295 296 Poster GM.qxd - Andra session_GM.pdf · p/gm/2 international meeting, september 17...>...18, 2007, lille, france page 299 clays in natural & engineered barriers for radioactive

Poster [GM]

Geochemistry & Mineralogy

r

P/GM/1

Page 297INTERNATIONAL MEETING, SEPTEMBER 17...>...18, 2007, LILLE, FRANCECLAYS IN NATURAL & ENGINEERED BARRIERSFOR RADIOACTIVE WASTE CONFINEMENT

INFLUENCE OF NATURAL SORBENTSIMMOBILIZATION OF SPENT IONEXCHANGE RESINS IN CEMENT

I. Plecas and S. Dimovic

Vinca Institute of Nuclear Sciences, P.O. Box 522, 11001 Belgrade,Serbia, ([email protected])

INTRODUCTION

Leach characteristics of 137Cs and 60Co radionuclides from spent mix bead ion exchange resins and bothordinary Portland cement and cement mixed with two kind of natural sorbents, (bentonite andclinoptilolite) have been studied using International Atomic Energy's (IAEA) standard leach method.

A study is undertaken to determine the waste immobilization performance of low-level wastes in cement-natural sorbents mixtures. The solidification matrix was a standard Portland cement mixed with 290-350(kg/m3) spent mix bead exchange resins, with or without 1-10 % of bentonite or/and clinoptilolite Theleaching rates from the cement-bentonite matrix as 60Co: (1.20-9.72)×10-5(cm/d) and for 137Cs: (1.00-9.22)×10-4(cm/d), after 300 days were measured. From the leaching data the apparent diffusivity of cobaltand cesium in cement-bentonite or/and clinoptilolite matrix with a waste load of 350 (kg/m3) spent mixbead exchange resins was measured as 60Co: (1.0-5.9)×10-6(cm2/d) and for 137Cs: (0.48-2.4)×10-4 (cm2/d)after 300 days. The compressive strength of these samples is determined following the ASTM standards.

These results are part of a 30-year mortar and concrete testing project which will influence the design ofradioactive waste management for a future Serbian radioactive waste disposal center.

Key words: solidification, leaching, cement, bentonite, clinoptilolite, cation exchange resins

Figure 1: De 60Co as a function of natural sorbents.

0

1

2

3

4

5

6

7

0 5 10 15 20 25

% sorbents

Dex106(cm2/d)

% bentonite (Co)

% clinoptilolite (Co)

% bentonite+clinoptilolite (Co)

P/GM/1

Page 298 INTERNATIONAL MEETING, SEPTEMBER 17...>...18, 2007, LILLE, FRANCECLAYS IN NATURAL & ENGINEERED BARRIERS

FOR RADIOACTIVE WASTE CONFINEMENT

EXPERIMENTAL CONCEPTSamples for leachability determination were prepared according to the IAEA standard Procedure.7 Leachant was exchanged and analysed for radioactivity after:1,2,3,4,5,6,7days, and thereafter every weekfor 1 month and from there on every month, until 300 th day. After each leaching period the radioactivityin the leachant was measured using EG&G- ORTEC spectrometry system and software.The volume of theleachant in every leaching period was 200 ml.Testing of concrete compressive strength is a classicalmethod which is practiced in civil engineering. Compressive strength (M), is expressed in MPa. A total of13 cube shaped concrete samples 10×10×10cm are prepared and tested at two different natural sorbentscontents (0,1,3,5,8,5 and 10 %) of cement. The compressive strength of these samples is determinedfollowing the ASTM standards. The results are presented in Table II. and Fig.1. where each point on graphrepresents the average of three tests.

RESULTS AND INTERPRETATIONTesting of mechanical characteristics of the cement-natural sorbents composite was performed with eachof the thirteen samples. We performed three measurements for each data point and the average value plus/minus standard deviation was determined for each data point. Table II gives compressive strength M(MPa)of thirteen samples after 28 days. Although mechanical characteristic is less important, it define additionalselective criteria for optimization of thirteen cement-natural sorbents based formulations.

The leaching rates from the cement-bentonite matrix as 60Co:(1.20-9.72)×10-5(cm/d) ,and for 137Cs:(1.00-9.22)×10-4(cm/d), after 300 days were measured. From the leaching data the apparent diffusivity of cobaltand cesium in cement-bentonite or/and clinoptilolite matrix with a waste load of 350 (kg/m3) spent mixbead exchange resins was measured as 60Co:(1.0-5.9)×10-6(cm2/d) and for 137Cs:(0.48-2.4)×10-4(cm2/d) after300 days. Table III gives the results of Incremental leach rate Rn(cm/d) and apparent diffusivity De(cm

2/d)60Co and 137Cs after 300 days, for spent ion exchange resins.

References:K. Andersson, B. Torstenfelt and B. Allard, 1981, Proceedings of the International Conference, ScientificBasis for NuclearWasteManagement, Vol.3, Plenum Press, New York, 235.

A. Atkinson and A.K. Nickerson, 1988, Nuclear Technology 81, 00.

R.H. Burns, Atomic Energy Rew., 1988, 9, 547.

H. Christensen, Nuclear and ChemicalWaste Management, 1982, 3,105.

F.P. Glasser F.P., Mat.Res. Soc. Proc., 2002, 713, JJ9.1.1-12.

I. Hashimoto, K.B. Deshpande, S.H. Thomas, I&EC Fundamentals, 1964, 3, 213.

E. D. Hespe, Atomic Energy Rev., 9 (1971) p.195.

I. Plecas, Lj.Mihajlovic and A. Kostadinovic, 1985, Radioactive waste management and nuclear fuel cycle,volume 6 (2), 161.

I. Plecas, J.Drljaca, A. Peric, A. Kostadinovic and S. Glodic, 1990, Radioactive Waste, Management andthe Nuclear Fuel Cycle, 14(3), 195.

I. Plecas , A. Peric, A. Kostadinovic, J. Drljaèa and S. Glodic , 1992, Cement and Concrete, Research anInternational Journal, 22, 937.

V. D. Marinin and G. N. Brown, 2000, Waste Management 20, 545-553.

A. Abusafa and H. Yucel, 2002, Sep. Purif. Technol. 28, 103-116.

H. Matsuzuru, N. Moriyama, Y.Wadachi and A. Ito A., 1997, Health Physics, 32, 529.

U.N. Yepimakhov and M.S. Oleinik, 2000, Ecological Chemistry, vol. 9, 116.

P/GM/2


CHEMICAL STABILITYOF RARE-EART DISILICATE

M.M. Orta,1 M.D. Alba,2 R. Alvero,2 A.I. Becerro,2 M.A. Castro,2 P. Chain,2

A. Escudero,2 M. Naranjo,2 E. Pavon,2 J.M. Trillo2

1. Centro de Investigación, Tecnología e Innovación de la Universidad de Sevilla. Avda. ReinaMercedes s/n. 41012 Sevilla, SPAIN

2. Instituto Ciencia de los Materiales de Sevilla. CSIC-Universidad de Sevilla. Avenida AmericoVespucio, 49. 41092 Sevilla, SPAIN

RE2Si2O7 is the final product of the chemical interaction of radioactive cations simulators [1] and thesilicates used in the engineering barrier of the deep geological repositories. One of the processes affectingthe stability of the disilicate phases is the dissolution of the silicate at different pH values. [2] The otherfactor that should be analyzed is the formation of cemented zones precipitation. It has been demonstratedthat the cement-bentonite interface has been found to be a chemically reactive zone. [3] Finally, it isimportant to establish the capacity of this phase to confine more lutetium cations as radioactive cationsimulators. [4]

It is the aim of this paper the study of the chemical and structural behaviour of Lu2Si2O7 (Lu as actinidesimulators) under several chemical attacks and can be summarized as follows: 1) Study of the dissolutionof lutetium disilicate in acid, neutral and alkali media. 2) Analysis of the stability of lutetium disilicate incement water. 3) Evolution of the lutetium disilicate under successive hydrothermal treatments with a newlutetium solution.

The conditions expected at the repositories are simulated in the laboratory with solutions of differentcompositions and pH values. Both the solid products and remnant solutions are analyzed by the use oftechniques that inform of the short- and long-range structural order.

The results show a good stability of Lu2Si2O7, under the studied conditions, and, therefore, theencapsulation of the radioactive actinide cations as a disilicate phase could be considered a method thathelps to ensure the long time safety of the repositories. We have demonstrated that Lu2Si2O7 is quiteinsoluble at a wide pH range although the solubility increases slightly at lower pH values. However, evenat pH values as low as 2 there is not appreciable leaching of actinide simulatior or phase recrystallization.The only observed effect is an increasing disorder at short range order.

We can establish that the degradation of cement at short-term does not affect the stability of the Lu2Si2O7and the changes observed is of a similar extent that those observed for the alkali attack. However, at long-term the cement degradation can affect more drastically its stability which can not be explained by thealkali conditions. We think that the presence of calcium in the solution induces a rearrangement of thestructure without the dissolution of the structural ions.

Finally, we have demonstrated the capacity of the new phase to immobilize the lutetium cation as an oxidephase, once all the silicon atoms have been consumed and the growing of Lu2Si2O7 is not possible.

ACKNOWLEDGMENT

We gratefully acknowledge financial support from DGICYT Projects no. CTQ2004-05113 and from theEuropean Commission for the project funded within the 6th Framework Programme as an HRM Activityunder contract number MRTN-CT-2006-035957.

P/GM/2



Reference[1] Chapman, N.A., Smellie, J.A.T., 1986. Introduction and summary of the workshop. Chemical Geology55, 167-173.

[2] Jozefaciuk, G., Bowanko, G., 2002. Effect of acid and alkali treatments on surface areas and adsorptionenergies of selected minerals. Clays and Clay Minerals 50, 771-783.Komadel, P., Schmidt, D.,Madejová, J., Èièel, B., 1990. Alteration of smectites by treatments with hydrochloric acid and sodiumcarbonate solutions. Applied Clay Science 5, 113-122.

[3] Berner, U.R., 1992. Evolution of pore water chemistry during degradation of cement in a radioactivewaste repository environment. Waste management 12, 201-219. De Cannière, P., Moors, H., Wang, L.,Put, M., Gens, R. “Effect of cement water on the chemistry of Boom Clay.” Conf.: Clays in natural andengineered barriers for radioactive waste confinement, 9-12 december 2002, Reims, France.Proceedings: Clays in natural and engineered barriers for radioactive waste confinement, pp 413-414.

[4] Leturcq, G., Berger, G., Advocat, T., Vernaz, E., 1999. Initial and long-term dissolution rates ofaluminosiloicate glasses enriched with ti, Zr and Nd. Chemical Geology 160, 39-62.

P/GM/3


MINERALOGICAL HETEROGENEITY OFROKLE BENTONITE AND RADIONUCLIDEADSORPTION: A CASE STUDY FOR CESIUM

J. Vejsada, A. Vokál

Nuclear Power and Safety Division, Nuclear Research Institute, Husinec Rez 130, 25068 Rez byPrague, Czech Republic ([email protected])

INTRODUCTION

Bentonite is worldwide studied as the most suitable clay material for sealing a deep geological repository(DGR). In most of countries, Na-bentonites (e.g. Wyoming MX-80 bentonite) were selected as a buffer/backfill material. Since in the Czech Republic there are abundant deposits of Ca-Mg bentonite, thismaterial has been proposed for Czech DGR reference concept. In contrast to Na-bentonites, Czechbentonite from the largest operating deposit Rokle represents complex mixture of clay and other mineralsthat may have a significant impact on bentonite properties and its suitability as a sealing material in DGR.For cesium adsorption, the mineral heterogeneity problem is related mainly to the presence of selective claysorbents (mainly illite and vermiculite). In order to investigate the effect of mineral composition changes,visibly different samples were selectively taken from the raw bentonite, analyzed and adsorption of cesiumon them has been determined using batch technique. The results were then compared to data for a sievefraction from a ground sample of whole material used in other geochemical and geotechnical laboratoryexperiments and also for a sample from the Rokle deposit used in previous experiments.

EXPERIMENTAL

In the first step, two predominant visibly different portions of bentonite, red (“R”) and green (“G”)bentonite, were taken from the raw material obtained directly from the deposit. Samples were dried andgrounded to fine grain. In the second step, more than 50 kg of the whole raw material has been milled incentrifugal mill and three fraction were collected: 0-2 mm (“F02”, used for laboratory tests), 2-4 mm andabove 4 mm (simulating possible rests removed during factory processing). Three clay samples (R, G, F02)were then analyzed in detail using wet chemical analysis and X-ray diffraction to determine theirmineralogical composition. Cation exchange capacity (CEC) has been determined using Cu-trien method(Meier and Kahr, 1999).

The sorption of cesium has been determined using batch-sorption method in 0.1 M CaCl2 solution atsolution-to-solid ratio 25 cm3.g-1 in cesium concentration range 10-2-10-7 M with 137Cs as a tracer and thedata were fitted using Freundlich adsorption isotherm. For capacity determination of cesium selectivesorption sites (Frayed Edge Sites, FES), present mainly on illite and vermiculite, the AgTU methoddescribed by de Koning et al., 2007 has been applied.

RESULTS AND DISCUSSION

The mineralogical analysis (Fig. 1) confirmed significant differences among selected samples. It was foundthat especially kaolinite, micas and carbonates are depleted in F02 sample, while these mineral phases areimportant admixtures to predominant smectite in sample G (carbonates) and R (micas, kaolinite). Sinceboth G and R bentonites were found accumulated in fraction above 2 mm, it can be said that sample F02,which consist of 0 – 2 mm fraction, is markedly depleted in these minerals and enriched in smectite bythe milling process. This has been confirmed by cation exchange capacity measurement; the CEC of F02fraction was found higher then CEC of G and R samples. In contrast to the sample from the same depositused in previous experiments (“V” sample), no vermiculite has been identified in X-ray diffraction patternsof the samples studied in this work

P/GM/3



Adsorption data were found to be in good agreement with mineralogical composition and CECmeasurements. Comparing the Kd trends, all these samples behave similarly. In comparison to the samplecontaining vermiculite (V sample), some distinctions were found that could be attributed mainly to thepresence of this mineral. The capacity of FES measured by the AgTU method in F02 sample was 36 µeq/g, in G sample 27 µeq/g, R sample 17 µeq/g and V sample 34 µeq/g. The results suggest that the biotiteinstead of illite represents main K-mica phase in R sample because of the lowest FES capacity value inrelation to total amount of micas in the sample. The positive correlation of FES capacity with smectitecontent for samples R, G and F02 also suggests possible effect of illite in illite/smectite mixed structures.But the FES capacity data should be taken and interpreted with care, with respect to known facts (see deKoning et al., 2007) and in our case also with respect to limitations of analytical methods (illite andvermiculite identification and quantification).

It follows from the results that mineral admixtures in bentonite from the Rokle deposit may play animportant role in sorption properties of this bentonite. Also sample preparation has significant effect onsample characteristics, especially smectite and other mineral phases content and their proportion.

References:de Koning A., Konoplev A. V., Comans R. N. J., 2007, Measuring the specific caesium sorption capacityof soils, sediments and clay minerals. Appl. Geochem. 22, 219–229

Meier, L.P., Kahr, G., 1999, Determination of the exchange capacity (CEC) of clay minerals using thecomplexes of copper (II) ion with triethylenetetramine and tetraethylenepentamine. Clays Clay Miner.47, 386–388.

Figure 1: Part of bulk air-dried diffraction patterns of R, G and F02 samples of Rokle bentonite. Arrowsindicate identified mineral phases.

P/GM/4


RHEOLOGICAL AND SORPTIONPROPERTIES OF CLAY-POLYMER

COMPOSITESH.P. Zadvernyuk, Yu.G. Fedorenko, B.P. Zlobenko, T.I. Koromyslichenko

Institute of Environmental Geochemistry of National Academy of Science of Ukraine

The properties of alkaline bentonite to swell in water, as well as capacity for sorption of radionuclides andorganic matters, determined use of bentonite as main component of engineered geochemical barriers.Recently, clay-polymer composites – new materials based on clays and polymers – are involved to createinsulant barriers. As a mineral component alkaline bentonites are used the most often. Many polymers aresuitable for preparation of clay-polymer composites. We focus on properties of clay-polymer compositesbased on bentonite and polyacrylamide.

It is known that addition of polyacrylamide solution into bentonite gel changes rheological properties ofclay-polymer suspension. Its yield value increases that is structural strength of swelling in the watercomposite. As a result, using composite in the mixture with ballast materials (sand, road metal, etc.) notonly decreases filtration coefficient but also increases resistance to hydraulic pressure. That means thatinsulant characteristics of barrier materials are improved. On the other hand, obtaining composites forbarrier materials based on stable suspension clay particles-polyacrylamide on the surface of these particlesor their aggregation forms layers of sorbing polymer molecules, which together with developed spatialstructure of polymer chains binds polymer with clay particles in composite. At the same time, active sites,involved in sorption of radionuclides, can be blocked. This process depends on many factors: correlationbetween bentonite and polyacrylamide in the system, water amount, hydrolysis degree of polyacrylamide,properties of radionuclides and etc.

Yield value is important rheological characteristic that shows degree of interaction between polyacrylamidemacromolecule with bentonite particles surface and state of polyacrylamide macromolecule in thecomposite suspensions.

The study of polyacrylamide effect on yield value of composite suspensions has shown that at theconcentration of anionic polyacrylamide from 7·10-3 to 20·10-3 %, yield value is the highest in the case ofthe use of anionic polyacrylamide with medium molecular weight 14·106 g mol-1 and hydrolysis degree24%. Yield value regardless of polymer type rises with increasing of its concentration. Low yield valuewas observed for composite based on cationic polyacrylamide with high molecular weight 39·106 g mol-1.In the composition with 1% content of anionic polyacrylamide yield value is 1,2 Pa.

The sorption of 137Cs at the correlation composite to the solution 1:100 and initial activity of the solution33,5 Bq ml was examined. At the concentration of different type polyacrylamide in the composites from7·10-3 to 20·10-3 % the degree of solution purification was high and independent of polyacrylamide typeand its concentration in the composite.

Sorption capacity of the composite with 1% polyacrylamide content with hydrolysis degree 30% andmolecular weight 26·106 g mol-1 was also estimated. Initial activity of the solution was 131Bq ml-1. It wasdetermined experimentally that at such polyacrylamide concentration sorption capacity of the compositewas also high (> 98,8%).

The sorption of 90Sr on clay-polymer composites based on alkaline bentonite and anionic polyacrylamideswas also investigated. Initial activity of the solution was 6,5 Bq ml-1, the correlation composite to thesolution was 1:100. The experiments have shown that at the polymer concentration to 1% sorption capacity

P/GM/4



of the composite for 90Sr was high (≥ 97,5%) and independent of polymer molecular weight and itshydrolysis degree.

The comparison of clay-polymer composites based on anionic and cationic polyacrylamide shows that atthe same their concentration in the composite effect on rheological characteristics of aqueous suspensionwas stronger in case of anionic polyacrylamide than cationic one. At investigated concentration differenttypes of anionic polyacrylamides has little influenced on high sorption capacity of clay-polymer composite.At the same time dependence of rheological characteristics – yield value – on physical and chemicalproperties and polyacrylamide concentration is observed. Yield value can control in a wide range from4,0·10-3 to 1,2 Pa. It allows on their base to create effective sorbent for radionuclides with high rheologicalcharacteristics.

Structural features of the composite explain the fact that at rather high content of anionic polyacrylamidesin the composite – to 1%, bentonite sorption capacity was remained. Dimensional network ofpolyacrylamide macromolecule practically did not blocked montmorillonite active sites did not preventpenetration of radionuclides to crystal surface.

Analysis of clay-polymer composites properties has shown that they are depending on composition.Rheological properties depend on polyacrylamide amount and macromolecule characteristics. Structuralfeatures of montmorillonite crystal determine sorption properties. Control rheological characteristics canconduct by using polyacrylamides with appropriate concentration, hydrolysis degree and molecular weightetc. Sorption characteristics can control by selecting and preparation of sorbent: disaggregating,purification of their surface, etc.

Clay-polymer composites are perspective at practical application as powder. Prepared clay-polymersuspension should be dried and ground.

P/GM/5


CLAYMINERAL INTERACTIONSWITH LEACHATE SOLUTIONS IN LANDFILLS

S. Battaglia M. Cervelli

Istituto di Geoscienze e Georisorse, Italian National Research Council, Via Moruzzi 1, 56124, Pisa,Italy

INTRODUCTIONNowadays, identifying suitable sites for disposing of urban waste in landfills has become an urgent priority.Clayey covers are often chosen as landfill sites because of their insulating properties. Consequently, studyof the interactions between the aqueous solutions circulating in the dump (leachate) and the clayey countryrock is of considerable relevance. Dump leachate is a liquid mixture formed by the interaction of theproducts of waste decomposition with the meteoric water infiltrating into the waste mass (Mora - Naranjoet al., 2004). The aim of the study presented herein is to verify the possible transformations undergone bythe clays due to their contact with dump leachate. The quantity and chemical composition of such productsvary widely, as they are dependent on a number of variables, including the composition of the waste mass,its time of residence in the landfill, the chemical, physical and biological processes taking place, thechemical and physical conditions within the fill, and lastly the qualitative and quantitative characteristicsof the percolating water.

The results of the study reveal that the leachate of the dump examined was essentially an ammoniumbicarbonate solution with a strong sodic component. The dump studied is located in the town ofBuriano, a few kilometres (2.5-3) from the village of Saline di Volterra; the geological setting of thestudy area is illustrated on the Region of Tuscany Website: http://sit.lamma.rete.toscana.it/scripts/sisterims.dll?Run?svr=MAPSERVER1&Func=open&map=%22geo250k%22&html= (sheet 295Pomarance). The area presents as a morphologically flat depression with a longitudinal extension ofabout 1.5 km. It is almost wholly contained within an Upper Pliocene clay formation. A further aimof the study is to apply the geochemical model (PHREEQC see Parkhurst &Appelo 1999) to simulatethe kinetics and thermodynamics of mineral dissolution-precipitation occurring in the landfill, andthereby arrive at determinations of the possible transformations in the clays due to their interactionwith the leachate solution. The results of such modelling are then compared with those determinedexperimentally on samples drawn from the area.

EXPERIMENTAL STUDYIn the preliminary stage of this study, four clay samples were taken from the landfill area: two from theimmediate perimeter around the dump (i.e., the tipface; samples 1 and 2) and two others from about 40mfrom the tipface in both the southeast (3) and direction northwest (4) directions.

The samples were analysed via powder X-ray diffractometry, according to the methodology set out inMoore & Reynolds (1997) with the aim of identifying the clayey phases present, scanning in both the air-dried (AD) and glycolated (EG) state . The results of the analyses are shown in Table 1 below reported:

Samples Clay minerals1 chlorite; illite; smectite-illite mixed layers (I-S)

2 chlorite; illite; smectite-illite mixed layers

3 chlorite; illite

4 chlorite; illite

Figure 1 presents the diffraction spectrum of sample 1, under EG conditions to highlight the formation ofI-S. Peak profile fittings were performed with a Pearson VII function using theWINFIT computer program

P/GM/5



(Krumm, 1996) by retaining the Kα 2 wavelength. SEM analysis of samples 3 and 4 yielded the followingchemical compositions respectively for the chlorite and the illite: Chl (Fe2.5Mg2.5Al2Si3O10(OH)8); Il(K0.6Mg0.25Al2.3Si3.5O10(OH)2). Such composition was used as the initial input for the simulation program.

THERMODYNAMIC DISSOLUTION-PRECIPITATION EQUILIBRIUMThe goal of the simulation was to verify the experimental XRD analysis data, that is, if the presence of I-S in the samples from the tipface (1 and 2) can be attributed to possible interactions between the clays andthe leachate. The minerals considered in the simulation were those found to be present in samples 3 and4, since the area belongs to a different hydrographical basin from the dump, and therefore not influencedby it. The results of the mineral dissolution–precipitation calculations carried out with PHREEQC aresummed up in the graph in figure 2. (For an explanation of the significance of the abscissas and ordinates,see Parkhurst and Appelo, 1999). The simulation reveals that the chlorite is in equilibrium with theleachate, while the illite is dissolved on contact with the leachate solution, thereby allowing ions to go intosolution. This, in turn, leads to the precipitation of other phyllosilicates, such as muscovite and smectite.

CONCLUSIONSThe dissolution-precipitation data obtained via simulation of the thermodynamic equilibrium are consistentwith the experimental data: interaction of the leachate with the clayey minerals of the dump basin,represented by chlorite and illite, lead to the dissolution of illite and the precipitation of smectite, asdemonstrated experimentally (illite reacts with the leachate to transform into a mixed I-S layer). The cationexchange capacity of the newly formed mineralogical phase is far superior to the initial phase. Interest insuch ion exchange capacity is based on the need to remedy environmental contaminants at disposal and/or accident sites. The current research study on clay minerals at a dumpsite points to likely contaminationof surrounding areas by dump leachate. However, such tentative findings remain to be confirmed bychemical and isotope analysis of the underground and surface waters.

ReferencesKrumm S. (1996): WINFIT 1.2: version of November 1966 XIII Conference on Clay Mineralogy andPetrology, Praha, 1994. Acta Universitatis Carolinae Geologica, 38, 253-261.

Moore M.D. & Reynolds R.C. (1997) X-ray Diffraction and the Identification and Analysis of ClayMinerals. Oxford University Press, Oxford-New–York, 378 pp.

Mora-Naranjo N., Meima J., Haarstrick A., & hempel D.C., (2004): Modelling and experimentalinvestigation of environmental influences on the acetate and methane formation in solid waste. WasteManagement & Research, 763-773.

Parkhurst D.L., Appelo C.A.J., (1999) User’s guide to PHREEQC Vers.2 A computer programfor speciation, batch-reaction one-dimensional transport and inverse geochemical calculations,www.xs4all/~appt/index.html.

Figure 1: Decomposition of the XRD patternsobtained from EG preparations of sample 1.

Figure 2: Mineral dissolution/precipitation.

0.0 0.2 0.4 0.6 0.8 1.0

Reaction progress

-5

-4

-3

-2

-1

0

1

2

3

Deltamoles

Muscovite

Smectite

Chlorite

Illite

P/GM/6


LITHIUM ISOTOPE FRACTIONATIONDURING ADSORPTION ONTO MINERAL

SURFACES

R. Millot, J.P. Girard

BRGM, Metrology, Monitoring and Analysis Division, 3 av. C. Guillemin BP 36009 45060 OrléansCedex 2, France, ([email protected], [email protected])

INTRODUCTIONLithium has two isotopes of mass 6 and 7 (with natural abundances of 7.5% and 92.5% respectively) andis a mobile element that tends to preferentially partition into the fluid phase during water/rock interaction.The relative mass difference between the two isotopes is considerable (17%) and is known to generate largemass dependent fractionation during geochemical processes. Indeed, the range of variation in Li isotopiccompositions is more than 50‰ in geological materials (Tomascak, 2004). Pistiner and Henderson (2003)showed that adsorption processes onto clay minerals may potentially induce significant Li isotopefractionation. However, the magnitude of adsorption-related Li fractionation remains poorly documented.We report here the results of a study conducted in conjunction to the ISOBLiFe research project betweenBRGM and ANDRA, in order to determine experimentally the extent of Li isotope fractionation as a resultof adsorption on clays and gibbsite.

EXPERIMENTAL DEVICE AND ANALYTICAL METHODSLaboratory experiments consisted in measuring the evolution through time of Li isotopic composition of asolution in equilibrium with a mineral powder. The following clay minerals specimen were used: smectite(SWy-2), kaolinite (KGa-2), illite (IMt-2), chlorite (CCa-2) from the Source Clays Repository and gibbsite(REFG) from Girard and Savin (1996).

Experiments were carried out at 25°C with a solution/mineral mass ratio of 40. 10 ml of reference solutionIRMM-016 (δ7Li = 0‰, pH = 7) and 250 mg of powdered mineral were placed in a screw-top Teflon® PFAbeaker. The beaker was kept in a temperature-controlled oven during several weeks, which temperaturewas maintained within 5% of target temperature over the total duration of the experiments. Aliquots of thesolution (100 µL after filtration at 0.20 µm) in contact with minerals were periodically sampled (from daysup to 55 weeks) and analyzed for lithium concentration and Li isotopic composition (Fig. 1).

Lithium isotopic compositions were measured using a Neptune Multi-Collector ICP-MS at BRGM'sIsotopic Geochemistry Laboratory (Millot et al., 2004). 7Li/6Li ratios were normalized to the L-SVECstandard solution (NIST SRM 8545) following the standard-sample bracketing method. Typical in-runprecision on the determination of δ7Li is about 0.1-0.2‰ (2σm). Before mass analysis, solution sampleswere prepared with chemical separation/purification by ion chromatography in order to produce a puremono-elemental solution. Chemical separation of Li was achieved using a cationic resin and HCl acidmedia (0.2N). Accuracy and reproducibility of the entire method (purification procedure + mass analysis)was estimated from repeated measurements of seawater (IRMM BCR-403), yielding a mean value inagreement with our long-term measurements (δ7Li = +31.0‰ ± 0.5, 2 σ, n=30, Millot et al., 2004) and withvalues reported in the literature.

