a mantle origin for paleoarchean peridotitic diamonds...

Lithos 112S (2009) 852–864

Contents lists available at ScienceDirect

Lithos

j ourna l homepage: www.e lsev ie r.com/ locate / l i thos

A mantle origin for Paleoarchean peridotitic diamonds from the Panda kimberlite,Slave Craton: Evidence from 13C-, 15N- and 33,34S-stable isotope systematics

Pierre Cartigny a,⁎, James Farquhar b, Emilie Thomassot a,c, Jeffrey W. Harris d, Bozwell Wing b,c,Andy Masterson b, Kevin McKeegan e, Thomas Stachel f

a Laboratoire de Géochimie des Isotopes Stables de l′Institut de Physique du Globe de Paris, UMR CNRS 7154, Franceb Earth System Science Interdisciplinary Center and Department of Geology, University of Maryland, USAc Department of Earth and Planetary Sciences and GEOTOP-UQAM-McGill, McGill-University, Canadad Department of Geographical and Earth Sciences, University of Glasgow, UKe Earth and Space Sciences, UCLA, Los Angeles, USAf Department of Earth and Atmospheric Science, University of Alberta, Edmonton, Canada

⁎ Corresponding author.E-mail address: [email protected] (P. Cartigny

0024-4937/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.lithos.2009.06.007

a b s t r a c t
a r t i c l e i n f o
Article history:Received 30 September 2008Accepted 3 June 2009Available online 23 June 2009

Keywords:MantleStable isotopesDiamondContinental lithosphere

In order to address diamond formation and origin in the lithospheric mantle underlying the Central SlaveCraton, we report N- and C-stable isotopic compositions and N-contents and aggregation states for 85diamonds of known paragenesis (73 peridotitic, 8 eclogitic and 4 from lower mantle) from the Pandakimberlite (Ekati Mine, Lac de Gras Area, Canada). For 12 peridotitic and two eclogitic sulfide inclusion-bearing diamonds from this sample set, we also report multiple-sulfur isotope ratios.The 73 peridotitic diamonds have a mean δ13C-value of −5.2‰ and range from −6.9 to −3.0‰, with oneextreme value at −14.1‰. The associated δ15N-values range from −17.0 to +8.5‰ with a mean value of−4.0‰. N-contents range from 0 to 1280 ppm. The 8 eclogitic diamonds have δ13C-values ranging from−11.2to −4.4‰ with one extreme value at −19.4‰. Their δ15N ranges from −2.1 to +7.9‰ and N-contents fallbetween 0 and 3452 ppm. Four diamonds with an inferred lower mantle origin are all Type II (i.e. nitrogen-free) and have a narrow range of δ13C values, between−4.5 and−3.5‰. The δ34S of the 14 analyzed peridotiticand eclogitic sulfide inclusions ranges from −3.5 to +5.7‰. None of them provide evidence for anomalousδ33S-values; observed variations in δ33S are from +0.19 to −0.33‰, i.e. within the 2 sigma uncertainties ofmantle sulfur (δ33S=0‰).At Panda, the N contents and the δ13C of sulfide-bearing peridotitic diamonds show narrower ranges thansilicate-bearing peridotitic diamonds. This evidence supports the earlier suggestion established from eclogiticdiamonds from the Kaapvaal that sulfide-(±silicate) bearing diamonds sample a more restricted portion ofsublithosphericmantle than silicate-(no sulfide) bearing diamonds. Our findings at Panda suggest that sulfide-bearing diamonds should be considered as a specific diamond population on a global-scale. Based on our studyof δ34S, Δ33S, δ15N and δ13C, we find no evidence for subduction-related isotopic signatures in the mantlesampled by Panda diamonds.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Within the last 30 years, using evidence from the study of deep-seated xenoliths and diamonds from southern African kimberlites,numerous models have been proposed to describe the formation andorigin of the continental cratonic lithosphere (e.g. Boyd and Gurney,1986; Haggerty, 1986; Pearson and Wittig, 2008). The discovery andmining of diamondiferous kimberlites on the Northern (Jericho),Central (Ekati, DO-27, Diavik) and southern Slave Craton (Snap Lake)provides new opportunities to test and refine these models.

).

ll rights reserved.

In the last few years, several first order distinctions between thesubcratonic lithosphericmantle beneath the Slave andKaapvaal Cratonshave been made. Diamonds from the Slave are less resorbed than theirSouth African counterparts (Stachel et al., 2003; Gurney et al., 2004)and, in this respect, are more similar to Siberian diamonds. Diamondsfrom the Slave also include a higher proportion of coated stones(commonly referred to as fibrous diamonds although this represents asimplification since some coats can actually be well-crystallised andnon-fibrousdiamond, c.f.Moore,1985). Eclogitexenoliths fromtheSlave(Jericho and Diavik) have been dated to ~2.1 Ga (Paleoproterozoic)(Schmidberger et al., 2005, 2007) and thus are generally younger thaneclogitexenoliths from theKaapvaal (andYakutia)which are principallyArchean in age (see Pearson et al., 1995a,b). In contrast, the formation ofperidotitic lithosphericmantle components in the Central Slave extends

mailto:[email protected]

http://dx.doi.org/10.1016/j.lithos.2009.06.007

http://www.sciencedirect.com/science/journal/00244937

Table 1δ13C, δ15N, N contents (determined by infrared spectroscopy and/or bulk combustion) and percentage of the nitrogen B species, averaged δ34S and Δ33S of sulfides in diamonds fromPanda.

Sample paragenesis Weight δ13C δ15N − + NCOMB NCOMB NFTIR B δ34S Δ33S

(mg) (‰) (‰) (ppm) (at.ppm) (at.ppm) (%) (‰) (‰)

PA01 p H 2.3437 −5.58 0.9 −0.6 0.6 451 387 348 22PA02 p H 3.0041 −5.22 1.2 −0.5 0.5 826 708 372 11PA03 p H 1.7632 −4.58 0PA04 p H 1.7632 −4.58 0PA05 p H 2.2344 −5.81 5.7 −0.6 0.8 267 229 225 19PA06 p H 2.3977 −5.51 26 3PA07 p H 0.7387 −5.21 67 47PA08 p H 0.6741 −5.19 0PA09 p H 2.7824 −5.28 −4.1 −0.6 0.5 744 638 631 89PA10 p H 1.3059 −5.65 1.2 −0.8 0.9 116 99 125 56PA11 p L 0.5774 −5.32 −10.8 −1.0 0.5 550 471 345 80PA12 p H 1.1805 −5.05 12PA13 p H 2.7696 −5.07 1.2 −0.8 0.9 50 43 56 40PA14 p H 2.9360 −6.93 −1.5 −0.6 0.5 498 427 296 41PA15 p H 1.5258 −14.05 0PA16 p H 0.9322 −5.00 −15.8 −2.4 0.1 96 82 131 61PA17 p H 1.5453 −5.45 17 0PA18 p H 0.4054 −4.99 106 71PA19 p L 1.0909 −5.71 3.4 −0.7 1.0 127 109 65 47PA20 p H 0.7688 −5.01 −15.4 −1.4 0.3 253 217 118 62PA21 p H 1.5788 −5.50 −0.9 −1.0 0.9 91 78 73 51PA22 p H 1.4411 −4.27 −17.0 −1.0 0.4 183 157 192 79PA23 p L 3.2667 −4.85 2.4 −0.5 0.5 680 583 543 57PA24 p 3.2115 −3.58 −0.5 0.5 577 495 464 7PA25 p 3.7332 −5.73 −6.4 −0.6 0.5 467 400 544 85PA26 p 1.2280 −5.14 −5.1 −0.8 0.6 245 210 269 11PA27 p 1.7429 −4.33 27 0PA28 p 1.7695 −4.98 −7.5 −1.2 0.6 90 77 174 6PA29 p 0.9011 −5.32 −10.8 −0.9 0.5 267 229 292 0PA30 p 1.2522 −4.77 −6.3 −0.7 0.6 375 321 78 9PA31 p 3.9765 −5.05 −6.9 −0.6 0.5 228 195 237 8PA32 p 1.1208 −2.98 148 2PA33 p 1.1550 −4.87 −3.5 −0.8 0.7 248 213 74 0PA34 p 0.9944 −5.31 −7.7 −1.0 0.6 226 194 271 13PA35 p 1.2376 −5.07 −6.8 −1.1 0.6 155 133 184 14PA36 p 0.8947 −4.73 −0.2 −0.6 0.6 1280 1097 609 8PA37 p 2.0017 −5.96 3.0 −0.6 0.7 135 116 79 45PA38 p 1.1500 −4.34 0.2 −0.6 0.6 572 490 404 26PA39 lm ? 1.4933 −3.55 0PA40 p L 2.5249 −5.75 0PA41 p 1.7900 −5.12 −4.0 −0.7 0.6 223 191 55 62PA42 p H 3.0033 −4.86 −14.5 −0.6 0.5 532 456 333 81PA43 p 1.3835 −5.15 −3.1 −0.7 0.6 330 283 118 37PA44 unk 1.6425 −5.74 3.8 −1.0 1.6 105 90 57 34PA45 p 2.3526 −4.87 8.5 −0.5 1.1 87 75 70 40PA46 p 2.4770 −5.71 3.8 −0.7 0.9 104 89 57 49PA47 p 1.2538 −4.27 0.4 −0.6 0.6 1072 919 782 7PA48 p 1.3928 −5.03 1.3 −0.6 0.6 995 853 834 22PA49 p 1.1653 −5.45 −1.8 −1.2 1.0 93 80 494 85PA50 lm ? 1.3593 −4.01PA51 p 2.8825 −5.57 1.3 −0.6 0.6 745 639 608 7.4PA52 p 0.7370 −5.71 0.7 −0.6 0.6 581 498 341 46PA53 unk 3.3445 −5.79PA54 lm ? 0.8656 −3.50PA55 lm ? 1.1777 −5.27 13PA56 p 0.6544 −6.36 −3.2 −0.6 0.6 1124 963 859 92PA57 p S 1.1167 −4.96 −1.1 −0.6 0.6 487 417 338 13 1.2 0.15PA58 p H 0.9998 −5.40 −0.8 −0.6 0.6 386 331 72 52PA59 p L 2.5684 −4.42 0PA60 p L 1.9878 −6.10 3.5 −0.5 0.5 1074 921 620 86PA61 p L 3.3804 −4.46 0PA62 p L 1.4952 −4.56 0PA63 e 0.9902 −11.21 0PA64 e 1.9441 −5.10 −2.1 −0.5 0.5 823 705 648 3PA65 e 0.9823 −19.40 7.9 −0.5 0.5 3452 2959 2720 100PA66 e S 0.8847 −9.90 2.5 −0.6 0.7 741 635 638 69 2.3 0.15PA67 p L 0.8318 −4.59 −2.4 −1.0 0.8 227 195 184 100PA68 p L 0.5623 −5.59 −1.2 1.2 275 236 262 100PA69 p L 1.4254 −5.29 −2.5 −0.6 0.6 1114 955 1028 14PA70 p L 0.2135 −5.07 184 158 64 11PA71 e S 1.1519 −8.36 14 0PA72 unk 1.0974 −4.46 0

