Journal of Conference Abstracts

Session J04:1B

J04 : 1B/33 : F2

Determination of Tg in Glasses Under High Pressure from Electrical Impedance Measurements

Nickolai Bagdassarov

(nickbagd@geophysik.uni-frankfurt.de)

Institut für Meteorologie und Geophysik, Feldbergstraße 47, 60323 Frankfurt/Main, Germany

The pressure dependence of the glass transition temperature (Tg) in silicates correlates with a pressure dependence of viscosity (Rosenhauer et al., 1979). For example, albite glass has a negative pressure dependence of shear viscosity and Tg, diopside and sodium trisilicate glasses have positive pressure dependence of Tg and viscosity. Calorimetric and rheological measurements of Tg at high pressures are complicated to carry out. Electrical conductivity measurements are easier to conduct at high pressures. By means of electrical impedance measurements Tg in glasses may be well estimated from the temperature dependence of impedance. Below Tg activation energy of electrical conductivity has smaller slope than at T> Tg. The intersection of two Arrhenious dependencies of electrical conductivity as function of 1/T defines Tg. In this study a pressure dependence of Tg has been estimated for anorthite from electrical impedance measurements in a 1/2-inch piston-cylinder apparatus under pressures up to 20 bar. The pressure cell consisted of Caf₂, graphite and bornitride. The cell for electrical represents a coaxial cylindrical capacitor made of Pt foil with a geometric factor about 8 cm. The pressure calibration has been done by measuring of melting points for NaCl and CsCl by means of electrical resistance measurements. Tg at 1 bar has been measured in a tube furnace purged with Ar-gas. Electrical impedance measurements were conducted in the frequency range 100 kHz - 0.02 Hz with "Solartron 1260" Gain-Phase Analyser. The measured Tg at 1 bar is 810°C and corresponds to the change of the activation energy of the electrical conductivity measured on 1 Hz from ca. 50 kJ/mol at T<Tg) to ca. 100 kJ/mol at T>Tg. At 7 kbar Tg is about 825°C, and the difference in activation energies of the electrical conductivity above and below Tg decreases. At pressures above 20 kbar the difference in activation energies becomes indistinguishable.

Rosenhauer M, Scarfe CM & Virgo D, Carnegie Institute Yearbook 1978-79, 557 - 559, (1979).

J04 : 1B/34 : F2

Geospeedometry of Natural Glasses

Donald B. Dingwell

(don.dingwell@uni-bayreuth.de) &

J. Gottsmann

BGI/Uni Bayreuth, Bayreuth, Germany

The generation of glasses through natural processes most commonly occurs through cooling from the melt phase. Knowledge of the rate of cooling involved in glass generation can provide a valuable constraint on the thermal history of the glassy phase. Such a constraint can be used in developing mechanistic models for the processes which produce the glassy phase from the higher temperature melt.

The geospeedometry of natural glasses can be accessed due to the fact that all glasses preserve structures and properties which correspond to the rate at which they have been cooled across the glass transition. Two possibilities exist for the quantification of this cooling rate; (1) the determination of quenched glass properties (density, refractive index) and (2) the quantification of the hysteresis observed during reheating of the glass across the glass transition.

The constraining of the thermal history using techniques that rely on observation of property hysteresis during reheating can be applied to glasses which are chemically pristine and phase stable up to and beyond the glass transition region. These techniques have been applied to volcanic and other natural glasses using the heat capacity as the relaxational property and with the obtainment of data using differential scanning calorimetry.

For the case of glasses generated in volcanic environments, the range of cooling rates to which melts can be subjected can be quite considerable. Calorimetric determinations of the cooling rates of natural glasses invesitigated to date confirm that these rates spanning some 4-5 orders of magnitude for an individual melt composition. Clear evidence of the apparent annealing of volcanic glasses in pyroclastic deposits corresponds to field evidence for extensive welding of such pyroclastic units.

It is hoped that work of this sort can lead us to a reliable facies-specific indicator for the provenance of volcanic and other glasses in the natural environment.

J04 : 1B/35 : F2

Memory of the Thermal History of Geoglasses and Medium Range Order

Cécile Chemarin (chemarin@pcml.univ-lyon1.fr)&

Bernard Champagnon

(champ@pcml.univ-lyon1.fr)²

LPCML-Univ Lyon 1, UMR 5620, 69622-Villeurbanne, France

The view of the glass structure changed in recent years due to the consideration of a medium range order, intermediate between the long range disorder and the well known short range order (silicate tetrahedra in a silicate glass for example). This new point of view is due to several experimental results, such as low temperature heat capacity, neutrons, X-rays or light inelastic scatterings measurements (1), and is also modelized in recent computer simulations. For geosciences it is important that the new parameter l, characterizing the medium range order, is sensitive to the thermal history of the glass (2, 3) and can influence some of its macroscopic properties. In the paper we present these correlations for synthetic glasses and demonstrate, from the comparison of several natural glasses (glasses from the Ries crater and moldavite for example) that, indeed, the memory of the thermal hisory of the glass is kept in its structure. Recent results showing a possible correlation between the evolution of the Boson peak in Raman scattering and the viscosity are discussed.

Champagnon B, Chemarin C, Richet P, Philosophical Magazine, 77, 663-669, (1998).

Champagnon B, Panczer G, Chemarin C, Chemie der Erde, 57, 280-296, (1997).

Chemarin C, Champagnon B, Panczer G, J. Non-Crystalline Solids, 216, 11-115, (1997).

J04 : 1B/36 : F2

Low Pressure, in Situ Viscosity Measurements of Hydrous Volcanic Melts

Alan Whittington (aw@ipgp.jussieu.fr)¹,

Pascal Richet (richet@ipgp.jussieu.fr)¹,

Harald Behrens (h.behrens@mbox.mineralogie.uni-hannover.de)² &

François Holtz

(f.holtz@mbox.mineralogie.uni-hannover.de)²

¹ Géomatériaux, IPGP, 4 place Jussieu, 75252 Paris cedex 05, France

² Institut für Mineralogie, Welfengarten 1, D-30167 Hannover, Germany

Melt viscosity is an extremely important property in the study of volcanology, particularly in controlling eruptive style. The most important influences on viscosity are temperature, chemical composition and volatile content (predominantly water). Most viscosity determinations of hydrous melts have been conducted under high pressures and temperatures, where volatiles are easily dissolved. In this study, viscosities were measured between 10⁹ and 10¹³ Pas, under low-pressure conditions relevant to volcanism, by the parallel-plate method. These temperatures are near the glass transition where the exsolution kinetics of water are slow, and equilibrium viscosities can be measured with minimal water loss. The compositions studied include phonolite, trachyte, and tephrite. We used iron-free synthetic liquids containing up to 5 wt.% H₂O, which allow spectroscopic investigation, and whose melt structure and rheology is very similar to their Fe-bearing counterparts. The addition of water has a very strong decreasing effect on viscosity, which is greatest at low water contents. Addition of only 1 wt.% H₂O reduces the temperature of the 10¹² Pas isokom from 919 to 790 K for phonolite, 972 to 822 K for trachyte, and 932 to 842 K for tephrite, while with 2 wt.% H₂O these temperature are 698, 740 and 804 K respectively. The reduction is similar for our phonolite and trachyte, and the andesite of Richet et al. (1996), which have similar melt structures (NBO/T about 0.2), but different ratios of alkali to alkaline earth elements. Very depolymerised compositions, e.g. tephrite, are much less affected by the addition of water, so that for example the haplogranite of Dingwell et al. (1996) is less viscous than our tephrite for equal water contents in excess of 1 wt.%.

Dingwell DB, Romano C, Hess K-U, Contribs Min Pet, 124, 19-28, (1996).

