteoriteswill be representativeof incompletemixinginthe primordialsolarnebula average cosmic matter,becausetheyhavenot isprovided byisotopic anomaliesfor avariety of beenaffected byphysicochemicalprocesses(e.g., elements inthe refractoryinclusions ofcarbon- meltingandcrystallization),although chondrules aceouschondritesandbydetection ofthe huge withinthemhaveexperienced meltingandcrystal- isotopeanomaliesofcarbon,,,and lization,the meteoritesasawholehavenot. In someheavy elementsintiny grains ofmeteorites, contrast,the crustofthe Earth, which formed by such assilicon carbide,nanodiamond, and the meltingofthe mantle, providesonly avery graphitegrains (e.g.,Anders andZinner,1993; biased samplingofelementalabundancesinthe Chapter1.02). Agoodexamplefor such bulkEarth.Goldschmidtcalculated the average anomaliesisgiveninFigure1,wherethe unusual concentrations ofelements incosmic matterby isotopic composition ofneodymium inanaggre- usingaweighted meanofelement abundancesin gateofSiC grains from the Murchison - meteoritephases:metal(two parts),sulfide (one aceouschondriteisshown (Richter etal .,1992). part),andsilicates(10 parts).Inthiswayhe Theneodymium isotopic compositions ofall other obtained the cosmic abundancesof66 elements. Itwasduringthe sametimethatastronomers begantoextractquantitativeinformation about elementalabundancesinthe Sun bysolarabsorp- tion spectroscopy anditwassoon realized thatthe compositions ofthe Sun andthe wholeEarthare similar,except for hydrogenandotherextremely volatileelements (see Russell,1941). Almost 20 years afterGoldschmidt,Suess and Urey(1956) publishedanewabundance table, which inpart relied on solarabundances. In addition,Suess andUrey(1956) introduced argumentsbased on nucleosynthesis. Theirso- called semiempiricalabundance rules,primarily the smoothabundance variation ofodd-mass nuclei withincreasingmass number,wereapplied to estimateabundancesfor elements for which analyticaldatafrom meteoriteswerenot available or had large errors. The Suess andUreycompi- Figure1 Nd isotopesinanaggregateofmeteoritic lation wasvery influentialfor theoriesofnucleo- SiC from Murchison. The deviation ofthe Nd isotopic composition from normalisgiveninpermil(d ). All synthesisandfor the development ofnuclear 144 astrophysicsingeneral. Latercompilationsby ratios arenormalized to Nd.Fullsymbolsare measured ratios. Error bars areinmost casessmaller Cameron (1973),Anders andGrevesse(1989), thansymbol sizes. Calculated s-process productions are PalmeandBeer(1993),andothers tookinto indicated.All previous analysesofNdisotopesin accountimproved analyticaldataon meteorites terrestrial,lunar,or meteoritesamplesfall alongthe line andthe moreaccuratedetermination ofelemental marked “average solarsystem,”which isused for abundancesinthe solarphotosphere.Overthe normalization (source Richter etal .,1992). Abundancesofthe Elementsinthe SolarNebula 43 solarsystemmaterials analyzed (i.e.,terrestrial, transition probabilitiesdetermined inlaboratory lunar,andmeteoritic samples) areindistinguish- experiments. The mainneed for improvingsolar ablewithinthe scaleofFigure1andtheseother abundance dataismoreaccuratetransition materialswouldfall on the linedesignated probabilities(GrevesseandSauval,1998). “averagesolarsystem.”Such s-process com- InTable1,the composition ofthe solar ponents havealsobeenfoundfor otherelements: photosphereasobtained byabsorption spec- For example, Nicolussi etal .(1998)identified troscopy isgiven. Abundancesarenormalized to nearlypures-process insomeSiC 1012 Hatoms,the usualpractice inastronomy. grains (see Chapter1.02,figure10). Thesefindings Most ofthe dataarefrom GrevesseandSauval confirm the presence ofmaterialofdistinct (1998),which isanupdateofthe photospheric nucleosynthetic origins atthe timeofaccretion of abundance tablebyAnders andGrevesse(1989). meteoriteparent bodies. However,such isotope For nitrogen,,silicon,andironnew anomaliesareconfined to avery small fraction (a photospheric abundancesfrom Holweger(2001) fewppm)ofthe bulkofameteorite, i.e.,this wereused.Thesedataaremarked HinTable1. materialistruly exotic.Moreover,itislikely that Theiruncertaintiesrange from about 30% for the morewidespread oxygenisotopeanomaliesare nitrogento12% for silicon. The accuracyofiron not ofnucleosynthetic originbut wereproducedby isgivenas20%. Thestandarddeviations listed by fractionation processeswithinthe solarnebulaor a Holweger(2001) are, inall casesexcept silicon, precursor molecularcloud(Chapter1.06),itisstill largerthanthosegivenbyGrevesseandSauval areasonableworkinghypothesisthatthe bulkof (1998). Holwegerascribesthistohismore the matterofthe solarsystemformed from a conservativeprocedurefor calculatingerrors. A chemically andisotopically uniform reservoir,the newdetermination ofthe solarlead isincluded primordialsolarnebula.The composition ofthis (Biemont etal .,2000)andmarked BinTable1. nebula, the average solarsystemcomposition,is Theoxygenabundance wastakenfrom apaperby well knownandcarriesthe signaturesofavariety of Allende Prieto etal .(2001),marked A1inTable1. nucleosynthetic processesinstellarenvironments. Thesolaroxygenabundance hasgonedown Although the elementalcomposition ofthe solar considerably,from8.93 ^ 0.35(Anders and systemisroughly similartothatofmany other Grevesse, 1989) and8.83 ^ 0.06 (Grevesseand stars,inparticularwithrespecttothe relative Sauval,1998)to8.736 ^ 0.078(Holweger,2001) abundancesofthe nongaseous elements,thereare, and8.69 ^ 0.05(Allende Prieto etal .,2001). This indetail,compositionaldifferencesamongstars 50% decreaseisimportant because, based on the andthereare, inaddition,truly exotic stars that oldvalue, itwasthoughtthatthe interstellar makethe term “cosmic abundancesofelements” medium (ISM)musthavehad adifferent H/Oratio questionable.We will thereforeusethe term “solar thanthe Sun (see below).The carbon abundance systemabundancesofthe elements”inthischapter. hasalso beenrevised downward, asindicated in Table1(Allende Prieto etal .,2002). The new carbon andoxygenlead to ahigherC/Oratioof 1.03.1.2.2 The composition ofthe solar 0.50 ^ 0.07 compared to the earlierratioof photosphere 0.43 ^ 0.06 (Anders andGrevesse, 1989), The quantitativedetermination ofelemental although bothvaluesoverlapwithinerror bars. abundancesinthe Sun involvesthree steps: Thisimpliesasomewhatmorereducingnebular (i)construction ofanumericalmodelatmosphere; gas,asall Cispresent asCO, andthe higherC/O (ii)calculation ofanemitted spectrumbased on ratiothus reducesthe numberofO2 molecules. the modelatmosphere;and(iii)comparison ofthis Raregaseshaveno appropriatelinesinthe solar spectrum withthe observed spectrum(Cowley, spectrum. The abundance ofGrevesseand 1995). Anotherassumption usually madein Sauval(1998)isderived from standardsolar calculatingsolarabundancesisthatoflocal models.The helium abundance inthe outerlayers thermodynamic equilibrium (LTE),i.e.,“the ofthe Sun seems to havedecreased overthe quantum-mechanicalstatesofatoms,ions,and lifetimeofthe Sun,from N He/ N H 0.098atthe moleculesarepopulated accordingto the relations beginningofthe solarsystemtothe¼ present value ofBoltzmann andSaha, valid strictly inthermo- of N He/ N H 0.085 correspondingto anabun- dynamic equilibrium” (Holweger,2001). Itisnot dance of10.93¼ ^ 0.004(GrevesseandSauval, clearif, inthe highly inhomogeneous anddynamic 1998),which isgiveninTable1. Adetailed plasmapermeated byanintense, anisotropic discussion ofthe solarhelium abundance canbe radiation,the LTE assumption isjustified.In foundinLodders (2003). recent calculations effects ofNLTE (nonthermal The abundance ofneon wascalculated from an localequilibrium thermodynamics) andof Ne/Mg abundance ratioof3.16 ^ 0.07 derived photospheric granulation aretakeninto account from emergingmagnetic flux regions observed (Holweger,2001). Anotherimportant factor inthe inSkylab spectroheliograms. Thisvalueisthought accuracyofsolarabundance determinations are to be representativeofthe Ne/Mg abundance 44 SolarSystemAbundancesofthe Elements Table1 Solarphotospheric abundancesandmeteoritederived solarsystemabundances(logabundance a (H) 12). ¼ Element Solar SD Meteorite SD Sun/meteorite photosphere (CI) 1H12.00 2He 10.99a 0.02 G b 3Li 1.10 0.10 G3.30 0.040.006 4Be1.400.09 G1.410.040.98 5B2.70 2 0.12, 0.21 C2.77 0.040.74 6C8.39 0.04Aþ 27.39 0.049.90 7N7.93 0.11 H6.32 0.0440.6 8O8.69 0.05A18.430.041.82 9F4.56 a 0.3 G4.45 0.06 1.29 10 Ne 8.00a 0.07 H 11 Na 6.33 0.03 G6.30 0.02 1.07 12 Mg 7.54 0.06 H7.560.01 0.94 13 Al6.470.07 H6.460.01 1.02 14Si7.54 0.05H7.55 0.01 0.99 15P5.45 (0.04)G5.44 0.041.02 16 S7.33 0.11 G7.19 0.041.37 17 Cl5.5 a 0.3 G5.26 0.06 1.74 18Ar6.40 a 0.06 G 19 K5.12 0.13 G5.11 0.02 1.03 20 Ca 6.36 0.02 G6.33 0.01 1.07 21 Sc 3.17 0.10 G3.080.01 1.22 22 Ti 5.02 0.06 G4.950.041.18 23 V4.00 0.02 G3.99 0.02 1.02 24Cr5.67 0.03 G5.67 0.01 0.99 25Mn5.39 0.03 G5.510.01 0.75 26 Fe 7.45 0.08H7.490.01 0.92 27 Co4.92 0.04G4.90 0.01 1.05 28Ni6.250.04G6.23 0.02 1.05 29 Cu4.21 0.04G4.280.040.85 30 Zn4.60 0.08G4.66 0.040.87 31 Ga 2.88 (0.10) G3.11 0.02 0.59 32 Ge 3.410.14G3.62 0.040.62 33 AsG2.350.02 34SeG3.400.04 35BrG2.61 0.04 36 Kr3.30a 0.06 P 37 Rb 2.60 (0.15)G2.400.041.59 38Sr2.97 0.07 G2.88 0.041.22 39 Y2.240.03 G2.21 0.041.06 40Zr2.60 0.02 G2.590.041.02 41Nb 1.420.06 G1.39 0.041.07 42Mo1.92 0.05G1.950.040.93 44 Ru1.84 0.07 G1.800.041.11 45 Rh 1.12 0.12 G1.10 0.081.05 46Pd 1.69 0.04G1.680.041.01 47Ag (0.940.25)G1.23 0.04 48 Cd 1.77 0.11 G1.750.041.05 49In(1.66 0.15)G0.800.04 50Sn2.00 (0.3) G2.12 0.040.76 51Sb 1.00 (0.3) G1.00 0.040.99 52Te G2.22 0.04 53IG1.500.08 54 Xe 2.16a 0.09 P 55 CsG1.12 0.04 56Ba 2.13 0.05G2.21 0.040.83 57La 1.17 0.07 G1.21 0.02 0.91 58 Ce 1.58 0.09 G1.62 0.02 0.90 59Pr0.71 0.08G0.800.040.81 60 Nd 1.500.06 G1.48 0.02 1.04 62 Sm1.01 0.06 G0.980.02 1.08 63 Eu0.510.08G0.55 0.02 0.92 (continued) Abundancesofthe Elementsinthe SolarNebula 45

