Journal of Conference Abstracts

Session E04:5A

E04 : 5A/01 : F1

Sources of Diamond Carbon

Oded Navon (oded@vms.huji.ac.il)

Institute of Earth Sciences, The Hebrew University, Jerusalem, 91904, Israel

Diamonds are commonly divided into two main parageneses, perodotitic and eclogitic, based on their mineral inclusions. This division reflects the two main diamond-bearing rock types in the mantle. Here I suggest that a different division should be used in order to distinguish the different sources of carbon for diamond formation. Such distinction may be based on the carbon isotopic composition of the diamonds.

It is well known that the two parageneses produce different histograms of carbon isotopic composition. Diamonds of the peridotitic suite produce a narrow peak between ¹³C values of -1 and -10, with maximum at -5 (all  values are in per mille deviation from PDB carbon or atmospheric nitrogen). Eclogitic diamonds span a much wider range, from +3 to -35, with maxima at -5, -11 and about -20. The main peak of the histogram at -5 overlaps with that of peridotitic diamonds.

It is suggested that eclogitic and peridotitic diamonds that form the peak around ¹³C =-5 share the same source of carbon, designated "mantle carbon". Those with more negative ¹³C values are dominated by a different source of carbon, suggested here to be subducted organic carbon. Diamonds formed from this second source are associated with much of the diversity in chemical and isotopic compositions of eclogites and eclogitic inclusions. For example, sulfide inclusions in diamonds with ¹³C < -10 exhibit a wide range of sulfur isotopic composition (Deines and Harris, 1995); garnet and clinopyroxene from eclogitic xenoliths that carry diamonds with ¹³C < -10 have ¹⁸O values higher than most mantle xenoliths (Deines et al., 1991). Such diverse values may reflect inhomogeneities in a subducted slab.

In their study of eclogitic diamonds from Jwaneng, Cartigny et al. (1998) argued the systematics of carbon and nitrogen isotopes precluded the association of these diamonds with subducted carbon. They based their interpretation mainly on the contrast between the negative ¹⁵N of Jwaneng diamonds and the positive ¹⁵N values of nitrogen in marine sediments and their metamorphic derivatives. I suggest that the negative correlation between carbon and nitrogen isotopic composition exhibited by the Jwaneng set is the result of mixing between graphite, formed by metamorphism of subducted organic matter ¹³C~-20, ¹⁵N~+5, C/N~1200) with mantle fluids ¹³C ~-6, ¹⁵N~-9, C/N~400).

It is suggested that diamonds with ¹³C < -10 were formed during interaction of graphite and fluids within discrete eclogitic micro-environments in a subducted slab. Those characterized by ¹³C~-5 were formed from fluids (or melts) that migrated through peridotitic and eclogitic rocks and acquired their carbon isotopic signature by exchange with mantle carbon. The nature of this "mantle carbon" is unclear. It may be the typical carbon scattered throughout the Earth's mantle. Alternatively, it may originate by devolatilization of carbonates and organic carbon from subducted slabs.

Cartigny P, Harris JW & Javoy M, Science, 280, 1421-1424, (1998).

Deines P & Harris JW, Geochim. Cosmochim. Acta, 59, 3173-3188, (1995).

Deines P, Harris JW, Robinson DN, Gurney JJ & Shee SR, Geochim. Cosmochim. Acta, 55, 515-524, (1991).

E04 : 5A/02 : F1

¹³C-n Diamond Systematic: Initial C/N Ratio of the Diamond Growth Medium and Relevance to Mantle Nitrogen Abundance and Mantle Fractionations

Pierre Cartigny (pcharti@ugcvax.dnet.gwdg.de)¹,

Jeff. W. Harris (J.Harris@geology.gla.ac.uk)² &

Marc Javoy (mja@ccr.jussieu.fr)³

¹ Lab. Géochimie Isotopes Stables, IPG, Université Paris VII, 4 Place Jussieu, 75251 Paris Cedex 05, France and Geochemisches Institut, Universität Göttingen, Goldschmidtstr. 1, 37 077 Göttingen, Germany.

² Dept of Geology and Applied Geology, University of Glasgow, Glasgow G12 8QQ, UK

³ Lab. Géochimie Isotopes Stables, IPG, Université Paris VII, 4 Place Jussieu, 75251 Paris Cedex 05, France

A compilation of more than 1200 ¹³C-N data from well characterised diamonds show a correlation of the maximum diamond nitrogen content (ie a limit sector) with ¹³C over the full diamond ¹³C range (ie more than 30 per mil). Diamonds with low ¹³C-values are characterised by rather low N-content (= 0 ppm at ¹³C < -30 per mil), whereas diamonds with high ¹³C have more variable nitrogen content, with a much higher upper limit (= 3500 ppm at ¹³C = -4.5 per mil). This correlation defines a concave trend that is therefore incompatible with a mixing relationship, such as would be produced by the admixture of a subducted component. This limit sector more likely reflects the evolution of mantle melts (or fluids) during their differentiation. The diamonds with nitrogen contents below that limit have fractionated the N/C ratio as a result of their slow growth conditions. Moreover the limit sector is applicable to every diamond parageneses (eclogitic, peridotitic and fibrous) suggesting that every diamond type may derive from melts of similar initial isotopic composition.

Considering a mantle ¹³C-value of -4.5 per mil, it is deduced that the initial C/N ratio of mantle melts (i.e diamond growth medium) from which diamonds crystallise ranges between 200 and 400, which is surprisingly similar to that of mid-ocean ridge basalts. Therefore, in spite of their different context and age, it appears that subcontinental and oceanic mantles give samples with similar ¹³C, ¹⁵N and C/N, suggesting an overall homogeneity of volatiles below continental lithosphere and within oceanic mantle since the Archean. Diamonds also allow to demonstrate that carbon and nitrogen do not behave similarly during the evolution of the diamond growth medium. However nitrogen concentration cannot be deduced in a simple way. If N behaved as an incompatible element during partial melting, a nitrogen concentration of about 2 ppm could be expected provided that the mantle carbon content is between 400 and 800 ppm. However, several lines of evidence suggest that nitrogen is not an incompatible element, implying, a higher mantle nitrogen concentration of at least 8 ppm and perhaps up to 40 ppm.

E04 : 5A/03 : F1

Nitrogen and Argon Isotopes in Carbonatites from the Kola Peninsula (Russia), an Open Window on the Deep Mantle

N. Dauphas (dauphas@crpg.cnrs-nancy.fr)¹,

B. Marty¹ &

I. Tolstikhin²

¹ CRPG, CNRS UPR 9046, rue Notre-Dame des Pauvres, BP 20, 54501 Vandoeuvre-lès-Nancy Cedex, France.

² Geol. Inst., Kola Sci. Center, Apatity 184200, Russia.

Stable isotopes (e.g., carbon) have long been studied in carbonatites (Deines 1989 & references therein), but surprisingly, investigations on nitrogen and rare gases in these rocks are scarce. High ³He/⁴He together with a steep ²⁰Ne/²²Ne-²¹Ne/²²Ne correlation (Marty et al., in press) indicate that a mantle plume contributed to the ca. 370 Ma old ultrabasic-alkaline-carbonatite magmatism in the Kola peninsula (eastern segment of the Baltic shield). Here, we report the first N-Ar measurements in ³He-rich inclusions released by stepwise vacuum crushing followed by on line N-Ar purification and static vacuum mass spectrometry.

A ²⁰Ne/²²Ne-⁴⁰Ar/³⁶Ar correlation (arising from mixing between the mantle and atmosphere end-members) allows to extrapolate a ⁴⁰Ar/³⁶Ar ratio of 5000±1000 for the plume source (Marty et al., in press). This ratio is used hereafter to constrain the isotopic and elemental characteristics of the Kola mantle source (KMS).

Data define a straight line in a N₂/³⁶Ar vs ⁴⁰Ar/³⁶Ar plot, which allows to estimate the N₂/³⁶Ar ratio of the KMS at 3.5 (0.7 (105, one order of magnitude lower than that of the upper mantle (Marty, 1995). Some data points are shifted to the right of this correlation, which may reflect a post-crystallisation radiogenic ⁴⁰Ar ingrowth (the samples are K-rich and ca. 370 Ma old). We plan to measure neon isotopes together with the N₂/³⁶Ar ratio to check this hypothesis.

Preliminary results suggest that the ¹⁵N value of the KMS is not significantly different from that of the convective upper mantle (as measured in MORBs, Marty & Humbert, 1997). If true, this would have profound implications on the Earth-atmosphere evolution. The present-day mantle is thought to have evolved from a primitive low ¹⁵N value (as recorded in diamonds, ¹⁵N=-25‰) to the present-day mantle value (¹⁵N~-4‰) as the result of the recycling of nitrogen with high ¹⁵N values (Cartigny et al., 1997 & references therein). Our results impose strong constraints on this scenario.

Cartigny, P et al, Terra Nova, 9/4, 175-179, (1997).

Deines, P, In Bell, K. (Ed.), Carbonatites. Unwin Hyman, London, 301-359, (1989).

Marty, B, Nature, 377, 326-329, (1995).

Marty, B & Humbert, F, Earth Planet. Sci. Lett, 152, 101-112, (1997).

Marty, B et al, Earth Planet. Sci. Lett. (in press).

E04 : 5A/04 : F1

Mantle Metasomatism by Fluids with Low Water Activity: Feldspar-Ti-Oxide-Bearing Peridotite Xenoliths in Basalts from SE Siberia

Dmitri Ionov (dmitri.ionov@mq.edu.au)

GEMOC, School of Earth Sciences, Macquarie University, (Sydney), 2109 N.S.W., Australia

The nature of fluids in the lithospheric mantle can be constrained by studies of modally metasomatised rocks produced by reaction of migrating fluids with host peridotite. The most common metasomatic products in mantle xenoliths and peridotite massifs worldwide are amphibole and mica. By contrast, xenoliths of spinel peridotite entrained in basaltic rocks in SE Siberia (Russia) have metasomatic aggregates of alkali feldspar and Ti-rich oxides (rutile, armalcolite, ilmenite) that replace amphibole and mica formed by earlier metasomatic episodes and make up cross-cutting veins. The evidence for breakdown of presumably hydroxyl-bearing amphibole and mica and their replacement by 'anhydrous' feldspar and Ti-oxides may indicate low water activity in the fluid responsible for metasomatism. Experimental results in the peridotite-CO₂-H₂O system at 10-15 kbar suggest that subsolidus dehydration of amphibole and mica may take place if CO₂/(CO₂+H₂O) ratios exceed 0.6-0.8 (Wyllie, 1978). The formation of the alkali feldspar and Ti-rich oxides is likely to be a relatively recent event because they are not in textural and chemical equilibrium with host peridotites. That process, however, is distinct from partial melting of hydrous minerals, spinel and pyroxenes in xenoliths shortly before and during their entrainment in host magma (e.g. Ionov et al., 1994) because the feldspar-bearing xenoliths do not contain silicate glass. It is not clear why the feldspar-Ti-oxide metasomatism is widespread in a few Siberian xenolith suites from different tectonic settings (Pacific continental margin, Baikal rift zone).

The unusual mineralogical composition of these metasomatic products defines a specific type of mantle metasomatism, distinct from those commonly attributed to H₂O-rich fluids, carbonate melts or Fe-Ti-rich silicate melts. Whole-rock and mineral analyses show that trace element enrichments in these xenoliths are largely hosted by alkali feldspar and Ti-rich oxides. Precipitation of this mineral assemblage can lead to unusual fractionations among incompatible elements both in the metasomatic assemblage and associated fluids (compared to precipitation of amphibole and mica). In particular, Ti-oxides may cause decoupling of high field strength elements (HFSE) from other elements in migrating fluids resulting in HFSE anomalies in mantle rocks and melt source regions. This type of metasomatism however is not likely to be responsible for negative Nb-Ta anomalies in arc magmas because it probably occurs only in shallow mantle (limited by the stability of alkali feldspars in peridotite) and in fluid environments with low water activity.

Ionov DA, Hofmann AW & Shimizu N, J. Petrol, 35, 753-785, (1994).

Wyllie PJ, J. Geology, 86, 687-713, (1978).

E04 : 5A/05 : F1

Low Copper and Sulfur Concentrations in Peridotites: A Result of Hydrous Melting in the Mantle?

Uwe Wiechert (wiechert@gl.ciw.edu)

Geophysical Laboratory, CIW, 5251 Broad Branch Rd., NW, Washington, DC, 20015, USA

Cu concentrations in Alpine peridotites from Lanzo, Balumuccia, Baldissero (Italian Alps) and Ronda (Spain) correlate with S concentrations indicating that sulfides host a considerable proportion of Cu in peridotitic rocks. In contrast, peridotite xenoliths exhibit lower Cu and S concentrations and no correlation between Cu and S. It has been suggested that the loss of S in xenoliths is due to an increase in oxidizing conditions in the host magma. Intergranular sulfides might be oxidized and sulfur lost as SO_x during the magma rises. However, magmatic oxidation of sulfides in xenoliths is not consistent with systematically lower Cu abundances compared to Alpine peridotites. Furthermore, the Italian Alpine massif of Finero is depleted in Cu (<20 ppm) and S (<100 ppm), thus excluding any relation between sulfide oxidation and magmatic processes near the surface. Therefore, it is most likely that oxidation of sulfides occurs in the mantle. For example, hydrous melting of peridotitic rocks may destabilize sulfides with Cu and S becoming highly incompatible elements. Restites of such a melting event would be low in Cu and S which is typical for peridotite xenoliths in alkali basalts. Thus, depletion of the upper continental mantle may have taken place under hydrous conditions. The Alpine peridotites of Finero also exhibit low Cu and S concentrations as well as high sulfur isotope ratios (+1.9 to +18.1‰ CDT). High ³⁴S values indicate that fluids originate from subducted crust/sediments. Therefore, we propose that hydrous fluids derived from these sources might "leach out" sulfides as they pass through the mantle wedge. This in turn would lead to high S and Cu concentrations in fluids derived from the slab and finally to high concentrations in subduction zone-related lavas assosiated with porphyry copper deposits. In contrast, the Alpine peridotites from Lanzo, Balmuccia, Baldissero (Italian Alps) and Ronda (Spain) are likely to be restites of anhydrous melting events e.g. melting below mid ocean ridges. Evidence for this includes a correlation between S and Cu indicating the presents of sulfides, high S abundances (>100 ppm), Cu content (>20 ppm), and sulfur isotope ratios between 0 and 1‰ (CDT).