RESULTSVariation of Li concentrations through time indicates that the amount of Li adsorbed onto clay minerals issmall (10-20%), in contrast to gibbsite (up to 80%). Solution δ7Li increases through time, indicating anenrichment of the solution in heavy isotope 7Li (Fig. 1) as a result of preferential adsorption of 6Li onmineral surface. This agrees with Pistiner and Henderson (2003). It is observed that after several weeks of

P/GM/6



reaction, solution δ7Li reaches plateau values of +4.0, +7.1, +0.9 and +3.8‰ for smectite, kaolinite, illiteand chlorite respectively. A significant higher δ7Li plateau value of +24.8‰ is observed for gibbsite.

These results indicate that Li isotope fractionation associated with adsorption processes is not significantfor illite, intermediate for smectite, chlorite and kaolinite, and elevated for gibbsite. This suggests thatminerals richest in OH groups show the greatest magnitude of fractionation. No relation was found betweengrain size and/or specific surface area and magnitude of fractionation. Our experiments illustrate thatadsorption processes on clays may influence the Li isotope composition of interstitial water in argillaceousrocks. This will be further discussed in light of additional data acquired in the ISOBLiFe project in orderto characterise the Li isotopic composition of pore water in the Callovo-Oxfordian argillite.

References:Girard, J.P., & Savin, S.M. (1996): Intracrystalline fractionation of oxygen isotopes between hydroxyl andnon-hydroxyl sites in kaolinite measured by thermal dehydroxylation and partial fluorination.Geochimica et Cosmochimica Acta, 60, 469-487.

Millot, R., Guerrot, C. & Vigier, N. (2004): Accurate and High-precision measurement of lithium isotopesin two reference materials by MC-ICP-MS. Geostandards Geoanalytical Research, 28, 153-159.

Pistiner, J.S. & Henderson, G.M. (2003): Lithium isotope fractionation during continental weatheringprocesses. Earth and Planetary Science Letters, 214, 327-339.

Tomascak, P.B. (2004): Developments in the Understanding and Application of Lithium Isotopes in theEarth and Planetary Sciences. In Reviews in Mineralogy & Geochemistry, 55, 153-195.

Figure 1: Evolution of Li isotopic composition of solution through time (initial δ7Li = 0‰) during isotopeexchange experiments with clay minerals and gibbsite.

P/GM/7


SORPTION OF SR2+ ONTO MIXEDSMECTITE / ILLITE CLAYSTiziana Missana, Miguel García-Gutiérrez, Úrsula Alonso

CIEMAT, Departamento de Medioambiente Avenida Complutense 22, 28040 Madrid, SPAIN([email protected]).

ABSTRACTThe radionuclide sorption is amongst the principal aspects to be evaluated for the assessment of clays asbarriers for contaminant isolation. In particular, clays are considered as adequate barriers in high-levelwaste radioactive geological repositories.

To interpret and quantify the interactions occurring at the clay/water interface by means of (quasi)thermodynamic models an experimental effort is needed: for the basic understanding of the sorptionmechanisms the effects of the most important physico-chemical parameters such as pH, ionic strength andradionuclide concentration have to be studied independently.

In this study we analyzed the sorption of Sr(II) onto illite and illite/smectite mixtures. Basically, wedetermined the sorption parameters by ion exchange and surface complexation models in the pure Na-smectite and pure Na-illite systems and tried to evaluate the capability of reproducing the behavior of themixed system verifying the degree of comprehension of the system. The final aim is to extrapolate theresults of the sorption models to natural clay-rock conditions, considering the real geochemicalenvironment.

The smectite, extracted from the FEBEX bentonite, and a “standard” and well-characterised illite (SilverHill, Montana, SH) were used in these experiments. Sorption experiments with 100% smectite and 100 %smectite were carried out and then two selected mixtures of Na-smectite and Na-Illite (namely 75%smectite + 25% illite and 50% smectite + 50% illite) were used.

Sorption edges (from pH 2 to 11) at different ionic strengths (from 1·10-3 M to 2·10-2 M) were carried outand liquid to solid ratio effects have been also evaluated. Sorption isotherms at 1-2 different pH anddifferent ionic strengths were also made.

The sorption data of Sr in Na-smectite were already satisfactorily modeled (Missana and Garcia-Gutierrez,in press). The main sorption mechanism for Sr on the Na-montmorillonite resulted to be the ionic exchangewith a linear sorption in the whole range of concentrations investigated. The mean logarithm of selectivitycoefficient, with respect to Na, obtained from the above-mentioned sorption studies, considering Sr traceconcentration, was 0.66 ± 0.06. A small contribution due to surface complexation at the clay edge sites(SOH) had to be considered to fit adequately the sorption results obtained at pH higher than 8 and higherionic strengths.

A similar model was used in this work for pure illite: again, the main sorption mechanism was the ionicexchange with a mean logarithm of the selectivity coefficient of 1.40 ± 0.09 and a minor contribution ofsurface complexation was observed at higher pH.

In order to test the Sr sorption behavior on the smectite/illite mixtures, sorption edges at 0.1M were carriedout, using a mixture of smectite 50 % and illite 50 % and a mixture of smectite 75 % and illite 25 %. Dataof the mixtures were compared with those of pure smectite and pure illite.

The sorption in the mixtures was also satisfactorily modelled using the sorption parameters previouslyobtained from the tests in pure Na-illite and Na-smectite systems.

P/GM/7



The simulations corresponded very well to the experimental data and the modelling allowed to show thatthe sorption due to ionic exchange is slightly higher when the montmorillonite content increase, but at pHhigher than 9, when surface complexation starts to make the difference, the higher the illite content, thehigher the sorption. Results showed that it is possible to model a system composed by a mixture of claysstarting from data extracted in the experiment with single minerals.

ACKNOWLEDGEMENTSThis work has been carried out in the frame of the ENRESA-CIEMAT association and partially funded bythe EU within the FUNMIG (Fundamental Processes of radionuclide Migration) Project (Ref. FP6-516514).

References:Missana, T. and Garcia-Gutierrez, M., in press. Adsorption of bivalent ions (Ca(II), Sr(II) and Co(II)) ontoFEBEX bentonite. Physics and Chemistry of the Earth.

P/GM/8


EH AND PH IN THE POREWATEROF COMPACTED BENTONITE

A. Muurinen and T. Carlsson

VTT Technical Research Centre of Finland Otakaari 3 A Espoo, P.O. Box 1000, FI-02044 VTT,Finland ([email protected]), ([email protected])

INTRODUCTION

Knowledge of the pore water chemistry in compacted water-saturated bentonite is important, e.g. whenassessing the corrosion of the waste canister or the behaviour of released radio nuclides in a repository fornuclear waste. However, the combination of low water content and high swelling pressure up to severalMPa makes direct measurements in the bentonite difficult. Mostly, measurements are therefore madeindirectly, which means that the water is separated from the bentonite, e.g. by squeezing prior to theanalysis. Such a method is always associated with uncertainty concerning the extent to which the chemistryof the squeezed pore water represents the ‘true’ chemistry of the pore water in the bentonite. Thisuncertainty can be avoided by measuring directly inside the wet bentonite, which requires that theanalytical instruments inside the bentonite can withstand the high swelling pressure and also function wellat low water contents. The following describes a technique that was developed for measuring pH and Ehin compacted water-saturated bentonite and results obtained for MX-80 bentonite in different conditions.

EXPERIMENTAL

The aim of the work was to study the evolution of the pore water chemistry of bentonite, especially pHand Eh, in anticipated repository conditions. The term porewater in this case means the free water in thelarge pores i.e. the part where chloride ions can enter. The focus has been in direct measurement of pHand Eh in compacted bentonite. Solid IrOx pH electrodes were prepared by means of high-temperatureoxidation method (Yao et al. 2001), in which an iridium oxide film is formed on Ir metal wire in lithiumcarbonate melt. The Eh electrodes (Pt and Au) consisted principally of pure metal wires. The reference

Figure 1: Squeezing cell used to study pH and Ehin compacted bentonite in closed conditions.

Figure 2: Diffusion cell experiment, where bento-nite is in contact with the external solution.

P/GM/8



electrode used was a commercially available Ag/AgCl electrode. The studies have comprised testing of pHand Eh electrodes in solutions, measurements in compacted clay in “squeezing” cells (Fig. 1), experimentsin “diffusion cells”, where the conditions in the external solution have been changed during the experiment(Fig. 2), and measurements of oxygen in bentonite.

The measured pH values in the squeezing cell experiment have been in line with those obtained bygeochemical modeling as seen in Figure 3. Oxygen measurements (Table 1) suggest that oxygen isconsumed quickly from compacted bentonite.

The results of the diffusion cell experiments, seen in Figures 4 and 5, have shown that the chemicalchanges of Eh in the external solutions in contact with the bentonite are quickly reflected as changes ofEh inside the bentonite, to the depth of 5 mm from the surface. After four months of interaction, it seemsthat the changes are buffered by some chemical reactions and will not progress quickly deeper into thebentonite, however.

ReferencesYao, S., Wang, M., Madou, M., 2001. A pH electrode based on melt-oxidized iridium oxide. Journal of TheElectrochemical Society, 148, (4) H29-H36.

Tableau 1: Oxygen in the porewateras a function of time.

Experimental time(days)

O2 in porewater (mg/l)

0.08 12.1

1 1.3

9 0

19 0

Figure 4: Eh at different depths of a diffusion cellsample after changing of the conditions in theexternal water from anaerobic to aerobic.

Figure 5: Eh at different depths of a diffusion cellsample after changing of the conditions in theexternal water from aerobic to anaerobic.

Eh with Pt electrode

-500

-400

-300

-200

-100

0

100

200

300

0 100 200 300 400 500

Time (days)

Eh(mV)

Eh 5 mm

Eh 10 mm

Eh 20 mm

Moved to aerobic conditions

Eh with Pt electrode

-100

0

100

200

300

0 100 200 300 400 500 600

Time (days)

Eh(mV)

Eh 5 mm

Eh 10 mm

Eh 20 mm

Moved to nitrogen glove-box

Figure 3: pH in bentonite in a squeezingcell measurement.

6

7

8

9

10

0 20 40 60 80 100

Time (day)

pH

Meas. d.i.water

Meas.saline sol.

Mod. d.i.water

Mod. Saline sol.

P/GM/9


CHEMICAL INTERACTION OF 152EUWITH THE CLAY BARRIER

P.Chain1, M.D. Alba1, A.I. Becerro1, M.A. Castro1, A. Escudero1,T. González-Carrascosa1, S. Hurtado2, E. Pavón1, M. Villa2

1. Instituto Ciencia de los Materiales de Sevilla. CSIC-Universidad de Sevilla. Avenida AmericoVespucio, 49. 41092 Sevilla, SPAIN

2. Centro de Investigación, Tecnología e Innovación de la Universidad de Sevilla. Avda. ReinaMercedes s/n. 41012 Sevilla, SPAIN

The interaction of 4f elements with natural and synthetic phyllosilicates during thermal and hydrothermaltreatments has been studied in a systematic and fundamental way. The existence of a reaction mechanism,which was not previously described, based on the chemical interaction between the lanthanide cations andthe orthosilicate anions of the lamellar structure has been identified [1] This finding has applied interestbecause lanthanides are used as simulators of high activity radionuclide (HAR) in agreement with theguidelines established in the bibliography [2] It has been observed that in conditions of moderatetemperature and pressure a chemical interaction exists between smectites and rare earth elements (RE) andphases of insoluble disilicate, RE2Si2O7, which would immobilize RE, are generated [3]. It is remarkablethat the reaction extends to all the set of the smectites [4], although they do not display the same reactivity,the saponite being the most reactive.

Although the structural changes suffered by the clay after the treatments have been extensively studied aquantification of the retention of the RE by the generation of the new insoluble phase, RE2Si2O7, has notbeen performed. Thus, it is the aim of this work thecombined study of the structural changes and the distributionof the RE cations (HAR simulators) between the solid andliquid part of the reaction. Eu3+ has been chosen, among theRE cations, because its physical properties (cation size,hydrolization constant, oxidation state…) make it a goodsimulator2 of Am3+ and Cm3+ and it is available as stable,151Eu y 153Eu, and radioactive isotope, 152Eu.

Different portions of saponite clay were hydrothermallytreated (at different temperatures and reaction times) with40 mL of Eu(NO3)3 7.9·10

-2M. Two sets of experimentswhere performed in parallel, the first one containing purelystable Eu isotope and the second one was enriched with 152Euisotope up to a total activity of 9.8 Bq. The solid samplescorresponding to the first set of experiments were analyzedby X-ray diffraction, scanning electron microscopy, andenergy dispersive spectroscopy. The experiments involvingthe radioactive 152Eu isotope were analyzed using low levelgamma-ray spectrometry.

X-ray diffraction has shown the transformation of thelayered structure of the starting clay to a new phasecontaining Eu, F-Eu2Si2O7.Figure 1 show the electron micrographs and the EDX analysis of the new phaseand the starting clay. The analysis of both solids and liquid phase of the reaction allow determining thedistribution of 152Eu between the solid phase and the solution, a quantification of the retention degree beingpossible.

These results will allow establishing a methodology for the study of the reaction between HAR cations(Am, Cm, Th, U, Pu..), where structural analysis is not possible, and the clay barrier.

Figure 1: Structure, SEM micrograph andEDX spectra of the original clay (a) and thesolid phase obtained alter the hydrothermaltreatment.

P/GM/9



ACKNOWLEDGMENTWe gratefully acknowledge financial support from DGICYT Projects no. CTQ2004-05113 and from theEuropean Commission for the project funded within the 6th Framework Programme as an HRM Activityunder contract number MRTN-CT-2006-035957.

Reference[1] J.M. Trillo, M.D. Alba, R. Alvero, M.A. Castro, A. Muñoz-Páez, J. Poyato. Inorg. Chem., 1994, 33,3861-3862

[2] N.A. Chapman, J.A.T. Smellie. Chem. Geol., 1986, 55, 167-173

[3] A. I. Becerro, M. Naranjo, M. D. Alba, J.M. Trillo, J. Mater. Chem., 2003, 13, 1835-1842

[4] M.D. Alba, A.I. Becerro, M.A. Castro, A.C. Perdigón, Amer. Miner., 2001, 86, 115-123

P/GM/10


MODELING THE ACID-BASE SURFACECHEMISTRY OF MONTMORILLONITE

I.C. Bourg1,2,3,4, G. Sposito2,4, A.C.M. Bourg3

1. Agence Nationale pour la Gestion des Déchets Radioactifs (ANDRA), Châtenay-Malabry,France

2. Department of Civil and Environmental Engineering, University of California, Berkeley, CA,USA ([email protected], [email protected])

3. Environmental HydroGeochemistry (LHGE), Université de Pau et des Pays de l’Adour, Pau,France ([email protected])

4. Geochemistry Department, Earth Sciences Division, Lawrence Berkeley NationalLaboratory, Berkeley, CA, USA

Proton uptake on montmorillonite edge surfaces can control pore water pH, solute adsorption, dissolutionkinetics and clay colloid behavior in engineered clay barriers and natural weathering environments.Knowledge of proton uptake reactions, however, is currently limited by strong discrepancies betweenreported montmorillonite titration data sets and by conflicting estimates of edge structure, reactivity andelectrostatics. In the present study, we show that the apparent discrepancy between titration data sets resultsin large part from the widespread use of an erroneous assumption of zero specific net proton surface chargeat the onset of titration. Using a novel simulation scheme involving a surface chemistry model to simulateboth pretreatment and titration, we find that montmorillonite edge surface chemistry models that accountfor the “spillover” of electrostatic potential from basal onto edge surfaces and for the stabilization ofdeprotonated Al-Si bridging sites through bond-length relaxation at the edge surface can reproduce keyfeatures of the best available experimental titration data (the influence of pretreatment conditions onexperimental results, the absence of a point of zero salt effect, buffer capacity in the acidic pH range).However, no combination of current models of edge surface structure, reactivity and electrostatics canquantitatively predict, without fitted parameters, the experimental titration data over the entire range of pH(4.5 to 9) and ionic strength (0.001 to 0.5 mol dm-3) covered by available data.

References:Baeyens, B., Bradbury, M.H., 1997. A mechanistic description of Ni and Zn sorption on Na-montmorillonite. Part I: Titration and sorption measurements. J. Contam. Hydrol. 27, 199-222.

Bourg I.C. (2004) Caractérisation du comportement d’une bentonite sodique pour l’isolement des déchets.Transport diffusif des traceurs ioniques (Na+, Sr2+, Cs+ et Cl-) dans la bentonite sodique compactée

Figure 1: Montmorillonite acid-base titration data of Baeyens and Bradbury (1997) and Duc et al. (2005)simulated with a model that accounts for the influence of montmorillonite pretreatment on the net protonsurface charge at the onset of titration.

-0.0006

-0.0004

-0.0002

0

0.0002

0.0004

0.0006

4 5 6 7 8 9 10

pH

[acid]-[base]

(mol dm-3) I = 0.5 M (Baeyens and Bradbury, 1997)

I = 0.1 M (Baeyens and Bradbury, 1997)

I = 0.5 M (simulation)

I = 0.1 M (simulation)

-0.0003

-0.0002

-0.0001

0.0000

0.0001

0.0002

4 5 6 7 8 9 10

pH

[acid]-[base]

(mol dm-3) I = 0.1 M (Duc et al., 2005)I = 0.01 M (Duc et al., 2005)I = 0.001 M (Duc et al., 2005)I = 0.1 M (simulation)I = 0.01 M (simulation)I = 0.001 M (simulation)

P/GM/10



saturée, et titration acide-base de la montmorillonite, Doctoral Thesis, Université de Pau et des Pays del’Adour, France.

Bourg I.C., Sposito G. & Bourg A.C.M. (2007) Modeling the acid-base surface chemistry ofmontmorillonite. J. Colloid Interface Science, submitted.

Duc, M., Gaboriaud, F., Thomas, F. (2005). Sensitivity of the acid-base properties of clays to the methodsof preparation and measurement. 2. Evidence from continuous potentiometric titrations. J. ColloidInterface Sci. 289, 148-156.

P/GM/11


A TIME RESOLVED LASER FLUORESCENCEAND X-RAY ABSORPTION SPECTROSCOPY

STUDY OF LANTHANIDE/ACTINIDESORPTION ON CLAY MINERALS:

INFLUENCE OF CARBONATECOMPLEXATION

M. Marques Fernandes1, Th. Rabung2, R. Dähn1, B. Baeyens1, M. H. Bradbury1

1. Laboratory for Waste Management, Paul Scherrer Institut, CH-5232, Villigen, Switzerland2. Institut für Nukleare Entsorgung, Forschungszentrum Karlsruhe, Postfach 3640, D-76021

Karlsruhe, Germany

INTRODUCTIONThe safety case for nuclear waste repository is, to a large degree, based on the physical and chemicalretention of radionuclides [1]. An important part of radionuclide retardation is the sorption on weatheredor secondary phase material of the waste and engineered structures around it, and natural minerals insediments along potential transport paths towards the biosphere. Retention is especially effective in clayminerals with their large surface and their high sorption capacities. Identifying and quantifying theradionuclide uptake processes occurring at clay/solution interface over a representative range of relevantconditions is indispensable for performance assessment.

In natural environments, the predominant aqueous phase reactions of trivalent actinides and lanthanides arehydrolysis and complexation with dissolved carbonates. The formation of strong carbonate complexes insolution can potentially lead to a decrease in metal ion sorption. For carbonate free systems, Eu(III)/Am(III) sorption edges and isotherms on montmorillonite are available [2, 3]. The sorption of Eu(III) andCm(III) on montmorillonite, illite and kaolinite, has been previously investigated by Time Resolved LaserFluorescence Spectroscopy (TRLFS) and X-ray absorption spectroscopy XAS at varying pH values in theabsence of CO2 [4-6]. Recent TRLFS study indicates the possibility that ternary carbonato surfacecomplexes are formed for Eu(III) on smectite and kaolinite under air equilibrated conditions [7]. Sorptionmodels neglecting the influence of carbonate may incorrectly predict radionuclide sorption in natural claysystems. The aim of this work is to improve the mechanistic understanding at the molecular level of thesorption of Eu(III)/Am(III)/ Cm(III) onto clay minerals by investigating the complexation of carbonate onsorption. In order to identify radionuclide surface species TRLFS and XAS will be applied.

BACKGROUND: MACROSCOPIC SORPTION RESULTSMacroscopic sorption edges for Eu(III) and Am(III) have been previously determined by means of classicbatch experiments. Purified and well characterized clay minerals were taken as sorbates. The influence ofcarbonate concentrations in equilibrium with the atmosphere (pCO2 = 10

-3.5 bar) on the sorption of Eu(III)/Am(III) in suspensions of Na-montmorillonite and Na-illite was investigated in 0.1M NaClO4 as a functionof pH at trace metal concentrations. The total carbonate concentration was fixed as a function of pH usingNaHCO3/Na2CO3 in buffered solutions from pH 6 up to a pH of ~10.3. Figs. 1 and 2 show the uptake ofEu as a function of pH in the absence and presence of carbonate (pCO2 = 10

-3.5 bar) on Na-montmorilloniteand Na-illite respectively. For the experimentally determined sorption edge of Eu(III) on Na-montmorillonite , no influence of carbonate was observed up to pH ~ 8 and the log Rd values were similarto those measured in carbonate-free montmorillonite suspensions. Above pH > 8, a strong decrease in thesorption compared to montmorillonite suspensions without carbonate was observed and the log Rd reducedto 2.0 ± 0.4 (L·kg-1) at pH ~10.3. The experimentally determined sorption edge of Eu(III) on Na-illite,

P/GM/11



however, show no influence of carbonate over the whole pH range; log Rd values were similar to thoseobserved in illite suspensions without carbonate except at the highest pH value investigated. At pH ~ 10.3a decrease in sorption was observed (total dissolved inorganic carbon concentration ~ 0.3 M).

MICROSCOPIC APPROACHIn order to achieve a molecular understanding of the radionuclide uptake mechanisms and the structure ofthe main surface complexes formed on Na-montmorillonite and Na-illite and to elucidate the differentbehavior of both clay minerals, XAS measurements will be performed on Am(III) loaded samples preparedin the absence and presence of carbonate. The overall aim is to identify the surface sorbed species in theactinide-carbonate-clay system. Extended X-ray absorption spectroscopy (XAS) yield information on thecoordination structure (coordination number, bond distance) of Am(III) sorbed onto clays. Theseexperiments require higher concentrations of actinide than was used in the sorption edge measurements. Inorder to ensure that the information obtained by XAS is not unduly influenced by the higher metal ionconcentrations, this study is combined with the application of TRLFS. For the TRLFS study Cm(III) isused because it has a much higher sensitivity for TRLFS measurements than Eu(III) or Am(III) and it isa good chemical analogue for Eu(III) and Am(III). Both the position and intensity of the fluorescenceemission peaks, normally measured as a function of pH, allow the identification of different species. Inaddition, the fluorescence lifetime provides sensitive information on the coordination environment,especially the presence of H2O/OH

- quenchers in the inner coordination sphere. This information, amongstothers, allows a distinction to be made between inner- or outer-sphere complexation. The informationprovided by the application of these techniques will show whether or not ternary complexes form on theclay surface. The spectroscopic results obtained will be presented in a poster at the conference.

References:[1] Nagra, Technical Report NTB 02-05. Project Opalinus Clay: Safety Report. Demonstration of disposalfeasibility (Entsorgungsnachweis) for spent fuel, vitrified high-level waste and long-lived intermediate-level waste. 2002, Nagra: Wettingen, Switzerland.

[2] Bradbury, M.H. and B. Baeyens, Modelling the sorption of Mn(II), Co(II), Ni(II), Zn(II), Cd(II),Eu(III), Am(III), Sn(IV), Th(IV), Np(V) and U(VI) on montmorillonite: Linear free energyrelationships and estimates of surface binding constants for some selected heavy metals and actinides(vol 69, pg 875, 2005). Geochimica Et Cosmochimica Acta, 2005. 69(22): p. 5391-5392.

[3] Bradbury, M.H., et al., Sorption of Eu(III)/Cm(III) on Ca-montmorillonite and Na-illite. Part 2: Surfacecomplexation modelling. Geochimica Et Cosmochimica Acta, 2005. 69(23): p. 5403-5412.

[4] Stumpf, T., et al., Time-resolved laser fluorescence spectroscopy study of the sorption of Cm(III) ontosmectite and kaolinite. Environmental Science & Technology, 2001. 35(18): p. 3691-3694.

[5] Rabung, T., et al., Sorption of Eu(III)/Cm(III) on Ca-montmorillonite and Na-illite. Part 1: Batchsorption and time-resolved laser fluorescence spectroscopy experiments. Geochimica Et CosmochimicaActa, 2005. 69(23): p. 5393-5402.

[6] Stumpf, T., et al., An EXAFS and TRLFS study of the sorption of trivalent actinides onto smectite andkaolinite. Radiochimica Acta, 2004. 92(3): p. 133-138.

[7] Stumpf, T., et al., Inner-sphere, outer-sphere and ternary surface complexes: a TRLFS study of thesorption process of Eu(III) onto smectite and kaolinite. Radiochimica Acta, 2002. 90(6): p. 345-349.

P/GM/12


STRUCTURE ELUCIDATIONAND OCCURRENCE OF TC(IV)PYROGALLOL COMPLEXES

Breynaert, E., and Maes, A.

Laboratory for Colloid Chemistry, Katholieke Universiteit Leuven, Kasteelpark Arenberg, 23,B-3001 Leuven, BELGIUM, ([email protected])

The redox-sensitive fission product technetium-99 (Tc) is of great interest in nuclear waste disposal studiesbecause of its potential for contaminating the geosphere due to its very long half-life (2.13×105 year) andhigh mobility under oxidising conditions, where technetium forms pertechnetate (TcO4

-). Under suitablereducing conditions, e.g. in the presence of an iron(II) containing solid phase which can act as anelectrondonor, the solubility can be limited by the reduction of pertechnetate followed by the formation ofa surface precipitate [1]. However, by association with mobile humic substances (HS) or other associating/complexing species, the solubility of reduced Tc species may be drastically enhanced [2]. The elucidationof the identity and geometrical structure of the species causing this enhanced solubility often remains adifficult issue.

EXAFS/XANES1 analysis can be very helpful to determine the molecular surroundings of reduced Tcorganic complexes or Tc(IV) colloid-HS-associations. The interpretation of the EXAFS spectra is howevernot a straightforward process due to the strong influence of multiple scattering paths on the spectra..Contrary to single scattering analysis which can be easily performed on the experimental EXAFS spectra,multiple scattering analysis can only be done, based on good approximations of the possible geometricalstructures of the species under consideration. As the structure of humic substances is not exactly knownand would be too complex for modelling, pyrogallol was used as a model compound for phenolicfunctional groups in natural organic matter.

Detailed multiple scattering analysis of EXAFS data of Tc(IV)/pyrogallol solutions based on DFT2

modelled reference structures revealed the existence of a stable Tc-pyrogallol complex which is readilyformed at pH 11 upon reduction of pertechnetate by hydrazine or dithionite in presence of pyrogallol. Thisinitial complex serves as precursor for a pH dependent series of pyrogallol complexes exhibiting areasonable stability towards technetium colloid or precipitate formation when lowering the pH from 11 to2 after synthesis.

The occurrence of stable, readily formed Tc-complexes with a humic substance model compoundpotentially changes current knowledge about Tc(IV) behaviour in natural systems containing dissolvedorganic matter. This observation is an indication for an extra competition between the formation ofimmobile Tc(IV) precipitates upon reduction of TcO4

- or potentially mobile organic Tc(IV) complexes andpreviously discovered Tc(IV) eigencolloids stabilised by dissolved HS through colloid-colloid interactions[3, 4]. The competition between pyrogallol and natural dissolved organic matter was tested using sizeexclusion chromatography for separation of the different colloidal species and complexes.

The authors acknowledge a grant from KULeuven University and financial support from the KULeuvenGeconcerteerde Onderzoeksacties (GOA2000/007).

1 Extended X-Ray Absorption Fluorescence Spectroscopy/ X-Ray Absorption Near Edge Spectroscopy2 Density Functional Theory

P/GM/12



References[1] Cui, D.Q. and T.E. Eriksen, Reduction of pertechnetate in solution by heterogeneous electron transfer

from Fe(II)-containing geological material. Environmental Science & Technology, 1996. 30(7): p.2263-2269.