Richet P, Lejeune A-M, Holtz F, Roux J, Chem Geol, 128, 185-197, (1996).

J04 : 1B/37 : F2

Effect of Water on the Local Structure Around Ni in Silicate Melts Exposed to High Pressure

François Farges (farges@univ-mlv.fr)¹,

Ralf Siewert¹,

Harald Behrens (h.behrens@mbox.mineralogie.uni-hannover.de)² &

Gordon E. Brown Jr. (gordon@pangea.stanford.edu)³

¹ Laboratoire des Géomatériaux, IFI-Université de Marne la Vallée, Marne la Vallée cedex 2, France

² Institut für Mineralogie - Universität Hannover, Welfengarten 1, D-30167 Hannover, Germany

³ Dept. of Geology, Stanford University, Stanford, CA 94305-2115, USA

The local structural environment around Ni(II) in albitic (ALB) and sodium trisilicate (NS3) glasses has been evaluated using high-resolution Ni K-edge XANES and EXAFS spectroscopies. The glasses were quenched isobarically from a melt equilibrated between 24 and 64 hours at 1100-1350°C and 1 bar-5 kbars. The glasses have Ni and water contents of about 1000-4000 ppm and 0-8.2 wt.%.

The pre-edges of Ni for ALB and NS3 glasses prepared at 1 bar are identical to that for 5-coordinated Ni (^[5]Ni), or a 50:50 mixture between ^[4]Ni and ^[6]Ni. Anharmonic analysis of the EXAFS spectra show relatively small disorder effects, inconsistent with the presence of an extended mixture of Ni-coordinations (^[4]Ni and ^[6]Ni may be present but below 25 atom.%). Si-second neighbors are detected near 3.30(5) Å.

At a synthesis pressure of 5 kbars, the pre-edge for both glasses indicates the presence of 1/3 of the Ni as ^[6]Ni. The average anharmonic Ni-O distance increases from 2.00 to 2.03(1) Å. When water is added to the high pressure melts, Ni is 6-coordinated in the quenched glasses with Ni-O distances of 2.06(1) Å. A Ni-Si contribution, near 3.20(5) Å, has increased in intensity. This medium range environment around Ni is different to that measured in aqueous solutions, in which the presence of H₂O units around Ni makes no contributions in the EXAFS signal above 3 Å.

However, above 3 wt.% water in ALB glass, the medium range environment of Ni is much more ordered than for the other glasses: large Ni-Ni contributions near 3.1, 4 and 6 Å are observed, similar to that known in nepouite, Ni₃Si₂O₅(OH)₄. TEM analysis of these glasses show the presence of diffraction features consistent with these for nepouite. Application of the random grain boundary theory to the EXAFS information suggests that the average diameter of these ordered clusters is =50 Å. These domains can be found in all ALB samples that are rich in water (>3%) whatever their synthesis conditions.

Based on these results, we suggest that Ni is not likely to form complexes with water in oxide glasses since Ni has nucleated in these glasses and the XAFS data for NS3 glass is clearly inconsistent with that for Ni-aqueous solutions. However, this data may support a possible complexation of Ni by OH, as suggested by simple valence considerations. Water seems then to enhance the transition of Ni to more highly (ie, 6-) coordinated environments (driven by crystal field effect), a transition that is clearly induced by pressure.

J04 : 1B/38 : F2

Stability of H and O Isotope Ratios in Natural Hydrated Silicic Glasses: Estimation of Diffusion Coefficient for Water at Ambient Temperatures

H. Albert Gilg (Albert.Gilg@geo.tum.de)¹ &

Simon M. F. Sheppard

(Simon.Sheppard@ens-lyon.fr)²

¹ Lehrstuhl fuer Angewandte Mineralogie udn Geochemie, Technische Universitaet Muenchen, Lichtenbergstr.4; 85747 Garching, Germany

² Laboratoire de Science de la Terre, CNRS-UMR 5570, Ecole Normale Supérieure de Lyon, 46 Allée d'Italie; 69364 Lyon cedex 07, France

Natural silicic glasses with high water content (>2 wt.%) are generally the result of secondary hydration by interaction with ground or surface waters at low temperatures. It has been suggested that hydrated glasses retain their initial hydrogen (and oxygen) isotope signatures even if the isotopic composition of the surrounding water has changed significantly. Thus hydrated glasses may be used to characterize the isotopic composition of old meteoric waters and thus might help to reconstruct paleoclimates and paleogeography. However, the timescale of stability of these isotopic signals are not known. We present new D/H and ¹⁸O/¹⁶O data from Mid-Miocene rhyolitic tuffs preserved in bentonites from the South German Molasse basin and Carboniferous perlites from the Meißen volcanic complex, Saxony, Germany, to constrain the retentivity of the stable isotope signals of hydrated glasses.

Separated and cleaned rhyolitic glass shards (average diameter 150 µm) from the Molasse basin show water contents from 4.9 to 5.8 wt.%, ¹⁸O_SMOW values of 14.1 to 16.3 permil and D_SMOW values of -115 to -128 permil. The glasses have equilibrated with low-temperature meteoric waters with D = -90±5 permil, that are isotopically depleted compared to present-day surface waters (D = -75±5 permil). Miocene meteoric waters deduced from isotopic composition of surrounding montmorillonites are even heavier than present-day ground waters. The isotopic composition of waters in equilibrium with these glasses is, however, identical to Pleistocene ground waters still present in deep aquifers of the Molasse basin (D = -90 permil). It is suggested that the glasses were initially hydrated during Miocene bentonite formation, but completely exchanged their hydrogen and oxygen isotope compositions with surrounding waters during Pleistocene times. In contrast, the Meißen perlites and overlying Early Cretaceous residual kaolins (¹⁸O = 20.1 to 21.4 permil and D = -79 to -88 permil) have equilibrated at ambient temperatures with waters that are isotopically much heavier (D = -45±5 permil) than present-day or Pleistocene waters. They show no indications for hydrogen or oxygen exchange during the Quaternary. The difference in comparison to the Molasse tuffs is explained by a larger diffusion length.

We use these empirical data to estimate the diffusion coefficient for water in hydrated silicic glasses at ambient temperatures. The value (log D = -22±1 [m²/s]) is in good agreement with extrapolations from high-temperature experiments but not lower than that for obsidians at low temperatures. It can be used to estimated the time necessary to change the oxygen and hydrogen isotope signals of hydrated glasses of a given size.

Session J04:1P

J04 : 1P/01 : PO

Structure and Crystallisation Behaviour of Barium Borate Glasses

Pasquale Pernice (pernice@unina.it)¹,

Vladimir N. Sigaev (sigaev@aha.ru)²,

Antonio Aronne (anaronne@unina.it)¹ &

Serena Esposito (serena@unina.it)¹

¹ Department of Materials and Production Engineering, University of Naples Federico II, Piazzale Tecchio, 80125- Naples, Italy

² D.I. Mendeleev University of Chemical Technology of Russia, Miusskaya sq.9, 125190- Moscow, Russia