E04 : 5A/06 : F1

Stability of Hydrous Phases in the Earth's Upper Mantle ­ Experimental 'Inconsistencies' Revisited

Peter Ulmer (pulmer@erdw.ethz.ch)

Department of Earth Sciences, ETH-Zentrum, Zurich, Switzerland

The stability of hydrous phases in mantle compositions is a fundamental parameter required for the evaluation of the transport, storage and mobility of hydrous fluids in the mantle. Numerous high-pressure experimental studies have been performed over the last 30 years to identify the potentially stable hydrous minerals and to delimit their stability fields. Despite the large number of studies performed with sophisticated experimental equipment, large, yet unresolved discrepancies exist in the stability of the principal hydrous phases likely to be stable in the earth's upper mantle: talc, serpentine, the 10-Å phase and phase A. The stability of the assemblage forsterite + talc, for example, varies by as much as 150°C at 2.5 GPa between different experimental studies which used very similar or even identical experimental techniques. Thermodynamically, the only variable which could explain some if not all of the observed 'inconsistencies' is the H₂O-fugacity: All experimental studies on the stability of hydrous phases assume that H₂O is the only fluid species present during the experiment and therefore fH₂O = p_total*<gamma>_H2O. The presence of other fluid species will lower the stability of the hydrous phases to a variable amount.

The following two mechanisms can decrease fH₂O in high-pressure experiments: (i) the well known diffusion of hydrogen in or out of the sample capsule and/or (ii) the diffusion of carbon through noble metal containers (recently demonstrated by Brooker et al., 1998). Hydrogen diffusion alone would either lead to H₂O-O₂ or H₂O-H₂ fluids in the capsule. Combination of carbon diffusion (from the furnace) and hydrogen loss would result in H₂O-CO₂ fluids. Both effects will lead to a reduction of the fH₂O and hence the hydrate stability.

An evaluation of the various experimental studies reveals that the stability of hydrous phases decrease with increasing run duration, the use of platinum capsules as opposed to gold capsules and the use of NaCl-assemblies compared to talc-pyrex or multi-anvil assemblies. In order to test and substantiate the proposed diffusional losses/gains as the prime reason for strongly varying hydrate stabilities in high-pressure experiments, a series of experiments is undertaken. The location of the reaction fo + tlc = en + H₂O is evaluated by running stoichiometric mixture of forsterite - talc - enstatite and excess H₂O in Au, Pt, and AgPd-capsules at 650 - 700C, 1.6 GPa and varying run duration (5 - 500 hours). Preliminary results indicate that for a given run duration (e.g. 44 hrs) the amount of educt phases (fo+tlc) is largest in the gold encapsulated sample and smallest in the platinum capsule. About 20% of the fluid is lost upon piercing of the capsule after the experiment (prior to heating), possibly indicating fluid species other than H₂O.

Brooker R, Holloway JR & Hervig R, Am. Mineral, 83, 985-994, (1998).

E04 : 5A/09 : F1

The Isotopic Composition of Argon in the Earth's Mantle

Mario Trieloff (Trieloff@kosmo.mpi-hd.mpg.de)¹,

Martina Falter (mfalter@kosmo.mpi-hd.mpg.de)¹ &

Elmar K. Jessberger (ekj@uni-muenster.de)²

¹ Max-Planck-Institut für Kernphysik, Saupfecheckweg 1, D-69117 Heidelberg, Germany

² Institut für Planetologie, Wilhelm-Klemm-Str.10, D-48149 Münster, Germany

Noble gas geochemistry is one tool to obtain information on the early history and composition of the Earth that otherwise is inaccessible. As contamination of mantle-derived rocks by atmosphere type noble gases is the most severe problem in elucidating the isotopic state of the mantle, we pursued a three-fold strategy to unravel the isotopic composition of argon in the Earth´s mantle: (1) separating mantle and atmospheric components by high resolution stepheating and stepcrushing, e.g. gas extraction in 15-22 heating steps instead of typically three (2) selecting those mantle samples that were already demonstrated to be most pristine. We analyzed submarine basalt glasses with the highest ²⁰Ne/²²Ne and ⁴⁰Ar/³⁶Ar ratios previously determined, including the prominent "popping rock" 2<Pi>D43 (Staudacher et al., 1989; Burnard et al., 1997; Moreira et al., 1998) and glasses dredged in 3-5 km depth at the southern rift zone of Loihi seamount (Valbracht et al., 1997). (3) Monitoring the components that are dissolved in the glass by a novel approach, stepheating after neutron irradiation. Neutron induced ³⁹Ar, ³⁷Ar and ³⁸Ar are derived from K, Ca and Cl, respectively. Their release pattern characterise Ar components homogeneously distributed within the glass. Heterogeneously distributed Ar is then identified by different degassing characteristics.

Our results show that mantle argon is partitioned between vesicles and basalt glass depending on vesicularity, while atmospheric Ar is heterogeneously distributed and most probably introduced at eruption. The high resolution stepheating and stepcrushing analyses improved the definition of the isotopic state of the Earth´s upper mantle [um] and the Hawaiian plume reservoir [pl] to ⁴⁰Ar/³⁶Ar~30,000 and ~6,000. We find the remarkable relationship: (⁴⁰Ar/³⁶Ar)_um/(⁴⁰Ar/³⁶Ar)_pl~(⁴He/³He)_um/(⁴He/³He)_pl~ (²¹Ne^*/²²Ne)_um/(²¹Ne^*/²²Ne)_pl where ²¹Ne^* is nucleogenic ²¹Ne. Such a relation between radiogenic (or nucleogenic) and primordial nuclides implies that the Earth's mantle reservoirs acquired the same initial He-Ne-Ar and K-Th-U compositions during accretion or that the mantle reservoirs were largely homogenized. While He and Ne were solar type, Ar was dominated by a planetary component. The latter is indicated by the ³⁶Ar/³⁸Ar ratio (e.g. Poreda and Farley, 1992) as well as the ²⁰Ne /³⁶Ar ratio (Moreira et al., 1998).

Solar He and Ne, and planetary Ar in the Earth's mantle can quite simply be explained by a mixture of solar and planetary rare gases. The planetary component that dominates the heavy rare gases was most probably carried by the planetesimals accreting to the Earth. The solar component could have been acquired by incorporation of noble gases from a dense solar nebula into an early magma ocean but the accompanying primary atmosphere must have been lost, probably along with dissipation of the dense solar nebula. Alternatively, the accreting planetesimals itself already carried the solar-planetary hybrid, e.g. such as some carbonaceous chondrites.

Burnard P, Graham D, Turner G, Science, 276, 568, (1997).

Moreira M, Kunz J, Allègre CJ, Science, 279, 1178, (1998).

Poreda R, Farley K, Earth Planet. Sci. Lett, 113, 129, (1992).

Staudacher Tet al, Earth Planet. Sci. Lett, 96, 119, (1989).

Valbracht PJ, Staudacher T, Malahoff A, Allègre CJ, Earth Planet. Sci. Lett, 150, 399, (1997).

E04 : 5A/10 : F1

Potassium and Sodium Distribution between Liquid Metal and Silicate Melt

Christine K. Gessmann (christine.gessmann@bristol.ac.uk) &

Bernard J. Wood

Department of Earth Sciences, Univeristy of Bristol, Wills Memorial Building, Queens Road, BS8 1RJ Bristol, U.K.

One of the most important events in the early history of the Earth was the separation of its iron-rich core. The rapid chemical fractionation associated with this event segregated most of the Earth's Fe and other 'siderophile' elements into the core, leaving the Earth's mantle depleted in siderophile elements relative to parental planetary material. The mantle is also depleted in a number of elements, which are not known to be siderophile, but whose abundances in the Earth provide important constraints on its evolution. Of particular importance are the alkali metals, whose depletion in the Earth is usually ascribed to their volatility but which, if present at modest concentrations in the core, would greatly change our ideas on Earth evolution. For example, if some K were to be present in the core, it would provide a source of energy for the geodynamo and an additional reservoir of ⁴⁰Ar.

Experiments to determine the distribution behaviour of alkali metals at conditions relevant to core formation processes were performed at 2.5 GPa and 1600ºC in a piston cylinder apparatus. Starting materials were FeS, FeS-Fe or FeS-FeSi metal mixtures intimately mixed with silicate powders that are enriched in K and Na. A range of redox conditions and S-contents has been imposed by the variation of the starting materials. Upon quench the liquid metal exsolves into intimately intergrown K, Na poor metal phases and K, Na-rich sulfides. Reintegrating the metal composition results in K and Na concentrations on wt% levels. The determined metal-silicate distribution coefficients for potassium and sodium are higher than previously reported values (for K) at similar pressures and temperatures. D_K is only slightly larger than D_Na, which is required by the similarity between K/Na in the Mantle and that in C1 chondrites. The distribution coefficients approach 1 with decreasing oxygen fugacity thus indicating increasingly siderophile behaviour of K and Na. Therefore the depletion of K and Na in the Earth's mantle may at least partly be accounted for by their siderophile behaviour rather than volatility during core formation processes.

E04 : 5A/11 : F1

Krypton Isotopes in the Earth's Mantle

Joachim Kunz (kunz@ipgp.jussieu.fr) &

Claude J. Allègre (zerbib@ipgp.jussieu.fr)

IPG, Géochimie et Cosmochimie, 4, place Jussieu, 75252 Paris, Cedex 05, France

Helium, neon, argon and xenon isotopes are in common use to investigate the composition of the Earth's interior. They give important constraints on the formation and the evolution of our planet and on the fluxes between its different volatile reservoirs (e.g. Allègre et al. 1986/87). Up to now, krypton isotopes have been of minor use only in the course of elemental ratio measurements (e.g. Moreira et al. 1998). Its isotopic ratios, except for specific geochemical conditions, were usually found to be atmospheric.

However, recent studies have put up the question, whether mantle krypton may reveal isotopic ratios slightly different from air. Two possible scenarios are discussed at present: a) A solar krypton component (Pepin, 1998). This suggestion is derived from an extrapolation of the well known presence of a solar neon component in the Earth's mantle towards the heavier rare gases. Although recent analyses of non-radiogenic argon and fission-shielded xenon isotopes disfavor this idea, a check for the krypton isotopes is needed. b) The presence of ²³⁸U- and ²⁴⁴Pu-fission xenon excesses in the MORB source (Kunz et al. 1998) should imply some ²³⁸U-fission krypton excesses, too. (²⁴⁴Pu-fission has rather no impact on krypton isotopes.)

For this reason we have carefully re-examined the exceptional mantle volatile rich MORB glass 2<Pi>D43, so-called popping rock, in a stepwise crushing experiment with special attention on the krypton isotopes. We find krypton isotopic ratios quite close to atmospheric values. Deviations for ⁷⁸Kr and ⁸⁰Kr can be explained by interferences from hyrocarbons (e.g. C₆H₆) and ⁴⁰Ar²⁺. Deviations for ⁸³Kr, ⁸⁴Kr and ⁸⁶Kr (relative to ⁸²Kr) are typically in the range of the mass discrimination effect. However, because the mass discrimination showed only small variations during the analytical period, we suggest that these deviations are real.

A solar component enriched in light isotopes (Pepin, 1991) is not found. In a three-isotope plot ⁸⁴Kr/⁸²Kr versus ⁸⁶Kr/⁸²Kr the data points follow roughly a linear trend, which is quite close to the mass fractionation line. This fractionation, which appears strongest in the last crushing steps opening the smallest vesicules, may have occured when volatiles (mainly CO₂) exsolved and formed the vesicules. The small offset between the data and the mass fractionation trend passing through air may indicate a ²³⁸U-fission contribution.

Allègre CL, Staudacher T & Sarda P, EPSL, 81, 127-150, (1986/87).

Kunz J, Staudacher T & Allègre CJ, Science, 280, 877-880, (1998).

Moreira M, Kunz J & Allègre CJ, Science, 279, 1178-1181, (1998).

Pepin RO, Icarus, 92, 2-79, (1991).

Pepin RO, Nature, 394, 664-667, (1998).