(continued on next page)

853P. Cartigny et al. / Lithos 112S (2009) 852–864

Table 1 (continued)

Sample paragenesis Weight δ13C δ15N − + NCOMB NCOMB NFTIR B δ34S Δ33S

(mg) (‰) (‰) (ppm) (at.ppm) (at.ppm) (%) (‰) (‰)

PA73 e 1.0030 −5.45 −1.7 −0.6 0.6 998 855 881 3PA74 e 0.7770 −4.37 −2.0 −0.6 0.6 1398 1198 1084 4PA75 p S 1.6073 −4.73 −7.6 −0.6 0.5 805 690 700 4 2.3 0.19

−1.2 0.14PA76 e S 3.3913 −4.91 −0.7 −0.5 0.5 1193 1023 985 11 2.0 0.19PA77 unk 2.0407 −3.91 0.9 −0.8 0.8 104 89 881 17PA78 p S 1.9185 −5.12 −6.8 −0.6 0.5 539 462 424 8 4.1 0.18PA79 p S 0.8211 –4.68 −12.4 −1.5 0.4 227 195 242 0 4.0 0.10PA80 p S 3.3774 −4.90 −5.0 −0.5 0.5 754 646 663 9 1.6 −0.18

3.6 0.06PA81 p S 1.3608 −5.07 −1.7 −1.1 0.9 98 84 114 23 −3.5 −0.03PA82 p S 0.8347 −4.33 −15.0 −1.8 0.2 148 127 123 5 1.3 −0.02PA83 p S 3.2817 −5.37 −9.6 −0.6 0.5 519 445 474 6 4.1 −0.12PA84 p S 2.2171 −5.14 −10.6 −0.8 0.5 205 176 86 22 2.9 −0.33PA85 p S 2.1668 −2.98 −2.5 −0.6 0.6 397 340 333 0 3.2 0.03

2.2 0.17PA86 unk 1.5134 −4.20 −7.2 −0.7 0.5 305 261 291 0PA87 p S 1.0606 −5.21 23 0 4.1 −0.10

4.1 −0.08PA88 p S 1.6038 −4.93 −7.6 −0.6 0.5 613 525 468 9PA89 p S 2.4886 −4.78 −0.5 −0.5 0.5 753 645 552 2 3.1 0.02PA90 p S 0.6945 −5.81 12 0 5.7 −0.11

E and P denote peridotitic and eclogitic diamonds, respectively, whereas S denotes eclogitic or peridotitic diamonds bearing sulfide; H and L stand for the harzburgitic and lherzoliticsubparageneses respectively. lm ? and unk. denote possible lower-mantle and unknown paragenesis samples, respectively.

854 P. Cartigny et al. / Lithos 112S (2009) 852–864

back to the late Paleoarchean (from 3.3 up to 3.5 Ga, Westerlund et al.,2006; Aulbach et al., 2004), apparently pre-dating Neoarcheanstabilization of the Slave crust at ~2.6 Gy (Davis et al. 2003) whereasin the Kaapvaal, this event has often been suggested to be penecon-temporaneous (Pearson et al., 1995a,b; but see Pearson and Wittig,2008). These latter observations may point to extended and complexprocesses associated with lithospheric mantle formation beneath theSlave. Furthermore, diamond formation as well differ beneath the twocratons both in terms of the timing of diamond formation relative tolithosphere crystallisation and with respect to the relative abundancesand ages of peridotitic and eclogitic sulphide inclusion-bearingdiamonds (reviewed in Aulbach et al., this issue).

Xenocryst and xenolith from beneath the Slave Craton reveal auniquely stratified lithosphere structure, with a downward transitionfrom ultra-depleted (60% harzburgite, 40% lherzolite) to deeper lessdepleted (15–20% harzburgite and 80–85% lherzolite) lithologiesoccurring at ~145 km depth (Griffin et al., 1999; Menzies et al., 2004).This compositional boundary approximately coincides with the gra-phite–diamond transition along a typical cratonic geotherm. Location ofdiamond sources in the less depleted deeper layer is reflected in themajor element compositions of silicate inclusions in peridotiticdiamonds from the central Slave Craton, which are much less depletedthan similar inclusions from the Kaapvaal Craton (Stachel et al., 2003).Griffin et al. (1999) proposed that the deep layer of the Slave lithosphereresults from a plume head, which rose from the lower mantle andunderplated the existing lithosphere (Griffin et al., 1999). Evidence insupport of this proposal includes the presence of ferropericlaseinclusions with an inferred lower mantle origin in diamonds fromseveral Central Slave kimberlites (Davies et al.,1999, 2004; Tappert et al.,2005; Donnelly et al., 2007). However, the presence of ferropericlaseinclusions on their own (i.e. without coexisting MgSiO3 and CaSiO3

inclusions), does not necesserily require a lower mantle origin asderivation from reduced dunitic lithospheric sources is possible (e.g.Stachel et al., 1998b; Brey et al., 2004; Tappert et al. 2005).

Osmium-data are available on peridotitic sulfide inclusions from thePanda kimberlite (Ekati) (Westerlund et al., 2006) and sulfide inclusionsin mantle xenocrysts from the A-154 kimberlite (Diavik; Aulbach et al.,2004). Both kimberlites show high initial Os-isotopic compositionsassociated with low Re/Os and comparatively high Os-contents. Theorigin of this signature is unclear, it may be attributable to inheritance

from subduction-related fluids (Westerlund et al., 2006) or to an ultra-deepmantle origin, i.e. from the lowermantle or the outer core (Aulbachet al., 2004) and, in both cases, would thus point to distinct diamondgrowth conditions or at least a distinct origin of sulfides at these twolocalities. Similarities in the C-isotope distribution of diamonds (Donnellyet al., 2007; see below) among A-154, DO-27 and Panda suggest that thischaracteristic might apply more generally to Central Slave diamonds.

The present study concentrates on the stable isotope characteristicsof diamonds from the Panda kimberlite and aims to improve under-standing of the origin and formation of diamonds at this locality – inparticular by using proxies of subducted components such as 15N/14N,33S/32S and 34S/32S– to gain insights into the origin ofmantlefluids andto evaluatewhether fluid derivation from a paleo-subductionwedge isrequired for peridotitic diamond formation at this locality.

2. Geological background and studied diamonds

The Panda kimberlite, located ~25 km north of Lac de Gras, was thefirst diatreme mined at the Ekati Mine Ekati and has been active since1998. The Panda kimberlitewas emplaced 53.2±0.3 Ma (Creaser et al.,2004) and belongs to the Lac de Gras kimberlite field, which includeamong others Diavik, Koala, Beartooth and Fox.

All the diamonds studied here (n=90) were previously charac-terised and discussed by Stachel et al. (2003; peridotitic diamonds)and Tappert et al. (2005; eclogitic and lower mantle diamonds),including descriptions of physical characteristics (size, shape, color,plastic deformation), inclusion type and chemistry, and infrared-characteristics (N-content and aggregation).

Excluding five samples of uncertain paragenesis, Tappert et al.(2005), observed that Panda mainly yields peridotitic diamonds(n=73, 86%), with a less abundant eclogitic (n=8, 9%) and lowermantle (n=4, 5%) diamond population. Most inclusions were silicatesand 17 diamonds contained at least one sulfide. Although somesulfide-bearing diamonds contained additional silicate inclusion(s),these diamonds will be referred to as sulfide-bearing diamonds.According to their high Ni-content (Yefimova et al., 1983, see alsoHarris, 1992), most sulfides (i.e. 14 of 17) can be assigned to theperidotitic paragenesis. Their Cr-contents (Stachel and Harris, 2008;Thomassot et al., 2009) or Os-contents (Pearson et al., 1999) havebeen shown to be good complementary paragenesis proxies. For


example, the peridotitic sulfides at Panda have Cr2O3-contentsbetween 0.15 and 0.32wt.% whereas the eclogitic sulfides have Cr-contents below the limits of detection (≤0.02 wt.%). Os-contents werenot measured on these samples. Finally, for silicate-bearing diamonds,the occurrence of pyroxene and the chemical composition of garnetinclusion-bearing diamonds (see Stachel and Harris, 2008) allow tobetter define the mineralogy of the host-peridotite, with 22 and 12belonging to harzburgitic and lherzolitic sub-parageneses, respec-tively (Stachel et al., 2003; Tappert et al., 2005) and these are referredto H and L in Table 1.

3. Analytical techniques

From the fragments left after diamond breakage, a diamond chip foreach of the 90 Panda diamonds (from 0.40 to 3.97 mg) was selected.These were characterised by micro-Fourier Transform infrared (FTIR)spectroscopy for N-contents (hereafter referred to as NFTIR) andaggregation states using a Nicolet Magna IR 550 coupled to a Spectra-Tech IR-microscope at the DTC Research Laboratories in Maidenhead(UK). After conversion to absorption coefficient, the spectra weredeconvoluted into the A, B and D components using least squaretechniques. Nitrogen concentrations were calculated using the absorp-tion coefficient at 1282 cm−1 of 16.5 and 79.4at.ppm.cm−1 for the A-and B-centres, respectively (Boyd et al., 1994a, 1995). Detection limitsdepend on the flatness of the cleaved diamond surface and rangebetween 10 and 20 ppm. Errors on NFTIR contents and aggregation statedeterminations are about 10–20% and±5%, respectively, as determinedfrom multiple deconvolutions of single spectra and considering errorson the absorption coefficients (see Discussion section). The NFTIR data inTable 1were already reported inTappert et al. (2005) but are listed hereagain to link FTIR and subsequent stable isotopedata thatwere obtainedon the same diamond fragment.

C- and N-stable-isotope measurements were undertaken at IPG-Paris following procedures described in Cartigny et al. (2004b). Type II(i.e. N-free) diamonds were combusted in a pure O2 atmosphere andthe produced CO2 was analysed for carbon isotopic compositionsusing a conventional dual-inlet mass spectrometer (Finnigan DeltaXP) with the data being expressed in delta notation relative to the PDBstandard, δ13C=(13C/12Csample / 13C/12CPDB−1)×103, with an accuracybetter than 0.05‰ (2σ) Type I (i.e. N-bearing, N20at.ppm) diamondswere combusted using a different vacuum-line to allow for simulta-neous characterisation of C- and N-stable isotopes and N-contents(hereafter referred to as Ncomb). After diamond combustion in a pureO2 atmosphere, the produced CO2 was trapped as calcium carbonateusing CaO. Trace amounts of carbon monoxide and water in theremaining nitrogen gas, which would cause isobaric interferences,were removed through a double purification using Cu held at 600 and450 °C, respectively. After quantification of the released amount ofnitrogen using a capacitance manometer with a precision of betterthan 5% (2σ), the 15N/14N-ratio of the nitrogen gas was analysed usinga purpose built triple collector static mass spectrometer with anaccuracy better than ±0.5‰ (2σ) and expressed in delta notationrelative to Air, δ15N=(15N/14Nsample/15N/14NAir−1)×103. For δ13Cmeasurements, CO2 was recovered after heating the CaO–CaCO3

mixture to 850 °C and analysed as described above with accuraciesbetter than ±0.1‰ (2σ). Although the analytical accuracy of δ15Ndeterminations is better than ±0.5‰ (2σ) – as attested by theanalyses of the international IAEA-N1 and -N2 standards – errorsreported in Table 1 can be larger and take into account a maximumblank contribution of 3.5 ng N with an assumed δ15N between−20 to0‰ (estimates are about −10±5‰).

