Journal of Conference Abstracts

Session O10:5A

O10 : 5A/01 : H1

Mixing in 3-D Convection and the Preservation of Mantle Heterogeneities: A Numerical Study of the Evolution of the Indian Ocean Mantle Domain

Joerg Schmalzl (joergs@uni-muenster.de)¹ &

Mark Rehkamper (markr@nwz.uni-muenster.de)²

¹ Institut fuer Geophysik, Universitaet Muenster, Corrensstr. 24, D-48149 Muenster, Germany

² Zentrallaboratorium fuer Geochronologie, Universitaet Muenster, Corrensstr. 24, D-48149 Muenster, Germany

Geochemical studies have firmly established that mid-ocean ridge basalts (MORB) from the Indian Ocean differ in isotopic and trace element composition from Atlantic and Pacific MORB, and these characteristics are generally thought to be the result of large-scale mantle contamination. The distinct geochemical signature of the Indian Ocean mantle domain has been interpreted to reflect contamination of a pristine Atlantic-Pacific type upper mantle by a recycled component consisting of altered oceanic crust with small amounts of sediment added. Lead isotopic systematics furthermore constrain the recycled component to be older than about 1.25 Gyr.

The >1.25 Gyr age inferred for the recycled contaminant requires that large-scale geochemical heterogeneities are not homogenized on a whole-Earth scale for at least this time span, whereas mixing within a single mantle domain is comparatively efficient. 2-D convection models, however, generally predict efficient homogenization of geochemical heterogeneities in the upper mantle between different cells of a convective regime within several hundred million years. This would appear to weaken the case for the creation of long-lived geochemical heterogeneities in the asthenosphere by simple recycling of crustal materials. Only recently, computational resources have become available for the investigation of mixing in three dimensional fully time-dependent flow as appropriate for the mantle of the Earth. The results of numerical experiments indicate that mixing in a 3-D system is dominated by processes operating at two different length scales. At the inner-cell scale heterogeneous material is rapidly homogenized, whereas cross-cell mixing operates at much larger time scales.

We will present the results of numerical experiments of time-dependent 3-D thermal convection carried out in a large aspect-ratio domain and address the question whether contamination of the upper mantle by a recycled component at 1-1.5 Gyr combined with slow cross-cell mixing can account for the unique geochemical signature of Indian Ocean MORB. To this end, we follow the dispersion of passive tracers that simulate the fate of the subducted lithospheric material. The tracers are characterized by Sr and Pb isotopic compositions and trace element concentrations that are distinct from the ambient mantle, such that the tracer distribution can be used to track the process of isotopic homogenization by convective mixing at different length and time scales.

O10 : 5A/02 : H1

Rhenium-Osmium Characteristics of Central Japan Arc Lavas

Sophie Alves (alves@ipgp.jussieu.fr),

Pierre Schiano &

Claude J. Allègre

Laboratoire de Géochimie et Cosmochimie, Institut de Physique du Globe, 4 Place Jussieu, France

In order to add constraints on the sources of arc volcanoes, we have measured Os isotopic ratios and Re and Os concentrations on 10 volcanic rocks from the Izu peninsula of Honshu island and from the Izu-Oshima island (Central Japan). In this region the volcanoes are built on oceanic crust and do not have radiogenic lead and strontium signatures, typical for crustal contamination. The rocks we analysed have variable degrees of differentiation from basalts to dacites but show no evidence of crustal contamination from trace elements and their ages range from 600 ka to recent time.

The Os concentrations of the samples are very variable, ranging from extremely low values (= 0.1 ppt) to 7.63 ppt; Re concentrations range from 131 to 727 ppt. An extremely large variation in the osmium isotopic composition is observed for the whole sample set, with ¹⁸⁷Os/¹⁸⁸Os ratios corrected for sample ages ranging from 0.139 to 0.690. These ratios are significantly higher than that of both MORB and OIB (e.g., Schiano et al, 1997; Roy-Barman and Allègre, 1995; Reisberg et al, 1993; Hauri and Hart, 1993), but they compare favourably with Os measurements we have previously obtained for arc lavas from Java island (Sunda arc), (Alves et al, submitted).

When comparing ¹⁸⁷Os/¹⁸⁸Os and 1/¹⁸⁸Os ratios, the samples display a well-defined positive linear correlation. This mixing relationship, also displayed by arc lavas from Java island, is likely to simply reflect a binary-mixing trend between a radiogenic end-member and the unradiogenic peridotitic upper mantle. However, correlations between Os and compatible elements (Ni) and differentiation indexes such as SiO₂, suggest that a fractional crystallisation process may explain part of the observations, although not the co-variations between the Os isotopic ratio, the Os content and the silica content.

Using geochemical tracers including trace and major elements and isotopes, these results will be discussed in terms of mixing relationships, differentiation and combined processes to explain the coherent behaviours between Os, major and trace elements.

P Schiano, J-LBirck and CJ Allègre, Earth Planet. Sci. Lett, 150, 363-381, (1997).

M Roy-Barman and CJ Allègre, Earth Planet. Sci. Lett, 129, 145-164, (1995).

L Reisberg, A Zindler, F Marcantonio, W White, D Wyman and B Weaver, Earth Planet. Sci. Lett, 120, 149-167, (1993).

EH Hauri and SR Hart, Earth Planet. Sci. Lett, 114, 353-371, (1993).

S Alves, P Schiano and CJ Allègre, submitted to Earth Planet. Sci. Lett.

O10 : 5A/03 : H1

Helium and Lead Isotope Geochemistry in the Azores Archipelago

Manuel Moreira (mmoreira@whoi.edu)¹,

Regis Doucelance (Doucelan@ipgp.jussieu.fr)²,

Bernard Dupre³,

Mark Kurz (mkurz@whoi.edu)¹ &

Claude Allegre²

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

² Laboratoire de Géochimie et Cosmochimie, URA CNRS 1758, Institut de Physique du Globe de Paris, Université Denis Diderot Paris 7, 4 Place Jussieu, 75252 Paris Cedex 05, France

³ Laboratoire de Geochimie, OMP, UMR-CNRS 5563, 38 rue des 36 ponts, 31400 Toulouse, France

We present new helium and lead isotopic data for basalts from the Azores archipelago (North Atlantic) which show that the Azores have ⁴He/³He ratios both higher and lower than MORB values. Good covariations of helium and lead isotopes are observed at the scale of the archipelago, and we propose the coexistence of two mantle components in the Azores, identified by Sao Miguel and Terceira island samples. The eastern part of Sao Miguel island displays radiogenic helium (⁴He/³He>140,000, R/Ra<5.1) and lead (20.00, 15.75 and 40.33 for ²⁰⁶Pb/²⁰⁴Pb, ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb). The ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb ratios for Sao Miguel are unusual for OIB. Terceira basalts contain relatively unradiogenic/primitive ⁴He/³He ratios, with a minimum value of 64,000 (R/Ra=11.3), and relatively high lead isotopic ratios (²⁰⁶Pb/²⁰⁴Pb=20.02, ²⁰⁷Pb/²⁰⁴Pb=15.64 and ²⁰⁸Pb/²⁰⁴Pb=39.35). We propose that the Terceira source has a composition produced by a mixing between recycled oceanic crust (high ²⁰⁶Pb/²⁰⁴Pb) and entrained lower mantle (high ³He) material. The Sao Miguel island isotopic signature may be due to sampling of very local (km-size) heterogeneity, located at relatively shallow depth. The preferred origin of this heterogeneity is the delamination of subcontinental lithosphere, which occured during rifting and opening of the north Atlantic in the Jurassic. The primitive helium ratios were also observed on the Mid Atlantic ridge at 38.5N, reflecting plume-ridge interaction, whereas radiogenic ratios (>100,000) were observed at latitude higher than 40N by Kurz et al. (1982) and may reflect the contamination of the upper mantle by the Sao Miguel component.

O10 : 5A/04 : H1

Determination of Li Contents and Isotopic Compositions in Various Oceanic Basalts by Ion Microprobe: Implication for the Li Geochemical Cycle

Sylvie Decitre (decitre@crpg.cnrs-nancy.fr) &

Etienne Deloule (deloule@crpg.cnrs-nancy.fr)

CRPG, 15 rue Notre-Dame des Pauvres, BP 20, France

Lithium concentration and isotopic composition have been measured on glass samples from N and E MORB(MAR, EPR and Red Sea), OIB (Galapagos) and BAB (Mariana and Manus Basin). Isotopic ratios were measured with the Cameca IMS1270 microprobe with a reproducibility of standard measurements of ¾ ±1‰. Concentrations were measured on the Cameca IMS3f microprobe using energy filtering, with a precision better than 0.5 ppm. Lithium concentrations range from 5 to 11 ppm in Manus basin, from 3 to 7 ppm in Mariana, from 3 to 7 ppm in Galapagos and from 4.5 to 5 ppm in MORB. Li contents correlate positively with K₂O and negatively with MgO, suggesting that the content variations are related to the fractionation process in the sources. The isotopic compositions measured on the MORB samples display only small variations, with ⁶Li values (⁶Li, relative to standard Lsvec) of -4.7 and -5.4 for NMORB (EPR and MAR respectively), -6.2 and -4.8 for EMORB (MAR and ReadSee), in agreement with results from (Chan et al, 1988, 1992). The Galapagos ⁶Li values on 8 different samples are quite constant (between -4.7 and -6.6) and in the same range as the MORB values. The Manus basin and Mariana values are both higher than MORB, between -1.5 and - 2.9 for Manus Basin and between -0.7 and -2.5 for Mariana. There is no correlation between Li content and ⁶Li, either among all the samples taken together, or within each locality. Therefore, no isotope fractionation is observed during the chemical fractionation process in the source, and the changes in ⁶Li are related to the initial composition of the source. For Mariana samples, Li isotope compositions are lighter than those of MORB with similar or lower Li contents, indicating that the source included high temperature altered oceanic crust. For Manus Basin samples, both the Li content and the ⁶Li values are higher than those of MORB or OIB, suggesting the addition of a light Li bearing component. The inverse correlation observed between ¹¹B (Chaussidon et Marty, 1995) and ⁶Li values suggests the addition of heavy Li during the assimilation of low temperature hydrothermaly altered crust.The similar content and isotopic composition of Li in MORB and OIB suggest that no Lithium with a sea water composition is subducted into the mantle. Indeed the high temperature altered component seems to be extracted in the back arc volcanism.

Chan LH & Edmond JM, Geochim. Cosmochim. Acta, 52, 1711-1717, (1988).

Chan LH, Edmond JM, Thompson G & Gillis K, Earth Planet Sci. Lett, 108, 151-160, (1992).

Chaussidon M & Marty B, Science, 269, 383-386, (1995).

O10 : 5A/05 : H1

Mg-Fe Partitioning between Plagioclase and Basaltic Liquids

Mario Aigner-Torres (mario@erdw.ethz.ch)¹,

Peter Ulmer (pulmer@erdw.ethz.ch)¹ &

Peter Kelemen (peterk@cliff.whoi.edu)²

¹ Institute of Mineralogy & Petrography, Swiss Federal Institute of Technology (ETHZ), Sonneggstrasse 5, Switzerland

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

Plagioclase is one of the most important minerals in partitioning studies, not only because of it "omnipresence" as a liquidus phase in various magmatic systems, varying from terrestrial and lunar rocks (Phinney, 1992), to some meteoritic magmas (Peters et al., 1994), but also for being a good candidate for retaining its original crystallized trace elements contents due to the slow character of CaAl-NaSi interchange.

Because trace element models are highly sensitive to the values chosen for the distribution coefficients; a poor understanding of the factors which control these values can lead to wholly erroneous interpretations. More rigorous investigations of relationships between intensive variables and trace element partition coefficients are needed.

Here we present a compilation of data of the compositions of plagioclase and coexisting glass from both natural and experimental samples. We also evaluate the role of composition, temperature and oxygen fugacity on the partitioning of Mg and Fe between plagioclase and basaltic liquids.

Plagioclase plays an enigmatic role in the differentiation of MORB and it is one of the dominant mineral in gabbros. It has approximately the same density as MORB liquids and potentially has a more complex fluid dynamic behavior during the crystallization of MORB than do olivines or pyroxenes. Because Mg behaves as a trace element in plagioclase, it may record the Mg concentration of the initial, parent liquid. Similarly, ferric and ferrous iron may have different partition coefficients, and the partition coefficient of total Fe can be used as an indicator of the redox state of the magma (Sato, 1989).

We attempt to establish a simple thermodynamic model for Mg and Fe partitioning between plagioclase and basaltic liquids. Finally, we consider some geochemical applications.

Peters MT, Shaffer E, Burnett DS & Kim, SS, Geochim. Cosmochim. Acta, 59, 2785-2796, (1994).

Phinney WC, Geochim. Cosmochim. Acta, 56, 1885-1895, (1992).

Sato H, Proc. Ocean Drilling Program, Scientific Results, 111, 17-26, (1989).