E04 : 5A/12 : F1

Noble Gases in the Iceland Plume

Darrell Harrison (mbexfdwh@stud.man.ac.uk),

Pete Burnard &

Grenville Turner

Dept of Earth Sciences, University of Manchester, Manchester, UK

Current geochemical models assume bulk transfer of matter from the deeper mantle to the upper mantle, which can be modelled at steady state. Accordingly, the upper mantle's noble gas inventory is a combination of a flux from the deeper mantle and addition of in-situ radiogenic decay products. We have analysed the noble gases trapped in Icelandic basaltic glasses and use the results to test the hypothesis of bulk mass transfer through a steady state upper mantle. Our results indicate that the Iceland samples have a lower primordial ³He/²²Ne ratio than MORBs, closer to estimates based on solar wind and meteorite analyses. The variability in helium relative to neon and argon suggests that noble gases are mass fractionated during transfer between mantle reservoirs. The Icelandic noble gases also have higher ²⁰Ne/²²Ne combined with lower ¹³⁰Xe/²²Ne than previously analysed mantle samples, yet do not exhibit the ¹²⁹Xe excesses characteristic of MORBs. Contrary to the usual explanation for the absence of ¹²⁹Xe excesses in ocean island basalts (OIBs), the effects of possible obscuration by atmospheric Xe is no greater in the Iceland samples than MORBs. If the steady state model is correct, this implies that Xe is excluded relative to Ne in the melting processes which generate plume-related basalts. Fractionation may result from the compatible behaviour of Xe at depth. The alternative explanation that ¹²⁹Xe has been preserved in the upper mantle for c. 4.5 Ga seems unlikely.

Using laser decrepitation techniques, selective noble gases and volatiles trapped within individual vesicles can be measured. Such techniques are currently being used to provide information into the mechanisms of atmospheric contamination of these Icelandic samples. This topic will be discussed briefly.

E04 : 5A/13 : F1

Crustal Contamination in the Lesser Antilles Island Arc; Evidence from Helium Isotope Variations in Arc Lavas

Matthijs C. van Soest (soet@geo.vu.nl)¹ &

David R. Hilton (drhilton@ucsd.edu)²

¹ Faculteit der Aardwetenschappen, Vrije Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands

² Scripps Institution of Oceanography, GRD, UCSD, San Diego, USA

We report a comprehensive study of helium isotope (³He/⁴He) variations on recent lavas from the Lesser Antilles island arc. We have analysed 55 mineral separates consisting of olivine, clinopyroxene, and orthopyroxene, separated from basaltic to andesitic lavas. Among the separates are 12 co-genetic pairs of ol and cpx and one cpx-opx pair. Helium was extracted by crushing under vacuum and isotopic variation and abundance are reported as R/R_A (R = (³He/⁴He); A = air) and [He] (ncm³STP/g).

Based on a previous study of helium isotopes in geothermal fluids from the arc (van Soest et al. GCA, in press) the arc can be divided into two regions: The southern arc (Grenada, St. Vicent, and St. Lucia) where crustal contamination (CC), is evident in the helium isotopes and the northern arc (Martinique, Dominica, Guadeloupe, Montserrat, and further north to Saba) where CC is not evident. The absence of evidence for CC in geothermal fluids from Martinique presents a dilemma, since CC has been well documented for lavas from this island (e.g. Davidson and Harmon, 1989). For this study Martinique will be treated separately.

There is a predominance of mantle helium throughout the arc with highest helium ratios in olivines coincident with the MORB ratio of 8±1R_A. For the northern arc (Guadeloupe to Saba) helium isotope ratios in olivine vary between 6.79-8.41R_A. In pyroxene, helium isotope ratios vary between 1.41-6.84R_A showing a correlation between (³He/⁴He) and [He]. For the southern arc (Grenada to St. Lucia) ratios in olivine vary between 3.56-7.64R_A, with the majority of the data between 3.56-4.90R_A. In pyroxene, ratios vary between 1.51-7.90R_A with the majority of the data in the range 1.51-4.99R_A. For Martinique only helium data from pyroxenes were obtained, and these show a range of 2.16-6.69R_A. Ol-Cpx pairs for the northern arc are all out of equilibrium, while for the southern arc most pairs are in equilibrium.

These results present compelling evidence that the helium isotopic composition of the Lesser Antilles mantle wedge is of MORB-like composition. It suggests the presence of a significant crustal contaminant in the southern arc. This contaminant has affected the helium isotope ratios of both olivine and pyroxene and is certainly present in the crust up to Martinique. No evidence for this contaminant was found north of Martinique, where MORB-like helium ratios are reported for all olivines. The correlation of (³He/⁴He) with [He] in pyroxene in this part of the arc suggests the facilitation of minor volatile contamination due to extensive degassing of the magma. Martinique seems to represent the point of transition of CC in the south to absence of CC in the north of the arc. This seems to be related to the tectonic situation of Martinique within the arc and positioning of Martinique's recent volcanic centers.

Soest MC van, Hilton DR & Kreulen R, Geochim. Cosmochim. Acta, in press, (1998).

Davidson JP & Harmon RS, Earth Planet. Sci. Lett, 95, 255-270, (1989).

E04 : 5A/14 : F1

Chemical Gas Monitoring on Merapi Volcano, Indonesia

Martin Zimmer (weihei@gfz-potsdam.de)¹,

Joerg Erzinger (erz@gfz-potsdam.de)¹ &

Yustinus Sulistiyo²

¹ GeoForschungsZentrum Potsdam, Telegrafenberg, D-14473 Potsdam,Germany

² BPPTK, Jalan Cendana 15, 55166 Yogyakarta, Indonesia

Since 1995 the GeoForschungsZentrum Potsdam (GFZ) has been actively working on the development and installation of a field observatory on Merapi volcano, Indonesia. In August 1998, within the framework of this project, we successfully installed a specially designed gas monitoring system for the continuous long time on-line gas analysis at the Merapi summit. This system consists of a strongly modified gas chromatograph, an alphaszintillometer and a temperature sensor. The complete unit operates automatically and can be remote controlled by radio link. Concentration levels of H₂O, CO₂, SO₂, N₂ and H₂ as well as the fumarole temperature were measured every 35 minutes over several weeks, and ²²²Rn and ²²⁰Rn (thoron) every 70 minutes. The data were transmitted via radio link to the Center for Volcano Research and Technology Development in Yogyakarta. In addition conventionally collected gas and fumarole condensate samples were analysed for He-, C-, H- and O-isotopes at the GFZ Potsdam.

Results of the on-line gas measurements to date show a significant variation of the volcanic gas composition. Periodically SO₂, CO₂ and H₂ concentrations increase while H₂O and N₂ amounts decrease. Simultaneously the ²²⁰Rn concentration increases due to a higher gas velocity. Lower fumarole gas temperatures occur during periods of high water concentrations in the gas resulting from a higher portion of cold meteoric water. These periodical gas pulses are correlated with higher seismic energies measured in the summit area. The ³He/⁴He-ratios in Merapi gas samples are between 8^-8.8*10^-6 which are typical for volcanic gases released at convergent plate boundaries. The ¹³C-value in CO₂ from Merapi gas is -4.9‰. This value is significantly heavier than average mantle material and could be explained by contamination of the magma with marine sediments. H- and O-isotope investigations on spring water and fumarole condensate samples indicate that the water present in the fumarole is mainly local rain water. This cold surface fluid circulates within the volcano edifice and mixes and dilutes with ascending volcanic fluids. A regular stronger degassing of the magma causes an increase in gas flux, gas velocity and gas temperature as well as an increase in the ratio of magmatic gas to meteoric water. This leads to higher CO₂, SO₂ and H₂ and lower H₂O and N2 concentrations in the fumarole gas.

The geochemical data series were correlated with geophysical parameters such as seismicity and deformation rates to create a seismo-geochemical model of the Merapi volcano and, thus, to deduce conclusions on the actual amount of ascending magma and herewith very crucial data on dome growth and volcanic activity.

Session E04:5B

E04 : 5B/21 : F1

Fluid Inclusions in Eclogites and Granulites from the Dabie Mountains, Central China

Bin Fu (fubi@geo.vu.nl)¹,

Jacques L. R. Touret (touj@geo.vu.nl)¹ &

Yong-Fei Zheng (yfzheng@ustc.edu.cn)²

¹ Faculteit der Aardwetenschappen Vrije Universiteit Amsterdam, De Bolelaan 1085, NL-1081 HV Amsterdam, The Netherlands

² Department of Earth and Space Sciences, University of Science and Technology of China, Hefei 230026, China

The burial of continental crust to depths of over 100 km and its subsequent exhumation during orogeny constitutes one of the most significant recent discoveries in present-day geology. The occurrence of microdiamond and coesite in eclogites from the Dabie Mountains in China makes this region one of the most important targets for studying the role of fluids during recycling of crustal materials in the processes of continental plate subduction and exhumation. Fluid inclusions were investigated from neighbouring (at a regional scale) ultrahigh-pressure (eclogites) and high-pressure metamorphic rocks (high-pressure granulites) from Qianshan and Luotian, respectively. Both have been metamorphosed during the same Mesozoic orogeny (220-240 Ma), but under quite different peak P-T conditions (P >= 27-28 kbar and T = 700±50°C for eclogites; and P = 10-14 kbar and T = 800-850°C for granulites) (e.g. Cong et al., 1995; Chen et al., 1998). Fluids trapped in inclusions are markedly different in composition and densities, implying different fluid regimes during peak metamorphism and subsequent retrograde evolution.In eclogites, fluids are dominantly aqueous, with very variable salinities from 0 to >30 wt% eq. NaCl. Extremely rare gaseous inclusions (CO₂ and, rarely N₂) have also been observed. Both fluids are found as primary inclusions in most eclogite-forming minerals (garnet, omphacite, zoisite and quartz inclusions in these minerals), with significant differences in the compositional of aqueous inclusions for each mineral. Aqueous and gaseous fluids remained immiscible during most of the metamorphic evolution, including peak conditions. Inclusion densities were constantly reset during most of the retrograde evolution, involving the formation of new inclusions (transposition features), but overall fluid migration did not exceed the scale of the host crystals. Granulites show prominent decompression textures, almost identical to those described e.g. in Limpopo, southern Africa. Inclusions are far more abundant than in eclogites; most of them are CO₂-rich (carbonic fluids). Some low-salinity aqueous inclusions are later, related to a rather late stage of retrogression. Primary inclusions in quartz-plagioclase assemblages within garnet, possibly issued from former melt, suggest that a carbonic fluid did exist prior to or during the crystallization of a syn-granulite partial melt.In eclogites, at the exception of CO₂ which may derive from neighbouring coesite-bearing marbles, fluids are essentially internally-derived, with three possible sources: remnants of pore fluids within the initial protoliths, fluids released from crystallizing melts, or release of hydroxyls initially dissolved in eclogite-forming minerals during ultrahigh pressure metamorphism. Total amount and fluid movements were extremely limited, resulting in the preservation of high-pressure mineral phases (e.g. coesite). Synmetamorphic free fluids (mainly CO₂) were more abundant and mobile in granulites, notably in high-temperature granulites (e.g., Sri Lanka or India).

Cong B, Zhai M, Carswell DA, Wilson RN, Wang Q, Zhao Z & Windley BF, Eur. J. Mineral, 7, 119-138, (1995).

Chen N-S, Sun M, You Z-D & Malpas J, J. Metamor. Geol, 16, 213-222, (1998).

E04 : 5B/22 : F1

Phase Relations in the K₂O-SiO₂-CO₂-H₂O System: Preliminary Results

Andy H. Shen (ahs26@esc.cam.ac.uk)¹ &

Haifei Zheng (hfzheng@geoms.geo.pku.edu.cn)²

¹ Department of Earth Sciences, University of Cambridge, Cambridge, CB2 3EQ, UK

² Department of Geology, Peking University, Beijing 100871, PRC

The nearly ubiquitous potassium enrichment in metasomatised mantle rocks and the vast presence of carbon dioxide in fluid inclusions from mantle xenoliths signify the importance of these two chemical components in mantle geochemical processes. In addition, the existence of carbonatites further demands clear knowledge of the phase relations of dry and hydrous carbonate-silicate systems when one considering the geochemical processes occurring in the mantle. Many experimental studies were carried out in the past using quenching method to preserve the phase assemblage. One recent study carried out by Wilkinson et al. (1996), using the synthetic fluid inclusion method, discovered a two-fluid immiscibility gap in the K₂O-SiO₂-CO₂-H₂O system within the low pressure-high temperature range.

We explored this system with the hydrothermal diamond-anvil cells to 700^oC and 9 kbar. The starting material consists of 0.72 molal K₂CO₃ aqueous solution and a synthetic quartz chip. The experimental procedure has been described in Shen and Keppler (1995, 1997). The temperatures were directly measured from the thermocouples and the pressures were estimated using NaCl-H₂O equation of state (Shen, 1994 and Shen et al., 1995) with equivalent electrolyte concentration. The following preliminary summaries were drawn from our 30 experimental runs. Firstly, the two-fluid immiscibility can be easily observed by the presence of two co-existing fluids and the approximate location of the immiscibility gap from our experiments agrees well with that determined by Wilkinson et al. (1995). Secondly, there seems to be a decarbonation reaction at high-pressure region (~6-9 kbar) and this boundary seems to be rather temperature-independent within 200 to 600 ^oC. However, we needs more evidence to confirm this idea. Finally, the dissolution of quartz into the aqueous solution near the "critical region" is very dramatic, as if the quartz becomes molten at 400-500 ^oC and 1 kbar.

Shen AH, Ph.D. Dissertation, Cornell University

Shen AH, Chou I-Ming, Bassett WA, Terra Abstracts, 7, 71, (1995).

Shen AH, Keppler H, Am. Mineral, 80, 1335-1338, (1995).

Shen AH, Keppler H, Nature, 385, 710-712, (1997).

Wilkinson JJ, Nolan J, Rankin AH, Geology, 24, 1059-1062, (1996).