For 18 sulfide inclusions from 15 diamonds (13 peridotitic+2eclogitic), sulfur isotope compositions were measured using a CamecaIMS-1270 ion microprobe at UCLA following techniques described inFarquhar et al. (2002). Sulfur isotope data for sulfide inclusions werecollected over the course of three analytical sessions (April 2005, August

2005 and August 2006). Analyses were undertaken using a primary Cs+

ion beam thatwasmass-filtered and accelerated to 20 keV impact energy.Typical operating conditions included a spot size of 20–25 micron and~2 nA beam current. A low energy, normal incidence, electron gun wasused for charge compensation. Isotope ratios were determined fromsimultaneous measurement of the 33−/32− and 34−/32− ion beams onFaraday cup detectors at a mass resolution of ~4000, allowing to separatethe 32SH− contribution from 33S−. Standardization was undertakenrelative to analyses of CanyonDiabloTroilite (CDT) determinedduring thesame session, but not mounted together with the sulfide inclusions.Potential drift in Δ33S was monitored using secondary sulfide standardsmounted together with the sulfide inclusions. Secondary standards andunknownsweremounted in 1inch epoxypucks thatwere prepared usingpreexisting (from the study of Tappert et al., 2005) samplemounts (½cmdiameter brass rings) for the sulfide inclusions in diamonds. Itwas judgedthat plucking of the inclusions from thesemounts and repolishing wouldhave resulted in unacceptable material loss (an unpublished experienceby the first and third author on sulfide inclusion from Koffiefontein) butthis approach has the drawback of larger variations in sample topographyand positioning of grains away from the center of the mount. It requiredmore extreme focussing settings of the primary and secondary ionbeams, leading to larger uncertainties in Δ33S. Sulfur isotope ratiosare expressed in delta notation relative CDT as δ34S=(34S/32Ssample/34S/32SCDT−1)×103 and deviation from the terrestrial fractionation array isdefined as Δ33S=δ33S−1000×[(δ34S/1000+1)0.515−1]. Two sigmauncertainties are estimated at ±0.4‰ and ±1.4‰ for single Δ33S andδ34S analyses respectively. No attempt was made to correct forinstrumental mass fractionation, but analyses of reference materialsundertaken during the course of the study suggest that differences in theIMF are ~7‰ and ~3‰ relative to CDT for pyrite and pyrrhotite,respectively. Compared to eclogitic sulfide inclusions which usuallyconsist of low temperature exsolutions of chalcopyrite, pentlandite andpyhrrotite, peridotitic sulfide inclusions show only fine-grained exsolu-tion of pyrrhotite and pentlandite (emulsion-like texture according toThomassot et al., 2009); the correction of IMF is thus considered to be lesscritical than for eclogitic inclusions and the δ34S-range is thus representa-tive of the population.

4. Results

4.1. Peridotitic diamonds

The 73 peridotitic diamonds (including 14 bearing-sulfide) have amean δ13C-value of−5.2‰ and range from−6.9 to−3.0‰, with onesingle extreme value at −14.1‰ (Fig. 1, Table 1). The associated δ15N-values of Type I diamonds (n=57) range from−17.0 to+8.5‰with amean value of −4.0‰ (Fig. 2). NFTIR-contents range from 0 to1280 ppm with an average NFTIR-content of 260±250 at.ppm and anabundance of Type II diamonds of ~15% (n=11, defined with Nb20at.ppm). N-aggregation states varywidely from0 to 100% (average ~32%)of nitrogen in the fully aggregated B center.

Lherzolitic (n=11) andharzburgitic (n=22)diamondshaveδ13C (seeFig.1), δ15N andN-ranges and distributions that are overall similar (Figs. 2and 4). The lherzolitic δ13C-distribution, although overlappingwith that ofharzburgitic diamonds, is, however, not well-defined and a highproportion (four out of 11, i.e. ~36%) of lherzolitic diamonds are Type II.

Compared to silicate-bearing diamonds, sulfide-bearing peridotiticdiamonds show a smaller range for δ13C (−5.8 to −3.0‰), for δ15N(−12.4 to −0.5‰), for N-aggregation states from 0 to 22.6 %B(average=7%, n=14) and a restricted range of overall higher N-contents (noType II, average NFTIR=325 ppm, n=14) (Figs. 1, 2 and 7).Sulfur isotope data (Table 1 for averaged values of single inclusion,Table 2 for individual analyses) have been obtained for 15 sulfideinclusions from 12 sulphide-bearing diamonds (Table 2). δ34S-valuesrange from −3.5 to +5.7‰ with an average of 2.5±2.2‰ (med-ian=3.1‰). The Δ33S of these inclusions range from −0.33 to 0.15‰

Fig. 1. δ13C histograms of Panda diamonds (right). Peridotitic diamonds are also subdivided into several sub-populations (left) and it can be noticed that they are all characterised bya main mode in δ13C at ~5‰. Note that the histograms on the right do not display the most extreme δ13C-value.


with an average value of 0.02±0.15‰ (the median value is 0.03‰). Ofthe analyses, 12 sulfide grains fall within 1 sigma, two grains within 2sigma and one grain (PA-84) within 2.4 sigma of Δ33S=0‰.

4.2. Eclogitic diamonds

The eight eclogitic diamondshave δ13C-values ranging from−11.2 to−4.4‰ (n=7) with one sample (PA-65) showing an extreme value at−19.4‰ (Fig. 1). Their δ15N ranges from −2.1 to +2.5‰ (n=5), withsample PA-65 showing a δ15N of +7.9‰ (Fig. 2). NFTIR-contents rangefrom 0 to 1090 at.ppm (n=7), with the exception of PA-65 with a N-content of 2720 at.ppm. Nitrogen aggregation states range from 0 to100%B (average is 27%).

The two eclogitic sulfide-bearing diamonds (PA-66 and PA-71,Table 1) show low δ13C-values of −9.9 and −8.4‰, respectively,associated with N-contents of 14 and 741 at.ppm and a δ15N-value of+2.5‰ for the latter. Their two sulfide inclusions have δ34S-values of+2.0 and +2.3‰ and Δ33S of +0.19 and +0.15‰, respectively(Tables 1 and 2). These values are within the near zero Δ33S range ofmass-dependant sulfur isotope fractionations.

4.3. Lower mantle diamonds

The four samples bearing ferropericlase inclusions are all Type II(nitrogen-free) and have a narrow δ13C interval between −4.5 and−3.5‰ (Fig. 1). These data are very similar to other lower mantle

Fig. 2. δ15N histograms of Panda diamonds compared to worldwide data. Note that all sulfide-bearing diamonds (darker color) are characterised by strictly negative δ15N–values.


diamonds worldwide falling within a restricted range of δ13C of ~3‰(e.g. Harte et al.,1999; Kaminskyet al., 2001; Stachel et al. 2002). In thedetail however, the Panda data are shifted to higher δ13C and are thuscloser to diamonds from the Kankan deposits in Guinea (Stachel et al.2002). In this study, these diamonds will not be addressed further.

4.4. Unknown paragenesis samples

Thefive samples of unknownparagenesis have C- (δ13C from−5.8 to−3.9‰) and N- (δ15N from −7.2 to +3.8‰) isotope values, NFTIR-contents (0 to 881 ppm, twoType II) and N-aggregation states (0 to 34%B) falling in the range of known-paragenesis samples (Table 1) andwillnot be discussed further.

4.5. NFTIR versus Ncomb

For all but one (PA-77) Type I diamond from Panda (n=66), agood agreement exists between N-contents determined by FTIR andvia bulk combustion (Fig. 3). Coupled NFTIR and Ncomb data fordiamonds from other locations (see Cartigny et al., 2003, 2004b andreferences therein) also display good agreement but with morescatter. The likely origin for the scatter lies in the different samplevolume for infrared and bulk combustion analyses. Micron scalezoning of N-content has previously been documented using e.g.cathodoluminescence (e.g. Harte et al., 1999), FTIR and SIMS (e.g.Bulanova et al., 2002; Zedgenizov and Harte, 2004). Such studies haveshown that diamonds can be highly heterogeneous in N-contents(from Type II to N-rich zones) on both sample (from core to rim) andmicron scale compared to fairly homogenous C-isotopes (usuallyshowing b2‰ of intra-variability). The present data document relativehomogeneity of Panda diamonds on a millimeter scale and two mainimplications can be derived from the good agreement between NFTIR

and Ncomb. Firstly, the present diamond dataset spans a large range ofIaB aggregation (from 0 to 100%, Table 1), thus confirming theaccuracy of the infrared absorption coefficients for the A and B-speciesdetermined by Boyd et al. (1994a, 1995). Secondly, overall Pandadiamonds might have grown over less variable conditions (super-saturation and/or oxygen fugacity) than southern African diamonds. Acomparative study including detailed diamondmapping is required toconfirm such an inference.

5. Discussion

Studies of eclogitic diamonds brought to the surface by kimberlitesand lamproites have been driven by two principal motivations. Eclogiticdiamonds are thought to increase in relative proportion among the largersize classes, providing an economic motivation (Gurney, 1989; Stacheland Harris, 2008). The second motivation is provided by the informationthey may yield for understanding the role of subduction in the originof cratonic eclogites (Liu et al., this issue; Jacob et al., this issue;Spetsius et al., this issue). Peridotitic diamonds, despite usually beingmore abundant (Stachel and Harris, 2008), have received less attention.Peridotitic diamonds are generally assumed (see Kesson and Ringwood,1989; Schulze,1986 foropposingviews) to represent a reference frame forlithospheric mantle that has remained largely unaffected by subduction-related processes and metasomatism by slab-derived fluids (e.g. Deinesand Harris 2004; Cartigny et al. 2004b and references therein).

The concept of old stable continental roots, remaining largelyisolated from the convecting mantle since the Archean (e.g. Walkeret al.,1989; Pearson et al.,1999) and being subsequently affected only bymetasomatic re-enrichment (e.g. Stachel andHarris,1997b; Burgess andHarte, 2004; Simon et al., 2007) and only rarely becoming remobilisedor thermally eroded (Zheng et al., 1998; Gao et al., 2002; Griffin et al.,this issue), is a widely accepted concept. The original idea that thelithospheric mantle represents the residue of high-pressure melting(≥40 kbar, e.g. Herzberg,1999) is not supported by the major and traceelement composition of peridotite xenoliths which points to derivationas the residue ofmelt extractionmainly at intermediate to lowpressures(b20–30 kbar; Stachel et al., 1998a; Kelemen et al., 1998; Walter, 1999;Pearson and Wittig, 2008). Recent models thus place the formation oflithopshericmantlewithin the framework of plate tectonics and involvea link with subduction (Helmstaedt and Schulze, 1989; Stachel et al.,1998a; Pearson and Wittig, 2008). The formation of peridotiticdiamonds might also be related to (or initiated by) subduction-relatedfluids (e.g. Kesson and Ringwood, 1989; Pearson and Wittig, 2008)although unambiguous evidence for this inference is still lacking.