Recently a remarkable interest is devoted to glass-ceramics materials containing polar phases characterised by non linear optical activity, ferroelectric and related properties. These materials exhibit a number of specific merits as compared with piezoceramics and pyroceramics. They combine the advantages of glass technology and the feasibility to produce textured structures by crystallisation under a high temperature gradient, hot extrusion or d.c. fields. The non-linear optical crystals of low temperature (ß) phase of barium metaborate ßBaB₂O₄ (ßBBO) exhibit an intensive second harmonic generation (SHG), a wide transparency range, a broad phase-matchable region and a high damage threshold. To study the growth of ßBaB₂O₄ crystals we have synthesised glasses containing different amount of glass forming oxides (SiO₂, TiO₂ and Al₂O₃) as it is difficult to prepare homogeneous glass from the stoichiometric composition BaO.B₂O₃. The compositions of the studied glasses are expressed by the general formulae: (50-x/2)BaO·(50-x/2)B₂O₃·xMO₂ with x=4, 8, 16; and (50-x/2)BaO·(50-x/2)B₂O₃·xAl₂O₃ with x=2, 4, 8. Transparent glasses have been obtained for all compositions. The glasses containing SiO₂ and TiO₂ are homogeneous while those containing Al₂O₃ are phase separated. The SiO₂ glasses devitrify in one step forming only ßBaB₂O₄ nanocrystals. The TiO₂ and Al₂O₃ glasses devitrify in one or two steps depending on the Al or Ti oxide content and on the specific surface area of the sample. In these systems the crystallising phases were ßBaB₂O₄ (main phase), BaB₄O₇ and BaTi(BO₃)₂. In all glasses, surface crystallisation was found to be dominant for the ßBBO phase. Any attempt to form bulk nuclei by heat treatments in a temperature range near the glass transition temperature (Tg) failed. Nevertheless, in TiO₂ glasses, when 16 hours heat treatments were performed on the as-quenched bulk samples a structural transformations occur in the glassy matrix. As a consequence two micro-region are formed: the one richer of more polymerised units, the other richer of less polymerised units than the as-quenched glass. For longer heat treatment times (3 days) only in the glass with the higher TiO₂ content (x=16) the nucleation of BaB₄O₇ crystals also occurs. On the contrary, the Al₂O₃ glasses become homogeneous after the heat treatment of 16 hours at Tg as theirs composition lie close to one side of the metastable immiscibility cupola. Al₂O₃ enhances the glass forming ability more than SiO₂ and TiO₂. Surface crystallisation of ßBBO occurs in all studied glasses. This is a major precondition for creation of pyroelectric glass-ceramic textures based on ßBBO crystals.

J04 : 1P/02 : PO

Comparison of Iron Environment in Volcanic and Synthetic Glasses

Yannick Lefrere (lefrere@lmcp.jussieu.fr),

Laurence Galoisy &

Georges Calas

LMCP, 4 place jussieu case 115, tour 16-26/2, France

Optical absorption spectra of Ferrous and Ferric iron have been compared in synthetic and volcanic glasses. Volcanic glasses spectra are different from those obtained on synthetic glasses depending on composition (silicic or basaltic). Optical absorption spectra of ferrous iron in synthetic glasses features usually a broad high-intensity asymmetric band located around 10000 cm^-1 while in silicic volcanic glasses a Gaussian shaped band is found with a similar intensity but at lower wavenumber (around 8500 cm^-1). The shape of this band depends on the composition and is related to five- and six- coordinated ferrous iron in synthetic glasses while in silicic volcanic glasses, the Gaussian shape of the band may be related to the pure octahedral character of the ferrous iron site.Another band of small intensity is always present around 5000 cm^-1 in the synthetic and basaltic glass spectra. The non correlated evolution of the two bands with glass composition suggests a four - coordinated ferrous iron origin. isolated ferric iron is responsible for low-intensity absorption bands laying on the front edge of a charge transfer band occuring in the UV range. It is possible to show that no scattering process exists in synthetic glasses while in silicic volcanic ones, these scattering mechanisms may be simulated by a Rayleigh function. In both types of glasses charge transfer phenomenon can be modeled by an Urbach law or by a Gaussian function depending on the glass.

J04 : 1P/03 : PO

Luminescence of the Fe³⁺ and Mn²⁺ in the Acid Volcanic Glasses

Bernard Piriou¹,

Yotzo Yanev (yotzo@geology.bas.bg)² &

Nikolai Zotov (Nikolay.Zotov@uni-bayreuth.de)³

¹ Ecole Centrale de Paris, Lab. P.C.M., Chatenay-Malabry 92295, France

² Geological Institute, Bulgarian Academy of Sciences, Sofia 1113, Bulgaria

³ Bayerisches Geoinstitut, Bayreuth D-95440, Germany

The luminescence spectra of synthetic and volcanic acid glasses (SiO₂ >65%wt) with different Fe³⁺ and water (0.2 - 6.7% wt) content have been investigated. Three luminescence peaks are established around 400, 550-600 and 720-740 nm. The intensity of the peaks decreases with increasing water content, which is attributed to depolymerization of the glass structure by water, so to increasing of the number of nonbridging oxigens. The first peak at about 400 nm is very rapid (relaxation time 20-40 ns) and is assigned by some authors to transition of Fe³⁺ from either the 4A1(G) or/and 4T2(D) energy level to the 6A1 ground state in the Tanabe-Sugano diagram. The luminescence peak at about 550 nm has longer relaxation time (110-250 ns). It is due to transitions of tetrahedrally coordinated Mn²⁺ from the excited 4T1(G)level to the 6A1 ground state. The peak at about 600 nm is very rarely observed and is assigned to Mn²⁺ in octahedral coordination. The strongest luminescence peak observed at 720-740 nm with relaxation time of few µs is due to the transition of tetrahedrally coordinated Fe³⁺ from the excited 4T1(G)level to the 6A1 ground state. Little displacement of this peak indicate a distortion of the Fe³⁺-O tetrahedra.

This study is supported by the project 528 of the Bulgarian National Scientific Foundation.

J04 : 1P/04 : PO

Structural Environment of Zr and Fe in Long Term Alteration Products of Nuclear Glass

Emmanuele Pelegrin (pelegrin@lmcp.jussieu.fr)¹,

Philippe Ildefonse (ildefons@lmcp.jussieu.fr)²,

Patrick Jollivet (jollivet@cea.fr)¹,

Stephane Gin (gin@amandine.cea.fr)¹,

Georges Calas (calas@lmcp.jussieu.fr)²,

Anne-Marie Flank (flank@lure.u-psud.fr)³ &

Armando Sole (sole@leo.esrf.fr)⁴

¹ CEA/SCD/LEAM, Centre Valrho, BP171, 30207 Bagnols/Ceze Cedex, France

² Universite Paris 6, LMCP, 4 place Jussieu, case115, T16-26 2eme, 75252 Paris cedex 05, France

³ LURE CNRS/CEA/MEN, Bat 209D, 91405 Orsay cedex, France

⁴ ESRF, BP220 F38043, Grenoble cedex, France

The understanding of the stability and the long term behavior of high activity nuclear matrix is of great importance for waste disposal. Alteration products of inactive analog of high activity nuclear glass (R7T7 type) are constituted of a XRD-amorphous phase enriched in Zr and Fe. A crystallochemical study (using X-ray Absorption Spectroscopy) was performed on nuclear alteration products obtained at 90°C, under static (SG) and dynamic flow (HP401, HP405), during 6 and 9 months respectively. For dynamic leaching, the silicon content of the alteration solution was gradually increased. Zr- and Fe-L_2,3 XANES together with Zr-, Fe-K EXAFS spectra were recorded at LURE/SuperACO (Orsay, France) and ESRF (Grenoble, France) facilities respectively. Reference compounds with known Zr and Fe structural environments and R7T7 glass were also investigated. Zr-L_2,3 XANES spectra evidence 2 significant behaviors according to alteration conditions. Zr is 6-fold coordinated to oxygens in both pristine glass and SG sample. On the contrary, under dynamic flow (HP401, HP405 samples), an increase in Zr coordination is observed. Derived Zr-1st neighbor and Zr-2nd neighbor distances indicate that, in R7T7 glass, octahedral Zr are sharing corner with silicon-bearing tetrahedra. In alteration products obtained under dynamic flow, Zr-O and Zr-2nd neighbors distances increase and the nature of near-nearest neighbors changes (Zr-enriched domains). These changes in local structure will be discussed according to model compounds. For iron, Fe-L_2,3 XANES and Fe-K EXAFS spectra evidence that, whatever the leaching conditions are, the local structure around Fe is modified. While Fe is tetrahedrally coordinated to oxygens in nuclear glass, Fe-bearing octahedra are edge-connected in alteration products. These results will be discussed in term of stability of structural model compounds.