O10 : 5A/06 : H1

The Hawaiian Plume Through the Pb Isotope Looking-Glass

W. Abouchami (wafa@mpch-mainz.mpg.de),

A. W. Hofmann (hofmann@mpch-mainz.mpg.de) &

S. J. G. Galer (sjg@mpch-mainz.mpg.de)

Max Planck Institut für Chemie, Postfach 3060, Mainz, Germany

The Hawaiian plume produces basalts with a large range of isotopiccompositions which have been modeled as mixtures of two or three components (Chen and Frey, 1983). The three partially independent radiogenic isotope systems of lead offer the unique opportunity to unravel mixing processes and separate endmembers - binary mixtures form straight linear arrays in Pb isotope space and deviations from straight lines require the existence of at least one additional component. But resolution of such binary mixing arrays does depend on the accuracy with which Pb isotopes can be measured. Here we report high precision Pb isotopic data (2<sigma>_ext < 100 ppm) obtained with a Pb triple spike, together with Sr and Nd isotopes, on lavas from Mauna Loa, Mauna Kea, Kohala, Lanai and Loihi. Pb isotope ratios are the lowest in Lanai and highest in Mauna Kea. Mauna Loa lavas have higher ²⁰⁸Pb/²⁰⁴Pb ratios than those from Kohala for comparable ²⁰⁶Pb/²⁰⁴Pb ratios. In Pb isotope space, samples from each volcano exhibit distinct and well-defined linear arrays, reflecting binary mixing within each volcano. However, the ²⁰⁷Pb/²⁰⁴Pb-²⁰⁶Pb/²⁰⁴Pb mixing lines of individual volcanoes are nearly parallel to one another, with a slope corresponding to an apparent age of ~2 Ga, suggesting that the true endmember compositions are far more extreme than the "average" signal observed in any of the lavas themselves. By contrast, the slopes defined by each volcano are quite distinct in ²⁰⁸Pb/²⁰⁴Pb vs. ²⁰⁶Pb/²⁰⁴Pb space. These Pb mixing relationships show that (1) the endmember components involved in each volcano are isotopically heterogeneous, and (2) that far more than three components are required for explaining the Pb isotope systematics alone. The three mixing components (Loihi, Kea, Koolau) defined by Eiler etal. (1996) using a principal component analysis are therefore insufficient to explain the sources of Hawaiian lavas - for example, the "Kea component" defined in that analysis cannot be a mixing endmember in any of the Mauna Kea lavas. Furthermore, the existence of the "Kea" and "Loa" trends should be reconsidered given that the Pb isotopic characteristics of Kohala, a member of the "Kea" trend, are quite different from those of Mauna Kea. We suggest that the Hawaiian plume is a heterogeneous mixture of several residual mantle and recycled crustal components. Within the relatively short time span of eruption for each of the suites analysed, there always seems to be an intimate mixture of a high ²⁰⁶Pb/²⁰⁴Pb and a low ²⁰⁶Pb/²⁰⁴Pb component, which are being coherently sampled in each individual volcano. Over the entire Hawaiian chain, however, these endmember components appear to be more heterogeneous.

O10 : 5A/09 : H1

Li Abundances in Eclogite Xenoliths from Orogenic Eclogites ­ Derivation from Subducted Oceanic Crust or Mantle Origin?

H. -M. Seitz

(mseitz@classic.min.uni-heidelberg.de) &

A. B. Woodland

(alan@classic.min.uni-heidelberg.de)

Mineralogisches Institut, Universität Heidelberg, D-69120 Heidelberg, Germany

Recent study of mantle peridotites and pyroxenites has demonstrated the utility of Li as a tracer of mantle processes (Seitz & Woodland 1998). This is due to the fact that the partitioning of Li between mantle minerals is effectively independent of temperature, pressure, and bulk chemistry. On this basis, we have measured the Li-contents in clinopyroxene (cpx) and garnet (gt) from eclogites in an attempt to characterise their origin. It is currently debated whether eclogites represent the high pressure equivalents of subducted oceanic crust or whether they can form by other means in the mantle (e.g. Jacob et al. 1994; Snyder et al. 1997). Two types of eclogites have been investigated: (1) eclogite xenoliths from several kimberlite localities, including Roberts Victor (South Africa), Obnazhennaya (Siberia) and the Eastern Finland Kimberlite Province, with some being diamondiferous; (2) orogenic eclogites from the Muenchberg Massif and Meidling, (Bohemian Massif) and occurrences in the central Alps (Alpe Arami, Trescolmen). Major element compositions of the xenolith samples yield gt-cpx temperatures (Krogh, 1988) in the range of 1300-1400°C, at 6.0 GPa for the Finnish eclogites, ~950 at 5 GPa for the Siberian and 1030°C and P(min) of 5 GPa for a diamondiferous eclogite from Roberts Victor. With the exception of one sample (Meidling), eclogites from the European Massifs record considerably lower temperatures of ~ 580-780°C at estimated pressures of 2-3 GPa. The eclogite from Meidling, which is believed to be of mantle origin, records conditions of =1250°C at 3-4 GPa.

The Li abundances in cpx and gt were determined by SIMS. Cpx is the dominant host for Li where as gt and other phases contain much less Li (max of 500 ppb in gt). Li contents in cpx are highly variable between samples from 370 ppb in a xenolith (Roberts Victor) to 75 ppm in an eclogite from the Muenchberg Massif. The xenoliths display only very limited inter-grain variation, where as the Li contents in orogenic eclogites can be much more variable. Two distinct groups can be defined: (1) Cpx in xenoliths and the eclogite from Meidling have Li abundances of 500-1350 ppb; (2) Cpx in orogenic eclogites always have >=10 ppm Li. This factor of 10 difference in Li content in cpx makes it unlikely that these two groups had the same protolith. The higher Li contents in the orogenic eclogites are consistent with an origin from subducted oceanic crust (Li in MORB is =10 times higher than in mantle perdidotites). We suggest that the much lower Li contents in group (1) are indicative of a mantle origin.

Krogh EJ, Contrib. Mineral Petrol, 99, 44-48, (1988).

Jacob D, Jagoutz E, Lowery D, Mattey D & Kurdrjavtseva G, Geochim. Cosmochim. Acta, 58, 5191-5207, (1994).

Seitz HM & Woodland AB, Proc. 7th. Int. Kimberlite. Conf (extended abstracts), Cape Town, S. Africa, 778-780, (1998).

O10 : 5A/10 : H1

Variable Oxygen Isotopic Compositions in Eclogites From Kimberlites: Could they Trace a Mantle Process?

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

K. Stephanus Viljoen (fviljoen@mhs7.tns.co.za)²,

Jochen Hoefs (jhoefs@gwdg.de)¹ &

Gerhard Wörner (gwoerne@gwdg.de)¹

¹ Geochemisches Institut, Universität Göttingen, Goldschmidtstr. 1, 37077 Göttingen, Germany

² De Beers Geosciences Center, P.O. Box 82232, South Vale, 2135 Johannesburg, South Africa

Deep-seated eclogite xenoliths found in kimberlites, including diamond bearing samples, are generally described as remnant subducted oceanic crust (eg Jacob et al., 1994). In such a model, oxygen stable isotopic composition is the central argument. Eclogites cover a large range of ¹⁸O-values, from +2 to +9 per mil, a range similar to that produced during low-temperature hydrothermal alteration. It is generally believed that no mantle process could produce such an isotopic range, as a consequence of the small O-isotopes fractionation factors at mantle temperatures. However, a simple subduction-related model fails to explain the overwhelming abundance of eclogites with high ¹⁸O-values (> +6.0 per mil) relative to low ¹⁸O-eclogites (< +5.0 per mil). The identification, therefore, of any mantle-process that could produce such anomalous ¹⁸O-values is an important and necessary prerequisite before validating the recycling model. Moreover, it must be emphasised that available data are very scarce, only few data from a single location being available.

In order to better constrain oxygen isotopic variability in eclogites and processes that control it, we selected a series of 100 eclogites from Orapa (Botswana), a mine known to display samples covering a great range of ¹⁸O-values (Deines et al., 1991). The sample set is composed of eclogites belonging to Group 1 (including diamond-bearing and non-diamond bearing rocks) and Group 2 (ie non-diamond bearing). Oxygen isotopes of both clinopyroxenes and garnets are measured, after petrographic, major and trace elements characterisations, by mass-spectrometry after extraction with an UV-laser ablation, a technique allowing intra-sample heterogeneity and isotopic equilibrium between minerals to be precisely investigated. This study is still in progress and we will present first results at the conference.

Deines P, et al., GCA, 55, 515-524, (1991).

Jacob D, et al., GCA, 58, 5191-5207, (1994).

O10 : 5A/11 : H1

K-Rich Clinopyroxenes as Mantle Conveyors of Crustal Components

Luc André (lucandre@africamuseum.be)¹,

Vladislav Shatsky (shatsky@mineralogy.uiggm.nsc.ru)²,

Katrien De Corte (kdecorte@africamuseum.be)³,

Nikolai Sobolev (sobolev@mineralogy.uiggm.nsc.ru)² &

Jacques Navez¹

¹ Department of Geology and Mineralogy, Royal Museum for Central Africa, Leuvensesteenweg, 13, B-3080, Tervuren, Belgium

² United Institute of Geology, Geophysics and Mineralogy, Novosibirsk, 630090, Russia

³ Department of Geology and Soil Science, University of Ghent, Krijgslaan, 281, B-9000 ghent, Belgium

Recycling of continental derived materials within the mantle has now found wide acceptance because it contributes to explain the isotopic and chemical diversity of Ocean Island Basalts and potassic intraplate igneous rocks. The nature of the mineral phases that convey these recycled components is still however a matter of conjecture. Here, we report the results of a geochemical investigation on K-rich clinopyroxenes from ultrahigh pressure (UHP) metasediments (Kokchetav Massif, Northern Kazakhstan) and from inclusions in the Argyle lamproitic diamonds (Australia) which may provide a solution to this question.

These K-rich diopsides and omphacites, stable in upper mantle conditions between 4 and 14 GPa (Schmidt, 1996; Harlow, 1997; Luth, 1997), display unusual crustal-level enrichments in Cs, Rb and Pb. They record an efficient decoupling of many continental-derived elements (K, Cs, Rb, Sr, Ba, Ce, Pb, Th, U) which results in unusual Rb/Cs (1-17), Ba/Rb (0.2-4), Ce/Pb (0.08-1.5) and K/Rb (>500). These geochemical properties are likely to be due to the interaction with K-rich fluids. A potential source for those fluids might be the phengite breakdown in the deep confines of subduction zones.

Subduction of K-cpx-rich rocks, formed either as residues from phengite and dolomite melting in the down-going slab (such as the Kokchetav garnet clinopyroxenites made up of 70-80 vol.% of K-rich diopside) or through the percolation of UHP K-rich fluids within eclogitic bodies, is proposed as an ideal mechanism to recycle supracrustal K, Rb, Cs, Pb, Sr, and Nd to the mantle, up to a pressure where clinopyroxenes gradually enter into solid solution with the garnet phase. The geochemical and isotopic implications of such a recycling will be discussed.

Schmidt MW, Science, 272, 1927-1930, (1996).

Harlow GE, American Mineralogist, 82, 259-269, (1997).

Luth RW, American Minerlogist, 82, 1198-1209, (1997).

O10 : 5A/12 : H1

No Chemical Mobility in one Century in Massive Lava Flows from Mount Cameroon

Marc Bulourde (bulourde@univ-rennes1.fr)¹,

Catherine Chauvel,

Aline Dia,

Bernard Déruelle²,

Philippe Ildefonse,

Martine Gérard³ &

François Chabaux⁴

¹ Géosciences Rennes, Campus Beaulieu, 35042 Rennes Cedex, France

² Université Paris 6, 4 Place Jussieu, 75252 Paris Cedex 05, France

³ LFS - ORSTOM, Bondy, France

⁴ CGS, 1 Rue Blessig, 67084 Strasbourg Cedex, France

Understanding mobility of trace elements in volcanic lavas is of prime importance in order to concentrate on their petrogenesis. Indeed, alteration processes can affect drastically the original composition of basaltic lavas. Element mobility and changes in isotopic composition have been reported in various weathered profiles but the effects of weathering on short time scale (< 10² years) are not well constrained. Basaltic lava flows erupted in 1909, 1922, 1959 and 1982 were therefore sampled on the flanks of Mount Cameroon in west Africa.Mount Cameroon is an active volcano located in a humid tropical area. The extreme rainfalls (2 to 10 m per year) and elevated temperatures (26 to 29°C at sea level) represent conditions that should favour rapid alteration of basaltic rocks. In addition, lavas are alkaline and rich in Cs, Rb and K, elements which are known to be highly mobile during alteration.For each flow, the fresh inner part and the altered outermost part were separated and analysed for major and trace elements. Sr, Nd and Pb isotopic compositions were also determined. Concentrations of most elements do not change significantly between inner and outer parts of the same flow. No variation is noticed for Th, Ti, Ta, Zr, Nb and other elements which are usually considered as immobile even in severe alteration conditions. The same applies to potentially more mobile elements such as REE, Sr and Ba. Moreover, inner and outer parts of the flows show nearly constant Cs/Ta, Rb/Ta and U/Th ratios. In all samples, these ratios have magmatic values. Similarly, the Ce/Pb ratios are unchanged and display mantle-like values. Consequently, the alkalis, U, and Pb do not show any sign of mobility.Isotopic compositions of the inner parts of flows are very uniform : ⁸⁷Sr/⁸⁶Sr ranges from 0.70331 to 0.70335, ¹⁴³Nd/¹⁴⁴Nd from 0.51278 to 0.51280 and ²⁰⁶Pb/²⁰⁴Pb from 20.28 to 20.39. These values are comparable to those reported in the literature for Mount Cameroon lavas (Halliday et al, 1988; Ballentine et al, 1997). Sr and Nd isotopic values obtained for the outer and the inner parts of the flows are indistinguishable. In contrast, Pb isotopic compositions vary slightly between the two parts but the changes are not systematic and are not yet understood.This study shows that, despite several decades of severe alteration conditions, chemical characteristics of recent basalts (< 100 years old) are untouched. They represent the original magmatic compositions even in the outermost parts of the flows. We suggest, therefore, that they can safely used to constrain mantle source compositions.

Ballentine CJ, Lee DC & Halliday AN, Chem. Geol, 139, 111-124, (1997).

Halliday AN, Dickin AP, Fallick AE & Fitton JG, J. Petrol, 29, 181-211, (1988).