The answer to the question of whether peridotitic diamond growthmay occur from a range of fluids with distinct origins, isotopic char-acteristics and chemistries (e.g. subduction- vs. mantle-related, highvs. low volatile contents, oxidised vs. reduced) is not clear. It is worthnoting that diamond characteristics show systematic variations withthe cratonic setting of their host kimberlites, for example from centralpositions to the rims of cratons, or to the relative distance to paleo-

Fig. 3. Comparison of N-contents measured by infrared spectroscopy and via bulkcombustion.

Table 2Individual δ34S and Δ33S measurements on sulfide in diamonds from Panda.

Sorted grain By session δ34S(‰)

δ33S(‰)

Δ33S(‰)

PA-57C c 2.7 1.4 −0.03PA-57C c 1.6 0.6 −0.21PA-57C c 0.8 0.5 0.07PA-57C e 1.1 1.1 0.57PA-57C e −0.1 0.3 0.36PA-66B a 4.7 2.6 0.16PA-75A d 1.7 1.1 0.19PA-75A d 2.8 1.6 0.19PA-75B c 0.9 0.6 0.16PA-75B c 0.3 0.5 0.35PA-75B c −1.7 −0.8 0.01PA-75B c −3.0 −1.6 −0.03PA-75B e −2.4 −1.0 0.19PA-76A c 2.7 2.0 0.56PA-76A c 2.1 1.1 −0.04PA-76A c 1.2 0.7 0.06PA-78A a 4.0 2.1 0.05PA-78A a 4.1 2.3 0.15PA-79A c 4.2 2.5 0.32PA-79A c 3.7 1.7 −0.25PA-79A c 3.0 1.7 0.11PA-79A e 5.1 2.9 0.23PA-80A c 1.6 0.7 −0.18PA-80B a 3.6 1.9 0.06PA-81A c −3.0 −1.6 −0.05PA-81A c −2.6 −1.4 −0.10PA-81A e −4.9 −2.4 0.07PA-82A c 2.7 1.4 0.01PA-82A c 1.2 0.7 0.04PA-82A c −0.6 −0.3 0.02PA-82A c 0.5 0.1 −0.16PA-82A e 2.5 1.3 −0.03PA-83A b 4.1 2.0 −0.12PA-84A d 4.8 2.2 −0.27PA-84A d 1.0 0.1 −0.39PA-85A a 3.3 1.8 0.07PA-85A a 3.1 1.6 −0.02PA-85B c 2.7 1.9 0.48PA-85B c 2.6 1.3 −0.08PA-85B c 2.1 1.0 −0.04PA-85B c 1.5 0.7 −0.07PA-85B e 3.4 2.1 0.36PA-85B e 1.3 1.0 0.36PA-87A b 4.1 2.0 −0.10PA-87B b 4.7 2.5 0.08PA-87B d 3.8 1.9 −0.07PA-87B d 3.8 1.7 −0.25PA-89A a 3.0 1.6 0.00PA-89A a 3.2 1.7 0.03PA-90A b 5.0 2.5 −0.10PA-90A d 5.3 2.6 −0.08PA-90A d 6.7 3.3 −0.14

a=April 26th, 2005, b=April 29th, 2005, c=August 2005, d=April 26th 2006.


subduction zones (e.g. Shirey et al., 2003; Stachel et al. 2004 andreferences therein) raising thepossibility of distinctmodes of peridotiticdiamond growth. The diamonds from Panda studied here provide avaluable resource to investigate this possibility. Sulfide inclusions inPanda diamonds show high initial Os-isotope compositions (Wester-lund et al., 2006), and contrastwith the fairly unradiogenic Osmeasuredfor other populations of peridotitic sulfides in diamonds and peridotitexenoliths worldwide (see Shirey et al., 2003 for a review). For thisreason, it has been suggested that Panda peridotitic diamonds grewfromsubduction-relatedfluids that allowedOs-addition and resettingofthe Re–Os isotope system before individual diamond inclusions wereentrapped (Westerlund et al., 2006).

The purpose of using nitrogen isotope compositions to studydiamonds lies in the significant differences between the average N-isotope values of mantle nitrogen (average δ15Nb0) and nitrogenderived from subducted sediments (average δ15NN0) (e.g. Cartignyet al., 1998a,b). Such a distinction in nitrogen isotopic signature

contrasts with δ13C values of sedimentary carbon that bracket theisotopic composition of mantle carbon and which, therefore, could re-equilibrate at depth to mantle-like values. δ13C close to −5‰ could,for example, result from either isotopic mixing. involving 80%carbonate (with δ13C~0‰) and 20% organic matter (with δ13C~−25‰), or carbonate devolatilisation. Alternatively, stable isotopefractionation in the Earth's mantle or crust has also been invoked toexplain δ13C-variations as large as observed in sediments (e.g. Deines,1968; Galimov, 1991) making δ13C, when considered by itself, a prioria non definitive proxy.

Nitrogen sequestered in sediments is initially of atmospheric origin,being trapped by N-fixing organisms such as cyanobacteria; a processassociated with very little fractionation resulting in δ15N~0±4‰(Fogel and Cifuentes, 1993). The present day marine biogeochemicalcycle of nitrogen is more complex and biogeochemical reprocessingresults in 15N enrichment in oceanic nitrogen (nitrate and ammo-nium). The nitrogen isotopic composition of modern sediments thusshows average positive δ15N-values, of ~+2‰ for Archean and ~+5‰for younger (recent) samples (Peters et al., 1976; Beaumont andRobert, 1999; Pinti et al., 2001; Garvin et al., 2009; review by Thomazoet al., in press). This observation that Archean sediments display loweraverage δ15N-values is thought to reflect either a simplified N-cyclewhere nitrates was absent (Beaumont and Robert, 1999) or mantle-related nitrogen in cherts (Pinti et al., 2001).

Nitrogen is fixed in sediments as ammonium ions (NH4+) which is

derived from metamorphosed organic matter originally containingpercent levels of nitrogen. Part of the nitrogen released duringdiagenetic maturation of organic matter enters silicate minerals asammonium ions substituting for potassium in K-bearingminerals suchas illite, smectite or phengite (e.g. Honma and Itihara, 1981). Withincreasing metamorphism, e.g. during subuduction, nitrogen alongwith other volatile elements may be partially volatilised (but seeBusigny et al., 2003), leading to further 15N-enrichement in residualnitrogen (Bebout and Fogel, 1992).

Of relevance to studies of nitrogen isotopes in the Panda diamonds aredata for early Archean sediments (Beaumont andRobert,1999; Pinti et al.,2001), which may display lower δ15N than most present-day sediments,down to −7‰. Marty and Dauphas (2003) proposed that subductednitrogen in the Archeanwould have had an average δ15Nof−5‰ (a valuefalling within the upper mantle δ15N-range) but neglected that theaverage δ15N-values of Archean sediments is ~+2‰. With increasingmetamorphism, these values would again have lead to further 15N-enrichments and thus positive average δ15N for subducted nitrogen(Cartigny and Ader, 2003). In other words, we argue that subducted


nitrogen is characterised by 15N-enriched average isotope compositionsand any subduction-related sedimentary component has to carry thistypical signature.

Another nitrogen reservoir to be considered is altered oceaniccrust. At some localities, present-day altered oceanic crust displaysδ15N as low as −11.6‰ with a weighted average of −2.6‰ (ODP 801hole; Li et al., 2007). Other localities, however, display positiveweighted average δ15N compositions (ODP 504B; Busigny et al.,2005b). Because too little δ15N data are available yet, it is difficult toprovide average values for both N-contents and δ15N. Support forpositive average δ15N-values for subducted altered oceanic crustcomes from observations that a.) altered oceanic crust is likely todehydrate about 98% of its original water (Dixon et al. 2002) leading toresidual 15N-enriched rocks, b.) metamorphic nitrogen from sub-ducted hydrothermally-altered oceanic crust (Busigny et al., 2005a;Halama et al., 2008), c.) subducted hydrothermally-altered oceaniclithosphere (Busigny et al., 2005b; Philippot et al., 2007), and d.)metamorphic diamonds formed within crustal rocks subducted toultra-high pressures (Cartigny et al., 2001a, 2004a), which all showpositive average δ15N-values (grey field in Fig. 2). Recent experimentalevidence shows that K-bearing phases are stable to greater depthsthan previously thought (Rapp et al., 2008) implying that surfacenitrogen could be subducted even into the deep mantle. We suggestthat the δ15N of subducted oceanic crust will be positive and thereforethat nitrogen isotopes are a good proxy of subducted material. If anysubducted component with negative average δ15N-exists, it yetremains to be found.

In contrast to subducted ammonium, mantle nitrogen is char-acterised by mostly (N70%) negative δ15N-values as illustrated by dataon fibrous (kimberlite-related) diamonds (e.g. Boyd et al., 1992,1994b; Cartigny et al., 2003), monocrystyalline diamonds withperidotitic and eclogitic inclusions (see Cartigny, 2005 for review),and mid-ocean ridge basalts (e.g. Marty and Zimmermann, 1999). Theoverall average value is −5±4‰. On the basis of co-variations withN-content, positive δ15N values, occurring most frequently amongmonocrystalline diamonds of peridotitic and eclogitic paragenesis, arethought to reflect high-temperature fractionations rather than sourceeffects. Little is known, however, about the N-bearing phases involvedand about nitrogen fractionation factors at mantle temperatures. Forthe latter, theoretical data suggest fractionations factors b1.5‰ (Richetet al., 1977) whilst empirical data suggest a fractionation of ±2‰between reduced/oxidised fluids and diamonds (see Thomassot et al.,2007 and ref. therein).

As illustrated in Fig. 2, peridotitic diamonds from Panda show alarge variation in δ15N which is similar to other diamond mines(Fig. 2), including peridotitic diamonds those from Pipe 50 (China),the coastal placers of Namibia, Udachnaya (Russia) and DeBeers Pool(South Africa) (Cartigny et al., 1997, 1998b, 2004b). The average andmode in δ15N is ~−4‰, similar to the overall mantle-value. Thesenegative δ15N-values are particularly well displayed by sulfide-bearingPanda diamonds which fall without exception below zero (δ15N from−12.4 to −0.5‰).

Panda peridotitic diamonds show δ13C-values in the range of othermines worldwide (δ13C mode between −3.5 and −5.5‰). Althoughδ13C-values below −10‰ typically occur more often among eclogiticdiamonds, both peridotitic and eclogitic diamonds have modes fallingwithin the mainmantle range (Fig. 1; e.g.; Galimov, 1991; Deines et al.,2001; Donnelly et al., 2007; Cartigny 2005 for review) and a negativetail, down to −19.4‰ in the present case. The whole Panda dataset ischaracterised by a very tight δ13C distribution (std dev is 1.3‰), whichoverall is very similar to peridotitic diamonds from A-154 at Diavik(std dev is ±0.7‰, Donnelly et al., 2007). These δ13C-values beingrather ambiguous when attempting to distinguish between mantleand subduction-related origins, the very tight carbon isotopicdistribution observed for Panda and A-154 would be consistent withprecipitation from mantle-derived volatiles.

