1

Revision21

Apetrologicalassessmentofdiamondasarecorderofthemantlenitrogencycle2

SamiMikhail1,2andDanielHowell2,33

1-DepartmentofEarth&EnvironmentalSciences,TheUniversityofSt.Andrews,UnitedKingdom4

2-SchoolofEarthSciences,WillsMemorialBuilding,TheUniversityofBristol,UnitedKingdom5

3-InstituteofGeoscience,GoetheUniversity,Frankfurt,Germany6

Abstract:7

Nitrogen is fundamental to the evolution of Earth and the life it supports, but for reasons poorly8

understood, it is cosmochemically themostdepletedof thevolatileelements.The largest reservoir in the9

bulksilicateEarthisthemantle,andknowledgeofitsnitrogengeochemistryisbiased,because≥90%ofthe10

mantlenitrogendatabasecomesfromdiamonds.However,itisnotcleartowhatextentdiamondsrecordthe11

nitrogencharacteristicsofthefluids/meltsfromwhichtheyprecipitate.Thereisongoingdebateregarding12

thefundamentalconceptofnitrogencompatibilityindiamond,andempiricalglobaldatasetsrevealtrends13

indicative of nitrogen being both compatible (fibrous diamonds) and incompatible (non-fibrous14

monocrystalline diamonds). A more significant and widely overlooked aspect of this assessment is that15

nitrogen is initially incorporated into thediamond latticeas singlenitrogenatoms.However, this formof16

nitrogenishighlyunstableinthemantle,wherenitrogenoccursasmolecularformslikeN2,orNH4+,bothof17

whichare incompatible in thediamond lattice.A reviewof theavailabledata shows that in classic terms,18

nitrogenisthemostcommonsubstitutionalimpurityfoundinnaturaldiamondsbecauseitisofverysimilar19

atomic size and charge to carbon. However, the speciation of nitrogen, and how these different species20

disassociateduringdiamond formation to create transientmonatomicnitrogen, are the factorsgoverning21

nitrogenabundance indiamonds.This suggests the counter-intuitivenotion thatanitrogen-free (Type II)22

diamondcouldgrowfromaN-richmediathatissimplynotundergoingreactionsthatliberatemonatomicN.23

Incontrast,anitrogen-bearing(TypeI)diamondcouldgrowfromafluidwithalowerNabundance,inwhich24

reactions are occurring to generate (unstable) N atoms during diamond formation. This implies that25

diamond’s relevance to nitrogen abundance in the mantle is far more complicated than currently26

2

understood.Therefore,furtherpetrologicalinvestigationsarerequiredtoenableaccurateinterpretationsof27

whatnitrogendatafrommantlediamondscantellusaboutthedeepnitrogenbudgetandcycle.28

29

1.Introduction30

The importance of the geodynamic nitrogen cycle should not be understated. Nitrogen is fundamental to the31

evolutionofEarthandthelifeitsupports.ThisisdemonstratedbytheEarth’satmospherebeingmadeupofroughly32

78%nitrogen(PorcelliandPepin,2003),andnitrogen isakeyelement in thestructureofmoleculesvital to life,33

includingaminoacids,proteins,andnucleicacids(BusignyandBebout,2013;referencestherein). Ithasalsobeen34

suggested that oxidation of ammonic nitrogen in themantle can generate water on the surface (Li and Keppler,35

2014),andthatnitrogenmayhaveplayedan importantrole inheatingEarth’ssurfaceabovethefreezingpointof36

waterdespiteafaintyoungsun(SaganandChyba,1997;Goldblattetal.,2009).Therefore,anunderstandingofthe37

geodynamicnitrogencycleisfundamentaltounderstandingthedevelopmentofthehabitableEarth(Canfieldetal.,38

2010).39

Placing firmconstraintson the fluxofnitrogenbetweentheEarth’sreservoirs is fundamental forquantitative40

models,butalsoverychallenging.EmpiricaldatausedtotracethebehaviorofnitrogeninEarth’smantleacrossdeep41

time (Ga timescales) come from geochemical studies of nitrogen-bearing mantle xenoliths and xenocrysts. The42

mantlemineralmostwidely used for this purpose is diamond (>90%of themantle nitrogen database), inwhich43

nitrogen is the most common and abundant lattice-bound impurity, with concentrations ranging from below44

detection to >5000 ppm (Fig.1a-c). However, there are distinctions between mantle diamond-types. The45

monocrystalline diamonds typically contain less nitrogen than coated diamonds (average = 200-300 ppm),46

polycrystalline, garnet-bearing diamondites are also known to contain higher nitrogen abundances than47

monocrystallinediamonds(average=600ppm;seeMikhailetal.,2013),andsublithosphericdiamondsaretypically48

found to contain nitrogen below detection limits (average <20 ppm; Harte, 2010). The analysis of nitrogen in49

diamondbynon-destructiveFTIRisrelativelysimple,whichhasledtothenitrogen-abundanceclassificationsystem50

(see Howell et al., 2012; references therein). Diamonds containing nitrogen are termed Type I, while Type II51

diamonds contain no detectable nitrogen (the lower limit depends upon the technique used; see Mikhail et al.,52

2014a). Kaiser and Bond (1959) were the first to show that differences in the FTIR spectra correlated with the53

detectionof nitrogenbymass spectrometry, leading to the concept that nitrogen is a lattice-bound substitutional54

3

impurityinTypeIdiamonds.However,itwasunknownifnitrogenwasaprimaryorsecondaryimpurity(Milledge55

and Meyer, 1962). Our current understanding is that single N atoms substitute for single C atoms on a growth56

interface, andbecome incorporated into the lattice as point defects. As diamond formation in themantle spans a57

largetemporalandspatialrange(from0.6to3.5Gaand<150to>600kmdepth;Gurneyetal.,2010;Shireyetal.,58

2013),thecarbonandnitrogengeochemistryofdiamondshavebeenusedtoplaceconstraintsonthenatureofdeep59

volatilecyclesthroughthemantle(seeCartignyetal.(2014)forareview).60

Tousethepresenceofnitrogenindiamondtoconstraintheextentandfluxofthemantle'snitrogenreservoir(s)61

requiresanunderstandingofthepartitioningbehaviourofnitrogenintodiamondasafunctionofP-T-X.Themost62

commonmethod applied to constrain nitrogen uptake into diamond is reverse modeling of empirical data. This63

methodassignsafluid-diamondpartitioncoefficientfornitrogenbyfittingtheco-variationsofcarbon-isotopevalues64

and nitrogen abundances to a curve for a given temperature (using theoretical equilibrium carbon-isotope65

fractionation factorsthatassumethespeciationofcarbonasmethaneorcarbonate: Javoyetal.,1984;Boydetal.,66

1987,1992;BoydandPillinger,1994;Bulanovaetal.,2002,2014;Cartignyetal.,1997,1998a,b,2001,2003,2004,67

2009;Klein-BenDavidetal.,2010;Harteetal.,1999;Haurietal.,2002;Howelletal.,2013,2015a;Gautheronetal.,68

2005;Hutchisonet al., 1997;Mikhail et al., 2013,2014a; Stachel andHarris,2009;Palot et al., 2009,2012,2014;69

Thomassot et al., 2007, 2009), where T is independently determined using the degree of nitrogen aggregation70

and/orgeothermometryonpairedsilicateinclusions(discussedbyStachelandHarris,2008).71