Table1 (continued). Element Solar SD Meteorite SD Sun/meteorite photosphere (CI) 64Gd1.12 0.04G1.080.02 1.10 65Tb( 2 0.1 0.3) G0.340.04 66 Dy1.140.08G1.16 0.02 0.96 67 Ho(0.26 0.16) G0.500.04 68Er0.93 0.06 G0.96 0.02 0.93 69 Tm(0.00 0.15)G0.150.04 70 Yb 1.08(0.15)G0.950.02 71 Lu0.06 0.10 G0.13 0.040.85 72 Hf 0.88 (0.08)G0.740.041.37 73 Ta G 2 0.140.06 74W(1.11 0.15)G0.66 0.04 75ReG0.29 0.04 76 Os1.45 0.10 G1.380.041.18 77 Ir1.35(0.10) G1.36 0.02 0.97 78Pt1.800.30 G1.67 0.06 1.36 79 Au(1.01 0.15)G0.84 0.02 80Hg G1.150.08 81Tl(0.9 0.2) G0.810.04 82Pb 2.00 0.06 B2.050.040.89 83Bi G0.69 0.06 90 Th 0.07 0.04 92 U( ,2 0.47) G 2 0.520.04

6 SD—standarddeviation index:0.1–12%,0.2–60%,0.3–100%. Meteoritedata: log a E 1.546; a E abundancesrelativeto 10 Si atoms (see Table3). Solardata: A1—Allende Prieto etal .(2001);A2—Allende Prieto etal .(2002); B—Bieþ mont etal .(2000);C—Cunha andSmith(1999); G—GrevesseandSauval(1998); H—Holweger(2001);P—PalmeandBeer(1993). a Abundancesarenot derived from the photosphere. b Average Sun; outerlayers ofSun are10% lowerinHe;valuesinparenthesisaredefined asuncertainbyGrevesseandSauval(1998). ratiointhe solarphotosphere(Reames,1998). be representativeofthe bulkparent planetesimal FollowingHolweger(2001) weadopt avalue from which the meteoriteswerederived.Differ- oflog A Ne 8.001 ^ 0.069 calculated from the entiated meteoritesarepiecesofplanetesimals that Ne/Mg andN¼ e/Oratios ofReames(1998). The weremoltenanddifferentiated into core, mantle, valueisbased on datafrom coronalspectra andcrust. Asamplefrom such abodywill not be andamorepreciseSEP (solarenergetic particles) representativeofthe bulkplanetanditisnot a valuebyReames(1998),asdiscussed inGrevesse trivialtaskto derive, from the samplesavailable, andSauval(1998). the bulkcomposition ofthe parent planet. For andxenon abundanceswere Undifferentiated meteoritesreflect,atleast to derived from computerfits of s N (neutroncapture somedegree,the composition ofthe solarnebula cross-section timesabundance)versusmass from which theyformed.Thereare, however, number. Nuclei thatareshieldedfromthe variabilitiesinthe composition ofundifferentiated r-process,so-called s-only nuclei, wereused for meteorites,which must reflectinhomogeneitiesin the fitandthe abundancesof 82 Krand 128 Xe were the solarnebulaor disequilibrium duringfor- calculated.From thesedata, andthe isotopic mation ofsolidsfrom gas,or both.Amore composition ofthe solarwind, the krypton and comprehensivediscussion ofmeteoriteclassifi- xenon elementalabundanceswerecalculated cation isgivenbyKrot etal .(Chapter1.05). (PalmeandBeer,1993)andarelisted inTable1. ThemeteoritedatagiveninTable1will be discussed inalatersection. 1.03.1.3.2 Cosmochemicalclassification ofelements 1.03.1.3 AbundancesofElements inMeteorites Many ofthe processesthatareresponsiblefor 1.03.1.3.1 Undifferentiated anddifferentiated the variablechemicalcomposition ofprimitive meteorites meteoritesarerelated to the temperatureof formation ofmeteoritic components. Although Therearetwodifferent groups ofmeteorites— itisdifficulttounambiguously ascertainthe undifferentiated anddifferentiated.The undiffer- condensation originofany singlemeteoritic entiated meteoritesarepiecesofplanetesimals that component,itisclearfrom the elementalpatterns haveneverbeenheated to meltingtemperatures. thatcondensation processesmusthaveoccurred in Theirchemicalandisotopic composition should the early solarsystemandthatthe volatilitiesof 46 SolarSystemAbundancesofthe Elements the elements inthe solarnebulaenvironment were meteoritesmaybe ascribed to the incorporation important inestablishingthe various meteorite ofvariableamounts ofanearlycondensed compositions. Condensation temperaturesprovide refractory phase. aconvenient measureofvolatility.Thesetem- (ii) Magnesium silicates .The major fraction of peraturesarecalculated byassumingthermodyn- condensiblematterisassociated withthe three amic equilibrium betweensolidsandacoolinggas mostabundant elements heavierthanoxygen— ofsolarcomposition. Major elements condenseas silicon,magnesium,andiron. Inthe reducing minerals whileminor andtrace elements condense environment ofthe solarnebulaironcondenses insolid solution withthe major phases. The almost entirely asmetal,whilemagnesium and temperaturewhere50% ofanelement isinthe silicon form forsterite(Mg2 SiO4 ),which is,to a solid phaseiscalled the 50% condensation large extent,converted to enstatite(MgSiO3 )at temperature(Wasson,1985;Lodders,2003). lowertemperaturesbyreaction withgaseous SiO. Withinthisframeworkfivemoreor less well- Asforsteritehasanatomic Mg/Si ratiotwicethe defined components thataccount for the variations solarsystemratio,loss or gainofaforsterite inthe elementalabundancesinprimitivemeteor- component isthe most simplewayfor producing itesmaybe defined (Table2). Inaddition,the state variations inMg/Si ratios.Thus,variationsin ofoxidation ofameteoriteisanimportant Mg/Si ratios ofbulkmeteoritesareproducedby parameterthataddstothe complextextural the incorporation ofvarious amounts ofearly- variability amongchondritic meteorites. Accord- formed forsterite. ingto the condensation temperaturesthe follow- (iii) Metallic .MetalcondensesasanFe– ingcomponents aredistinguished: Ni alloy ataboutthe sametemperatureas (i) Refractory component.The first phasesto forsterite, the sequence dependingon pressure. condensefrom acoolinggasofsolarcomposition Atpressuresabove102 4 bariron metalcondenses arecalcium,aluminum oxides,andsilicates beforeforsteriteandatlowerpressuresforsterite associated withacomparatively large numberof condensesahead ofmetal(GrossmanandLarimer, trace elements,such asREEs,,haf- 1974). Variations inthe concentrations ofiron and nium,andscandium. Theseelements areoften othersiderophileelementsinmeteoritesare named refractory lithophileelements (RLEs),in producedbythe incorporation ofvariablefrac- contrast to the refractory siderophileelements tions ofmetal. (RSEs) comprisingmetals withlow vapor press- (iv) Moderately volatileelements .These ures,e.g.,,,andiridium conden- havecondensation temperaturesbetweenthose singatsimilarly high temperaturesas ofmagnesium silicatesandFeS (troilite). The multicomponent metalalloys. BothRLEsand most abundant ofthe moderatelyvolatile RSEsareenrichedinCa–Al-richinclusions elements issulfur,which condensesbyreaction (Chapter1.08)byafactor of20 on average, ofgaseoussulfur withsolid ironat710 K, reflectingthe factthatthe refractory component independent ofpressure.Othermoderately makesup , 5%ofthe totalcondensiblematter volatileelements condenseinsolid solution with (GrossmanandLarimer,1974). Variationsin major phases. Moderately volatileelements are Al/Si,Ca/Si,etc.,ratios ofbulkchondritic distributed amongsulfides,silicates,andmetals.

Table2 Cosmochemicalclassification ofthe elements. Elements

Lithophile(silicate)Siderophile chalcophile (sulfideþ metal) þ Refractory T c 1,850–1,400 K Al,C¼ a, Ti, Be, Ba, Sc, V, Sr,Y, Re, Os,W,Mo,Ru,Rh, Ir, Zr,Nb, Ba, REE, Hf, Ta, Th, U, Pu Pt,Rh

Maincomponent T c 1,350–1,250 K Mg,¼ Si, Cr,LiFe, Ni, Co,Pd

Moderately volatile T c 1,230–640 K Mn¼ ,P,Na, B, Rb, K, F, ZnAu,Cu,Ag, Ga, Sb, Ge, Sn, Se, Te, S Highly volatile T c , 640 K Cl,Br,I,Cs,Tl,H,C,N,O,He, Ne, Ar, In,Bi, Pb, Hg Kr,Xe