E04 : 5B/23 : F1

Elemental Fractionation at Convergent Margins: Evidence from Glass Compositions from the Central Izu Arc (NW Pacific)

Susanne M Straub (sstraub@geomar.de)¹ &

Graham D Layne (glayne@whoi.edu)²

¹ GEOMAR Research Center, 24148 Kiel, F.R.G.

² Woods Hole Oceanographic Institution, Woods Hole MA 02543, U.S.A.

We analyzed major and trace element (Cl, B, Li, Be, F, H) contents of arc melts of the intra-oceanic Izu arc utilizing glass shards and melt inclusions (=glasses) from distal tephra fallout. The tephra has been drilled at the outer fore arc (ODP site 782A) and ranges from mid-Miocene to Quaternary in age. Previous bulk work established the provenance from depleted arc front sources (Nb~0.2-1.5 ppm, (La/Sm)_n~0.6). The glasses investigated have a large range in silica (50-75 wt% SiO₂) that corresponds to only a small range in MgO (=6-1 wt%). At the high MgO end, the glasses are offset from low-MgO MORBs: Ti, P, Na, and Be are lower, but Al and the incompatible elements B, K, Cl, H, F are comparatively enriched. Si, Fe and Ca overlap with MORBs. A key observation is that Be follows a tight trend in the glasses, but increases only from ~0.4 to ~0.6 ppm with increasing silica. This precludes the compositional range of the glasses having been produced by fractional crystallization because Be should behave incompatibly (D_bulk/~0.5) during crystallization of the phases observed in the tephra. The low Be and the linear trends of other incompatible elements suggest that the glass compositions reflect mixing of high- and low -silica endmembers which are both depleted in Be. A simple genetic model is that a HFSE-depleted mantle wedge is fluxed by a Be-poor hydrous siliceous component from the subducting slab. The slab component is enriched in Si, B, K, Li, Cl, F, H (and possibly Na and Al) but is not homogenous, as demonstrated by the significant range in B/K (0.004-0.075), Cl/K (0.20-1.60) and B/Li (2.5-5). Because shallow degassing can cause loss of volatile elements H, Cl, Li and F, determining the elemental budget of individual slab-components is difficult. However, andesitic melt inclusions in anorthitic plagioclase that are enriched in Li, Cl and F, but not in B or K, provide clear evidence that B can be decoupled from Li and the halogens.The large range in ratios of LLE/Be (e.g. B/Be~20-100) implies considerable fluctuations in the magnitude of the slab-flux from a single arc segment. Presently, we cannot distinguish whether the hydrous components derive from dehydration of the subducted sediment or the basaltic crust. Temporal fluctuations, however, suggest that - next to source composition - the slab flux at any given time is also influenced by processes like fractionation and mixing.

E04 : 5B/24 : F1

Fluid and Element Cycling During Subduction of Serpentinite

M. Scambelluri (msca@dister.unige.it)¹,

G. L. Frueh-Green²,

G. B. Piccardo¹,

E. Rampone³ &

F. Vallis¹

¹ Dipartimento di Scienze della Terra, Genova, Italy

² Institut fur Mineralogie und Petrographie, ETH-Zentrum, Zuerich

³ Dipartimento di Scienze della Terra, Milano, Italy

Serpentinized peridotites are a significant component of the oceanic lithosphere and can contain 10-13 wt% bulk water fixed in serpentine and other associated hydrous phases. Serpentine may survive to considerable depths during cold subduction and thus serpentinites are the most effective transporters of water into the mantle. At some stages of burial serpentinites partially dehydrate and originate metamorphic fluids. In such cases the textural and compositional relationships between the oceanic assemblages and the high-pressure minerals + fluids become crucial to assess the fluid and element cycling by serpentinites during subduction. These aspects are investigated here in serpentinites from the Western Alps (Erro-Tobbio peridotite unit), representing fragments of mantle emplaced at high crustal levels during opening of the Mesozoic Alpine Tethys, and successively recrystallized at eclogite-facies conditions during alpine plate convergence and subduction of the oceanic lithosphere. Eclogitization caused partial dewatering and development of olivine + diopside + Ti-clinohumite + antigorite in the rocks and in vein systems. Shear localization during subduction produced high-pressure mylonitic shear zones, surrounding cores of undeformed serpentinized peridotite with static and incomplete high-pressure recrystallization which still preserve abundant relics of the pre-subduction assemblages. Previous studies showed that high-pressure breakdown of chlorine and alkali-bearing hydrothermal serpentines and phyllosilicates was accompanied by partitioning of these elements into high-pressure brines. This suggests deep cycling of shallow fluids during eclogitization. The shallow serpentinization was also accompanied by decrease in bulk-rock Ca, coupled with increase in H₂O and Sr as a function of the intensity of alteration. Bulk analyses of high-pressure veins display much higher Sr concentrations than the precursor serpentinites and than their host rocks, thus indicating Sr mobilization in the high-pressure fluid. Vein-filling diopside has trace element compositions comparable to those of primary mantle clinopyroxene still preserved in the host rocks, the main difference being related to Sr positive anomaly shown by the vein diopside. Oxygen isotope ratios of relict mantle clinopyroxene and of hydrothermal serpentine (respectively between 5-7 per mil and 6.4-7.6 per mil) suggest that the shallow alteration took place at low-temperatures. High-pressure diopside and antigorite from samples fully recrystallized at eclogitic conditions, display ¹⁸O signatures similar to those of clinopyroxene and serpentine from hydrothermally altered samples. Moreover, the ¹⁸O compositions of coexisting diopside and antigorite vary significantly over centimetre scales in both low and high-strain eclogitized rocks. The recognition of O-isotope heterogeneities at various scales indicates overall disequilibrium and suggests limited fluid production during high-pressure metamorphism; interaction with the high-pressure fluid did not reset the original oceanic imprint. The above results suggest that at eclogite-facies conditions the analyzed peridotite internally cycled fluid and elements such as chlorine, alkalis and strontium that were uptaken during alteration in surface environments close to the seafloor.

E04 : 5B/25 : F1

Origin of Deep Crustal Waters in the Hellenic Volcanic Arc (Greece)

Volker J. Dietrich (wumme@erdw.ethz.ch)¹,

E. Gartzos²,

C. Leonis³,

Th. Lyberopoulou³,

F. M. Schwandner (florimax@erdw.ethz.ch)¹,

C. Moor (christoph.moor@empa.ch)⁴,

R. Kipfer (kipfer@eawag.ch)⁵ &

G. Bernasconi-Green (gretli@erdw.ethz.ch)¹

¹ Inst. Mineralogy & Petrography, ETH Zürich, Switzerland

² Inst. Mineralogy & Geology, Agricultural University, Athens, Greece

³ Inst. Geology & Mineral Exploration, IGME, Peania (Attica), Greece

⁴ EMPA Dübendorf, Switzerland

⁵ EAWAG, Dübendorf, Switzerland

Cold CO₂-rich spring waters, geothermal and hydrothermal waters and fumarolic gases related to the Hellenic Volcanic Arc and deep crustal faults were choosen for full geochemical and isotopic analyses. The water chemistry (including major and trace elements) as well as stable isotope ratios of C, H, O and noble gas isotopic ratios of He, Ne, Ar, Kr, and Xe allow a comprehensive discussion on a mantle or magmatic origin, on the influx of metamorphic, meteoric, seawater, hydrothermal and surface waters. The Hellenic volcanic arc is regarded as a magmatic expression of the still-active subduction of the African plate beneath the Aegean plate, which started around 4 Ma at the beginning of Pliocene. It extends over 600 km from the Volos/Atalanti area to Corinth and Sousaki, Methana, Poros and Aegina islands in the Saronic gulf, to the islands of Santorini and Milos, Kos, Yali and Nisyros in the South Aegean sea. All volcanic fields and islands are associated with major tectonic lineaments and active NE/SW and E/W striking faults. In many cases, cooling magmatic bodies at deeper crustal levels seem to provide the required thermal energy for heating deep circulating water, the generation of hydrothermal systems (e.g. Milos and Nisyros) and ascend to the surface.The temperatures of the thermal waters range from 24^oC (Sousaki) to 75^oC (Edipsos) and TDS range from 500 ppm to 32 000 ppm respectively. Processes, affecting the water composition, such as mineral-fluid equilibria, mixing, boiling and conductive cooling, have been taken into consideration and calculated. The source regimes of cold mineral, geothermal and hydrothermal waters were discriminated by the use of halogenide ratios of Cl, F, Br and I. Mixtures of seawater and meteoric water could be calculated using ²H and ¹⁸O. ¹³C values from bicarbonates show the contribution of metamorphic components through decarbonation of marine sediments as well as biogenic and atmospheric components.Helium, as well as the other noble gases are dissolved mainly in CO₂-rich waters and gases and are accompanied by CH₄ and H₂ emanations. The ³He/⁴He ratios of 7.5 to 4.4 x 10^-6 of the Nisyros fumarolic condensates and of some hydothermal waters from Milos reflect well the high amount of mantle derived primordial ³He, which in the case of Nisyros, may be related to magma degassing. The lower ratios of 3.4 to 1.6 x 10^-6 in the hydrothermal waters from Milos and in the Methana thermal springs represent a mixing with radiogenic ⁴He due to crustal contamination during ascent of primordial ³He through the crust.

E04 : 5B/26 : F1

Evaporite is an Important Boron Source for B-Rich Granites and Pegmatites

Shao-Yong Jiang (sjiang@mpch-mainz.mpg.de)

Max-Planck-Institut fuer Chemie, Postfach 3060, 55020 Mainz, Germany

Magmatic volatiles-water, halogens, phosphrous, and boron-are of great importance in the generation and evolution of granitic magma. The existence of boron in melt can significantly reduce liquidus and solidus temperatures, decrease viscosity, and increase the solubility of H₂O, and thus modify the melt properties. Boron may also affect the fractionation of other petrologically and economically important metals in the melt. However, the origin of boron in the magmatic systems is still a matter of great debate.

Tourmaline has a wide P-T stability and can persist in metasedimentary rocks at least up to anatexis conditions. Therefore, it has been suggested that tourmaline represents a most likely reservoir of B for S-type granitic magmas. Most common rock-forming silicate minerals contain only minor to trace amounts of B (up to several hundred ppm in muscovite, for example), but their high modal abundance may still make them as a significant B inventory for granitic magmas. Evaporites, especially non-marine evaporites, contain high concentrations of B, and are therefore potentially important boron sources, in particular for those extremely high-B granites and pegmatites. However, previous studies have not yet obtained conclusive data to support this hypothesis as evaporites are extremely soluble and not readily preserved in the geologic record.

Boron isotope composition is a relatively new geochemical tool to trace the source of boron. The ¹¹B values of the major reservoirs of boron are well characterized. Of particular interest is the large difference in ¹¹B values between marine and non-marine evaporites (+18 to +32‰ and -30 to +10‰, respectively). In this paper, I report ¹¹B data for several B-rich granites and pegmatites from the Czech Republic. The very low ¹¹B values (-15 to -36‰) in the studied granitic and pegmatitic tourmalines provide strong evidence to support derivation of boron from non-marine evaporites, either in the source regions or during ascent of the magma. Non-marine evaporites are the only boron reservoir known in nature with an ¹¹B value less than -20‰. There is an overlap of ¹¹B values between most crustal rocks and the upper data range of non-marine evaporites (0±10‰). Hence, it is possible that many B-rich granitic bodies with higher ¹¹B values (around -10‰) such as Cornwall granites in SW England may also have non-marine evaporites as a major source for their boron.

E04 : 5B/29 : F1

Fluid Exchange between the Lower Continental Crust and Upper Mantle

Lieselotte J. A. Bolder-Schrijver (schl@geo.vu.nl) &

Jacques L. R. Touret (touj@geo.vu.nl)

Faculteit der Aardwetenschappen, Vrije Universiteit Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands

Situated at the interface between the continental crust and the mantle, the lower continental crust is a complex domain where a number of fluid/melt/mineral interaction processes take place. A limited number of fluid types are repeatedly observed in all lower crustal segments now exposed at the Earth's surface. Composition, possible source and transport pathways of these fluid types, have been documented by fluid inclusion and stable isotope studies. They are represented by high density gases (mainly CO₂), to a much lesser extent N₂ and other species) and aqueous solutions of variable salinity (mainly high-salinity brines).

The nature of dominant fluids and the scale of fluid transport appear to be essentially controlled by the maximum pressure reached during peak metamorphic conditions. In eclogites, fluids are dominantly aqueous with very variable salinities, more rarely gaseous (N₂ dominant) (Nadeau et al., 1993). These fluids are internally generated during the prograde metamorphic evolution, resulting in a limited amount of fluids of variable composition in crystal-size domains. On the contrary, large quantities of fluids with a relatively constant composition (CO₂ and high-salinity brines) are generated in granulites, due to the increasing importance of partial melting in granulites (from high- toward low-pressure), as well as the systematic occurrence of synmetamorphic intrusives (especially in high-temperature granulites) (Touret & Huizenga, 1998). These fluids are able to percolate on rather large distances along channelized pathways.

Detailed discussion of a number of occurrences in former Gondwana lands such as Madagascar, India and Sri-Lanka (Bolder-Schrijver et al., submitted) indicate that most CO₂ is mantle-derived, transported in the lower crust by CO₂-saturated melts of variable composition (silicate and carbonate).

Nadeau S, Phillipot P & Pineau F, Earth and Planetary Science Letters, 114, 431-448, (1993).

Touret JLR & Huizenga JM, Journal of African Earth Sciences, (in press), (1998).