However,thisapproachhasnotexplainedhowsingleatomsofnitrogenhavebeengeneratedinthemantlefor72

incorporationintothediamondlattice.AsmonatomicNishighlyunstable,nitrogenwouldbeexpectedtooccurin73

othermorestable forms (Mikhail andSverjensky,2014).Therefore, to interpret themeaningof thenitrogendata74

recordedindiamonds,andtomodelanypossibleequilibriumstable-isotopefractionationof15N/14Ntoexplainthe75

large rangeofδ15Ndata (Mikhail et al.,2014a)anunderstandingof thepartitioningbehaviourofnitrogenduring76

diamondformationisrequired(e.g.Howelletal.,2015a).Inthispaperweevaluatethebehaviorofnitrogenduring77

diamondformation,andshowthattheincorporationofnitrogenintodiamondisprimarilycontrolledbydiamond-78

formingreactionsinvolvingnitrogeninthegrowthmedium.79

80

2.Nitrogencompatibilityindiamond81

4

Considering the rules of trace element partitioning, first laid down by Goldschmidt (1937), the presence of82

nitrogenasastructuralconstituentindiamondisnotunexpected,becausecarbonandnitrogenhavesimilarcharges83

andionicradii.Thefirstrulestatesthatatomsofthesamechargeandradiuswillenteracrystallatticewithequal84

ease. This provides the basis for modern elemental substitutionmodels, and in the simplest sense remains an85

accurateassumption,e.g.NisubstitutingforFeinolivine(BlundyandWood,2003).Anyresultingchargeimbalance86

will require balancing via a double substitution. For example, in diopside at 11.5 GPa and 750 °C, monovalent87

ammoniumcanbe incorporated (up to1000ppm) throughadouble-substitutionmechanism,witha trivalent ion88

balancingthecharge:(Ca2+)M2+(Mg2+)M1⇔(NH4+)M2+(Al3+orCr3+)M1(Watenphuletal.,2010).89

Thecompatibilityofanelementismostcommonlyexpressedinoneoftwoways(foramoredetaileddiscussions90

seeBlundyandWood,2003).Assumingatwo-phasesystem(diamond+fluid)therelativepartitioningofNandC91

can be represented by a partition coefficient (D), which reflects the compatibility of N/C relative to fluids/melts92

(Eq.1).However, due to the crystal symmetry of diamond,N can be incorporated on any crystallographic site. In93

addition, C + Nmix on the same sites, meaning a chemical control (i.e., stoichiometric constraint) will drive the94

partitioningbehavior(Kd–Eq.2).Therefore,thepartitioningbehaviorcouldbeconsideredastheratioofpartition95

coefficients(Kd–Eq.2).Thisapproachissimilartowhatistypicallyimplementedfordescribinge.g.Fe-Mgexchange96

equilibriumbetweene.g.olivineandsilicatemelts(Toplis,2005).NotethatififDNorKN≥1,theelementisdescribed97

ascompatible.98

Eq.1 𝐷"#$%&'()*+,-$) = "0123456

"78916 99

Eq.2 𝐾";<=#$%&'()*+,-$) = "

78916÷"0123456

?78916÷?0123456 100

101

2.1.Previousmodels102

Thereisnoconsensusintheliteratureregardingtheoverallcompatibilityofnitrogenindiamond.Instead,there103

are threecontradictoryarguments thatare (confusingly)all supportedbyempiricaldatasets.Thesecontradictory104

arguments are as follows: [1] nitrogen is incompatible in diamond (Boyd et al., 1994; Cartigny et al., 2001), [2],105

nitrogen is compatible in diamond (Stachel and Harris, 2009), and finally, [3] the compatibility of nitrogen in106

diamondisredox-sensitive(Deinesetal.,1989;Thomassotetal.,2007;SmithandKopylova,2014).107

5

Theevidencefornitrogenbeingincompatible:Theglobaldatasetfortheabundanceofnitrogenindiamondshowsa108

distributionskewedtowardszero fornon-coatedmonocrystallinemantlediamonds(Fig.1a-b),andthispattern is109

characteristic of an incompatible element (Ahrens, 1954). If N is incompatible, the incorporation of nitrogen in110

diamondmust be governed by a kinetic process rather than by equilibriumdistribution, despite the similarity in111

charge and ionic radius between atomic N and C. This would mean that nitrogen-rich diamonds occur as a112

consequenceofrapiddisequilibriumgrowthandtheN/Cratioofdiamondapproachesthatoftheprecipitatingfluid113

or melt. In contrast, Type II diamonds would be produced by slow growth under near-equilibrium conditions,114

regardlessofthenitrogencontentofthegrowthmedium(Cartignyetal2001).Inaddition,arecentstudybyPalotet115

al.(2013)thatdeterminedtheco-variationsforδ13CvaluesandN-abundancesinsituusingSIMS,arguedthatthese116

data are consistent with bulk-sample data (e.g. Cartigny et al., 2001), and concluded that nitrogen behaved117

incompatiblyduringdiamondformation.118

Theevidencefornitrogenbeingcompatible:StachelandHarris(2009)arguethatnitrogeniscompatibleindiamond119

irrespective of fO2, citing the decreasing nitrogen contents from core to rim observed in high-pressure high-120

temperature(HPHT)syntheticdiamondsgrownunderreducingconditionsinFe-Nisolventcatalysts(e.g.Reutskyet121

al2008)andundermoreoxidizingconditions(de-carbonationofcarbonate)inthepresenceofsilicates(Pal’yanovet122

al2002).The conceptofnitrogenbeing compatible is also consistentwith recentdata fromsinglepopulationsof123

diamonds(e.g.Thomassotetal.,2007),andwithinsingle(zoned)diamonds(Smartetal.,2011;Palotetal.,2014;124

Pettsetal.,2015).Thomassotetal.(2007)analyzedasuiteofdiamondsfromasingleperidotitexenolith(assumed125

to have formed in a single event) and found co-variations between the carbon-isotope values and nitrogen126

abundancesthatfittheequilibriumRayleighfractionationmodeloutlinedbyCartignyetal.(2001).Thesedataresult127

in a calculated KN of 2 between diamond and a hypothetical methanogenic fluid; meaning nitrogen is twice as128

compatible in the diamond relative to the fluid. Similar studies used high spatial-resolution SIMS profileswithin129

individualdiamonds that showedpronouncedoscillatoryzoningandrecordedco-variationsofNcontentsvs. δ13C130

(Smartetal.,2011;Palotetal.,2014;Pettsetal.,2015).Byapplyingthesamereverse-modelingapproachasCartigny131

etal.(2001)andThomassotetal.(2007)aKNvalueof5wascalculatedfordiamondprecipitationfromacarbonatitic132

fluid(Smartetal.,2011).Collectively, thesetwodatasetssuggestthatnitrogeniscompatibleunderbothoxidizing133

and reducing conditions conducive to diamond-formation, assuming a simple two-phase relationship between134

diamondandfluid(i.e.notaccountingforthepossibilityofammoniumpartitioningintosilicatesorotherphases).135

6

Theevidencefornitrogencompatibilitybeingredoxsensitive:Severalstudieshaveproposedthatthecompatibility136

ofnitrogenindiamondisredox-sensitive(Deinesetal.,1989;Thomassotetal.,2007).Themostrecentmodelargues137

thatnitrogeniscompatibleindiamondonlyintheabsenceofFe0orFe-Nialloys(SmithandKopylova,2014).Thisis138

largely based on the observation that the decrease in average nitrogen abundance in diamonds with increasing139

depthoforigincanbecorrelated(crudely)withthepredictedincreasingmolefractionofFe0 inperidotiticmantle140

withdepth(Frostetal.,2004;Rohrbachetal.,2007).SmithandKopylova(2014)citeexperimentaldatafordiamond141

synthesis in a sealed capsule with a strong redox gradient where one end contained Fe-metal and the other142

containedcarbonate(Palyanovetal.2013).ThediamondsincontactwithFe0contained100–200ppmN,whereas143

thediamondsinthecarbonate-meltportioncontained1000–1500ppmN.However,despiteonepotentialexception144

(Howell et al., 2015b), there is littleevidence to justifymodelingdiamond-formation in themantle inequilibrium145

withametallicsolventcatalyst,especiallyconsideringthatthetheoreticalmetalsaturationinthemantleisonly1wt146