2 4 T c —Condensation temperaturesatapressureof10 bar(Wasson,1985;for B, LaurettaandLodders,1997). Abundancesofthe Elementsinthe SolarNebula 47 Theirabundancesareinmost casesbelow solar, fugacity ofameteoriteis,however,not well i.e.,theyhavelowerelement/silicon ratios than defined.Large variations inoxygenfugacity are the Sun or CI chondrites,theyaredepleted (see oftenrecorded inindividualcomponentsofa below). InFigure2,abundancesofmoderately singlemeteorite.Thevarious componentsof volatileelements inCV3meteoritesrelativeto primitivemeteoritesapparently represent extreme thoseinCI meteoritesareplotted.Increasing disequilibrium.Oxygenisotopesgiveasimilar depletions correlatewithdecreasingcondensation picture(Clayton,1993). Ithasbeensuggested that temperaturesbut areindependent ofthe geochem- variationsin D 17Owereproducedbyreaction of icalpropertiesofthe elements.Depletions of 16O-rich materialwithagasrich in 17Oand 18 O moderately volatileelements inmeteoritesare (Clayton,1993). Thegasphasemaybe considered produced byincompletecondensation. The anadditionalindependent component ofmeteor- amount andthe relativeabundancesofthese ites. Thus,the extent ofthe gas–solid reaction at elements inmeteoritesareprobably the result of various temperatures,andpossibly also fluid– removalofvolatilesduringcondensation (Palme solid reactions,on aparent bodydeterminethe etal .,1988). degree ofoxidation andthe oxygenfugacity of (v) Highly volatileelements.Thesehavecon- meteoritic componentsandbulkmeteorites densation temperaturesbelowthatofFeS (Chapter1.06). (Table1). Thegroup ofhighly volatileelements compriseselements withvery different geochem- 1.03.1.4CIChondritesasStandardfor Solar icalaffinity,such asthe chalcophileleadandthe Abundances atmophileelementsnitrogenandraregases. Similarprocessesasthoseinvokedfor the 1.03.1.4.1 Chemicalvariationsamong depletion ofmoderately volatileelements are chondritic meteorites responsiblefor variationsintheseelements. In InFigure3the variations ofselected element addition,heatingon small parent bodiesmaylead ratios inthe different groupsofchondritic to loss ofhighly volatileelements. meteoritesareshown. All ratios arenormalized (vi) Oxygenfugacity andoxygenisotopic to the best estimateofthe average solarsystem composition.The oxygenfugacitiesrecorded in ratios,the CI ratios (takenfrom Table3,see meteoritic minerals areextremely variable, from below).Meteoritegroups arearranged inthe the high oxygenfugacity recorded inthe magne- sequence ofdecreasingbulkoxygencontents,i.e., titeofcarbonaceous chondritestothe extremely decreasingaverageoxygenfugacity. Element reducingconditionsinenstatitechondrites, ratios inthe solarphotospheredetermined by reflected inthe presence ofsubstantialamounts absorption linespectroscopy areshownfor ofmetallic silicon dissolved inFeNi.The comparison (Table1). InFigure3,aluminum isrepresentativeof refractory elements ingeneralandthe Al/Si ratios indicatethe sizeofthe refractory component relativeto the major fraction ofthe meteorite.Itis clearfrom thisfigurethatthe Al/Si ratioofCI meteoritesagreesbest withthe solarratio, although the ratios inCM (Type2carbonaceous chondrites) andevenOC (ordinary chondrites) are almost withinthe error barofthe solarratio. The errors ofthe meteoriteratios arebelow 10%,in many casesbelow 5%. Avery similarpattern as for aluminumwouldbeobtained for other refractory elements (,,, REEs,etc.),asratios amongrefractory elements in meteoritesareconstant inall classesofchondritic meteorites,atleast within , 5–10%. Theaverage Sun/CI meteoriteratioof19 refractorylithophile elements (Al,Ca, Ti, V, Sr,Y,Zr,Nb, Ba, La, Ce, Pr,Nd, Sm,Eu,Gd, Dy,Er,Lu,see Table2) is Figure2 Abundancesofvolatileelements inCV3 1.004withastandarddeviation of0.12. The chondrites(e.g.,Allende)normalized to CI chondrites andSi.Thereisacontinuous decreaseofabundances accuracyofthe solarabundance determination of withincreasingvolatility asmeasured bythe conden- each oftheseelements isbetterthan25%(see sation temperature.The sequence contains elements of Table1). Elements withlargererrors werenot very different geochemicalcharacter,indicatingthat considered.Calculatingthe standarderror ofthe volatility isthe only relevant parameterinestablishing meanfor the average refractory element/silicon thispattern (source Palme, 2000). ratioleadstoanuncertainty ofabout ^ 0.03 inthe 48 SolarSystemAbundancesofthe Elements

Figure3 Element/Si mass ratios ofcharacteristic elements invarious groups ofchondritic (undifferentiated) meteorites. Meteoritegroups arearranged accordingto decreasingoxygencontent. The best match betweensolar abundancesandmeteoritic abundancesiswithCImeteorites(see text for details). ratioofrefractoryelements betweenCI meteorites however,less diagnostic, asall groups ofcarbon- andthe Sun. Thus the absolutelevelofrefractory aceouschondriteshavethe sameMg/Si ratio elements(measured byrefractoryelement/silicon (WolfandPalme, 2001). The OC andthe EC ratios) andratios amongrefractoryelements are havesignificantly less magnesium.The error of the sameinCI meteoritesandthe Sun. Thelevel the solarratiois , 19% (combiningthe errors of ofrefractory elements inotherchondritic meteor- silicon andmagnesium,Table1) andthus covers itesishigherinCM (by13%) andinCV (by25%) the range ofall classesofchondritesexcept EH andlowerinHchondrites(by10%)andenstatite chondrites. chondrites(by20%). Thus,the agreement Untilrecently,the range ofthe bulkironcontent betweenrefractory elements inthe Sun andin inchondritic meteoritesvaried byabout afactor of CI meteoritesisstatistically significant andall 2,withmost meteoritegroups beingdepleted in othergroups ofmeteoriteswill not match solar iron(Figure3). Therearenew,recentlydiscov- refractoryelement abundances. TheCRmeteor- ered subgroups ofcarbonaceous chondrites,with ites(Renazzo typemeteorites) haveCIratios of large excessesofiron (see Chapter1.05). These Al/Si (Bischoff etal .,1993) (not shownin CH andCBchondritesindicatethatmetalbehaves Figure3). However,thesemeteoritesaredepleted asanindependent component,somegroups of involatileelements relativeto CI meteorites, chondritic meteoritesareenrichedinironand disqualifyingthemasasolarsystemstandard. othersiderophileelements,others aredepleted. The Mg/Si ratios ofCIchondritesalso match Theexcellent agreement ofCImeteoriteabun- withthe solarabundance ratio(Figure3). Thisis, dance ratios withsolarabundance ratios is Abundancesofthe Elementsinthe SolarNebula 49 Table3 Solarsystemabundancesbased on CI meteorites. PalmeandBeer(1993,updated)Anders andGrevesse(1989)

Element MeanCI Lit. Estimated Atoms per10 6 MeanCI Atoms per10 6 abundance accuracy atoms ofSi abundance atoms ofSi (byweight) (%) (byweight) 1H2.02 (%) p 10 5.27 106 2.02 (%) 2He 56 (nLg2 1 ) p £ 56 (nLg2 1 ) 3Li 1.49 (ppm) 10 56.51.50 (ppm) 57.1 4Be0.0249 (ppm) 10 0.727 0.0249 (ppm) 0.73 5B 0.69 (ppm) Z13 16.80.870 (ppm) 21.2 6C3.22 (%) p 10 7.05 105 3.45 (%) 1.01 107 7N3,180 (ppm) p 10 5.97 £ 104 0.318 (%) 3.13 £ 106 8O 46.5 (%) p 10 7.64 £ 106 46.4 (%) 2.38 £ 107 9F58.2 (ppm) 15806£ 60.7 (ppm) 84£ 3 10 Ne 203 (pLg2 1 ) p 203 (pLg2 1 ) 11 Na 4,982 (ppm) 55.70 104 5,000 (ppm) 5.74 104 12 Mg 9.61 (%) 31.04 £ 106 9.89 (%) 1.074 £ 106 13 Al8,490 (ppm) W38.27 £ 104 8,680 (ppm) 8.49 £ 104 14Si10.68 (%) 3 ; 10£ 6 10.64 (%) ; £106 15P 926 (ppm) W77.86 103 1,220 (ppm) 1.04 104 16 S5.41 (%) D54.44 £ 105 6.25 (%) 5.15 £ 105 17 Cl698 (%) 155.18 £ 103 704 (ppm) 5.24 £ 103 18Ar751 (pLg2 1 ) p £ 751 (pLg2 1 ) £ 19 K544 (ppm) 53.66 103 558 (ppm) 3.77 103 20 Ca 9,320 (ppm) W36.12 £ 104 9,280 (ppm) 6.11 £ 104 21 Sc 5.90 (ppm) 334.£55.82 (ppm) 34£ .2 22 Ti 458 (ppm) W42.52 103 436 (ppm) 2.40 103 23 V54.3 (ppm) 5280£ 56.5 (ppm) 293£ 24Cr2,646 (ppm) 31.34 104 2,660 (ppm) 1.35 104 25Mn1,933 (ppm) 39.25 £ 103 1,990 (ppm) 9.55 £ 103 26 Fe 18.43 (%) 38.68 £ 105 19.04 (%) 9.00 £ 105 27 Co506 (ppm) 32.26 £ 103 502 (ppm) 2.25 £ 103 28Ni1.077 (%) 34.82 £ 104 1.10 (%) 4.93 £ 104 29 Cu131 (ppm) 10 5421£ 26 (ppm) 5£22 30 Zn323 (ppm) 10 1.30 103 312 (ppm) 1.26 103 31 Ga 9.71 (ppm) 536.6£ 10.0 (ppm) 37.£ 8 32 Ge 32.6 (ppm) 10 11832.7 (ppm) 119 33 As1.81 (ppm) 56.351.86 (ppm) 6.56 34Se21.4 (ppm) D571.3 18.6 (ppm) 62.1 35Br3.50 (ppm) 10 11.53.57 (ppm) 11.8 36 Kr8.7 (pLg2 1 )8.7 (pLg2 1 ) 37 Rb 2.32 (ppm) 57.142.30 (ppm) 7.09 38Sr7.26 (ppm) 521.87.80 (ppm) 23.5 39 Y1.56 (ppm) J234.61 1.56 (ppm) 4.64 40Zr3.86 (ppm) J2211.1 3.94 (ppm) 11.4 41Nb 0.247 (ppm) J230.699 0.246 (ppm) 0.698 42Mo0.928 (ppm) 52.54 0.928 (ppm) 2.55 44 Ru0.683 (ppm) J131.780.712 (ppm) 1.86 45 Rh 0.140 (ppm) J130.358 0.134 (ppm) 0.344 46Pd 0.556 (ppm) 10 1.37 0.560 (ppm) 1.39 47Ag 0.197 (ppm) 10 0.4800.199 (ppm) 0.486 48 Cd 0.680 (ppm) 10 1.590.686 (ppm) 1.61 49In0.0780 (ppm) 10 0.1780.080 (ppm) 0.184 50Sn1.68 (ppm) 10 3.72 1.720 (ppm) 3.82 51Sb 0.133 (ppm) 10 0.2870.142 (ppm) 0.309 52Te 2.27 (ppm) 10 4.682.320 (ppm) 4.81 53I0.433 (ppm) 20 0.897 0.433 (ppm) 0.90 54 Xe 8.6 (pLg2 1 )8.6 (pLg2 1 ) 55 Cs0.188 (ppm) 50.372 0.187 (ppm) 0.372 56Ba 2.41 (ppm) 10 4.61 2.340 (ppm) 4.49 57La 0.245 (ppm) 50.4640.2347 (ppm) 0.4460 58 Ce 0.638 (ppm) 51.20 0.6032 (ppm) 1.136 (continued) 50 SolarSystemAbundancesofthe Elements Table3 (continued). PalmeandBeer(1993,updated)Anders andGrevesse(1989)