J04 : 1P/05 : PO

A New Analog Family for Vitrified Wastes: Paleometallurgic Slags

Cecile Mahe-Le Carlier

(cmahe@crpg.cnrs-nancy.fr),

Alain Ploquin

(ploquin@crpg.cnrs-nancy.fr),

Christian Le Carlier de Veslud (carlier@crpg.cnrs-nancy.fr) &

Jean-Jacques Royer (royer@crpg.cnrs-nancy.fr)

C.R.P.G.-C.N.R.S, 15 rue Notre Dame des Pauvres, BP 20, Vandoeuvre-Les-Nancy, France

Vitrification is a more and more used process for stabilization of municipal or nuclear wastes. Frequently, It is important to predict their long term durability of these wastes. The study of natural or anthropic analogs is a method to estimate their behavior in natural conditions. We propose anew family of analogs for periods of time ranging from 100 to 4000 years.This period is intermediate between those of laboratory tests and those of natural analogs like basalts. Two kinds of slags have been studied, coming from two significant types of metallurgy: iron metallurgy and polymetallic metallurgy (for example, copper and lead). The slags present generally a vitrified upper surface. The study of altered zones on these surfaces makes it possible to propose mechanisms and determine the rate of alteration under natural conditions. It is also possible to characterize the toxicelements behavior. In the iron metallurgical slags, toxic element concentrations are low. However, it is useful to study the alteration of these materials because their chemical composition are very similar to those of vitrified wastes (municipal wastes named "vitrifiats de REFIOM", A and B grade nuclear wastes). Because of the large variation of chemical composition of this type of slags, it is also possible to investigate the effect of composition on the rate of alteration.Compositions of polymetallic slags are similar to those of iron metallurgical slags, but this material presents in addition high content of toxic elements (Pb, Zn, Co, Ni, As, Sb, Sn, Cu, Ba) sometimes up to several percent. Observations show that the natural alteration of these ancient slags corresponds to the first step of alteration (formation of an altered layer andoxy-hydroxides), relative to what is observed in natural rocks, such as basalts (formation of clays and zeolites). Polluant elements (Ba, Pb or Zn)are not retained in this altered layer and are released in the surrounding environment. The average rate of alteration ranges from 20 to 180 µm/1000y, a relatively low value, compared to those obtained from studies of natural rocks. This rate seems to be correlated with the polymerisation state of the material.

J04 : 1P/06 : PO

Structural Study of Synthetic Glasses by ²³Na and ²⁷Al MQ MAS NMR Spectroscopy

Frédéric Angeli (angeli@scm.saclay.cea.fr),

Thibault Charpentier,

Pascal Faucon,

Jean-Claude Petit,

Georges Calas &

Joseph Virlet

CEA Saclay, SCM-bat. 125, 91191 Gif sur Yvette, France

We describe a first approach to apply the Multiple Quantum Magic-Angle Spinning Nuclear Magnetic Resonance spectroscopy (MQ MAS NMR) to the amorphous materials using inversion procedure in order to obtain quantitative information. Taking into account the overall efficiency of the MQ MAS sequence, it allows to represent the distributions of each interaction in two dimensions: the isotropic chemical shift and the quadrupolar frequency dimensions. Thus, the local structure of several aluminosilicate and aluminoborosilicate glasses (containing the main components of the French nuclear waste glass, R7T7, and of a basaltic glass) has been studied using ²³Na and ²⁷Al 3Q MAS NMR. All glasses contain aluminum as network former and sodium as network modifier, which can be used in NMR as probes of their local environment. These nuclei allowed to investigate the short-range and the intermediate-range order in glasses observing chemical shifts and quadrupolar interactions. The interpretation of the different site distributions is discussed in term of topological disorder, i.e. the distribution of specific geometrical parameters like bond distances and angles. Using semi-empirical relationships previously established with the structure of crystalline silicate compounds, a quantification of Na-O distances and Al-O-Si bond angle distributions is given with the distributions of sodium and aluminum isotropic chemical shift. This approach is easily transferable to a large wide of disordered materials, notably for complex structures which were difficult to investigate until now.

J04 : 1P/07 : PO

Geochemistry of the Rhyolitic Magmas Associated with the Streltsovka Uranium Deposits (Transbaikalia, Russia): A Magmatic-Inclusion Study

Aliouka Chabiron

(chabiron@cregu.cnrs-nancy.fr)¹,

Alexei P. Alyoshin²,

Michel Cuney¹,

Etienne Deloule³,

Viatcheslas N. Golubev²,

Vasilii I. Velitchkin² &

Bernard Poty¹

¹ CREGU, BP23, 54501 Vandoeuvre-les-Nancy, France

² IGEM, Staromonetny per. 35, 109017 Moscow, Russia

³ CRPG-CNRS, BP 20, 54500 Vandoeuvre-les-Nancy, France

The U-Mo deposits of the Streltsovka caldera are located in Transbaikalia near the Chinese-Mongolian border. They represent the largest uranium field associated with volcanics in the world. About 20 U-Mo deposits have been discovered in this caldera. They were formed immediately after volcanic activity. All the rocks, volcanics and granites, were deeply altered by hydrothermal fluids and their original chemistry was strongly modified, particularly the primary uranium concentrations. The study of melt inclusions allows to characterize the original chemistry of the silicate melts before alteration. Melt inclusions, preserved in quartz phenocrysts from strongly altered rhyolites of the Streltsovka caldera, were analyzed by electron microprobe for Si, Al, Fe, Na, K, Ca, F, Cl, Mn, Ti, P, S and Mg, by fission tracks for U and by ionic microprobe for H, U, Th, La, Ce, Pr, Nd, Sm, Eu, Gd, Dy, Er, Yb, Lu, Sc, Ti, V, Sr, Y, Nb, Ba, Hf, Th U and H⁺. The presence of daughter minerals, such as fluorite, F-rich mica and REE fluo-carbonate, in melt inclusions and the melt analyses (after homogenization) indicate that the Streltsovka rhyolites correspond to mildly peralkaline magmas (1.04 < Na+K/Al < 1.10), rich in F (from 1.4 to 2.7 wt%), REE and with 0.2 wt% Cl. Melt temperatures obtained from homogenization of the inclusions are relatively high for such magma (1015 ± 30°C) and may be surestimated because of a possible volatile loss since rhyolite formation. Present water content in the homogenized silicate melt is between 0.6 and 2.4wt%. The high U contents of the melts (14 to 23 ppm) indicate that the rhyolites are one of the possible sources for the Streltsovka uranium deposits.

Session J04:2A

J04 : 2A/01 : F2

Meteoric Alteration of Icelandic Volcanic Glasses: Change in the Long Term Mechanism

Xavier Le Gal (xlegal@illite.u-strasbg.fr)¹,

Jean Louis Crovisier (jlc@illite.u-strasbg.fr)²,

Francois Gauthier-Lafaye

(gauthier@illite.u-strasbg.fr)¹,

José Honnorez (honnorez@illite.u-strasbg.fr)¹,

Walter Bernotat (bernotat@ine.fzk.de)³ &

Bernd Grambow (grambow@nanhp7.in2p3.fr)⁴

¹ Institut für Nukleare Entsorgungstechnik, Forschungszentrum Karlsruhe, Postfach 3640, D-76021 Karlsruhe, Centre de Géochimie de la Surface, UMR n° 7517, CNRS et Université Louis Pasteur, 1, rue Blessig 67084 Strasbourg Cedex, Germany

² Centre de Géochimie de la Surface, UMR n° 7517, CNRS et Université Louis Pasteur, 1, rue Blessig 67084 Strasbourg Cedex, France

³ Institut für Nukleare Entsorgungstechnik, Forschungszentrum Karlsruhe, Postfach 3640, D-76021 Karlsruhe, Germany

⁴ Ecole des Mines de Nantes, Laboratoire SUBATECH, 4, rue Alfred Kastler BP 20722 F 44307 Nantes Cedex 3, France

Fresh water altered Icelandic basaltic hyaloclastites from Hengill, Husafell and Vatnajökull were studied. They range in age from 0.09 to 3.1 My and their basaltic glassy clasts are surrounded with various alteration layers.