For the dominant P-type diamonds (silicate and/or sulfide — seeTable 1), nitrogen contents also provide a means to evaluate possiblesubduction components. For Panda, the distribution of NFTIR-contents for all diamonds studied here fall within the upper rangeof diamonds worldwide (showing a Poisson-type distribution,where average~standard deviation, e.g. Cartigny et al. 1997). Theaverage NFTIR-content of 260±250 at.ppm falls within the upperrange of worldwide diamond populations. Worldwide, peridotiticdiamonds have average NFTIR-contents of 123at.ppm at DeBeers Pool,133at.ppm at Akwatia in Ghana, 144at.ppm at Koffiefontein; 330 at.ppm at Premier, 339at.ppm at A-154 and up to 380at.ppm at Pipe 50in China (Deines et al., 1989, 1991; Stachel and Harris, 1997a;Cartigny et al., 1997 and unpub. data; Donnelly et al., 2007). Theabundance of peridotitic Type II diamonds, however, is low at Panda(~15%) compared to kimberlites worldwide, which mostly rangefrom 20 to 40% and up to 55% at De Beers Pool (Harris and Spear,1986; Stachel and Harris, 1997a; Deines et al., 2001; Viljoen, 2002;Cartigny et al., unp. data). An even lower value (2%) is recorded forperidotitic diamonds from A154 (Donnelly et al., 2007), suggestingthat Slave kimberlites carry diamonds that are richer in N thanobserved, e.g., for the Kaapvaal Craton. This nitrogen-rich characteris unlikely to be connected with subduction-related N additionthough, as that would also imply a higher average δ15N-value, whichis not present in the data (see above).

Peridotitic diamonds from Panda (harzburgitic, lherzolitic orundifferentiated peridotitic) display no clear coherent covariationsin δ13C–δ15N–N-contents although a possible trend of decreasing δ13C(from−4‰ to−6‰) with increasing δ15N and decreasing N-contentsmay exist (Fig. 4a). This trend resembles one identified amongdiamonds precipitated via methane oxidation (Thomassot et al.,2007). However, to fit a methane precipitation hypothesis to thepresent data, at least three fluids with distinct initial δ15N–N-contents(Fig. 4b) would be required, and evenwith these prescribed fluids, theδ13C–N fit would be poor (Fig. 4c). Alternatively, poorly correlatedδ13C–N co-variations may be a consequence of precipitation ofdiamonds from oxidised fluids, as suggested by Palot et al. (thisissue) to explain data for diamonds from a single eclogite xenolith (seealso Boyd et al., 1994b; Cartigny et al., 2001b for earlier suggestionsand Stachel et al., this issue, for an alternative view). A furtherpossibility is that the chemical and isotopic variability, in particular ofthe δ15N-values, reflects fluid–rock interaction and differentiationprior to diamond formation (Cartigny et al. 2001b). At present, it is notpossible to conclusively deduce the C-speciation of the peridotiticdiamond precipitating fluid(s), but it appears that, at Panda, similar/identical fluid(s) contributed to lherzolitic, harzburgitic, and undiffer-entiated peridotitic diamonds.

The same discussion applies to eclogitic diamonds and therefore isnot repeated. However, it is worth noting that sample PA-65 exhibitshigh N-contents at low δ13C (Fig. 5) and positive δ15N (Table 1) whichdeviates from the mantle field and the predicted fractionation/differentiation trend of Cartigny et al. (1998a, 2001b). These actuallyfits the values expected for a subduction-related formation. From thisperspective, PA-65 is the only sample compatible with a subductionmodel.

The purpose of using multiple sulfur isotopes for sulfides includedin diamonds lies in the fact that Archean surface derived sulfur mayshow significant deviations (i.e. non-zero Δ33S) from the terrestrial/theoretical fractionation line that is defined, in simplified form, asδ33S~0.52×δ34S (for the full equation, see Section 3). It has beenshown that other relevant processes, including biological and high orlow temperature geological (i.e. abiotic) fractionations, would onlyinduce some δ34S-variability but no significant Δ33S anomalies. So far,the only known mechanism that may produce significant anomalies(i.e. Δ33SN0.2‰) requires UV and photolysis of gaseous S-bearingmolecule(s) (Farquhar et al. 2001; Lyons, in press; Thomassot et al.,2009). In the anoxic atmosphere, which prevailed in the Archean

Fig. 4. δ13C– δ15N–N co-variations among Panda diamonds. Since only the maincompositional ranges are displayed here some samples with extreme δ13C are omitted.The trends presented indicate that even several distinct, reduced fluids evolving duringdiamond crystallisation are unlikely to account for the whole range in δ13C–δ15N–Ncovariations.

Fig. 5. δ13C–N variations among Panda diamonds. The nearly constant δ13C over a largerange of N-contents is quite usual and this was considered as an argument for theincompatible behavior of nitrogen during diamond growth Cartigny et al. (2001b). Notethat one sample (PA-65) falls above the limit sector (straight line) defined by Cartignyet al. (2001b) and cannot be explained using the previously proposed mantlefractionation model. Also showing the most positive δ15N of all Panda samples, thisdiamond might record formation from subduction-related material. Lower mantlederived samples do not appear on the figure owing to their N-deprived character.


before atmospheric oxygen rose at ~2.4 Gy ago (Farquhar et al., 2000,2007; Kasting, 2001), chemical reactions induced byUV-photolysis ledto the production of two S-sinks, one oxidised (sulfate aerosols with

Δ33Sb0 down to several ten's of permill) and another reduced (S0 to S8molecules with Δ33SN0 up to several 10's of per mill) (Farquhar et al.,2001).

The recognition of significant anomalous sulfur isotope composi-tions, i.e. ofΔ33S outside the 2σ ion-probe uncertainty, among eclogiticsulfide inclusions in diamonds from the Orapa (3 out of 12, Farquharet al., 2002) and Jwaneng (11 out of 19, Thomassot et al., in press)mines in Botswana, together with their absence among peridotiticdiamonds (Jwaneng, n=2, Thomassot et al., 2009) indicates thepresence of recycled sedimentary sulfur in sulfide inclusions ineclogitic diamonds. The range of Δ33S-values (from −0.5 to 1.1‰)for eclogitic sulfide inclusions is smaller than observed in Archeansediments (from ~−2 to ~+11‰, see Farquhar et al., 2007) and itremains to be determined whether these values reflect samplingeffects, dilution of the signal by mixing between sulfide andsedimentary sulfate pools or withmantle-sulfur, or are a characteristicof the sulfur pool that is recycled (Thomassot et al., 2009). These sulfurisotope anomalies assume homogenous near-zero values of δ34S andΔ33S for juvenile upper mantle sulfur and such values are inferred onthe basis of cosmochemical (e.g. Hulston and Thode, 1965) andterrestrial data (e.g. Chaussidon et al. 1991; Thomassot et al. 2009).

Westerlund et al. (2006) provided evidence for an Archean age forsulfide inclusions in peridotitic diamonds from Panda. This allows thepossibility that these inclusions may carry a Δ33S-anomaly, if asubduction related fluid carrying recycled, atmosphere derived sulfurwas involved in their formation. Table 1 shows that for Panda a casefor anomalous Δ33S cannot be made for peridotitic diamonds sinceΔ33S-values, at the 2 sigma analytical uncertainty, are not significantlydifferent from zero. The absence of a clear Δ33S anomaly does notprovide definitive evidence that sedimentary subducted sulfur isabsent (it could be rehomogenised), but it is worth noting that therange in δ34S falls within the range of upper mantle samples and issmall compared to eclogitic diamonds (e.g. Chaussidon et al., 1987)and Archean sedimentary sulfur (Farquhar et al. 2007 and ref. therein)(see Fig. 6). Os-mobility via C-bearing subduction-related fluids, assuggested by Westerlund et al. (2006), implies an associated mobilityfor sulfur and nitrogen and thus, associated signatures for subductionderived, (meta-)sedimentary-related carbon, nitrogen and sulfur. C-,N-, and S-isotope systematics, however, are not suggestive of asurficial origin. It is worth noting that the sample PA-84 (characterised

Fig. 6. (a) δ34S–Δ33S and (b) δ15N–Δ33S diagrams. Filled triangles and filled diamonds correspond to peridotitic and eclogitic (sulphide-bearing) diamonds, respectively. Previouslypublished data (open diamonds) correspond to eclogitic diamonds from Orapa and Jwaneng, both in Botswana (Farquhar et al., 2002; Thomassot et al., 2009 respectively). Note thatonly two samples fall within the Archean subduction field.


by the greatest deviation from Δ33S=0) has the lowest δ15N of−12‰, strongly suggesting that the sample is mantle-related (Fig. 6).

In summary, peridotitic diamonds from Panda with silicate and/orsulfide inclusions show a series of mantle-like features, includingδ13C-, δ15N-, δ34S- and Δ33S-values, as expected of diamond formationsolely frommantle volatiles, implying that most, if not all, the isotopicvariability is related to fluid differentiation/evolution e.g. throughwall–rock interactions. In the absence of significant δ13C–variability,we do not favor the role of primordial heterogeneity of carbon(e.g. Deines et al., 1989, 1991). The absence of significant Δ33S-anomalies and the occurrence of average negative δ15N-values suggestthat the ranges in both δ34S and δ15N are more likely the result ofprocesses occurring in the mantle, before or during diamondprecipitation, rather than a strong influence of Archean sedimentsubduction.

Given that no correlation between δ34S and δ15N exists, duringformation of diamonds through metasomatic processes some decou-pling might occur between diamond and its inclusion(s) and between

Fig. 7. diagram of %B versus NFTIR-content for diamonds from (a) Panda (Westerlund et al., thidiamonds from the Central Slave are largely restricted to low-temperature (shallower) dom

S and C (Thomassot et al., 2009). Given the requirement for S-mobility(i.e. sulfur uptake via metasomatism) to carry Os (Westerlund et al.,2006), we conclude that if some sulfides pre-existed, either these hadΔ33S~0‰ or the sulfur budget in peridotitic diamonds from Panda hadbeen buffered by metasomatic sulfur.

For many years, an over-abundance of sulfide-bearing diamondsreaching up to 60% of inclusion-bearing diamonds in single kimber-lites (Harris, 1992) has been noted. It led several authors (e.g. Meyer,1987; Deines et al., 1987) to classify sulfide-bearing diamonds as aseparate diamond paragenesis, usually referred to as S-type. Thissubdivision has not received much attention, most-likely becausesulfide-bearing diamonds may be of both peridotitic and eclogiticparageneses, as determined on the basis of co-existing silicateinclusions or other chemical criteria (see above). However, recentevidence based on a larger dataset shows that eclogitic sulfide-bearingdiamonds from southern African kimberlites (Koffiefontein, De BeersPool and Jwaneng) have distinct ranges and averages in δ13C, δ15N, N-contents and aggregation states compared to eclogitic silicate-bearing

s study) and (b) A-154 (Donnelly et al., 2007) illustrating that sulfide-bearing peridotiticains of the continental lithosphere.


diamonds from the same kimberlite (Thomassot et al., 2008, 2009). Indetail, sulfide-bearing diamonds show narrower ranges in δ13C andδ15N, both of which are centered around mantle-values, high N-contents (usually N350 ppm, almost no Type II) and show littleevidence for advanced N-aggregation (usually b50% IaA diamonds)(Deines et al., 1987; Thomassot et al., 2008, 2009). The reason whysulfide-bearing diamonds are distinct remains unclear, but their lownitrogen aggregation states suggest residence in the lithosphericmantle at lower temperatures and/or for shorter times than theirsilicate-bearing counterparts.