[2] Maes, A., et al., Quantification of the interaction of Tc with dissolved boom clay humic substances.Environmental Science & Technology, 2003. 37(4): p. 747-753.

[3] Maes, A., et al., Evidence for the Formation of Technetium Colloids in Humic Substances by X-RayAbsorption Spectroscopy. Environmental Science & Technology, 2004. 38(7): p. 2044-2051.

[4] Geraedts, K., et al., Evidence for the existence of Tc(IV) - humic substance species by X-ray absorptionnear-edge spectroscopy. Radiochimica Acta, 2002. 90(12): p. 879-884.

P/GM/13


GEOCHEMISTRY OF SE(0) UNDER BOOMCLAY CONDITIONSC. Bruggeman1, †, A. Maes1, J. Vancluysen1

1. Laboratory for Colloid Chemistry, Katholieke Universiteit Leuven, Kasteelpark Arenberg 23,B-3001 Leuven, Belgium ([email protected])

†. Current affiliation: SCK•CEN, R&D Waste & Disposal, Boeretang 200, B-2400 Mol, Belgium([email protected])

INTRODUCTIONMetal selenium or Se(0) is an important Se sink which is predicted to be thermodynamically stable undermildly reducing conditions. Because of the limited availability of Fe2+ and kinetic implications whichimpede the formation of metal-selenide precipitates, Se(0) is predicted to be the dominant solubility-controlling phase for Se to be accounted for in Performance Assessment calculations [1]. However,thermodynamic calculations showed that this solubility is very dependent on redox potential and pH [1]Therefore it is of utmost importance to study the geochemistry of Se(0) under repository (in this case,Boom clay) conditions.

EXPERIMENTAL CONCEPTAmorphous Se(0) precipitates spiked with 75Se were prepared by a reduction procedure in dialysis bagsstarting from Se(IV) and using ascorbic acid. The solubility of these precipitates was measured in asynthetic Boom clay water background electrolyte (SBCW) in presence and absence of dissolved humicsubstances (HS). To evaluate the effect of solid phases, Se(0) precipitates were also contacted for up to sixmonths with Boom clay suspensions at various solid-to-liquid ratios. In all these systems the solutionspeciation was checked with a combination of ultrafiltration (30 kDa MWCO), ion chromatography (todistinguish Se oxyanions) and gel permeation chromatography (to measure small Se(0) colloids).

RESULTS AND INTERPRETATIONThe total Se solution concentration in SBCW in presence of Se(0) precipitate ranged from 8.4×10-8 mol·l-1to 1.6×10-7 mol·l-1 and was constant from 1 hour up to 5 days equilibration time (pH 8.56±0.02, Eh 100±5 mV vs. SHE). After ultrafiltration at 30 kDa MWCO, the total Se solution concentration decreased in allsamples to 3.5(±0.6)×10-8 mol·l-1. Analysis of the Se speciation in solution showed that 50-80% of the totalSe solution concentration after ultrafiltration was consisting of SeO3

2- ions, while the remainder consistedof colloidal Se(0).

Upon contacting 75Se(0) with dissolved Boom Clay HS, the total Se solution concentration increased withincreasing dissolved HS concentration and values ranging between 3.5×10-8 mol·l-1 (0 mg·l-1 HS) and1.9×10-7 mol·l-1 (175 mg·l-1 HS) were reached already after one day equilibration time (figure 1A). Withtime, an increase of the total Se solution concentration for each sample was observed and after 1 monthtotal Se solution concentrations were as high as 8.9×10-7 mol·l-1 (175 mg·l-1 HS) (figure 1B). Analysis ofthe Se solution speciation showed that the increase in total Se concentration with increasing HSconcentration was predominantly attributable to an increase of the colloidal Se fraction. SeO3

2-

concentrations were in the range of those observed in blank SBCW. After ultrafiltration at 30 kDa MWCO,Se concentrations in the permeate were in the same range for all samples irrespective of the HSconcentration and amounted to 8.4(±1.1)×10-8 mol·l-1.

When 75Se(0) was added to Boom Clay suspensions (with SBCW) of different solid-to-liquid ratio (rangingfrom 0 to 240 g·l-1, which resulted in dissolved HS concentrations ranging from 0 to 237 mg·l-1), the totalSe solution concentrations showed a tendency to slightly increase over time but more variation was notedcompared to systems in which Se(0) was contacted with dissolved Boom Clay HS only. Also, the increase

P/GM/13



proceeded less rapidly indicating the existence of interaction processes with the Boom Clay solid phase.After 6 months, the total Se solution concentration after 0.20 µm filtration ranged from 1.8×10-7 mol·l-1 upto 4.8×10-7 mol·l-1. Similarly to the systems containing dissolved HS only, the Se solution concentrationincreased with increasing dissolved HS concentration, but now both after 0.20 µm filtration and after 30kDa ultrafiltration. Therefore, the influence of dissolved HSon the operational solubility of Se(0) solidphases under repository conditions can be potentially quite significant.

ACKNOWLEDGEMENTWe gratefully acknowledge financial support from NIRAS-ONDRAF and IP FUNMIG

References[1] J. van der Lee and L. Wang, Topical report on: Speciation and solubility calculations for uranium,plutonium and selenium under Boom clay conditions, R-3400, SCK•CEN, Mol, Belgium.

Figure 1: Total Se solution concentration (mol·l-1) as a function of dissolved Boom Clay organic matterconcentration (mg·l-1) after 24 hours equilibration time (figure A - left), and as a function of equilibrationtime (h) (figure B - right), in a solubility-type experiment starting from Se(0). Red data points areconsidered to be outliers.

0 40 80 120 160

4,0x10-8

8,0x10-8

1,2x10-7

1,6x10-7

2,0x10-7

A

dissolved organic matter concentration (mg.l–1)

totalSesolutionconcentration(mol.l–1)

0 1 2 3 4 5 6

0

1

2

3

4

FT Magnitude

R +Δ (≈)

Illite pH3

Illite pH4

Illite pH6

HSeO-

3

B

P/GM/14


EXPERIMENTAL AND MODELLING STUDYOF PURE SECONDARY SILICATE MINERALS

REACTIVITY IN GEOLOGICAL CO2SEQUESTRATION CONDITIONS

A. Credoz1, O. Bildstein1, M. Jullien1, J. Raynal1, J.-C. Pétronin1, L. Trotignon1, O. Pokrovsky2

1. CEA Cadarache, DTN/SMTM/LMTE, 13108 Saint Paul Lez Durance ([email protected])2. LMTG – UMR 5563 CNRS – UPS – IRD, Observatoire Midi-Pyrénées, 14, avenue Edouard

Belin - 31400 Toulouse ([email protected])

INTRODUCTIONFor one decade, international authorities and scientists have agreed with the irreversible effects ofanthropogenic emissions of carbon dioxide on global climatic changes. Among others possibilities, CO2capture and storage (CCS) in deep aquifers and depleted oilfields appears to be one of the main solutionsto reduce greenhouse gases release to the atmosphere.

The French scientific group involve in the “ANR-Géocarbone” project is looking at the feasibility, theperformance, and the safety of such a solution. They focused the efforts on potential injection sites in theParis basin (PB) and defined a pilot site in Saint Martin de Bossenay (SMB, in the South-East of BP) forthe CO2 injection, storing and monitoring. In the deep geological conditions of the sequestration (ca. 1500m depth, 80°C, and 150 bar), CO2 is in a supercritical state (CO2-SC). After injection into the Doggerreservoir, the CO2-SC bubble will rise due to gravitational forces, reach the Callovo-Oxfordian (COx) cap-rock layer, and trigger geochemical interactions with the minerals. A specific project called “Géocarbone-Integrity” is looking specifically at the integrity of argillaceous cap-rocks. An important task of this projectis to perform experimental studies and numerical modelling to evaluate the geochemical reactivity andsafety of the COx cap-rock at SMB with CO2-rich fluids.

In this study, we looked at the reactivity of pure minerals found in the mineralogy of the COx at SMB.This simplifying approach of the reactive rock composition will enable us to better understand thegeochemical mechanisms involved in the CO2-water-rock interactions in deep-geological sequestrationconditions (Regnault et al. 2005). Since a lot of work has already been dedicated to carbonate minerals,we focused on the main silicate minerals, I/S illite/smectite and illite, as minerals of interest for theassessment of the cap-rock integrity.

MATERIAL AND METHODThe target silicate mineral is the interstratified illite/smectite mineral (I/S). Purified reference I/S sampleswere used in the experiments and chosen because of their characteristics which are close to those of the I/S found in SMB-COx (where they represent 25% of the total composition). Pure illite, representative ofthe SMB-COx (2% of total composition) and pure smectite are also potential candidates for experimentssince they may facilitate the understanding of the mechanisms of alteration of the I/S compound.

After selection, the minerals were crushed and reduced to < 500µm fraction to maximize the reactivespecific surface. Each mineral sample was introduced in a titanium autoclave where fluids (CO2 and/orbrine) are maintained at constant temperature and pressure. The experimental solution was synthesizedfrom the composition of the Dogger brine given by Azaroual et al. (1997).

For each mineral, four sets of experiments were carried out: pure mineral with brine, pure mineral withbrine acidified with CO2(g), pure mineral with CO2-SC, and pure mineral with brine and CO2-SC.

P/GM/14



Experiments with CO2-SC were performed at 80/150°C and 150 bars during 30/90 days. Experiments withdissolved CO2 were performed at 80/150°C and maintained at atmospheric pressure.

Before and after reaction, the minerals were analysed by X-ray diffraction (XRD) and scanning electronmicroscopy (SEM) coupled to an energy-dispersion spectrometer (EDS). For each tests, three replicates ofbrine were analysed before and after reaction by inducted coupled Plasma – Atomic Emission Spectrometer(ICP-AES).

A COMBINED EXPERIMENTAL AND MODELLING APPROACHTo predict and compare the geochemical reaction pathways, including the parameters for the reactionkinetics of pure minerals with CO2(aq)/CO2(g)/CO2-SC/Na-Cl solutions, reactive transport modelling wasperformed with the Chess code (Van der Lee, 2005) and the Crunch code (Steefel, 2001). Two scenarioswere simulated representing both the conditions at the SMB pilot injection site and experimental conditionspresented above. The first one involves CO2-SC interactions with pure minerals equilibrated with a typicalDogger brine. The second one involves CO2-rich brine interactions with pure minerals.

This modelling study focused on the behaviour of minerals at the timescale of the experiments (30 to 90days) and calculations were also performed for duration up to 10,000 years which is the relevant timeframefor the industrial application. Kinetic of mineral precipitation and dissolution were included in the modelto understand geochemical pathways as the thermodynamic equilibrium hypothesis was found to beinadequate (Gaus et al. 2005).

Data from the experimental and modelling approach on pure minerals will be compared and integrated intothe SMB caprock performance and safety assessment for long-term CO2 sequestration.

References:Azaroual M, Fouillac C., Matray J.M., 1997. Solubility of silica polymorphs in electrolyte solutions, II.Activity of aqueous silica and solid silica polymorphs in deep solutions from the sedimentary ParisBasin. Chemical Geology, 140, 3-4, pp. 67-179.

Gaus I., AzaroualM., Czernichowski-Lauriol I., 2005. Reactive transport modelling of CO2 injection on theclayey cap rock at Sleipner (North Sea). Chemical Geology 217, pp. 319-337.

Regnault O., Lagneau V., Catalette H., Schneider H., 2005. Etude expérimentale de la réactivité du CO2supercritique vis-à-vis de phases minérales pures. Implications pour la séquestration géologique duCO2. C. R. Geoscience 337, pp. 1331-1339.

Steefel, C.I., 2001. CRUNCH, Lawrence Livermore National Laboratory, pp. 76.

Van der Lee J., 1998. Thermodynamic and mathematical concepts of CHESS. Technical Report LHM/RD/98/39, CIG, Ecole des Mines de Paris, Fontainebleau, France, 99pp.

CO2-SC injectionPure mineral + brine

Brine + CO2(aq) injection

Pure mineral

P/GM/15


EXPERIMENTAL EVALUATIONOF A RETENTION MODEL

FOR MAJOR GROUNDWATER ELEMENTSON THE TOURNEMIRE ARGILLITE

P. Jacquier1, C. Beaucaire1, A.L. Vuillaume2, Ch. Wittebroodt2, J. Ly1,J. Page1, S. Savoye2, H. Pitsch2

1. CEA – Commissariat à l’Energie Atomique, Saclay, DEN/DANS/DPC/SECR/L3MR, Bât. 450,91191 Gif-sur-Yvette, France([email protected], [email protected], [email protected], [email protected])

2. IRSN - Institut de Radioprotection et de Sûreté Nucléaire, BP n˚17, 92262 Fontenay-aux-roses,France ([email protected], [email protected], [email protected])

INTRODUCTIONThis work deals with the behaviour of Tournemire argillite towards major elements. Rock samples comefrom the IRSN Tournemire experimental facility, located in Aveyron, Southern France. Their “intrinsic”ion-exchange properties, i.e. the concentrations of the different types of adsorption sites and the associatedselectivity coefficients for H, Na, K, Ca and Mg were previously determined (Jacquier et al., 2004). Thisset of data constitutes a retention model which enables to describe the behaviour of major cations and H+

during water-rock interaction. The objective of the present study is to estimate the robustness of this modelby comparing experimental data derived from several types of leaching experiments carried out onTournemire samples with the corresponding calculated data.

EXPERIMENTAL CONCEPTThe occurrence of pyrite in the argillite led us to adopt two types of experimental protocols, since pyriteoxidation may drastically change the chemistry of leachates by producing a large amount of sulphate anddecreasing pH. So, the first approach consisted in carrying out leaching experiments in a glove box withvery low oxygen content (∼1 ppm of O2) in order to limit oxidation disturbances as much as possible. CO2partial pressure was also fixed at a value close to 10-2.4 atm which is assumed to meet the in situ condition(Beaucaire et al., 2007).

Conversely, in the second approach, the O2 content of the glove box atmosphere was fixed close to 1000ppm in order to estimate the effect of an oxidising disturbance on the leachate chemistry. In this case, CO2was not injected into the glove box, allowing pCO2 to evolve freely.

In both protocols, four solid/liquid ratios, i.e. 0.04, 0.25, 0.6 and 1 g/g were investigated in order to covera large range of chemical conditions. The concentrations of major dissolved cations (Na+, K+, Ca2+, Mg2+)and anions (Cl-, SO4

2-, HCO3-/CO3

2-) were monitored at several times’ steps, i.e. 24 hours, 7 and 21 days.The adsorbed cationic populations on the argillite samples were also determined by cation displacementwith Ni-ethylenediamine.

RESULTS AND MODELLINGAmong the elements monitored during these experiments, different behaviours may be distinguished:

– Cl behaves as a mobile element, meaning that it is not controlled by an equilibrium with secondaryminerals. In such leaching experiments, mobile elements diffuse out of the matrix porosity and theirconcentrations in the external fluid increase in proportion to the solid/liquid ratio. The case of SO4 is

more complex because, in sedimentary rocks, at least three different reservoirs exist: matrix porosity,sulphide- and sulphate-bearing minerals. In the present study, the SO4 behaviour cannot be accounted

for by considering a single source;

P/GM/15



– Ca andMg are quickly constrained by carbonates, regardless of the liquid/solid ratio;

– other elements such as Na and K are released in solution following the ion-exchange equilibria withsurface minerals.

Finally, it is possible to compare the measured concentrations of major elements in the leachates withtheoretical concentrations derived from the ion exchange properties of the rock and the dissolutionprecipitation equilibria of selected mineral phases like carbonates. Calculated concentrations are generallyin good agreement with measured ones, whatever the amount of oxygen and even when the partial pressureof CO2 is not fixed (See for example figure 1). So, we can consider that the retention model built up forthe Tournemire argillite is sufficiently robust to describe the chemical reactivity of the argillite. Therefore,it is possible to calculate the chemical composition of the pore-water present in the Tournemire argillite in“equilibrium” conditions, as well as the composition of fluids resulting, for example, from an oxidizingdisturbance.

ReferencesJacquier P., Ly J. and Beaucaire C. (2004): The ion-exchange properties of the Tournemire argillite I -Study of the H, Na, K, Cs, Ca & Mg behaviour. Applied Clay Science 26, 163-170.

Beaucaire C., Michelot J.-L., Savoye S. and Cabrera J. (2007): Groundwater characterization andmodelling of water-rock equilibria in the argillaceous formation of Tournemire (Aveyron, France).Submitted to Applied Geochemistry.

Figure 1: Comparison of calculated concentrations with measured ones in leaching experiment, with O2 ~1000 ppm and pCO2 not fixed, for two solid/liquid ratios (1.0 and 1.5 g/ml).

-5

0

5

10

15

20

pH

HCO3

pCO2 C

LSO4

Na K C

aMg

C(mM);LogpCO2(atm).

Cal. S/L 1.0 Ox

Calc. S/L 1.5 Ox

Meas. S/L 1.0 Ox

Meas. S/L 1.5 Ox

P/GM/16


MODELLING THE LONG TERMINTRACTION OF CEMENTITIOUS PORE

WATER WITH BOOM CLAY

D. Jacques, L. Wang

SCK•CEN, Boeretang 200, B-2400 Mol, Belgium (corresponding author: [email protected])

INTRODUCTION

In Belgium, Boom Clay is studied as a potential host formation for geological disposal of radioactivewaste. The current reference design of the engineered barrier system (‘supercontainer design’) plans to usea considerable amount of cementitious materials as construction material, buffer and backfill. Interactionsbetween the alkaline pore fluids from the concrete engineered barriers with the Boom Clay may changethe retention properties of the Boom Clay in the vicinity of the interface with the engineered barriers. Theobjective of this work is to assess the extent of an alkaline disturbed zone in the Boom Clay for a periodup to 105 years using reactive-diffusion model simulations. The sensitivity of different model parametersand of (major) model assumptions regarding the Boom Clay mineralogy and the choice of the secondaryphases is assessed.

MODEL APPROACH

The diffusion of an alkaline plume is simulated for a period up to 105 years with the PHREEQC-2.12geochemical code (Parkhurst and Appelo, 1999) using the llnl.dat database in that code. To account forthe mass balance and variations in the composition of the alkaline plume with time, the concrete materialin the near field is explicitly included in the model. The concrete material is located between 0.22 m and1.62 m from the centre of a radial simulation geometry. The steel components in the supercontainer designare ignored, and an initial homogeneous concrete material was assumed.

A reference model is defined including five minerals for the Boom Clay (quartz, kaolinite, illite, Na-montmorillonite, and calcite), a pH-independent (clay minerals) and dependent (organic matter) cationexchange complex and surface acidity reactions on the illite and Na-montmorillonite (after Bradbury et al.,2005). All mineralogical reactions are assumed to be in equilibrium. In addition to the reference case, othersimulations assess the sensitivity of the capacities of the exchange and surface sites, the initial amount ofprimary minerals (plus and minus 25%) and the diffusion coefficient in the Boom Clay (five times smallerand larger). In addition, some alternative model formulations for the Boom Clay mineralogy and the choiceof the secondary phases are tested.

In all these simulations, there is no feedback of mineral precipitation and dissolution to changes in porosityand, subsequent, diffusion. For example, clogging of the pore space in the concrete due to calciteprecipitation will certainly have an effect on the diffusion in and out the concrete. Nevertheless, the currentsimulations will give an idea of the possible extent of the alkaline plume perturbation in the Boom Clay.


Figure 1 shows the pH evolution in the reference model.Within the first 1000 years, the Na- and K-oxidesin the concrete are depleted resulting in a pH decrease from 13.5 to 12.5. Portlandite is completely depletedup to 0.3 and 1.0 m from the concrete – Boom Clay interface after 25000 and 105 years, respectively (pHis smaller than 12.5). At the concrete – Boom Clay interface, a substantial amount of calcite is precipitated(data not shown). Overall, the disturbance of the Boom Clay by the diffusion of an alkaline plume islimited up to 0.4 – 1.2 m from the concrete – Boom Clay boundary after 25000 years (for different modelformulations) and up to 2.0 – 2.5 m after 105 years (for the reference case). After 25000 years, the volume

P/GM/16



occupied by minerals is increased by a maximum of 15 percent for the reference case (taken into accountparameter uncertainty) and of 25 percent for the different model formulations.

Parameters defining the capacities of the cation exchange complexes or the surface acidity sites have noeffect on the extent of changes in pH. The initial amount of the primary minerals has a slightly largereffect, especially kaolinite. The most crucial parameter is the diffusion coefficient. Although this parameteris relatively well defined for non-disturbed Boom Clay, it is still an uncertain parameter with respect tothe altered Boom Clay by an alkaline plume. The pore diffusion coefficient depends on the porosity of theporous medium and the alkaline plume may change the porosity. However, these effects were not takeninto account in the present simulations.

The selection of mineral reactions and sequence is a crucial factor for assessing the spatial extent of thealkaline plume perturbation. Decreasing (e.g., neglecting exchange and surface acidity reactions orallowing primary minerals only to dissolve) or increasing (adding dolomite-dis as a primary mineral) thetotal buffer capacity in the model increase or decrease the spatial extent of the disturbed zone, respectively.An alkaline plume disturbed zone is most limited if the partial pressure of CO2 in the Boom Clay isbuffered by an assemblage of minerals as proposed by De Craen et al. (2004). In that case, the disturbedzone is limited to 0.4 m after 25000 y.

References

Bradbury, M.H., B. Baeyens, H. Geckens, and Th. Rabung, 2005. Sorption of Eu(III)/Cm(III) on Ca-montmorillonite and Na-illite. Part 2: Surface complexation modelling. Geochim. Cosmochim. Acta,69:5403-5412.

De Craen, M., L.Wang, M. Van Geet, and H. Moors, 2004. Geochemistry of Boom Clay pore water at theMol site. SCK•CEN-BLG-990, 179 p.

Parkhurst, D.L., and C.A.J. Appelo, 1999. User's guide to PHREEQC (version 2) – A computer programfor speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations. Water-Resources Investigation Report 99-4259, Denver, Colorado, 312 p.

ACKNOWLEDGEMENTS

This work is undertaken in close co-operation with, and with the financial support of NIRAS-ONDRAF,the Belgian Agency for Radioactive Waste and Enriched Fissile materials. The critical follow-up by andfruitful discussions with Mrs. A. Dierckx , Mr. M. Van Geet and Mr. R. Gens are very much appreciated.

Figure 1: pH evolution near the concrete – Boom Clay boundary for the reference model (vertical lineindicates the concrete – Boom Clay interface).

-1 0 1 2 3

8

10

12

14

pH

initial

250 y

500 y

1000 y

5000 y

10000 y

25000 y

Distance from concrete - Boom Clayboundary (m)

-1 0 1 2 3

8

10

12

14

pH

25000 y

50000 y

75000 y

100000 y

Distance from concrete - Boom Clay boundary (m)

P/GM/17


SORPTION-DESORPTIONOF RADIONUCLIDES AND ANALOGUESIN CLAYS SUITABLE FOR BARRIERS

E. Galunin,1 P. Chaín,2 M.D. Alba,2 M. Vidal1

1. Department of Analytical Chemistry - Universitat de Barcelona. Martí i Franqués 1-11, 08028,Barcelona, Spain ([email protected])

2. Instituto Ciencia de los Materiales de Sevilla. CSIC-Universidad de Sevilla. Avda. AmericoVespucio, 49. 41092 Sevilla, Spain

INTRODUCTIONThe chemical and physical nature of the clay is a key issue in the design of engineered barriers. The claygeometry, total charge and isomorphical substitutions, among other factors, have a significant effect on thesorption of pollutants on the clay phase. Here a set of clays were examined as potential candidates for thebuild up of engineered barriers. The performance of the clays were tested with respect to the sorption ofa set of lanthanides (La and Lu), which are usually considered as natural analogues of actinides (Chapmanand Smellie, 1986) and other radionuclides present in high level radioactive waste, and of radionuclides(95Zr and 241Am) that can be present in the radioactive waste. After estimating the sorption pattern of theelements in the clays, the reversibility of the sorption step was also evaluated.

EXPERIMENTAL CONCEPTClay samples. Five clays were tested in this study, including a reference sample of silica (see Table 1).Clays differed in geometry and in isomorphic substitution.

Sorption tests. Sorption isotherms were obtained for La3+ and Lu3+ by the quantification of the distributioncoefficients (Kd) obtained in the range of initial concentrations of 10

-2, 10-3, 10-4, and 10-5 mol l-1, while for95Zr and 241Am single measurements of the Kd were obtained, with activity concentrations high enough toensure the quality of the measurements. For all cases, sorption tests were carried out in H2O and in Ca(OH)20.01 mol l-1 at pH = 7.

The sorption tests were performed equilibrating 30 ml of the respective element solution with 0.2 g of clay.The contact time was 24 hours, and then the supernatants were decanted off after centrifugation (25 min;10000 rpm; 10 oC). The final pH was monitored in all cases. The quantification of the concentration ofevery element in the final solution allowed us to quantify the sorption Kd.

Desorption tests. These tests were applied to determine the desorption Kd and the percentage of elementreversibly sorbed. The desorption tests were performed with the clay residue from the sorption step, underthe same experimental conditions, but without the element.

Table 1: Clay samples used in the experiments

Clay Formula GeometryIsomorphicsubstitutions

Silica (SIL) SiO2TexasMontmorillonite(STx-1)

(Ca0.27Na0.04K0.01){Si8}{Al2.41Mg0.71Fe0.09Ti0.03}O20(OH)4 Dioctahedral Octahedral sheet

MontmorilloniteOtay(SCa-3)

(Mg0.45Ca0.15Na0.26K0.01){Si7.81Al1.19}{Al2.25Mg1.31Fe0.12Ti0.02}O20(OH)4

Dioctahedral Tetrahedral andoctahedral sheet

Saponite(SAP)

Na0.8{Si7.2Al0.8}{Mg5.79Fe0.14}O20(OH)4 Trioctahedral Tetrahedral sheet

Laponite (LAP) Na0.7{Si8}{Mg5.5Li0.4}O20(OH)4 Trioctahedral Octahedral sheet

P/GM/17



Analytical measurements. The La3+ and Lu3+ concentrations were determined with ICP-OES and ICP-MSdepending on their concentration range. The concentration activities of 95Zr and 241Am were measured witha Ge semiconductor.

RESULTS AND INTERPRETATIONSIt was observed for all clays a significant decrease in the sorption Kd with the increase in the initial elementconcentration, thus indicating the existence of specific sites with a high affinity for these elements (withKd over 15000 ml g

-1 in some cases, and up to 50000 ml g-1) when La3+ and Lu3+ were present at lowconcentrations. In fact these are the scenarios of most interest, since they are more representative of thewaste management. The STx-1 and especially SCa-3 clays appeared to be much better sorbents for La3+

and Lu3+ than silica. This is consistent with the fact that in the montmorillonites two sorption mechanismstake place: surface complexation at the edge of the clay particles and exchange with the compensatinginterlayer cations. Only the first mechanism can operate in silica. The better sorption capacity of SCa-3 vs.STx-1 is due to its higher layer charge deficit that facilitates the second sorption mechanism and moreover,the tetrahedral layer deficit confers more specific sorption. The Kd values obtained for Lu

3+ were in generalhigher than those obtained for La3+. This agrees with the fact that the Kd values follow a reverse order ofhydration energy, as previously reported (Nagasaki et al., 1997). Regarding the role of the ionic scenarioin the sorption step, both the Kd(La) and the Kd(Lu) were in most cases higher in H2O than in the Ca(OH)2solution (up to one order of magnitude), thus indicating the significant role of pH and concentration ofmajor elements in the contact solution in the Kd quantification. A clear exception to this pattern was theLa-SiO2-H2O scenario, where the Kd could not be quantified, while the Kd(La) in the Ca solution was lowbut quantifiable.

Desorption data confirmed the suitability of the montmorillonites for engineered barriers, since in mostcases either the sorption reversibility could not be quantified, or the desorption yields were low. In general,the sorption reversibility decreased when decreasing the element concentration, but it also increased in theCa solution. This indicated that sorption at the high affinity sites was also less reversible, which is anadditional good indicator for the immobilisation of pollutants by these clays. However, the competing roleof Ca for the specific sites confirms that it is mandatory to simulate the ionic media of the leachingsolutions when assessing the performance of these clays.

References:N.A. Chapman, J.A.T. Smellie. Chem. Geol., 55 (1986) 167-173.