In the youngest samples (<1 My old), an anisotropic and porous «palagonite» layer is formed by precipitation of poorly crystalline smectite as a result of a congruent dissolution of the glassy network. Similar layers have been reproduced during laboratory experiments. In older hyaloclastites, zeolites (chabazite, phillipsite, analcime) fill the initial porosity and are systematically associated to an isotropic residual glass which is hydrated (10% of water), and depleted in Na and Ca. In addition, Fe is oxidized: FeO is about 10% in fresh glass and 2% in residual glass. Measurements of ¹⁸O_SMOW for glass (5⁰/₀₀), palagonite (15⁰/₀₀), residual glass (15⁰/₀₀) and intergranular clays (15⁰/₀₀) indicate major modifications of the parental glass structure. This isotopic results are corroborated by iron oxidation. Finally, traces elements analyses (ICP-MS and ICP-AES) on the individual alteration products were performed after separation in heavy liquids or in a magnetic separator. These data permit us to calculate the alteration mass balance calculations.

The formation of zeolites has an important implications for the long-term alteration mechanism: (1) the dissolution of the glass is selective, probably as the result of a zeolite screen between glass and the altering water; (2) zeolites do not constitute a cement but a network of micro-channels (< 2.5nm) leading to the hydration and the selective leaching of some glass components, mainly Na and Ca. These observations of natural samples of basaltic glasses can also help us to understand the long term behaviour of nuclear waste form glasses stored in natural environment where the formation of zeolites occurs.

J04 : 2A/02 : F2

Application of High Resolution Fe K- Pre-Edge Spectroscopy to Iron Environment in Volcanic Glasses

Laurence Galoisy (galoisy@lmcp.jussieu.fr),

Georges Calas (calas@lmcp.jussieu.fr) &

Marie-Anne Arrio (arrio@lmcp.jussieu.fr)

LMCP, Case 115, 4 place jussieu, FRANCE

The sites occupied by ferrous and ferric cations in volcanic glasses may be related to the cooling conditions and to the coexisting phases. However, the geometry of these sites is still debated. We present high resolution XANES spectra recorded at the Fe K-edge in minerals with various Fe-coordinations, as well as in synthetic and natural glasses. The volcanic glasses investigated include basaltic compositions, andesitic and more differentiated silicic glasses of rhyolitic and pantelleritic compositions. A pre-edge peak is observed on the low energy side of the main edge and is usually assigned to transitions towards 3d-like electronic levels. The shape, position and relative intensity of the various components of the pre-edge depend on oxidation states, Fe- coordination number, mixing of sites and site distortion. Under high resolution conditions, the complex structure of the pre-edge peak makes it difficult to determine an average position which could be used to determine the redox state of Fe in minerals. The set of data obtained on crystalline references is used to discuss the oxidation state and local surrounding of Fe in volcanic glasses. The coordination geometry of ferric ions in basaltic glasses and obsidians is different. The surrounding of ferrous ions is highly dependent on the glass composition and on the polymerization of the glassy matrix. The comparison with synthetic glasses gives important constraints on the nature of the Fe-sites in volcanic glasses.

J04 : 2A/03 : F2

The Special Case of Iron in Tektites

Stephanie Rossano (sros98@esc.cam.ac.uk)¹,

Aline Ramos (ramos@lmcp.jussieu.fr)²,

Etienne Balan (balan@lmcp.jussieu.fr)²,

Guillaume Morin (morin@lmcp.jussieu.fr)²,

Georges Calas (calas@lmcp.jussieu.fr)² &

Christian Brouder (brouder@lmcp.jussieu.fr)²

¹ Earth Sciences department, University of Cambridge, Downing street Cambridge CB2 3EQ, England

² Lab. de Mineralogie-Cristallographie, 4 place Jussieu, tour 16 case 115, 75252 Paris cedex 05, France

Tektites are silica-rich glasses (SiO₂ > 70 wt%) found on different places at the Earth surface and supposed to be formed under a giant meteorite impact on Earth sediments (Bouska, 1993). The samples of the different fields show color variation but have very similar spectroscopic features considering optical absorption and Mössbauer spectra, for example. In comparison, these glasses are chemically different from typical rhyolitic glasses in that they show a high concentration in alkaline-Earth and alkaline elements. Besides, iron is almost entirely under the Fe²⁺ chemical form (Fe²⁺/Fe_tot » 92%) in tektites while terrestrial volcanic glasses (Monocrater type) contain a greater proportion of ferric iron (Fe²⁺/Fe_tot » 60%). The work presented here is an exhaustive spectroscopic study of two types of tektites : one moldavite from Tchecoslosvaquia (bottle green colored, 13.5 My, associated with the Ries crater) and one tektite from Thailand (brown colored, 0.8 My, no crater association). Mössbauer spectra analysis in terms of "site" distribution raised up the continuous character of iron site geometry in these glasses around the 4-fold and 5-fold preponderant coordinated sites (Rossano et al., submitted). This new interpretation allows to explain the contradictions derived from optical absorption spectra analysis. No ferric iron is detected. To confirm the results obtained for ferrous iron and get a more quantified view of local structure around Fe²⁺ ion, we performed EXAFS measurements at iron K-edge and molecular dynamics simulation (Delaye and Ghaled, 1996) of a tektite of a simplified composition. MD empirical potential parameter for iron-oxygen pair (Belashenko, 1996) has been adjusted in regard of EXAFS first shell distance while MD simulation helped making assumptions about next-nearest neighbors nature. The combined use of a structural model and of the GNXAS data analysis package (Westre, 1995) leads to a precise determination of local and medium range order around iron. In particular, iron "polyedra" obtained by MD simulation are in good agreement with what has been deduced from Mössbauer data analysis. The new interpretation of iron surrounding proposed in this study unifies the results obtained by the different techniques on ferrous iron in glasses. In particular, EXAFS spectroscopy is shown to be sensitive to small distances only and thus is not able to detect the presence of 5-fold coordinated iron while Mössbauer spectra analysis in terms of distribution shows the presence of a continuous distribution of iron site between the 4-fold and 5-fold coordinated site. The coupling of these spectroscopic techniques with molecular dynamics simulation adds a structural dimension that lacks in most of natural glasses studies.

Bouska V, Natural glasses, Ellis Horwood, (1993).

Belashenko DK, J. Non Cryst. Solids, 205-207, 212-215, (1996).

Delaye J-M & Ghaleb D, Mater. Sci. Eng, B37, 232-236, (1996).

Rossano S & Balan E & Morin G & Bauer J-Ph & Calas Gand Brouder Ch, Phys. chem. Miner, (submitted 1998).

Westre TE & Di Cicco A & Filipponi A & Natoli CR & Hedman B & Solomon EI & Hodgson KO, J. Am. Chem. Soc, 117, 1566-1583, (1995).