O10 : 5A/13 : H1

Magma Mixing and Chemical Interdiffusion in the Westerwald (Germany) and the Massif Central (France) ­ A Petrographic and Geochemical Comparison

Joachim Koppen

(joachim.koppen@uni-essen.de) &

Ulrich Schreiber (ulrich.schreiber@uni-essen.de)

FB 09 Geologie Universität Essen, Universitätsstr. 5, D-45141 Essen, Germany

The genesis of the alkali-rich volcanics (basanites, alkali-rich trachytes, phonolites) of the undersaturated alkaline series in the Tertiary Westerwald (Germany) can be explained by a special kind of magma mixing combined with diffusive exchange processes. During the intrusion of a hot basic guest melt into a cooler more evolved host, the guest melt are dispersed as small spherical melt droplets (globules). Due to diffusive exchange between both melts, the globules become enriched in alkalis while the contents of alkaline earth element, Fe, and Ti are reduced. The result of these exchange processes is a phonolitic composition of the most evolved globules. These processes are documented by geochemical investigations in a subvolcanic latite/trachyte-intrusion of the Westerwald (Schreiber et al., in press). Due to the diffusive exchange, density, melt tempera-ture, and viscosity of the globules has been reduced in respect to the host melt. Therefore, the diffusive altered globules are able to rise up within the host melt and collect themselves beneath the top of the magma chamber generating an alkali-rich extraction melt. This process is documented in a microglobular texture of the alkali-rich vol-canics of the undersaturated series. Similar phenomena can be ob-served in the subvolcanic complex of the Puy Griou in the late Tertiary Cantal province (M. Central, France). The phonolitic host rock contains similar spherical inclusions With a tephritic composition. According to the Westerwald volcanics, a mineral exchange between both melts is indicated by strongly zoned phenocrysts (e.g. green core pyroxenes). Furthermore, microprobe measurements in the glasses and the microlithic matrix of the inclusions clearly show a diffusive migration of the alkaline elements (mainly sodium) in the inclusions. Anyway, the center of the inclusions has not been affected by the diffusive enrichment indicating a rapid cooling down of the whole magma body. The original composition of the globules of the Puy Griou is unknown. But the measured profiles clearly indicate a similar diffusive exchange as in the Westerwald volcanics. Additionally, the phonolitic host shows a microglobular texture according to the Westerwald phonolites. Slightly higher alkali contents in the center of the microglobules indicate a generation of the phonolitic host melt by mixing, diffusive exchange and extraction of the alkali-rich globules. After this first mixing event, a second mixing must have taken place producing the greater tephritic inclusions. The dif-fusive exchange between these inclusions and the phonolitic host must have been interrupted by rapid solidification of the subvolcanic intrusion.

Schreiber U, Anders D & Koppen J, Lithos, (in press).

O10 : 5A/14 : H1

Geochemical Evidence for Nature and Evolution of the Rift Volcanism Related to the Opening of the Black Sea, Central Pontides, Turkey

Mehmet Keskin (keskin@istanbul.edu.tr) &

Okan Tuysuz (tuysuz@itu.edu.tr)

Istanbul University, Muhendislik Fakultesi, Jeoloji Bolumu, 34850 Avcilar, Istanbul Turkey

In the Central Pontides, the Cide-Inebolu area contains an almost complete record of opening of the Black-Sea basin during the Upper Cretaceous period. The area lies parallel to the Black Sea cost as a long, narrow, E-W extending belt and contains a thick volcano-sedimentary succession deposited in an extensional regime. Our palaeontologic data indicate that rifting started in Turonian with a widespread volcanism, giving rise to the deposition of a clastic succession, termed the Derekoy formation. The unit consists basically of marl, shale, sandstone, occasional chert and debris flows intercalated with porphyritic lava flows and agglomerates. It also contains limestone olistolithes of Jurassic age and debris flows of basically volcanic origin, reflecting intensive extensional deformation of the crust associated with volcanism. During the period between Upper Santonian and the beginning of Campanian, both volcanism and clastic sedimentation stopped and pelagic sedimentation dominated all over the region. The Unaz formation, composed of red pelagic limestone, precipitated during that period. Volcanism and clastic sedimentation restarted in the Campanian, and produced a thick succession of coarse-grained volcaniclastic beds intercalated with columnar-jointed and pillow lava flows, termed the Cambu formation. During Maastrichtian, volcanism ceased and gave way to the sedimentation of distal flysches named the Akveren and Gursoku formations respectively. Our geochemical data indicate that lavas of the Derekoy formation span a compositional range from andesite to rhyolite whereas those of the Cambu formation cover a wider range from basalt to rhyolite. Lavas of the Derekoy formation are calc-alkaline in character and contain anhydrous phases (POAM). In contrast, lavas of the Cambu formation consist of polybaric crystallisation assemblages: (1) POAM and (2) hydrous (PA: plagioclase + amphibole). Lavas containing PA assemblage are all calc-alkaline (CA) in character, while those containing POAM phases are either alkaline or display an alkaline affinity with a within plate signature. Trace element trends (e.g. Hf or Zr vs Sc, V, Mg and Fe) provide evidence for magma mixing between alkaline and CA magmas. Lavas from both Dereköy and Cambu formations contain a distinctive subduction signature represented by selective enrichment of LILs and LREEs. Since this signature is present even in the most primitive lavas, we believe that it may have been inherited from a previous subduction event in the region. Our trace, major and REE data are consistent with a model where magma generation is associated with lithospheric thinning which caused derivation of magma from progressively deeper zones in the mantle during the opening of Black Sea.

Session O10:5B

O10 : 5B/21 : H1

The Genesis of Volcanic Rocks from Central Italy: In Situ UV-Laser ¹⁸O Data

Peter Barnekow (pbarnek@ugcvax.dnet.gwdg.de)¹,

Jochen Hoefs (jhoefs@gwdg.de)¹ &

Angelo Peccerillo (pecceang@unipg.it)²

¹ Geochemisches Institut, University of Göttingen, Goldschmidtstr. 1, D - 37077 Göttingen, Germany

² Dipartimento di Scienze della Terra, University of Perugia, Piazza Università, I - 06100 Perugia, Italy

A common feature of the Pliocene to Pleistocene volcanic rocks from Tuscan and Roman Magmatic Province (central Italy) is the very high K₂O content and the enrichment in ¹⁸O, ⁸⁷Sr (Ferrara et al., 1986), and several incompatible trace elements. This enrichment results in a crustal signature even for the most primitive mantle melts. However there is discussion on whether crustal geochemical signatures result from mantle contamination processes or by assimilation of crustal rock during uprise to the surface.

Primitive mantle melts occur as lamproites in the Tuscan and as SiO₂ undersaturated Roman-type rocks in the Roman Province. The chemical compositons of these rocks permit only small degrees of assimilation and differentiation (Conticelli et al., 1991; Conticelli et al., 1992) but the ¹⁸O values are uncommonly high - 12.6‰ for a phlogopite from a lamproite and 8.0‰ to 8.6‰ for clinopyroxenes from Roman Province rocks.

Transitional rocks between lamproites and Roman-type rocks occur at Radicofani and Torre Alfina. Radicofani shows lower ¹⁸O values (cpx: 7.0 ± 0.2‰; ol: 7.1 ± 0.3‰) than Roman Province, whereas rocks from Torre Alfina display high ¹⁸O values of 13.1 ± 0.3‰. A high-pressure ultramafic xenolith from Torre Alfina lavas shows uncommon high and heterogeneous ¹⁸O values for the olivines indicating an ¹⁸O enriched mantle (Barnekow et al., 1998).

Crustal anatectic rocks and hybrids between mantle- and crustal-derived melts also occur in the Tuscan Province. Anatectic rocks show typically high oxygen isotopic compositions (fsp: 11.5 ± 0.2‰; qz: 12.4‰).

Oxygen isotope ratios of hybrid rocks are intermediate between mantle and crustal values (latite: cpx: 9.0 ± 0.2‰, fsp: 10.3 ± 0.3‰; trachyte: fsp: 12.3 ± 0.2‰), in agreement with the nature of these rocks.

In summary, oxygen isotopic data support the hypothesis of an anomalous metasomatic mantle source for Italian potassic magmatism. However, low-pressure interaction between mantle-derived magmas and crustal wall rocks also appear to have played a role in determining the oxygen isotopic composition of the rocks investigated. Source metasomatism and magma contamination processes worked at different degrees in the various volcanic centres generating a complex pattern of isotopic variation in volcanic rocks.

Barnekow P, Hoefs J & Peccerillo A, Mineral Mag, 62A, 120-121, (1998).

Conticelli S, Francalanci L & Santo AP, J. Volcan. Geotherm. Res, 46, 187-212, (1991).

Conticelli S, Manetti P & Menichetti S, Eur. J. Mineral, 4, 1359-1375, (1992).

Ferrara G, Preite-Martinez M, Taylor HP, Tonarini S & Turi B, Contr. Mineral. Petrol, 92, 269-280, (1986).

O10 : 5B/22 : H1

Ethiopian Ignimbrites: Relation with Oligocene Continental Flood Basalts and Potential Impact on the Atmosphere

Dereje Ayalew (dereje@crpg.cnrs-nancy.fr)¹,

Bernard Marty (bmarty@crpg.cnrs-nancy.fr)²,

Pierre Barbey (barbey@crpg.cnrs-nancy.fr)²,

Gezahegn Yirgu (dgg@telecom.net.et)¹,

Laurie Reisberg (reisberg@crpg.cnrs-nancy.fr)² &

Endale Ketefo¹

¹ Addis Ababa University, Geology & Geophysics Department, P.O.Box 1176, Addis Ababa, Ethiopia

² CRPG-CNRS, 15 rue notre dame des pauvres, 54501 Vandoeuvre les nancy, France

The Ethiopian traps, located near the Afar triple junction, cover an area of > 6x10⁵ km² with an estimated volume of at least 3x10⁵ km³. The traps related to the development of the Afar plume, belong to one of the youngest CFB provinces (30 ± 1 Ma, Hofmann et al., 1997). The volcanic succession is bimodal with the basalts being dominant in volume over the rhyolites (> 60 000 km³). The rhyolites form thick sequences (up to 500 m, exceptionnally 1000 m in the SW) of ignimbrites and tuffs overlying the CFB and often capped by obsidians. Individual flows commonly reach tens of meters in thickness and tens of kilometers in lateral extent, indicating voluminous fissural eruptions. The rhyolites are variably porphyritic containing phenocrysts of anorthoclase, sanidine, quartz, ilmenite, clinopyroxene, amphibole and locally plagioclase (An_23-39), pigeonite and titanomagnetite.

Whole-rock and mineral (anorthoclase and amphibole) Rb-Sr isotopic data indicate that silicic volcanism migrated from the plateau towards the rift. An isochron age of 30.2 ± 1.5 Ma has been obtained for the large ignimbrite unit from the northern part of the plateau (Wegel Tena). The rhyolites from the northern Debre Birhan area yield a whole-rock isochron age of 15.5 ± 1.3 Ma, whereas a much younger isochron age (8.01 ± 1.6 Ma) has been obtained from a rhyolite in the southern Debre Birhan area close to the rift. The rhyolites from southwestern Ethiopia (Jima area) yield an isochron age of 30.7 ± 0.5 Ma, similar to that in the north, with an initial ⁸⁷⁸⁷Sr/⁸⁶Sr ratio of 0.70379 ± 72. The ¹⁴³Nd/¹⁴⁴Nd ratios are very similar to those of the associated basalts. The ¹⁸O values are in the range of rhyolites derived from mantle-like protholith. This suggests that (i) the Oligocene rhyolites in SW Ethiopia was contemporaneous with those of N Ethiopia and once formed a continuous unit, (ii) they did not derive from crustal anatexis, and (iii) they are closely linked to the associated basalts through extensive fractional cristallisation coupled with limited amounts of crustal contamination. Such fractionation implies the presence of huge mafic cumulates beneath the Ethiopian crust or lithosphere. The Oligocene rhyolites have high volatile contents, such as S (2500 ppm) and Cl (6400 ppm). This together with their large volume extruded over a short time span, and their explosive character might have had a significant effect on the climate at the time of their formation, i.e. on the major global cooling around 30 Ma ago as inferred from the ¹⁸O record (Peckard and Miller, 1996).

Hofmann C, Courtillot V, Féraud G, Rochette P, Yirgu G, Ketefo E & Pik R, Nature, 389, 838-841, (1997).

Pekar S & Miller KG, Geology, 24, 565-570, (1996).

O10 : 5B/23 : H1

Organic Geochemistry of Slope Sediments and Abyssal Turbidites: A 15 Million Year Continually Catastrophic Environment?

Sten Lindblom (sten.lindblom@geo.su.se)¹,

Sofie Kruckenberg¹,

Ulf Järnberg (ulf.jaernberg@itm.su.se)² &

Paul Frogner¹

¹ Department of Geology and Geochemistry, Stockholm University, S-10691 Stockholm, Sweden

² Institute of Applied Environmental Research, Stockholm University, S-10691 Stockholm, Sweden

The Canary Basin contains a varied geolological setting with the continental shelf and slope off the Morroccan coast, the Canary islands and the Madiera Abyssal Plain. The abyssal plain is made up of a repeated sequence of turbidites with a large part high in organic matter. The bottom environment of the plain is oxic, altering the organic matter downwards into the bottom sediment until the next turbidite is emplaced. The primary accumulation of organic matter has been thought to occur in a source area along the Northwest African continental margin and that a steady sedimentation of regular and identical organic-rich turbidites from this area has taken place over a long time (Schmincke, Weaver, Firth et al., 1995).

Using organic geochemistry of different biological markers will further elucidate source areas and alteration of the turbidites. GC-MS -analyses of both abyssal plain and continental slope sediments indicate proportions of terrestrial versus marine origin of the organic matter (Venkatesan, 1987).

Turbidite samples from Site 951 and 952 (ODP Leg 157) has an organic carbon content between 0.75 and 1.68% C and a C/N ratio of around 10 (indicating a mainly marine source). n-alkane distributions between C15 and C34 indicate a bimodal distribution with a prevalence of the C15 to C20 n-alkanes (mainly marine) and a clear but minor C21 to C34 fraction (terrestrial part). Samples originate at a depth below sea floor of from 60 to 250 meters (mbsf) and up to 140 km apart and have almost identical distribution of n-alkanes.

Samples from the continental slope show very similar n-alkane distributions although these samples have slightly less organic carbon and are much younger (1.6 to 6.6 mbsf), probably around 30 000 to 50 000 years old depending on the sedimentation rate. Earlier work on the abyssal plain has given lower organic carbon values on nearer surface samples of similar age (Weaver and Rothwell, 1987).

Fatty acids show similar distributions of carbon atom species with a higher content in nearer surface sediments of the slope samples.

An investigation of aromatics indicate very low values representative of normal preindustrial concentrations compared to other than marine environments.