Bolder-Schrijver LJA, Kriegsman LM & Touret JLR

E04 : 5B/30 : F1

Solubility and Diffusivity of Noble Gases in Synthetic Phlogopite: An UVLAMP Investigation

Knut Roselieb (kroseli@gwdg.de)¹,

Jo-Anne Wartho (J.Wartho@open.ac.uk)²,

Heinz Büttner (hbuettn@gwdg.de)¹,

Albert Jambon (jambon@ccr.jussieu.fr)³ &

Simon Kelley (S.P.Kelley@open.ac.uk)²

¹ Mineralogisch-Petrologisches Institut, Goldschmidtstr. 1, 37077 Göttingen, Germany

² Department of Earth Sciences, The Open University, Milton Keynes MK7 6AA, England

³ Laboratoire MAGIE, 4 Place Jussieu, 75252 Paris, France

Noble gas partitioning between crystalline phases and melts is a prerequisite to understanding terrestrial outgassing. Knowledge of the partition coefficient requires analyses of the solubility in both crystal and melt. Solubilities in melts are now well known, whereas solubility in minerals is very low and mechanisms are not well understood. Since bulk analytical methods tend to overestimate partition coefficients (Roselieb et al., 1997; Brooker et al., 1998), we have employed the UV laser ablation microprobe (UVLAMP), a sensitive method with high spatial resolution for noble gas analysis. We used this method to determine simultaneously the solubility and diffusivity of noble gases in a synthetic iron-free F-phlogopite, which was synthesised by Hammouda et al. (1995). Moreover, knowledge of the diffusivity of Ar in K-bearing minerals is needed in order to evaluate if K-Ar or Ar-Ar techniques yield reliable geochronological information.

Pt-capsules were filled with single grains of the sample and the noble gas (pure Ne, Ar or a mixture of 35% Ne, Ar and 10% He, Kr, Xe) and welded closed following the method of Boettcher et al. (1989). The capsules were run in an internally heated pressure vessel. Experiments with Ne and Ar were performed at a pressure of 6 kbar and temperatures of 1200°C and 1300°C for 24 h. During the Ar-analyses both surface layer and cleavage effects were encountered and Ar concentrations up to 600 ppm were detected from these regions. These effects were avoided by removing the top surface layer and controlling the laser power and ablation rate so as not to ablate an underlying cleavage trace. Profiles were analysed, perpendicular to the cleavage traces (after removal of the surface layer), along two approximately 400 x 200-250 µm long and 2 µm deep traverses. From the measured Ar profiles the diffusion coefficient can be calculated employing the solution for radial diffusion in a cylinder (Crank, 1975) yielding a diffusivity of 1.6x10^-9 and 1.1x10^-9 cm²/s at 1300°C. The solubility of Ar, which is represented by the surface concentration, can be calculated from the profiles and yields a value of 11 ppm. Using the Arrhenius-parameters of the bulk analysis from Giletti (1974), an extrapolation to 1300°C yields a diffusivity of 6.9x10^-9 cm²/s, which is in reasonable agreement with our results. Assuming that the solubility obeys Henry's law, an Ar partition coefficient can be estimated. For phlogopite an argon solubility of 0.0018 ppm/bar was calculated. This solubility can be compared to the Ar solubility in an ultrabasic potassium rich melt (e.g., leucite-basanite or ugandite) in equilibrium with phlogopite. Ar solubilities in these melts at 1350°C range from 0.22 to 0.08 ppm/bar (Lux, 1987). The comparison yields an estimate of the partition coefficient ranging from 0.008 to 0.02.

Boettcher SL, Guo Q & Montana A, Am Min, 74, 1383-1384, (1989).

Brooker RA, Wartho J-A, Carroll MR, Kelley S & Draper DS, Chem. Geol, 147, 185-200, (1998).

Crank J, The Mathematics of Diffusion. Clarendon, Oxford, 2nd ed, (1975).

Giletti BJ, Geochemical transport and kinetics. Carn. Inst. Washington, 634, 107-115, (1974).

Lux G, Geochim. Cosmochim. Acta, 51, 1549-1560, (1987).

Roselieb K, Blanc Ph, Büttner H, Jambon A, Rammensee W, Rosenhauer M, Vielzeuf D & Walter H, Geochim. Cosmochim. Acta, 61, 533-542, (1997).

E04 : 5B/31 : F1

Vesicle Loss and Rare Gas Concentrations in MORB Glasses

Philippe Sarda (sarda@geol.u-psud.fr)¹ &

Manuel Moreira (mmoreira@whoi.edu)²

¹ Département des Sciences de la Terre, Université Paris 11, 91405 Orsay Cedex91405 Orsay Cedex, France, France

² Woods Hole Oceanographic Institution, Woods Hole MA, 02543, USA

The aim of this study is to better understand the variations of rare gas concentrations in Mid Ocean Ridge Basalt (MORB) glasses. We use our data for the Discovery and Shona hotspots zone, in the Southern Atlantic Ocean, selecting samples with no plume influence. These data were generated at I.P.G. Paris.

In 1990, Sarda & Graham proposed that most MORB samples are highly degassed by loss of the vesicles generated at equilibrium during magma ascent or in the magma chamber, i.e. following a simple Henry's law of solubility. This view was challenged by Matsuda & Marty (1995), who presented a positive correlation in a ⁴He/⁴⁰Ar vs ⁴He diagram in contradiction with this hypothesis. However, their correlation was given by OIB data, while MORB did not appear to exhibit any correlation at all.

Here, we use He, Ne, Ar, Kr, Xe data. We show that the model of vesicle loss is supported by the variations of the ²⁰Ne/⁴⁰Ar* ratio, which indeed increases when Ar concentration decreases. What is seen is mainly a two component mixing between vesicles and fractionated, gas-poor magma. Helium appears perturbed and the possibility that it was lost by an additional diffusion process is discussed.

The ⁴He/⁴⁰Ar* and ³He/²²Ne elemental ratios vary coherently, which is consistent with vesiculation and vesicle loss to be the mechanism that controls the concentrations and elemental ratios of the rare gases in MORB glasses.

Sarda P, Graham D W, Earth Planet Sci. Lett, 97, 268-289, (1990).

Matsuda J, Marty B, Geoph. Res. Lett, 22, 1937-1940, (1995).

E04 : 5B/32 : F1

Degassing of Submarine Basalts: The OIB Case

Manuel Moreira (mmoreira@whoi.edu)¹ &

Philippe Sarda (sarda@geol.u-psud.fr)²

¹ Woods Hole Oceanographic Institution, 360 Woods Hole road, MS25, Woods Hole, MA 02543, USA

² Groupe de Géochimie des Gaz Rares, Departement Sciences de la Terre et U.M.R. Orsayterre, Université Paris 11 - Sud, 91405 Orsay Cedex, France

We compare the rare gas elemental ratios measured in glass samples of Mid-Oceanic Ridge Basalts and submarine basalts from mantle primitive plumes to current estimates of the same ratios for upper and lower mantle reservoirs. The combined use of He, Ne and Ar elemental systematics allows to differentiate clearly between the degassing process at mantle plumes, which appears to be controlled by distillation, and at normal mid-oceanic ridges, where vesiculation occurs at equilibrium (Sarda and Graham, 1990; Sarda and Moreira, 1998). Such a distillation has important consequences for the interpretation of the measured rare gas isotopic ratios in primitive plume-derived materials and specially for the so-called "helium paradox" that says the helium content of plumes glasses is always lower than MORB glasses even if they seem to be less degassed.

Sarda P, Graham D W, Earth. Planet. Sci. Lett., 97, 268-289, (1990).

Sarda P, Moreira M., J. Conf. Abs., 4 (1999).

E04 : 5B/33 : F1

A Mantle Plume in Transylvania - Noble Gas Isotope Studies of Mantle Xenoliths from the Persani Mts., Romania

Tilmann Althaus,

Samuel Niedermann &

Jörg Erzinger

GeoForschungsZentrum Potsdam Department 4.2 Telegrafenberg D-14473 Potsdam, Germany

In the Persani Mts., Romania, alkali-basalts carrying abundant ultramafic mantle xenoliths were erupted in Plio-Pleistocene time (2.5 - 0.7 Ma). These eruptions were contemporaneous with calc-alkaline volcanic activity which was caused by the subduction of oceanic crust and the convergence of the Tisia-Dacia-microplate and the European platform. The volcanic areas are separated by 40 km distance only. Samples investigated in this work include spinel-lherzolites and clinopyroxenites, mineral separates thereof and metasomatic kaersutitic amphiboles. They were recovered from two quarries at Bogata and from a roadcut exposure at La Gruiu Fintina. We found clear evidence for a mantle origin in He, Ne and Ar isotopic compositions, whereas Kr and Xe show no deviation from atmospheric ratios.

Two distinct mantle He components can be discerned. The first and more abundant one is characterised by ³He/⁴He of 6.5 - 7.3 times the atmospheric value (RA). This is similar to ratios observed in xenoliths from other European localities by Dunai & Baur (1995) and is regarded as being characteristic of the shallow lithospheric mantle. This component is released at ¾1200°C and by mechanical crushing, thus it is most probably stored in fluid inclusions. The second He component showing extremely high ³He/⁴He up to > 20 R_A was observed in some olivines, orthopyroxenes and one amphibole during the final (1800°C) heating step. The abundance of this component, which is obviously dissolved in the crystal lattice of the samples is very low (⁴He < 5 x 10^-10cm³ STP / g). A spallogenic origin due to cosmic irradiation during surface exposure is not indicated, because the samples were taken from the bulk of the lava flows exposed only recently in the active quarries. Comparably high ³He/⁴He have been observed in the Hawaiian mantle plume.

An atmosphere-like Ar-component with ⁴⁰Ar/³⁶Ar close to the air value of 295.5 is released during lower temperature heating steps and in crushing experiments. At high temperature (1800°C) Ar with high ⁴⁰Ar/³⁶Ar of up to 10500 is observed indicating that mantle-derived Ar is predominantly bound in the crystal lattice.

Neon is characterised by three distinct components. Most abundant is atmosphere-like Ne. Ne with high ²⁰Ne/²²Ne and with correlated ²¹Ne/²²Ne ratios that are higher than those observed in MORB glasses (Sarda et al., 1988) is indicative for the shallow lithospheric mantle. The ²¹Ne excesses relative to MORB have not been produced by an in-situ nuclear process but are inherited from the mantle. Moreover in two samples we have identified Ne with high ²⁰Ne/²²Ne but lower ²¹Ne/²²Ne interpreted to be of mantle plume origin.

Dunai T J & Baur H, Geochim. Cosmochim. Acta, 59, 2767 - 2783, (1995).

Sarda P et al, Earth Planet. Sci. Lett., 91, 73 - 88, (1988).

E04 : 5B/34 : F1

The Isotopic Composition of Argon in Xenolith Nodules from La Réunion Island, Indian Ocean

Jens Hopp (hopp@kosmo.mpi-hd.mpg.de) &

Mario Trieloff (trieloff@kosmo.mpi-hd.mpg.de)

MPI für Kernphysik, Postfach 103980, Heidelberg, Germany

The determination of the noble gas isotopic signature of the Earth´s interior is important for modelling mantle degassing and evolution. Noble gas data point to the existence of two different mantle reservoirs: the strongly degassed MORB mantle and a more primitive mantle source represented by oceanic islands, characterised by lower proportions of radiogenic noble gases. In particular the occurrence of solar type neon in these mantle reservoirs cannot be reconciled with atmospheric or crustal contamination. Signatures for the most primitive helium and neon were obtained for the Loihi hot spot (Valbracht, 1997), while previous He and Ne data of Réunion (Staudacher, 1990) suggest a more radiogenic, i.e. less primitive mantle source. Here we show that the isotopic composition of argon provides additional evidence for a less primitive mantle reservoir. Contrasting previous attempts using basalt glasses and olivine phenocrysts, we analysed xenoliths and yield a better preservation of mantle volatiles. Further separation of atmospheric contamination (introduced through meteoric waters or hydrothermal activity) was achieved by high resolution stepheating and stepcrushing of neutron irradiated samples (inclusion rich and inclusion poor sub-mm-sized chips, or precrushed grain size fractions). Neutron induced argon isotopes ³⁷Ar, ³⁸Ar, and ³⁹Ar derived from Ca, Cl, and K, respectively, allow identification of carrier phases of argon of different isotopic composition (Trieloff, 1997). Upon stepcrushing, dunite xenolith ILR84-4 released argon with ⁴⁰Ar/³⁶Ar-ratios between 5200 and 9600, hosted by abundant fluid inclusions. Similarly ⁴⁰Ar/³⁶Ar-ratios of 5000-10800 were obtained for argon extracted at temperatures <900°C from a further inclusion rich split, while an inclusion poor split gave significantly lower values (3000-5800). This is most probably due to low temperature decrepitation of fluid inclusions, that dominate the inclusion rich split, but are suppressed in the inclusion poor split. At high temperatures of olivine melting above 1450°C, high ⁴⁰Ar/³⁶Ar-ratios persisted in both stepheated samples, interpreted as release from inclusions still not degassed upon olivine melting. At intermediate temperatures argon with lower ⁴⁰Ar/³⁶Ar-ratios of 1000-4000 was released. Correlated degassing of ³⁹Ar and analyses with electron microprobe show this (atmospheric influenced) component to be hosted by K-bearing magmatic veins incorporated during the contact with the xenolith's host magma. We suggest a lower limit of 8000-10000 for the ⁴⁰Ar/³⁶Ar-ratio of the Réunion mantle source confirming the less primitive character compared to Loihi. We will discuss possible explanations, e.g. a stronger degassing of a primitive progenitor material, large scale admixture of MORB type noble gases or recent production of "recycled" radiogenic parent nuclides. The latter explanation has to be taken into account especially because Réunion belongs to the DUPAL-anomaly.