An important question to consider is whether the conclusioninferred from eclogitic sulfide-bearing diamonds applies to theperidotitic paragenesis and also whether these characteristics arefound in other cratonic areas. It is not possible to test whetherperidotitic sulfide-bearing diamonds from southern African kimber-lites are distinct from their silicate bearing counterparts becausediamonds with peridotitic sulfides are exceedingly rare on theKaapvaal craton (Shirey et al., 2003; Richardson et al., 2004;Thomassot et al., 2008, 2009). However, for two of the Slavekimberlites (Panda and A-154), significant numbers of peridotiticsulfide- or silicate-bearing diamonds have been analysed (Westerlundet al., 2006 and this study, Donnelly et al., 2007, respectively) allowingsuch an assessment to be made for the Central Slave Craton.

Fig. 7 shows that at both Panda and A-154 pipe peridotitic sulfide-bearing diamonds have more restricted ranges in both N-contents andN-aggregation state relative to peridotitic silicate-bearing diamonds.As noted earlier (Westerlund et al., 2006), peridotitic diamonds fromPanda with sulfide inclusions have higher N-contents and loweraggregation states than their peridotitic silicate bearing counterparts(Fig. 7a). The same conclusionmay be drawn from the A-154 diamonddata shown in Fig. 7b (Donnelly et al., 2007). These observationssupport the conclusion based on eclogitic diamonds from southernAfrican kimberlites. From a stable isotope perspective, δ15N-valuesfor sulfide included diamonds show a narrower range (from −12.4to −0.5‰) than silicate-bearing diamonds (from −17.0 to +8.5‰).This observation is compatible with the concept of diamonds withhigh N-contents generally showing mantle like C and N-isotope ratios(Cartigny et al. 2001b). Because of a similar small range in δ13Cbetween sulfide and silicate-bearing peridotitic diamonds from bothPanda and A-154, a distinction based on carbon isotopes is howevernot obvious. Fig. 7 also suggests that for the limited set of eclogiticdiamonds from Panda and Diavik, a similar difference in nitrogencharacteristics may exist between sulfide and silicate includedsamples. The differences between sulfide-bearing and silicate-bearingdiamonds, therefore, appear significant on a broader scale and notdependent on source paragenesis. If the low N-aggregation states ofthe sulfide-bearing diamonds derived from the coolest (shallowest)part of diamond stable lithosphere whereas silicate-bearing diamondsextended to deeper parts (Westerlund et al., 2006; Thomassot et al.,2009), then one can somewhat predict that subcratonic lithosphericsulfides should become scarce with increasing depths. Evidence for anoverall decreasing degree of depletion with depth within the Slavelithospheric mantle (Griffin et al., 1999) is, however, inconsistent withsuch an interpretation since the less depleted deeper lithosphereshould be a priori richer in volatile elements and hence more likely tocontain sulfides. Furthermore, the study of peridotitic sulfides inmantle xenocrysts from Diavik (Aulbach et al., 2004) shows these tooriginate from depths exceeding the restricted interval from whichsulfide-bearing diamonds appear to derive.

Although at this stage the reason(s) why sulfide and silicate-bearing diamonds are distinct is unclear, sulfide-bearing diamondsappear to sample a more restricted range of P–T conditions thansilicate-bearing diamonds, and this suggests that diamond populationcharacteristics defined on the basis of studies of sulfide-bearingdiamonds are likely biased. By extension, the diamond genesis age(s)defined from sulfide-bearing diamond are may not extend to the

entire population (Thomassot et al., 2009). It is worth noting that atLac de Gras, available data on sulfides in olivines (Aulbach et al., 2004)that come from a large depth range match however quite well thesulfide in diamond ages (Westerlund et al., 2006).

6. Conclusions

To constrain thepossible involvementof subduction-relatedmaterialduring the growth of peridotitic diamonds from Panda (Ekati Mine,Slave Craton, Canada), we applied two powerful proxies for subduction-relatedmaterial in the Earth'smantle (namely δ15Nand, for theArchean,Δ33S) and combined them with additional compositional and isotopicdata. For the 90 samples examined in this study,we found evidence for asubduction-related in only one (eclogitic) diamond (PA 65). Data forδ13C, δ15N, δ34S, Δ33S and N-contents of peridotitic diamonds areentirely consistent with an origin from mantle volatiles.

We also show that peridotitic sulfide- and silicate-bearing diamondsare distinct (with the former showing more restricted ranges in δ15Nand N content and aggregation state), extending previous evidenceestablished for eclogitic diamonds from the Kaapvaal to the Slave Cratonand to the peridotitic paragenesis. These differences thus are likely aworldwide feature.

For representative sampling of diamond source regions, we suggestthat future diamond studies should aim to mirror the sulfide/silicateratio derived from inclusions.

Acknowledgments

We are grateful to the Diamond Trade Company (DTC) for diamonddonation. We wish to acknowledge financial support from CNRS(through the Dyeti program) and a donation from De Beers to supportthe maintenance of our experimental equipments. JF acknowledgessupport of an NSF Career grant (0348382 ) and the GuggenheimFoundation. Andreas Pack and an anonymous reviewer are thanked fortheir constructive and supportive reviews. IPGP contribution 2531.

References

Aulbach, S., Griffin, W.L., Pearson, N.J., O'Reilly, S.Y., Kivi, K., Doyle, B.J., 2004. Mantleformation and evolution, Slave Craton: constraints from HSE abundances and Re–Os isotope systematics of sulfide inclusions in mantle xenocrysts. Chemical Geology208, 61–88.

Aulbach, S., et al., 2009, this issue. Sulphide survival and diamond genesis duringformation and evolution of Archaean subcontinental lithosphere: A comparisonbetween the Slave and Kaapvaal cratons. Proceedings of the 9th InternationalKimberlite Conference. Lithos 112S, 747–757.

Beaumont,V., Robert, F.,1999.Nitrogen isotope ratios of kerogens inPrecambrian cherts: arecord of the evolution of atmosphere chemistry? Precambrian Research 96, 63–82.

Bebout, G.E., Fogel, M.L., 1992. Nitrogen-isotope compositions of metasedimentaryrocks in the Catalina Schist, California: Implications for metamorphic devolatiliza-tion history. Geochimica et Cosmochimica Acta 56, 2839–2849.

Boyd, F.R., Gurney, J.J., 1986. Diamonds and the African lithosphere. Science 232,472–477.

Boyd, S.R., Pillinger, C.T., Milledge, H.J., Mendelssohn, M.J., Seal, M., 1992. C and Nisotopic composition and the infrared absorption spectra of coated diamonds:evidence for the regional uniformity of CO2–H2O rich fluids in lithospheric mantle.Earth and Planetary Science Letters 109, 633–644.

Boyd, S.R., Kiflawi, I., Woods, G.S., 1994a. The relationship between infrared absorptionand the A defect concentration in diamond. PhilosophicalMagazine. B 69,1149–1153.

Boyd, S.R., Pineau, F., Javoy, M., 1994b. Modelling the growth of natural diamonds.Chemical Geology 116, 29–42.

Boyd, S.R., Kiflawi, I., Woods, G.S., 1995. Infrared absorption by the B nitrogen aggregatein diamond. Philosophical Magazine. B 72, 351–361.

Brey, G.P., Bulatov, V., Girnis, A., Harris, J.W., Stachel, T., 2004. Ferropericlase — a lowermantle phase in the upper mantle. Lithos 77, 655–663.

Bulanova, G.P., Pearson, D.G., Hauri, E.H., Griffin, B.J., 2002. Carbon and nitrogen isotopesystematics within a sector-growth diamond from the Mir kimberlite, Yakutia.Chemical Geology 188, 105–123.

Burgess, S.R., Harte, B., 2004. Tracing lithosphere evolution through the analysis ofheterogeneous G9–G10 garnets in peridotite xenoliths, II: REE chemistry. Journal ofPetrology 45, 609–634.

Busigny, V., Cartigny, P., Philippot, P., Ader, M., Javoy, M., 2003. Massive recycling ofnitrogen and other fluid-mobile elements (K, Rb, Cs, H) in a cold slab environment:evidence from HP to UHP oceanic metasediments of the Schistes Lustres nappe(western Alps, Europe). Earth and Planetary Science Letters 215, 27–42.


Busigny, V., Ader, M., Cartigny, P., 2005a. Quantification and isotopic analysis of nitrogenin rocks at the ppm level using sealed tube combustion technique: a prelude to thestudy of altered oceanic crust. Chemical Geology 223, 249–258.

Busigny, V., Laverne, C., Bonifacie, M., 2005b. Nitrogen content and isotopic com-position of oceanic crust at a superfast spreading ridge: a profile in altered basaltsfrom ODP Site 1256, Leg 206. Geochemistry Geophysics Geosystems 6. article n°Q12O01.

Cartigny, P., 2005. Stable isotopes and the origin of diamond. Elements 1, 79–84.Cartigny, P., Ader, M., 2003. A comment on “The nitrogen record of crust–mantle

interaction and mantle convection from Archean to Present” by B. Marty and N.Dauphas [Earth Planet. Sci. Lett. 206 (2003) 397–410] — Discussion. Earth andPlanetary Science Letters 216, 425–432.

Cartigny, P., Harris, J.W., Javoy, M., 1998a. Formation of eclogitic diamonds at Jwaneng:no room for a recycled component. Science 280, 1421–1423.

Cartigny, P., Boyd, S.R., Harris, J.W., Javoy, M., 1997. Nitrogen isotopes in peridotiticdiamonds from Fuxian, China : The mantle signature. Terra Nova 9, 175–179.

Cartigny, P., Harris, J.W., Phillips, D., Girard, M., Javoy, M., 1998b. Subduction-relateddiamonds ? — the evidence for a mantle-derived origin from coupled δ13C–δ15Ndeterminations. Chemical Geology 147, 147–159.

Cartigny, P., De Corte, K., Shatsky, V.S., et al., 2001a. The origin and formation ofmetamorphic microdiamonds from the Kokchetav massif, Kazakhstan: a nitrogenand carbon isotopic study. Chemical Geology 176, 265–281.

Cartigny, P., Harris, J.W., Javoy, M., 2001b. Diamond genesis, mantle fractionations andmantle nitrogen content: a study of δ13C–N concentrations in diamonds. Earth andPlanetary Science Letters 185, 85–98.

Cartigny, P., Harris, J.W., Taylor, A., Davies, R., Javoy, M., 2003. On the possibility of akinetic fractionation of nitrogen stable isotopes during natural diamond growth.Geochimica et Cosmochimica Acta 67, 1571–1576.