S. Nagasaki, S. Tanaka, A. Suzuki. J. Nucl. Mater. 244 (1997) 29-35.

P/GM/18


MODELLING OF THE REDOX EVOLUTIONIN THE TUNNEL BACKFILL OF A HIGHLEVEL NUCLEAR WASTE REPOSITORY

F. Grandia1, C. Domènech1, D. Arcos1, L. Duro1, J. Bruno1

1. Enviros Spain, S.L., Passeig de Rubí, 29-31, 08197 Valldoreix, Barcelona- Spain([email protected])

INTRODUCTIONReducing conditions are required to ensure the stability of both metallic containers and spent fuel storedin a high level nuclear waste (HLNW) repository at depth. These conditions are initially present in the hostrock but are disturbed by the excavation of the deposition tunnels and holes. Oxidising conditions persistduring the operational phase of the repository, up to about 100 years. During this time, atmospheric oxygenis able to diffuse into the wall rock pores affecting the water-mineral equilibrium. Precipitation of ironoxyhydroxides in the walls during this stage will have a significant influence on further evolution of theredox conditions of the system. At the end of the operational stage, molecular oxygen is expected to remainin the system, mainly in the pores of the wall rocks and in the backfill porosity.

Timing of oxygen consumption is of special concern in the safety assessment studies of the near field ofa HLNW repository. To gain an insight of this timing, reactive transport modelling of the redox evolutionin the backfill has been evaluated in this work.

REDUCING CAPACITY OF THE BACKFILLOne of the concepts for backfilling of the deposition tunnels in the KBS-3 concept of HLNW storage is amixture of crushed host rock (70%) and bentonite (30%) (SKB, 2006). Initially, air fills 42% of theporosity of the backfill, and it is progressively dissolved in porewater as regional groundwater saturates thebackfill material. Due to the low hydraulic conductivity of the rock mixture, oxygen is predominantlytransported by diffusion. Oxygen can be consumed by a number of processes, mainly (1) iron sulphideoxidation (r.1), microbial respiration (r.2), siderite dissolution (r.3), and dissolution of Fe-biotite (annite)(r.4). In processes 3 and 4, oxygen is effectively removed by subsequent Fe2+ oxidation to Fe(III), whicheventually precipitates as Fe(III)-oxyhydroxydes.

FeS2 (s) + 3.5 H2O + 3.75 O2 → Fe(OH)3(s) + 2 SO42- + 4 H+ r.1

CH2O + O2 → HCO3- + H+ r.2

FeCO3 + H+ → Fe2+ + HCO3- r.3

KFe3AlSi3O10(OH)2 +10 H2O → K+ +3 Fe2+ +Al(OH)4- +3 H4SiO4 + 6 OH- r.4

Siderite and pyrite are found in trace amounts in bentonite (MX-80) and annite is abundant in granite.Microbial activity is possible in the backfill, although the very small pore volume in the bentonite portionmay prevent microorganism development.

MODELLING OF THE OXYGEN CONSUMPTION IN THE BACKFILLReactive transport modelling in one and two dimensions has been performed to simulate the oxygenconsumption in the backfill. The simulations are based on kinetic rate laws for the reactions above defined.Results from the calculations reveal that oxygen in equilibrium with backfill porewater is quicklyconsumed (less than a month) if pyrite and/or siderite are considered as reactive minerals (Figure 1).Microbial activity, if present, is also a very effective mechanism of oxygen removal. If not present, and

P/GM/18



assuming that accessory minerals in bentonite donot play an oxygen-consuming role, annite (as partof the granite clasts in the backfill mixture) is ableto consume oxygen in ~200 years. All theseprocesses are much more efficient to lowerdissolved oxygen concentration in backfillporewater than simple diffusion to host rock,which is expected to last more than 5x104 years.

MODELLING OF THE REDOX FRONTFROM TUNNEL WALLS TO BACKFILLHigh redox potentials remain in the pores andmicrofractures of tunnel walls due to theoperational stage of the repository. Modelling offluid-rock interaction during this stage shows thatoxic conditions are expected to penetrate into therock as much as 50 cm, leading to theprecipitation of iron oxyhydroxides throughoxidation of Fe(II) present in the environment.According to the results of the modelling, anoxic conditions are re-established deeper into the host rockdue to the inflow of reducing groundwater and the redox buffering exerted by Fe(II) minerals, basicallypyrite. After sealing the tunnel, oxygen remaining in the pores of tunnel walls is quickly removed and theredox state is controlled by the Fe(OH)3(am) and Fe

2+ equilibrium. Pyrite and siderite in the backfill arethe main solid phases contributing to the buffering of the oxidising intrusion.

CONCLUSIONS

Reactive transport simulations performed to evaluate the extent and persistence of high REDOX conditionsin the backfill of HLNW repository show that accessory mineral phases in backfill materials are able toprovide enough Fe(II) to maintain favourable redox conditions. These conditions are quickly attained aftertunnel sealing.

References:SKB (2006). Long-term safety for KBS-3 repositories at Forsmark and Laxemar – a first evaluation MainReport of the SR-Can project. Swedish Nuclear Fuel andWaste Management Co Report TR-06-09.

Figure 1: Timing for oxygen consumption in thebackfill porewaters from calculations in 1D reactivetransport model.

13 d

6.5 d8 d

27 d

0

5

10

15

20

25

30

35

40

45

50

190 y

> 5.104y

Time(d)for[O2(aq)]<10–6M

Pyrite

Siderite

Py+Sid

O.Matter

Fe-biotite

Diffusion

P/GM/19


REACTIVITY OF NITRATESIN THE DIFFERENT STORAGE

COMPARTMENTS OF TYPE-B WASTES

L. André1, H. Pauwels1, M. Azaroual1, A. Albrecht2, M.-A. Romero2

1. BRGM, Water Division, 3 Avenue Claude Guillemin, Boite Postale 6009, F-45060 Orléans Cedex 2([email protected]; [email protected]; [email protected])

2. Andra, Parc de la Croix Blanche, 1/7 Rue Jean Monnet, F-92298 Châtenay-Malabry Cedex([email protected]; [email protected])

INTRODUCTION

B2-type waste consists of a mixture of an organic (bitumen) and a mineral phase (mainly sodium-nitrateand barium-sulphate salts mixed with radionuclides) proposed for storage in an underground storage cells.Re-saturation initiates after cell closure.When sufficient relative moisture is reached, salts start to dissolveand diffuse within the various compartments of the storage cell, provoking chemical disturbances. Afraction of the bitumen is solubilized and mobilized within the high concentration nitrate plume (up to10M). The development of a bacterial activity in this medium, in particular in the near field, may becomepossible. Bacterial activity accelerates nitrate reduction leading to N2 (denitrification) or NH4 formation(dissimilative reduction). This study is intended to gain a better comprehension of the redox reactionsinvolving nitrates and bacteria and their possible impact on the future evolution of the B2 waste cells inthe context of storage within Callovo-Oxfordian (COx).

MODELLING CONCEPT

Denitrification processes are modelled with Phreeqc (Parkhurst and Appello, 1999), a code which has thecapabilities to combine geochemical simulations with dispersive and diffusive transport calculations. Thegeometrical model used for these calculations is a 1D column with a nitrate source on one side (the primarywaste container) followed by two major zones (cement and argillite) in which only diffusion of dissolvedspecies (nitrate salts) is considered. The water/mineral interactions are modelled including possibledissolution or precipitation of minerals within the cement and argillite barriers. Mineral reactions involvedare based on thermodynamic equilibrium.

Particular consideration was given to the role of nitrate as an electron acceptor within the cement andargillite compartments in the presence of different electron donors (i.e. iron, pyrite, organic matter orhydrogen). Generally, these denitrification reactions are catalysed by micro-organisms which live in thischemical environment and use the surplus energy of nitrate reduction for their growth. Under theseconditions, bacterial breathing can be represented by a coupled process between the nitrate reduction andthe reactions of ATP synthesis. In order to consider the impact of microbial respiration on the kinetics ofdenitrification, Jin and Betkhe (2002; 2003; 2005) established a comprehensive law which considers twokinetic factors for the effects of the concentrations of the chemical species implied in the half reactions,and one factor, which takes into account the thermodynamic potential, driving force of the total reaction.This original approach was implemented into Phreeqc in order to model both reactive chemical transportand the associated microbial activity.

RESULTS AND INTERPRETATION

Firstly, a series of batch experiments with nitrates in the presence of different electron donors wasmodelled with Phreeqc and the implemented microbial approach. These preliminary simulations focusedon the validation of the model setup and were based on laboratory experiments and literature data. Themodelling results of denitrification in the presence of organic matter, represented by acetate (CH3COOH),

P/GM/19



show a good coherence with experimental data (Figure 1). Simulation results with other electron donors(residual oxygen, iron, hydrogen…) present a similar satisfactory correlation.

Secondly, we carried out a combined reaction/transfer model using a 1D column geometry. We performeda long-term simulation (100,000 years) in order to estimate the influence of electron donor (acetate)availability on the extent of the nitrate diffusion zone. Results indicate that under the simulated conditions,the nitrate reduction to N2 can cause significant pressure increases (up to 10 bars), particularly in the vicinityof the storage compartment, where organic matter concentrations are sufficiently important to maintain theprocess. Furthermore, these simulations show that electron donors, issued from the waste container, aremost often the limiting factors for denitrification. Due to the high content of dissolved nitrates, NO3

- ionscan migrate further into the cement and toward the COx provoking further redox reactions with new electrondonors (i.e., Fe(II), pyrite) present in these zones. These reactions will be implemented for futuresimulations. Nevertheless, even if a limited number of denitrification reactions is modelled, probablemodifications of the mineralogical composition of the cement and the COx can be identified such asdissolution of oxides, silicates and aluminates and precipitation of calcite, ettringite or clinochlore in thecement; dissolution of siderite, dolomite, illite-smectite and quartz and precipitation of calcite, illite,beidellite-Na and saponite-Na in the COx, the extent of width will be experimentally quantified.

ReferencesDevlin, J.F., R. Eedy, and B.J. Butler (2000): The effects of electron donor and granular iron on nitratetransformation rates in sediments from a municipal water supply aquifer. Journal of ContaminantHydrology, 46, 81-97.

Jin, Q., and C.M. Bethke (2005): Predicting the rate of microbial respiration in geochemical environments.Geochimica et Cosmochimica Acta 69: 1133-1143.

Jin, Q., & Bethke, C.M., (2002): Kinetics of electron transfer through the respiratory chain. BiophysicalJournal, 83, 1797-1808.

Jin, Q., & Bethke, C.M., (2003): A new rate law describing microbial respiration. Applied andEnvironmental Microbiology, 69, 2340-2348.

Parkhurst D.L. & Appelo C.A.J. (1999): User's guide to PHREEQC (version 2): A computer program forspeciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations. U.S.Geological SurveyWater-Resources Investigations Report 99-4259, 312 p.

Figure 1: Evolution of nitrate concentration in batch experiment in the presence of micro-organisms andorganic matter (acetate). Initial concentration of nitrates and micro-organisms are, respectively, 10 and1 mg L-1 (modelling performed with Phreeqc; data from Devlin et al, 2000).

0

2

4

6

8

10

0 5 10 15 20 25 30 35 40

Time (days)

[NO3-](mg/L)

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

[Biomass](mg/L)

Experimental data

Simulated nitrate concentration

Simulated biomass concentration

[X]0=1.00 mg/L

[Acetate]0= 33.00 mg/L

Ybiomasse =0.13 mg biomass/mg NO 3

Kacetate =1.20 mg/L

Knitrate =0.66 mg/L

k=1.00 mg NO3/mg biomass/day

Decreasing rate (K D)=0.06 day-1

P/GM/20


INVESTIGATION INTO THE BIODIVERSITYOF SULPHATE REDUCING BACTERIA

IN BOOM CLAYS. Aerts1, P. Boven1, M. Van Geet1,2, P. De Boever3

1. SCK•CEN – Waste & Disposal Department, Boeretang 200, 2400 Mol, Belgium([email protected])

2. Present address: ONDRAF/NIRAS – Belgian Agency for Radioactive Waste and EnrichedFissile Materials, Kunstlaan 14, 1210 Brussel, Belgium

3. Present address: VITO – Environmental Toxicology, Retieseweg, 2440 Geel, Belgium

INTRODUCTIONThe Belgian program for High Level Waste considers geological disposal in clay as the primary option forfinal disposal. In this frame Boom Clay is studied as the reference host rock. Boom Clay is present at adepth of 190 to 293 m at the reference site in Mol, which is located in the northeast of Belgium. BoomClay is a sedimentary deposit with a water content of 20 % and siliciclastic minerals, fossils and organicmatter (1-5 wt%) as main components. The Boom Clay formation contains substantial amounts of pyrite(1-5 wt%) [De Craen et al., 2004], which is partly oxidized during excavation and ventilation of the galleryresulting in high amounts of sulphate in the excavation disturbed zone [De Craen et al., 2007]. The sulphatecan be reduced by Sulphate Reducing Bacteria (SRB) yielding sulphide. A good understanding of this cycleand the impact of SRB on microbially influenced corrosion under these particular environmental conditionsis important as it will affect the containment safety function. Therefore, the focal point of our researchconcerns the presence and activity of SRB in Boom Clay.

In order to estimate the influence of SRB on corrosion under repository conditions (high pH, hightemperature), it’s necessary to know the SRB population present in Boom Clay. Based on literature dataof SRB present in the population and data from specific experiments, a prediction of the sulphideproduction, and hence the expected corrosion, should be possible.

EXPERIMENTAL CONCEPTEarly measurements showed high concentrations of hydrogen sulphide (150 ppb) in sampled clay water,which pointed to the presence of SRB. This has been investigated with culture-based most probable number(MPN) determinations, using Postgate medium. The microbial diversity is analyzed by means of the DNA-based method of terminal restriction length polymorphism (T-RFLP) using 16S rDNA, aps-A and dsr-ABprimers. The aps-A gene codes for APS reductase (key enzyme in the conversion of sulfate to sulfite) anddsr-AB gene codes for dissimilatory sulfite reductase (key enzyme in the conversion of sulfite to hydrogensulfide).

Eleven water samples from different filters of the MORPHEUS piezometer were used in the analysis. Thispiezometer is positioned vertically under the HADES laboratory and allows to study the variability of theBoom Clay pore water composition within the different layers in the Boom Clay: organic rich layers,carbonate rich layers and more silty layers. The filter number is an indication for the position of the filter,the lower the number the deeper the filter is situated.

RESULTS AND INTERPRETATIONImmediately after sampling, the eleven MORPHEUS water samples were stored under anoxic conditionsand the MPN was determined using Postgate medium. All samples showed values higher than 1,1 +E03CFU/ml (colony forming units).

Both the original water and the water samples that were first cultured in Postgate medium (MPN samples)were used in the community analysis. PCR reactions were set up to amplify part of the aps-gene using the

P/GM/20



aps-7F and FAM labeled aps-8R primers [Friedrich, 2002]. The resulting DNA fragment (~900 bp) wasdigested using the HhaI restriction enzyme and subsequently analysed on a ABI-310 Genetic Analyzer. Theobtained T-RFLP patterns were analysed using GeneScan software and subsequently used in communityanalysis (BioNumerics, cluster analysis using Pearson correlation). The bacterial populations werecompared in order to see which samples were most similar. The resulting dendrogram is shown in Figure1, together with two electropherograms.

The dendrogram shows that there are distinct differences between the original and the cultivated samples.All water samples (except W20) are grouped into 2 clusters (W10, W12, W13, W18, W23, W8 and W4,W9, W6, W2) and also the cultivated samples (except C23) are grouped into 2 clusters (C6, C13, C12, C18,C20 and C4, C10, C9, C8, C2). The apparent difference between the water and the cultured samples islikely due to the fact that culture based techniques always focus on a limited part of the bacterial populationbecause not all bacteria will grow in the same medium.

Analysis of the different T-RFLP patterns shows that the different samples have a diverse population, sincethe number of detected peaks (Genescan), i.e. the number of SRB species, ranges from 1 up to 24.

Additional analyses (e.g. 16S and dsr-AB based T-RFLP, aps-A based T-RFLP with additional restrictionenzymes and cloning experiments) are necessary. They should provide the basis for a more detailedcommunity analysis.

These tests have proven that viable SRB are present in Boom Clay and that the SRB population in BoomClay is very diverse.

References:De CraenM., Wang L., Van GeetM., Moors H. (2004): Geochemistry of Boom Clay pore water at theMolsite: SCK•CEN Scientific Report

De Craen M., Van Geet M., Honty M., Wang L., Weetjens E. (2007): Extent of oxidation in Boom Clay asa result of excavation and ventilation of the HADES URF: Experimental and modelling assessments,Abstract at Lille 2007

Friedrich, M.W. (2002): Phylogenetic analysis reveals multiple lateral transfers of adenosine-5’-phosphosulfate reductase genes among sulfate-reducing microorganisms, J. Bact., 184 (1), 278-289

ACKNOWLEDGEMENTS:This work is undertaken in close co-operation with, and with the financial support of NIRAS-ONDRAF,the Belgian Agency for Radioactive Waste and Enriched Fissile materials. The critical follow-up by andfruitful discussions with Mrs. A. Dierckx and Mr. R. Gens are very much appreciated. Discussions withand the advice of Prof. K. Pedersen are also valued highly.

Figure 1: Electropherogram of the water sample of filter 18 (top left) and of culture of filter 12 (bottomleft). Dendrogram showing similarities in populations between the different original water (W) and culture(C) samples of the MORPHEUS piezometer.

P/GM/21


COLLOID GENERATION MECHANISMSFROM COMPACTED BENTONITE UNDERDIFFERENT GEOCHEMICAL CONDITIONS

U. Alonso, N. Albarran, T. Missana, M. García-Gutiérrez

1. CIEMAT, Environmental Department, Avda. Complutense 22, Edif. 19 28040 Madrid, Spain([email protected])

Compacted bentonite is considered as engineered barrier in a high-level radioactive waste repository. Thebentonite barrier can generate colloids (Missana et al., 2003) which can absorb radionuclides (RN) andinfluence their transport.

To assess the relevance of bentonite colloids on RN migration, amongst other processes, is necessary toknow the amount of colloids that can be generated in a given environment and the degree of their stability.

The aim of this study is to establish which physico-chemical factors are playing a mayor role in thegeneration of colloids from the compacted bentonite and to quantify the parameters affecting the bentonitecolloid “source term”. Additionally, the stability of colloid generated under different geochemicalconditions will be analysed.

An experimental set-up was designed with the aim of quantifying the bentonite colloid generation ratesfrom compacted bentonite under static conditions (no flow). The experimental set-up, basically consistedon introducing a compacted bentonite tablet in a closed cell, sandwiched between two sintered stainless-steel filters (20 mm diameter, 3 mm thick and 100 µm porous size).

The experimental variables were selected accounting for different mechanism that can lead to colloidgeneration (Pusch, 1999; Ryan and Elimelech, 1996), under static conditions. The parameters of thebentonite varied were: bentonite compaction density and main exchangeable cation (natural bentonite andhomoinised clays were used). Apart from the characteristics of the clay, the solution in contact with thebentonite was also varied (natural ground waters and electrolytes at different ionic strengths).

Bentonite colloid generation is primarily related to the hydration of the clay. Due to hydration, a swellingpressure is developed, that induces the clay intrusion in available spaces (in this case the filter porous)promiting colloid generation because of local loss in density. If a local hydraulic gradient does not exist,the driving force for colloid mobilisation is the diffusion towards the stationary water layer (Kallay et al.,1987). To analyse this contribution on colloid generation rates, experiments were performed at differentcompaction densities: 1.2, 1.4 and 1.6 g/cm3. Results showed that the initial formation of bentonite colloidsis increased if the compaction is higher.

In addition, a favourable chemical environment (i.e. continuous incoming of low ionic strength waters)promotes the peptisation (defloculation mechanism) and dispersion of the clay gel, that can occursspontaneously and favoured at high pH and low ionic strength (Sen and Khilar, 2006).

To account for chemical effects on colloid generation, the experimental cells were immersed in differentaqueous solutions. The electrolytes were selected accounting for its ionic strength, from the mostfavourable conditions (deionized water, i.e. the lowest ionic strength) to Grimsel groundwater (ionicstrength 10-3 M) and NaCl 10-2 M.

First results demonstrated that the aqueous solutions were, as expected, a critical parameter for the colloidsstability, but less critical for the initial generation rates. This suggested that changes in the internal pore

P/GM/21



water ionic strength, caused by the hydration, were playing an important role. Indeed, the adsorption ofcationic surfactant on a charged surface leads to significant modification of the charge distribution in theelectrical double layer, and, thus, in the interaction of particles with each other. To analyse the effects ofthe interchangeable cations, several cells were installed with homoionised bentonite prepared withmonovalent and bivalent cations (K+, Na+, Ca2+ and Mg2+).

Results showed that initial generation rates were mostly controlled by the density of the compacted clayand the solution chemistry while, the characteristics of colloids and their stability were mainly chemistry-controlled. Colloids generated in favourable conditions presented the lower size (around 250 nm),equivalent to the bentonite colloids prepared in the laboratory, while at higher ionic strengths colloids werebigger (500 nm), despite porous sizes were hundred times higher. The experimental set-up allowedperforming stability evaluation at the same time and that after months the colloids generated in the lowerstrength electrolytes remained stable.

The mechanisms responsible of colloid generation are discussed here according to the obtained results, indifferent experimental conditions.

References:Kallay, N., Barouch, E. and Matijevic, E., 1987. Diffusional detachment of colloidal particles from solidsolution interfaces. Advances in Colloid and Interface Science, 27: 1-42.

Missana, T., Alonso, U. and Turrero, M.J., 2003. Generation and stability of bentonite colloids at thebentonite /granite interface of a deep geological radioactive waste repository. Journal of ContaminantHydrology, 61: 17-31.

Pusch, R., 1999. Clay colloid formation and release from MX-80 buffer. Technical Report TR 99-31, SKBTechnical Report TR 99-31 (1999).

Ryan, J.N. and Elimelech, M., 1996. Review: Colloid mobilization and transport in groundwater. Colloids& Surfaces A:, 107: 1-56.

Sen, T.K. and Khilar, K.C., 2006. Review on subsurface colloids and colloid-associated contaminattransport in saturated porous media. Advances in Colloid and Interface Science., 119: 71-96.

Yariv, S. and Cross, H., 1979. Geochemistry of colloid systems. Springer-Verlag, Berlin, 450 pp.

P/GM/22


EXPERIMENTAL REDUCTIONOF AQUEOUS SULPHATE BY HYDROGEN

IN THE CONTEXTOF THE CALLOVO-OXFORDIAN ARGILLITE

L. Truche1, G. Berger1, D. Guillaume1, E. Jacquot2

1. LMTG, Université de Toulouse, CNRS, IRD, OMP, 14 Av. Edouard Belin, 31400 Toulouse,France

2. Andra, 1-7 rue Jean Monnet, Parc de la Croix Blanche, 92298 Châtenay-Malabry Cedex, France

INTRODUCTIONThe Callovo-Oxfordian argillite is under investigation as a potential host rock for a high-level radioactivewaste repository in the Andra Underground Research Laboratory in Meuse/Haute-Marne (France).Radionuclides migration through the argillite formation depends among others on the pore water chemistrywhich is still under investigation (Gaucher and al., 2004, Altmann and Jacquot, 2005). In deep geologicalrepositories, hydrogen is produced both by radiolysis of organic matter contained in some waste packagesand by corrosion of the steel container protecting the nuclear waste. The Hydrogen pressure increase withtime represents a risk of mechanical damage for the surrounding engineered and geological barriers (Ortizand al., 2002; Lassabatère and al., 2004). One of the remaining uncertainties concerns the role of H2 as areducer. More particularly, the reduction of aqueous/mineral sulphates and other oxidized species presentin the site can induce a change in the geochemical conditions prevailing in the initial argillite. Hence, thepurpose of this study is to investigate nature and rate of sulphate reduction under significant H2 pressure.It should be noted that sulphate reduction, largely studied with organic matter as a reducing agent, is a slowreaction at ambient temperature. As it may be the same with H2, experiments have been performed first,in the 150-250 temperature range. The objective is to quantify the kinetic parameters of the reaction as afunction of temperature and gas pressure for a further evaluation of the extent of reaction in the nuclearwaste repository conditions, at lower temperature.

MATERIALS AND METHODSAll the experiments were conducted in a 300 cm3 stirred Autoclave Engineer reactor made of Hastelloy.The experimental apparatus enabled periodic on-line sampling of reaction fluids to monitor reactionprogress, and control of gas pressure at a maximum value of 350 bars and 400°C.

The first experiments reported here were conducted in the 200-250°C temperature range, using 0,027MNa2SO4 solutions in accordance with the sulphate concentrations of the Callovo-Oxfordian formation water;and under H2 partial pressure constrained at 17-20 bars with an Ar-H2 mixture. The reacting solution wasprepared in an anoxic glove box and we paid attention all along the procedure and sampling to avoidatmospheric oxygen contamination.

pH and Eh were measured in glove box immediately after the on-line sampling, and sulphateconcentrations were measured by ionic chromatography (IC) using a Dionex ICS-2000 operating as agradient ion-exchange chromatography-system.

FIRST RESULTS AND PERSPECTIVESThe concentration of sulphate remaining in the experimental apparatus decreased with increasing time ineach individual experiment. We also observed a first initial stage of very slow or absent reaction. The rateat which the decrease in sulphate concentration occurs seems to depend on the H2 pressure and thetemperature (Fig.1. A). A sulphate half-live of 10 days at 200-250°C can be extrapolated from theseexperiments. The phenomena leading to the observed initial plateau are under investigation.

P/GM/22



These first kinetic data can be compared to rate values reported in the literature for ThermochemicalSulphate Reduction (TSR) in the sedimentary context (Goldhaber and Orr, 1995; Worden and Smalley,1995; Bildstein and al, 1999; Cross and al., 2004). Although TSR involves mineral sulphate and morecomplex electron donors such as kerogen or soluble organic compounds, the kinetics of SO4

2- reduction isin the same order of magnitude in the different systems (Fig. 1. B).

The further investigations will consist in better exploring the effect of temperature and H2 pressure on thekinetic rate, and to experiment interactions with Callovo-Oxfordian argillite.

References:Altmann, S., Jacquot, E., 2005. La chimie des eaux interstitielles dans la couche du Callovo-Oxfordien àl’état initial (site Meuse/Haute-Marne). Note technique Andra C.NT.ASTR.03.023

Bildstein, O., Worden, R.H., Brosse, E., 2001. Assessment of anhydrite dissolution as the rate-limiting stepduring thermochemical sulphate reduction. Chemical Geology 176, 173-189.

Cross, M.M., Manning, D.A.C., Bottrell, S.H., Worden, R.H., 2004. Thermochemical sulphate reduction(TSR) experimental determination of kinetics and implications of the observed reaction rates forpetroleum reservoirs. Organic Geochemistry 35, 393-404.

Descostes, M., Tevissen, E., 2004. Definition of an equilibration protocol for batch experiments onCallovo-Oxfordian argillite. Physics and Chemistry of the Earth 29, 79-90.

Gaucher, E., Blanc, Ph., Barot, F., Lassin, A., Crouzet, C., Moussay, A., Braibant, G., Breeze, D., 2004.Simulation de la chimie des eaux du Callovo-oxfordien. Rapport BRGM-RP-52039-FR.

Goldhaber, M.B., Orr, W.L., 1995. Kinetic controls on thermochemical sulphate reduction as a source ofsedimentary H2S. In: Vairavamurthy, M.A., Schoonen, M.A.A. (Eds.), Geochemical Transformationsof Sedimentary Sulphur. American Chemical Society Symposium Series 612. American ChemicalSociety, pp. 412-425.

Lassabatère, T., Dridi, W., Servant, G., 2004. Gas transfer and mechanical incidence on storage barriers.Applied Clay Science 26, 511-520.

Ortiz, L., Volckaert, G., Mallants, D., 2002. Gas generation and migration in Boom Clay, a potential hostrock formation for nuclear waste storage. Engineering Geology 64, 287-296.

Worden, R.H., Smalley, P.C., Oxtoby, N.H., 1995. Gas souring by thermochemical sulphate reduction at140 C. American Association of Petroleum Geologists Bulletin 79 (6), 854-863. Symposium Series612. American Chemical Society, pp. 412-425.

Figure 1: A- Variation in sulphate concentration as a function of time at 200 and 250°C; B- Inverse ofreaction temperature (1000/T with T in Kelvin) against the log of sulphate half-live and comparison withother published TSR experiments.