%Fe0(Frostetal.,2004).Therefore,theapplicationofthestudybySmithandKopylova(2014)tonaturalsystemsis147

limited. The total nitrogen abundance of the charges used in all diamond-synthesis experiments is unknown (i.e.148

neverreported).Ergo,itisnotpossibletodetermineDNorKNvaluesduringexperimentaldiamondformationusing149

theexistingpublisheddatasets.150

151

3.Anewpetrologicalassessmentoftheincorporationofnitrogeninmantlediamonds152

Theprecedingreviewdemonstrates,quitesurprisingly,thatthecompatibilityofnitrogenindiamondremainsa153

highlydebatedtopic.Thomassotetal.(2007)andSmartetal.(2011)showconvincingdatathatindicatenitrogenis154

compatibleindiamond,withKNvaluesof2and5duringdiamondprecipitationfromCH4-richandCO32--richfluids155

respectively (at 1200°C). However, these conclusions do not explain why the global distribution of nitrogen156

concentrationsforinnon-coatedmonocrystallinediamondsisskewedtowardszero(akintoincompatiblebehavior;157

Fig.1a-b),whereasthedistributionforthefibrousgrowthofcoateddiamonds isGaussian,peakingatca800ppm158

(akintocompatiblebehavior;Fig.1c).Equallyperplexingarethelownitrogenconcentrationsobservedindiamonds159

from the deeper parts of themantle (seeHarte (2010) and references therein). Despite the question of nitrogen160

compatibilityindiamondbeingcontested,weproposeittobeamootpointforthefollowingreasons.Accordingto161

Goldschmidt’s (1937) rules for element compatibility, nitrogen is of the similar atomic size and charge to carbon,162

making N compatible in the diamond lattice, as is borne out by nitrogen being themost common substitutional163

7

impurityindiamond.However,nitrogenisincorporatedintodiamondinamonatomicstate,eventhoughitdoesnot164

occur inthemantle inthishighlyunstable form(MikhailandSverjensky,2014;LiandKeppler,2014).Nitrogenis165

stable under equilibrium conditions as a variety of molecules, where each nitrogen complex exhibits radically166

different chemical affinities. Therefore, the real focus of investigation should be into understanding how these167

various nitrogen species behave, and which potential chemical reactions can disassociate these compounds to168

produce(unstable)monatomicNduringdiamondformation.169

Historically,nitrogenhasbeenconsideredanatmophileelementandaccordinglygroupedwiththenoblegases.170

Atmophileelementsaredefinedas‘thoseelementsthatremainmostlyonorabovethesurfacebecausetheyare,or171

occurin, liquidsand/orgasesattemperaturesandpressuresfoundonthesurface’(Goldschmidt,1937).However,172

underconditionsofhighPandT,nitrogencanbehavelikeanoblegas(N2),asiderophileelement(Fe3N,TiN,BN),an173

alkalimetal(NH4+)oranorganicreactant(e.g.nitrosyl;NO-).Infact,insamplesfromthelowercrustandtheupper174

mantle,nitrogenhasbeenfoundasmolecularN2(Andersenetal.,1995;Smithetal.,2014),ammonium(impliedby175

Yokochietal.,2009),metallicnitride(Fe3N,TiN,BN;Dobrzhinetskayaetal.2009),animpurityincarbide(Kaminsky176

&Wirth,2011)andasalattice-boundcomponentwithindiamond(NC4;Kaiser&Bond,1959).Recentexperimental177

andtheoreticaldatahaveshownthatpressure,temperature,redoxstate,pH,andthemolarabundanceofnitrogen178

canhavesignificanteffectsonthespeciationofnitrogen,withredoxbeingthemostimportant(LiandKeppler,2014;179

Mysenetal.,2014;MikhailandSverjensky,2014;Roskoszetal.,2006).Forconditionsrelevanttomost(bysample180

mass)diamondformation(1000-1300°Cand4-7GPaacrossaLOGfO2rangeofQFM+2to-4),nitrogenwillbestable181

indiamond-formingfluidsasammonic(NH4+/NH30)ormolecular(N2)forms(Fig.2&3).182

None of the nitrogenmolecules listed above are compatible in diamond, because they are far too large to fit183

withinthediamondlattice(orneutrallychargedandthereforeinert,as isthecaseforN2andNH3).Therefore,the184

only state inwhichnitrogen canbepartitioned intodiamond is asmonatomicN-3,which asnoted above is not a185

stableformofnitrogen.ThisrequiresthatnitrogenisincorporatedintodiamondwhenmonatomicNisproducedby186

coupledoxidation/reductionandacidity/basicity reactions involvingN2orNH4+ andCO2orCH4.Three simplified187

idealreactionsareshowninequations3-5,whereCO2isinterchangeablewithCO32-(butwouldrequireadifferent188

mass balance). Noteworthy, the "NC4" molecule is not an independent species, but instead represent diamond189

containingNasapointdefectinthecrystalstructure:190

Oxidation/reductionreactions:191

8

Eq.3 4CH4+NH4++5O2=(NC4)++10H2O192

Eq.4 8CO2+N2+2H2=2(NC4)++2H2O+7O2193

Acidity/basicityreaction:194

Eq.5 4CO2+NH4+=NC4+2H2O+3O2195

For the above reactions to form diamond requires that the nitrogen-bearing species in the diamond-forming196

mediumtobecomeunstableatuppermantleconditions.Thismeanstheywillreactduringdiamond-formation,thus197

enablingmonatomicnitrogentopartitionintodiamondfollowingoxidationofNH4+(Eq3),thereductionofN2(Eq.4),198

ordehydrogenationofNH4+(Eq.5).Equilibriumconstantsforthestabilityofammonium/molecularnitrogenunder199

conditionsconducive todiamond formationshowthat the transitionbetweenammonicandmolecularnitrogenat200

1200°CoccurswhenLOGfO2isbetweenΔQFM=0and-2(LiandKeppler,2014).Theseconditionsareverysimilarto201

thoseunderwhichcarbonate/methanetransitionstodiamond(Stagnoetal.,2010;Sverjenskyetal.,2014;Frostand202

McCammon, 2008) (Fig.2). In short, diamond formation can occur under conditions where ammonium is203

thermodynamicallystable,whichwouldinhibitnitrogenpartitioningintodiamond.Ourmodelimpliesthatnitrogen204

uptakeintodiamondisnotakineticallydrivenprocessbaseduponNbeingcompatibleinthediamondlattice,but205

insteadrequirescoupledoxidation/reduction(Eq.3-4)oracid/base(Eq.5)reactionsto liberatemonatomicNfrom206

ammoniumormolecularnitrogenduringdiamondformation. Inthismodel, the formationofTypeIIdiamondcan207

occur under a variety of conditions, and requires no single mechanism/environment for formation. From a208

petrological standpoint it is not necessary for Type II diamonds to precipitate inN-free domains or fromN-poor209

fluids;itisonlyrequiredthatTypeIIdiamondprecipitateintheabsenceofmonatomicnitrogen.Diamondgrowing210

inequilibriumwithFe3N,TiN,NH4,NH3,orN2wouldbeTypeII(becausethesenitrogenmoleculesareincompatible211

inthediamondlattice).212

Diamonds thatare thought togrow in the lowerpartsof theuppermantle, the transitionzone,and the lower213

mantle (based on theirmineral inclusion assemblages) are commonly found to beType II (seeHarte (2010) and214

references therein). One possible explanation could be that nitrogen becomesmore compatible in othermineral215

phasesthatarepresentatthesereduceddepthsinthemantle(suchasFe-alloys,andK-bearingmineralssuchasK-216

hollandite). In a sense, this is a similar process to that used in industrial HPHT diamond synthesis of Type II217

diamonds; the reduction of nitrogen in the presence of native metals such as Fe, Cr, or Ti (nitrogen getting /218

nitriding;StachelandHarris,2009).TentativeevidenceforthisprocessinnaturecomesfromanunusualsuiteofFe-219