Cartigny, P., Chinn, I., Viljoen, K.S., Robinson, D., 2004a. Early Proterozoic ultrahighpressure metamorphism: evidence from microdiamonds. Science 304, 853–855.

Cartigny, P., Stachel, T., Harris, J.W., Javoy, M., 2004b. Constraining diamondmetasomaticgrowth using C- and N-stable isotopes: examples fromNamibia. Lithos 77, 359–373.

Chaussidon, M., Albarède, F., Sheppard, S.M.F., 1987. Sulphur isotope heterogeneity inthemantle from ionmicroprobemeasurements of sulphide inclusions in diamonds.Nature 330, 242–244.

Chaussidon, M., Sheppard, S.M.F., Michard, A., 1991. Hydrogen, sulphur and neodymiumisotope variations in the mantle beneath the EPR at 12°50′N. In: Tayor, H.P., O'Neill,J.R., Kaplan, I.R. (Eds.), Stable isotope geochemistry: a tribute to Samuel Epstein.InGeochemical Society, St Louis, pp. 325–337.

Creaser, R.A., Grutter, H., Carlson, J., Crawford, B., 2004. Macrocrystal phlogopite Rb–Srdates for the Ekati property kimberlites, Slave Province, Canada: evidence formultiple intrusive episodes in the Paleocene and Eocene. Lithos 76, 399–414.

Davies, R.M., Griffin, W.L., Pearson, N.J., Andrew, A.S., Doyle, B.J., O'Reilly, S.Y., 1999. In:Gurney, J.J., Gurney, J.L., Pascoe, M.D., Richardson, S.H. (Eds.), Diamonds from thedeep: Pipe DO-27, Slave craton, Canada. In: Proceedings of the 7th InternationalKimberlite Conference, vol. 1. Red Roof Design, Cape Town, pp. 148–155.

Davies, R.M., Griffin,W.L., O'Reilly, S.Y., Doyle, B.J., 2004.Mineral inclusions and geochemicalcharacteristics of microdiamonds from the DO27, A154, A21, A418, DO18, DD17 andRanch Lake kimberlites at Lac de Gras, Slave Craton, Canada. Lithos 77, 39–55.

Davis, W.J., Jones, A.G., Bleeker, W., Grutter, H., 2003. Lithosphere development in theSlave craton: a linked crustal and mantle perspective. Lithos 71, 575–589.

Deines, P., 1968. The carbon and oxygen isotopic composition of carbonates from amicaperidotite dike near Dixonville, Pennsylvania. Geochimica et Cosmochimica Acta32, 613–625.

Deines, P., Harris, J.W., 2004. New insights into the occurrence of C13-depleted carbon inthe mantle from two closely associated kimberlites: Letlhakane and Orapa,Botswana. Lithos 77, 125–142.

Deines, P., Harris, J.W., Gurney, J.J., 1987. Carbon isotopic composition, nitrogen contentand inclusion composition of diamonds from the Roberts Victor kimberlite, SouthAfrica: evidence for a 13C depletion in the mantle. Geochimica et CosmochimicaActa 51, 1227–1243.

Deines, P., Harris, J.W., Spear, P.M., Gurney, J.J., 1989. Nitrogen and 13C content of Finschand Premier diamonds and their implications. Geochimica et Cosmochimica Acta53, 1367–1378.

Deines, P.,Harris, J.W.,Gurney, J.J.,1991. Thecarbon isotopic compositionandnitrogencontentof lithospheric and asthenospheric diamonds from the Jagersfontein and Koffiefonteinkimberlites, South Africa. Geochimica et Cosmochimica Acta 55, 2615–2626.

Deines, P., Viljoen, F., Harris, J.W., 2001. Implications of the carbon isotope and mineralinclusion record for the formation of diamonds in the mantle underlying a mobilebelt: Venetia, South Africa. Geochimica et Cosmochimica Acta 65, 813–838.

Dixon, J.E., Leist, L., Langmuir, C., Schilling, J.G., 2002. Recycled dehydrated lithosphereobserved in plume-influenced mid-ocean-ridge basalt. Nature 420, 385–389.

Donnelly, C.L., Stachel, T., Creighton, S., Muehlenbachs, K., Whiteford, S., 2007.Diamonds and their mineral inclusions from the A154 South pipe, Diavik DiamondMine, Northwest Territories, Canada. Lithos 98, 160–176.

Farquhar, J., Bao, H.M., Thiemens, M., 2000. Atmospheric influence of Earth's earliestsulfur cycle. Science 289, 756–758.

Farquhar, J., Savarino, J., Airieau, S., Thiemens, M.H., 2001. Observation of wavelength-sensitivemass-independent sulfur isotope effects during SO2 photolysis: implicationsfor the early atmosphere. Journal of Geophysical Research-Planets 106, 32829–32839.

Farquhar, J., Wing, B.A., McKeegan, K.D., Harris, J.W., Cartigny, P., Thiemens, M.H., 2002.Mass-independent sulfur of inclusions in diamond and sulfur recycling on earlyearth. Science 298, 2369–2372.

Farquhar, J., Peters, M., Johnston, D.T., Strauss, H., Masterson, A., Wiechert, U., Kaufman,A.J., 2007. Isotopic evidence for Mesoarchaean anoxia and changing atmosphericsulphur chemistry. Nature 449, 706–711.

Fogel, M.L., Cifuentes, L.A., 1993. Isotope fractionation during primary production. In:Engel,M.H.,Macko, S.A. (Eds.), Organic Geochemistry. Plenum,NewYork, pp. 73–98.

Galimov, E.M., 1991. Isotope fractionation related to kimberlite magmatism anddiamond formation. Geochimica et Cosmochimica Acta 55, 1697–1708.

Gao, S., Rudnick, R.L., Carlson, R.W., McDonough,W.F., Liu, Y.S., 2002. Re–Os evidence forreplacement of ancient mantle lithosphere beneath the North China craton. Earthand Planetary Science Letters 198, 307–322.

Garvin, J., Buick, R., Anbar, A.D., Arnold, G.L., Kaufman, A.J., 2009. Isotopic evidence foran aerobic nitrogen cycle in the latest Archean. Science 323, 1045–1048.

Griffin,W.L.,Doyle, B.J., Ryan, C.G., et al.,1999. Layeredmantle lithosphere in the LacdeGrasarea, Slave Craton: composition, structure and origin. Journal of Petrology40, 705–727.

Griffin, W.L., et al., 2009, this issue. A translithospheric suture in the vanished 1-Galithospheric root of South India: Evidence from contrasting lithosphere sections inthe Dharwar Craton. Proceedings of the 9th International Kimberlite Conference.Lithos 112S, 1109–1119.

Gurney, J.J., 1989. Diamonds. In: Ross, J. (Ed.), Kimberlites and Related Rocks: Vol. 2.Their Mantle/crust Settings, Diamonds and Diamond Exploration. GeologicalSociety of Australia Inc, Perth, pp. 935–965.

Gurney, J.J., Hildebrand, P.R., Carlson, J.A., Fedortchouk, Y., Dyck, D.R., 2004. Themorphological characteristics of diamonds from the Ekati property, NorthwestTerritories, Canada. Lithos 77, 21–38.

Haggerty, S.E.,1986. Diamond genesis in amultiply-constrainedmodel. Nature 320, 34–38.Halama, R., Bebout, G., John, T., Schenk, V., 2008. Recycling of nitrogen during

subduction of oceanic crust: insights from high-pressure and ultra-high pressuremetamorphic rocks: Eos Trans. AGU, vol. 89 (53).

Harris, J.W., 1992. Diamond geology. In: Field, J.E. (Ed.), The Properties of Natural andSynthetic Diamond. InAcademic Press, pp. 345–393.

Harris, J.W., Spear, P.M., 1986. Systematic studies of nitrogen in diamonds from knownsources. In: Ross, J., et al. (Ed.), Extended Abstract of 4th International KimberliteConference Geological Society of Australia: Geological Society of Australia, SpecialPublications, Perth, Australia, pp. 398–400.

Harte, B., Harris, J.W., Hutchison, M.T., Watt, G.R., Wilding, M.C., 1999. Lower 907mantlemineral associations in diamonds from Sao Luis, Brazil. In: Fei, Y., Bertka, C.M.,Mysen, B.O. (Eds.), Mantle Petrology: Field Observations and High PressureExperimentation. A Tribute to Francis R. (Joe) Boyd: The Geochemical SocietySpecial Publication no. 6, The Geochemical Society, Houston, TX, pp. 125–154.

Herzberg, C., 1999. Phase equilibrium constraints on the formation of cratonic mantle. In:Fei, Y., Bertka, C.M., Mysen, B.O. (Eds.), Mantle Petrology: Field Observations and HighPressure Experimentation. A Tribute to Francis R. (Joe) Boyd, TheGeochemical SocietySpecial Publication n°6, The Geochemical Society, Houston, TX, pp. 241–257.

Helmstaedt, H., Schulze, D.J., 1989. Southern African kimberlites and their mantlesample— implication for Archean tectonics and lithosphere evolution. In: Ross, J., etal. (Ed.), Kimberlites and Related Rocks: Their Composition, Occurrence, Origin andEmplacement. Geological Society of Australia Inc, Perth, pp. 358–368.

Honma, H., Itihara, Y., 1981. Distribution of ammonium in minerals of metamorphic andgranitic rocks. Geochimica et Cosmochimica Acta 45, 983–988.

Hulston, J.R., Thode, H.G., 1965. Variations in the 33S, 34S, and 36S contents of meteoritesand their relation to chemical and nuclear effects. Journal of Geophysical Research70, 3475–3484.

Jacob, D.E., et al., 2009, this issue. Eclogite xenoliths from Kimberley, South Africa — Acase study of mantle metasomatism in eclogites. Proceedings of the 9thInternational Kimberlite Conference. Lithos 112S, 1002–1013.

Kaminsky, F.V., Zakharchenko, O.D., Davies, R., Griffin, W.L., Khachatryan-Blinova, G.K.,Shiryaev, A.A., 2001. Superdeep diamonds from the Juina area, Mato Grosso State,Brazil. Contributions to Mineralogy and Petrology 140, 734–753.

Kasting, J.F., 2001. The rise of atmospheric oxygen. Science 293, 819–820.Kelemen, P.B., Hart, S.R., Bernstein, S., 1998. Silica enrichment in the continental upper

mantle via melt/rock reaction. Earth and Planetary Science Letters 164, 387–406.Kesson, S.E., Ringwood, A.E., 1989. Slab–mantle interactions. 2. The formation of

diamonds. Chemical Geology 78, 97–118.Li, L., Bebout, G.E., Idleman, B.D., 2007. Nitrogen concentration and delta 15N of altered

oceanic crust obtained on ODP Legs 129 and 185: insights into alteration-relatednitrogen enrichment and the nitrogen subduction budget. Geochimica etCosmochimica Acta 71, 2344–2360.

Liu, Y., et al., 2009, this issue. Metasomatic origin of diamonds in the world's largestdiamondiferous eclogite. Proceedings of the 9th International Kimberlite Con-ference. Lithos 112S, 1014–1024.

Lyons, J.R., in press. Atmospherically-derived mass-independent sulfur isotopesignatures, and incorporation into sediments. Chemical Geology. doi:10.1016/j.chemgeo.2009.03.027.