The processes on the continental slope, involving mass transfer to the plain, appear to have been going on for a very long time. The record of repeated turbidites goes back 15 million years to the Early Miocene and there is a strong indication from organic geochemistry that these processes has had a regular occurrence over this time period. (This is a contribution within the European CANIGO-project, G. Parilla, coordinator)

Schmincke H-U, Weaver PPE, Firth JV et al., Proc. ODP, INIT. Reps. 157 (Ocean Drilling Program), 157, 843 pp, (1995).

Venkatesan MI, Ruth E, Steinberg S & Kaplan IR, Marine Chemistry, 21, 267-299, (1987).

Weaver PPE & Rothwell RG, Geology and Geochemistry of Abyssal Plains, Geol. Soc. Spec. Publ, 31, 71-86, (1987).

O10 : 5B/24 : H1

Geochemical Study of Pakistani Natural Gases Combining Rare Gases and Hydrocarbon Gases Analysis

Anne Battani (battani@ccr.jussieu.fr)¹,

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

Alain Prinzhofer (alain.prinzhofer@ifp.fr)³

¹ UPMC, case110, T26, 4 place Jussieu, 75252 paris cedex 05, France

² groupe de geochimie des gaz rares, Universite Paris 11 bat. 504, 91405 Orsay Cedex, France

³ IFP, 1&4 avenue, de bois-preau, 92852 Rueil-Malmaison Cedex, France

We collected 25 natural gas samples in the lower Indus basin and in the middle Indus basin, Pakistan. All were analyzed for their chemistry and carbon isotopes, while 15 of them were also analyzed for their He, Ne and Ar contents, as well as He and Ar isotopic ratios. This study was carried out in order to understand the dryness of the Pakistani natural gases, and the logic of distribution, in the different reservoirs, of CO₂ and N₂. Indeed, in some of the studied fields, the proportions of these two species are negligible, while they can reach very high values in some others, as 69% of CO₂ and 23% of N₂. Comparing the composition of the gases with what is obtained from laboratory pyrolysis in close and open systems, it seems that Pakistani gases have high maturities, from the end of the oil window to the middle of the gas window. They present a small efficiency of accumulation. This may be linked with the dryness of the gas, with the lack of any associated oil or condensate, as well as for large proportions of nitrogen, as this gas is assumed to be generated during the catagenesis. Moreover, the fields with the worst yield of accumulation are also the dryer ones. This proves that the oils generated earlier have been lost out of the reservoirs. The helium isotopic ratios indicate that there is no mantle-derived helium in the accumulations, so that nitrogen and carbon dioxide may therefore originate from another source. The ¹³C of CO₂ are heavy, implying that CO₂ does not come from an organic source, but most probably come from the thermal degradation of carbonates. On the contrary, organic matter maturation can account for Nitrogen production during catagenesis, with its high proportion being related to the openness of the system, as mentioned earlier. The rare gas data demonstrate the importance of mixing processes during gas generation and accumulation, with end-members clearly related to geographic areas. Moreover, using alternatively the three major species CH₄, N₂ and CO₂ with ³⁶Ar data in mixing diagrams, we define the same mixing families as those found using only rare gas data.Given these results, we built a mixing model using CH₄, N₂ and CO₂ with the assumption that a first pulse of relatively homogeneous thermogenic gas was generated in the whole Indus basin, and that different second pulses arrived latter, with compositions related to the respective locations of the fields. We explain the relative proportions of CO₂ and N₂ with the relative proportions of these different late pulses.

O10 : 5B/25 : H1

Biomarker Maturity Parameters: The Role of Kerogen-Bound Components

Paul Farrimond (paul.farrimond@ncl.ac.uk)¹,

Gordon Love,

Andy Bishop (bishoan@texaco.com)²,

Colin Snape (C.E.Snape@strath.ac.uk)³ &

Nils Telnaes (Nils.Telnas@nho.hydro.com)⁴

¹ Fossil Fuels & Env. Geochemistry, University of Newcastle upon Tyne, Newcastle upon Tyne, United Kingdom

² Texaco Inc., Houston, United States

³ Dept. Applied Chemistry, University of Strathclyde, Glasgow, United Kingdom

⁴ Norsk Hydro, Bergen, Norway

Hydrocarbon biomarkers have been used to determine the extent of thermal maturation of organic matter in the geosphere (principally oils and their source rocks) for over 20 years. However, quantitative investigations of the response of biomarker composition to progressive maturation are rare, and the processes underlying biomarker maturity parameters have been poorly constrained. This study presents quantitative data from two suites of samples: (1) a carbonate-cemented horizon heated by an igneous intrusion (Isle of Skye, Scotland), and (2) a Tertiary section (1495-2840 m) in a well from the Barents Sea heated by natural progressive burial.

Progressive changes in conventional biomarker maturity parameters were compared with the concentrations of individual biomarkers to identify the possible mechanisms controlling the changes with increasing maturity. Our data confirm previous indications that the observed changes in biomarker maturity parameters result primarily from the combined effects of biomarker generation (from the kerogen, or other non-hydrocarbon fraction of the sedimentary organic matter) and subsequent thermal degradation (e.g. Bishop and Abbott, 1993; Requejo, 1994). Most biomarkers increase dramatically in concentration just before, and within, the effective oil window. The concentrations of these generated biomarkers far outweigh the original biomarker concentrations in the bitumen, indicating a dominant origin from a non-hydrocarbon source. Consequently, direct isomerisation between biomarker isomers in the bitumen appears to be a quantitatively unimportant effect in the operation of the maturity parameters.

Kerogen-bound biomarkers were determined for the Isle of Skye (intrusively heated) sample suite using hydropyrolysis, a method with the unique characteristics of liberating maximum yields of biomarkers with minimal alteration (Love et al., 1995). The concentrations of kerogen-bound biomarkers beyond the heating effects of the dyke were found to be comparable to the amounts of newly generated biomarkers within the oil window, suggesting that the kerogen may indeed be the major source of the biomarkers in the mature samples. Examination of the kerogen-bound biomarkers indicates that the bound hopanes become progressively more mature with increasing proximity to the dyke, indicating that some isomerisation reactions are occurring within the kerogen. However, the isomerisation of kerogen-bound biomarkers appears to generally lag behind the biomarkers in the bitumen, and we propose that isomerisation between less mature and more mature isomers is principally associated with the cleavage of biomarkers from the kerogen.

Bishop AN & Abbott GD, Geochim. Cosmochim. Acta, 57, 3661-3668, (1993).

Requejo AG, Org. Geochem, 21, 91-105, (1994).

Love GD, Snape CE, Carr AD & Houghton RC, Org. Geochem, 23, 981-986, (1995).

O10 : 5B/26 : H1

Hydrochemical Results of the Lakes Burdur and Acygöl, Western Anatolia, Turkey

Fikret Suner (suner@itu.edu.tr),

Atasever Gedikoglu (gedikoglu@itu.edu.tr),

Mustafa Kumral (kumral@itu.edu.tr),

Vildan Esenli (esenliv@itu.edu.tr),

Murat Budakoglu (budakoglu@itu.edu.tr) &

Ahmet Celenli (celenli@itu.edu.tr)

ITU Mining Fac., Dep. of Geology, 80670 Maslak, Turkey

The hydrochemical investigation of the lake- complex in Western Anatolia were performed. Two lakes, named Acogöl and Burdur lakes, were studied in terms of chemical precipitaion paths. In the studied area many lakes had been formed during the new tectonic regime which was responsible for the formation of geologic mosaic in Turkey. Based on the preliminary results, obtained from water, brines and crust sediments and rocks samples'analyses, it is postulated that the lakes were subjected to the similar chemical precipitation trends. The data obtained from the lake of Burdur reveal that Burdur lake waters are exhibited a Ca-Mg free character while in anionic point of view there is no distinctive property. The waters and interstitial brines carry a high Na content; and ClSo₄ HCO₃ + CO₃ anionic order were determined. These properties are yielded to a brine type as observed in Deep Springs and/or Walley playas. On the other hand, the data of Acigöl showed that the waters of this lake are Ca-MgHCO₃-CO₃ and Cl/SO₄ ratio is quite high. This compositional framework is postulated that the waters of Acigöl Lake exhibit a similarity partly to Death Valley and Great Salt Lake in terms of monomineralic evaporation path.

O10 : 5B/29 : H1

A Geochemical Approach to the Genesis of Miocene Gypsum Occurrences in Eskisehir Basin (Turkey)

Fazli Coban (coban@itu.edu.tr),

Fikret Suner (suner@itu.edu.tr) &

Isik Ece (ece@itu.edu.tr)

itü mining fac. 80670, Turkey

In the presented paper, it was studied formation conditions of gypsum occurrences in terms of geochemical aspects. Three types of gypsum formation were identified in Eskisehir - Sivrihisar Basin (Central Anatolia) where Upper Paleozoic metamorphic rocks and Upper Cretaceous ophiolitic complex are the basement rocks which are overlain by Miocene sedimentary rocks. Upper Miocene - Pliocene stratigraphic succession consists of basal conglomerates, sandstones, claystones, limestones, dolomites that are interbedded with cyclic evaporation zones where the mentioned gypsum occurrences were formed. Within these cyclic evaporation zone, the gypsums are alternated with clayey carbonates and partly unconsolidated very soft dolomite levels, which are conformably underlain by sepiolite bearing clayey carbonate and disconformably overlain by Pleistocene alluvium and conglomerate series. In the studied area, three different types of gypsum were classified based on the data of sedimentary and chemical investigations: 1) laminated gypsum beds, which are interbedded with dolomites and green clays; 2) rosette gypsums, which are determined within thin reddish and greenish clays and carbonates 3) scattered gypsum, which are observed in carbonates mainly. As a result of the field observations, macro and microscopic studies and chemical data, it is postulated that the studied basin had been subjected to a multi-periodic evaporative episode. During these episodes a meromictic type lake had been developed and various physicochemical conditions had worked; Eh, pH and salinity levels in the lake waters and pore solutions were fluctuated. Under these conditions different sub-environments had been possibly yielded and they were postulated in the following forms: 1) subaqueous playa environment during short evaporation period 2) marginal swamps under the effect of cyclic fluctuation of lake waters and 3) extenuated closed evaporative lake episode.Within these subenvironments various gypsum morphologies within different sedimentary subfacies had been formed in arid, semi-arid and evaporative conditions with cyclic climatic changes and lake water fluctuations. A very large interval of major, minor and trace elements contents, determined by means of chemical analyses, are also supported the presence of postulated subenvironments and a multiperiodic meromictic type of playa lake.

O10 : 5B/30 : H1

Factors Controlling Pollution Migration from Khibiny Apatite Mines

Dmitry Malinovsky (sharov@inep.ksc.ru)

INEP, 14 Fersman st., 184200, Apatity, Russia

Intensively developed mining industry is one of the main sources of water pollution on areas surrounding Khibiny mountains. Nitrogen compounds, Sr, Al, F, SO₄, suspended substances are predominant pollutants entering into natural waters as a result of mining activity. Objective of the work was to study formation of surface and ground water chemical compositions on the areas affected by the mines and to define the factors controlling distribution of pollutants. During 1996-1998 there was done the regular sampling of water and sediments in surface streams, lakes and wells in area affected by two apatite mines. In water samples there were determined basic cations and anions, trace metals that are of importance in the area (Sr, Al, Fe, Mn, Pb, Cd), content and chemical composition of suspended substances. Special attention was paid to determination of metal's speciation in water. Using fractionation procedure the suspended, non-labile and labile forms of metals were determined. Contents of suspended substances were analyzed by sequential filtering of large water volume through 1.00 and 0.45 mm filters. In sediment samples the total and bioavailable contents of metals were determined using extraction methods. In order to assess the adsorption capacity of the aquifer there were done laboratory experiments on adsorption of Sr, Pb, Cd, Mn, Al on Quaternary deposits containing ground water. Field observations over distribution of the pollutants in surface water enable to assess the influence of dilution of polluted flux by runoff from the catchment and sorption of contaminants onto streambed sediments. Predominant forms of metals migration are suspended. Most of metals are associated with particles of size 0.45-1.00 mm. Dissolved metals, except Sr, are present mainly in non-labile form bound with inorganic and organic ligands. Technogenic input of elevated concentrations of Cl and SO₄ into surface water in points of wastewater discharges strongly increases metal`s mobility. Results of the adsorption experiments indicate that under conditions of ground water aquifer intensive adsorption of Pb, Cd, Sr, Mn is observed. Adsorption is strongly affected by pH-conditions and presence of complexing substances. Experimental data on adsorption provide a good explanation for field observations. Water sampling in wells indicate that ground water more rapidly immobilizes fluxes of pollution compared to surface water. In whole it can be concluded that weakly alkaline waters of the research area effectively decrease the concentrations of Cd, Pb, Fe, Sr and Mn. Migration of these metals occurs mostly in suspended and colloidal forms. Compounds having high migration capacity and representing a real threat to water ecosystems are nitrogen compounds and Al.

O10 : 5B/31 : H1

Do Amphiboles Behave as a Single Phase in Mineral/Melt Partitioning?

Massimo Tiepolo (tiepolo@crystal.unipv.it)¹,

Roberta Oberti (oberti@crystal.unipv.it)²,

Richard Brumm (rbrumm@gwdg.de)³,

Riccardo Vannucci (vannucci@crystal.unipv.it)¹ &

Stephen F. Foley (sfoley@gwdg.de)³

¹ Dip. di Scienze della Terra, Università di Pavia, Via Ferrata 1, I-27100 Pavia

² CNR-CS per la Cristallochimica e la Cristallografia, Via Ferrata 1, I-27100 Pavia

³ Mineralogisch-Petrologisches Institut, Universität Göttingen, D-37077 Göttingen

Calcic (kaersutite/pargasite) and sodic-calcic (K-richterite) amphiboles occur as primary magmatic phases in alkaline lavas and in lamproites/kimberlites, and are likely to play a significant role in determining the trace element and radiogenic characteristics of different lithospheric mantle domains. The availability of consistent sets of amphibole/melt partition coefficients for a large number of geochemically important elements is therefore crucial for modelling the melting behaviour of metasomatically enriched amphibole-bearing mantle sources. New partitioning experiments were performed under conditions relevant to mantle processes; they were planned a) to highlight the different partitioning behaviour of calcic and sodic-calcic amphiboles, and b) to unravel the complex mechanisms of trace element incorporation in amphibole. These aims were pursued by combining the lattice-site elastic-strain models with site populations and dimensions determined by SREF. The results relevant for geochemical modelling are:

1) ^Amph/LD values for REE are lower by about one order of magnitude in K-richterite than in kaersutite/pargasite, in keeping with the limited capability of the former of hosting (major-element) trivalent cations and with the significant presence of Na at the M4 site, which would locally imply a jump of +2 in the ionic charge. Contrary to kaersutite/pargasite, in K-richterite HREE are strongly partitioned relative to LREE, for which a nearly flat variation of ^Amph/LD is observed. LREE and HREE are incorporated within the same cavity (M4) but in different positions in kaersutite/pargasite, whereas partition between different sites (M2 and M4) in K-richterite. The ^Amph/LD for HREE in the calcic amphiboles correlate with Fe and Mg occupancy in the M4 cavity.