9

carbide inclusionswithin a diamond from theAmazon craton (Juina, Brazil; Kaminsky andWirth, 2011). The Fe-220

carbidehasahighNcontent(termednitro-carbide;73000–91000ppmN)whilethehostdiamondcontainsonly221

44ppmnitrogen. Thiswould give aKNdiamond-carbide of only 0.0005.However,Mikhail et al., (2014c) also described222

several Fe-carbide inclusions from two Southern African diamonds (from Jagersfontein, Kaapvaal craton, South223

Africa) in which neither the diamond nor the syngenetic carbide inclusions contained detectable nitrogen.224

Interestingly,arecentstudybyTsunoandDasgupta(2015),demonstratingtheeffectofSonthestabilityofdiamond225

in equilibriumwith Fe-Ni alloys, undermines the argument for co-existing nitride formation. They found that the226

presenceofSresultsintheassemblagediamond+sulfide,whereasintheabsenceofStheassemblageisFe-carbide227

(cementite)+Fe-Nialloy(Lordetal.,2009).Assulfidesare themostcommon inclusion indiamond(Shireyetal.,228

2013),itislikelythattheformationofsulfideprecludestheformationofnitrides,asitdoesforcarbides.Although229

thisaspectisthesubjectofdebateandongoingstudy,itisimportanttonotethattheexistenceoflithosphericTypeII230

diamonds(i.e.thoseformedatmoretypicaldepthsof120-200km)andthelackofmetallicinclusionsinalmostall231

sub-lithosphericdiamondsdemonstratethatthepresenceofnativemetalsandexceedinglyreducedconditionsare232

not required to explain the formation of Type II diamonds in the sub-lithospheric mantle. In addition, there is233

evidence that sub-lithospheric diamond-formation occurs in equilibrium with relatively oxidizing subducted234

carbonatemelts in the sublithosphericmantle, ergo, inhibiting carbideornitride stabilityduring ‘deep’ diamond-235

formation(Walteretal.,2008,2011;Thompsonetal.,2014).236

237

4.Implicationsfordiamond-formation238

Despite the similarity in charge and ionic radius between N and C, data from extensive studies of natural239

diamondsandHPHTexperimentsprovideconflictingevidenceregarding thecompatibilityofnitrogen indiamond240

(e.g.Deinesetal.,1989;Cartignyetal.,2001;Thomassotetal.,2007;StachelandHarris,2009;Smartetal.,2011;241

Smith and Kopylova, 2014). It is clear that when the matter is considered from a petrological standpoint, the242

behavior of nitrogen during diamond formation is complex. For nitrogen to be in an atomic form, capable of243

incorporation into thediamond latticeduringgrowth, requirescoupledoxidation/reduction(Eq.3-4)oracid/base244

(Eq.5) reactions during diamond formation to liberate N in the monatomic state. This fundamental concept245

underminesourentirethinkingaboutdiamondasarecorderofmantlenitrogen.Forexample,itispossibletoargue246

thataTypeIIdiamondcouldhavegrownfromafluidwithahighnitrogenconcentration,whereconverselyaTypeI247

10

diamondcanprecipitate froma fluidwithamuch lowerNcontent -acounter-intuitivenotion.Ergo, thenitrogen248

abundanceofmantlediamonds ispotentially controlledby the fO2 conditionsduringdiamond formation, andnot249

justthenitrogenconcentrationofthediamond-formingfluid.Byinference,usingtheglobaldatasets(Fig.1a-c)we250

predict that the formationof the fibrousovergrowthsofcoateddiamondsmayrequirea largeredoxgradientthat251

transitions across the stability of ammonium and molecular nitrogen (which is dependent upon P, T, and X),252

generating abundant monatomic nitrogen that is readily incorporated into the fibrous diamond. Conversely, the253

formation of non-fibrous monocrystalline diamond occurs across a large range of redox states, that do not254

necessarilycrosstheboundarybetweennitrogenspecies,meaningmanysamplesmaygrowunderconditionswhere255

ammoniumormolecularnitrogenarecompletelystablespeciesduringdiamondformation(andaredissolvedinthe256

fluid).Thismodelexplainswhythedistributionofnitrogenconcentrationsinnon-fibrousmonocrystallinediamonds257

isskewedtowardszero(akintoanincompatibleelement;Fig.1a-b)andthedistributionforthefibrousovergrowths258

ofcoateddiamondsisGaussian(akintoacompatibleelement;Fig.1c).259

It is importanttoreiteratethepointthatbroadstatementsregardingthecompatibilityofnitrogenindiamond260

are too simplistic.Atomicnitrogen is compatible,whereasmolecularnitrogen, ammonium/ammonia andmetallic261

nitridesarenot.Therefore,weneedtounderstandtheeffectsofthemantleP-T-X-fO2conditionsonthespeciationof262

nitrogenovertherangerelevanttodiamondformation.Thediversityofstablenitrogenmoleculesunderthisrange263

ofmantleconditionscanexhibitradicallydifferentsolubility,stability,andpartitioningbehaviorinfluidsandmelts,264

meaningageneralviewonnitrogencompatibilityindiamondisirrelevantanditneedstobeconsideredonacase-265

by-casebasis.Therefore,amuchmorepertinentquestionis‘whatisthebehaviorofnitrogeninthemantle,andhow266

isdiamondformationrecordingit?’267

4.1NinHPHTSyntheticDiamonds268

This review of N behavior during natural diamond growth raises obvious questions when it comes to synthetic269

diamondgrowth.WhendiamondsaresynthesizedunderHPHTconditionsfrommetalsolvents(e.g.FeNialloy),they270

commonly contain substitutionalN unless special steps are taken (i.e. the addition of N getters such as Ti0). The271

sourceof theN is assumed tobe atmospheric contamination (i.e.N2;Boyd et al., 1988). Ifwe consider the redox272

reactionsabove (eq.3-5)whichaffectNspeciation, then it is likely that theN2willbe subjected tovery reducing273

conditionswithintheHPHTcapsule(inequilibriumwithC0andaFe-Nialloy).ThismeansthatN2willbeconverted274

11

to iron nitride or NH4, during which time monatomic N will be generated, which can be incorporated into the275

diamondlattice.276

5.Broaderimplicationsandfuturedirections277

5.1Stableisotopefractionation278

The use of stable isotopes to trace subducted material through the mantle is widespread, and based on the279

thermodynamicprinciple thatequilibriumstable isotope fractionation is largeat low temperaturesanddecreases280

greatlyasafunctionofT(Urey,1947).Therefore,thelightelementstableisotoperatios(e.g.carbonandnitrogen)281

are fractionated inEarth's surficial reservoirs,primarily theatmosphereandhydrosphere,bymuch larger factors282

than are possible in the mantle. This makes the stable isotopes of light atmophile elements powerful tracers of283

materialsubductedthroughthemantle(Hiltonetal.,2002).However,becausecarbonisotopescanbesignificantly284

fractionated in themantle under open-systemRayleigh conditions (Cartigny et al., 2001), or reducing conditions285

involvingFe-carbides(Mikhailetal.,2014c),theuseofdiamonds’carbonisotopesalonecannotconclusivelyascribe286

amantleorigintothediamond-formingcarbon(Cartignyetal.,1997).Asnitrogenispresentinmostdiamonds,the287

15N/14N ratios indiamondhavebeen coupledwith the 13C/12C ratios to verify or challenge the conclusionsbased288

solelyoncarbon-isotopedata(Javoyetal.,1984;Boydetal.,1987,1992;BoydandPillinger,1994;Cartignyetal.,289

1997,1998a,b,2001,2003,2004,2009;Harteetal.,1999;Bulanovaetal.,2002,2014;Haurietal.,2002;Howellet290

al.,2015a;Gautheronetal.,2005;Hutchisonetal.,1997;Thomassotetal.,2007,2009;Pettsetal.,2015;Palotetal.,291

2009, 2012; 2014; Klein-BenDavid et al., 2010; Mikhail et al., 2013, 2014a). However, it is unclear whether the292