A

0

500

1000

1500

2000

2500

3000

0 100 200 300 400 500 600

Time (hours)

SulphateConcentration(mg/l) Run 1 : 250°C, 250 bars

Run 2 : 200°C, 185 bars

B

0

1

2

3

4

5

6

7

8

9

10

1,5 2 2,5 3

1000/T (K)

Goldhaber et Orr (1995)

Berger et al. (2005)

Cross et al. (2004)

ce travail Run 1

ce travail Run 2

Logt1/2sulphate(hours)

P/GM/23


CATION EXCHANGED FE(II) AND SRAS COMPARED TO OTHER DIVALENT

CATIONS (CA, MG) IN THE CALLOVIAN-OXFORDIAN FORMATION. IMPLICATIONS

FOR POREWATER COMPOSITIONMODELING

Christophe Tournassat1, Catherine Lerouge1, Jocelyne Brendlé2, Jean-Marc Greneche3,Stéphane Touzelet1, Philippe Blanc1 and Eric C. Gaucher1

1. French Geological Survey (BRGM), France, [email protected]. LMPC, Mulhouse, France3. LPEC, Le Mans, France

ABSTRACTFe and Sr bearing phases were thoroughly investigated by the mean of spectrometric and microscopictechniques in Callovian-Oxfordian (COX) samples originating from the ANDRA underground researchlaboratory (URL) at Bure (France). Sr was found to be mainly associated with celestite mineral phaseswhereas Fe was found to be distributed over a large panel of mineral phases. Fe is mainly in its +IIoxidation form in the studied samples (~93 % fromMössbauer results). Fe(II) is found to be present mainly

Figure 1: A: Micrograph of pyrite (Py) and sphalerite (Sph) (backscattered electron image – sample12436); B: Macrograph of a celestite vein crosscutting the Callovian-Oxfordian formation (Sample EST20714); Micrograph of an euhedral grain of dolomite with an iron-rich rim (backscattered electron image– sample EST 21400); D: Micrograph of an euhedral grain of carbonate. The core is a iron-rich calcite(Cc), whereas the rim is a siderite (Sd) (backscattered electron image – sample EST 25687).

P/GM/23



in pyrite, siderite/ankerite and clay minerals. Iron +III, if present, is associated to clay minerals (probablyillite, IS and chlorite). No Fe(III) oxi(hydro)xide could be detected in the samples. Strontianite was notobserved too. Based on these observations it is likely that the COX porewater is in equilibrium with thefollowing carbonate minerals, calcite, dolomite and ankerite/siderite, but not with strontianite. It is thenshown that these phase (dis)equilibrium constrains can be combined with the clay cation exchangecomposition information in order to give direct estimates or constrains on the solubility products of thesecarbonates minerals: dolomite, siderite and strontianite. As a consequence an experimental method wasdeveloped to retrieve the cation exchanged Fe(II) in very well preserved COX samples.

It has then been found that the very homogeneous cation exchange composition of the formation iscompletely in agreement with an homogeneous presence of calcite and dolomite minerals whoseequilibrium reactions control part of the porewater composition. Amongst the broad range of valuesavailable for dolomite solubility products in thermodynamics database, the value of log Kdolomite = -3.56 isthe most appropriate for the one present in the COX formation. With regards to strontianite that is neverobserved in the formation it appears that the equilibrium constant tabulated in the Llnl database is not validfor clayey natural systems. The value given in BUSENBERG et al., 1984, used by most of other availablethermodynamic databases seems to be far more appropriate. Concerning Fe(II) and siderite/ankeriteequilibrium, it was found that the measured Fe/Ca ratio on the clay exchanger (~0.01) could only beconsidered as a maximum value due to experimental possible bias, leading to the following constrain forthe solubility of the siderite-like phase present in the COX: Fe0.7Mg0.2Ca0.1CO3 + H

+ ⇔ 0.7 Fe2+ + 0.2 Mg2+

+ 0.1 Ca2+ + HCO3- log Ksiderite_COX < 0.42. This constrain is in agreement with the estimated solubility of

this phase based on a simple ideal solid solution model: log Ksiderite_COX = -0.21. These information will helpto refine the clay rock porewater model already developed in GAUCHER et al., 2006 (see Gaucher et al, thisissue).

ReferencesBusenberg E., Plummer L. N., and Parker V. B. (1984) The solubility of strontianite (SrCO3) in CO2-H2Osolutions between 2 and 91°C, the association constants of SrHCO+3(aq) and SrCO

03(aq) between 5 and

80°C, and an evaluation of the thermodynamic properties of Sr2+(aq) and SrCO3(cr) at 25°C and 1 atmtotal pressure. Geochim. Cosmochim. Acta 48, 2021-2035.

Gaucher E. C., Blanc P., Bardot F., Braibant G., Buschaert S., Crouzet C., Gautier A., Girard J.-P., JacquotE., Lassin A., Negrel G., Tournassat C., Vinsot A., and Altmann S. (2006) Modelling the porewaterchemistry of the Callovian-Oxfordian formation at a regional scale. C.R. Geosci. 338, 917-930.

Figure 2: Triangular Ca-Fe-Mg diagramshowing the main chemical compositions ofthe carbonates in the Callovian-Oxfordianformations. Black symbols correspond to thesample EST25687 in which siderite grewafter iron-rich calcite. Open symbols corre-sponds to all other samples.

100

90

80

70

60

50

40

30

20

10100

90

80

70

60

50

40

30

20

10

100 90 80 70 60 50 40 30 20 10

Ca

Fe Mg

mmmmmmmmmmmmmmm

mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

mmmmmmmmmmmm

mm

m

m

m

mm

mm

mmmmmmmmmmmm

m

mm

mmm

u

uuu

uu

u

u

u

u

u

u

u

ummmmmmmm

mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

mmmmmmmmmmmmm

mmm

mmm

m

Calcite

Siderite

DolomiteAnkerite

100

90

80

70

60

50

40

30

20

10100

90

80

70

60

50

40

30

20

10

100 90 80 70 60 50 40 30 20 10

Fe Mg

mmmmmmmmmmmmmmm

mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

mmmmmmmmmmmm

mm

m

m

m

mm

mm

mmmmmmmmmmmm

m

mm

mmm

u

uuu

uu

u

u

u

u

u

u

u

ummmmmmmm

mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

mmmmmmmmmmmmm

mmm

mmm

m

P/GM/24


MONTE CARLO SENSITIVITY ANALYSISOF MODELLED OPALINUS CLAY

POREWATERS FROM THE MONT TERRIROCK LABORATORY

T. Thoenen1

1. Waste Management Laboratory, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland([email protected])

INTRODUCTIONOne of the reasons why clay formations are chosen as potential host rocks for radioactive waste repositoriesis their very low permeability. This great advantage turns into a disadvantage when it comes tocharacterizing the in situ clay porewater composition, because it is exceedingly difficult to obtainrepresentative water samples. In the case of the Mont Terri Rock Laboratory, water data from boreholesamples, squeezed rock samples, and laboratory studies were compared and, combined with mineralogicaldata, a synthesis was developed using geochemical modelling (Pearson et al., 2003). Since severalparameters that constrain the geochemical model are not very well known (among many others pCO2 andthe exchangeable cation populations) it is important to assess the sensitivity of modelling output parametersto uncertainties in the input. “Classical” sensitivity analyses are typically conducted by varying an inputparameter at a time and performing a separate calculation for each variation. While it is certainly possibleto obtain the maximal range of output parameters in this way, the procedure is cumbersome if several inputparameters are to be varied systematically. A more efficient and informative method is the Monte Carlomethod.

MONTE CARLO METHODThe Monte Carlo method for quantifying uncertainties in geochemical calculations was introduced to thegeological literature by Anderson (1976). The method consists of repeated calculations of a geochemicalmodel, whereby all uncertain input data are changed randomly and simultaneously within the stateduncertainty limits from calculation to calculation. Statistical evaluation of the variation in the resultingoutput parameters yields their uncertainty. Sources of uncertainty in the input can be thermodynamic dataor analytical data needed to constrain the geochemical model. For our purposes, we concentrated onanalytical uncertainties.

In order to perform the sensitivity analysis we set up a simple Perl script that is centered on thegeochemical code PHREEQC by Parkhurst & Appelo (1999). The script (1) chooses random data forselected input parameters within specified limits using a specified probability distribution (uniform,Gaussian etc.), (2) writes an input file, (3) runs a batch version of PHREEQC, and (4) writes specifiedoutput data to an output file. Steps (1) to (4) are repeated ad nauseam, that is, until running averages ofoutput parameters have stabilized to constant values. The resulting distributions of output parameters canbe considered as a measure of their uncertainty.

PRELIMINARY RESULTSPearson et al. (2003) modelled water from the BWS-A1 and BWS-A3 boreholes at the Mont Terri RockLaboratory. The modelling assumption was that, with the exception of certain free or non-reactiveconstituents (Cl-, Br-, and SO4

2-), the porewater composition is controlled by saturation reactions withminerals (mainly carbonates) and by cation exchange reactions with clay minerals. We tested our MonteCarlo procedure on the model for the BWS-A3 porewater. The saturation minerals and their controlledelements were calcite (Ca), dolomite (Mg), siderite (Fe), rhodochrosite (Mn), quartz (Si), kaolinite (Al),UO2 (U), while pyrite (coupled with SO4

2-) provided redox control. Variations were considered for the non-

P/GM/24



reactive constituents Cl- (0.119 – 0.137 mol/kg H2O), Br- (fixed Br-/Cl- ratio), and SO4

2- (fixed SO42-/Cl-

ratio); for the cation exchange population (controlling Na, K, and Sr) with NaX (495 – 500 mol/kg H2O),KX (20 – 80 mol/kg H2O), MgX2 (65.5 – 85 mol/kg H2O), CaX2 (120 – 160 mol/kg H2O), and SrX2 (5 –8 mol/kg H2O); and for pCO2 (-2.7 – -1.5). Uniform probability distributions were used for thesecalculations.

The variability of selected output parameters is illustrated by the histograms in Figure 1, which werecalculated from 25 000 Monte Carlo iterations. Notable is the skewness in the distributions of total C andtotal U, the explanation of which will be the subject of further investigations. Such skewed distributionscan easily be missed by “classical” sensitivity analysis.

CONCLUSIONSThis feasibility study has shown that the proposed Monte Carlo method is an efficient and computationallycheap way to assess the variability of output parameters due to uncertainties in the analytical input of aporewater model. In addition, the flexibility of the PHREEQC input file structure easily allows theextension of our approach to uncertainties in thermodynamic data (e.g., cation selectivity constants andsolubility constants). Monte Carlo sensitivity analysis is a promising tool for further investigations into theuncertainties inherent in porewater models at the Mont Terri Rock Laboratory.

References:Anderson, G.M. (1976): Error propagation by the Monte Carlo method in geochemical calculations.Geochimica et Cosmochimica Acta, 40, 1533-1538.

Parkhurst, D.L. & Appelo, C.A.J. (1999): User’s guide to PHREEQC (version 2) – a computer program forspeciation, batch reaction, one-dimensional transport, and inverse geochemical calculations. U.S.Geological Survey, Water-Resources Investigations Report 99-4259, 312 pp.

Pearson, F.J, Arcos, D., Gaucher, E. & Waber, H.N. (2003): Pore water chemistry and geochemicalmodeling. In: Pearson, F.J., et al. (2003): Mont Terri Project – Geochemistry of Water in the OpalinusClay Formation at the Mont Terri Rock Laboratory. Reports of the Federal Office for Water andGeology (FOWG), Geology Series No. 5, 67-105.

Figure 1: Histograms of selected output parameters (concentrations in mol/kg H2O), calculated from25 000 Monte Carlo steps.

P/GM/25


MICROSTRUCTURAL INVESTIGATIONOF OPALINUS CLAY – PROPOSAL

OF A CARBONATE DISTRIBUTION MODELM. Klinkenberg1,2, S. Kaufhold1, R. Dohrmann1, S. Siegesmund2

1. BGR - Bundesanstalt für Geowissenschaften und Rohstoffe, Stilleweg 2, 30655 Hannover,Germany([email protected], [email protected], [email protected])

2. GZG – Geowissenschaftliches Zentrum der Universität Göttingen, Goldschmidtstr. 3, 37077Göttingen, Germany ([email protected])

INTRODUCTIONThe micro fabric of a rock represents the spatial arrangement of minerals, their shape, size and orientation.Therefore the porosity and texture are a result of the micro fabric. The influence of clay micro fabric onrelevant physical properties for barrier systems of clays is more often claimed than proved (Bauer-Plaindoux et al., 1998). One reason might be that the characterisation of the micro fabric still is a challengebecause of the very fine grained nature of the material and its heterogeneity on various scales. Whenstudying physical properties of clays, effects of wetting, drying, and oxidation during sample preparationand storage on micro fabric have to be considered.

This study shows a comparison of two methods which proved to be suitable for the investigation of claymicro fabric on the micro scale: Scanning Electron Microscopy (SEM) combined with EDX mapping andinfrared microscopy. The aim of this study is to explain the different mechanical behaviour of the Opalinusclay samples with respect to mineralogical composition and micro fabric. A number of mechanical uniaxialand triaxial compression tests were performed on shale samples of the Opalinus Clay formation, which arederived from bore holes of the Laboratory Temperature Testing (LT) Experiment (Schnier, 2004). For “s-samples” (sample axis perpendicular to bedding) different failure values could be distinguished. Inconclusion, two different types of Opalinus Clay with respect to mechanical properties were identified. Themost important result is that not only the carbonate content but also the grain size distribution of carbonatesseems to have an influence on the mechanical behaviour of the clay. For this reason size and shape ofcarbonates are investigated and correlated with the mechanical parameters.

EXPERIMENTAL CONCEPTFor the comparison of the two microanalysis methods (SEM/EDX and IR-microscopy) different samplespreparation techniques were used: freshly broken pieces of Opalinus Clay, uncovered thin sections, andpolished sections.

After mechanical testing the samples for the SEM and IR measurements were taken perpendicular to thebedding by sawing. Polished sections produced by a dry polish method to avoid artefacts due to wateruptake and to minimize artefacts caused by drying proved to be optimum for optical characterization. Forthe analysis of SEM images and EDX-mapping the ImageSXM® software and a number of macros(download from http://pages.unibas.ch/earth/micro/) were used. Carbonates were segmented and thefollowing parameters were determined by image analysis: grain size, aspect ratio (longest axes/shortestaxes), shape factor (perimeter/equivalent perimeter), and angle of orientation. The results of the imageanalyses are compared to the results of the mechanical tests.

RESULTS AND INTERPRETATIONThe EDX mapping provides chemical information about the sample. Some combinations of elementsindicate phases (e.g. Fe & S = pyrite). The infrared microscopy provides intensity distributions ofabsorption bands which are characteristic for specific minerals.

P/GM/25



The advantages and disadvantages of both methods are discussed in detail. Although different analyticalprincipals are applied the observed microstructures in the mapped area are comparable. These methodsprovide a suitable basis for subsequent quantitative image analysis.

The above mentioned image analysis values were used for the explanation of yet unexplainable differencese.g. of mechanical behaviour, since they are expected to be determined by the micro fabric. We concludethat shape and grain size distribution of carbonates are supposed to play an important role. A method wasdeveloped and the hypothesis is that Clays of the same formation such as Opalinus clay with a high contentof coarse grained carbonate show lower failure values than fine grained.

References:Bauer-Plaindoux, C., Tessier, D., and Ghoreychi, M. (1998) Propriétés méchaniques des roches argileusescarbonateés: importance de la relation calcite-argile. C. R. Acad. Sci. Paris, Sciences de la Terre et desplanètes / Earth and Plantary Sciences, 326, 231-237.

Schnier, H. (2005): LT Experiment: Strength tests on cylindrical specimens, documentation and evaluation,(phases 6 and 7). Technical note 2002-50, 86 p., 17 fig., 8 tab., 3 app., Hannover.

P/GM/26


INFLUENCE OF NATURAL ORGANICMATTER ON THE SORPTION BEHAVIOUR

OF EU ON ILLITE(AS MODEL COMPONENT FOR BOOM CLAY)

D. J. Liu, C. Bruggeman, N. Maes

SCK-CEN, Boeretang 200, B-2400 Mol, Belgium

INTRODUCTIONThe elucidation of the role of natural organic matter on metal and radionuclide mobility is of considerableimportance when assigning transport parameters (diffusion, retention, operational solubility) in organicmatter rich clay host rocks (e.g. Boom Clay). On one hand, binding of components to solid organic matteror immobile colloidal organic matter enhances retention while on the other hand, the mobility of thecomponent might be increased due to complexation to mobile organic matter. In the present study, thesolid-to-liquid distribution of europium (studied as a chemical analogue for long-lived trivalent actinides),in ternary systems containing illite (Silver Hill illite) and different batches of Boom Clay organic matteris investigated. The different NOM batches reflected different molecular weight fractions which arepresumed to be either mobile or immobile under in situ conditions.

EXPERIMENTALBatch sorption experiments were designed to evaluate the influence of organic material on the sorption ofEu3+ on illite under conditions respresentative for the Boom Clay. Batches of illite were pre-equilibratedin dialysis bags with Synthetic Boom Clay Water (SBCW, ~0.014 mol/l NaHCO3) before use. Eu sorptionisotherms (10-9 mol·l-1-10-7 mol·l-1) were measured in samples containing 1.6 g·l-1 illite in SBCWbackground electrolyte (pH 8.3) under controlled atmosphere conditions (99,6% Ar; 0,4% CO2), in thepresence of 40 ppm dissolved organic matter with different molecular weight distributions. The solutionswith natural organic matter were obtained either from sampling filters in the HADES URL (RBCW, RealBoom Clay Water) or by leaching Boom Clay (BC) cores with SBCW or RBCW (Figure 1). Sorption ofdissolved organic matter on illite was investigated in a separate experiment. Reversibility and theestablishment of chemical equilibrium were checked by varying the order of mixing (and pre-equilibrating)of the three different components (Eu, organic matter, and illite). Phase separation was carried out by aconsecutive procedure of centrifugation (2 hrs at 21 000g) and ultrafiltration (30 kDa) to distinguishimmobile (larger colloids) from mobile (small colloids and dissolved species) Eu fractions.

RESULTS AND DISCUSSIONIn absence of humic substances, the measured KD values (LogKd = 3.6±0.2) could be modelled using the2 SPNE SC/CE surface complexation model of Bradbury and Baeyens (Bradbury et al., 2005) andthermodynamic complexation constants taken from literature (Hummel et al., 2002). For these systems, theEu distribution was also independent of the phase separation procedure, indicating that the formation ofinorganic colloids containing Eu is negligible.

With respect to the influence of humic substances on the Eu distribution, the sorption of dissolved organicmatter on illite was very weak but dependent on the size distribution with larger organic matter moleculesbeing more likely to interact with the illite phase (LogKd~0.8-1.5).

The Eu distribution in all ternary systems, measured both after centrifugation and after ultrafiltration, couldbe well described by linear sorption relation. There was no observed influence of the order of mixing onthe measured Eu distribution, indicating that chemical equilibrium was established in all samples and thatsufficient reversibility was attained. As expected, the presence of dissolved organic matter in solution after

P/GM/26



centrifugation decreased the Eu distribution coefficients by an order of magnitude because of complexationreactions between Eu and organic matter.

After ultrafiltration, the amount of dissolved organicmatter andEu concentration decreased relative to the sizedistributions of the different organic matter batches wich resulted in increased calculated distributioncoefficients (logKd~3.4-4.1) (Figure 2).This observation leads to the unexpected conclusion that the decreasedsorption onto illite due to the formation of Eu-NOM colloids,may be translated into an increased retention ofEu under in situ conditions because of the immobility of the large-sizeNOMmolecules (> 100 kDa).

At present, the influence of dissolved organic matter on the observed Eu distributions is being incorporatedinto geochemical modelling, allowing also to make distinction between larger- and smaller-size organiccolloids. With this incorporation, it is attempted to draw conclusions with respect to the in situ situation,based on the size distribution of autigenic organic matter.

CONCLUSIONBecause of the high affinity of (trivalent) radionuclides to natural organic matter, sorption of theseradionuclides to the inorganic mineral assemblage will be decreased under conditions relevant for BoomClay. However, the overall retention may be increased because of the immobility of high molecular-weightNOM molecules under in situ conditions.

ReferencesBradbury M. H., Baeyens B., Geckeis H. et al. (2005): Sorption of Eu(III)/Cm(III) on Ca-montmorilloniteand Na-illite. Part 2: Surface complexation modelling. Geochim. Cosmochim. Acta, 69(23): 5403-5412.

Hummel W., Berner U., Curti E. et al. (2002): Nagra / PSI Chemical Thermodynamic Data Base 01/01.Technical Report 02-16. July 2002

ACKNOWLEDGMENTMr. D. J. Liu acknowledges the financial support of the Belgian Government via FOD wetenschapsbeleidas a visiting scientist. This work is undertaken in close co-operation with, and with the financial supportof NIRAS-ONDRAF and EC in the frame of the 6FP FUNMIG project. The critical follow-up by andfruitful discussions with Mrs. A. Dierckx, Mr. R. Gens (N/O) are very much appreciated.

Figure 1: The size distribution of different naturalorganic matter

Figure 2: Relationship between remaining TOCafter phase separation and sorption distribution ratioin different systems

0%

20%

40%

60%

80%

100%

<1KD

(1-10)KD

(10-30)KD

(30-100)KD

100KD-0.22um

(0.22-0.45)um

>0.45um

RBCW

RBCW + BC (10:1)

SBCW + BC (10:1)

SBCW + BC (5.5:1)

0%

20%

40%

60%

80%

100%

120%

Euonillite

Eu-RBCWonillite

Euonillite-RBCW

Eu-RBCW+BConillite

Euonillite-RBCW+BC

Eu-SBCW+BConillite

Euonillite-SBCW+BC

remainingTOCafterseparation[%]

2.5

2.8

3.1

3.4

3.7

4

4.3

logKd[L/Kg]

remaining TOC after centrifugationremaining TOC after ultra-filtrationlogKd after centrifugationlogKd after ultra-filtration

P/GM/27


FATE OF BORON AND DYNAMICSOF REACTIVE TRANSPORT PROCESSES

IN THE NEAR FIELD OF A HLW DISPOSALT.Weber1, L. Trotignon1, C. Pozo1, O. Bildstein1, G. de Combarieu2, P. Frugier2, D. Menut3

1. CEA Cadarache, DTN/SMTM/LMTE, 13108 Saint Paul Lez Durance ([email protected]),2. CEA Valrhô Marcoule, DEN/VRH/DTCD/SECM/LCLT, 30207 Bagnols-sur-Cèze,3. CEA Saclay, DPC/SCPA/LALES, 91191 Gif Sur Yvette.

INTRODUCTIONThe near field of a high activity nuclear waste glass repository will undergo physical and chemicaltransformations that couple degradation processes of canisters, alteration of glass packages and diffusivetransfers of water and reactants through the engineered barrier system and in the first decimeters of thegeological medium. The time span of such transformations extends from 103 to 106 years and will controlthe retention and transfer properties of key radionuclides (eg. long lived fission products like Tc, Se oractinides like Np) through the near field.

Performance assessment of such a near field system will in part rely on numerical simulations enabling toreproduce the spatial and temporal evolution of transformations affecting the porous media and degradingmaterials. In the case of nuclear waste glasses, which are borosilicates, the understanding of the fate ofboron in the near field is one of the keys for the understanding and predicting of several features relatedto coupled processes.

EXPERIMENTAL AND MODELLING APPROACHBoron is one of the most mobile constituent of the altering glass. This element is in addition not presentas a major element in the surrounding materials (clays, claystone). Up to present day calculations, boronhas been considered as an inert element in the near field.

Recent literature survey however reveals that boron may be sorbed on clays (Su and Suarez, 1995) and oncorrosion products such as iron oxides (Goldberg, 1997) and may also substitute to other elements inaluminosilicate minerals.

Previous calculations coupling the interaction between glass, canister and clay, have pinpointed the factthat the pH level around the waste package was dependent on the fate of boron (Bildstein et al., 2007).Dissolved boron (as boric acid and/or polyborates) acts in fact as an efficient pH buffer. In the case ofsignificant boron adsorption, the scenario of pH evolution around waste packages may be modified.

Moreover, boron could compete with silicon for the same sorption sites (Goldberg, 1997; Philippini et al.,2006). This phenomenon is therefore crucial since it would directly impact the alteration rate of glass.

Reactive transport modelling of boron in the near field was performed both with the Alliances code(Bengaouer et al., 2003) and with the Chess and the HYTEC codes (van der Lee et al., 2003).

Main processes affecting boron fate in the near field, identified from literature survey, were included intothe speciation chemistry of boron:

• adsorption by iron oxides formed during canister corrosion (e.g. mainly magnetite)

• adsorption by clays

• adsorption by calcium carbonates (e.g. calcite)

P/GM/27



Kd values for this processes were, when possible, extracted from literature experiments (i.e. in the case ofkaolinite, montmorillonite and calcite) or determined from batch experiments (magnetite) when no otherdata were available.

We will present results of modelling boron fate when interacting with corrosion products and with bothcorrosion products and clay minerals and compare these predictions with spectroscopic studies of samplesfrom integrated experiments.

References:Bengaouer, A., P. Montarnal, L. Loth and Gaombalet (2003): ALLIANCES Project: integration, analysisand design software environment for nuclear waste storage and disposal. International Conference onSupercomputing in Nuclear Applications, Paris, France.

Bildstein, O., L. Trotignon, P. Claudine and J. Michel (2007): Modelling glass alteration in an alteredargillaceous environment. Journal of Nuclear Materials in press.

Goldberg, S. (1997): Reactions of boron with soils. Plant and Soil 193(1-2): 35-48.

Philippini, V., A. Naveau, H. Catalette and S. Leclercq (2006): Sorption of silicon on magnetite and othercorrosion products of iron. Journal of Nuclear Materials 348: 60-69.

Su, C. and D. L. Suarez (1995): Coordination of Adsorbed Boron: A FTIR Spectroscopic Study.Environmental Science and Technology 29(2): 302-311.

van der Lee, J., L. De Windt, V. Lagneau and P. Goblet (2003): Module-oriented modeling of reactivetransport with HYTEC. Computers & Geosciences 29(3): 265-275.

P/GM/28


SULFATE-REDUCING BACTERIA(DESULFOVIBRIO. DESULFURICANS)

ACTIVITY MONITORED BY MAGNETICMEASUREMENTS IN BURE CLAYSTONES

(FRANCE) AND MONT TERRI CLAYSTONES(SWISS)

C. Aubourg1, D. Janots2 and J-P. Pozzi2

1. CNRS. Laboratoire de Tectonique. University Cergy Pontoise France 2) CNRS. Laboratoire deGéologie. ENS Paris France

With modern magnetic equipment, we aim to detect subtle variations of iron and sulphide oxides thatoccurred during biological activity. Action of bacteria on radionucleides transit and/or transformationvector is one of the major research trend in radioactive wastes disposal studies. Within the excavateddamaged zone (EDZ) of the galleries leads to rapid weathering of iron sulphides such as pyrite (FeS2) inthe host rock. Under the joint action of oxygen and water, many sulphates are produced. In addition,organic matter (~1%) and anaerobiotic micro-conditions may favour the development of Desulfovibri.desulfuricans species (SRB). During thermal stress at 95°C, we show that few ppm of newly magneticgrains as well as sulphate are produced (Pozzi et al., this volume). We take advantage of thesetransformations to compare SRB activity for natural and heated (95°C) claystones.

We carried out 4 in vitro experiments at 32°C (18 samples) on callovian-oxfordian claystones from theParis basin (Bure, France) and from Opalinus claystones (Mont Terri, Swiss). These claystones, similar intheir chemical composition, have however different geological stories. Mont Terri claystones sufferedalpine orogeny during Cainozoic time. As a result, these claystones are relatively mature (~90°C) comparedto Bure claystones which are immature (~60°C).We placed SRB-rich and SRB-free claystones in a specialincubator where a vertical magnetic field of 2 mT (~40 times the Earth magnetic field intensity), is applied.We measured the remanent magnetization (RM) with a SQUID magnetometer 2G and counted the SRBoptically. RM is a combination of :

• A visquous magnetization imprinted in initial magnetic grains (VRM),

• A chemical remanent magnetization (CRM) imprinted when newly formed magnetic grains passed a cri-tical volume (from 50 nm to several µm);

• An eventual loss of imprinted magnetization in relation with SRB activity.

Note that magnetic measurement is not destructive, and that it is not necessary to remove sample fromholder. In addition, few ppb of magnetic grains can be detected.

During (A) experiment, we take natural claystones and test the activity of SRB in a sea-water like mediumregularly supplemented in lactate and sulphate. We observed a good SRB activity (few millions/ml) allalong the ~7 months experiments. However, we saw only similar VRM-like magnetization for SRB-richand SRB-free claystones. Thus, we do not see particular trend of any SRB magnetic activity for naturalclaystones during A experiment.

For B, C and D experiments, we initially heated claystones at 95°C for several weeks. B experiment is likeA experiment for non heated claystones. No sulphate is supplemented in experiment C. Experiment D islike experiment C, but no magnetic field was applied, and samples were shielded against the earth magneticfield. In heated Mont Terri claystones (C experiment), we see again a VRM-like magnetization, just as

P/GM/28



observed in experiment A. There isno obvious magnetic signature ofSRB activity in heated Mont Terriclaystones.In heated Bure clay-stones, we observed either VRM-like magnetization just as we see inMont Terri, or remarkable changesof magnetization due to SRB activ-ity. We present some examples ofSRB-rich monitoring.