Staudacher et. al., Chem. Geol., 89, 1-17, (1990).

Trieloff et. al., Geochim. Cosmochim. Acta, 61, 5065-5088, (1997).

Valbracht et al, Earth Planet. Sci. Lett, 150, 399-411, (1997).

Session E04:5P

E04 : 5P/01 : PO

Fluid and Mineral Micro-Inclusions in Cloudy Diamonds

Elad Izraeli (eladi@earth.es.huji.ac.il)¹,

Marcus Schrauder (A8611GAD@AWIUNI11.EDVZ.UniVie.AC.AT)² &

Oded Navon (oded@vms.huji.ac.il)¹

¹ Institute of Earth Sciences, The Hebrew University, Jerusalem, 91904, Israel

² Institute of Geochemistry, University of Vienna, Vienna, a-1090, Austria

It is well accepted that diamonds grow from fluids or melts, but fluid or melt inclusions are rarely found in diamonds. Most inclusions carry discrete mineral phases. Fluid inclusions were found in fibrous diamonds, but these diamonds are different from the common, single crystal, octahedral or dodecahedral diamonds. They are probably younger and represent unique growth conditions.

We report the finding of fluids in micro-inclusions that form cloudy regions within octahedral diamonds from Koffiefontein and Udachnaya. The clouds are ~1 mm in size, consist of millions of sub-micrometer inclusions and commonly occupy the central zone of the diamonds.

Infrared spectroscopy reveals the presence of water, carbonate, garnet, clinopyroxene, phlogopite, and an unidentified silicate, but no single diamond contains the entire assemblage. Careful analyses reveal no detectable hydrocarbons. When examined with a 300 µm aperture, some clouds reveal internal zoning. Electron probe analyses of individual inclusions reveal the presence of wide range of fluid compositions, discrete mineral inclusions, and mixtures of the two. The mineral inclusions belong to either the eclogitic suite: garnet, clinopyroxene and carbonate, or to the pridotitic one: olivine, Cr-rich pyrope, and phlogopite. Except for their microscopic size, the mineral inclusions have similar composition to that of the large discrete inclusions found in diamonds.

In eclogitic diamonds, we distinguish three types of fluids: 1) Silica-rich fluids broadly similar to those trapped in fibrous diamonds. 2) Fluids rich in K, Cl and water. 3) Fluids rich in carbonate and water. Peridotitic diamonds carry only fluids rich in K, Cl and water. High concentrations of K, Cl, Ba and Sr were detected in the inclusions of a peridotitic diamond from Udachnaya.

The unique feature of cloudy diamonds is the presence of both mineral and fluid inclusions in individual diamonds, sometimes in a single inclusion. This strongly suggests that at least some octahedral diamonds, which grew in either eclogitic or peridotitic host rocks, were formed in equilibrium with carbonate- and water-rich fluids.

E04 : 5P/02 : PO

Raman Barometry of Diamonds

Elad Izraeli (eladi@earth.es.huji.ac.il)¹,

Jw Harris (jwh@geology.gla.ac.uk)² &

Oded Navon (oded@vms.huji.ac.il)¹

¹ Institute of Earth Sciences, The Hebrew University, Jerusalem, 91904, Israel

² Department of Applied Geology, University of Glasgow, Glasgow, G12 8QQ, Scotland

Pressures and temperatures of the diamond source region are commonly estimated using chemical equilibria between coexisting mineral inclusions. Such determinations are destructive, possible only for diamonds with bi- or poly-mineralic inclusion assemblage and allow estimation of pressure only if orthopyroxene or spinel is present.

Source conditions can also be inferred from the internal pressure within diamond inclusions and from the deformation of the diamond lattice around them (Rosenfeld and Chase, 1961). Upon transition to the earth's surface, both the inclusion and the diamond expand as pressure is released and contract with falling temperatures. Since the bulk modulus (K) and thermal expansion coefficient (<alpha>) of the inclusion are different from that of the diamond, the expected volume of the inclusion at the surface is different from that of its hole (which behaves as the diamond itself). If the final volume of the inclusion is smaller than the hole in the diamond, then the inclusion is under no pressure. If instead, the final volume of the inclusion is larger than the hole, then the inclusion is under compression and the diamond surrounding it deforms elastically. We used Raman spectroscopy to measure the internal pressure of olivine inclusions. In ON-UDC-201 a single olivine inclusion is under pressure of 0.72 GPa. The average pressure of 13 inclusions in ON-UDC-203 is 0.36 GPa. Using K and a of diamond and olivine and the above pressures, we calculated the pressure during entrapment of the inclusion as a function of source temperature. In the case of ON-UDC-201, temperature can be estimated from the aggregation state of nitrogen (83% in B form). If the diamond is older than 100 Ma, calculated source temperatures are 1100-1200°C (Taylor at al., 1990). The corresponding pressure at the source is 5.8 GPa (using K_olivine of Matsui and Manghnani, 1985) or 4.5 GPa (K_olivine of Gillet et al., 1991). These conditions are within the diamond stability field and are in general agreement with previous P-T estimations based on chemical equilibria. If the diamond is younger, calculated temperatures and corresponding pressures are higher. No nitrogen was detected in ON-UDC-203, so source temperature could not be estimated.

We demonstrated that it is possible to estimate source region pressure and temperature for a single, old, nitrogen-bearing diamond that carries an olivine inclusion. Such estimation is non-destructive as it involves only IR and Raman spectroscopy. Other inclusions may be used as well when accurate data for their K and a are available. In diamonds with more then one inclusion pressure and temperature can be determined using Raman barometry and the chemistry of the inclusions.

Rosenfeld JL & Chase AB, Amer. J. Sci, 259, 519-541, (1961).

Gillet P, Richet P, Guyot F & Fiquet G, J. Geophys. Res., 96, 11805-11816, (1990).

Matsui T & Manghnani MH, Phys. Chem. Mineral, 12, 201-210, (1885).

E04 : 5P/03 : PO

The Isotope Composition of Metamorphic Fluids of Graphite Occurrence of Lapland Granulite Belt

Anaid A. Avedisyan (root@ksc-gi.murmansk.su),

Nikolai Ye. Kozlov (kozlov@ksc-gi.murmansk.su),

Valevtin A. Pripachkin (root@ksc-gi.murmansk.su),

Serafim V. Ikorskiy (root@ksc-gi.murmansk.su) &

Igor L. Kamenskiy (root@ksc-gi.murmansk.su)

Geological Inst., Kola Science Center RAS, Apatity, Russia

A study of the granulite stage in the development of the Kola Collision zone and the origin of lower crustal granulites, is one of the aspects of origin of ancient active zones of the Earth, and stimulates the search for new approaches in an attempt to reveal the primary nature of the Lapland Granulite Belt (LGB). At present, opinions differ as to the genesis of the Lapland granulites, which are interpreted to be either intrusive, or volcanic-sedimentary rocks, or a tectonically uplifted fragment of the granulite-basic crust layer.

A significant influx of CO₂ from the lower crust, resulting in the formation of high-density CO₂ inclusions at the peak of granulite metamorphism, has been recently considered (Newton et al., 1980) as a mechanism responsible for the formation of "dry" granulite complexes. However, this mechanism fails to explain the origin of reduced gases in products of deep degassing. According to the most reliable marks, i.e. helium and carbon isotopes, it is unlikely that large masses of reduced gases could have come from the mantle. This study is based on the data on graphite contained in Late Archean rocks of the Kola Peninsula, where sheet-like graphite-bearing bodies are abundant, occurring commonly at the contact between granulite and anorthosite. To ascertain whether the carbon and hydrocarbon compounds are of mantle or metamorphic origin, the authors employed data on carbon and helium isotopes. There is no general agreement about the origin of graphite in metamorphic rocks, although many researchers hold to the idea that it has a biogenic nature (Weis et al., 1981; Yaroschuk et al, 1993). At the same time, the location of graphite bodies close to tectonic faults suggests that large masses of fluids might arrive from under-crustal levels.

Carbon isotope studies show that ¹³C of graphite from kyanite-graphite-garnet rock varies within narrow limits, from -18.4 to 21.3, typical of the organic matter of sedimentary rocks. However, carbon isotope data can be interpreted differently, as they fall within the interval of intermediate ¹³C values for carbon bearing (average ¹³C 5-10%) and ordinary (¹³C from -20 to -30%) chondrites. To evaluate the contribution of deep fluids more confidently, the authors used data on helium isotopes. ³He/⁴He ratios in the same samples are within the limits (8.4-9.1x10^-8) typical of crustal rocks, indicating a negligible involvement of mantle fluids in the formation of graphite occurrences.

Newton R S, Smith J V & Windley B F, Nature, 288, 45-50, (1980).

Weis P, Friedman I & Gleason J D, Geoc. et Cosm. Acta, v.45, 12, 2325-2332, (1981).

Yaroschuk M A, Savchenko L T, Kramskikh E P, Geologitcheskiy J, Naukova Dumka, N2, (1993).

E04 : 5P/04 : PO

Pegmatoid Pyroxenite Dike Complex of West Kieway Lobe of the West Panski Layered Intrusion: Implication for Fluid Dynamics of the PGE Reef

Rayssa Galimzyanova (tamara@geo.kolasc.net.ru),

Felix Mitrofanov &

Tamara Bayanova

Geological Institute of the Kola Science Centre, RAS, 14 Fersman St., 184200 Apatity, Russia

Feldspathic pyroxenite pegmatoid is one of the enigmatic rocks of the lower PGE-reef of the West Kieway lobe of the Pana Pt-bearing intrusion. The latter is situated in the north-eastern Baltic Shield between Early Proterozoic and Archean structures. Pegmatoid bodies (2470±9 Ma) occur among gabbronorites (2491±1,5 Ma) and spotted gabbro. They consist of 70% magnesian bronzite, 10% diopside and 10% plagioclase and are distinguished from resident rocks by F (<0,20), Al/Si(<0,11) and high contents of gases in the plagioclase. The latter are (in ml/g): H₂-0.28, CH₄-0.009, CO-0.10, CO₂<0.13, H₂O-2.21 (Orsoev et al., 96). The same contents of gases are typical of rocks occurring in a 30 cm thick interval of the vein walls. Our study was focused on field relationships between the pegmatoids and gabbroid cumulates. The testing showed that they have an intrusive origin and are discordant veins and thin layers (0,03-1,35 m thick) connected by feeders. They are located in a 150 m wide block of semiconsolidated cumulates which have a subsidence amplitude of up to 15 m relative to the roof of the reef. This block is overlain by trachytoid, but ore-free cumulates with xenolith of related rocks. We consider that the emplacement of PGE-bearing fluid-saturated magma took place during postcumulus stage at the time of the origin of the subsidence block. Local diffusion of fluid from these bodies to neighbouring cumulates was 0,30 m. These data were for the first time obtained for fluid dynamics of the West Panski PGE-reef.

This study was financially supported by Russian Foundation for Basic Research, grant 98-05-64321.

Orsoev DA, Konnikov EG, Zigankov AA & Kislov EV, DAN, 347, 228-231, (1996).

E04 : 5P/05 : PO

Effect of Mantle Fluids in Ore Deposition in Black Shales

Vladimir Rusinov (rusinov@igem.msk.su)¹ &

Olga Rusinova²

¹ Inst.Geol.Ore Deposits RAS, Staromonetnyi, 35, Moscow, Russia

² Central Sci. Geol.Prospecting Inst., Varshavskoe shosse, 129-b, Moscow, Russia

Mantle fluids consist mainly of C-O-H-compounds and supply the crust with Au, Pt, Cr, V, Te, Hg and some other metals within global or regional geodynamically active zones. These fluids are an important agents affecting redox conditions (fO₂) of crustal ore deposition and of the soluble ore metal complexes stability. Fluid conduits penetrate into the crust along long active tectonic structures that join upper mantle with higher levels. Three marginal types of such regional structures occur (in order of the fO₂ upper limit increasing): 1) straight deeply penetrating tectonic zones, reaching the mantle (regional shear zones, continental rifts, etc.) - it allow quick lifting of the mantle fluids without significant fluid-rock interaction; 2) series of disconnected vertical faults, interrupted with magmatic chambers, initiating fluid-rock interaction and transformation of the mantle fluids into postmagmatic solutions; and 3) volcanic structures, where fluids are mixed with meteoric waters. Black shales relate commonly to the regional shear zones, and the environments of Au, Pt deposition there correspond to the first structural type.

So the reduced conditions (fO₂ less than QFM buffer) indicate the predominance of mantle fluids in black shales ore deposition. The ¹⁸O of the fluid correlates with its redox conditions and with the fraction of mantle fluids in ore solution. The reduced mantle fluids depositing gold in black shales and shear zones like as at the Muruntau deposit (Uzbekistan) have ¹⁸O =10 - 15 ppm (near to the fluid from mantle spinel lehrzolite), in contrary with oxidized fluids of the epithermal deposits (¹⁸O < 8-10 ppm). The type of wall rock alteration also depends on the redox parameter of the ore fluids. At the Muruntau deposit the most abondant wall rock alteration are biotitization, local albitization, K-feldspatization, and sericitization (with minor quartz). There are no products of acidic metasomatism as at the Muruntau deposit as at the other black shale deposits. The fCO₂ affecting fluid acidity there depends on the mole fraction of CO₂ in fluid and on the total pressure. At the deep levels of shear zones (10 - 5 km depth) and in black shales reduced conditions provide low CO₂/(CO₂ + CH₄) ratio and low CO₂ fugacity. So this is the reason of neutral to slightly alkaline reaction of the ore solutions.