Element MeanCI Lit. Estimated Atoms per10 6 MeanCI Atoms per10 6 abundance accuracy atoms ofSi abundance atoms ofSi (byweight) (%) (byweight) 59Pr0.0964 (ppm) 10 0.1800.0891 (ppm) 0.1669 60 Nd 0.474 (ppm) 50.8640.4524 (ppm) 0.8279 62 Sm0.154 (ppm) 50.269 0.1471 (ppm) 0.2582 63 Eu0.0580 (ppm) 50.100 0.0560 (ppm) 0.0973 64Gd0.204 (ppm) 50.3410.1966 (ppm) 0.3300 65Tb0.0375 (ppm) 10 0.0621 0.0363 (ppm) 0.0603 66 Dy0.254 (ppm) 50.411 0.2427 (ppm) 0.3942 67 Ho0.0567 (ppm) 10 0.09040.0556 (ppm) 0.0889 68Er0.166 (ppm) 50.261 0.1589 (ppm) 0.2508 69 Tm0.0256 (ppm) 10 0.0399 0.0242 (ppm) 0.0378 70 Yb 0.165 (ppm) 50.2510.1625 (ppm) 0.2479 71 Lu0.0254 (ppm) 10 0.03820.0243 (ppm) 0.0367 72 Hf 0.107 (ppm) 50.158 0.104 (ppm) 0.154 73 Ta 0.0142 (ppm) J26 0.0206 0.0142 (ppm) 0.0207 74W 0.0903 (ppm) J140.129 0.0926 (ppm) 0.133 75Re0.0395 (ppm) J140.0558 0.0365 (ppm) 0.0517 76 Os0.506 (ppm) J1 p 20.699 0.486 (ppm) 0.675 77 Ir0.480 (ppm) J140.6570.481 (ppm) 0.661 78Pt0.982 (ppm) J141.32 0.990 (ppm) 1.34 79 Au0.148 (ppm) J140.1980.140 (ppm) 0.187 80Hg 0.310 (ppm) 20 0.406 0.258 (ppm) 0.34 82Tl0.143 (ppm) 10 0.184 0.142 (ppm) 0.184 82Pb 2.53 (ppm) 10 3.21 2.470 (ppm) 3.15 83Bi 0.111 (ppm) 150.1400.114 (ppm) 0.144 90 Th 0.0298 (ppm) 10 0.03380.0294 (ppm) 0.0335 92 U0.00780 (ppm) 10 0.00860.0081 (ppm) 0.0090