We show the monitoring of COXBure claystones at level C2B1 (675m) (left figure).While SRB-free shows only a VRM-type magnetization, SRB-rich shows: 1) a large increaseof RM after 5 days, also correlated to a high SRB concentration.We call this new RM : CRM1. 2) CRM1 isprogressively destroyed from 5 to 69 days; 3) then an increase-to-stabilisation of RM.We labelled this RMCRM2. The drop of CRM1 is not necessarily correlated to a drastic chemical alteration of CRM1 magneticcarriers. A slight change of Fe versus S in iron sulphide can change dramatically the magnetic properties. Bycontrast, CRM2 is likely correlated to input of newly formed magnetic grains that passed the critical volume.For this monitoring, we account thus for three distinct signatures of SRB metabolism. Preliminary rockmagnetic investigation shows that magnetic assemblage is changing from natural, heated and SRB-richclaystones.

For COX Bure claystones of level C2b1 (845 m) (right figure), we ran D experiment, without sulphate andwithout magnetic field. After 80 days, SRB activity removed about 70% of the initial magnetization. It isknown that SRB activity through hydrogen sulphide production can reduce iron oxides. Large iron oxidesreduction can thus explain the observed loss of magnetization. However, the large drop suggests that thisaction took place not only on the surface of sample, but also in the center, where porosity is too small tolet SRB live correctly.

We suggest that the action of moderate heating of poorly matured claystones as the Bure claystones triggerssome aspects of the SRB metabolism. We think that the reduction of iron oxides as well as production ofiron sulfides is part of SRB metabolism in claystones.

Figure 1: Magnetic monitoring of SRB-free and SRB-rich Bure claystones and bacterial counting of SRB-rich claystones. These claystones were initially resaturated at 95°C following Pozzi et al. (this meeting)technique. In A), experiment took place with a magnetic field to evidence newly formed minerals resultingfrom SRB activity. In B) the experiment took place in magnetic field free space to measure the activity ofSRB on former magnetic grains.

0.E+00

1.E-07

2.E-07

3.E-07

4.E-07

5.E-07

6.E-07

7.E-07

8.E-07

0 50 100 150 200 250

Days

Remanentmagnetization(Am2/Kg)

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

Counting

SRB-free EST-211-675

SRB-rich EST-211-675

SRB counting

EST-211-845 m (C2b2)

0.0E+00

5.0E-07

1.0E-06

0 20 40 60 80 100 120

Days

Magnetization(Am2/Kg)

1.E+04

1.E+05

1.E+06

1.E+07

Counting

SRB-rich EST-211 845

SRB-free EST-211 845

SRB counting

Figure 2: SEM picture ofSRB-rich natural claystonesafter 10 months of bacterialactivity. One can see void onthe left which is likely theremnant of a framboidal pyriteafter alteration. Small filamentsare not yet understood and theylikely result from SRB activity.

P/GM/29


INFLUENCE OF EDTAAND ISOSACCHARINATE ORGANIC

LIGANDS ON THE UPTAKEAND MIGRATION OF EUROPIUM IN THE

CALLOVO-OXFORDIAN ARGILLITEJ. Klein and I. Pointeau

Laboratoire de Mesures et Modélisation de la Migration des Radionucléides, Service d’Etude duComportement des Radionucléides, Département de Physico-Chimie. CEA Saclay91191 Gif-sur-Yvette, France.

INTRODUCTIONIn the design of deep long-lived intermediate level waste repositories, argillaceous rock is considered aspotential host medium for the far-field environment and cement is an important component of the near-field materials. Waste package organic matters and their degradation products can alterate the uptake andthe transport of radionuclides (RN) in these geological and engineered barriers. Until now, literature studieshave been focused on the influence of the organic ligands on the RN solubility and distribution coefficient(Kd) values in cementitious environment. For example, results of Wieland et al. (2002) indicate thatisosaccharinate (ISA), the main degradation product of cellulose under alkaline environment, is a strongcomplexant of actinides in cementitious systems.

The objective of the present work was to investigate the effect of the complexing effect of these organicligands on the uptake and migration of RN in clay environments as it has not been studied yet.

A preliminary set of calculations was performed toevaluate the decrease of Eu(III) Kd values as afunction of organic concentration. Complexformation in solution and RN uptake by argillaceousrock have been considered as competing reactions.Available thermodynamic complexing constantvalues have been selected in literature for furtherorganic ligands and the results are expressed infigure 1.

A first set of batch sorption experiments has beenperformed in order to check the evolution of Kdvalues as a function of EDTA and ISA concentra-tion. Callovo-Oxfordian argillite (COx) has beenused as it is the geological rock present in theMeuse/Haute-Marne underground research labora-tory. Then, through-diffusion experiments will beperformed in order to study the influence on the complexation of Eu(III) by organic on the Eu(III) migra-tion parameters (distribution coefficient).

EXPERIMENTAL WORKThe argillite used in this study is issued from a COx core sampled at a depth of (-500.8)/(-501.0) m in the“Est207 borehole”. Slices of 1cm have been sawed for through-diffusion experiments and the other part ofthe core has been manually crushed and screened. The fraction inferior to 63 ìm has been characterized byXRD and BET, washed several times with the artificial COx groundwater, and then used for batch sorption

Figure 1: Evolution of calculated Kd values ofEu(III) as a function of organic concentration. It canbe noted that available complexing constants ofEu(III) with ISA have only been determined in theliterature in high pH range.

1.E-09 1.E-08 1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02

[Orga]

EDTA

Gluconate

ISA

Acetate

Oxalate

KdEu,mL/g

P/GM/29



experiments. The artificial COx groundwater has been synthesized with a final pH of 7.35 and equilibratedwith a pCO2 of -2.15.

Kd values have been measured for152Eu, 14C-EDTA and 14C-ISA with different concentrations and Kd(Eu)

values have been measured as a function of organic concentrations. The 14C-ISA tracer has beensynthesised in this study applying Whisthler and Bemiller (1963) protocol with a 14C-lactose solution.

RESULTS AND INTERPRETATIONThe initial Kd value of Eu(III) in COx (Kd°) was initially estimated with literature values on illite (about4.105mL/g, from Poinssot et al., 1999) corrected from the complexing effect of HCO3

-, SO42- and H3SiO4.

The calculated value is lower than the measured value and it can be supposed that Eu(III) inorganiccomplexes might sorb on the COx surface.

The measured feature of Kd(Eu) values vs. [EDTA] is in good agreement with the estimated Kd valuesobtained by correcting the Kd° value with a reducing factor including thermodynamic complexing data. Itis not the case for Kd(Eu) vs. [ISA], where the calculated effect of the organic is underestimated. It maybe due to the lack of thermodynamic data for Eu:ISA complexes in the COx pH range.

The uptake of EDTA and ISA has also been investigated and through-diffusion experiments are in progressfor both organic ligands. An experiment is also running with Eu(III), strongly complexed by EDTA inorder to decrease the Kd value to ~ 5 mL/g.

References:Poinssot, C., Baeyens, B., and Bradbury, M.H. (1999) Experimental studies of Sr, Ni, Eu sorption on Illiteand the modeling of Cs sorption. Nagra Technical Report NTB 99-04.

Wieland, E., Tits, J., Dobler, J.P., and Spieler, P. (2002) The effect of á-isosaccharinic acid on the stabilityof and Th(IV) uptake by hardened cement paste. Radiochim. Acta 90, 683-688.

Whistler, R.L., BeMiller, J.N. (1963) ‘‘á’’-D-Isosaccharino-1,4-lactone. Action of lime water on lactose.In: Wolfrom, M.L., BeMillers, J.N. (Eds.), Methods in Carbohydrate Chemistry. Reactions ofCarbohydrates, vol. 2. Academic Press, New York, 477–479.

ACKNOWLEGEMENTS:The authors would like to thank F.Goutelard and M.Descostes for fruitful discussions and B.Grenut for hiscontribution to the rock sample cutting. This study has been financially supported by ANDRA.

P/GM/30


EXPERIMENTAL DETERMINATIONOF THERMODYNAMIC PROPERTIES

OF A CHLORITE

H. Gailhanou1, J. Rogez2, J.C. van Miltenburg3, A. van Genderen3, J.M. Grenèche4,E.C. Gaucher1, C. Crouzet1, S. Touzelet1 and P. Blanc1.

1. BRGM – The French Geological Survey – 3 av. Claude Guillemin BP6009 45060 Orléans cedex2, France ([email protected])

2. TECSEN – Université Paul Cézanne, av. Escadrille Normandie-Niemen, 13397 Marseille cedex20, France ([email protected])

3. Chemical Thermodynamics Group, Utrecht University, Padualaan 8, 3584 CH Utrecht, TheNetherlands ([email protected])

4. LPEC - Laboratoire de Physique de l’Etat Condensé – Université du Maine, av. OlivierMessiaen, UFR Sciences et Techniques, 72085 Le Mans cedex 9, France([email protected])

INTRODUCTIONIn the context of nuclear waste repositories in argillaceous formations, it is necessary to know thegeochemical behaviour of natural and engineered clay barriers considering long-term periods. However,thermodynamic data of clay minerals, which are required for geochemical modelling, are rare. In particular,there is a lack of data for chlorite minerals. Chlorite is present in the Callovo-Oxfordian formation(Gaucher et al., 2004), and may precipitate due to the alteration of smectite in contact with iron from steelcontainers (Guillaume et al., 2003). Chlorite may also precipitate due to cement/clay interactions (Gaucherand Blanc, 2006).

In this study, all the thermodynamic properties of a natural trioctahedral chlorite were determined, usingcalorimetric methods, between 0 K and 520 K. This method was previously used for studying illite andsmectite minerals (Gailhanou, 2005; Gailhanou et al., 2007). In parallel, solubility experiments werecarried out in order to determine the solubility product of the chlorite and to compare it to calorimetricresults. All these new data will be included in the ANDRA database THERMOCHIMIE in progress.

MATERIALSThe chlorite sample is the internationally referenced ripidolite CCa-2, from Flagstaff Hill (California, USA;Post and Plummer, 1972), provided by the Clay Mineral Society. No impurity was detected by X-rayDiffraction; only a few TiO2 impurities were observed by Transmission ElectronMicroscopy. The chemicalcomposition of the ripidolite was obtained using X-ray fluorescence spectrometry. The Fe2+/Fe3+ rate wasdetermined both by 57Fe Mössbauer spectroscopy and by wet chemical procedure, with a good agreementbetween the two results.Microprobe analyses showed evidence of the chemical homogeneity of the chloritesample. Finally, the mean structural formula of the ripidolite CCa-2 is (Si2.633Al1.367)(Al1.116Mg2.952Fe

2+1.712Fe

3+0.215Mn0.012Ca0.011)O10(OH)8. Before calorimetric measurements, the chlorite sample

was dehydrated at 120°C during 24 hours.

CALORIMETRIC RESULTSEnthalpy of formation of the mineral was determined from enthalpies of dissolution in a HF-HNO3 solution,at 25°C, of (i) the sample (mineral + impurities) and (ii) the mixing of oxide or hydroxide constituents (ofthe mineral) + impurities.

Heat capacity measurements were realized using adiabatic calorimetry (5 K - 380 K) and differentialscanning calorimetry (300 K – 520 K). Heat capacities were extrapolated to 0 K using the Debye

P/GM/30



approximation at low temperatures. The Cp(T) curve shows an anomaly with a small peak at 377 K, whichmay correspond to a lambda-type transition. The enthalpy of transition is estimated to 58.5 J.mol-1.K-1.

From heat capacity measurements, entropy and heat content of ripidolite were calculated between 0 K and520 K. The standard entropy of formation of the mineral was then determined from the entropies of themineral and of the elements.

Finally, Gibbs free energy of formation of the chlorite was obtained from enthalpy and entropy offormation, at any temperature. The thermodynamic data of the ripidolite CCa-2 at 1 bar and 298.15 K aregiven in table 1. These values are close to the data obtained for another ripidolite (Wolery and Daveler,1992) and a clinochlore (Helgeson et al., 1978) (Table 1).

ReferencesGailhanou H. (2005) Experimental determination of thermodynamic properties and study of nanostructuresof clay minerals. Ph. D. Thesis, Aix-Marseille III Univ, 262pp.

Gailhanou H., Miltenburg van J.C., Rogez J., Olives J., Amouric M., Gaucher E.C. and Blanc P. (2007)Thermodynamic properties of anhydrous illite IMt-2, smectite MX-80 and mixed-layer illite-smectiteISCz-1 by calorimetric methods. Part I. Heat capacities, heat contents and entropies. Submitted toGeochim. Cosmochim. Acta.

Guillaume D., Neaman A., Cathelineau M., Mosser-Ruck R., Peiffert C., Abdelmoula M., Dubessy J.,Villiéras F., Baronnet A. and Michau N. (2003) Experimental synthesis of chlorite from smectite at300°C in the presence of metallic iron. Clay Minerals, 38, 281-302.

Gaucher E.C., Robelin C., Matray J.M., Negrel G., Gros Y., Heitz J.F., Vinsot A., Rebours H.,Cassagnabère A., Bouchet A. (2004) ANDRA underground research laboratory: interpretation of themineralogical and geochemical data acquired in the callovian-oxfordian formation by investigativedrilling. Physics and Chemistry of the Earth Journal, 29, 55-77.

Gaucher E.C. and Blanc P. (2006) Cement/clay interactions - A review: Experiments, natural analoguesand modelling. Waste Management, 26, 7, 776-788.

Helgeson H.C., Delany J.M., Nesbitt H.W. and Bird D.K. (1978) Summary and Critique of thethermodynamic Properties of Rock-forming Minerals, Am. Jour. Sci., 278A, 229 p.

Post, J.L., and Plummer, C.C. (1972) Chlorite series of the Flagstaff Hill area, California: a preliminaryinvestigation, Clays and Clay Minerals, 20, p.271-283.

Wolery T.J. and Daveler S.A. (1992) EQ3/6, a software package for geochemical modelling of aqueoussystems. Lawrence Livermore National Laboratory, UCRL-MA-110772PT I-IV.

Table 1: Thermodynamic data of ripidolite CCa-2 and two other trioctahedral chlorites(ripidolite (Wolery and Daveler, 1992) and clinochlore (Helgeson et al., 1978)),

at 1 bar and 298.15 K.

Cp°J.mol-1.K-1

S°J.mol-1.K-1

ΔS°fJ.mol-1.K-1

ΔH°fkJ.mol-1

ΔG°fkJ.mol-1

Ripidolite CCa-2 546.09± 0.16

468.40± 0.31

-2170± 2

-8253.70± 22.75

-7607.1± 23.4

Ripidolite (Mg3Fe2Al)(Si3Al)O10(OH)8 -7518.6

Clinochlore (Mg5Al)(Si3Al)O10(OH)8 465.3 -8856.2 -8207.8

P/GM/31


STRUCTURAL INCORPORATIONOF TRIVALENT f ELEMENTS

INTO THE TRIOCTAHEDRAL CLAYMINERAL HECTORITE

N. Finck1, K. Dardenne1, M. L. Schlegel2, D. Bosbach1

1. Institut für Nukleare Entsorgung, Forschungszentrum Karlsruhe, Postfach 3640, D-76021Karlsruhe, Germany.

2. DEN/DPC/SCP Centre d’Etudes Nucléaires de Saclay, BP 11, 91191Gif sur Yvette, France.

SCIENTIFIC BACKGROUNDClay formations are contemplated as potential host rocks for high-level nuclear waste repository in severalcountries. Also, clay-based materials may be used to build geotechnical barriers surrounding the wastecanister [1]. They have a high affinity for trivalent actinides, and several distinct molecular levelmechanisms of actinide retention can operate, as shown Figure 1: outer-sphere and inner-spherecomplexation, cation-exchange (interlayer), and structural incorporation. These mechanisms differ not onlyin structural location of the sorbate ion, but also in terms of long-term reversibility. For performanceassessment, it is therefore of paramount importance to identify actual retention mechanisms to predictradionuclide immobilization and remobilization.

Radionuclide immobilization by incorporation into the bulk structure of clay minerals may occur viacoprecipitation. Radionuclide trapped in structural sites of the clay octahedral sheet would be effectively“blocked” from further migration within the geosphere. However, the size mismatch (e.g. Cm(III)[VI]= 0.97Å [2]) with the cations which typically occur in the octahedral sites of sheet silicates (e.g. Mg2+, Fe2+, Al3+,Fe3+; r = 0.75 Å) would result in large lattice strains which are energetically unfavourable [3]. From acrystal chemical point of view, trivalent actinides could occupy a 6-fold oxygen coordinated lattice site(Pauling’s first rule). Hectorite, a trioctahedral smectite, is one of the secondary phases identified withinthe alteration layer of corroded HLW glass, which may represent a significant retention potential forradionuclides. Recent time-resolved laser fluorescence spectroscopy (TRLFS) data for Eu(III)/Cm(III)

Figure 1: Molecular level sorption mechanisms for trivalent f elements and clay minerals.

P/GM/31



coprecipitated with hectorite at 90˚C [4] suggest that such a substitution mechanism can operate [5,6]: dueto the sensitivity of the Cm(III) emission to changes in the ligand field, distinct steps during hectoriteformation were characterized.

Conventional EXAFS spectra collected on powders of Eu-containing hectorite allowed to identify thepresence of neighbouring oxygen shells at distances suggesting that Eu(III) was hexacoordinated withoxygens, as in a Mg structural site [7]. Neighbouring structural cations were not detected in these powderEXAFS spectra, maybe due to cancellation effects between EXAFS waves backscatterred by out-of-planeSi and in-plane (Mg) cations [8]. As the clay minerals are oriented through their layered structure,Polarized-EXAFS (P-EXAFS) experiments on self-supported films of these oriented clay minerals werecarried out to confirm this interpretation.

The principle of P-EXAFS measurements is detailed in several recent works [8-10]. The amplitude of theEXAFS contribution for a given absorber-backscatterer atomic pair depends on its angle with the electricfield vector E: it is enhanced for E parallel to the absorber-backscatterer pair and weakened for Eperpendicular to this pair. Consequently, contribution of neighbouring cations can be selected by changingthe angle á between E and the particle plane. For example, for Eu located at Mg structural sites, an angleá = 0˚ magnifies the EXAFS contribution of the in-plane octahedral Mg neighbours at ~3.1 Å and the Sishell is weakened. Conversely, the contribution of the more distant Si shell at ~3.3 Å is magnified for á =90˚ and the Mg contribution disappears.

RESULTSEuropium- and lutetium-containing organo-hectorite have been synthesised for varying lanthanide contents,as surrogates for trivalent actinides, in order to determine the influence of the size mismatch (Eu(III)[VI] =0.95 Å, Lu(III)[VI]= 0.86 Å [3]) and of the degree of substitution on the local crystal structure in a first part.EXAFS spectra were collected for several values of á, and for different temperatures (15K, 298K). Thequantitative information on the bond lengths allows us to estimate the structural compatibility of Eu(III)/Lu(III) for the Mg sites, and consequently yield some structural information required to assess thethermodynamic stability of lanthanides/actinides incorporation. The identification and characterization ofthe incorporation mechanism of trivalent lanthanides/actinides helps to complete the molecular-levelunderstanding of sorption reactions on clay minerals. This will definitely improve the predictive quality ofthe migration behaviour of radionuclides with respect to long-term safety aspects of a nuclear wasterepository system.

References[1] D. Mallants, J. Marivoet, et al., J. Nucl. Mater. 298, 125-135 (2001).

[2] R.D. Shannon, Acta Cryst. A32, 751-767 (1976).

[3] N.L. Allan, J.D. Blundy et al., Solid-Solutions in Silicates and Oxides, EMU Notes in Mineralogy, Vol.3, Eötvös University Press, Budapest.

[4] K.A. Carrado, L. Xu et al., Chem. Mater. 12, 3052-3059 (2000).

[5] H. Brandt, D. Bosbach et al., Geochim. Cosmochim. Acta 71 (1), 145-154 (2007).

[6] H. Pieper, D. Bosbach et al., Clays Clay Miner. 54, 47-55 (2006).

[7] H. Pieper, Wissenschaftliche Berichte FZKA 7188, Forschungszentrum Karlsruhe, Germany (2005).

[8] M.L. Schlegel, A. Manceau et al., J. Colloid Interface Sci. 215, 140-158 (1999).

[9] M.L. Schlegel, A. Manceau, et al., Geochim. Cosmochim. Acta 65, 4155-4170 (2001).

[10] A. Manceau and M.L. Schlegel, Phys. Chem. Minerals 28, 52-56 (2000).

P/GM/32


HEAVY METALS MIGRATIONIN ARGILACEOUS ROCKS: ON THE USE

OF LASER-INDUCED BREAKDOWNSPECTROSCOPY MICROPROBE

(LIBS MICROPROBE)AS A MICROANALYSIS TOOL

Menut D., Lacour J.L., Salmon L.

CEA Saclay, Nuclear Energy Division, DPC/ SCP/ LRSI, Bâtiment 391 point courier 33, 91191 Gifsur Yvette, France

INTRODUCTIONA thorough assessment of the long-term behaviour of deep nuclear repositories entails an accurateprediction of the fate for radionuclides in geological formations. In the argillite formation probed by Franceto host such a repository, diffusion is assumed to be the main transport mechanism governing radionuclidesmigration. Consequently, macroscopic through-diffusion experiments were conducted to determine thetransport mechanisms [1]. However, due to the intrinsic mineralogical complexity of the clay material, theparameters of these studies are often difficult to interpret. This difficulty can be circumvented by the useof microscopic techniques such as the LIBS microprobe, which allow to perform local spectrochemicalanalysis of nonconductive sample. Here, we show how this alternative microanalytical technique canbe to confirm, the observations obtained by through diffusion technique [1].

EXPERIMENTAL DEVELOPMENTThe Laser-Induced Breakdown Spectroscopy (LIBS) technique has recently been developed as a versatileand sensitive probe for spectrochemical analysis of various materials. The study of elemental compositionof material surfaces by LIBS is based on the analysis of the optical emission from plasma created by afocused laser beam. The distribution of elements can be obtained by scanning the laser beam over theinvestigated surface. Unlike many other conventional analytical techniques, this laser-based technique hasthe advantages of being applicable to any possible object and providing a non-contact, rapid, multi-elementanalysis under a wide variety of ambient conditions with minimal sample preparation. The micro LIBSequipment is the combination of a laser microprobe, a powerful XY driver that can move quickly andprecisely which is used at 20 Hz (corresponding to the laser shot frequency) with a 3-µm displacementbetween consecutive laser shots. The system is shown in figure 1.

NEW PERSPECTIVES TO EVIDENCE HEAVY METALS DIFFUSION PROFILEIN HIGHLY HETEROGENEOUS MEDIAAs a example of the unique potential of LIBS microprobe analysis to address fundamental issues,migrations fronts in clay probes were probed for qualitative and quantitative analysis. Clay coupons wereimmersed in different content heavy metals solutions, cross-cut, and further subject to a LIBS microprobeanalysis to investigate the diffusions and mineralogical affinities of the heavy metals tracers. Quantitativeelemental mapping were made by LIBS microprobe that showed the actual distribution of the various microareas observed on the Callovo-Oxfordian sample’s surfaces. By images processing, the most relevantmineralogical phases were identified and their statistical distribution was determined. Averaging the tracerconcentrations parallel to the diffusion front yield a diffusion profile, from which diffusion parameterscould be readily obtained (figure 2) [3].

Some new perspectives to reveal heavy metals diffusion profile in highly heterogeneous media, asargillaceous rocks, and, maybe, cements and concrete are opened with the use of the LIBS microprobe. The

P/GM/32



LIBS microprobe arrangement developed provides rapid access to quantitative surface mapping of multipleelements for non-conductive samples. Statistical studies of the results obtained with high resolution overa wide area can be made, and information such as aggregation modal repartition and chemical compositionscan be deduced. Diffusion phenomena and spectroscopic findings by LIBS microprobe will play a key rolewith actual numerical modelling codes and their validations.

References:[1] M. Descostes, V. Blin, B. Grenut, P. Meier, E. Tevissen: HTO Diffusion in Oxfordian Limestone andCallovo-Oxfordian Argillite Formation.. Scientific Basis for Nuclear Waste Management XXVIII,MRS Spring Meeting, San Francisco, USA (2004).

[2] D. Menut, P. Fichet, J.L. Lacour, A. Rivoallan, P. Mauchien: Micro Laser-Induced BreakdownSpectroscopy technique : a powerful method for performing quantitative surface mapping onconductive and nonconductive sample. Appl. Opt. Vol 42 (30): 6063-6071.

[3] Menut D., Descostes M., Meier P., Radwan J., Mauchien P., Poinssot P. : Europium migration inargilaceous rocks : on the use of micro Laser-Induced Breakdown Spectroscopy (micro LIBS) as amicroanalysis tool. Scientific Basis for Nuclear Waste Management XXIX, Symposium ProceedingsVol. 932: 913-919.

Figure 1: General view of the LIBS microprobe experiment set-up.

Figure 2: Eu diffusion profil with erfc fit as anexemple.

P/GM/33


ISOTOPIC ANOMALIES OBSERVEDAT THE VICINITY OF FRACTURES

IN POREWATER OF TOURNEMIRE SHALES:EXPERIMENTAL ARTEFACTS

OR LOCAL PALAEO-CIRCULATIONS?S. Savoye1, J.L. Michelot2, M.V. Altinier1,2, S. Lemius1

1. IRSN - Institut de Radioprotection et de Sûreté Nucléaire, Av. du Gen. Leclerc BP n°17, 92262Fontenay-aux-roses, France([email protected], [email protected], [email protected])

2. UMR “IDES” CNRS-Université de Paris-Sud, Bât 504,91405 Orsay, France([email protected])

INTRODUCTIONPatriarche et al. (2004) acquired in the Toarcian argillaceous formation of Tournemire a vertical profile ofstable isotope content of interstitial water. They found a good agreement between these experimental dataand calculated values obtained from a pure diffusion model except for samples collected less than onemeter from fractures. Samples closely located to fractures displayed a systematic increase of their waterstable isotope contents. Two types of hypotheses were proposed for accounting for these discrepancies:(i) some local circulations could have occurred around the fractures and may have affected stable isotopeconcentrations and (ii) the vacuum distillation technique used for determining the isotope contents mayhave induced experimental artefacts on these particular samples.

In the present study, the influence of fractures on the stable isotope contents was addressed by acquiringan isotopic profile along a fracture by means of several techniques (extraction and equilibrationapproaches) coupled with mineralogical and petrophysical characterisations of rock samples.

EXPERIMENTAL CONCEPTCore samples were collected from an horizontal air-drilled borehole intercepting a sub-vertical fracture.Two methods were applied for determining the water stable isotope contents in pore water (see Altinier etal., 2007, for details). The vapour exchange method consists in equilibrating by diffusion the pore-waterof crushed sample with a spiked solution in water-tight container. In the other hand, we performed vacuumdistillation technique at two extraction temperatures (at 50°C, as Patriarche et al. did and at 150°C). At last,several petrophysical determinations were carried out: water contents by oven-drying at 105°C and 150°C,bulk densities by kerosene immersion and grain densities by helium pycnometry. Mineralogicalcharacterisations were performed by means of XRD analyses and total chemistry.

RESULTS AND INTERPRETATIONSFig. 1 shows the oxygen-18 content of pore water as a function of the distance to the fracture. Dataobtained from vapour exchange experiments display a slight scattering, which is not linked to the presenceof the fracture. The same trend can be highlighted from data obtained from distillation at 150°C. Inversely,the profile derived from distillation at 50°C displays a clear increase of oxygen-18 contents at less thanone meter from the fracture. These results suggest that the discrepancies observed by Patriarche et al.(2004) would be due to artefacts induced by vacuum distillation technique at 50°C and not to naturalphenomena associated to the presence of the fracture.

In order to try answering the question of the artefact origin, the variation of several mineralogical andpetrophysical parameters was studied. Neither the mineralogy of the bulk rock nor the composition of theclay fraction display any change as a function of the distance to fracture.

P/GM/33



In fact, the increase of water content at the vicinity of fracture (Fig. 2), which is also associated with anincrease of the proportion of large pores in samples (from 10 nm to 180 nm, BET-N2 analyses) couldaccount for the isotopic anomaly observed on data derived from vacuum distillation at 50°C. This higherproportion of large pores at the vicinity of the fracture could indeed make the water extraction easier,limiting the effects of incomplete distillation. This would lead to obtain water isotope contents close tothose derived from both vacuum distillation at high temperature and vapour exchange method.

ReferencesAltinier, M.V., Savoye, S., Michelot, J.L., Beaucaire, C., Massault, M., Tessier, D. & Waber, H.N. (2007):The isotopic composition of argillaceous-rocks pore water: an intercomparison study on the Tournemireargillite (France). Physics and Chemistry of the Earth, 32, 209-218.