J04 : 2A/04 : F2

Glass Homogeneity Tests on a Microscale: Use of Bulk Analysis Standards for Microanalysis

Richard Hinton (rhinton@glg.ed.ac.uk)

Geology and Geophysics, University of Edinburgh, Edinburgh EH9 3JW

The use of glass standards for trace element microanalysis requires an knowledge of their homogeneity on all scales from microns to cm. Whereas determination of absolute concentrations is best achieved by conventional analysis, important information on the homogeneity of a glass can be gained by microanalytical methods. Any observed variations in signal/concentration across a single sample which cannot be explained by limitations of the analytical method may be assumed to be due to sample inhomogeneity. If, however, the signal is constant (to the known limit of the method, or to the counting statistics precision) then limits can be set on the possible level of inhomogeneity within the sample. Testing on a larger scale can be achieved by demonstrating that point analyses of random samples are identical to repeat analyses on a single sample (Hinton et al., 1995). A more stringent test of homogeneity can be achieved by comparing different standards against each other. Both have to be homogeneous, to the smallest analytical volume, to given consistent results.

The NIST 600 series of glasses were made as cane drawn from 100 kg of melt and represent the most widely available multi-element glass standards. Previous measurements made on single samples of NIST SRM 613 and SRM 611 (nominally 50 ppm and 500 ppm, but actually approximately 37 and 450 ppm, of 61 trace elements) are contrasted with new analyses of SRM 612 and SRM 610 (same composition but different thickness/source of glass wafer). The average trace element abundance ratios, SRM 611/SRM 613 and SRM 610/SRM 612, are close to 12 and the two glass pairs give agreement to better than 1% for many elements. Since the precision of method was at best 0.5%, the individual glasses tested must be homogenous to better than 0.5%. Analyses across the edges of a whole wafer demonstrates that only the most volatile elements (e.g. Tl and F) are lost from the outer 100 microns during the drawing process. Although all analyses strongly support a high degree of homogeneity between random samples of these glass fragments over scales of at least cm, it is not known how far this can be extended to the 100 kg of glass cane manufactured. The level of homogeneity observed over scales of 10's microns to cm suggest that differences observed between individual techniques (e.g. Rocholl et al., 1997 and Pearce et al. 1997), and with earlier certified values, cannot be readily ascribed to sample inhomogeneity.

Hinton, R.W., Harte, B. and Witt-Eickschen, G., Analyst, 120, 1315-1319, (1995).

Rocholl, A.B.E, Simon, K., Jochum, K.P, Bruhn, F., Gehann, R., Kramar, U., Lueke, W., Molzahn, M., Geostandards Newsletter, 21, 101-114, (1997).

Pearce, N.J., Perkins, W.T., Westgate, J.A. Gorton, M.P., Jackson, S.E., Neal, C.R. and Chenery, S.P., Geostandards Newsletter, 21, 115-144, (1997).

J04 : 2A/05 : F2

Electron Probe Microanalysis of Volcanic Glass: Patterns and Causes of Inter- and Intra-Laboratory Variation

John B. Hunt (jhunt@chelt.ac.uk)¹ &

Peter G. Hill (phill@glg.ed.ac.uk)²

¹ Geography and Environmental Management Research Unit Cheltenham & Gloucester College of Higher Education, Francis Close Hall Cheltenham, GL50 4AZ, U.K

² Department of Geology & Geophysics, University of Edinburgh, King's Buildings, Edinburgh EH9 3JW, U.K

The analysis of glass shards erupted during explosive volcanism forms the basis of tephrochronology and tephrostratigraphy - the identification and utilization of rapidly deposited layers of volcanic ash as a means of interpreting the timescale of palaeoenvironmental change. The utilization of these ash [tephra] layers is dependent upon the recognition of signatures or fingerprints that can be defined as characteristic of a particular eruption. Major and minor element quantitation by electron probe microanalysis (EPMA) is an established technique in this field and is effective because glass, being rapidly quenched erupted magma, is geochemically representative of the volcanic event

Problems associated with EPMA analysis of glasses, including mobility of alkali metals and effective over-representation of high abundance components, have long been recognised, though poorly understood. In tephrochronology, more than in petrology and general igneous geology, these problems are potentially acute owing to the small scale differences, nearing limits of instrumental precision, between different eruptions of individual volcanoes

In investigating the processes and problems associated with alkali mobility a well characterised secondary glass standard (Lipari obsidian) has been employed both to investigate inter-laboratory variability and to examine anomalous responses of glass to variations in electron beam diameter

Twelve electron microprobe instruments in Europe and the U.S. have contributed to the interlaboratory variability programmes with results, in some cases, showing considerable deviation from the independently verified composition, largely as a result of the electron-beam-induced negative relationship between Na₂O and SiO₂. The potential geochronological consequences arising from these discrepancies are discussed. In addition, it has been established that a positive relationship holds between Na₂O and SiO for smaller electron beams (larger electron flux). This previously unrecognised phenomenon may relate to differential pitting of the glass during electron bombardment with the conse-quent uncorrected absorption of selected wave-length X-rays. The potential industrial applications of this effect will be discussed in addition to their significance for the analysis small surface-area geological glasses.

Sodium-Silicon Relationship for Differing Beam-Paster Dimensions

J04 : 2A/06 : F2

Comparison of Microanalytical Techniques for Hydrous Rhyolite Glasses ­ Results from A Melt Inclusion Study

Karl Schmitt Axel (axelk@gfz-potsdam.de),

Dieter Rhede (rhede@gfz-potsdam.de) &

Rolf Emmermann

(emmermann@gfz-potsdam.de)

Geo. Forschungs. Zentrum Potsdam, Telegrafenberg B124, 14473 Potsdam, Germany

Melt inclusions have been investigated in order to assess pre-eruptive volatile contents of ignimbrite-forming magmas from the Central Andes. Quartz and plagioclase hosted melt inclusions were analyzed by electron microprobe (EMPA), ion probe (SIMS) and infrared techniques (FTIR) following established methods (Ihinger et al. 1994, Devine et al. 1995). Melt inclusion glasses are rhyolitic in composition similar to pumice matrix glasses. Water was found to be the dominant volatile component in melt inclusions. CO₂ was generally below 200 ppm. We compare here the results of SIMS and FTIR water analysis with H₂O contents estimated from EMPA data. Two EMPA approaches were used. One is to add all oxide compounds and assume that the difference from 100 wt.% represents H₂O (EMPA 100-oxides). A second approach is to measure oxygen directly by electron microprobe and to subtract from this the stoichiometric oxygen from major element concentrations. The difference in oxygen is then calculated as H₂O (EMPA O_meas-O_stoi). Both methods are indirect, and precision based on counting statistics is comparable. An internal reference material was prepared from natural rhyolitic glass from the study region and analyzed for water by bulk rock thermal extraction methods (IR and Karl-Fischer titration). H₂O contents were 2.86 ± 0.05 wt.% (1<sigma> error) for IR and 2.69 ± 0.12 wt.% for Karl-Fischer. SIMS water analyses on the same glass show considerable scatter, but are close to the bulk rock values (2.54 ± 1.02 wt.%). Both EMPA approaches yield generally higher water values than SIMS or bulk-rock analysis. Furthermore, the EMPA 100-oxides values are higher and more variable than the EMPA O_meas-O_stoi water estimates (4.00 ± 0.38 wt.% and 3.40 ± 0.30 wt.%, resp.). Analysis of melt inclusion glasses with higher water contents underscore the differences between EMPA and direct determinations. Glasses of quartz-hosted, petrographically similar melt inclusions from ignimbrite pumice yield concordant H₂O values for the direct methods (SIMS: 3.7±0.2, n = 15 and FTIR: 3.8±0.4, n = 17). The EMPA H₂O estimates are 7.1±0.7, n = 20 (EMPA 100-oxides) and 5.0±1.7, n = 20 (EMPA O_meas-O_stoi). We can assume that the samples were homogeneous within analytical precision because of the agreement between the SIMS and FTIR results and their minor scatter. In conclusion, a systematic overestimation of water seems inherent in EMPA water estimates. To explain this we invoke sample coating, surface imperfections, sample charging and inadequate correction procedures applied to raw data in EMPA. Caution is therefore urged in the use of EMPA estimates for water in hydrous silicate melt inclusion glasses.