15N/14Nratioindiamondrecordsthe15N/14Nratioofthediamond-formingfluids.Themodelpresentedhereimplies293

thatpartitioningofnitrogen intodiamondcanoccurvia threemainreactions, involvingeitherN2orNH4(Eq.3-5).294

Duetosignificantdifferencesinthevibrationalfrequenciesfornitrogenbondsinreducedandoxidisedspecies(e.g.295

C-N, bonds,N-Hbonds, andN-Nbonds;Richet et al., 1977), deviations from the typical 15N/14N ratios inmantle-296

derivedrocksandwithinsingleminerals(e.g.diamond)maybeaconsequenceofredoxandpHchangesinupper-297

mantle fluids(Pettsetal.,2015). Insteadofrequiringspecialmechanismssuchas tectonic injectionof isotopically298

heterogeneouscrustalsources(BoydandPillinger,1994;Mikhailetal.,2014a)orprimordialheterogeneities(Javoy299

etal.,1986;Cartignyetal.,1997;Palotetal.,2012),theobservedvariationsinthenitrogenisotopiccompositionsof300

diamondsandothermantlematerialsmightbetheresultofintra-mantleequilibrium(Pettsetal.,2015)orkinetic301

(Yokochietal.,2009;Lietal.,2009)stable-isotopefractionationassociatedwithchangesinfluidchemistryduring302

12

metasomatic reactions in the mantle. Therefore, qualitative understandings of equilibrium stable-isotope303

fractionationfactorsfornitrogenisotopesduringthereactionsshowninEq.3-5arerequired.304

5.2Themantlenitrogenbudget305

Mantlediamondsshowadecreaseinaveragenitrogenabundancewithincreasingdepthofformation.Thiscould306

reflectadecreaseinthenitrogencontentofthesilicateEarthwithdepth,aswouldbepredictedifnitrogenwereto307

behave like a noble gas, or if nitrogen ismore compatible in other phaseswith increasingdepth. Several lines of308

directandindirectevidenceshowthatnitrogendoesnotnecessarilybehavelikeanoblegas,andthatthemantleisa309

significantreservoirfornitrogen:[1]Nitrogenshowsthelargestdepletion(relativetochondrites)inthebulksilicate310

Earth+atmospherecomparedwiththeenrichmentlevelsofothervolatileelements,includingthenoblegases(H,C,311

N,Ne, Ar, Kr, and Xe;Marty, 2012;Halliday, 2013), [2] The calculated flux for nitrogen between the surface and312

interior (subductionvs. volcanism) implies thatmorenitrogen isbeingout-gassed from the interior than isbeing313

returned into the mantle (Busigny et al., 2011), and [3] the solubility of nitrogen in enstatite and forsterite314

demonstrates that the reduced parts of the upper mantle can store >50 times more nitrogen than the surface315

reservoirs(includingtheatmosphere;LiandKeppler,2013;Watenphuletal.,2010).Collectively,thesedatastrongly316

imply the existence of a deep reservoir. Paradoxically, diamond is the mantle mineral with the highest average317

nitrogen abundance, butmay not be the largest reservoir ofmantle nitrogen. By far themost important storage318

mechanism for nitrogen in the mantle is probably through the exchange equilibria between positively charged319

ammoniumcations(NH4+)andpositivelychargedalkalimetals(e.g.Rb+&K+).ThisimpliesthatthemantleN-Hcycle320

couldfollowthesamepathwaysasalkalimetals.Thus,basedonthestabilityoftheknownK-bearingphases(Harlow321

andDavies,2004)itislikelythatammonium-bearingmantlephasesarestablethroughouttheentiresilicateportion322

of theEarth.Thenatureof thedeepnitrogen reservoir thereforedependsupon the stabilityof ammonium in the323

mantle.Themolefractionofammonic/totalnitrogenispredictedtodecreasewithincreasingtemperature(e.g.Fig.3324

a&b).However, therethemolefractionofammonic/totalnitrogenshouldincreasewithincreasingpressure(e.g.325

Fig.3b&d),albeitthisrelationshipisknownwithalimitationofonly4GPaforΔPandmaximumTofonly1400°C326

(Mikhail&Sverjensky,2014;Li&Keppler,2014).Whatisnowrequiredaredataonthesolubilityandpartitioningof327

ammonicnitrogenbetweenfluids,melts,andhigh-pressureK-bearingphases, toaddressthedepletionofnitrogen328

relativetotheothervolatileelements,andbyinferencetoexpress(mechanically)whyaveragenitrogenabundance329

in diamonds decreases with increasing depth of formation. Addressing this question will provide amore robust330

13

understanding of the fluxing of nitrogen during subduction, and the N-H storage capacity of planetary silicate331

mantlesasawhole.332

6Summary333

Nitrogenisinitiallyincorporatedintothediamondlatticeassinglenitrogenatoms,aformofnitrogenwhichis334

highlyunstable.However,thestableformsofnitrogeninthemantlearedominatedbyN2,andNH4+,bothofwhich335

are incompatible in thediamond lattice.Areviewof theavailabledatashowsthat inclassic terms,nitrogen is the336

mostcommonsubstitutionalimpurityfoundinnaturaldiamondsbecauseitisofverysimilaratomicsizeandcharge337

to carbon. However, the speciation of nitrogen, and how these different species disassociate during diamond338

formationtocreatetransientmonatomicnitrogen,arethefactorsgoverningnitrogenabundanceindiamonds.Our339

modelsuggestsacounter-intuitivenotion;nitrogen-freediamondcouldgrowfromaN-richmediathatissimplynot340

undergoing reactions to liberatemonatomicN fromN2, orNH4+ during coupledoxidation/reductionor acid/base341

reactionsduringdiamond-formation(Eqs.3-5).Incontrast,anitrogen-bearingdiamondcouldgrowfromafluidwith342

alowerNabundance,inwhichreactionsareoccurringtogenerate(unstable)Natomsduringdiamondformationare343

occurring.344

Themostlogicalconclusionwecanderiveisthat,atpresent,wedonotunderstandhowdiamondsarerecording345

themantlenitrogencycle.However,webelievediamondsstill representourbest tool for investigatingnumerous346

aspects of the mantle C-N cycle, especially when considering carbon and nitrogen stable isotope data. However,347

furtherinvestigationsintothespeciation,partitioning,andequilibriumstable-isotopefractionationofnitrogenwill348

enable a more accurate interpretation of the nitrogen data from mantle diamonds, and lead to a better349

understandingoftheEarthdeepvolatilecyclesandfluxes.350

Acknowledgements351

We are grateful to W.L Griffin for providing valuable scientific and editorial comments on an earlier version of this352

manuscript.ReviewersD.PettsandC.M.Breeding,andeditorD.Hummerarethankedforusefulcommentsthatfurther353

improved thismanuscript.We are also grateful tomembers of the Bristol University petrology group (M.JWalter, S.C354

Kohn,R.ABrooker,O.TLord,SMcMahon,LSpeich,LZiberna,C.BSmith,andG.PBulanova)fordiscussingthistopicaspart355

of the internal diamond group meetings. During part of this study, SM was employed at Bristol on NERC grant356

NE/J008583/1 to Michael J Walter. DH is funded by a DAAD PRIME fellowship. SM is grateful to the Deep Carbon357

14

Observatoryforfundingthefortnightly‘DeepCarbon’meetingattheCarnegieInstitutionofWashington(USA),wherethe358

ideaforthisstudyfirstoriginated.359

360

Referencescited361

Ahrens, L.H. (1954) The log-normal distribution of the elements (a fundamental law of geochemistry and its362

subsidiary).GeochimicaetCosmochimicaActa,5,49–73.363

Andersen,T.,Burke,E.A. J.,andNeumann,E.R. (1995)Nitrogen-rich fluid intheuppermantle: fluid inclusions in364

spinelduniteformLanzarote,CanaryIslands.ContributionstoMineralogyandPetrology,120,20-28.365