Datafrom PalmeandBeer(1993),except:raregasesinnLg2 1 or pLg2 1 at(STP)from Anders andGrevesse(1989),elements marked D (Dreibus etal .,1995),J1(Jochum,1996),J1 p Osiscalculated from the averagecarbonaceous chondriteratioIr/Osof0.0949from Jochum (1996); J2(Jochum etal .,2000); W(WolfandPalme, 2001); Z(Zhai andShaw,1994); p elements incompletely condensed inCI meteorites. Average CI abundancesfrom Anders andGrevesse(1989),Table1,columns 6and2,except C, N, OandraregasesHe, Ne, Ar,Kr,Xewhich areonly from Orgueil. obvious,although the formalerror ofthe solariron cannot be representativeofsolarabundances abundance isquitelarge (Table1). However,the becauseoftheirlow refractory element contents sameargument thathasbeenapplied to refractory andtheirfractionated Mg/Si ratios (Figure3). elementscanbe used here.The patterns ofnickel Thenewphotospheric determinations andcobalt areidenticaltothatofironandthe (Table1) show thatthereisnow agreement errors inthe solarabundancesofnickelandcobalt betweenthe CI leadandthe solarlead abundance. areonly halfofthe error inthe ironabundance Thus the excellent match ofCIchondriteswiththe determination (Table1).Thecombined error solarphotospherecanbe extended to someofthe associated withthe solar/CI meteoriteratiofor highly volatileelements. thesethree metals isthenbelow 10%,assuming The CI chondritesnot only havethe highest constant ratios betweenthe three metals. Thus contents ofvolatileelements but also the highest only CI, H, andEHchondriteshavesolarFe/Si content ofoxygen(alarge fraction inthe form of ratios. AsHandEHchondriteshavevery different water) ofall chondritic meteorites. However,in refractoryelement contents,theyarenot suitable contrast to otherhighly volatileelements,such as meteoritesfor representingsolarabundances. lead, the amount ofoxygencontained inCI InFigure3,,, andsulfur are meteoritesisstill afactor of2below thatofthe representativeofthe abundancesofmoderately solarphotosphere(Figure3),implyingthatwateris volatileelements (Figure2andTable2). Abun- not fully condensed.The concept ofincomplete dancevariationsreach afactor of5for and condensation ofvolatileelements inmost meteor- 10 for zinc.All three elements show excellent itegroups issupported bythe observation thatthere agreement ofsolarwithCIabundances,incontrast isnogroup ofmeteoritesthatisenriched inthese to othergroups ofchondritic meteorites,except elements relativeto the average solarsystemabun- for the enstatitechondrites,which reachthe level dances. Therewasnolateredistribution ofvola- ofCIabundances. However,enstatitechondrites tiles,for example, byreheating.InCI meteorites, Abundancesofthe Elementsinthe SolarNebula 51 the moderately volatileelements andsomeofthe 1.03.1.4.3 The CI abundance table highly volatileelements arefully condensed;other groups ofchondritesacquired lowerfractions of The CI abundance tablegivenhere(Table3) volatilesbecausethe solargasdissipated during islargely based on the compilation byPalme condensation (Palme etal .,1988). andBeer(1993).Relevant datapublished Insummary,thereisonlyonegroup of after1993 areincorporated asdescribed below. meteorites,the CI chondrites,thatclosely match Inaddition,the widely used CI dataofAnders solarabundancesfor elementsrepresenting andGrevesse(1989) arelisted for comparison. the various cosmochemicalgroups andexcluding Inthe 1993 compilation wasthe only the extremely volatileelements such asthe rare element thathad notbeendetermined inthe gases,,carbon,oxygen,andnitrogen Orgueilmeteorite.Jochum (1996)reported a andalsothe element ,whichwill be valueof134ppbfor hisOrgueilanalysis,in discussed inSection 1.03.1.5.3.All otherchon- goodagreement withthe valuelisted inthe 1993 dritegroups deviatefrom solarabundancesand compilation,estimated from rhodium inother the deviations canbe understood, atleast in meteorites. The valuesfor ,tungsten, principle, bygas–solid fractionation processesin ,osmium,,,andgold the early solarsystem. wereall takenfrom the paperbyJochum (1996) andaremarked J1inTable3. Thedifference betweenthe newdatafromJochum (1996) andthe olddatalisted inthe Palmeand 1.03.1.4.2 CI chondrites Beer(1993) compilation isinall casesbelow5%. Amongthe morethantwentythousandrecov- Newsulfur andselenium datafor CI chondrites ered meteoritesthereareonly fiveCImeteorites: wereobtained byDreibus etal .(1995)inastudy Orgueil,Ivuna, Alais,Tonk,andRevelstoke. ofsulfur andselenium inchondritic meteorites. Thesemeteoritesarevery fragileandareeasily Withthe measured CI contentsof5.41% for sulfur fragmented on atmospheric entry. Inaddition, and21.4ppm for inOrgueilaCIS/Se theirsurvivaltimeon Earthisshort. All fiveCI ratioof2,540isobtained.Anearly identicalS/Se meteoritesareobserved falls. Most analyseshave ratioof2,560 ^ 150wascalculated asaverage of beenperformed on the Orgueilmeteorite, simply carbonaceouschondrites,consideringonly becauseOrgueilisthe largest CI meteoriteand observed meteoritefalls. The average ratioofall materialiseasily availablefor analysis. However, meteoritefalls analyzed byDreibus etal .(1995) problems withsamplesize, samplepreparation, was2,500 ^ 270. Thesedatasuggest thatthe new andthe mobility ofsomeelements arereflected in sulfur andselenium content ofOrgueilprovidesa the chemicalinhomogeneitieswithinthe Orgueil reliableestimatefor the solarsystemaverage.The meteorite, andoftenmake acomparison ofdata newCI sulfur content isslightlyhigherthanthe obtained bydifferent authorsdifficult. This sulfur content giveninthe PalmeandBeer(1993) contributessignificantly to the uncertaintiesin compilation andsignificantly lowerthanthe sulfur CI abundances. content of6.25%inthe compilation ofAnders and The chemicalcomposition ofCIchondritesas Grevesse(1989).Thecorrespondingchange in showninFigure3isthe basisfor designatingCI selenium isto21.4from 21.3 ppm inPalmeand meteoritesasprimitiveor unfractionated.Textu- Beer(1993) andfrom 18.6 ppm inAnders and rally andmineralogically theyarefarfrom being Grevesse(1989). primitive.CI meteoritesaremicrobrecciaswith Newphosphorus andtitanium XRF datafor CI millimetertosubmillimeterclasts withvariable chondriteswerereported byWolfandPalme composition. Latestage fracturesfilled with (2001).The change inphosphorus issignificant. carbonates,hydrous calcium,andmagnesium The newCI abundance is926 ppm,which ismuch sulfatedemonstratethatlow temperaturepro- lowerthanthe valuesinthe oldercompilations, cesseshaveaffected the meteorite.TheCI 1,105ppminPalmeandBeer(1993) and meteoriteshaveno chondrulesandthus consist 1,200ppm inAnders andGrevesse(1989), almost entirely ofextremely fine-grained hydrous respectively. Thenewphosphorus contents are silicateswithsomemagnetite.High temperature considered to be morereliable.The changesin phasessuch asolivineandpyroxenearefrequently titanium aresmall. WolfandPalme(2001) also found(Dodd,1981).Although the CI meteorites reported major element concentrationsofCI undoubtedly match solarabundancesvery closely, meteoritesandothercarbonaceous chondrites. processesthatoccurred lateon the Orgueilparent Theirmagnesium andsilicon contentswere bodyhavelargely establishedtheirpresent texture almost identicaltothosegivenbyPalmeand andmineralogy. Onacentimeterscalethese Beer(1993):10.69% versus 10.68%(for silicon) processesmusthavebeenessentially isochemical, inPalmeandBeer(1993) and9.60% versus otherwisethe composition ofOrgueilwouldnot 9.61% (for magnesium)inPalmeandBeer(1993). be solarfor so many elements. The aluminum,calcium,andiron concentrations 52 SolarSystemAbundancesofthe Elements ofPalmeandBeer(1993) were, however,slightly (ii)ThereisacleartrendofdecreasingFe/Mg different.The average betweenthe newdataand ratios withthe increasingdepletion ofmoderately thoselisted earlierinPalmeandBeer(1993) was volatileelements,suggestingthatthe variations in used inthe present compilation. ironcontentsareprobably related to the volatility Figure4isaplot ofsomerefractory and ofiron. Thus indefiningthe CI ratiofor ironone moderately volatileelements inthe various types cannot rely on othercarbonaceous chondrites. ofcarbonaceous chondrites. Calculated conden- Thereisanuncertainty of , 5%inthe CI Fe/Mg sation temperaturesatapressureof102 4 bar ratioof1.92 (byweight) giveninTable1. This (Wasson,1985)areindicated to the right. Thereis uncertainty reflects variableiron contentsinCI atrendfor increasingdepletions(decreasing samplesanditwill be difficult to obtainmore abundances) withdecreasingcondensation tem- accurateCIironcontents. peratures(see alsoFigure2).However,the (iii)Asimilarproblemisencountered with sequence ofdecreasingcondensation tempera- refractory elements.WolfandPalme(2001) turesfor silicon,,iron,andphosphorus noted a20% variation inthe calcium content of doesnot match withthe depletions observed in various CI meteoritesamples. Thecorresponding carbonaceous chondrites. The differencesare aluminum contents arefortunately much more small; however,the condensation temperatures constant andthe Al/Mg ratioof0.0865is ofphosphorus andchromium arenot well known estimated to be accurateto within2%. Asa andthe condensation temperatureofiron(con- result thereissomeuncertainty inthe solar densingasmetal) relativeto silicon andother systemCa/Alratio,from , 1.07to1.10. lithophileelementsisdependent on nebular Members ofthe reduced subgroup inCV pressure(GrossmanandLarimer,1974). These chondriteshaveCa/Alratios atleast 10% minordiscrepanciesdonot change the basic belowthoseofothercarbonaceous chondrites conclusion thatthesedepletions arerelated to (WolfandPalme, 2001). Ingeneral,however, condensation temperatures. Anumberofcon- ratios amongrefractory elementsinother clusionswithregardto the CI abundancescanbe chondritic meteoritesmaybe used to improve drawn from Figure4. the accuracyofchondritic refractory element (i)TheMg/Si weightratio,0.90,isconstant ratios,which arevariableinCI meteoritesdueto withinonepercent inbulkcarbonaceous chon- inhomogeneousdistribution.Anextreme drites,except for athree percent depletion of examplefor CI variability isU.Rocholl and magnesium inCV chondrites. Thereisno Jochum (1993) foundTh/Uratios inCI chon- apparent trendthatwouldsuggest volatility dritesvaryingfrom 1.06 to 3.79. The variability related depletion ofsilicon,comparableto the isprimarily aresult oflow temperaturemobil- depletions observed for chromium,, ization ofuranium underaqueous conditions. As zinc, etc. otherchondritic meteoritesarealso variable, although to alesserextent,the CI content isnot too well determined.Rocholland Jochum (1993) suggest avalueof3.9 ^ 0.2. Otherelements,such asbarium,,and ,mayalso be affected byalteration processesonparent bodies,although to alesser extent (RochollandJochum,1993). (iv) The geochemically similar,but cosmo- chemically dissimilarelements magnesium and chromium arefractionated incarbonaceous chon- drites:chromium behavesasaslightlymore volatileelement thanmagnesium (Figure4). The Fe/Crratioisapparently less variablethanthe Mg/Crratioreflectingthe volatility related behavior ofironincarbonaceous chondrites Figure4 Abundancesofrefractory andmoderately (Figure4). WolfandPalme(2001) foundan volatileelements invarious groups ofcarbonaceous average Mg/CrratioinCI of36.52compared to chondrites,normalized to CI andMg.Refractory 40.75inCO chondrites,but Fe/Crratios of70.89 elements increasefrom CI to CV3chondriteswhile for CI and71.05for CO.Much strongervariations Mg/Si ratios areconstant inall groups ofcarbonaceous arefoundfor othermoderately volatileelements, chondrites. Although the elementsCr,Fe, andP aresignificantly less depleted thanMnandZn,they such asmanganese, sodium,,sulfur, show asimilarbehavior,suggestingvolatility related zinc, selenium,etc.(see Figures2–4). The depletions ofCr,Fe, andPincarbonaceous chondrites concentrations oftheseelements inothercarbon- ofhighermetamorphic grades(source Wolfand aceouschondritescannot be used to inferor Palme, 2001). improvetheirCI abundances. Abundancesofthe Elementsinthe SolarNebula 53 1.03.1.4.4Comparison withAnders and atoms/meteoriteabundance per10 6 silicon atoms Grevesseabundance table is1.546 ^ 0.045.Aseachofthe solarand meteoriteabundance measurements areindepen- Acomparison withthe Anders andGrevesse dent the error ofthe meanmaybe used which (1989) compilation (Table3) showsthatthereare gives1.546 ^ 0.008, correspondingto aratioof fewelements for which the difference between 35.16 ^ 0.65.Thus, thiscompilation andthatofAnders andGrevesse (1989) exceedsfivepercent.Thisisnot surprising log A ast log A met 1 : 546 asbothcompilations rely,atleast inpart,on the ¼ þ samesourcesofdata.Differencesabove10% are Thisyieldsasilicon abundance on the astro- nomicalscaleoflog A (Si) 7.546anda foundfor ,sulfur,andselenium. The ast ¼ newvalueslisted, asdiscussed above, aremore hydrogenabundance on the meteoritic scaleof log A (H) 10.45 or 2.84 1010 which is reliable.Differencesbetween5%and10% are met ¼ £ foundfor carbon,,antimony,, giveninTable3. Anders andGrevesse(1989) ,rhenium,andmercury. The rare calculated avalueof1.554 for the ratioofsolarto earthabundacesinCI meteoritesreported hereare meteoritic abundances,which leadstoahydrogen abundance of2.97 1010 on the meteoritic scale. the sameasthoseinPalmeandBeer(1993),which £ arebased on databyEvenson etal .(1978). There Lodders (2003) used aconversion factor of1.540 isanexcellent agreement ofthe REE dataof based on the ratioofphotospheric andmeteoritic Evenson etal .(1978)withthoseintwo Orgueil silicon. samplesdetermined byBeer etal .(1984),except Incolumn 6ofTable1,the meteoritedataare for asmall difference ofafewpercent inthe light givenonthe astronomicalscaleandincolumn 8 REE.Becauseofthisgeneralagreement the the ratios ofthe photospheric abundancestothe ,,strontium,,andcae- meteoritic abundancesarelisted.Theseratios are sium concentrations ofBeer etal .(1984)were displayed inFigure5.The solarabundancesof used inthe present compilation. The Anders and carbon,nitrogen,andoxygenarehigherbecause Grevesse(1989) REE andstrontium datawere theseelements areincompletely condensed inCI obtained byaveragingdatasets ofvarious authors. meteorites. Although the newsolaroxygen Theabsoluteconcentrations ofthe REEsare abundance is50% lowerthanthe oldvalue, the slightlylower,exceeding5%for strontium, solaroxygenabundance isstill afactor of2above cerium,andpraseodymium. The differencesin the meteoritevalue, which is,however,signifi- carbon andantimony concentrationsbetween cantly less thanthe depletion ofnitrogen(factor PalmeandBeer(1993) andAnders andGrevesse 41) andofcarbon (factor 9.9). Amajor revision in (1989) arerelated to the selection ofdata photospheric abundanceswasfoundfor literature. abundancesinCI chondrites which isnow ingoodagreement withmeteoritic areextremely variable.The high contents of abundances(GrevesseandSauval,1998). Asthe mercury inOrgueilmayreflectcontamination photospheric abundance ofboron appears to fit (PalmeandBeer,1993). Asingleanalysisof withmeteoritedata, atleast withinafactor of2, Ivuna(0.31 ppm) wasthereforeused inthe present which isalso the error assigned to the photo- compilation. The valueof0.31ppm isinagree- spheric determination byGrevesseand ment withamercurycontent inferred from nuclear Sauval(1998),lithium isthe only lightelement abundance systematics(PalmeandBeer,1993). thatisstrongly depleted inthe Sun. “The Li–Be–B problemisnow reduced to explaininghow the Sun candepleteLibyafactor of160 whereasBe andB 1.03.1.5SolarSystemAbundancesofthe arenot destroyed”(GrevesseandSauval,1998). Elements The recent revision ofthe photospheric lead 1.03.1.5.1 Comparison ofmeteoriteandsolar abundance broughtthiselement into agreement abundances withmeteoritic lead abundancestowithin10% (Biemont etal .,2000). InTable1the Si-normalized meteoriteabun- Incomparingmeteoritewithsolarabundances dancesofTable3(log A 6) areconverted to inFigure5only ratios areplotted wherethe error of Si ¼ the H-normalized abundances(log A H 12). The the photospheric abundancesisbelow0.1 dex,i.e., conversion factor betweenthe two s¼ caleswas below 25%. Inthe Figure5^ 10% variation is calculated bydividingthe H-normalized solar indicated.Thelargest discrepancybetween abundancesbythe Si-normalized meteoriteabun- solarandmeteoritic abundancesisnow found dances. The comparison wasmadefor all ele- for sulfur,scandium,strontium,andmanganese, mentswithanerrorofthe corresponding the only elements wherethe difference between photospheric abundance ofless than0.1 dex, solarandmeteoritic abundancesexceeds20%. i.e.,less than , 25%. Thirty-four elements Thesolarsulfur andscandium abundances qualified for thisprocedure, andthe logofthe hadbeenrevised byGrevesseandSauval average ratioofsolarabundance per10 12 H (1998). Theirnewvaluesarehigherthanthoseof 54 SolarSystemAbundancesofthe Elements

Figure5 Comparison ofsolarandmeteoritic abundances(see Table1). The elements C, N, andOareincompletely condensed inmeteorites. Li isconsumed byfusion processesinthe interior ofthe Sun,but not Be andB.Solarand meteoritic abundancesagree inmost caseswithin10%. Only the four elements S, Mn,Sc, andSrdifferbymorethan 20% from CI abundances. The difference isbelow 10% for 27 elements. Only elements withuncertaintiesofless than 25%inthe photosphereareplotted.