Ihinger PD, Hervig RL & McMillan PF, Reviews in Mineralogy, 30, 67-121, (1994).

Devine JD, Gardner JE, Brack HP, Layne DG & Rutherford MJ, American Mineralogist, 80, 319-328, (1995).

J04 : 2A/09 : F2

Kinetics of Pollutant Element Release During Glass Alteration

Sterpenich Jerome (jerome@crpg.cnrs-nancy.fr) &

Guy Libourel (libou@crpg.cnrs-nancy.fr)

CRPG-CNRS, 15 rue ND des Pauvres BP20, 54501 Vandeouvres, France

The study of the long-term behaviour of glasses is of great interest for the storage of vitrified wastes. At present, mechanisms and kinetics of alteration as well as elemental behaviour during glass leaching are mainly deduced from experimental studies. However, information on natural alteration of glass over longer periods of time is also essential in order to check the validity of experimental models of dissolution. In this paper we use the alteration of medieval stained-glasses to evaluate not only the global kinetics of alteration but also the release rate of elements, in particular the polluting ones.

The samples studied are archaeological stained glasses excavated from different sites in France (Marseille, Digne and Rouen) which date from the IX to the XIIIth century. All glasses are rich in silica, calcium and alkalis. ICP-MS analyses also reveal significant amounts of trace elements which correspond either to impurities in the starting materials (e.g. Ba, Sr, Pb, Zn, REE) or colour forming elements (e.g. Co, Cu, Fe, Mn). It is of note that these metals are the same as the polluting elements of present day vitrified wastes.

Each sample is characterised by the presence of an altered layer which has developed on its surface. The bulk chemical composition of the corroded crust was determined by ICP-AES and ICP-MS whereas the behaviour of elements between the pristine and the leached glass was studied by electron and ion microprobe.

Assuming that the initial surface of the studied glasses is still present and that the kinetics of alteration are a linear function of time, the release rate of each element can be quantified if the compositions of pristine and weathered glass are known as well as their density. We show that the release rate is strongly dependent on the element considered. In general, network forming cations (NF) are systematically less leachable than network modifying cations (NM). For instance, Na, K and Mg are released at around 0.01 g/m²/day whereas Zr, Ti and Al are released close to 0.001 g/m²/day and Th near 0.0001 g/m²/day. Among NF elements, we show also that the kinetics of release increase with decreasing ionic radius, implying for example that Zr is less leachable than Si.

Moreover, since medieval stained-glasses present different compositions, we show that it is possible to quantify the influence of the glass composition on its long term behaviour.

J04 : 2A/10 : F2

The Transformation of Heterogeneous Municipal Solid Wastes into Homogeneous Raw Materials Using Innovative Melting Processes: Comparison with Natural Analogues

Peter Kruspan (peter.kruspan@eawag.ch)¹,

Bernd Zimanowski

(zimano@gaia.geologie.uni-wuerzburg.de)² &

Thomas Lichtensteiger (lichtens@eawag.ch)¹

¹ EAWAG/ETH, Ueberlandstrasse 133, CH - 8600 Duebendorf, Switzerland

² Physikal.-Vulkanolog. Labor, Universität Würzburg, D - 97070 Würzburg, Germany

Municipal solid waste (MSW) is treated today in mainly two ways: direct deposition or incineration at 900°C in plants with a grate system. Incineration allows maximum recovery of the energy content in MSW. Therefore this process is thermally optimised. On the other hand, the waste products have to be landfilled because they are not suitable enough for the application as a secondary mineral resource. Both, primary and secondary mineral resources can only be generated by processes containing two fundamental technical concepts: separation and homogenisation of the different fractions obtained. Thus the processes 'direct deposition' as well as 'incineration' are not materially optimised. In both cases, the chemical and petrological heterogeneity is not substantially changed.Because of these reasons, technical processes have been recently developed in order to transform MSW into valuable secondary raw materials. A large-scale plant operated according to the Von Roll/Holderbank RCP/HSR process has been studied in detail. Similarly to the primary mineral industry, MSW here passes through separation and homogenisation (cf. Zeltner, 1998). The core of the plant is a high-temperature (1500°C) reactor where a basaltic melt enriched in lithophilic elements (98%) is separated metallurgically from a sulfidic melt enriched in chalcophilic (8%) and siderophilic (91%) elements (i.e. heavy metals). Both types of melts leave the reactor separately passing through a cooling step, so that a Fe-Cu-Ni-Sn-alloy and a silicate product can be sold to the mineral industry. 20% of the total input into the RCP/HSR process is converted to the silicate product which can form different shapes depending on the type of cooling step installed. The main goal is to develop an appropriate cooling step to optimise product quality and price. The easiest way is the installation of a conventional water cooling, where the silicate melt is fragmented hydroclastically so that a completely glassy granulate of sand to gravel size is obtained. This granulate possesses middling pozzolanic properties i.e. it contributes to strength development in a cement environment and can therefore act partially as a substitute for Portland cement clinker. Pozzolanicity is directly dependent on the structure of the glass network and thus dependent mainly on melt polymerisation or viscosity. A way to enhance pozzolanicity is an explosive fragmentation of the melt according to the techniques described by Zimanowski et al. (1997). Generally, the formation and alteration of pyroclastic rocks can serve as helpful natural analogues. Tuff, for example, is used worldwide as a material with excellent pozzolanic properties because of phreatomagmatic formation and subsequent alteration of the glass to zeolites.

Zeltner, Ch, PhD thesis ETHZ No. 12688, (1998).

Zimanowski B, Büttner R, Lorenz V, Häfele HG, J. Geophys. Res, 102, 803-814, (1997).

J04 : 2A/11 : F2

Structural Data in a Light Water Waste Containment Glass

Mikael Le Grand (legrand@lmcp.jussieu.fr)¹,

Laurence Galoisy (galoisy@lmcp.jussieu.fr)²,

Georges Calas (calas@lmcp.jussieu.fr)²,

Aline Ramos (ramos@lmcp.jussieu.fr)²,

Dominique Ghaleb (ghaleb@compote.cea.fr)¹,

Jean-Marc Delaye (delaye@compote.cea.fr)¹,

Bruno Boizot³ &

Bruno Reynart⁴

¹ CEA Valrho-Marcoule, DCC/DRRV/SCD - B.P.171, 30207 Bagnols/CEZE Cedex, France

² 4 place jussieu, Tour16 2ème étage - case 115, 75252 Paris Cedex 05, France

³ CEA Saclay, DSM/DRECAM/SRSIM, 91191 Gif sur Yvette Cedex, France

⁴ LST - ENS Lyon, 46 allée d'Italie, 69634 Lyon Cedex 05, France

X-ray absorption (XAS) and Raman spectroscopies together with x-ray diffraction (XRD), electronic microprobe analysis and molecular dynamics (MD) calculations were used on glass compositions representative of a light water waste containment matrix to investigate the structure of the glass, of its precipitates and the interaction between the glass and those precipitates. Noble metals (NMs) occur as RuO₂ and Pd₉₀Te₁₀ precipitates, the Pd-Te bond being non metallic as suggested by Pd K-edge EXAFS. Even though NMs are known to have low solubilities in borosilicate glasses, we show that they modify the medium range order of the glass network. Indeed, Si K-edge XANES spectra suggest that NMs modify the Si-O-Si angles. These data are supported by complementary Raman spectra, which suggest a change in the ring statistics, in NM containing glasses.The structural effect of NMs on the glassy matrix have also been investigated indirectly by EXAFS measurements around cations which are known to have well defined medium range environments, such as Zr and Zn. Zr and Zn are 6- and 4-fold coordinated respectively. Zn-Si and Zr-Si distances suggest that ZrO₆ as ZnO₄ polyhedra share corners with SiO₄ tetrahedra. Unlike Zr, Zn is interpreted as being in a network forming position. A third shell of neighbours around Zr was evidenced only in NM bearing glasses indicating a modification of the medium range order when NMs are present. Conversely, the Zn environment is insensitive to the structural modifications of the glassy network, suggesting that Zn atoms are not homogeneously distributed.This matrix may also contain precipitated alkaline and alkaline-earth molybdates whose unit cells and chemical compositions have been determined by Rietveld refinement and electronic microprobe analysis, respectively. In the absence of precipitates, Mo K-edge EXAFS and bond-valence considerations show that Mo occurs as molybdate groups (4-fold Mo⁶⁺) non connected to the silicate framework. The location of these molybdate complexes in the glassy matrix will be discussed.