2) HFSE are preferentially incorporated in kaersutite/pargasite than in K-richterite, in which the large M2 site does not favour the entrance of short-radius high-charge cations. HFSE are ordered at M1 in richterite, especially in the presence of significant amounts of Ti balancing for partial dehydrogenation. A complex behaviour is observed in kaersutite/pargasite, in which HFSE R⁴⁺ cations (i.e. Zr and Hf) follow the fraction of Ti⁴⁺ which is not related to dehydrogenation (that occurring at M2), whereas HFSE R⁵⁺ cations (i.e. Nb and Ta) follow Ti in M1 and cooperate to the balancing of dehydrogenation.

3) Rb, Ba, Cs (in the A cavity) have lower ^Amph/LD values in K-richterite than in kaersutite/pargasite. Sr is more compatible for the larger and less constrained M4 sites of K-richterite.

Contrasting mineral/melt partitioning behaviour of calcic and sodic-calcic amphiboles mainly results from peculiar ordering patterns strongly controlled by the relative dimensions of the structural sites. This implies that the melting behaviour of amphibole-bearing mantle sources can be modelled with confidence only if appropriate assumptions on the possible amphibole composition are made.

O10 : 5B/32 : H1

Minor and Trace Element Determinations in Li₂B₄O₇ Fused USGS Standard Materials Calibrated Without Matrix-Matched Standards Using Laser Ablation-ICP-MS

Albrecht von Quadt (quadt@erdw.ethz.ch)¹,

Detlef Günther (guenther@inorg.chem.ethz.ch)²,

Rolf Frischknecht (frischknecht@erdw.ethz.ch)¹ &

Volker Dietrich (dietrich@erdw.ethz.ch)¹

¹ Institue of Isotope Geology and Mineral Resources, ETH Zurich, Sonneegstr. 5, Switzerland

² Lab. Anorganic Chemistry, ETH Zurich, 8092 Zurich, Switzerland

XRF and ICP-MS are the major techniques used for whole rock analysis. This study was performed to evaluate the LA-ICP-MS as a complementary technique to electron microprobe and XRF for trace elements in Li₂B₄O₇ without time consuming digestion procedures necessary for solution analysis using ICP-MS. In the last few years the potentional of LA-ICP-MS has been demonstrate for trace element determinations in minerals. However, matrix matched standards are often required to perform accurate analyses.

Table 1:

Examples of selected elements determined (AGV-1) using LA-ICP-MS in comparison to XRF-values, ICP-MS, SSMS.

element XRF ICP-MS SSMS LA-ICP-MS stdev.

Na₂O 4.26 n.m. n.m. 4.12 0.07

CaO 4.94 n.m. n.m. 5.36 0.71

Rb 0 67.3 68.9 65.4 1.05

Zr 229 227 242 208.3 4.53

La 36.3 38.0 40.8 38.38 0.78

Nd 29.3 33.0 5.7 32.36 1.16

Dy 3.26 3.60 4.09 3.54 0.09

Yb 1.54 1.72 1.83 1.81 0.11

Hf 5.32 5.10 5.43 5.28 0.29

Pb 36.7 36.0 47.0 30.4 0.41

Th 5.5 6.5 6.97 6.0 0.20

n.m.= not measured

An excimer laser ablation-ICP-MS system was used for multi-element analysis in lithium tetraborate fusion pellets. Different Li₂B₄O₇ fused USGS standards (BHVO-1, AGV-1, RGM-1, NIM-P, NIM-S, STM-1, DNC-1, BCR-1) were analysed using the LA-ICP-MS. NIST 612 was used as external calibration standard and SiO₂ was tested as internal standard for correction of the different masses ablated. 40 isotopes, including a few major, minor and trace elements were measured.

The reproducibility as an expression of the quality of the fused pellets were determined, as shown for AGV-1 standard in Table 1, with better 5% for most of the major and minor elements and better 8% for most of the trace elements. The determined values are compared to different reference values measured using XRF, spark source MS (SSMS), and solution nebulization-ICP-MS. The deviation of the measured concentration in comparison to certified or recommend values demonstrate the potential of laser ablation ICP-MS as a complementary technique tool for trace element determinations. Reproducibility, the deviation from the reference values, the limits of detection, and the advantage of the fast laser ablation-technique using non-matrix-matched standards will be extensively discussed.

O10 : 5B/33 : H1

Isotopic Imaging of Individual Aerosol Microparticles with the Cameca IMS 1270 Ion Probe

Jérôme Aleon (aleon@crpg.cnrs-nancy.fr),

Marc Chaussidon (chocho@crpg.cnrs-nancy.fr),

Bernard Marty (bmarty@crpg.cnrs-nancy.fr),

Michel Champenois (champ@crpg.cnrs-nancy.fr) &

Denis Mangin (mangin@crpg.cnrs-nancy.fr)

CRPG-CNRS, Vandoeuvre-lès-Nancy, France

In climate studies the accurate modelling of the climate response to the modification of atmospheric conditions requires that the role of the different parameters is well understood. With this respect the role of aerosol although intensively studied is not yet well constrained (Alpert et al., 1998). Stable isotopes could bring informations on the sources and the transport mechanisms of these microparticles (Clayton et al., 1972).

Several conditions are needed for an accurate measurement of oxygen isotope ratios in these microparticles. Information must be obtained on individual particles and this implies that a high number of particles should be measured. Chemical techniques are unable to meet these requirements since their sensitivity is several orders of magnitude too low. Thus we developped ion probe analysis of single grains. Spatial resolution must be better than 1µm and a high transmission efficiency is required due to the weak signal. Analyses are performed by isotopic imaging on several microparticles with a high resolution and image processing to extract the isotopic ratios directly from the images.

The sample is rastered over 25µm by 25µm. Secondary ions are counted on an electron multiplier. Spatial resolution and ion yield are linked to the primary beam, so a <1µm focused Cs+ beam with a primary current around 5-10 pA is used. Image processing is automated, to compute the images at high speed just after acquisition, to extract geometrical parameters of the particles, sort them and eliminate abberations. Acquisition time are 1 hour to obtain both ¹⁶O and ¹⁸O images. To enhance the signal it could be repeated until particles are completely eroded by the beam. With these settings a counting statistic of around ±7‰ is obtained on 0.3 - 1 µm diameter feldspar grains. The major difficulties are due to the strong instrumental fractionations which occur due to charging and relief effects which if not corrected induce an error of ±10-15‰.

Alpert P, Kaufman YJ, Shay-El Y, Tanre D, da Silva A, Schubert S & Joseph JH, Quantification of dust-forced heating of the lower troposphere, Nature, 395, 367-370, (1998).

Clayton RN, Rex RW, Syers JK & Jackson ML, Oxygen isotope abundance in quartz from Pacific pelagic sediments, J. Geophys. Res, 77, 21, 3907-3915, (1972).

Metamorphic Fluids Sampled by Microdiamonds

Katrien De Corte (kdecorte@africamuseum.be)¹,

Wayne Taylor (wayne.taylor@anu.edu.au)²,

Paul De Paepe (paul.depaepe@rug.ac.be)¹ &

Vladislav Shatsky (shatsky@uiggm.nsc.ru)³

¹ Department of Geology and Soil Science, University of Ghent, Belgium

² Research School of Earth Sciences, Australian National University, Canberra, Australia

³ Institue of Mineralogy and petrography, Novosibirsk, Russia

The application of FTIR spectroscopy to microdiamonds of ultrahigh pressure metamorphic rocks of the Kokchetav Massif (Northern Kazakhstan) revealed the presence of water- and carbonate-rich inclusions indicating diamond formation from a hydrous/carbonatitic fluid (De Corte et al., 1999). To collect additional information about the fluid composition, diamonds (<120µm) from garnet clinopyroxenite and garnet-clinopyroxene dolomitic rock were polished and subsequently studied using microprobe and UV-laser-ICPMS techniques. The new data suggest that the aqueous component of the diamond-related fluids are metamorphic brines and that the ionic components (K, Ba, Na, Ca, Fe, Zn, Pb etc) most probably derive from the local rock environment. Detailled information on oxygen fugacity ranges and diamond precipitation reactions will be discussed during the conference.

De Corte K, Cartigny P, Shatsky VS, Sobolev NV & Javoy M, GCA, in press, February, (1999).

Session O10:5P

O10 : 5P/01 : PO

Relationships between Pedogenesis and Trace Elements Geochemistry in Nested Alterites ­ An Example from Normandy, France

Damien Butaeye (butaeye@geos.unicaen.fr)¹,

Anne-Veronique Walter (walter@geos.unicaen.fr)¹,

Jean-Pierre Camuzard (camuzard@geos.unicaen.fr)¹ &

Jean Le Gall (+33 (0)2 31 56 53 78)²

¹ UPRES-A CNRS 6143, Centre de Géomorphologie, Université de Caen, 24 rue des Tilleuls, 14000 Caen, France

² GEGER, Université de Caen, Esplanade de la Paix, 14032 Caen cedex, France

Detailed granulometric, mineralogical and geochemical studies have been carried out on a weathering profile developed on paleozoic quartzite called «Grès Armoricains». Samples have been collected in the Ecouves Massif, the highest point (417 m) of the Basse-Normandie region (France). The weathering profile is 3 meters thick and exhibits six main horizons, affected by recent cryoturbation processes. The lowermost horizon corresponds to the fresh, unweathered, parent quartzite. The second horizon is a thin layer solely made up of illite. The third horizon is a truncated paleosoil (so-called paleosoil 1), with brecciated structures, ferralitic characteristics and alternations of silica dissolution/cristallisation processes. The fourth horizon is another truncated paleosoil (so-called paleosoil 2) that presents brown soil and leaching characteristics. The boundary between paleosoil 1 and paleosoil 2 is sharp and well defined by a sandy layer (fifth horizon) and hardening of the top of paleosoil 1. The uppermost horizon is a podzol.Major, trace and rare earth elements analyses show that the two paleosoils are different but both enriched compared to the quartzite and the sandy layer (fifth horizon). Principal components analysis of trace elements amounts allows us to distinguish the two paleosoils. Paleosoil 1, the fifth horizon and the quartzite have close trace elements distributions, confirming that paleosoil 1 and the fifth horizon may be the result of the quartzite in-situ weathering, even if the fifth horizon has been further leached. In return, paleosoil 2 and the podzol show two distinct and clearly different trace elements distributions. Rare earth elements (REE) patterns show that the fifth horizon is REE-depleted compared to the quartzite, with a slight relative enrichment in light-REE (LREE) compared to heavy-REE (HREE). In return, paleosoil 1 is REE-enriched compared to the quartzite, with a slight relative enrichment in middle-REE compared to LREE and HREE. Paleosoil 2 shows a slight LREE- and a marked HREE-enrichment compared to the quartzite. Except Ce and Eu, the REE distributions are homogeneous within each paleosoil. These analyses confirm the truncations of paleosoil 1 and paleosoil 2 (lack of a LREE-enriched layers) and their distinct origins.

O10 : 5P/02 : PO

Hf Isotope Analysis of Kimberlite Megacrysts by Laser Ablation- and Solution-Mode Plasma Ionisation Multicollector Mass Spectrometry (PIMMS): Evidence for a Non-Asthenospheric Deep Mantle Component in Kimberlites

Geoff Nowell (gmn@wpo.nerc.ac.uk)¹,

Graham Pearson (d.g.pearson@durham.ac.uk)²,

Richard Carlson (carlson@dtm.ciw.edu)³,

Dave Bell (DBELL@geology.uct.ac.za)⁴,

Robert Zartman (RZART@geology.uct.ac.za)⁴,

Phill Shaw⁵ &

Ian Bowen (vgsectors@compuserve.com)⁵

¹ NIGL, Kingsley Dunham Centre, Keyworth, NG12 5GG, UK

² Dept. of Geological Sciences, University of Durham, Durham, DH1 3LE, UK

³ Carnegie Institution of Washington, 5241 Broad Branch Road N.W., Washington, D.C., 20015, USA

⁴ Dept. of Geological Sciences, University of Cape Town, Rondebosch 7700, SA

⁵ VG Elemental, Ion Path, Road 3, Winsford, CW7 3BX, UK

Silicate and oxide minerals of the Cr-poor megacryst assemblage (CPMA) are characteristic of group I kimberlites worldwide. High P-T equilibrium conditions and trace element compositions of megacrysts suggest they crystallised from an OIB-like parental magma at the base of the lithospheric mantle. The parental megacryst magma, sometimes referred to as a 'protokimberlite', may eventually evolve toward kimberlitic compositions through the interaction and assimilation of enriched lithospheric mantle. Megacrysts therefore provide an important opportunity to determine the isotopic composition and evolution of precursor kimberlite magmas prior to lithospheric contamination. Previous isotopic studies on the CPMA have focussed on mainly Sr and Nd and were restricted in their coverage because only garnet and clinopyroxene had sufficient concentrations of these elements to permit isotopic analysis. Such restrictions do not apply to Hf, which is present in most megacryst mineral phases at concentrations sufficient for straight forward isotopic analysis of virtually the whole CPMA with the new generation Plasma Ionisation Multi-collector Mass Spectrometers (PIMMS).