Blundy, J., andWood, B.J. (2003) Partitioning of trace elements between crystals andmelts. Earth and Planetary366

ScienceLetters,210,383-397.367

Boyd,S.R.,Pillinger,C.T.,Milledge,H.J.,Mendelssohn,M.,andSeal,M.(1988)Fractionationofnitrogenisotopesina368

syntheticdiamondofmixedcrystalhabit.Nature331,604-607.369

Boyd, S.R., Pillinger, C.T., Milledge, H.J., and Seal, M.J. (1992) C and N isotopic composition and the infrared370

absorptionspectraofcoateddiamonds:evidence for theregionaluniformityofCO2-H2Orich fluids in lithospheric371

mantle.EarthandPlanetaryScienceLetters,108,139-150.372

Boyd, S.R.,Mattey,D.P., Pillinger, C.T.,Milledge,H.J.,Mendelssohn,M., and Seal,M. (1987)Multiple growth events373

during diamond genesis: an integrated study of carbon and nitrogen isotopes and nitrogen aggregation state in374

coatedstones.EarthandPlanetaryScienceLetters,86,341-353.375

Boyd,S.R.,andPillinger,C.T.(1994)Apreliminarystudyof15N/14Ninoctahedralgrowthformdiamonds.Chemical376

Geology,116,43-59.377

Bulanova,G.P.,Pearson,D.G.,Hauri,E.H., andGriffin,B.J. (2002)Carbonandnitrogen isotopesystematicswithina378

sector-growthdiamondfromtheMirkimberlite,Yakutia.ChemicalGeology,188,105-123.379

BulanovaG.P.,WiggersdeVriesD.,PearsonD.G.,BeardA.,MikhailS.,SmelovA.P.andDaviesG.R.(2014)Carbon380

recycling and polygenetic origin of eclogitic diamonds recorded by a single crystal from the Mir pipe (Yakutia).381

ChemicalGeology,381,40–55.382

Busigny, V., Cartigny, P., and Philippot, P. (2011) Nitrogen isotopes in ophioliticmetagabbros: A re-evaluation of383

15

modern nitrogen fluxes in subduction zones and implication for the early Earth atmosphere. Geochimica384

CosmochimicaetActa,75,7502-7521.385

Busigny,V,.andBebout,G.E.(2013)NitrogeninthesilicateEarth:Speciationandisotopicbehaviourduringmineral-386

fluidinteractions.Elements,9,353-358.387

Canfield,D.E., Glazer,A.N., andFalkowski, P.G. (2010)The evolution and futureofEarth's nitrogen cycle. Science,388

330,192-196.389

Cartigny, P., Boyd, S.R., Harris, J.W., and Javoy,M. (1997) Nitrogen isotopes in peridotitic diamonds from Fuxian,390

China:themantlesignature.TerraNova,9,175-179.391

Cartigny,P.,Harris, J.W.,and Javoy,M. (1998a)EclogiticDiamondFormationat Jwaneng:NoRoom foraRecycled392

Component.Science,280,1421-1424.393

Cartigny,P.,Harris, J.W.,Phillips,D.,Girard,M.,andJavoy,M.(1998b)Subduction-relateddiamonds?Theevidence394

foramantle-derivedoriginfromcoupleddeltaδ13C-δ15Ndeterminations.ChemicalGeology,147,147-159.395

Cartigny,P.,Harris,J.W.,andJavoy,M.(1999)Eclogitic,PeridotiticandMetamorphicdiamondsandtheproblemsof396

carbonrecycling—thecaseofOrapa(Botswana).7thInternationalKimberliteConferenceExtendedAbstract,117–397

124.398

Cartigny,P.,Harris,J.W.,andJavoy,M.(2001)Diamondgenesis,mantlefractionationsandmantlenitrogencontent:a399

studyofδ13C-Nconcentrationsindiamonds.EarthandPlanetaryScienceLetters,185,85-98.400

Cartigny,P.,Harris, J.W.,Taylor,A.,Davies,R., and Javoy,M. (2003)On thepossibilityof akinetic fractionationof401

nitrogenstableisotopesduringnaturaldiamondgrowth.GeochimicaetCosmochimicaActa,67,1571-1576.402

Cartigny,P.,Stachel,T.,Harris,J.W.,andJavoy,M.(2004)ConstrainingdiamondmetasomaticgrowthusingC-andN-403

stableisotopes:examplesfromNamibia.Lithos,77,359-373.404

Cartigny,P.(2005)Stableisotopesandtheoriginofdiamond.Elements,1,79-84.405

Cartigny,P.,Farquhar,J.,Thomassot,E.,Harris,J.W.,Wing,B.,Masterson,A.,McKeegan,K.,andStachel,T.(2009)A406

mantleorigin forPaleoarcheanperidotiticdiamonds from thePandakimberlite, SlaveCraton:Evidence from13C-,407

15N-and33,34S-stableisotopesystematics.Lithos,112,852-864.408

16

Cartigny, P. (2010)Mantle-related carbonados? Geochemical insights from diamonds from the Dachine komatiite409

(FrenchGuiana).EarthandPlanetaryScienceLetters,296,3-11.410

Cartigny, P., Palot, M., Thomassot, E., and Harris, J.W. (2014) Diamond formation: a stable isotope perspective.411

AnnualReviewsinEarthandPlanetarySciences,42,699–732.412

DeinesP.,HarrisJ.W.,SpearP.M.andGurneyJ.J.(1989)Nitrogenand13CcontentofFinschandPremierdiamonds413

andtheirimplications.GeochimicaetCosmochimicaActa,53,1367–1378.414

Dobrzhinetskaya,L.F.,Wirth,R.,Yang, J.,Hutcheon, I.D.,Weber,P.K., andGreen,H.W. (2009)High-pressurehighly415

reducednitridesandoxidesfromchromititeofaTibetanophiolite.ProceedingsoftheNationalAcademyofSciences,416

106,19233-19238.417

Frost,D.J.,Liebske,C.,Langenhorst,F.,McCammon,C.A.,Tronnes,R.G.,andRubie,D.C.(2004)Experimentalevidence418

fortheexistenceofiron-richmetalintheEarth'slowermantle.Nature,428,409-412.419

Frost,D.J,andMcCammon,C.M. (2008)TheRedoxStateofEarth'sMantle.AnnualReviews inEarthandPlanetary420

Sciences,36,389-420.421

Gautheron,C.,Cartigny,P.,Moreira,M.,Harris,J.W.,andAllègre,C.J.(2005)Evidenceforamantlecomponentshown422

by rare gases, C andN isotopes inpolycrystallinediamonds fromOrapa (Botswana). Earth andPlanetary Science423

Letters,240,559-572.424

Goldblatt, C., Claire, M.W., Lenton, T.M., Matthews, A.J., Watson, A.J., and Zahnle, K.J. (2009) Nitrogen-enhanced425

greenhousewarmingonearlyEarth.NatureGeoscience,2,891–896.426

Goldschmidt,V.S. (1937)Theprinciplesofdistributionof chemical elements inminerals and rocks. Journalof the427

ChemicalSociety,140,655–673.428

Gurney, J.J.,Helmstaedt,H.H.,Richardson,S.H.,andShirey,S.B.(2010)DiamondsthroughTime.EconomicGeology,429

105,689-712.430

Halliday,A.N.(2013) Theoriginsofvolatilesintheterrestrialplanets.GeochimicaCosmochimicaetActa,105,146-431

171.432

HarlowG.E.,andDavis,R.(2004)StatusreportonstabilityofK-richphasesatmantleconditions.Lithos,77,647-653.433

17

Harte,B. (2010)Diamondformation in thedeepmantle: therecordofmineral inclusionsandtheirdistribution in434

relationtomantledehydration.MineralogicalMagazine,74(2),189-215.435

Harte,B.,Fitzsimons, I.C.W.,Harris, J.W.,andOtter,M.L. (1999)Carbon isotoperatiosandnitrogenabundances in436

relation to cathodoluminescence characteristics for some diamonds from the Kaapvaal Province, S. Africa.437