Anders andGrevesse(1989),which had agreed to thesedatathe abundancesfor individualmass within5%withCIabundances. The error inthe numbers canbe calculated usingthe elemental photospheric abundance ofstrontium is17% abundancesasgiveninTable1. Figure6isa accordingto GrevesseandSauval(1998)andthe plot ofabundancesversus mass number. The difference isthus not significant. The only element generally higherabundance ofevenmassesis thathasconsistently shown amajor deviation apparent. Plots for evenandodd mass numbers are withinthe stated error ismanganese(Figure5). In moreor less smooth, the latterforminga the compilation ofGrevesseandSauval(1998)an considerably smoothercurvethanthe former. uncertainty of7% isassigned to the manganese Historically,the so-called abundance rules,estab- abundance.Recently Prochaska andMcWilliam lished bySuess (1947),postulatingasmooth (2000) pointed out thatthereareso farunrecog- dependence ofisotopic abundancesonmass nized problems inthe photospheric determination number A ,especially ofodd-A nuclei, played an ofmanganesebyincorrecttreatment ofhyperfine important roleinestimatingunknown or badly splitting.Theseauthors alsomention thatsimilar determined abundances. Laterthisrulewas problems maybe involved withthe photospheric modified andsupplemented bytwo additional abundance determination ofscandium. The errors rules(Suess andZeh, 1973)inordertomakethe associated withthe photospheric abundancesofthe concept applicableto the now moreaccurate otherelements areall inexcess of10%,someare abundance data.However,the smoothness of considerably higher,e.g.,praseodymium with odd-A nuclei abundancesitselfhasbeenques- 20%. tioned (Anders andGrevesse, 1989; Burnett and The agreement betweenmeteoritic andsolar Woolum,1990). Figure7(a)isanenlargement ofa abundancesmustbe considered excellent and part ofFigure6. Thehigh abundance of 89 Y thereisnot much room leftfor furtherimprove- (Figure7(a)),anapparent discontinuity (Burnett ments. Obvious candidatesfor redetermination of andWoolum,1990),reflects the low neutron the solarabundancesaremanganeseandsulfur. capturecross-section ofadominantly s-process nucleus withamagic neutronnumber(50). Also,therearemajor breaksinthe abundance curveofodd-A nuclei inthe region ofmolyb- 1.03.1.5.2 Solarsystemabundancesversus denum,rhodium,,,andantimony mass number (Figure7(a)).Indetailthereisthereforeno smoothness ofodd-A nuclei withmass number. Theisotopic compositionsofthe elements Similararguments apply to the abundancesof arenot discussed inthischapter. The compilation odd-A isotopesofneodymium,,and byPalmeandBeer(1993) contains alist of (Figure7(b)).Theydonot follow a isotopesandtheirrelativeabundances. Lodders smoothtrendwiththe lowerabundance of (2003) hasprepared anewcompilation.From samarium. The Abundancesofthe Elements inthe ISM 55

Figure6 Solarsystemabundancesbymass number. Atoms withevenmassesaremoreabundant thanthosewith odd masses(Oddo–Harkins rule)(source PalmeandBeer,1993).

1.03.1.5.3 Othersourcesfor solarsystem authors also reported very high enrichmentsof abundances (29 CI)andarsenic (7.4 CI),perhaps acquired inthe£ Earth’s atmosphere.£ Emission spectroscopy ofthe solarcorona, Theseparticlesprobably comefrom the aster- solarenergetic particles(SEP)andthe compo- oid belt (Flynn,1994). TheyarebroughttoEarth sition ofthe solarwindyieldinformation on the bythe action ofthe Poynting–Robertson effect. composition ofthe Sun. Solarwinddatawereused Perhaps theyarederived from sourcesthatcontain for isotopic decomposition ofraregases. Coronal uncondensed volatilesfrom the innerpart ofthe abundancesarefractionated relativeto photo- nebula.Thiswouldbethe only exampleofaclear spheric abundances. Elementswithhighfirst enhancement ofmoderately volatileelements in ionization potentialaredepleted relativeto the solarsystemmaterial. rest (see Anders andGrevesse, 1989for details). The composition ofdust grains ofcometHalley hasbeendetermined withimpactionization time- of-flightmass spectrometers on boardthe Vega-I, 1.03.2 THE ABUNDANCES OF Vega-II, andGiotto spacecrafts.The abundances THE ELEMENTS IN THE ISM of16 elementsandmagnesium,which isused for normalization,areon average CI chondritic to 1.03.2.1 Introduction withinafactor of2–3,except for hydrogen, The solarsystemwasformed asthe result ofthe carbon,andnitrogenwhich aresignificantly collapseofacloudofpre-existinginterstellargas higherinHalleydust,presumably dueto the anddust.We shouldthereforeexpectaclose presence oforganic compounds(Jessberger etal ., compositionalrelationshipbetweenthe solar 1988). Thereisnoevidence for aclearenhance- systemandthe interstellarmaterialfrom which ment ofvolatileelements relativeto CI. itformed.If wemakethe assumption thatthe Many ofthe micron-sized interplanetary dust composition ofthe ISM hasremained unchanged particles(IDPs) haveapproximately chondritic since the formation ofthe solarsystem,wecanuse bulkcomposition (see Chapter1.26 for details). the localISM asameasureofthe originalpre- Porous IDPsmatch the CI composition betterthan solarcomposition. Differencesbetweenthe solar nonporous (smooth)IDPs. Onanaverage, IDPs systemandcurrent localISM wouldimply that show someenhancement ofmoderately volatile fractionation occurred duringthe formation ofthe andvolatileelements (see Palme, 2000). Arndt solarsystem,thatthe localISM composition etal .(1996) foundsimilarenrichmentsintheir changedaftersolarsystemformation or thatthe suiteof44chondritic particles(averagesize solarsystemformed inadifferent part ofthe 17.2 ^ 1.2 m m). The elements , , galaxy andthenmigrated to its present location. zinc, ,selenium,andrubidium were StudiesofsolarsystemandlocalISM composition enriched byfactors of2.2–2.7. Inaddition,these arethereforefundamentaltothe formation ofthe 56 SolarSystemAbundancesofthe Elements

Figure7 Enlarged parts ofFigure6:(a)mass range 70–140and(b)mass range 138–209. The abundancesofodd mass nuclei arenot asmoothfunction ofmass number,e.g.,YandSn. solarsystem,the natureofthe localISM andthe 1.03.2.2 The Natureofthe ISM generalprocessesleadingto low-mass star formation. The ISM isthe medium betweenthe stars. For The discovery ofpresolargrains inmeteorites present purposes,wewill also considerthe media has,for the first time, enabled the precisechemical immediately surroundingstars (generally con- andisotopic analysisofinterstellarmaterial(e.g., sidered ascircumstellarmedia)aspart ofthe ISM. Anders andZinner,1993;Chapter1.02). The huge Most ofthe matterinthe ISM isinthe form ofa variationsinthe isotopic compositionsofall the very tenuous gaswithdensitiesofless thanone elementsanalyzed inpresolargrains isinstark hydrogenatom percm 3 to perhaps amillion contrast to the basically uniform isotopic compo- hydrogenatoms percm 3 .For comparison the sition ofsolarsystemmaterials (see Figure1). terrestrialatmospherecontains about 1019 hydro- Thisuniformity wouldhaverequired aneffective genatoms percm 3 . isotopic homogenization ofall the materialinthe Interstellarmatteriscomprised ofbothgasand solarnebula, i.e.,gasanddust,duringthe early dust.The gasconsists ofatomic andpolyatomic stagesofthe formation ofthe solarsystem. ions andradicals,andalsoofmolecules. Itisthe The Abundancesofthe Elements inthe ISM 57 form ofhydrogeninthe ISM thatisused to last column the ratios ofthe z Ophabundances describe its nature, i.e.,ionized (H þ ),atomic (H), to the solarabundancesaregiven. InFigure8 or molecular(H 2 ),withincreasingdensity from theseratios areplotted againstcondensation the ionized to the moleculargas. The dust is temperatures(see Savage andSembach, 1996 primarily composed ofamorphouscarbons and for details).The abundancesofmany ofthe silicatesandmakesup , 1% ofthe mass ofthe highly volatileandmoderately volatileelements ISM.The dustparticleshavesizesinthe up to condensation temperaturesofaround900K nanometertomicrometerrange (e.g.,Mathis, (at10 2 4 bar) are, withinafactor of2,the samein 1990).The totalamount ofabsorption and the ISM andinthe Sun,independent ofthe scatteringalongany givenlineofsightiscalled condensation temperatures. Thissuggeststhat the interstellarextinction. The degree ofextinc- theseelements predominantly reside inthe gas tion dependsupon the wavelength, the sizeofthe phaseinthe ISMandthattheirelemental dust grains andtheircomposition. abundancesinthe ISM aresimilartothoseof the solarsystem. Athighercondensation tem- peraturesacleartrendofincreasingdepletions withincreasingcondensation temperaturesis 1.03.2.3 The ChemicalComposition ofthe ISM seen. The morerefractory elements arecondensed into grains inthe outflows ofevolved stars or 1.03.2.3.1 The composition ofthe interstellar perhaps inthe ejected remnantsassociated with gasandelementaldepletions supernovae explosions. The elements incorporated into grains inthe Thereareseveralelements whoseabundances ISM cannot be observed directly. Theyareunder- deviatefrom the generaltrend(e.g.,phosphorus represented inordepleted from the gasphase. andarsenic inFigure8). Itwill be important to Onlythe fraction ofanelement thatremains inthe findout duringthe courseoffutureworkwhether gasphasecanbe detected inthe ISM, provided thesedeviations reflectproblems withthe extre- thatthe element hasaccessibleandobservable mely complexanalysesorwhethertheyaretrue transitions. Initialmeasurements ofthe abun- variationsthatindicateparticularchemicalpro- dancesofelements inthe gasphasedateback to cessesinthe ISMor arecharacteristic of the mid-1960s,whenwiththe advent ofspace condensation instellaroutflows. missions,itbecamepossibleto eliminatethe Itshouldbeemphasized thatthe depletion absorbingeffects ofthe Earth’s atmosphere(e.g., pattern inthe ISM doesnot reflectthermodynamic Morton andSpitzer,1966). Withthe sensitive equilibrium betweendust andgas. Thetempera- spectrographsonboardthe InternationalUltra- tureinthe ISMissolow thatvirtuallyall violetExplorer(IUE)andthe HubbleSpace elements,includingthe raregases,shouldbe Telescope(HST, e.g.,the GoddardHigh Resol- condensed ingrains if thermodynamic equilibrium ution Spectrometer,GHRS,andmorerecently the betweendust andgasisassumed.The depletion Space TelescopeImagingSpectrograph, STIS) pattern ratherreflects conditions athighertem- many linesofsighthavenow beenstudied for peraturesestablished duringthe condensation of alarge numberofelements (e.g.,Savage and minerals inthe outflows ofdyingstars or super- Sembach, 1996;Howk etal .,1999;Cartledge novae explosions. Thispattern isthenfrozenin etal .,2001). the coldISM. Hydrogenandhelium arethe most abundant elements inthe ISM gasphase.Todatesome 30–40elements heavierthanhelium havebeen 1.03.2.3.2 The composition ofinterstellar observed andtheirgasphaseabundancesdeter- dust mined.Based on the existingdata(e.g.,Savage andSembach, 1996;Howk etal .,1999;Sofia and Asconcluded inthe previous section,the dust in Meyer,2001)the localISM sampled out to afew the ISM isprimarily composed ofthe elements kiloparsecsfrom the Sun appears to be rather carbon,oxygen,magnesium,silicon,andiron. uniform inchemicalcomposition. Thisargument isbased on the elementalmake-up Asanexampleweshow the resultsof ofthe solid phaseinthe ISM determined from the abundance determinations alongthe lineofsight depletions. However,the exactchemicaland towards z Oph( z Ophiuchus),amoderately mineralogicalcomposition ofthe dust inthe reddened starthatisfrequently used asstandard ISM canbe determined through infrared obser- for depletion studies. Moleculesareobserved vations ofthe absorption ofstarlightbycolddust alongthislineofsightandthe materialisa ( T < 20 K)alonglinesofsighttowarddistant blendofcool diffusecloudsandalarge cold stars,andalso bythe emission featuresfrom hot cloud.The atomic hydrogencolumn density is dust ( T < afewhundred kelvin) inthe regions log N (H) 21.12 ^ 0.10. InTable4 all dataare closeto stars.Such observationsrevealthe normalized¼ to 1012 atoms ofhydrogenandinthe spectralsignaturesofamorphousaliphatic and 58 SolarSystemAbundancesofthe Elements Table4 Abundancesofelements inthe gasphaseofthe ISM inthe direction of z Ophiucus.