Previous Hf isotopic studies of South African kimberlites reveal a unique source component that plots well below the mantle array, conjectured to represent a deep Earth reservoir (Nowell et al in press). To further constrain the isotopic composition of the 'protokimberlite' magma we have determined the Hf isotopic composition of a variety of CPMA mineral phases from the Monastery and other southern African kimberlites on a VG-Elemental P54. Analyses of ilmenite, garnet and clinopyroxene megacrysts were carried out by solution mode-PIMMS while mono-mineralic zircons and zircon-ilmenite intergrowths were analysed by both solution- and laser ablation-PIMMS. The first Nd isotope analyses of mantle ilmenites, obtained by thermal ionisation mass spectrometry, will also be presented. The CPMA has a limited range in <epsilon>Hf, -2.8 to 2.9. There is no clear systematic isotopic variation within the megacryst crystallisation sequence. Furthermore, like their host kimberlites, the CPMA is characterised by some of the most extreme displacements of any terrestrial samples below the mantle array in <epsilon>Hf-<epsilon>Nd space (Nowell et al., 1998).

Our Hf data gives no indication for any significant isotopic evolution of the parental CPMA magma during crystallisation of the CPMA due to lithospheric contamination. The CPMA therefore appear to give a true reflection of the isotopic composition of a magma, perhaps a 'protokimberlite', which was derived from sub-lithospheric mantle depths. It is also clear that this 'protokimberlite' magma was isotopically unlike any melt known to be derived from the convecting mantle. We argue that a distinct component, lying below the mantle array in <epsilon>Hf-<epsilon>Nd space, must exist within the deep mantle. The most plausible candidate for this component will be shown to be old subducted oceanic crust, necessarily generated in the presence of residual garnet, stored in a deep mantle boundary layer, perhaps at the 650 km discontinuity or probably D" layer.

Nowell GM, Pearson DG, Kempton PD, Noble SR, & Smith CB, 7th Int. Kimb. Conf. volume. Cape Town, (in press).

Nowell GM, Pearson DG, Kempton PD, Noble SR, & Smith CB, Ext. Absts. 7th Int. Kimb. Conf. Cape Town, 640-642, (1998).

O10 : 5P/03 : PO

Genesis of Late Proterozoic Iron Formations in Namibia

Tim Vietor (tvietor@gwdg.de) &

Klaus Weber (kweber@gwdg.de)

IGDL, Goldschmidtstr.3, D-37077 Goettingen, Germany

The iron-formations at the base of the Late Proterozoic Damara Sequence in Namibia formed between 742 and 600 Ma. Geochemical and mapping results suggest a genesis by upwelling of suboxic, iron-rich deep basin waters onto the continental shelf during periods of global glaciation. Field evidence prooves the glaciogenic origin of the up to 150 m thick iron-formations: most of the iron-rich horizons are dropstone-bearing. The lithofacies of the surrounding sediments indicates that their deposition coincides with the beginning of a transgression after a major eustatic regression. REE data of the iron-formations lacking a negative Ce-anomaly indicate precipitation of iron hydroxides from suboxic waters which in turn require a stratified water column. Hydrothermal input appears to be insignificant compared to Archean and Early Proterozoic iron formations since a positive Eu anomaly is absent in the data. The chemical properties of iron limit the range of possible conditions for its transport in sea water. Sulfate concentrations in the Late Proterozoic oceans had already reached today´s values. High organic productivity in the upper portions of a stratified water column will invariably lead to the reduction of sea water sulfate and buildup of euxinic conditions by oxidation of organic matter. Under these conditions the low solubility of iron-sulfides prevents the transport of Fe(II). Thus, low organic productivity is a prerequisite for the transport of iron and the deposition of iron formations in the Late Proterozoic. Late Proterozoic "snowball earth" conditions hampered organic productivity in the oceans in different ways. Widespread continental glaciation limited transport of nutrients into the sea and potential highly productive continental shelves were covered with ice sheets even at low latitudes. The phosphorus concentrations of the iron formations are up to three times higher than in the surrounding sediments and are positively correlated with the iron concentrations of the rocks. It appears possible that during the deposition of the iron formations the scavenging of phosphate on the hydroxide surfaces further limited the amount of phosphorus available in the surface waters. Not only the stagnant sea water bodies under widespread ice-sheets but also the limitation of organic input during glacial periods were required for the deposition of iron formations in the Late Proterozoic.

O10 : 5P/04 : PO

Stable Carbon and Nitrogen Isotope Composition of Organic Matter from the Toarcian of the Italian Northern Apennines (Central Italy, Umbria-Marche Basin)

Antonio Bacheca (antonio.bacheca@imp.unil.ch)¹,

Elena Morettini²,

Peter Baumgartner² &

Johannes Hunziker¹

¹ Inst. Minèralogie et Petrographie, BFSH-2, 1015 Lausanne - CH, Switzerland

² Inst. Gèologie et Paleontologie, BFSH-2, 1015 Lausanne - CH, Switzerland

Stable isotope composition of carbon and nitrogen, C/N ratios in organic matter of sediments from the Lower Toarcian (Tenuicostatum zone) show that the Umbria-Marche black-shale horizon was deposited under conditions of reduced fertility/productivity with respect to the above well oxygenated formations. The Toarcian organic-carbon-rich beds were found to have lower ¹⁵N and lower ¹³C values with respect to organic-carbon-poor beds, that might represent higher water fertility at times of more vigorous circulation and of efficient recycling of nutrients in the water column. The range of organic ¹³C shows distinct very low (¹³C_org-33 per mill) and heavier (¹³C_org-24 per mill) values, that we interpret as reflecting abundance variations in the phytoplankton communities, rather than changes in the source. Paleoproductivity variation was the main controlling factor over organic matter source and accumulation.

O10 : 5P/05 : PO

Light and Transition Element Compositions of Garnet and Clinopyroxene from Eclogite Xenoliths from the Eastern Finland Kimberlite Province

A. B. Woodland

(alan@classic.min.uni-heidelberg.de)¹,

H. -M. Seitz

(mseitz@classic.min.uni-heidelberg.de)¹ &

P. Peltonen (petri.peltonen@gsf.fi)²

¹ Mineralogisches Institut, Universität Heidelberg, D-69120 Heidelberg, Germany

² Geological Survey of Finland, FIN-02150 Espoo, Finland

Four eclogite xenoliths have been recovered from the Lahtojoki kimberlite pipe (no.7) that intruded the late Archean craton in eastern Finland during the early Paleozoic. This pipe lies close to the SW margin of the craton, adjacent to the Proterozoic (=1.9 Ga) Svecofennian mobile belt. The four xenoliths are composed of garnet (gt) and clinopyroxene (cpx); two are diamondiferous.

Abundances of Li, B, P, Sc, Ti, V, Cr, Mn, Co and Ni in gt and cpx were measured by SIMS. In addition, Fe³⁺-contents of gt and cpx were determined from Mössbauer spectroscopy. Preliminary measurement of Fe³⁺/Fe²⁺ by EELS using a TEM (van Aken et al. 1998) is in excellent agreement with the Mössbauer results, suggesting this microscopic approach can be applied to natural samples. Major and trace element contents are relatively homogeneous, but differ quite significantly between samples; two samples are group A and two are group B eclogites. Except for the diamondiferous samples being somewhat richer in total Fe, no correlation exists between any major or trace element in gt or cpx and the occurrence of diamond. The xenoliths yield gt-cpx temperatures (Krogh,'88) of 1300-1400°C, at 6.0 GPa. Sc contents, 44-125 ppm in gt and 6-21 ppm in cpx, overlap with, but tend to be higher than those reported for Yakutian eclogites (Snyder et al. 1997). Thermometry based upon Sc-partitioning systematics between gt and cpx (Seitz, unpubl) yields temperatures basically in agreement with the thermometer of Krogh (1988). Gt and cpx contain similar amounts of V (100-300 ppm), but Ni and Ti are 2-3 times higher in cpx (Ni: 425-575 ppm in cpx vs. 150-225 ppm in gt). Co contents of both gt and cpx have the least variation of any trace element; =60 ppm in gt and =37 ppm in cpx. Gt contains 0.06-0.19 wt% Na, which far exceeds P (120-200 ppm), but there is ample Ti to provide a charge balanced substitution. The Fe³⁺/(sum)Fe ratios in gt (0.06-0.08) are like those obtained for gt-peridotites from the same pipe. In contrast, Fe³⁺/(sum)Fe ratios in cpx (0.09-0.17) are less than those observed from the gt-peridotites (Woodland & Peltonen 1998), which is consistent with increased partitioning of Fe³⁺ into gt with increasing temperature; the eclogites record higher temperatures than the peridotites. Unfortunately no reliable estimate of ƒO₂ can be made for the eclogite assemblage. Cpx is the dominant host for Li (500-800 ppb). The Li contents (Seitz & Woodland this meeting), along with the high equilibration temperatures point to a mantle origin for these eclogites, rather than subducted oceanic crust.

Krogh EJ, Contrib. Mineral. Petrol, 99, 44-48, (1988).

Snyder GA, Taylor LA, Crozaz G, Halliday AN, Beard BL, Sobolev VN & Sobolev NV, J. Petrol, 38, 85-113, (1997).

van Aken PA, Liebscher B & Styrsa VJ, Phys. Chem. Minerals, 25, 323-327, (1998).

Woodland AB & Peltonen P, Proc. 7th. Kimberlite Conf, Cape Town S. Africa, (in press).

O10 : 5P/06 : PO

Determination of Non-Ideal Solution Parameters of Hydrous Melts from Phase Equilibria in the Haplogranite System Ab-Or-Qtz-H₂O

Marcus Kirschen (kirschen@cnrs-orleans.fr)¹,

Christian DeCapitani (capi@therion.minpet.unibas.ch)² &

Michel Pichavant¹

¹ CNRS-CRSCM, 1A rue de la Ferollerie, F-75071 Orleans Cedex 2, France

² Min.-Pet.Institut, Bernooullistrasse 30, CH-4056 Basel, Switzerland

The computation of equilibrium phase diagrams of hydrous silicate systems requires a model of the Gibbs free energy of the melt defining the activity of the solution end-members at given composition, pressure, temperature. There is limited experimental indication showing that Henry's law is obeyed for volatile components like H₂O and CO₂. Extension to more complex systems was made assuming ideal mixing of the melt species normalized to 8 oxygens (Burnham & Nekvasil 1986). Recent developments include water speciation (e.g. Stolper 1989), higher order excess polynomials (e.g. Blencoe 1992) and non-ideal solution parameters between melt species including water (Ghiorso 1995). Systematic data have been obtained on the effect of water on solid-liquid equilibria and the effect of composition, pressure and temperature on water solubility in the system Ab-Or-Qtz-H₂O (e.g. Holtz et al. 1995). Therefore it seems appropriate to use this system (1) to test existing solution models and (2) to develop thermodynamic modelling of hydrous melts.Non-ideal solution parameters for the silicate liquid were derived simultaneously from liquidus and solubility data using mathematical programming techniques (e.g. Murtagh & Saunders 1987). For this purpose a Fortran program was written that converts experimental liquidus (and solvus) data to linear constraints of the optimization problem using bracketing relations _rG <> 0. The use of a highly symbolic input code of experimental results improves significantly the transparency of the optimization problem. The linear problem is changed to a linearly constrained optimization when constraints from solubility and/or activity measurements are available. Margules-type polynomials were used to fit the excess free energy of the hydrous melt as a first approach. Predictions of existing solution models will be compared with results from this study.

Blencoe JG, Trans. Royal Soc. Edingb., 83, 423-428, (1992).

Burnham CW & Nekvasil H, Am. Mineral, 71, 239-263, (1986).

Ghiorso M, Contrib. Mineral Petrol, 119, 197-212, (1995).

Holtz F, Behrens H, Dingwell DB & Johannes W, Am. Mineral, 80, 94-108, (1995).

Stolper E, Am. Mineral, 74, 1247-1257, (1989).

O10 : 5P/07 : PO

Geochemistry and Mineralogy of the Zone of Interformational Contact on the Pre-Polar Urals

Irina Kozyreva (yudovich@geo.komi.ru),

Yakov Yudovich (yudovich@geo.komi.ru),

Irina Shvetstova (yudovich@geo.komi.ru) &

Marina Ketris (yudovich@geo.komi.ru)

Syktyvkar, Pervomayskaya st., 54, Institute of Geology Komi SC, Russia

The zone of interformational contact (ZIC) of the Baikalian complex of pre-uralides (foundation) and Caledonian-Hercinian complex of uralides (mantles) is observed on the Pre-Polar Urals (Upper Kozhym river). The upland rocks in the ZIC are represented by metamorfic shales different in composition and genesis. These are rhyolites and diabases of foundation, modified in different intensity in Cambrian crust of weathering (CW), also Alkesvozh Series of the basement of uralides (Cm₃-O,al) and superposed quartzite-sandstone of Telposs Series (O₁tp).The substance of metamorphites in ZIC, is original geochemical phenomenon because of its unique paragenesis of the chemical elemens. In particular, shales shaped on the substratum of rhyolites include paradoxical combinationof volatile elements of greisen paragenesis (Ge, Sn, W, Mo, Be, As, Bi, F, B), elements-hydrolizates typical for CW (Fe³⁺, Al, Ga, TR, Th), elements of mantle paragenesis (Mn, Cr, Pd, Hg, Au). The sum of TR reaches 1% and content of Au is n·g/t. The shales shaped on the substratum of basic rocks show powerfull content of Fe₂O₃, TiO₂ and K₂O. The unique Mn-Zr-chrome-spinelli have been discovered in auriferous Alkesvozh Series. The aporhyolitic shales are characterized by uniform complex of accessory minerals: hematite, chloritoid, pyrophyllite and diaspore (sometimes they are rock-forming), leucoxene, zircon, epidote (sometimes allanite), monazite, yttrotitanite, xenotime, tourmaline, cassiterite. These rocks also contain series of rare-earth minerals, very rare (for example arsenate of yttrium - chernovite) or even new minerals related to little-know series: phosphates - arsenates, tungstates (molybdates) - arsenates, alumophosphates - alumoarsenates. The apobasic shales contain hematite, apoilmenitic leucoxene and its crystallization products, allogenic and authigenic zircons, two varieties of apatite, epidotes, tourmaline, chloritoid. The extremely high content of TR and new varieties of high-manganous epidote have been found in unique monazite-epidote-quartz concretions of unknown genesis (glacial circus of Grubependity lake). All of it speaks about the presence of the material of old (Cambrian) CW on basic rocks and part on rhyolites in shales of ZIC. Presence of Late Paleozoic (?) metasomatism is also probable. The established geochemical and mineralogical features of this rocks are result of the long (Cambrian - Permian) and multistage exogenetic and endogenetic processes.