MineralogicalMagazine,63,829-829.438

Hauri, E.H., Wang, J., Pearson, D.G., and Bulanova, G.P. (2002) Microanalysis of δ13C, δ15N, and N abundances in439

diamondsbysecondaryionmassspectrometry.ChemicalGeology,185,149-163.440

Hilton,D,R., Fischer,T.P., andMarty,B. (2002)Noblegasesandvolatile recyclingat subductionzones. InPorcelli,441

D.P.,Ballentine,C.J.,andWieler,R.(Eds.).NobleGasinGeochemistryandCosmochemistry,MineralogicalSocietyof442

America,WashingtonDC,319–370.443

Howell,D.,O’Neill,C.J.,Grant,K.J.,Griffin,W.L.,Pearson,N.J.,andO’Reilly,S.Y.(2012)μ-FTIRmapping:distributionof444

impuritiesindifferenttypesofdiamondgrowth.DiamondandRelatedMaterials,29,29–36.445

Howell,D.,Griffin,W.L.,Piazolo,S.,Say,J.,Stern,R.A.,Stachel,T.,Nasdala,L.,Rabeau,J.,PearsonN.J.,andO’Reilly,S.Y.446

(2013) A spectroscopic and carbon-isotope study of mixed-habit diamonds: Impurity characteristics and growth447

environment.AmericanMineralogist,98,66-77.448

Howell,D.,Stern,R.A.,Griffin,W.L.,Southworth,R.,Mikhail,S.,andStachel,T.(2015a)Nitrogenisotopesystematics449

andoriginsofmixed-habitdiamonds.GeochimicaCosmochimicaetActa,157,1-12.450

Howell,D.,Griffin,W.L.,Yang,J.,Gain,S.,Stern,R.A.,Huang,J.,Jacob,D.E.,Xu,X.,Stokes,A.J.,O’Reilly,S.,Pearson,N.,451

(2015b)Diamondsinophiolites:contaminationoranewdiamondgrowthenvironment.EarthandPlanetaryScience452

Letters,430,284-295.453

Hutchison,M.T.,Cartigny,P.,Harris, J.W.(1997)CarbonandNitrogenCompositionsandPhysicalCharacteristicsof454

Transition Zone and LowerMantle Diamonds from Sao Luiz, Brazil, in: Gurney, J.J., Gurney, J.L., Pascoe, S.H., and455

Richardson, S.H. (Eds.), Proceedings of the VII International Kimberlite Conference. RedRoofDesign, Cape Town,456

372-382.457

Javoy,M.,Pineau,F.,andDemaiffe,D., (1984)Nitrogenandcarbon isotopiccomposition in thediamondsofMbuji458

Mayi(Zaire).EarthandPlanetaryScienceLetters,68,399-412.459

18

Kaiser,W.,andBond,W.L.(1959)Nitrogen,AMajorImpurityinCommonTypeIDiamond.PhysicalReviewLetters,460

115,857-863.461

Kaminsky,F.V.,andWirth,R.(2011)Ironcarbideinclusionsin lower-mantlediamondfromJuina,Brazil.Canadian462

Mineralogist,49,555-572.463

Klein-BenDavid,O.,Pearson,D.G.,Nowell,G.M.,Ottley,C.,McNeill,J.C.R.,andCartigny,P.(2010)Mixedfluidsources464

involved in diamond growth constrained by Sr–Nd–Pb– C–N isotopes and trace elements. Earth and Planetary465

ScienceLetters,289,123–133.466

LiY.andKepplerH.(2014)Nitrogenspeciationinmantleandcrustalfluids.GeochimicaCosmochimicaetActa,129,467

13–32.468

Li,L.,Cartigny,P.,andAder,M.(2009)Kineticnitrogenisotopefractionationassociatedwiththermaldecomposition469

of NH3: Experimental results and potential implications to natural-gas and hydrothermal systems. Geochimica470

CosmochimicaetActa,73,6282-6297.471

Lord,O.T.,Walter,M.J.,Dasgupta,R.,Walker,D.,andClark,S.M.(2009)MeltingintheFe-Csystemto70GPa,Earth472

andPlanetaryScienceLetters,284,157–167.473

Marty, B. (2012) The origins and concentrations of water, carbon, nitrogen and noble gases on Earth. Earth and474

PlanetaryScienceLetters,313,56–66475

Mikhail,S.andSverjensky,D.A.(2014)NitrogenspeciationinuppermantlefluidsandtheoriginofEarth’snitrogen-476

richatmosphere.NatureGeoscience,7,816-819.477

MikhailS.,VerchovskyA.B.,HowellD.,HutchisonM.T.,SouthworthR.,ThomsonA.R.,WarburtonP., JonesA.P.and478

Milledge H.J. (2014a) Constraining the internal variability of the stable isotopes of carbon and nitrogen within479

mantlediamonds.ChemicalGeology,366,14–23.480

Mikhail,S.,McCubbin,M.F.,andHowell,D.(2014b)Evidenceformultiplediamondite-formingeventsinthemantle.481

AmericanMineralogist,99,1537-1543.482

Mikhail,S.,Guillermier,C.,Franchi,I.A.,Beard,A.D.,Verchovsky,A.B.,Wood,I.,Jones,A.P.,andMilledge,H.J.(2014c)483

Empirical evidence for the fractionation of carbon isotopes between diamond and iron carbide from the Earth’s484

mantle.GeochemistryGeophysicsGeosystems,DOI:10.1002/2013GC005138.485

19

Mikhail, S., Dobosi, G., Verchovsky, A.B., Kurat, G., and Jones, A.P. (2013) Peridotitic andwebsteritic diamondites486

providenew information regardingmantlemelting andmetasomatism induced through the subductionof crustal487

volatiles.GeochimicaetCosmochimicaActa,107,1-11.488

Milledge,H.J.,andMeyer,M.O.A.(1962)Nitrogen-14inNaturalDiamonds.Nature,195,171-172.489

Mysen,B.O.,Tomita,T.,Ohtani,E.,Suzuki,A.(2014)SpeciationofandD/Hpartitioningbetweenfluidsandmeltsin490

silicate-D-O-H-C-N systems determined in-situ at upper mantle temperatures, pressures, and redox conditions.491

AmericanMineralogist,99,578-588.492

Palot,M.,Cartigny,P.,andViljoen,F.(2009)Diamondoriginandgenesis:ACandNstableisotopestudyondiamonds493

fromasingleeclogiticxenolith(Kaalvallei,SouthAfrica).Lithos,112,758–766.494

Palot M., Cartigny P., Harris J.W., Kaminsky F.V., and Stachel T. 2012. Evidence for deep mantle convection and495

primordialheterogeneityfromnitrogenandcarbonstableisotopesindiamond.EarthandPlanetaryScienceLetters,496

357,179–193.497

Palot, M., Pearson, D.G., Stern, R.A., Stachel, T., and Harris, JW. (2014) Isotopic constraints on the nature and498

circulation of deep mantle C–H–O–N fluids: Carbon and nitrogen systematics within ultra-deep diamonds from499

Kankan(Guinea).GeochimicaetCosmochimicaActa,139,26-46.500

Pal’yanov Y.N, Sokol A.G, Borzdov Y.M, Khokhryakov A.F and Sobolev N.V. (2002) Diamond formation through501

carbonate–silicateinteraction.AmericanMineralogist,871009–1013502

Pal’yanovY.N.,BatalevaY.V.,SokolA.G.,BorzdovY.M.,KupriyanovI.N.,ReutskyV.N.,andSobolevN.V.(2013)Mantle-503

slabinteractionandredoxmechanismofdiamondformation.ProceedingsoftheNationalAcademyofSciences,110,504