Element T c Solarsystem z Ophiucus cool z Ophiucus cool/solar (K) logX SD (%) logX SD (%) Lit. Highly volatileelements Ar256.4056.0845(1) 0.48 Kr253.30 152.97 15(1) 0.47 C758.39 30 8.1435(1) 0.56 N120 7.93 30 7.90 15(1) 0.93 O1808.69 20 8.48 15(1) 0.62 Pb 427 2.05151.3440(1) 0.19 Cd 429 1.77 30 1.67 10 (2) 0.79 Tl448 0.81601.27 30 (1) 2.9 Moderately volatileelements S648 7.19 30 7.45 90 (1) 1.8 Zn660 4.66 20 3.9830 (1) 0.21 Te 6802.22 , 3.01 (1) , 6.2 Se 684 3.403.45 70 (1) 1.1 Sn720 2.12 90 2.16 25(1) 1.1 F736 4.45 90 4.26 60 (3) 0.65 Ge 8253.62 403.01 10 (1) 0.25 Cl863 5.26 90 5.27 60 (1) 1.0 B9082.87901.9525(1) 0.12 Ga 9183.11 251.99 15(1) 0.076 Na 970 6.30 55.36 25(1) 0.11 K1,000 5.11 354.0480(1) 0.085 Cu1,037 4.2810 2.92 5(1) 0.044 As1,1352.35?? 2.16 25(1) 0.65 P1,1515.44 10 5.07 75(1) 0.43 Mn1,190 5.5154.085(1) 0.037 Li 1,2253.30 10 1.73 15(1) 0.027 Mg-silicatesandmetallic FeNi Cr1,301 5.67 53.45(1) 5.3 10 –3 Si 1,311 7.55 156.245(1) 4.9 £ 10 –2 Fe 1,337 7.49205.245(1) 5.7 £ 10 –3 Mg 1,3407.56156.33 5(1) 5.9 £ 10 –2 Ni 1,354 6.23 10 3.515(1) 1.9 £ 10 –3 Co1,3564.90 10 2.1530 (1) 1.8 £ 10 –3 £ Refractory elements V1,455 3.99 5 , 2.06 (1) , 1.2 10 –2 Ca 1,5186.33 52.61 15(1) 1.9 £ 10 –4 Ti 1,5984.95141.91 10 (1) 9.2 £ 10 –4 £ 2 4 12 T c —condensation temperaturesat10 bar(Wasson,1985),except B(LaurettaandLodders,1997); log X —logofabundancesrelativeto 10 atoms ofH;SD’s aregiventothe nearest 5%. (1) Savage andSembach (1996); (2) Sofia etal .(1999); (3) Snow andYork(1981). aromatic hydrocarbons andamorphous silicates. indicatesthattheyhaveanolivine-typestoichi- Indenseclouds(densitiesofthe orderof103 –105 ometry wherethe depletions arelargest,anda hydrogenatoms percm 3 ),wherethe matteris mixed oxide/silicatestoichiometry wherelower well-shieldedfromthe destructiveeffects of depletions indicatethatsomedust erosion has stellarUV-light,weobservemolecularspecies occurred (e.g.,SavageandSembach,1996; such asH 2 ,H2 O, CO, CO2, andCH 3 OH. Jones,2000 andreferencestherein).In The interstellarcarbon grains containboth the lowerdensity regions abovethe galactic aliphatic andaromatic C—HandC—C bonds; planeweclearly see adifferent dust stoichi- bothareobserved inabsorption andemission in ometry (e.g.,Savage andSembach, 1996). This the ISM, but beyondthistheirexactcomposition change incomposition ispresumably areflection isnot known. Giventhe seemingly uniform ofthe effects ofshock wavesthathavelifted elementalabundancesinour localISM wemight interstellarcloudshigh abovethe planeandthat expectthatthe dust composition wouldalso be have, atthe sametime, eroded anddestroyed chemicallyuniform. However,the inferred somefraction ofthe dust incorporated into these elementalcomposition ofsilicatesinthe ISM clouds. The Abundancesofthe Elements inthe ISM 59

Figure8 Abundancesofelements alongthe lineofsighttowards z Oph( z Ophiuchus),amoderately reddened starthatisfrequently used asstandardfor depletion studies. The ratios of z Ophabundancestothe solarabundances areplotted against condensation temperatures. The abundancesofmany ofthe highly volatileandmoderately volatileelements up to condensation temperaturesofaround900 Kare, withinafactor of2,the sameinthe ISM andin the Sun. Athighercondensation temperaturesacleartrendofincreasingdepletions withincreasingcondensation temperaturesisseen. Itisusually assumed thatthe missingrefractory elements areingrains (source Savage and Sembach, 1996).