O10 : 5P/08 : PO

Sedimentation Rates in the Lake Chapala (Mexico) by Radiometric Methods

Hermes Ulises Ramirez Sanchez (ramirez@cerege.fr)¹,

Pedro F. Zarate del Valle (zarate@cencar.udg.mx),

Francois Michaud (micho@ccrv.obs-vlfr.fr)²,

Francois Fernex (fernex@unice.fr)³,

Josette Dalmasso (dalmasso@unice.fr)⁴ &

Genevieve Barci-Funel (barci@unice.fr)⁴

¹ CEREGE Pôle de activite commercialle de l'arbois, 13545 BP 80 Aix en Provence CEDEX 4, France

² Observatoire oceanologique de Villefranche, bp 48 - 06230 Villefranche sur mer, France

³ Laboratoire de Geochimique isotopique, Universite de Nice

⁴ Laboratoire de Radioecologie et Radiochimie, Universite de Nice

Chapala Lake is the most important mexican lake (80 km x 20 km) wich has only 6 m de deep. The very shallow deep is a particuler case for a tropical lake located in a volcanic region. A research to characterize the sediments and the sediments and the sedimentation rate (SR) of Chapala Lake is in progress. Sediments has roughly 70-80% of humidity and 20% of organic matter. 12 gravitational logs covering the whole lake were analized by radiometric methods: ²¹⁰Pb, ¹³⁷Cs, and ^239-240Pu. The major quantity of ²¹⁰Pb is present in atmosphere by decay of ²²⁶Rn. ²¹⁰Pb profils shows an exponentiel form tipical of radioactivite decay with a periode of 22.3 years. This fact let us to evaluate the SR in four logs. ¹³⁷Cs y ^239-240Pu are artificial radionucleids that resultat from the atmospheric nuclear essays developping in 50 and 60's with a maximal activity in 1963. six longs covering the lake were analized by these radionucleids and its profils show maxima activity picks wich levels correspond to sediments deposited during this period. SR trough Chapala lake is not uniform and it varies as follows: on east side at Lerma river mouth and 10 km inside the lake, the SR has a value of 4 mm/yr,whereas 17 km inside it reaches 6 mm/yr. In central part we have determined a SR of 2 mm/yr, having a little variation in central south part where we have gotten a SR of 3.5 mm/yr. Finally on western part SR has a value <2 mm/yr. On the other hand and based on the statigraphic description of 6 logs and helped by magnetic susceptibility values, we have identified irregularites in sedimentation through the whole lake. These irregularites are linked to slumping, bioturbation and fishing activities. After data, SR on est side of Chapala lake es higher in three times than in western side, so deep on western side must be minor, fact thet doesn't occur. So we conclude that there must be a tectonic compensation that explain why the Chapala lake botttom has an uniform deep in spite of the different SR determined.

O10 : 5P/09 : PO

Isotopic and Trace Element Geochemistry of Quaternary Volcanism in Kula (Western Anatolia, Turkey): Implications for the Involvement of Asthenospheric and Lithospheric Mantle in Their Genesis

Pinar Alici (pinar@jeo.hun.edu.tr)¹,

Abidin Temel &

Alain Gourgaud

(gourgaud@opgc.univ-bpclermont.fr)²

¹ Hacettepe University, Geological Engineering Department, 06532, Beytepe-Ankara, Turkey

² Université Blaise Pascal, 5 rue Kessler, 63038, Clermont-Ferrand Cedex, France

The Quaternary alkaline volcanism occuring in Kula region is situated in the western Anatolian graben system which developed as a result of Aegean extensional regime. Kula volcanic rocks are Ne-normative silica undersaturated, and can be classified as basanite, tephrite and tephriphonolite. The K₂O/Na₂O ratio of the lavas ranges between 0.3 and 0.7 that indicates their sodic nature. Certain features of the silica and magnesium variation diagrams are consistent with fractional crystallization trends, and the Kula volcanic rocks have Mg-number <0.63 and Ni<97 ppm contents too low to represent primary magmas. Almost all having undergone fractionation of olivine + clinopyroxene + amphibole ± plagioclase.

To determine the source characteristics of the Kula volcanic rocks, major, trace and rare earth elements as well as Sr, Nd and Pb isotopic compositions have been analysed. They are characterized by significant enrichment in HFS elements (such as Nb and Ti) relative to LILE. ⁸⁷Sr/⁸⁶Sr, ¹⁴³Nd/¹⁴⁴Nd, ²⁰⁶Pb/²⁰⁴Pb, ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb isotopic compositions for Kula range between 0.703029 - 0.70349; 0.512773 - 0.512941; 18.689 - 19.064; 15.606- 15.683; 38.56 - 39.113, respectively. The Sr isotopic compositions are depleted relative to the bulk-earth, whereas Nd isotopes are relatively enriched. All the samples plot close to the mantle array, indicating derivation from an isotopically depleted mantle source. Trace element ratios and isotopic compositions compare well with those of OIB (such as St. Helena and Tubuai) and alkalic basalts from the Basin and Range. The chemical characteristics of the Kula volcanics indicate that they are derived from a within-plate source region with low Y/Nb <0.45 ratios. The Zr/Ba ratio is also consistent with derivation from an OIB-like source but the range of Y/Nb, as well as Zr/Ba ratios are lower than the typical alkalic-ocean island basalts. Thus, the low Y/Nb and Zr/Ba ratios relative to the OIB in the Kula volcanics imply that their source region is more enriched than the OIB-like source. Furthermore, incompatible trace elements exhibit patterns similar in many respect to those of St. Helena, Tubuai and Basin and Range, but enrichment in Sr, Ba, Rb and K in the Kula volcanics is higher than that of those suites. We, therefore, attribute these high contents of various incompatible elements and low Y/Nb and Zr/Ba ratios relative to the ocean-island basalts, to the mixing of ascended melts mainly from OIB-like asthenospheric mantle with enriched lithosphere.

O10 : 5P/10 : PO

Meliata Ophiolites: the Only Representant of Mesozoic Oceanic Crust in the Western Carpathians

Dusan Hovorka (mineralogia@fns.uniba.sk)¹ &

Ján Spisiak (spisiak@gu.bb.sanet.sk)²

¹ Comenius University, Mlynská dolina, 842 15 Bratislava, Slovak Republic

² Geological Institute of the Slovak Academy of Sciences, Severná 5, 974 01 Banská Bystrica, Slovak Republic

For the Western Carpathians Mesozoic great diversity among lithofacial development is reported. The majority of the Mesozoic units in this segment of the Alpine belt is characterized by shallow water sedimentation on the slopes of the European continent as well as on the slopes of Apulia promontory of the African block. The Meliata unit Mesozoic represents the only one of deep water sedimentation (oceanic) type. Oceanic crust of the Meliaticum, except of the oceanic (deep water) sedimentary pile, is represented by dismembered ophiolite complex. Its members are namely those of intensively serpentinized peridotites, less abundant gabbros, tholeiitic basalts and radiolarites. Part of the Meliata unit rock pile recrystallized under high-pressure (blueschist) conditions. Dismembered Mesozoic Meliaticum ophiolites crop out in several partial tectonic units, in which various vertical segments of the ophiolite complex are present. The Meliata Unit is known to occur in both, the Slovak and Hungarian territories. Stratigraphic span of the Meliata unit is Permian-Jurassic. But it should be stressed here that during the Permian and the lowermost Triassic (Skytian to Pelsonian) geological development of the Meliaticum s. s. and units in its periphery, was identical. Since that time development underwent in different environmental condition. Except of pronouncedly oceanic development of the axial zone of newly opened rift structure (e.g. Meliaticum s. s.), on slopes and shelves of the adjacent continental blocks several transitional faciesses developed. Stratigraphy of the Meliata Unit is based mostly on conodont and radialaria. From the stratigraphical and lithological points of view it is considered that development of the Meliata oceanic lithosphere begun in the Middle Triassic by the carbonate platform collapse - since that time individualized development of the Meliata rift zone (basin) is recorded. Meliata unit represents an olistostroma with the Middle to Upper Triassic mostly carbonates olistoliths. The matrix of this olistostroma unit consists of turbidite shale and aleurolite of the Upper Triassic to Middle Jurassic - Oxfordian stratigraphy. Closure of the Meliata ocean is documented by the blueschist recrystallization of its sedimentary and magmatic sequences (= 160 Ma).

O10 : 5P/11 : PO

Distribution of Methane in the Southern Atlantic Ocean

Katja Heeschen (kheeschen@geomar.de)¹,

Gregor Rehder (grehder@geomar.de)¹,

Robin Keir (rkeir@geomar.de)¹,

Christian Rodehacke (c04@zfn.uni-bremen)² &

Erwin Suess (esuess@geomar.de)¹

¹ Geomar Research Center for Marine Geosciences, Wischhofstr. 1-3, 24148 Kiel, Germany

² Department for Environmental Physics, University of Bremen, P.O.Box 330440, 28334 Bremen, Germany

The mixing ratio of methane in the atmosphere has been rising over the last 150 years, and if preformed concentrations in deep water formation areas are linked to the atmosphere, methane in young deep waters may be affected by the atmospheric rise. During April and May 1998 dissolved methane and chlorofluorocarbon CCl₃F (CFC11) were measured on hydrographical sections in the Weddell Sea and Weddell-Scotia Confluence and along the Greenwich meridian between Antarctica and Capetown. In this region, a mixture of ventilated shelf waters and old Warm Deep Water (WDW) contributes to the formation of Weddell Sea Deep Water (WSDW) and Weddell Sea Bottom Water (WSBW). CFCÕs have only been released into the atmosphere over the last 50 years and are chemically inert in the ocean. By comparing the methane and CFC distribution patterns, one obtains an indication of both the atmospheric effect and the rate of methane oxidation. The highest concentrations of methane in deep waters were found in WSBW formed along the Antarctic Peninsula. Elevated values were also observed in bottom waters of the Weddell Sea and in young water masses in the Powell and Jane basins. The distribution patterns of methane and CFC11 were remarkably similar, with the methane saturations being somewhat greater than those of CFC11. Surface water concentrations of methane as well as CFC11 were undersaturated relative to their atmospheric partial pressures. This is in contrast to most other surface waters, where methane is generally oversaturated or in equilibrium with the atmosphere.

O10 : 5P/12 : PO

Formation of Trona Deposite in Beypazari District (Turkey): An Experimental Approach

Y. Burkut (oredgeo@itu.edu.tr),

F. Suner (Suner@itu.edu.tr) &

V. Esenli (Esenli@itu.edu.tr)

Mining Fac. Tech Univ. of Istanbul, Turkey

In the presented paper, it was studied precipitation conditions of Beypazari deposite, which is one of the most interesting Trona formation in the World with 200 million metric tons of proven reserve, under the control of temperature and Eh - Ph parameters. For this aim, Trona, nahcolite and natron, the main minerals in the deposits, were dissolved in the purewater and saturated lake-water conditions were obtained approximately. Then, steps of precipitation were controlled by a thermo-element containing pH-Eh meter and calculated Eh/pH values were plotted on the following diagram below. Obtained data show that natron, rarely observed in the deposit, had been precipitated oxic and anoxic environments at a temperature interval of 60-17°C: The pH and Eh intervals of this formations are about 10-10.5, -80+240 mV respectively. Moreover, the main precipitation was determined at pH of 10.5. This result was supported by the examination of thin section, the presence of Natron which was crystallised at first from the saturated solution. Nahcolite, determined more often in the deposits, and the dominant mineral Trona exhibit a similar and co-precipitation period. Although the crystalisation were happened under anoxic conditions and the intervals of temperature, pH and Eh were 17-60°C, 9.60-10.2 and 60-240 mV, respectively; the most favourable pH level for co-precipitation is 9.70-10.0, which is supported by the field and microscopic observations.

O10 : 5P/13 : PO

The Anisotrophy of Ino-Silicates: A Tool in the Interpretation of Dissolution Mechanisms

Paul Frogner (paul.frogner@geo.su.se)

Department of Geology and Geochemistry, Stockholm University, S-106 91 Stockholm, Sweden

Low temperature dissolution experiments of Amphibole (Frogner et al., 1998) suggests that the overall dissolution rate, which is measured in terms of Si release, is influenced by the formation of a leached layer. Freshly fractured surfaces initially have a composition and structure equivalent to that of the unreacted bulk, and consequently they have higher initial dissolution rates. However, as dissolution proceeds, a leached near-surface layer develops between the fluid/surface interface and the bulk, unreacted mineral. This near-surface layer based on calculated depleation depth from solution data becomes thicker as a function of time. The observed decrease in the overall rate of dissolution with time is hypothesised to be due to a continuous restructuring of the leached layer as it becomes more and more deficient in Al, Mg, and Fe. It is important to note that these leached layers are not only chemically distinct from the bulk, but their structures differ also. As the dissolution reaction progresses, this leached layer will finally reach a steady-state thickness; this occurs when the flux of cations diffusing through the leached layer equals the surface flux of Si, which is simply a function of the dissolution rate of the layer itself. This near-surface layer must thus be both tightly bound and continuous in order to provide an effective diffusion barrier (Blum, 1994). In addition preliminary SEM results also shows that the dissolution of actinolite (an amphibole) probably is favoured with respect to certain crystallographic directions. This means that leaching depth calculations for these kinds of minerals should take this parameter into consideration. It also connects the dissolution mechanism to the crystallographic direction that should be noticed with care as it indicates a weakness in the structural element elongated in one direction.

Blum AE, Feldspars and Their Reactions, Parsons I., (ed) NATO Advanced Study Inst., Kluwer Academic Pub., Netherlands, 595-629, (1994).