20408-20413.505

Petts,D.C.,Chacko,T.,Stachel,T.,Stern,R.A.,Heaman,L.M.(2015)Anitrogen isotope fractionation factorbetween506

diamondanditsparental fluidderivedfromdetailedSIMSanalysisofagemdiamondandtheoreticalcalculations.507

ChemicalGeology,410,188-200508

Porcelli,D.,andPepin,R.O.(2003)TheOriginofNobleGasesandMajorVolatilesintheTerrestrialPlanets.Treatise509

onGeochemistry,319-347.510

20

Reutsky,V.N.,Harte,B.,EIMF,Borzdov,Y.M.,Palyanov,Y.N.(2008)Monitoringdiamondcrystalgrowth,acombined511

experimentalandSIMSstudy.EuropeanJournalofMineralogy,20,365-374.512

Richet,P.,Bottinga,Y., and Javoy,M. (1977)Areviewofhydrogen, carbon,nitrogen,oxygen, sulphurandchlorine513

stableisotopefractionationamonggaseousmolecules.AnnualReviewsinEarthandPlanetarySciences,5,65–110.514

Rohrbach,A.,Ballhaus,C.,Golla-Schindler,U.,Ulmer,P.,Kamenetsky,V.S.,andKuzmin,D.V.(2007)Metal-saturation515

intheuppermantle.Nature,449,456-458.516

Sagan, C., and Chyba, C. (1997) The early Sun paradox: organic shielding of ultraviolet-labile greenhouse gases.517

Science,276,1217–1221.518

ShireyS.B.,CartignyP.,FrostD.J.,KeshavS.,NestolaF.,NimisP.,PearsonD.G., SobolevN.V., andWalterM.J.2013.519

Diamondsandthegeologyofmantlecarbon.ReviewsinMineralogyandGeochemistry,75,355-421.520

Smart K.A., Chacko T., Stachel T., Muehlenbachs K., Stern R.A., and Heaman L.M. (2011) Diamond growth from521

oxidized carbon sources beneath theNorthern Slave Craton, Canada: a δ13C–N study of eclogite-hosted diamonds522

fromtheJerichokimberlite.GeochimicaCosmochimicaetActa,75,6027–6047.523

Smith,E.M.,andKopylova,M.A.(2014)Implicationsofmetallicironfordiamondsandnitrogeninthesublithospheric524

mantle.CanadianJournalofEarthScience,51,510–516.525

Smith,E.M.,Kopylova,M.G.,Frezzotti,M.L.,andAfanasiev,V.P.(2014)N-richfluidinclusionsinoctahedrally-grown526

diamond.EarthandPlanetaryScienceLetters,393,39–48.527

Stachel,T.,andHarris,J.W.(2009)FormationofdiamondintheEarth'smantle.JournalofPhysics:CondensedMatter528

21,36-26.529

Stachel, T., and Harris, J.W. (2008) The origin of cratonic diamonds - Constraints from mineral inclusions. Ore530

GeologyReviews,34,5-32.531

Stagno,V.,andFrost,D.J.(2010)Carbonspeciationintheasthenosphere:experimentalmeasurementsoftheredox532

conditionsatwhichcarbonate-bearingmeltscoexistwithgraphiteordiamondinperidotiteassemblages.Earthand533

PlanetaryScienceLetters,30,72–84.534

Sverjensky,D.A.,Stagno,V.,andHuang,F.(2014)Importantrolefororganiccarboninsubduction-zonefluidsinthe535

deepcarboncycle.NatureGeoscience,7,909–913.536

21

Thomassot, E., Cartigny, P., Harris, J.W., and Viljoen, K.S. (2007) Methane-related diamond crystallization in the537

Earth's mantle: Stable isotope evidences from a single diamond-bearing xenolith. Earth and Planetary Science538

Letters,257,362-371.539

Thomassot, E., Cartigny, P., Harris, J.W., Lorand, J.P., Rollion-Bard, C., and Chaussidon, M. (2009) Metasomatic540

diamond growth: A multi-isotope study (13C, 15N, 33S, 34S) of sulphide inclusions and their host diamonds from541

Jwaneng(Botswana).EarthandPlanetaryScienceLetters,282,79-90.542

Thomson, A.R., Kohn, S.C., Bulanova. G.P., Smith, C.S., Araujo, D.A., EIMF., Walter, M.J. (2014) Origin of sub-543

lithospheric diamonds from the Juina-5 kimberlite (Brazil): constraints from carbon isotopes and inclusion544

compositions.ContributionstoMineralogyandPetrology,168,1081-1100.545

Toplis,M.J.(2005)Thethermodynamicsofironandmagnesiumpartitioningbetweenolivineandliquid:criteriafor546

assessing and predicting equilibrium in natural and experimental systems. Contributions to Mineralogy and547

Petrology,149,22-39.548

Tsuno,K.,andDasgupta,R.(2014)Fe–Ni–Cu–C–Sphaserelationsathighpressuresandtemperatures–Theroleof549

sulfur incarbonstorageanddiamondstabilityatmid- todeep-uppermantle.EarthandPlanetaryScienceLetters,550

412,132-142.551

Urey,HC.(1947)Thethermodynamicpropertiesofisotopicsubstances:JournaloftheChemicalSociety,562-581.552

Watenphul, A., Wunder, B., Wirth, R., and Heinrich, W. (2010) Ammonium-bearing clinopyroxene: A potential553

nitrogenreservoirintheEarth'smantle.ChemicalGeology,270,240-248.554

Walter,M.J.,Bulanova,G.P.,Armstrong,L.S.,Keshav, S.,Blundy, J.D.,Gudfinnsson,G., Lord,O.T., Lennie,A.R.,Clark,555

S.M.,Smith,C.B.,andGobbo,L.(2008)Primarycarbonatitemeltfromdeeplysubductedoceaniccrust.Nature,454,556

7204-7208557

Walter,M.J.,Kohn,S.C.,Araujo,D.,Bulanova,G.P.,Smith,C.B.,Gaillou,E.,Wang,J.,Steele,A.,Shirey,S.B.(2011)Deep558

mantlecyclingofoceaniccrust:evidencefromdiamondsandtheirmineralinclusions.Science,334,54–57.559

Yokochi,R.,Marty,B.,Chazot,G., andBurnard,P. (2009)Nitrogen inperidotitexenoliths:Lithophilebehaviorand560

magmaticisotopefractionation.GeochimicaCosmochimicaetActa,73,4843-4861.561

562

22

Figure1.Histogramshowingthedistributionofnitrogencontents(ppm)ofmantlediamonds.Data frommultiple563

sources(Boydetal.,1987,1992;Bulanovaetal.,2002,2014;Cartignyetal.,1997,1998a,1998b,2001,2003,2004;564

2009;Gautheron et al., 2005;Harte et al., 1999;Howell et al., 2015a;Hauri et al., 2002; Javoy et al., 1984;Klein-565

BenDavid et al., 2010; Mikhail et al., 2013, 2014a, 2014b; Palot et al., 2009; 2012, 2014; Smart et al., 2011;566

Thomassotetal.,2007,2009)567

568

Figure2.Aschematicrepresentationoftheaqueousspeciationofnitrogenandcarbonandthesolidstateofcarbon569

asafunctionofredoxstateat5GPaand1200°C(*denotescalculationsareinequilibriumwithdiamond).Datafor570

aqueousnitrogenspeciationfromMikhailandSverjensky(2014),Mysenetal.,(2014),andLiandKeppler(2014),571

dataforaqueouscarbonspeciationfromSverjenskyetal.(2014)andFrostandMcCammon(2008),andthedatafor572

solidcarbonspeciationarefromStagnoetal.(2010)andRohrbachandSchmidt(2011).573

574

Figure3.Diagramsshowingtheaqueousspeciationofafluidwith1wt.%nitrogenasafunctionofthefluidP,T,fO2,575

andpH(a-d)usingdatafromMikhailandSverjensky(2014).576