Patriarche, D., Ledoux, E., Michelot, J.L., Simon-Coincon, R. & Savoye, S. (2004): Diffusion as the mainprocess for mass transport in very low water content argillites Part 2. Fluid flow and mass transportmodeling.Water Resources Res., 40, W01517, 2004.

Figure 1: Profile of δ18O as a function of the distance to the fracture.

Figure 2: Profile of water content as a function of the distance to the fracture.

-12

-11

-10

-9

-8

-6

-5

-360 -330 -300 -270 -240 -210 -180 -150 -120 -90 -60 -30 0 30 60 90 120

Distance to the fracture (cm)

Vapour diffusive exchange Vacuum distillation at 150°CVacuum distillation at 50°C Fracture

to the head

borehole

-7

O(‰

vsSMOW)

3.6

3.7

3.8

3.9

4.0

4.1

4.2

4.3

4.4

-350 -300 -250 -200 -150 -100 -50 0 50 100

W 105°C

W 150°C

Distance to the fracture (cm)

Watercontent(wt-%)

P/GM/34


EFFECT OF TEMPERATUREON THE RETENTION CAPACITYOF COMPACTED BENTONITE:

AN EXPERIMENTAL AND NUMERICALINVESTIGATION

A. Jacinto1, R. Gómez-Espina2, M.V. Villar2, A. Ledesma1

1. Tech. University of Catalunya – UPC, Jordi Girona, 1-3 (D2), 08034 Barcelona, Spain,([email protected]; [email protected])

2. CIEMAT – Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas, Avd.Complutense 22, 28040 Madrid, Spain ([email protected]; [email protected])

INTRODUCTIONA current design for engineered barriers in the context of high level radioactive waste includesbentonite compacted blocks initially unsaturated. The heat released by the waste may induce hightemperatures in the bentonite barrier. As an example, in the Temperature Buffer Test (TBT) valueswell above 100°C are reached in the bentonite rings surrounding the heaters (Åkesson et al, 2007, thisConference), in order to simulate real conditions in a particular disposal concept. The work presentedhere investigates a fundamental issue concerning the influence of high temperatures on the retentioncapacity of compacted bentonite. Previous works present water retention curves of bentonite measuredfor different densities, and traditionally under free volume conditions. It has been shown, however,that the experiments should be performed under constant volume conditions in order to simulate theconfinement of the barrier (Villar et al. 2005). In addition to that, the effect of temperature should beconsidered as an independent variable. Thus retention properties not only depend on density but alsoon temperature.

The material used in this investigation is the MX-80 bentonite, compacted to high densities with differentinitial water contents and tested at temperatures ranging from 20 to 100°C. The results have beenananalysed and compared with predictions based on a capillary model. Some discrepancies betweenmeasurements and predictions are discussed and explained in the paper. Also, several 1D THM analysesusing the computer code “Code_Bright” have been performed in order to analyse the consequences ofconsidering or not temperature effects in the water retention curve.

EXPERIMENTAL RESULTSThe MX-80 powder has been mixed with different quantities of deionised water and, after equilibration,blocks have been manufactured by uniaxial compaction to dry densities 1.5, 1.6 and 1.7 g/cm3. Theseblocks have been placed in stainless steel hermetic cells that can be heated and in which a hygrometerinserted in the block allows the measurement of the relative humidity, which is converted to suction bytaking into account the temperature, which is also measured.

For each dry density and water content, the suctions obtained are lower the higher the temperature,especially when the water content is low. Also, for a given water content and temperature, the suctionobtained is slightly higher the higher the dry density. However, this trend clearly inverts for suctions below1 MPa. Some of the results are plotted in Figure 1.

MODELLING RESULTSThe decrease of suction as temperature increases in a soil is usually explained by the decrease of surfacetension with temperature, as suggested by Philip & de Vries (1957). This change has been computed for

P/GM/34



each sample tested and some of the results are plotted in Figure 2, together with the experimental points.It becomes clear that the change of suction with temperature is higher than the one predicted by the changein surface tension (capillary model). Other additional processes invoked to explain the change of suctionwith temperature are the changes in the wetting coefficient caused by variations in the contact angle, thethermal expansion of the trapped air, and the evacuation of the air trapped caused by heating. Besides, inclayey soils heating is considered to trigger the transference of microstructural water to the macrostructure.

Several THM analysis of a simple 1D geometry have been performed taken into account temperatureeffects on the water retention curve. To do that, a modification of the van Genuchten expression includingtemperature effects as observed in this experiments, is proposed. That expression was incorporated into thecode “Code_Bright”. The results indicate that when temperatures reach values above 100°C, temperatureeffects on retention properties should be taken into account. In particular, hydraulic variables (suction orrelative humidity) obtained in the analyses are sensitive to those effects.

References:Åkesson, M., Jacinto, A., Ledesma, A., Combarieu, M. (2007). Temperature Buffer Test: experimental and

model results. This Conference.

Philip J.R., de Vries D.A. (1957). Moisture movement in porous materials under temperature gradients.Trans. Am. Geophys. Union, 38 (2), 222-232.

Villar, M.V., Martín, P.L. & Lloret, A. (2005): Determination of water retention curves of two bentonites athigh temperature. In: Tarantino, A., Romero, E. & Cui, Y.J. (eds.): Advanced experimental unsaturatedsoil mechanics. EXPERUS 2005. A.A. Balkema Publishers. London, pp 77-82.

Figure 1: Retention curves for the MX-80 bentonite compacted at different dry densities (indicated ing/cm3) and obtained at 40°C (left) and 100°C (right).

Figure 2: Change of suction withtemperature for the MX-80 claycompacted at dry density 1.6 g/cm3 withdifferent water contents (indicated in %).The predictions of the capillary model areindicated with dotted lines. The slope ofthe lines relating suction and temperatureare shown.

0

50

100

150

200

250

3 8 13 18 23

Water content (%)

Suction(MPa)

39, 1.75

41, 1.60

40, 1.50

0

50

100

150

200

250

3 8 13 18 23

Water content (%)

Suction(MPa)

99, 1.75

101, 1.60

100, 1.50

-0.3731

-0.2455

-0.2821

-0.2390

0

20

40

60

80

100

20 40 60 80 100 120 140

Temperature (°C)

Suction(MPa)

11 13

15 19-0.1867

-0.0426

-0.1413

-0.0880

P/GM/35


RETENTION OF CS IN BOOM CLAY:COMPARISON OF DATA FROM BATCH

SORPTION TESTS AND DIFFUSIONEXPERIMENTS ON CLAY CORES

N.Maes1, S. Salah1, D. Jacques1, M. Aertsens1, M. VanGompel1, P. De Cannière1, N. Velitchkova2

1. SCK•CEN, Boeretang 200, BE-2400 Mol, Belgium (corresponding author: [email protected]).2. Bulgarian Academy of Science – Geological Institute, Acad. G. Bontchev St., Bl. 24, BU-1113 Sofia,

Bulgaria.

INTRODUCTIONThe Long-term safety of a geological disposal of high-level waste (HLW) and spent fuel in argillaceousformations relies very much on the retardation safety function of the clay host formation. Sorption and ion-exchange are the main mechanisms that ensure this retardation. Two distinct ways can be used to obtaintransport parameters: migration experiments on non-reshaped clay cores enable to determine the retardationfactor (R) while batch-sorption experiments on clay suspensions enable to determine the distributioncoefficient (Kd or Rd). In transport calculations, distribution coefficients obtained from batch-sorptionexperiments are indeed commonly used to account for the retardation of sorbed species at lowconcentration. However, the conversion of the distribution coefficients determined on diluted claysuspensions in terms of retardation factors truly applicable to the transport in compacted clay is still amatter of debate. The objective of this study is to compare caesium sorption data obtained from Boom Claysuspensions with transport parameters determined from long-term migration experiments on clay cores inorder to evaluate if this conversion is justified in low permeability clay.

EXPERIMENTALA Cs adsorption isotherm on Boom Clay in equilibrium with Boom Clay porewater(~ 0.014 mol dm-3 HCO3

-) is determined under simulated in situ conditions (pH = 8.2, Ar / 0.4 % CO2oxygen-free atmosphere) at a solid-to-liquid ratio of 6.7 g dm-3. The equilibrium concentration of Cs+ wasdetermined after 14 days contact time. Cs+ migration parameters were obtained from percolationexperiments on compact clay cores: a known activity of Cs-137 was spiked on a cellulose filter sandwichedbetween 2 clay cores of a total length of 72 mm and 38 mm diameter. Boom Clay porewater was injectedthrough the clay cores at a pressure of ~ 1MPa and collected at the outlet for Cs-137 analysis by gammacounting. After 13.7 years, the clay cores were cut in thin slices (0.5 mm) and the Cs-137 activity profilein the clay was also determined via gamma analysis.

RESULTS AND DISCUSSIONThe concentration dependent uptake of Cs+ on Boom Clay could be successfully modelled (with theGeochemist’s Workbench and PhreeqC codes) using the generalized 3-sites cation-exchange model forillite developed by Bradbury and Baeyens (2000). Only at high Cs+ concentrations (above 5 × 10-4 mol dm-3) the modelled isotherm deviates from the experimental data. At trace concentrations (below 1 × 10-6mol dm-3) relevant for disposal conditions, the experimentally determined log Kd values rangebetween 3 and 4 (Kd in dm

3 kg-1), which could be well reproduced by the calculations. During the wholeduration (13.7 years) of the column migration experiments, no Cs-137 activity could be detected in theoutflowing water. At the end of the experiments, a migration profile with a Gaussian shape (dispersion) isobtained with a slight shift in the position of the maximum of the bell curve with respect to the initialCs-137 source location attributed to a minor contribution of advection. However, the Gaussian profile isslightly disturbed at the source position by an unknown artefact. This could be due to an unexpected effectof the paper filter on which the Cs-137 solution was spiked. By discarding these aberrant points for thefitting, the two remaining hemi-profiles could be successfully fitted and consistent parameter values

P/GM/35



calculated. The value of the apparent diffusion coefficient (Dapp) is obtained from the width of the Gaussiancurve while the ηR product (diffusion accessible porosity times retardation factor) depends on the peakdisplacement. The value of Dapp = 1.6 – 1.9 × 10-13 m2 s-1 is robust and consistent with other results obtainedfrom a previous in situ experiment (De Cannière et al., 1996) and a series of electromigration tests (Maeset al., 2001). The values obtained for ηR = 2 500 – 7 100 are however much less robust (error of ± 100 %)since the displacement of the profile is very small, and in order to obtain a reliable number, it should beat least larger than the resolution of slicing (0.5 mm, corresponding to ηR = 1 300). In transportcalculations of dissolved species, the following relationship is commonly used to relate the retardationfactor R to Kd (considering that sorption is a fast, linear, and reversible process):

(for Boom Clay: considering that the diffusion accessible porosity for Cs+ equals that

of HTO, η = 0.37, ρdry = 1.7 kg dm-3).

The log Kd derived from the ηR values of migration experiments, log Kd=3.2-3.6, are consistent with thevalues from the batch sorption tests, but the uncertainty on ηR is high. If we consider a minimum ηR valueof 1 300, corresponding to a displacement on the thickness of one clay slice, a minimum value of log Kdis 2.9.

The retardation factor R can also be estimated from the Dapp value with a few hypotheses:

with the diffusion coefficient of Cs+ in pure water, Daq = 2.05 × 10-9 m2 s-1; and considering the same rockfactor value as for HTO (Rf ~ 10, Maes et al., 2001) to globally account for the pore geometry effects

(tortuosity and constrictivity). This leads to values of R ~ 1 100 and log Kd ~ 2.4.

As the sorption of Cs+ onto Boom Clay is not linear, this commonly used relationship between R and Kdis in principle not valid for linear reversible sorption. Considering also that the Cs+ retardation factor (R)obtained with these migration experiments is subject to a large uncertainty, the values hereabove mentionedare rather consistent. We can however not draw firm conclusions whether batch-derived sorption data forCs+ can be extrapolated to compacted clay. To elucidate this, in-diffusion experiments, to obtain morereliable retardation factors, and simulations of the migration experiments with chemical-coupled transportcodes will be conducted.

ACKNOWLEDGEMENTSThis work is scientifically and financially supported by ONDRAF/NIRAS the Belgian NationalRadioactive Waste Agency and the European Commission in the frame of the 6th FP FUNMIG project.

ReferencesBradbury M. and Baeyens B. (2000) A generalised sorption model for the concentration dependent uptakeof caesium by argillaceous rocks. Journal of Contaminant Hydrology 42, 141–163.

De Cannière P., Moors H., Lolivier P., De Preter P., and Put M. (1996) Laboratory and in situ migrationexperiments in the Boom Clay. Nuclear Science and Technology report EUR16927EN, EC, Luxembourg.

Maes N., Moors H., Dierckx A., Aertsens M., Wang L., De Cannière P., and Put M. (2001) Studying themigration behaviour of radionuclides in Boom Clay by electromigration. In: Schriftenreihe AngewandteGeologie Karlsruhe Vol 63: EREM 2001 – 3rd symposium and status report on electrokineticremediation (Czurda K., Hötzl H., Eds.).

R 1ρdryKd

η----------------+=

RDaq

DappRf----------------=

P/GM/36


SORPTION AND ENGINEERINGCHARACTERISTICS OF SOME CLAY/SHALE

DEPOSITS FROM NIGERIAAS LANDFILL LINER

M.N. Tijani , O.P. Bolaji

Dept. of Geology, University of Ibadan, Ibadan, Nigeria

INTRODUCTIONUnlike the concerns for radioactive waste disposal in the developed countries, the concerns about theincreasing rate of solid waste production and the need for proper disposal and management to ensuresustainable environmental management and development has been the concerns of many developingcountries like Nigeria. Hence the need for properly engineered sanitary landfill is imperative for asustainable waste management and healthier environment, as the present common method of open dumpingcould expose the population to serious health hazards and the possible contamination of shallowgroundwater system. However, a major constraint to the development of engineered landfills is the highcost of synthetic liners and its scarcity in the local markets, thus there is the need for other readily availablesource materials for landfill liner. Hence, this study focuses on the assessment of sorption and engineeringproperties of selected shale/clay samples from Nigeria for possible usage as a landfill liner or as acomponent of a landfill barrier system.

MATERIALS AND METHODSThe shales used in this study were collected from five different locations within the Lokoja and AnambraBasins. The samples were air-dried, crushed, and pulverized in order to reduce the effect of clods onhydraulic conductivity. Subsequently, engineering tests, which included grain-size analysis, consistencylimit tests, compaction, and hydraulic conductivity tests were performed on the samples following theBritish Standard (BS 1377) for material testing. In addition, the cation exchange capacity (CEC) of thesamples were determined using the Ag-Tu method while sorption/uptake measurements using the batchmethod was also carried out, with respect to Cu2+, Zn2+, Mn2+, Pb2+ as the contaminant cations.

RESULTS AND DISCUSSIONThe results of sorption tests and CEC measurements performed on the clay/shale samples as presented inTable 1 reveal that two of the analyzed samples (L8 and L27) meet the minimum CEC requirement of 10cmol/kg as suggested by previous studies (Taha and Kabir, 2003), since high values of CEC result ingreater sorption/removal of contaminant from leachate flowing through the shale. Clay mineral type, theirrelative amounts in clay samples, as well as other factors like grain size usually define the magnitude ofCEC (Sezer et al., 2003). The values exhibited are typical for soil materials dominated by kaolinite. AlsoF-test confirms a significant correlation between the activity and CEC of the shales at 0.05 level (F = 4.52,Fcr =3.99), which is a clear indication of active influence of clay fractions on the sorption.

As presented in Fig. 1, heavy metal uptake follows the general order of Pb2+ > Zn2+> Cu2+> Mn2+, withaverage uptake of 20.58, 16.17, 14.97, and 14.11 mg/g respectively. However, the order Pb2+ > Zn2+> Mn2+

> Cu2+ for samples L8 and L27 imply lower uptake with respect to Cu which can be attributed to commonion effect, as these two samples exhibited higher geogenic concentration of Cu.

The results of the basic engineering properties of the crushed shales as shown in Table 1 revealed that allthe samples are of low to medium plastic silts (ML) according to the Unified Soil Classification System(USCS). In addition, the estimated activity of the clay fractions, based on grain size distribution and Atter-berg limits, range from 0.75 to 2.76, suggesting inactive to active clays according to Skempton classifica-tion. However, sample L8 and L27 clearly exhibit high activity, which is consistent with the respective

P/GM/36



higher CEC values of 10.71 and 10.14 cmol/kg respectively, Furthermore, the hydraulic conductivity (k)obtained from the samples at the Standard Proctor level and the West African level of compaction rangefrom 8.32 × 10-5 to 9.82 × 10-5 cm/s and 6.74 ×10-5 to 8.07 ×10-5 cm/s respectively. These values are sig-nificantly higher than the minimum value of 10-7 cm/s recommended for liner materials, which is a con-straint to direct useage of the study clay/shale materials as landfill liner or barrier system.

CONCLUSIONThis study had clearly shown the significance of both chemical sorption and engineering hydrauliccharacters of fine-grained materials as regard application as landfill liner or barrier system. The results andevaluation of engineering characteristics presented in this study have shown that the crushed shale/claysamples, generally, do not satisfy the basic engineering properties for clay liners in terms of the requiredhydraulic properties. However, the assessment of sorption characteristics revealed that samples L8 (Ahoko)and L27 (Enugu), with CEC values of 10.71 and 10.14 cmol/kg, exhibited acceptable uptake property withrespect to Pb, Zn, Cu, and Mn, and hence could be utilized as component of barrier design in sanitarylandfills subject to appropriate amendment technology (e.g. mixing with bentonite, cement and otherbinding materials) to enhance their hydraulic properties to the required specification. Consequently, thepositive environmental implication of this study lies in the fact that there is a high prospect of utilizationof clay/ clay-based materials in developing countries, as cheap landfill seal/liner materials.

ReferencesSezer, G.A., Turkmenoglu, A.G., Gokturk, E.H. (2003): Mineralogical and sorption characteristics ofAnkara Clay as a landfill liner. Journal of Applied Geochemistry, 18, 711-717.

Taha, M.R., Kabir, M.H. (2003): Sedimentary residual soils as a hydraulic barrier in waste containmentsystems. Second International Conference on Advances in Soft Soil Engineering Technology,Putrajaya, Malaysia, 895-904.

Table 1: CEC, Sorption/Uptake and basic engineering characteristics.

Sample CEC Zn Cu Mn Pb LL PL PI ActK

SP WA

L3 (Asheni)L8 (Ahoka)L11 (Ahoka-2)L19 (Ojodu)L27 (Enugu)

3.5210.714.133.8310.14

13.3117.4816.7017.6315.75

17.109.7818.7118.8110.44

17.6413.1111.5213.3114.97

20.5720.6120.5620.5920.57

43.134.738.724.038.0

29.526.125.921.024.3

13.68.612.82.9713.8

0.751.410.890.992.76

8.98.39.18.99.8

8.16.76.87.97.7

LL= Liquid limit (%); PL= Plastic limit (%); PI= Plasticity Index (%); Act. = Clay activityK= Hydraulic conductivity (cm/s); SP= Standard Proctor Compaction; WA= West African Compaction.

Figure 1: Profiles of Uptake, CEC and Activity of the analyzed samples.

0

5

10

15

20

25

L3 L8 L11 L19 L27

Samples

Uptake(mg/g)

0

5

10

15

20

%Clay&ActivityZn Cu Pb % Clay Activity

P/GM/37


INFLUENCE OF ORGANIC MATTERCOMPOSITION ON IODINE SORPTION:

FROM FRESH TO THERMALLYMATURED PEAT

Sophie Maillant1.2, Pierre Faure1, and Elisabeth Leclerc2,

1. G2R, Nancy-Université, CNRS, BP 239, 54506, Vandoeuvre-les-Nancy.2. Andra, Direction Scientifique, Service Transferts, 1-7 rue Jean Monnet, 92298 Châtenay-

Malabry, France.

The soil organic matter (OM) retains some organic and metallic contaminants. Most often, only thequantity of OM is known to influence the sorption of the element/molecule. In the case of iodine, thesorption has been related to some specific chemical characteristics of the OM. In deed, it has beensuggested that the phenolic groups of organic molecules are involved in the sorption of iodine onto OMin soils and geological materials [1]. To obtain more information on the nature of the bounds involved inretention of iodine, we assess the relation between the composition (elemental and molecular) of organicmaterials and their ability to sorb iodine.

The organic materials used in this study vary from fresh terrestrial organic matter (highly oxygenated) tothermally matured organic matter (highly aromatic). These materials are obtained through the progressive“thermal denaturation” of peat in high-pressure autoclaves (confined pyrolysis [2]). The initial peat is ablack peat sample from a boreal peat bog. The “denaturation” is carried out for 24 hours under pressure(700 bars) at six temperatures: 150°C, 200°C, 250°C, 300°C, 350°C and 400°C. Following this treatment,the chemistry of the material is then thoroughly analyzed: C, H, O contents, molecular composition of boththe extractable organic fraction (saturated and aromatic hydrocarbons and polar compounds by GC-MS)and the insoluble organic matter (by THM (Thermally assisted hydrolysis and methylation) – GC-MS),initially separated by solvent extraction. Spectroscopic characterisation is also carried out using FourierTransformed Micro-Infrared Spectroscopy on the extractable organic matter. These analyses (molecularand spectroscopic) allow the identification of oxygenated groups (such as phenolic and polyphenolicgroups) and aromatic structures and their evolution during the “thermal denaturation”. In parallel, sorptionexperiments are carried out on raw (initial to thermal matured peat) and pre-extracted matured (removal ofextractable organic matter by organic solvent) materials. The materials are shaken for 10 days with a KIsolution at a 1/10 solid to solution ratio. At the end of the contact period, the solutions are extracted andanalysed for total I content. A distribution coefficient (Kd) is then calculated to assess the sorption capacityof the material.

The “thermal denaturation” of peat produces gas and solid residues. The gas production increases slowlyup to 250°C and then sharply up to 400°C. Initial and pyrolysed material analyses show two trends. Firstly,the extract rate increases up to 200 mg/g of material at 300°C (oil window). This extract contains mainlypolar compounds (90% of the extract, figure 1). Beyond that temperature, the extraction rate decreasesdramatically. This extract contains mainly aliphatic hydrocarbons. Amounts of aromatic molecules aresignificants only at the highest temperature. Results are similar to those obtained by Yao et al. [3].

Efficiency of iodine sorption increases with the thermal maturation of the peat suggesting that sorption iscontrolled by the aromaticity of the organic matter. Moreover, sorption experiments carried out after theremoval of the extractable organic fraction (EOF) reveals a more important iodine sorption. Such resultssuggest a physical (increase of peat porosity after EOF removal which favour iodine/organic mattercontact) and/or chemical (solid residu more aromatic than EOF [4] and more able to sorb iodine) controlof the iodine sorption.

P/GM/37



References[1] J.A.Warner, W.H. Casey, and R.A. Dahlgren. Environmental science & technology. 34 (2000) 3180.

[2] M. Monthioux, P. Landais and B. Durand. Organic Geochemistry. 10 (1986) 299.

[3] S.Yao, C.Xue, W.Hu, J. Cao, and C. Zhang. Journal of Coal Geology. 66 (2006) 108.

[4] R. Michels, V. Burkle, L. Mansuy, E. Langlois, O. Ruau, and P. Landais. Energy & Fuel. 14 (2000)1059.

Figure 1: Molecular composition of the organic extract obtained after confined pyrolysis of peat deducedfrom GC-MS analyses.

0

10

20

30

40

50

60

70

80

90

100

Pe at 150 °C 200 °C 250 °C 300 °C 350 °C 400 °C

Aliphatic hydrocarbons (%)

Aromatic hydrocarbons (%)

Aliphatic polar compounds (%)

Aromatic polar compounds (%)

Organic extract

P/GM/38


OBSERVATION OF MICROSTRUCTUREOF COMPACTED BENTONITE

BY X-RAY MICRO CT METHOD OPTIMIZEDWITH COMPUTER SIMULATION

H. Takamatsu1, N. Noda1, T. Kozaki1, M. Kitaichi2, S. Tomioka2, S. Sato1

1. Division of Energy and Environmental Systems, Graduate School of Engineering, HokkaidoUniversity, Kita 13 Nishi 8, Sapporo 060-8628, Japan(corresponding author: [email protected])

2. Division of Quantum Science and Engineering, Graduate School of Engineering, HokkaidoUniversity, Kita 13 Nishi 8, Sapporo 060-8628, Japan

INTRODUCTIONCompacted bentonite is a candidate buffer material for high-level radioactive waste disposal in Japan.Microstructure of compacted bentonite is considered to influence radionuclide migration in the buffer. Toclarify the migration behavior of the radionuclide, nondestructive observation inside a bentonite sample isneeded. An X-ray micro CT method is one of the most powerful techniques for microstructural studybecause of its high spatial resolution [1]. However, in order to acquire fine cross-sectional images,experimental conditions, such as voltage applied to an X-ray tube and filtering to attenuate the low energyX-rays, are needed to be optimized with taking into account size and elemental composition of sample. Inthis study, the optimum conditions in the X-ray micro CT observation for the compacted bentonite wereestimated from the results of the computer simulation of X-ray transmission using a three-dimensionalMonte Carlo Code (EGS5 [2]). The conditions evaluated were examined in the observation, and fine cross-sectional images were obtained from the observation under the condition.

SIMULATION OF X-RAY TRANSMISSIONTransmission of X-ray in the compacted bentonite was simulated in regard to the X-ray micro CTexperiment. The simulation was performed by a Monte Carlo code, EGS5 (Electron Gamma Shower 5),which enables us to calculate electromagnetic cascade and interaction of charged particles in materials suchas simples, compounds, and mixtures. The virtual system in the simulation corresponds to the experimentalsetup of the X-ray micro CT apparatus under atmospheric condition; an X-ray beam of 1 x 1 mm2 in sectionfrom a source reaches a detector through a filter (Al or Ta) and a sample holder with bentonite sampleunder dry or water-saturated states. The dry density of the bentonite was 1.0 Mg m-3 and the sizes of thesample holder and the bentonite sample were the same as those in the experiment described below. Theeffect of the cavity on the X-ray transmission was evaluated for the bentonite samples with and without acavity cube of 1 mm3 at the center of bentonite sample. Moreover, effect of the filter (Al or Ta) on thetransmission and its dependence on the X-ray energy were examined.

X-RAY CT OBSERVATIONThe microstructures of compacted dry and water-saturated bentonites were observed with an X-ray microCT system, SkyScan-1172 (SKYSCAN), having about 1 ìm resolution under the best condition. It consistsof a sealed microfocus X-ray tube, 10M pixel cooled CCD fiber-optically coupled to scintillator, a rotatablestage to place a sample, and a data-processing system.

The bentonite samples used in this study was Kunipia F, Kunimine Industries Co., Ltd., Japan. The Na-bentonite power of 75-150 ìm in grains size was prepared by homoionization followed by sieving. The Na-bentonite powder was compacted into a sample holder (glassy carbon tube with 7-mm- outer and 5-mm-inner diameters, and with 10-mm high). A small amount of glass beads was introduced into the bentonite

P/GM/38



sample as reference. The water-saturation of the compacted sample was carried out by contacting thesample with de-ionized water through a sintered stainless filter.

RESULTS AND DISCUSSIONFrom the computer simulation with EGS5, it turned out that X-ray transmission in the lower energy regionunder 50 keV was sensitive to the presence of the cavity in the bentonite sample. In addition, the simulationindicated that the use of Al filter with thickness of 0.5-1.0 mm could effectively attenuate X-rays under30 keV. It is said that artifact rings, which often appear in CT images, can be removed by using amonochromatic X-ray. Accordingly, it was concluded that the applying voltage of 50 kV and Al filter of0.5 mm in thickness were an appropriate condition in the X-ray micro CT observation for the bentonitesample prepared in this study. Figure 1a shows the cross-sectional image of the dry bentonite sampleobserved under the condition. Well-defined cross-sectional images of the bentonite can be seen in thefigure; both flake-like montmorillonite particles (Montmorillonite particles of elliptic shape) and circularglass beads are clearly observed. Especially the long axis of the elliptic montmorillonite particles in thefigure is around 100 ìm, which is in good agreement with that determined with SEM. In addition, the shapeof glass beads is not distorted at any position inside the circle of the sample holder. Therefore, it isconsidered that the image reflects the information of the microstructure of the bentonite sample. Figure 1bshows the image of the water-saturated sample. Although glass beads are observed as clearly as in the drysample, the montmorillonite particles are not seen, suggesting grain size of bentonite becomes less thanspatial resolution of X-ray micro CT due to the water-saturation.

ACKNOWLEDGEMENTSThe authors would like to thank Professor T. Enoto, of Hokkaido University, and Dr. S. Suzuki, of Instituteof Research and Innovation. The authors express their appreciation to Mr. T. Ohyagi, Ms. Y. Mizuguchi,TOYO Corporation, providing the opportunity to use the micro CT apparatus and their technicalsuggestions. Financial supports were provided by Japan Atomic Energy Agency (JAEA), and Japan,Radioactive Waste Management Funding and Research Center (RWMC).