The study is supported by the Integration State Program, 2-1, Project k528.

E04 : 5P/06 : PO

The Water Content of Trachytic to Trachyphonolitic Glasses and Related Sanidine Phenocrysts from the Eruption of the Breccia Museo (Campi Flegrei, Italy): An IR Spectroscopy Study

Giuseppina Balassone (balasson@unina.it)¹,

Carmine Amalfitano (evidente@unina.it)¹,

Anton Beran (anton.beran@univie.ac.at)²,

Leone Melluso (melluso@unina.it)¹,

Vincenzo Morra (vimorra@unina.it)¹,

Annamaria Perrotta (scarpati@int0828.it)²,

Gennaro Ricci (vimorra@unina.it)¹ &

Claudio Scarpati (scarpati@int0828.it)²

¹ Dipartimento di Scienze della Terra - Università di Napoli Federico II - I-80134 Napoli, Italy

² Institut Mineralogie Kristallographie - Geozentrum - Universitat Wien A-1090 Wien, Austria

The Breccia Museo Member (BMM) is a pyroclastic deposit formed by an explosive eruption that occurred in the SW sector of Campi Flegrei area (southern Italy) about 20 ka ago (Perrotta and Scarpati, 1994; Melluso et al., 1995). Campi Flegrei is an active volcanic complex formed by monogenic volcanoes and a caldera depression. The eruptive sequence of BMM consists of the Lower Pumice Flow Unit and the overlying Upper Pumice Flow Unit with its associated lithic Breccia Unit. Interlayered with Breccia Unit is the Spatter Unit. Chemical composition of products range from trachytic to trachyphonolitic, and the phenocryst assemblage is dominated by sanidine (Or_88-86 to Or_67-63), with minor amounts of Na-plagioclase, salitic clinopyroxene, biotite, apatite and opaque oxides. Infrared spectroscopic analyses of volcanic glasses sampled from the different Units of BMM were carried out in order to determine the amount of water dissolved in these tephra and in sanidine crystals. It is know that exolution of dissolved H₂O during magma decompression provide a major driving force for subaereal explosive volcanism and the determination of the volatile behaviour can be used to construct quantitative model of magmatic systems and eruptive processes (Bursik, 1993; Carrol and Blank, 1997). IR spectroscopy provide a sensitive tool for the detection of trace amounts of H₂O molecule or hydroxyl groups in minerals, glasses and melts (Beran, 1986; Libowitzky and Rossman, 1997). Double polished glass samples (obsidian, pumice, spatter free of inclusions, bubbles and devitrified zones) and (010) sanidine plates were used for FTIR investigations (for sanidine polarized IR spectra were also obtained). For each spectrum, peaks were measured at the band of interst, mostly in the OH stretching mode of mid-IR (centred at 3550 cm^-1 for glasses and 3400 cm^-1 for sanidine). The concentration of water was determined using a modification of the Beer-Lambert law: C=18.02.Abs/<epsilon-rho>d (C=water concentration in weight fraction, Abs=absorbance at the band of interest in absorbance units, <epsilon>=molar absorptivity at the band of interest in litre mole^-1 cm^-1, <rho>=density in g/l). Preliminary data show that the water contents of unaltered glass samples are mostly in the range 1.20-2.90 wt%, while the water amount detected in the sanidine crystals is generally lower than 0.10 wt%. In some case, a correlation between high-H₂O glassy matrices and sanidines can be pointed out.

Beran A, Phys. Chem. Min, 13, 306-310, (1986).

Bursik M, J. Volc. Geoth. Res, 57, 57-70, (1993).

Carrol MR & Blank JG, Am. Min, 82, 549-556, (1997).

Libowitzky E & Rossman GR, Am. Min, 82, 1111-1115, (1997).

Melluso L, Morra V, Perrotta A, Scarpati C & Adabbo M, J. Volc. Geoth. Res, 68, 325-339, (1995).

Perrotta A & Scarpati C, J. Volc. Geoth. Res, 59, 335-355, (1994).

E04 : 5P/07 : PO

The Model of Fluid Movement in the Crust Waveguides and its Geophysical Application

Andrej Karakin (karakin@glas.apc.org)¹ &

Yury Kuznetsov (karakin@glas.apc.org)²

¹ VNIIGeosystem Inst., Varshavskoe shosse 8, 113105 Moscow, Russia,

² Joint-Stock Research Industrial Company GERS Apt. 62, b.4, 2-nd Pereulok Truzhenilov, 119121, Moscow, G-121, Russia

Introduction. The analysis of geoelectric data (Feldman, 1976), seismic cross-sections (Sharov, 1997) and the results of the study of the deep and superdeep boreholes (The Kola superdeep borehole, 1984) show that continental crust is divided with many cracked layers. Many specialists think that this kind of structure is typical for the continental crust and take place in very often cases. The most popular hypothesis is that the waveguide is a fracture-porous media filled with the mineralized water.

Waveguide model. Some mechanism of fluid movement able to produce self-exited oscillation regime providing the state of dynamical equilibrium is proposed (Karakin, 1990). The model describes two layer-system. The moving forces are tectonic horizontal stresses. The upper layer is elastic. There are two regimes in the lower, second layer: compaction and dilatation competing with themselves. Their interaction consists in some oscillate or wave process with both regimes working in turn. Stress-strain fields in the system plus fluid motion in the lower layer are calculated. Maybe this mechanism is mainly responsible for stress state and fluid regime in the continental crust during geological times corresponding waveguide size scale. This mechanism is important for understanding human-life valuable processes: earthquakes and hydrocarbons transportation from mother-rocks to the deposits. Knowledge of fluid movements during geological time is also important if we are looking for nuclear waste places.

Superdeep drilling proposal on the basis of this model. Superdeep boreholes can help to know: 1. Whether there are the crust waveguides really? 2. What is their fluid regime? 3. The parameters of the fluid and temperature regime.

The matter of question is to determine a full set of the constitutive parameters and their values. It is desirable to compare this model with the alternative analogous models (the model of thermal convection in porous media for example) for selection of the best model. Recently the modern tectono-physical and physical methods were worked out to investigate the stress state and fluid movement. These methods are grounded on the tectonic jointing investigation and neutronographic analysis. As a rule they can determine the axes of principal recent and paleostresses, and the fluid velocity vector. They can be use to identify a model.

Feldman I S, Conf. Geoelec. and Geotherm. Stud. KAPG Geophys. Monogr. Bp., 721-745, (1976).

Karakin A V, Matematicheskoe modelirovanie (Mathematical modelling), 2 (3), 31 - 42, (1990).

E04 : 5P/08 : PO

Evaluating Crustal Contamination of Lesser Antilles Island Arc Lavas by a Study of ⁸⁷Sr/⁸⁶Sr and ¹⁴³Nd/¹⁴⁴Nd from Olivine and Pyroxene Phenocrysts

Matthijs C. van Soest (soet@geo.vu.nl),

Melanie Griselin (grim@geo.vu.nl) &

Tim Elliott (ellt@geo.vu.nl)

Faculteit der Aardwetenschappen, Vrije Universiteit Amsterdam, De Boelelaan 1085, 1081 HV, Amsterdam, The Netherlands

The Lesser Antilles island arc is one of the few oceanic arcs with sample suites that clearly show evidence of crustal contamination (CC). Nevertheless, there also is compelling evidence for a large subducted component (SC) in the lavas throughout the arc, and it remains difficult to distinguish between respective contributions of CC and SC in many cases.

In an attempt to address this problem we have undertaken a Sr and Nd isotope study of olivine and clinopyroxene separates and corresponding whole-rock basalt and andesite samples from the Lesser Antilles, which have been well characterised for evidence of CC by a study of their helium isotopes (van Soest and Hilton, this volume). It is observed that lavas show clear evidence of CC south of the island of Martinique, with (³He/⁴He) clearly lower than mantle ratios that would be expected. North of Martinique there is little or no evidence for CC in the helium isotope data. This study allows us to make observations concerning the isotopic evolution of the magma after the onset of olivine crystallisation and put constraints on timing of CC.

Nd isotopic ratios were measured utilising the NdO technique. The chemical separation technique required for this method was carried out according to new procedures developed at the VU. Sr concentrations in olivine obtained by ID measurements range from 1.3 to 21.7 ppm. Nd concentrations determined by ICP-MS in two olivine samples are 0.22 and 0.36 ppm. These concentrations are too high to be attributed solely to olivine so we are presumably looking at Sr and Nd from melt inclusions trapped at the moment of crystallisation.

Results obtained so far for Sr on 30 olivine and pyroxene samples and for Nd on 6 olivine samples show equilibrium between most mineral separates and whole rocks for both Sr and Nd. In some cases ⁸⁷Sr/⁸⁶Sr ratios of olivines are higher (up to 2 units in the 3rd decimal) than whole rock ratios, although we believe this is probably due to problems resulting from fine scale alteration along cracks within otherwise very fresh mineral grains. In those samples we have checked, leaching of the olivine reduces the difference between whole- rock and olivine. This emphasises the importance of the Nd results which are less susceptible to alteration or diagenesis which is difficult to detect visually. The presence of whole-rock olivine equilibrium in Sr and Nd isotopes, is irrespective of position in the arc and helium isotope results. This implies that for those samples, that show evidence for CC in helium isotope results, CC must have happened before the onset of crystallisation of olivine.

E04 : 5P/09 : PO

Crustal Dynamics and Advective Heat Transport in Continental Rift Zones Indicated by Thermal and Helium Tracers

Christoph Clauser (c.clauser@gga-hannover.de)¹,

Erika Griesshaber (e.griess@uni-tuebingen.de)² &

Horst J. Neugebauer (neugb@geo.uni-bonn.de)³

¹ Joint Geoscientific Research Institute (GGA), Stilleweg 2, D-30655 Hannover, Germany

² Angewandte Geologie, Geologisches Institut, Universität Tübingen, Sigwartstr. 10, D-72076 Tübingen, Germany

³ Geodynamics - Physics of the Lithosphere, Bonn University, Nussalle 8, D-53115 Bonn, Germany

Continental rift zones are characterized by increased permeability due to crustal extension and brittle block faulting of their basement. Numerical simulations of groundwater and heat flow in a two-dimensional type-model of the upper crust show that the dominant mode of heat transport changes from conduction to advection within the permeability range 5*10^-18 m² - 3*10^-17 m², values typically reported for crystalline rocks. Depending on the host rock permeability, faults and fracture zones may either serve as preferred conduits for pervasive infiltration or free convection systems may develop within these structures. Three characteristic patterns evolve in the temperature and flow fields, depending on the relative strength of forced and free convection. Inspection of a real continental rift zone, the Rhine graben, shows that these patterns cannot be distinguished on the basis of thermal data alone. The ³He/⁴He-ratio of the stable helium isotopes provides an additional, independent tracer for transport. Mantle helium varies from 15-25% around the Miocene Kaiserstuhl volcano to less than 3% close to the thermal anomalies of Landau (Germany) and Soultz sous Forêts (France). Finally, a combination of numerical simulations and a joint interpretation of thermal, helium isotope, and hydochemical data shows that: (1) The thermal anomalies of the Rhine graben are caused by an E-W groundwater flow system and not directly linked to a mantle intrusion; (2) The ³He/⁴He anomaly around the Kaiserstuhl volcano is not accompanied by a thermal anomaly. The present-day helium signature there evidently trails behind the prior thermal signal and is not accompanied by thermally relevant mass flow rates.

E04 : 5P/10 : PO

Noble Gas Investigations of Ancient Subcontinental Mantle Samples (kimberlites and Lamproites) from the Eastern Baltic Shield

Thomas Wiersberg (wiers@gfz-potsdam.de)¹,

Samuel Niedermann (nied@gfz-potsdam.de)¹,

Jörg Erzinger (erz@gfz-potsdam.de)¹,

Lev K. Levsky (lev@ad.iggp.ras.spb.ru)² &

Kirill I. Lokhov (kirill@KL1407.spb.edu)²

¹ GeoForschungsZentrum Potsdam, Department 4.2, Telegrafenberg, D-14473 Potsdam, Germany

² Institute for Precambrian Geology and Geochronology, Russian Academy of Sciences, Makarova emb. 2, RUS-199034 St. Petersburg, Russia

Kimberlites and lamproites may provide undisturbed compositional information about the ancient subcontinental mantle at the time of their genesis, as their rapid ascent through the crust renders them very resistant to crustal contamination. We investigated lamproites from Kostamuksha (Karelian craton) and kimberlites and lamproites from Poria Guba (Kola craton); both cratons were formed before the late Archean (2.7 Ga). Rb-Sr and Sm-Nd dating of whole rock samples yield a crystallisation age of 1.24 Ga for Kostamuksha lamproites (Belyatskii et al, 1997). K-Ar dating of micas provides the same age, whereas the K and radiogenic ⁴⁰Ar concentrations of whole rock samples would imply an age of up to 2,000 Ma. The obvious excess of radiogenic ⁴⁰Ar in whole rock samples could be due to complete degassing of the micas at the time of eruption, while other minerals (e.g. olivines and pyroxenes) retained excess argon. Further investigations on mineral separates are in progress to confirm this.

Fission-produced ¹³²Xe, ¹³⁴Xe and ¹³⁶Xe in whole rock samples could be used for dating purposes also, but calculated ages, assuming that all excess xenon is produced by ²³⁸U spontaneous fission are too high.