Marty, B., Dauphas, N., 2003. The nitrogen record of crust–mantle interaction andmantle convection from Archean to Present. Earth and Planetary Science Letters206, 397–410.

Marty, B., Zimmermann, L., 1999. Volatiles (He, C, N, Ar) in mid-ocean ridge basalts:assessment of shallow-level fractionation and characterisation of source composi-tion. Geochimica et Cosmochimica Acta 63, 3619–3633.

Menzies, A., Grütter, H., Gurney, J.J., Carlson, J., Fung, A., Nowicki, T., 2004. Peridotiticmantle xenoliths from kimberlites on the Ekati Diamond Mine property, NWT,Canada: major element compositions and implications for the lithosphere beneaththe central Slave craton. Lithos 77, 395–412.

Meyer, H.O.A., 1987. Inclusions in diamond. In: Nixon, P.H. (Ed.), Mantle Xenoliths. JohnWiley & Sons Ltd, New York, pp. 501–522.

Moore, M., 1985. Diamond morphology. Industrial Diamond Review 45, 67–71.Palot, M., et al., 2009, this issue. Diamond origin and genesis: A C and N stable isotope

study on diamonds from a single eclogitic xenolith (Kaalvallei, South Africa).Proceedings of the 9th International Kimberlite Conference. Lithos 112S, 758–766.


Pearson, D.G., Wittig, N., 2008. Formation of Archaean continental lithosphere and itsdiamonds: the root of the problem. Journal of the Geological Society 165, 895–914.

Pearson, D.G., Carlson, R.W., Shirey, S.B., Boyd, F.R., Nixon, P.H., 1995a. Stabilisation ofArchean lithospheric mantle: a Re–Os isotope study of peridotite xenoliths from theKaapvaal craton. Earth and Planetary Science Letters 134, 341–357.

Pearson, D.G., Shirey, S.B., Carlson, R.W., Boyd, F.R., Pokhilenko, N.P., Shimizu, N., 1995b.Re–Os, Sm–Nd, and Rb–Sr isotope evidence for thick Archean lithospheric mantlebeneath the Siberian craton modified by multistage metasomatism. Geochimica etCosmochimica Acta 59, 959–977.

Pearson, D.G., Shirey, S.B., Bulanova, G.P., Carlson, R.W., Milledge, H.J., 1999. Dating andparagenetic distinction of diamonds using the Re–Os isotope system: application tosome Siberian diamonds. In: Gurney, J.J., Gurney, J.L., Pascoe, M.D., Richardson, S.H.(Eds.), Proceedings of the 7th International Kimberlite Conference, vol. 2. Red RoofDesign, Cape Town, pp. 637–643.

Peters, K.E., Sweeney, R.E., Kaplan, I.R., 1976. Correlation of carbon and nitrogen stableisotope ratios in sedimentary organic-matter. Limnology and Oceanography 23,598–604.

Philippot, P., Busigny, V., Scambelluri, M., Cartigny, P., 2007. Oxygen and nitrogenisotopes as tracers of fluid activities in serpentinites and metasediments duringsubduction. Mineralogy and Petrology 91, 11–24.

Pinti, D.L., Hashizume, K., Matsuda, J.-I., 2001. Nitrogen and argon signatures in 3.8 to2.8 Ga metasediments: clues on the chemical state of the Archean ocean and thedeep biosphere. Geochimica et Cosmochimica Acta 65, 2301–2315.

Rapp, R.P., Irifune, T., Shimizu, N., Nishiyama, N., Norman,M.D., Inoue, J., 2008. Subductionrecycling of continental sediments and the origin of geochemically enriched reservoirsin the deep mantle. Earth and Planetary Science Letters 271, 14–23.

Richardson, S.H., Shirey, S.B., Harris, J.W., 2004. Episodic diamond genesis at Jwaneng,Botswana, and implications for Kaapvaal craton evolution. Lithos 77, 143–154.

Richet, P., Bottinga, Y., Javoy, M., 1977. A review of hydrogen, carbon, nitrogen, oxygen,sulphur and chlorine stable isotope fractionation among gaseousmolecules. AnnualReview of Earth and Planetary Sciences 5, 65–110.

Schmidberger, S.S., Heaman, L.M., Simonetti, A., Creaser, R.A., Cookenboo, H.O., 2005.Formation of Paleoproterozoic eclogitic mantle, Slave Province (Canada): insightsfrom in-situ Hf and U–Pb isotopic analyses of mantle zircons. Earth and PlanetaryScience Letters 240, 621–633.

Schmidberger, S.S., Simonetti, A., Heaman, L.M., Creaser, R.A., Whiteford, S., 2007. Lu–Hf,in-situ Sr and Pb isotope and trace element systematics for mantle eclogites fromthe Diavik diamond mine: evidence for Paleoproterozoic subduction beneath theSlave craton, Canada. Earth and Planetary Science Letters 254, 55–68.

Schulze, D.J., 1986. Calcium anomalies in the mantle and a subducted metaserpentiniteorigin for diamonds. Nature 319, 483–485.

Shirey, S.B., Harris, J.W., Richardson, S.H., et al., 2003. Regional patterns in theparagenesis and age of inclusions in diamond, diamond composition, and thelithospheric seismic structure of Southern Africa: Lithos A Tale of Two Cratons: TheSlave-Kaapvaal Workshop, vol. 71, pp. 243–258.

Simon, N.S.C., Carlson, R.W., Pearson, D.G., Davies, G.R., 2007. The origin and evolution ofthe Kaapvaal cratonic lithospheric mantle. Journal of Petrology 48, 589–625.

Spetsius, Z.V., et al., 2009, this issue. Combined C isotope and geochemical evidence for arecycled origin for diamondiferous eclogite xenoliths from kimberlites of Yakutia.Proceedings of the 9th International Kimberlite Conference. Lithos 112S,1032–1042.

Stachel, T., Harris, J.W., 1997a. Syngenetic inclusions in diamond from the Birim field(Ghana)— a deep peridotitic profile with a history of depletion and re-enrichment.Contributions to Mineralogy and Petrology 127, 336–352.

Stachel, T., Harris, J.W., 1997b. Diamond precipitation and mantle metasomatism —

evidence from the trace element chemistry of silicate inclusions in diamonds fromAkwatia, Ghana. Contributions to Mineralogy and Petrology 129, 143–154.

Stachel, T., Harris, J.W., 2008. The origin of cratonic diamonds— constraints frommineralinclusions. Ore Geology Reviews 34, 5–32.

Stachel, T., Viljoen, K.S., Brey, G., Harris, J.W., 1998a. Metasomatic processes in lherzoliticand harzburgitic domains of diamondiferous lithospheric mantle: REE in garnetsfrom xenoliths and inclusions in diamonds. Earth and Planetary Science Letters 159,1–12.

Stachel, T., Harris, J.W., Brey, G.P.,1998b. Rare and unusual mineral inclusions in diamondsfrom Mwadui, Tanzania. Contributions to Mineralogy and Petrology 132, 34–47.

Stachel, T., Harris, J.W., Aulbach, S., Deines, P., 2002. Kankan diamonds (Guinea) III: δ13Cand nitrogen characteristics of deep diamonds. Contributions to Mineralogy andPetrology 142, 465–475.

Stachel, T., Harris, J.W., Tappert, R., Brey, G.P., 2003. Peridotitic diamonds from the Slaveand the Kaapvaal cratons— similarities and differences based on a preliminary dataset. Lithos 71, 489–503.

Stachel, T., Viljoen, K.S., McDade, P., Harris, J.W., 2004. Diamondiferous lithosphericroots along the western margin of the Kalahari Craton — the peridotitic inclusionsuite in diamonds from Orapa and Jwaneng. Contributions to Mineralogy andPetrology 147, 32–47.

Stachel, T., et al., 2009, this issue. Sources of carbon in inclusion bearing diamonds.Proceedings of the 9th International Kimberlite Conference. Lithos 112S, 625–637.

Tappert, R., Stachel, T., Harris, J.W., Shimizu, N., Brey, G.P., 2005. Mineral inclusions indiamonds from the Panda kimberlite, Slave Province, Canada. European Journal ofMineralogy 17, 423–440.

Thomassot, E., Cartigny, P., Harris, J.W., Viljoen, K.S., 2007. Methane-related diamondcrystallization in the Earth'smantle. Earth and Planetary Science Letters 257, 362–371.

Thomassot, E., Cartigny, P., Harris, J.W., 2008. Sulfide-bearing diamonds: the exception,not the rule. Geochimica et Cosmochimica Acta 70, A648.

Thomassot, E., Cartigny, P., Harris, J.W., Lorand, J.-P., J., F., M., C., 2009. Metasomaticdiamond growth and its implications on diamond dating: a multi-isotope study(13C, 15N, 33S, 34S) of sulfide bearing diamonds from Jwaneng (Bostwana). Earth andPlanetary Science Letters 282, 79–90.

Thomazo, C., Pinti, D.L., Busigny, V., Ader, M., Hashizume, K., 2009. Biological activityand Earth's surface evolutions: Insights from carbon, sulfur, nitrogen and ironstable isotopes in the rock record. Comptes Rendus Palevol., in press.

Viljoen, K.S., 2002. An infrared investigation of inclusion-bearing diamonds from theVenetia kimberlite, Northern Province, South Africa: implications for diamonds fromcraton-margin settings. Contributions to Mineralogy and Petrology 144, 98–108.

Walker, R.J., Carlson, R.W., Shirey, S.B., Boyd, F.R., 1989. Os, Sr, Nd, and Pb isotopesystematics of southern african peridotite xenoliths - implications for the chemicalevolution of subcontinental mantle. Geochimica et Cosmochimica Acta 53,1583–1595.

Walter, M.J., 1999. Melting residues of fertile peridotite and the origin of cratoniclithosphere. In: Fei, Y., Bertka, C.M., Mysen, B.O. (Eds.), Mantle Petrology: FieldObservations and High Pressure Experimentation: A Tribute to Francis R. (Joe)Boyd. Special Publication. InThe Geochemical Society, Houston, pp. 225–239.

Westerlund, K.J., et al., 2006. A subduction wedge origin for Paleoarchean peridotiticdiamonds and harzburgites from the Panda kimberlite, Slave craton: evidence fromRe–Os isotope systematics. Contributions to Mineralogy and Petrology 152, 275–294.

Yefimova, E.S., Sobolev, N.V., Pospelova, L.N., 1983. Sulfide inclusions in diamond andspecific features of their paragenesis. Zapiski Vsesroyuznogo MineralogischeskogoObshchestva 112, 300–310 (in Russian).

Zedgenizov, D.A., Harte, B., 2004. Microscale variations of δ13C and N content within anatural diamond with mixed-habit growth. Chemical Geology 205, 169–175.

Zheng, J.P., O'Reilly, S.Y., Griffin, W.L., Lu, F.X., Zhang, M., 1998. Nature and evolution ofCenozoic lithospheric mantle beneath Shandong peninsula, Sino-Korean craton,eastern China. International Geology Review 40, 471–499.

a mantle origin for paleoarchean peridotitic diamonds...

Documents