J04 : 2A/12 : F2

Long Term Behaviour of Ultimate Glassy Wastes: Alteration of the Matrix and Mobility of Heavy Metals

Mikael Motelica-Heino

(mikael.motelica@univ-pau.fr)¹,

Olivier Donard (olivier.donard@univ-pau.fr)¹,

Arnaud Gauthier

(arnaud.gauthier@campus.univ-poitiers.fr)² &

Philippe Le Coustumer (philippe.lecoustumer@campus.univ-poitiers.fr)²

¹ LCBIE EP CNRS132, Hélioparc Pau Pyrénnées 2 av. Pierre Angot, 64053 Pau cedex 9, France

² Laboratoire des Matériaux et Géologie Environnementale UMR6532, 40 av Recteur Pineau, 86022 Poitiers, France

Management of nuclear wastes is a major issue for environmental protection as regards to the long-term behaviour of these materials and potential release of hazardous species. Because of the time scales involved, risk assessment studies are often based on theoretical models and numerical simulations. However besides those computational studies, evaluation methodologies based on experimental predictive models can also be useful. Solid waste incinerator fly ashes concentrate heavy metals (Zn, Cu, Pb, Sn, Cr and Ni). They follow a similar stabilisation process in a glass matrix using plasma technologies similarly to nuclear wastes. Therefore vitrified domestic wastes can be used to simulate nuclear wastes considering heavy metals as geochemical analogues for radionucleides.

In this work experimental simulations of the alteration of ultimate glassy wastes containing heavy elements were conducted. Several representative vitrified wastes materials with different composition and mineralogy were selected. The evolution of the glassy matrix and the mobility of heavy elements trapped within the glass were investigated under controlled altering conditions. Leaching tests were carried out to simulate weathering in different environmental conditions (aqueous, oxidising or reducing) using altering solutions. Thermal treatment at high temperature (above 1000°C) were also conducted to investigate the thermo-stability of these matrices. A global analytical strategy was developed to monitor the physical evolution of the matrix together with the distribution of the heavy metals trapped in the matrix. Physical characterisation techniques such as electron microscopy and XRD were used to study the structural and textural patterns of the original and altered materials. In situ microanalysis of heavy metals within the solid matrix was performed with laser ablation ICP-MS.

These experiments could estimate the long term evolution of the matrix and the behaviour of heavy metals. Physical alteration features and mineralogical evolution of the matrix were investigated and the partition of the heavy metals within the different mineralogical phases was established. Influence of the alteration of the matrix on the potential mobility of the heavy elements was then discussed.

J04 : 2A/13 : F2

Partial Melting in Fulgurites from Little Ararat (Turkey) and Dissistock (Swiss Alps)

Dieter Goedeke (dgoedek@gwdg.de) &

Kirsten Techmer (ktechme1@gwdg.de)

Geochem. Institut, Goldschmidtstrasse 1, Goettingen, Germany

Fulgurites (lat. Fulgur = lightning) are one of the rarely found glasses on earth, because their sizes are in the range of mm up to some cm. They originate from lightning strikes with peak temperatures in the return stroke of 30.000 K (Uman, 1987). The chemical and microstructural study of fulgurites as an undercooled glass derived from the most rapid melting process on earth, comprises the unique possibility, to study disequilibrium melting processes in natural environments. Rock and sand fulgurites together with their host rocks were studied by microprobe, SEM and X-ray techniques. The rock fulgurites derived from the summit of Little Ararat, Eastern Turkey and from Dissistock, Swiss-Alps. The hollow fulgurite tubes from Mount Little Ararat are about 1 mm in width and consist of a glassy rim which is 0,4 mm in width. The sand fulgurites are from Dixiana (South Carolina, USA).

Commonly, the chemical composition of the rock fulgurites can be compared with their host rocks or sometimes with one the major components of the adjacent rocks. The glass matrix of fulgurites from Little Ararat, show high Fe²⁺/Fe³⁺ ratios (about 5 wt-%/wt-%) in contrast to lower Fe²⁺/Fe³⁺ ratios of 0,5 wt-%/wt-% in the host rock, which is a hypersthene-diopside bearing andesite. All fulgurite glasses studied show a marked loss of H₂O (wt%) compared to the host rock. These ratios document an extreme reductive environment during the formation of the fulgurites, when lightning hits a rock. During this event, lasting only a few microseconds, the melted parts have no time to reach homogenization. In the fulgurite glass from Dissistock, threads of nearly pure SiO₂ (lechatelierite) appear in the glass matrix, probably resulting from congruent melting of quarz grains and indicating temperatures well above 1713°C as the melting point of beta-cristobalite.

There is no sharp boundary between lechatelierite threads and adjacent glass matrix. "Reaction rims" of some micrometer in diameter show a chemical composition, which is a mixture between lechatelierite and the glass composition. The glass matrix itself shows turbulent flow structures. Gaseous bubbles can be found in all fulgurites, indicating that boiling and vaporisation occur parallel to the partial melting process. The plagioclase lats at the rim of the glass matrix are partly melted and show intense microcracking. The cracks are orientated parallel to the boundary between the glass and host rocks.

Uman MA, Int. Geophys. Series, 39, (1987).

J04 : 2A/14 : F2

Characterization and Origin of Obsidians from Colombian and Ecuadorian Archaeological Sites

Ludovic Bellot-Gurlet (lbellot@ujf-grenoble.fr),

Gérard Poupeau (poupeau@ujf-grenoble.fr) &

Olivier Dorighel

UPRES-A 5025 CNRS and University Joseph Fourier, Institut Dolomieu, 15 rue Maurice Gignoux, 38031 Grenoble, France

We characterized by PIXE (Particle Induced X-Ray Emission) 175 archaeological obsidians from 50 Colombian and Ecuadorian sites (13 major and trace elements contents determined). Four compositional groups appeared, of which three correspond to those of known obsidian sources, also analysed by PIXE. The fourth one corresponds to three artefacts found in nearby sites on the South Colombian Pacific coast.FTD (Fission Track Dating), which gives the formation ages of obsidian glass, was applied to 57 of these artefacts. This geochronological characterization showed that in some artefacts compositional groups one may find discrete age groups. In each of these cases, one age group corresponds to that of a known obsidian source of similar composition and a source attribution can be proposed. The presence of other age groups implies different -presently unknown- origins for the artefacts belonging to these groups. A more detailed chemical characterization, by ICP-AES/-MS (36 major and trace elements) of obsidian sources and selected artefacts does not allow to resolve more chemical groups than PIXE analyses. This suggests a purely geochemical characterization of obsidian might be of limited use for sourcing studies.Thus we suggest for archaeological obsidian provenances studies in the Northern Andean area a combined PIXE/FTD approach.