Frogner P & Schweda P, Chemical Geology, 151, 169-179, (1998).

O10 : 5P/14 : PO

Trace-Element Site-Partitioning in Olivine: Further Insights from Structure Refinement and Lattice-Site Elastic-Strain Modelling

Massimo Tiepolo Tiepolo (tiepolo@crystal.unipv.it)¹,

Piero Bottazzi (bottazzi@crystal.unipv.it)²,

Roberta Oberti (oberti@crystal.unipv.it)² &

Alberto Zanetti (zanetti@crystal.unipv.it)²

¹ Dip. di Scienze della Terra, Università di Pavia, Via Ferrata 1, I-27100 Pavia

² CNR-CS per la Cristallochimica e la Cristallografia, Via Ferrata 1, I-27100 Pavia

In order to characterise the trace-element site-preference in olivine and to decipher the structural control on trace-element olivine/melt partition coefficients, the lattice-site elastic-strain models were applied to a new experimental set of ^Ol/LD and the calculated lattice-site parameters were compared with the results of single-crystal X-ray structure refinement (SREF).

Olivine-melts pairs were obtained by crystallisation of basanitic composition doped with a trace-elements mixture at a pressure of 1.4 GPa and temperature of 1040°C. Olivine composition is close to Fo75. Trace element concentrations in both olivine and melt were determined both by SIMS and LAM-ICP-MS analyses. Cation partitioning for major elements was obtained by SREF. Among trivalent cations, V and Sc have the highest affinity for olivine, their partition coefficients approaching unity for the former and 0.5 for the latter. Among REE, only M-and HREE have concentrations above the detection limits. ^Ol/LD for HREE are much higher than for MREE. ^Ol/LD_Yb is roughly 10^-2 whereas ^Ol/LD_Sm is roughly 10^-4. Although ^Ol/LD_REE values are very low, REE fractionation is greater than in other upper-mantle phases such as clinopyroxene and garnet.

The ^Ol/LD obtained for all trivalent cations (Al, Sc, V and REE) were then plotted against ionic radii for 6-fold coordination; the results cannot be fitted by a single parabolic trend, suggesting that these elements partition themselves between the two independent octahedral sites (M1 and M2) in the olivine structure, which have different site dimensions (<M2-O> = 2.147 Å and <M1-O> = 2.117 Å, as determined by SREF). ^Ol/LD were fitted according to the elastic-strain theory under the assumption of a distribution of the trivalent cations among two different octahedral sites, and the lattice-site parameters were calculated. Although REE lay on one side of the parabola with the larger r0, we obtained a reasonable agreement between the calculated ideal ionic radii (r0) and the observed ones (r* = <cat-O> - 1.38 Å). A higher value of the Young's modulus was obtained for M1 than for M2, in keeping with the results for divalent cations.

We can thus conclude that REE are almost completely ordered at the larger M2 site in olivine. This is a further confirmation that site dimensions exert a strong control over REE partitioning. Conversely, transition elements distribute between the two independent octahedral sites, with a marked preference for the smaller M1 site (at least at high temperature); this is in agreement with what observed for Fe²⁺ (Artioli et al., 1995) and with the well known affinity between these chemical species.

Artioli G, Rinaldi R, Wilson C C and Zanazzi P F, Am. Mineral., 80,, 197-200, (1995).

O10 : 5P/15 : PO

REE Distribution between Aqueous Liquid and Water-Saturated CO₂ Gas at 140°C and 4 bar

Peter Möller (pemoe@gfz-potsdam.de)¹,

Nevzat Özgür² &

Marc A. Conrad (conrad@gfz-potsdam.de)¹

¹ GeoForschungsZentrum Potsdam, Am Telegrafenberg, Germany

² Freie Universität Berlin, Malteserstr. 74-100, 12249 Berlin, Germany

During four campaigns in the years 1995 to 1998 the artesian wells of the geothermal pilot plant at Kizildere/Turkey were sampled. According to the geology, the Na-HCO₃-CO₃-SO₂ type of water originate from the basement consisting of mica schists, gneisses and intercalated marbles. The derived bottom hole temperatures are 200 - 212°C at 60 to 70 bars. The wells are cased to depths of about 800 m and filter the CO₂-rich water at depths between 800-1200 m. The artesian wells produce a water-steam mixture of about 13 bar and about 140°C at the well head. Exsolution of CO₂ starts at depth of about 500 m and Sr-rich aragonite/calcite together with some silica precipitates. From the well head the water steam mixture passed a separator were the steam was separated from the liquid at pressures of 4.5 bars and 145°C. The water at the well head contains 2% CO₂. About 10 litres of water and condensate from the water-saturated CO₂ phase were collected in PE Carboys. Directly after collecting the fluids they were pressure-filtered into another PE carboy in which the fluids were adjusted to about pH 2 and a spike of 1 ml of 100 ppm Tm was added. The steam filters were white, where are most of the filters of the water were light grey, thereby indicating that some solid material is also transported. Since contamination of the water-saturated CO₂ phase by tiny water droplets cannot be excluded, Na and Ca contents were also determined. Na contents in the water and water-saturated CO₂ were about 1200 and <1 ppm, whereas Ca contents were about 1 and <0.5 ppm, respectively. Thus, Na indicates that the water-saturated gaseous phase at 140°C is contaminated by tiny droplets of water by less than 1‰ only. In contrast, Ca is present in the gaseous phase at rather high levels. Since solubility of Na is expected to be low in the gas phase under the given conditions, it is assumed that all residual Na present in the water condensed from the CO₂- saturated gas phase is due to contamination with the original liquid. Although this is an over-correction yielding too low distribution coefficients, it is considered to be a safe guess. The ratio of the REE abundances in the corrected condensate from the water-saturated CO₂ phase and the aqueous phase are in the range of 0.4 to 0.2 and show some continuous decrease from La to Lu. This indicates that light REE are preferentially enriched in the water-saturated CO₂ phase. This result could have some importance for REE transport in the earth's crust or even mantle. Carbonatites and giant REE-phosphate deposits (e.g. Bayan Obo) show steep REE patterns dominated by the light REE, gas transport of REE might have some bearing on their genesis.

O10 : 5P/16 : PO

Nb-Ta Geochemistry: New Constraints from ICP-MS Analyses

Florence Kalfoun

(kalfoun@dstu.univ-montp2.fr)¹,

Jean-Louis Bodinier

(bodin@dstu.univ-montp2.fr)¹,

Hans G. Barsczus

(barsczus@dstu.univ-montp2.fr)¹,

J. C. Lassiter (lassiter@clrs1.ciw.edu)²,

Liliane Savoyant (icpms@dstu.univ-montp2.fr)¹,

Olivier Alard (oalard@laurel.ocs.mq.edu.au),

T. Kogiso (kogiso@geo.titech.ac.jp)³,

Marguerite Godard

(godard@dstu.univ-montp2.fr)¹,

Rosa Maria Bedini (bedini@dstu.univ-mon) &

Claude Merlet (merlet@dstu.univ-mon)

¹ ISTEEM, CNRS et Université de Montpellier 2, Place E. Bataillon, Cc57, 34095 Montpellier cedex 05, France

² Department of Terrestrial Magnetism, Carnegie Institution of Washington, Washington, DC 20015, USA

³ Department of Earth and Planetary Sciences, Tokyo Institute of Technology, 2-12-1 Ookayama, Meguro, Tokyo 152-8551, Japan

We have analyzed Nb and Ta by ICP-MS in OIBs, continental crustal rocks and mantle peridotites, using surrogate calibration based on Zr and Hf measurements.

OIBs from French Polynesia, selected for their wide range of isotopic compositions, have Nb/Ta ratios in the range 13 - 18. Similar to the MORB value of 15.5 (± 1) proposed by Jochum and Hofmann (1998), our average OIB value (15.3 ± 1.1) is significantly lower than the chondritic, silicate-Earth value of 17.6. As suggested by Stolz et al. (1995), the low (subchondritic) Nb/Ta values of the oceanic basalts may be complementary of the high (suprachondritic) values found in several island-arc basalts. The hypothesis that the Nb/Ta ratio was lowered in the OIB source by addition of recycled oceanic lithosphere would account for the existence of positive correlations between Nb/Ta and radiogenic isotopes. For instance, in the Marquesas Islands, Nb/Ta decreases from 17 in the EM2-type basalts to 15 in the HIMU ones (Woodhead and Barsczus, in prep.).

Preliminary results confirm the subchondritic Nb/Ta signature of the continental crust (10.9 - Rudnick and Fountain, 1995). Yet, young (Phanerozoic) continental crust tends to show higher Nb/Ta values than older crust. In the Variscan belt of French Massif Central, the lowest Nb/Ta ratios (~ 11) are found in S-type granites and the highest values (16.5 - 17) in lower-crust xenoliths. The narrow range of subchondritic Nb/Ta values in continental crust contrasts with the extreme variation of this ratio in arc volcanics (12 to > 100 according to literature).

More than 300 samples of mantle rocks were analyzed for Nb and Ta, down to very low concentration levels (Ta ~ 0.4 ppb). Basalt-borne xenoliths and ophiolitic peridotites show a systematic variation of Nb/Ta in a range (8 - 90) similar to that of arc volcanics. In xenoliths Nb/Ta tends to be higher in LILE-enriched, HFSE-depleted samples while in the Oman ophiolite it is correlated with peridotite foliation dip. Abyssal peridotites show a comparatively restricted range of Nb/Ta values (13 ± 4), in spite of their high degree of alteration marked by elevated U/Th ratios (50 - 5000 x chondrites).

In conclusion, our results indicate that: (1) Nb/Ta is rather insensitive to high-temperature melt processes occuring in the convective mantle, as well as to low-temperature alteration; (2) yet, Nb and Ta were substantially fractionated from each other during silicate-Earth differentiation; (3) Nb-Ta fractionation occured in shallow, intra-plate and subarc mantle, and was associated with the circulation of volatile-rich small melt fractions - possibly saturated in Ti-oxides (Bodinier et al., 1996); (4) this process was probably involved in the segregation of the continental crust from the mantle, but the continents did not grow merely by accretion of island arcs.

Bodinier J-L, Merlet C, Bedini RM, Simien F, Remaidi M, Garrido CJ, Geochim. Cosmochim. Acta, 60, (1996)

O10 : 5P/17 : PO

Organic Substance in Natural Salts

Svetlana Shanina (shanina@geo.komi.ru)

54, Pervomayskaya st., Syktyvkar, 167610, Russia

Organic substance up to oil are widespread in natural salts. Analysed was, the composition of gases released at heating minerals from several salt deposits: Verkhnekamsk deposit of potassium salt (Western Urals); Silurian evaporates in Kochmes district in the Kosiyu-Rogovskaya depression (Komi); Seregovo salt dome (Komi); Karvakh salt dome (Yemen); salt from the salt lake of Larnaca (Cyprus); Suria deposit of potassium salt (Spain).

The following characteristics of volatile compounds were determined: the amount of (µg/g) nitrogen, carbon monoxide, carbon dioxide, water, hydrocarbons from methane to heptane, the sum of reduced gases, the degree of oxidation of the volatiles, n-butane/ i-butane ratio.

It is established, there is a good correlation between methane contents and the total amounts of heavy hydrocarbons up to heptane for Verkhnekamsk, Kochmes and Seregovo deposits. The ratio of n-butane to isobutane is found to be higher for Seregovo and Kochmes deposits (0.5-2) compared toVerkhnekamsk deposit (0.1-0.5). The diagram of hydrocarbon distribution suggest increased methane contents as compared with heavy hydrocarbons in Verkhnekamsk deposit, whereas heavy hydrocarbons prevail in Kochmes and Seregovo deposits.

The obtained data brings us to the following conclusions: (1) Organic matter in salts from Karvakh deposit is of heterogeneous nature: except their own organic matter, they contain hydrocarbons from the underlying oil deposits. (2) Organic substance in the samples from Verkhnekamsk and Suria salts is of homogeneous nature: here the source of organic gases is organic substance, scattered between the grains of salt. (3) The high contents of methane and reduced gases for the samples of Cyprus salt may be accounted for by the micro-organisms activity, since Larnaka salt lake is an open saline basin. (4) Organic substance in Kochmes and Seregovo salts, in all probability, is epigenetic, as their data points are situated close to those for Yemen and Cyprus samples.

O10 : 5P/18 : PO

Geological Standard Glasses for In-Situ Microanalysis

Brigitte Stoll (stoll@mpch-mainz.mpg.de)¹,

Klaus Peter Jochum (kpj@mpch-mainz.mpg.de)¹,

Ingrid Raczek¹,

Albrecht W. Hofmann

(Hofmann@mpch-mainz.mpg.de)¹ &

Donald Dingwell

(Don.Dingwell@uni-bayreuth.de)²

¹ MPI für Chemie, Postfach 3060, 55020 Mainz, Germany

² Universität Bayreuth, Postfach 101251, 95440 Bayreuth, Germany

Standard reference samples are important for the primary standardization of the analytical results of most microanalytical techniques. In geochemistry, mainly synthetic NIST glasses have been used for such purposes to correct for differences in ion formation, ablation behavior, transmission and detection of the various elements. However, the major element composition of the NIST glasses is quite different from any geological matrix.

Therefore, we have prepared eight glasses by direct fusion of 50 - 100 g chips of different rock type (basalt, andesite, rhyolite, peridotite, tonalite, komatiite) at temperatures in the range of 1400 - 1600°C. Glasses were held at temperature for 1 hour in a platinum crucible and then placed in a second furnace. During the second fusion, the melts were stirred for up to 12 hours. The melts were then quenched by placing the bottom of the Pt crucible in water.

The geological glasses were subsequently investigated by various bulk (MIC-SSMS, TIMS, INAA, XRF) und microanalytical (EMP, LA-ICPMS, SIMS, PIXE, LIMS, SY-XRF) techniques in different laboratories. Most elements are homogeneously distributed in the glass samples. Major and trace elements are constant within the precision of the microanalytical techniques used that is about 1 - 3%. Possible exceptions may be some elements (e.g. Pt, Mo, U) which were introduced during glass preparation.

We calc