The neon isotopic composition reveals the presence of nucleogenic ²¹Ne and ²²Ne, produced by the reactions ¹⁸O(<alpha>,n)²¹Ne and ¹⁹F(<alpha>n)²²Na(ß⁺)²²N<epsilon>, respectively. When plotted in a three isotope diagram, with the exception of one sample all data points are consistent with either of two linear fits, representing simple two-component mixtures between atmospheric Ne and two nucleogenic endmembers with distinct (²¹Ne/²²Ne)n ratios. These components probably reflect two minerals with distinct O/F ratios. Using the empirical correlation between the (²¹Ne/²²Ne)_n production ratio and the ¹⁸O/¹⁹F ratio (cf. Kennedy et al., 1990), total O/F volume ratios of ~33 for (²¹Ne/²²Ne)_n = 0.16 and ~ 81 for (²¹Ne/²²Ne)_n = 0.40 are calculated.

In one kimberlite from Poria Guba we found a ³He/⁴He ratio of 0.28*10^-6 using the stepwise heating technique and 0.45*10^-6 by crushing the sample. Both values are clearly higher than the crustal ratio (~2*10^-8). Because cosmogenic ³He cannot be released by crushing and contamination by atmospheric ³He is negligible, the excess ³He must reflect a mantle component.

Belyatskii BV, Nikitana LP, Savaa EV, and Levsky LK, Geochem. Int, 35 (6), 575-579, (1997).

Kennedy BM, Hiyagon H & Reynolds JH, Earth Planet. Sci. Lett, 98, 277-286, (1990).

E04 : 5P/11 : PO

Mantle-Derived He and C End-Member for European Volcanism Identified in Natural Fluids of Pantelleria Island, Southern Italy

Francesco Parello (parello@unipa.it)¹,

Patrick Allard²,

Walter D'Alessandro³,

Cinzia Federico¹ &

Philippe Jean-Baptiste²

¹ CFTA, Università di Palermo, via Archirafi, 27, Palermo, Italy

² LMCE, CNRS-CEA, Orme de Merisiers, Gif/Yvette, France

³ IGF, CNR, via La Malfa, 153, Palermo, Italy

Helium is a trace rare gas commonly occurring in magmatic and volcanic gases. Together with carbon dioxide, helium rapidly partitions into the exsolved gas phase when magmas ascend towards the surface and both He and CO₂ remain closely associated during volcanic degassing. Reliable information upon their magmatic source can then be inferred from their chemical and isotopic ratios in surface fluids. Here we report new results on both He and carbon (plus H and O) isotopes in groundwaters and volcanic fluids of Pantelleria, a southernmost active volcano in Italy, providing the first detailed isotopic investigation of He on this island. This study has led to the following observations and conclusions: a) Thermal waters and gases contain mantle-derived helium and carbon with ultimate ³He/⁴He ratio of 7.3 Ra and ¹³C of -4.5‰. These values indicate a MORB-type mantle component feeding volcanoes in this distensive area. This deep gas interacts with both the host rocks and with groundwaters in the thermal aquifer, which occasionally fractionates helium from CO₂ and adds organic-type carbon, lowering both ¹³C and R/Ra ratio of the residual dissolved gas; b) Gaseous manifestations with highest R/Ra are located within the caldera, whereas, outside the caldera and along the coast, gases and thermal waters display lower R/Ra values. This feature agrees with a higher heat flow below the central part of the island; c) ¹³C of CO₂ and R/Ra ratios of gas emissions in Pantelleria plot into the field of MORB-type and related fluids basalts. They are consistent with Nd and Sr isotope ratios in lavas (Civetta et al., 1984; Esperança & Crisci, 1995), suggesting an E-MORB type mantle source beneath the area. It remains to account for the more variable Pb isotope ratios (Esperança & Crisci, 1995) from a depleted MORB-type mantle composition to a High-µ (HIMU) one.Hence, Pantelleria represents a new isotopic pole for both helium and carbon in the Mediterranea, representative for the parent source of volcanism in southern-western Europe. The respective northward decrease (R/Ra) and increase (¹³C) of these parameters indicate increasing crust-mantle interactions as one moves from the distensive regime in the Sicily Channel rift to the compressive and subduction context of the Aeolian Arc and volcanism in central Italy.

Civetta L., Cornette Y., Crisci G., Gillot P.Y., Orsi G. & Requeiros C.S., Geol. Magazine, 121, 541-562

Esperança S. & Crisci G.M., Earth Planet. Sci. Lett., 136, 167-182

E04 : 5P/12 : PO

Helium Isotopes in Lavas and Fumarolic Gas of the Lesser Antilles Arc

Anselmo Pedroni (pedroni@zedat.fu-berlin.de),

Konrad Hammerschmidt

(chemie@zedat.fu-berlin.de) &

Hans Friedrichsen (hfriedri@zedat.fu-berlin.de)

FU-Berlin, Geochemie, Malteserstrasse 74-100, 12249 Berlin, Germany

The ³He/⁴He ratio of fumarolic and vesicular gases has been investigated on summit fumaroles and lavas of four volcanoes of the Lesser Antilles Arc. The volcanoes are the Soufrière (Guadeloupe), the Micotrin Trois Pitons (Dominica) the Mont. Pelèe (Martinique) and the Soufrière (St.Vincent). A pattern common to all four volcanoes clearly emerges from this study. The highest air-corrected ³He/⁴He ratio, which is typically close to the MORB ratio (³He/⁴He = 8-8,5 Ra), is found for summit fumaroles. Similar or slightly lower values are found for vesicular gases in olivine, whereas distinctly lower values characterize vesicular gases in pyroxene (3-8 Ra) and plagioclase (2-3 Ra). The helium concentrations are highest in olivine (10^-6cm³ STP/g), 3 to 60 times lower in pyroxene and 40 to 750 times lower in plagioclase. The variations in helium isotope ratio may be attributed to pre-eruptive fractional degassing of cooling magma bodies. The concentrations suggest minimum degassing in the order of 99% during crystallization and, according to the Rayleigh destillation model, (e.g. Hilton et al., 1993), maximum ³He/⁴He ratios of 4 Ra are expected for plagioclase. High-level crustal assimilation and/or in situ accumulation of radiogenic helium (magma aging) may have contributed in lowering further the ³He/⁴He ratio in plagioclase whereas post-eruptive accumulation of radiogenic Helium is ruled out by the fact that most studied samples are historic or Recent. Summit fumaroles, with ³He/⁴He ratios even higher than vesicular gas in olivine, appear to tap the magma at very early stage of degassing. The former thus are the choice to gather the most pristine ³He/⁴He signature of the magma. Later degassing is either not channeled to summit fumaroles or overwhelmed by degassing of more juvenile sources or, due to higher magma viscosity, intermittent e.g. explosive.

Hilton D.R., Hammerschmidt K., Loock G. & Friedrichsen H., Geochim. Cosmochim. Acta, 57, 2819-2841, (1993).

E04 : 5P/13 : PO

Is the Neon Solubility in Silicate Melts Underestimated? Constraints from Noble Gas Measurements in Tektite

Daniele Pinti (pinti@geol.u-psud.fr)¹,

Takuya Matsumoto

(matsumoto@ess.sci.osaka-u.ac.jp)²,

Jun-ichi Matsuda

(matsuda@ess.sci.osaka-u.ac.jp)³ &

Zong Fang⁴

¹ Dept. Sciences de la Terre, Université Paris XI, 91405 Orsay Cedex, France

² Dept. of Earth and Space Science, Osaka University, Toyonaka, Osaka 560-0043, Japan

³ Dept. of Earth Sciences, Nanjiin University, Nanjiin, Popular Republic of China

Noble gas solubility in silicate melts is an important parameter to quantify mantle degassing and atmosphere formation, but it is still poorly constrained. Recent works have shown that noble gas solubility is dependent of multiple factors such as the pressure (Chamorro-Perez et al., 1998) and the degree of polymerization of silicate melts (Shibata et al., 1998) and it could be several orders of magnitude higher than that measured in earlier studies. We suggest that neon solubility in natural glasses could be 5-10 times higher than previously measured (Jambon et al., 1986; Lux, 1987). This conclusion arises from Ne measurements in tektite, a terrestrial silicate glass produced by the melting of surface rocks during the impact of meteorites. Previous measurements showed large amounts of atmospheric ²⁰Ne (ª1x10^-7 cm³STP/g) which cannot be explained by solubility equilibrium of the tektite with the atmosphere. Indeed the amount of ²⁰Ne dissolved in glass should be ª1-2x10^-8 cm³STP/g, by using solubility data from Lux (1987). Previous hypothesis suggested that small-size bubbles (<150 µm) concentrated on the surface of the tektite could be responsible for the storage of the ²⁰Ne excess (Matsubara & Matsuda, 1995). However, this hypothesis cannot explain the near constant ²⁰Ne concentration observed in tektites (mean = 1.0±0.3x10^-7 cm³STP/g on 23 literature data). Furthermore vesicularities up to 0.5%, needed to store the ²⁰Ne excesses, are rare in tektites (Müller & Gentner, 1968). This suggests that neon is dissolved in the glass. To test this hypothesis, we measured neon in five tektites from China. Initially we extracted the neon by pyrolysis at 1800°C. The results showed a near-constant amount of ²⁰Ne (0.95±0.13x10^-7 cm³STP/g) in agreement with previous studies. The ²⁰Ne has been successively extracted by crushing from a second set of fresh samples and the results showed a lower amount of ²⁰Ne (0.3-4x10^-8 cm³STP/g). For completeness we re-measured ²⁰Ne in a third set of samples, extracting the gas by crushing and then by pyrolysis. The results confirm that less than 3-40% of ²⁰Ne is extracted by crushing. To explain these results we should hypothesize i) that crushing is incomplete (and Ne resides in small-size vesicles as suggested by Matsubara & Matsuda, 1995), or ii) that most of Ne is residing in glass. This latter hypothesis suggests that ²⁰Ne solubility in silicates could be 5-10 times higher than previously supposed. Solubility dependence to the degree of polymerization of the silicate liquids and the occurrence of three-dimension Si units accomodating noble gases (Shibata et al., 1998) could explain the observed higher solubility of Ne in natural silicate glasses.

Chamorro-Perez E, Gillet P, Jambon A, Badro J & McMillan P, Nature, 393, 352-355, (1998).

Jambon A, Weber H & Braun O, Geochim. Cosmochim. Acta, 36, 401-408, (1986).

Lux G, Geochim. Cosmochim. Acta, 51, 1549-1560, (1987).

Matsubara K & Matsuda J, Geochem. J, 29, 293-300, (1995).

Muller O & Gentner W, Earth. Planet. Sci. Lett, 4, 406-410, (1968).

Shibata T, Takahashi E & Matsuda, J Geochim. Cosmochim. Acta, 62, 1241-1253, (1998).

E04 : 5P/14 : PO

¹³C Anomalies in the Archean and Proterozoic: Organic Anomaly or Mile Stone of Global Tectonic Reorganization

Dieter Rammlmair (dieter.rammlmair@bgr.de)

BGR, Stilleweg 2, 30655 Hannover, Germany

Since the Archean, the bulk of ¹³C of organic carbon ranges between -10 permil (rel. to PDB) and -40 permil (mean between -20 permil and -30 permil). Twice, in the Archean (Fortescue Group: approx. 2.7*10⁹ yr.) and in the Proterozoic (Francevillian Series: approx. 2.1*10⁹ yr.) anomalous low values reaching -50 permil and -45 permil, respectively, are reported (Schidlowski, 1988). A number of investigations has shown that such low values are widespread in the Archean (Strauss et al., 1993). Extremely low values are also reported from living communities around active brine seeps on the continental slopes at several locations (Paull et al., 1985). Carbon-rich sediments (0.1 to 3.5 wt.% TOC) from the Archean Sukumaland Greenstone Belt have been investigated in detail (Rammlmair et al., 1996). By means of microscopy, SEM, grain size fractionation, Raman spectroscopy, C-H-N-contents and pyrolysis first time two distinct carbon sources could be identified.The first carbon source is of a primary kerogen (¹³C approx. -25 permil; 0.1 to 0.3 wt.% TOC; grain size >0.1 to 200 µm) which suffered local metamorphic overprint. Carbon from the second source (¹³C approx. -60 permil; 0.1 to 3.5 wt.% TOC; grain size <0.1µm) was introduced post peak metamorphism in varying amounts, via active shear zones to the primary kerogen. The second carbon most probably derived from a thermogenic, highly fractionated methane. In the two-source-model, the past interpretation of anomalous low ¹³C bulk values is questionable. Were Archean methane-consuming organisms responsible for these low values or did methane-derived carbon move into sheared, more or less metamorphosed black shales and precipitate as minute particles? Both, the organic and the inorganic trapping of methane requires an at least locally higher methane flux in these time periods. This higher flux could be provided by deeply rooted mega-shear zones which do point towards major changes in the global tectonic style.

Paull CK, Jull AJT, Toolin LJ & Linick T, Nature, 317 (6039), 709-711, (1985).

Rammlmair D, van den Kerkhof AM, Faber E & Reutel Ch, V.M. Goldschmidt Conference, Heidelberg, J. Conf. Abs. 1, 495, (1996).

Schidlowski M, Nature, 333 (6171), 313-318, (1988).

Strauss H, DesMarais DJ, Hayes JM & Summons RE, The Proterozoic Lithosphere: A multidisciplinary Study: JW Schopf & C Klein, Cambridge Univ. Press, 117-127, (1992).