1.03.2.3.3 Did the solarsysteminherit two elements overhundredsofmillions ofyears the depletion ofvolatileelements withoutthorough remixingatthe beginningofthe from the ISM? solarsystem,wouldproduce alarge range ininitial 87 Sr/86 Srratios atthe beginningofthe solarsystem Most meteoritesaredepleted inmoderately andshouldhavebeenobserved insolarsystem volatileandhighly volatileelements (see Figures materials (see Palme, 2001). 2–4). The terrestrialplanets Earth, Moon,Mars, andthe asteroid Vestashow similaroreven strongerdepletions (e.g.,Palme etal .,1988; 1.03.2.3.4The ISM oxygenproblem Palme, 2001). The depletion patterns inmeteorites andinthe innerplanets arequalitatively similarto Snow andWitt (1996) andothers argued that thoseinthe ISM.Itisthus possiblethatthe the composition ofthe ISM isdifferent from the materialinthe innersolarsysteminherited the composition ofthe Sun.Based on stellar depletionsfromthe ISMbythe preferential compositionaldata, theseauthors concluded that heavy elements inthe Sun areonly two-thirdsof accretion ofdust grains andthe loss ofgasduring the solarcomposition,withthe implication that the collapseofthe molecularcloudthatled to the the heavy elementseitherfractionated from formation ofthe solarsystem. Thereis,however, hydrogenduringformation ofthe Sun (Snow, littlesupport for thishypothesis: 2000)orthatthe Sun wasformed inadifferent (i)The generaluniformity ofthe isotopic place inthe MilkyWaywherethe heavy element compositions ofsolarsystemmaterials andthe abundanceswerehigher. Anotherpossibility was extremevariations inpresolargrains suggest thattherewereadditional“hidden” reservoirs. In isotopic andelementalhomogenization atthe particular,therewasaproblemoftoo much beginningofthe solarsystem. interstellaroxygenwhenusingthe solaroxygen (ii)Thisapplieslikewiseto all systems invol- abundance asstandard.Wherecouldthatexcess vingradioactivenuclei thatareused for dating.For oxygenbe stored?Some20% ofthe solaroxygen example, the extremely uniform strontium isotopic abundance must havebeencombined with composition inall solarsystemmaterials atthe magnesium,silicon,andiron,etc.inthe form beginningofthe solarsystem,4.566 billion years ofthe amorphous silicatesobserved inthe ISM. ago,indicatesthatthe strontium isotopeswere Another40% wasdirectly observed inthe gasas homogenized atthattime, including 87 Sr,the decay atomic oxygen(Meyer etal .,1998). The “miss- productof 87 Rb ( T 5 1010 yr). Inthe ISM, ing”40% ofoxygenremained elusive;itcould 1 = 2 ¼ £ the refractory strontium isingrains andthe volatile not be inthe form ofmolecules,e.g.,H2 O rubidium inthe gasphase.The separation ofthese andCO, becausetheywerenot observed in 60 SolarSystemAbundancesofthe Elements sufficient abundance.The recent re-evaluation of ArndtP.,BohsungJ.,Maetz M.,andJessbergerE.(1996) The the solaroxygenabundance (Holweger,2001; elementalabundancesininterplanetary dust particles. MeteoriticsPlanet. Sci. 31,817–833. Allende Prieto etal .,2001),andofinterstellar BeerH.,WalterG.,MacklinR.L.,andPatchett P.J.(1984) oxygen(Sofia andMeyer,2001),hasresolved Neutron capturecross sections andsolarabundancesof thisproblem. The newsolaroxygenisnow only 160,161Dy, 170,171Yb, 175,176Lu,and 176,177Hf for the s-process , 60% ofits previous valueandso thereis,and analysisofthe 176 Lu. Phys.Rev. C30, indeed neverwas,aproblem. Recent interstellar 464–478. Begemann F.(1980) Isotopeanomaliesinmeteorites. Rep. abundance determinations usingobservations of Prog.Phys. 43 ,1309–1356. solar-typeFandGstars inthe galactic neighbor- Biemont E.,GarnirH.P.,PalmeriP.,LiZ.S.,andSvanbergS. hood(Sofia andMeyer,2001)suggest thatmany (2000) Newf-valuesinneutrallead obtained bytime- elementalabundancesinthe ISM arecloseto resolved laserspectroscopy,andastrophysicalapplications. solarabundances. Mon. Not. Roy. Astronom. Soc. 312,116–122. Bischoff A.,PalmeH.,AshR.D.,Clayton R.N.,Schultz L., Herpers U.,Sto¨fflerD.,GradyM.M.,PillingerC.T.,Spettel B.,WeberH.,GrundT.,Endreß M.,andWeberD.(1993) 1.03.3 SUMMARY Paired Renazzo-type(CR)carbonaceous chondritesfrom the Sahara. Geochim. Cosmochim. Acta 57 ,1587–1603. Updated solarphotospheric abundancesare Burnett D.S.andWoolum D.S.(1990) The interpretation of compared withmeteoritic abundances. Itis solarabundancesatthe N 50neutron shell. Astron. Astrophys. 228 ,253–259. ¼ shown thatonly onegroup ofchondritic meteor- Cameron A.G.W.(1973) Abundancesofthe elements inthe ites,the CI chondrites,matchessolarabundances solarsystem. Space Sci.Rev. 15 ,121–146. inrefractory lithophile, siderophile, andvolatile Cartledge S.I.B.,MeyerD.M.,andLauroesch J.T.(2001) elements. All otherchondritic meteoritesdiffer Space telescopeimagingspectrographobservationson from CI chondrites. The agreement betweensolar interstellaroxygenandkrypton intranslucent clouds. Astrophys. J. 562,394–399. andCIabundancesfor all elements heavierthan Clayton R.N.(1993) Oxygenisotopesinmeteorites. Ann. Rev. oxygenandexcludingraregaseshasconstantly EarthPlanet. Sci. 21,115–149. improved since Goldschmidt(1938)publishedhis CowleyC.R.(1995) Anintroduction to cosmochemistry. first comprehensivetableofcosmic abundances. Cambridge University Press,Cambridge, 480pp. Cunha K.andSmithV.V.(1999) Adetermination ofthe solar The abundancesof39 nongaseous elements in photospheric boron abundance. Astrophys.J. 512, the Sun haveassigned errors below 30%. Only the 1006–1013. four elements sulfur,manganese, scandium,and Dodd R.T.(1981) Meteorites,apetrologic-chemicalsynthesis . strontium differbymorethan20% from CI Cambridge University Press,Cambridge, 368pp. abundances. The difference isbelow 10% for 27 Dreibus G.,PalmeH.,SpettelB.,ZipfelJ.,andWa¨nke H. (1995)Sulfur andselenium inchondritic meteorites. oftheseelements. The agreement between Meteoritics 30,439–445. meteoritic andsolarabundancesmust therefore Evenson N.M.,Hamilton P.J.,andO’Nions R.K.(1978) be considered excellent andthereisnot much Rare-earthabundancesinchondritic meteorites. Geochim. room leftfor furtherimprovements. Obvious Cosmochim. Acta 42 ,1199–1212. candidatesfor redetermination ofsolarabun- Flynn G.J.(1994)Interplanetary dust particlescollected from the stratosphere: physical,chemical,andmineralogical dancesaremanganeseandsulfur. The limiting propertiesandimplicationsfor theirsources. Planet. factor inthe accuracyofmeteoriteabundancesis Space.Sci. 42 ,1151–1161. the inherent variability ofCIchondrites,primarily GoldschmidtV.M.(1938)Geochemische Verteilungsgestze the Orgueilmeteorite. derElementeIX.Die Mengenverha¨ltnissederElementeund The ISM from which the solarsystemformed derAtom-Arten. SkrifterUtgitt avDetNorske Vidensk. Akad.SkrifterI.Mat. Naturv. Kl. No. 4, Oslo,1937,148pp. hasvolatileandmoderately volatileelements, GrevesseN.andSauvalA.J.(1998)Standardsolar withinafactor of2,the samecomposition asthe composition. Space Sci.Rev. 85,161–174. Sun. The morerefractory elements ofthe ISM are GrossmanL.andLarimerJ.W.(1974)Earlychemical depleted from the gasandareconcentrated in history ofthe solarsystem. Rev. Geophys.Space Phys. 12, grains. 71–101. Harkins W.D.(1917) The evolution ofthe elements andthe stability ofcomplexatoms. J.Am. Chem. Soc. 39,856–879. HolwegerH.(2001) Photospheric abundances:problems, REFERENCES updates,implications. In SolarandGalactic Composition (ed.R.F.Wimmer-Schweinsgruber).AmericanInstituteof Allende Prieto C.,Lambert D.L.,andAsplundM.(2001) The Physics,pp. 23–30. forbiddenabundance ofoxygeninthe sun. Astrophys. J. 556 , HowkC.,Savage B.D.,andFabianD.(1999) Abundances L63–L66. andphysicalconditionsinthe warm neutralmedium Allende Prieto C.,Lambert D.L.,andAsplundM.(2002) A toward m Columbae. Astrophys. J. 525 ,253–293. reappraisalofthe solarphotospheric C/Oratio. Astrophys. J. JessbergerE.K.,ChristoforidisA.,andKisselJ.(1988)Aspect 573,L137–L140. ofthe major element composition ofHalley’s dust. Nature Anders E.andGrevesseN.(1989) Abundancesofthe elements: 332,691–695. meteoritic andsolar. Geochim. Cosmochim. Acta 53 , Jochum K.P.(1996) Rhodium andotherplatinum- 197–214. elements incarbonaceouschondrites. Geochim. Cosmochim. Anders E.andZinnerE.(1993) Interstellargrains inprimitive Acta 60,3353–3357. meteorites:diamond, silicon carbide,andgraphite. Jochum K.P.,Stolz A.J.,andMcOrist G.(2000) and Meteoritics 28 ,490–514. incarbonaceouschondrites:constraintsonthe solar References 61

systemandprimitivemantleniobium/tantalum,zirconium/ bitmorestrontium. LunarPlanet. Sci. XXIII, niobium,andniobium/uranium ratios. MeteoriticsPlanet. 1147–1148. Sci. 35 ,229–235. Rocholl A.andJochum K.P.(1993) Th, Uandothertrace JonesA.P.(2000) Depletion patterns anddust evolution inthe elements incarbonaceous chondrites:implications for the interstellarmedium. J.Geophys. Res. 105 ,10257–10268. terrestrialandsolarsystemTh/Uratios. Earth.Planet. Sci. LaurettaD.S.andLodders K.(1997) The cosmochemical Lett. 117,265–278. behavior ofberyllium andboron. Earth.Planet. Sci.Lett. Russell H.N.(1941) The cosmicalabundance ofthe elements. 146 ,315–327. Science 94 ,375–381. Lodders K.(2003) Solarsystemabundancesandcondensation SavageB.D.andSembach K.R.(1996)Interstellar temperaturesofthe elements. Astrophys. J. 591,1220–1247. abundancesfrom absorption-lineobservationswiththe MathisJ.S.(1990) Interstellardust andextinction. Ann. Rev. HubbleSpace Telescope. Ann. Rev. Astron. Astrophys. 34 , Astron. Astrophys. 28 ,37–70. 279–329. MeyerD.M.,JuraM.,andCardelliJ.A.(1998)The definitive Snow T.P.(2000) Composition ofinterstellargasanddust. abundance ofinterstellaroxygen. Astrophys.J. 493, J.Geophys. Res. 105 ,10239–10248. 222–229. Snow T.P.andWitt A.N.(1996) Interstellardepletions Morton D.C.andSpitzerL.(1966) Linespectraofdeltaandpi updated: whereall the atoms went. Astrophys. J. 468 , Scorpii inthe far-ultraviolet. Astrophys. J. 144,1–12. L65–L68. NicolussiG.K.,PellinM.J.,LewisR.S.,DavisA.M.,Amari Snow T.P.andYorkD.G.(1981) The detection ofinterstellar S.,andClayton R.N.(1998)Molybdenumisotopic fluorineinthe lineofsighttowardDeltaScorpii. Astrophys. J. 247 ,L39–L41. composition ofindividualpresolarsilicon carbide grains Sofia U.J.andMeyerD.M.(2001) Interstellar from the Murchison meteorite. Geochim. Cosmochim. Acta abundance standardsrevisited. Astrophys.J. 554, 62,1093–1104. L221–L224. PalmeH.(2000) Aretherechemicalgradients inthe innersolar Sofia U.J.,MeyerD.M.,andCardelliJ.A.(1999) The system? Space Sci.Rev. 92,237–262. abundance ofinterstellartinandcadmium. Astrophys. J. PalmeH.(2001) Chemicalandisotopic heterogeneity in 522,L137–L140. protosolarmatter. Phil. Trans. Roy. Soc.London A359 , Suess H.(1947) U¨ berkosmische Kernha¨ufigkeitenI.Mitteilung: 2061–2075. EinigeHa¨ufigkeitsregeln undihreAnwendungbei der PalmeH.andBeerH.(1993) Abundancesofthe elements in Abscha¨tzungderHa¨ufigskeitwertefu¨rdie mittelsschweren the solarsystem. InLandolt-Bo¨rnstein,Group VI: Astron- undschwerenElemente: II.Mitteilung: Einzelheiteninder omy andAstrophysics:Instruments; Methods; SolarSystem Ha¨ufigskeitverteilungdermittelschwerenundschweren (ed.H.H.Voigt). Springer,Berlin,vol. 3(a),pp. 196–221. Kerne. Z.Naturforsch. 2a ,311–321,604–608. PalmeH.,LarimerJ.W.,andLipschutzM.E.(1988) Suess H.E.andUreyH.C.(1956) Abundancesofthe elements. Moderately volatileelements. In Meteoritesandthe Early Rev. Mod.Phys. 28 ,53–74. SolarSystem (eds. J.F.Kerridge andM.S.Matthews). Suess H.E.andZeh H.D.(1973) The abundancesofthe heavy University ofArizonaPress,Tucson,pp. 436–461. elements. Astrophys.Space Sci. 23,173–187. Prochaska J.X.andMcWilliamA.(2000) Onthe perils of Wasson J.T.(1985) Meteorites:TheirRecordofEarly Solar- hyperfinesplitting: areanalysisofMnandScabundance systemHistory.W.H.Freeman,NewYork, 267pp. trends. Astrophys. J. 537,L57–L60. WolfD.andPalmeH.(2001) The solarsystemabundancesofP ReamesD.V.(1998)Solarenergetic particles:sampling andTiandthe nebularvolatility ofP. MeteoriticsPlanet. coronalabundances. Space Sci.Rev. 85,327–340. Sci. 36,559–571. RichterS.,Ott U.,andBegemann F.(1992) S-process Zhai M.andShawD.M.(1994)Boron cosmochemistry:Part 1. isotopeanomalies:,samarium anda Boron inmeteorites. Meteoritics 29,607–615.

q 2003,ElsevierLtd.All rights reserved TreatiseOnGeochemistry Nopart ofthispublication maybe reproduced, stored inaretrievalsystemor ISBN (set):0-08-043751-6 transmitted inany form or byany means,electronic, mechanical,photocopying, recordingor otherwise, without prior writtenpermission ofthe Publisher. Volume1; (ISBN: 0-08-044336-2); pp. 41–61