References:[1] T. Kozaki, S. Suzuki, N. Kozai, S. Sato, and H. Ohashi, J. Nucl. Sci. Technol., 38(8), 697-699 (2001).

[2] H. Hirayama, Y. Namito, A. F. Bielajew, S. J.Wilderman, andW. R Nelson, “The EGS5 code System”,SLAC-R-730. : Stanford Linear Accelerator Center (2006).

Figure 1: Cross-sectional images of (a) dry and (b) water-saturated bentonite samples observed with X-raymicro CT.

(a)

1mm

(b)

1mm

P/GM/39


HYDRATION AND HYDROLYSISOF SM3+ AND EU3+ IN CLAY INTERLAYER:

NEUTRON DIFFRACTION STUDY

O. Sobolev1, L. Charlet1, G. J. Cuello2, A. Gehin1, J. Brendle3

1. LGIT-OSUG, University of Grenoble I and CNRS, B.P. 53, F-38041 Grenoble Cedex 9, France2. Institut Laue Langevin, 6, rue Jules Horowitz, B.P. 156, F-38042 Grenoble Cedex 9, France3. LMPC, UMR CNRS 7016 Mulhouse Cedex

INTRODUCTIONThe following problems are fundamental for understanding and correct estimation of the retardation of theradionuclides in the clay barrier:

• The radionuclides are often present in water in the form of hydrolyzed neutral aqueous species. Howeverprevious study on for sorption of ammonia (as ammonium) and pyridine [1] have shown the interlayerwater to be more acidic than the water surrounding the clay aggregates. The excess acidity in the inter-

layer space may cause the radionuclides to be rather in the ionic form Mn+. This acidity of the interlayerwater has never been studied for actinides and lanthanides. This question is extremely important because

Mn+ species are much more favourably adsorbed in clay interlayer than.

• Fe(II) contained in natural montmorillonite and arising from the corrosion of the container can affect thesorption ability of the redox sensitive radionuclides in the interlayer space through the redox reaction

Mn+ →Mn(n-1)+.

For our investigation we have chosen Eu3+ and Sm3+ cations. Eu3+ is interesting because of its chemicalsimilarity to Pu3+ and Am3+. Sm3+ is also similar to Eu3+ and it is more suitable for neutron diffractionstudies, but its redox potential is different.

EXPERIMENTThe neutron diffraction experiment was curried out with using diffractometer D4 at the Institute Laue-Langevin (Grenoble, France). The aim of this experiment was to investigate by means of neutrondiffraction with isotopic substitution the structural parameters of the hydration, hydrolysis and sorption ofSm and Eu, in the interlayer space of the montmorillonite and their dependence on chemical and physicalfactors, which are important for radioactive waste management. The experimental results are obtained forthe hydrated samples of Sm-montmorillonite, prepared at low pH and pH>7 in order to find out whetherSm is present as aqueous Sm(OH)o3, Sm

3+, or intermediate hydrolyzed species at pH>7, and how it isconnected to the clay surface. The measurements with hydrated samples of Eu-montmorillonite, preparedat low pH were carried out for the estimation of the feasibility of our future neutron diffractionexperiments, having the aim to study the possible changes of Eu coordination in clay due to theM3+ →Mn2+, Fe2+ → Fe3+ reaction. This estimation had to be done, because of some uncertainties in theavailable information on the scattering length of 151Eu isotope.

RESULTS AND INTERPRETATIONIn Fig. 1 the radial distribution functions gSmX(R), describing the correlations between Sm atom and otherneighbour atoms, obtained with D2O and H2O hydrated montmorillonite are shown. The peak at 2.5 Å canbe identified as corresponding to Sm – O distance. The peak around 3.1 Å corresponds to Sm – Hcorrelations. In the case of H2O hydrated sample this peak has negative amplitude due to the negativescattering length of hydrogen. gSmX(R) obtained for different pH values looks similar, but in the case of pH> 8, the Sm – O distance is slightly shorter than that for the low pH (Fig. 2).

P/GM/39



The numbers of oxygen and hydrogen atoms in the vicinity of the Sm were estimated integrating theintensity of Sm – O and Sm – H peaks of the gSmX(R) functions. It was found that the number ofhydrogen atoms is equal or even slightly smaller than those of oxygen atoms. This means that Sm3+

Sm3+ is bind to the clay surface and it is probably partially hydrolyzed.

For the Eu-montmorillonite samples we obtained curves very similar to those shown in Fig. 1. This resultprovides the proof of the feasibility of our future experiments on Eu – Fe-doped montmorillonite systems.

References:[1]M.M. Mortland and K.V.Raman, Clays and Clay Minerals 16 (1968) 393.

Figure 1: Composite pair distribution functions gSmX(R) obtained for pH = 5.

Figure 2: gSmX(R) compared for different pH.

P/GM/40


MOLECULAR SIMULATIONSOF NA- AND CA-MONTMORILLONITE

Magnus Hedström, Martin Birgersson, Ola Karnland

Clay Technology AB, IDEON Research Center, S-223 70 Lund, Sweden

Over the past 15 years there has been a steadily increasing activity of computer simulations of swellingclays at the atomic level. In the present work we employ two different simulation techniques Monte Carlo(MC) and Molecular Dynamics (MD) in order to gain understanding at the fundamental atomic level ofclay swelling and also counter-ion diffusion in the interlayer pore water.

The clay mineral is represented by a rigid structure based on pyrophyllite and with partial charges, neededfor the evaluation of the Coulomb interactions between the mineral and the pore water including counterions, given by Skipper et al. 1995. Substitutions in the tetrahedral or octahedral layers are assumed to leavethe geometric structure unchanged, but do alter the partial charge.

In addition to electrostatic interactions the models used also include van derWaals interactions in the formof 6-12 Lennard-Jones potentials.

Several factors can influence the montmorillonite properties and in this work we have decided toinvestigate the effects of monovalent contra divalent counterions represented by Na+ and Ca2+ respectively,as these two species have similar ionic radii. In addition to the question of counter ion valences we alsostudy the effect of charge distribution between the octahedral and tetrahedral layers.We make comparisonsbetween two different montmorillonite structures both with layer charge equal to 1e-, Si8(Al3Mg)O20(OH)4and (Si7.75Al0.25)(Al3.25Si0.75)O20(OH)4.

The MC calculations are performed in the grand canonical ensemble which keep (µ,V,T) fixed, µ beingthe chemical potential and V and T the volume and temperature respectively. The MC simulations in the(µ,V,T)-ensemble mimic an experimental situation where the clay is in contact with bulk water andconfined to the volume V. The simulations give the water content and the swelling pressure. In the presentwork all simulations are done at T=300 K.

Diffusion coefficients are obtained fromMD simulations in the canonical (N,V,T) fixed and the isothermal-isobaric ensemble (N,P,T) fixed, where N is the number of water molecules and P is the external pressure(1 bar).

References:Skipper N.T., Chou Chang F.-R. and Sposito G. 1995. Monte Carlo simulation of interlayer molecularstructure in swelling clay minerals. 1. Methodology. Clays and Clay Minerals. Vol. 43: p. 285-293.

P/GM/41


EFFECT OF TEMPERATURE ON THEPROTON CHARGE OF MONTMORILLONITE

M. Duc1*, F. Thomas2, F. Gaboriaud1.

Laboratory of Physical Chemistry and Microbiology for the Environment, Nancy University,CNRS, 405 Rue de Vandoeuvre, F-54600 Villers-lès-Nancy, France([email protected]).

Laboratory Environment and Mineral Processing, Nancy University, CNRS, P.O. Box 40, F-54501Vandoeuvre-lès-Nancy Cedex, France, ([email protected])

*new adress: French Public Works Research Laboratory, 58 bd Lefebvre, 75732 Paris Cedex 15,France, ([email protected]).

In future deep geological depositories of nuclear wastes, the clay matrix used as confinement materialaround the radioactive containers will experience a temperature gradient up to 80-100°C, which maysignificantly alter the structural and physico-chemical properties of the constituting montmorillonite, andsubsequently its capacity to retain escaping radionuclides. The present work aimed at investigating theeffect of temperature in the range 25-80°C on the acid-base properties of montmorillonite. This remains achallenging problem, since the protonation-deprotonation of the silanol and aluminol sites on the edges ofthe clay platelets is coupled with dissolution-readsorption processes, which increase with increasingtemperature. Therefore, these processes were quantified and taken into account.

The studied montmorillonite was extracted from the MX80 bentonite by dispersion and decarbonatation,homoionized by Na exchange and stored as wet concentrated dispersions.

The proton charge was quantified by continuous and batch potentiometric titration. Continuous titrationwas performed using a conventional automated titrator (Duc et al., 2005), in a thermostated teflon reactorunder argon atmosphere. Aliquots of 0.01mL of 0.1M NaOH or HNO3 were added every 10 minutes.Thermodynamic calculations were applied according to the Berubé and De Bruyn’s approach (1968) basedon the correlation between the PZC variation versus the temperature and the enthalpy and entropyassociated to the reaction:

≡SO- + H+(aq) + H2O > ≡SOH2+ + OH-(aq), and .

For batch titration (Duc et al., 2006), the purified Na-montmorillonite was dispersed in separate reactorsin 0.1M or 0.01M NaNO3. In each reactor the pH was brought to an initial value ranging from 4 to 10 byaddition of HNO3 or NaOH. The reactors were hermetically closed and shaken in a thermostated bath at25°C, 40°C or 80°C for 24 hours or 1 month. Then, the pH was recorded at the experimental temperaturein a glove box flushed with argon. The suspensions were centrifuged, and the supernatants were analyzedfor Al, Mg, Si, Fe released by dissolution of the clay. The cations re-adsorbed on the clay were displacedby lanthanum exchange and analyzed (Turner et al., 1996). The net titrant consumption was calculated bydifference between the amounts of added titrant and the final proton concentration calculated from theequilibrium pH. Titrant consumption due to dissolution of the clay was calculated from the amount ofsoluble and re-adsorbed ions, and subtracted from the total consumption (Duc et al., 2006). Themineralogical and morphologic changes undergone by the clay during the batch experiments werecharacterized by XRD, IR, TEM and SEM.

Previous studies (Duc et al., 2005, 2006) have shown that the titration curves of 2:1 clays at different ionicstrengths do not show the expected characteristic intersection point indicating the PZC. Therefore thePZNPC, the points of zero net proton consumption obtained in continuous titration, were plotted versus theinverse temperature (Figure 1). The intercepts of the straight lines clearly give access to the entropy of thedeprotonation reaction (- 0.8 kJ/mol.°K). The negative slopes show the endothermic character of the

2R 10ln PZC 12---pKw–

ΔpS°ΔpH°

T-------------–=×

P/GM/41



protonation-deprotonation reactions on montmorillonite,in contrast to the expected exothermic character. Thecorresponding enthalpy is close to 10 to 20 kJ/mol islower than the expected 30 or 40 kJ/mol (in absolutevalue) for oxides, which would indicate a physisorptionrather than a chemisorption. Moreover, the ionic strength,which strongly influences the coulombic interactions, hasalmost no effect on the enthalpy and entropy of thedeprotonation, which would indicate that the observedPZNPC are close to the PZC of the studiedmontmorillonite (pH 5 ± 0.5).

However, such observations must be tempered by theresults of the batch titration experiments, where theimportance of dissolution was clearly shown in the onemonth equilibrium experiment, by the significant increaseof the net titrant consumptions recorded at acid and basicpH for all temperatures. Therefore, the exothermicdeprotonation could be overwhelmed by the stronglyendothemic dissolution. After correction according to thesoluble and readsorbed cations, the proton consumptioncurves (Figure 2) still show a tendency of the PZNPC todecrease with increasing temperature.

References:Berube Y. G. and De Bruyn P. L. (1968): Adsorption at the rutile-solution interface. I. Thermodynamic andexperimental study. J. Colloid Interface Sci., 27, 305-318.

Duc M., Gaboriaud F., Thomas F. (2005): Sensitivity of the surface charge of clays to the methods ofpreparation and measurement. 2. Evidences from continuous potentiometric titrations, J. ColloidInterface Sci., 289,148-156

Duc M., Thomas F., Gaboriaud F., (2006): Coupled chemical processes at clay/electrolyte Interface. Abatch titration study of Na-montmorillonite. J. Colloid Interface Sci, 300, 2, 616-625.

Turner G. D. et al. (1996): Surface-charge properties and UO22+ adsorption of a subsurface smectite.

Geochim. Cosmochim. Acta, 60, 3399-3414.

Figure 1: According to the Berubé and De Bruyn’s approach, variation of the PZNPC in 0.1 and 0.01MNaNO3 electrolyte versus the inverse of the temperature.

-2.5

-2

-1.5

-1

-0.5

0

0.0026 0.0028 0.003 0.0032 0.0034

1/T (K-1)

y = -0.17605 -265.64 (1/T)

y = -0.16869 -507.5 (1/T) R = 0.93306I = 0.1M

I = 0.01M

PZN

PC-1

/2p

Kw

Figure 2: Corrected surface charge curvesfrom batch titrations (suspension at 4.35g/L)agitated 24h at (Δ) 25°C, (c) 40°C, (r) 80°Cand after 1 month at 80°C (n) (suspension at10g/L).

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

3 4 5 6 7 8 9 10 11

-log[H3O+]

80°C 1 mois 10g/L

80°C

40°C

25°CQmol/kg

P/GM/42


TRANSFORMATION OF ORGANIC MATTERIN THE PRESENCE OF SMECTITE:

FATE OF LANTHANIDESDURING POLYMERIZATION REACTIONS

T. Schäfer1, P. Michel1, F. Claret2, G. Buckau1, M. Elie3

1. Institut für Nukleare Entsorgung (INE), Forschungszentrum Karlsruhe, P.O. Box 3640, 76021Karlsruhe, Germany ([email protected])

2. CEA Saclay, CEA/DPC/SECR/LSRM, Bâtiment 391, piece 39B, 91191 Gif sur Yvette, France([email protected])

3. UMR G2R-CREGU, BP239, 54506 Vandoeuvre lès Nancy Cedex, France([email protected])

Studies on argillaceous rocks under discussion for the storage of high level nuclear waste (i.e. Callovo-Oxfordian argillite and Opalinus Clay) have indicated that the low amount of organic matter (OM) presentin these natural rock formations (~0.2-1.5%) can have a significant influence on the reactivity of clays(Claret et al., 2002). It could been demonstrated by means of X-Ray spectromicroscopy (STXM) that evenafter two years reaction time in hyper alkaline solutions OM was still partly associated with clay edge sitesgiving hint that the natural OM is not solely associated via surface sorption (Schäfer et al., 2003).Furthermore alkaline extraction studies showed that only a very small amount (1-7% of total organiccarbon) of this OM could be classified as hydrophilic humic (HA) or fulvic (FA) acids (Claret et al., 2005).The combination of principle component- (PCA), cluster- and target spectra- analysis revealed differentsources for mobilized HA/FA, namely smectite rich regions for HA and illite dominated regions for FA(Schäfer et al., 2005). In addition, these investigations showed a low functional group content of themineral associated OM. Conceptual models exist in the literature that describe the OM mineral interactionas a multi-step process (Collins et al., 1995). One of the questions that arise from the above-mentionedinvestigations is, how the clay mineral type or exchangeable cation composition might influence theorganic matter polymerization process and how rare earth element (REE, homologues for trivalentactinides) are bond in such systems.

Condensation of low molecular weight compounds (glucose and glycine or catechol) on Na-, Ca-, or Fe-exchanged smectite via theMaillard reaction were performed as a first step to generate humic-like material.These samples were pyrolysed in a second step to reach a maturity comparable to the natural clay organicmatter (kerogen) found in the Callovo-Oxfordian formation.

UV/VIS analysis of the humic-like material shows a significant increase of polymerisation withtemperature (25˚C to 80˚C). A general tendency to have higher amounts of aromatic rings with substitutedpolar groups in the samples (Fe3+>Ca2+>Na+) could be identified. Infrared spectroscopy shows in the reactedglycine + glucose system still bands of low molecular weight (LMW) compounds. The broad absorptionin the region around 3650-3250cm-1 shifts in the presence of clay towards vibrations of primary amine.Absorption at the low end of the region 1850-1650cm-1 show amide probably in conjugation with a doublebond or aromatic ring. An increase of absorption in the region typical for aromatics can be identified. TheµFTIR results point out a catalytic effect of the mineral surface and especially the importance of the initialstep of organic matter cation bridging on the mineral surface for the formation of aromatic structures andmacromolecules. The polymerisation reaction seems to be triggered in the first step by the OM sorptiondensity.

The XRD patterns of the reacted clay shows a shift toward higher values in the d001 basal spacing can beobserved after the polymerization reaction. However, the observed d001 increase of ~0.6 Å is small and

P/GM/42



could also be explained by changes in the size of the coherent scattering domains (CSD). In the catecholsystem the asymmetrical d001 peak and the shift in d003 indicates the non-homogeneous distribution oforganics in the interlayer.

XANES investigations at the C(1s)-edge using STXM revealed for the polymerization reaction withoutsmectite at 80˚C almost no aromatics (C=C π*-transition at 285.2eV) low absorption in the phenol-typegroup range (286.6eV) and a high content of carboxylic (288.4eV) and carbonyl-type (289.2eV) groups.The Na-exchanged smectite associated OM shows a significantly lower content of oxygen-containingfunctional group and an increase in overall OM aromaticity. Further results show a strong impact of claysize on polymerization pathway (changes in OM functionality). In the catechol system a formation ofaromatic networks via C-O-C condensation was observed in the supernatant of Fe-exchanged clay(286.4eV), but C=C (285.2eV) decrease indicates as major process ring cleavage and dealkylation andcatalysis by clay surface.

The results also show that the smectite associated synthetic OM polymers after Maillard reaction still haveconsiderable higher oxygen containing functional group content in comparison to the clay mineralassociated OM found in natural formations, i.e. Opalinus Clay and Callovo-Oxfordian argillite. Therefore,in a second step pyrolysis were used to reach comparable maturity to the natural formations. Results onSTXM measurements of these samples as well as the fate of REE during the polymerization reaction willbe presented.

ACKNOWLEDGEMENTWe are grateful for beamtime allotment by BNL/NSLS. The results presented were partly supported by IPFUNMIG under the contract number FP6-516514.

ReferencesClaret, F., Bauer, A., Schäfer, T., Griffault, L. and Lanson, B. (2002): Experimental investigation of theinteraction of clays with high-pH solutions: A case study from the Callovo-Oxfordian formation,Meuse-Haute Marne underground laboratory (France). Clays and Clay Minerals, 50(5): 633-646.

Claret, F., Schäfer, T., Rabung, T., Wolf, M., Bauer, A. and Buckau, G. (2005): Differences in propertiesand Cm(III) complexation behavior of isolated humic and fulvic acid derived from Opalinus clay andCallovo-Oxfordian argillite. Applied Geochemistry, 20: 1158-1168.

Collins, M.J., Bishop, A.N., Farrimond, P.: Sorption by Mineral Surfaces - Rebirth of the ClassicalCondensation Pathway for Kerogen Formation Geochim. Cosmochim. Acta 59(11), 2387-2391 (1995).

Schäfer, T., Claret, F., Bauer, A., Griffault, L., Ferrage, E. and Lanson, B., 2003. Natural organic matter(NOM)-clay association and impact on Callovo-Oxfordian clay stability in high alkaline Solution:Spectromicroscopic evidence. Journal de Physique IV, 104: 413-416.

Schäfer, T., Claret, F., Lerotic, M., Buckau, G., Rabung, T., Bauer, A. and Jacobsen, C., (2005): Sourceidentification and characterization of humic and fulvic acids from Oxfordian argillite and OpalinusClay. In: E.A. Ghabbour and J. Davies (Editors), Humic Sustances: Molecular Details and Applicationsin Land andWater Conservation. Taylor & Francis, New York, Chapter 4.

P/GM/43


SURFACE COMPLEXATION MODELLINGUSING MULTI-COMPONENTS APPROACH.SORPTION OF SELENIUM ON ARGILITE

N. Marmier and C. Hurel

LRSAE, University of Nice Sophia Antipolis, Faculté des Sciences, 28 avenue Valrose, 06108 Nicecedex 2, France ([email protected], [email protected]).

INTRODUCTIONThe French underground research laboratory for the assessment of a deep radwastes disposal feasibility willbe located in a calloxo-oxfordian argilite layer. Major components in the argilite material are clays andcalcite, but numerous accessory phases, such as iron oxides and pyrite, could also be present. The retentionproperties of the argilite surface are of course an important parameter which can limited migration ofradioelement in the far field and are investigated by Kd measurements. Concurrently to this work, surfacecomplexation model using the multi-components (or components additivity, CA) approach is used as anhelp for interpretation. Previously determined mass action law (MAL) data, concerning pure mineralphases, are used to account for experiments performed using natural argilite and Se(IV). The surfacespeciation curves obtained by calculation help us to vizualise active phases for argilite with selenium, andthe role of the major aqueous species, including dissolved silicates, in the sorption processes.

EXPERIMENTAL CONCEPTIn a first step, the ability for prediction of the component additivity concept (ADTS model) was tested byconfrontations with experimenets. To simulate an argilite surface, different mixtures of previouslycharacterized bentonite, calcite, iron oxydes, and pyrite are suspended in an synthetic groundwatercontaining Se(IV) in various aqueous conditions (with varying pH, Se(IV) concentrations), and relativemineral amounts to simulate a spatial heterogeneity. In the second step, the same kind of experiments wereperformed on natural argilite suspensions, in the same aqueous conditions and masse/volume ratio.

RESULTS AND INTERPRETATIONResults show that mass action law (MAL) data for active phases (bentonite and iron oxides), includingsurface complexation constants and sites densities determined in independent experiments, can be used incalculations without any change in the case of the pure mineral phases mixtures. Surfaces of calcite andpyrite do not required an explicit parameterization in the model for the tested experimental conditions (lowM/V ratios), but dissolved ions (such as calcium, carbonates,...), competing in surface and/or solutionreactions, are taken into account with their own quantified reactivity. Dissolved silicates and theirinterference with aqueous and surface species are also considered. The same global behaviour is observedin the case of argilite suspensions, and the same MAL dataset can be used to account for the Se(IV)rétention. On the other hand, surfaces sites densities for argilite are different from those determined on purephases, and have to be reevaluated for the calculations, consistently to the CEC value for argilite.

References:Buerge-Weirich D., Hari R., Xue H., Behra P., Sigg L.(2002) adsorption of Cu, Cd, and Ni on goethite inthe presence of natural groundwater ligands, Environ. Sci. Technol., 36, 328-336

Davis J.A., Coston J.A., Kent D.B., Fuller C.C.(1998) Application of the surface complexationconcept to complex mineral assemblages, Environ. Sci. Technol.,, 32, 2820-2828.

Hurel C., N. Marmier N. (2006) Sorption of selenium on mineral mixtures : role of minor phases in themodelling part. Scientific Basis for Nuclear Waste Management XXIV, Materials Research SocietySymposium Proceedings, 932 , 967-974.

P/GM/44


IN-SITU CHARACTERIZATIONOF THE CALLOVO-OXFORDIANPOREWATER COMPOSITION

A. Vinsot1, S. Mettler2

1. Andra, Laboratoire de recherche souterrain de Meuse/Haute-Marne, RD 960, F-55290 Bure,France ([email protected])

2. NAGRA, Hardstrasse 73, CH-5430 Wettingen, Switzerland ([email protected])

INTRODUCTIONThe study of the mechanisms governing the composition of the interstitial fluids in clay formations withvery low permeability constitutes a part of the scientific programmes associated with the safety analysesof deep repositories for the geological disposal of long-lived radioactive waste. Up to 2005, the knowledgeabout the Callovo-Oxfordian porewater composition was entirely built on geochemical modeling. Areference geochemical model had been defined from measures performed on bore cores drilled on the siteof the Meuse/Haute Marne Underground Research Laboratory (URL) (Gaucher et al. 2006, Jacquot andAltmann 2005). Since 2005, porewater has been sampled in situ in several URL boreholes equipped withvarious setups. Results obtained from the porewater analyses are presented here and compared to thegeochemical model.

EXPERIMENTAL CONCEPTExperiments called “PAC” were designed to characterize the Callovo-Oxfordian porewater. Theseexperiments were performed in boreholes drilled from the URL drifts (Delay et al., 2006). Theexperimental concepts were based on feedback from the Mont Terri rock Laboratory (Pearson et al., 2003).They are of two types: one based on water circulation and the other combining gas circulation and watersampling. In addition, sampling of seepage water was possible in an isolated borehole at -490 m (called“SUG”) devoted to hydrogeological measurements. The PAC borehole test intervals were drilled withnitrogen to avoid oxidation of the formation. Furthermore, maximum precautions were taken to avoidbacterial contamination (disinfection of the equipment, of the injected water, etc.). The SUG borehole wasdrilled with air, without antimicrobial precautions. The water circulation experiment consists in circulatingsynthetic porewater in a closed loop between the test interval located at 10 to 15 m depth in a boreholeand devices for sampling and online measuring installed in the URL drift. The circulating watercomposition is alike the real porewater and traced with D and Br-. It was defined on the basis of thegeochemical model. In contact with the rock, the water evolves towards equilibrium with the interstitialwater by rock-water reactions, diffusive exchange and inflow. Two water circulation experiments wereinstalled: PAC460 at -460 m and PAC505 at -505 m. The water sampling and gas circulation experimentconsists of a vertical ascending borehole with a 5 m long test interval at its far end (Vinsot et al., 2007).After installation of the equipment, the test interval was filled with pure argon at a pressure of 1 bar. Theborehole equipment allows the circulation of the gas in contact with the rock in a closed circuit. Due tothe large hydraulic gradient between the test interval and the surrounding rock, the water flows into theinterval and is pumped out at a rate of 20 to 40 mL/day. The borehole equipment allows the sampling ofthe water produced in the test interval. Two water sampling experiments were installed: PAC430 at -430m and PAC475 at -475 m.

GEOCHEMICAL MODELThe geochemical model is described in Gaucher et al. (2006). It assumes thermodynamic equilibriumbetween the minerals of the formation and the porewater. The mineral assembly (calcite, dolomite, siderite,celestite, quartz, daphnite, illite, pyrite and either a chlorite-Mg or a fixed pCO2) constrains the elementconcentrations. Also cation exchange reactions are considered.

P/GM/44



RESULTS AND INTERPRETATIONThe experimental results show an overall convergence of the measured water compositions independentlyof the experimental concepts (collection of seepage water or water circulation), and of the sampling depths(Figure 1).

Comparison of the observed and calculated compositions (Figure 1) shows a discrepancy below a factorof three for all the major species except potassium and strontium. Attempts at explaining the observeddifferences suggested a review of the Sr solubility constraints as well as the cation exchange parameters.Based on these experimental data and further developments, an improved model is now being proposed(Gaucher et al. 2007).

References:Altmann S., Jacquot E. (2005) La chimie des eaux interstitielles dans la couche du Callovo-Oxfordien àl’état initial (Site Meuse/Haute-Marne). Note technique Andra C.NT.ASTR.03.023.

Delay J., Vinsot A., Krieguer J.M., Rebours H. and Armand G. (2006) Making of the undergroundscientific experimental programme at the Meuse/Haute-Marne underground research laboratory NorthEastern France. J. Phys. Chem. Earth (2006), doi:10.1016/j.pce.2006.04.033

Gaucher E. C., Blanc P., Bardot F., Braibant G., Buschaert S., Crouzet C., Gautier A., Girard J.-P., JacquotE., Lassin A., Negrel G., Tournassat C., Vinsot A., and Altmann S. (2006) Modelling the porewaterchemistry of the Callovian-Oxfordian formation at a regional scale. Comptes Rendus Geosciences338(12-13), 917-930.

Gaucher E. C., Tournassat C., Jacquot E., Altmann S. and Vinsot A. (2007) Improvements in the modellingof the porewater chemistry of the Callovian-Oxfordian formation. Clays in natural and engineeredbarriers for radioactive waste confinement - 3rd International Meeting. Lille, France

Pearson F.J., Arcos D., Bath A., Boisson J.Y., Fernández A.M., Gäbler H.E., Gaucher E., Gautschi A.,Griffault L., Hernán P. etWaber H.N. (2003) Geochemistry ofWater in the Opalinus Clay Formation atthe Mont Terri Rock Laboratory. Report of the Swiss Federal Office for Water and Geology, GeologySeries. N˚ 5. 319 pp. ISBN 3-906723-59-3.

Vinsot A. and Delay J. (2007) In situ sampling and characterization of Callovo-Oxfordian pore water.Water Rock Interaction WRI-12 Kunming China. (submitted).

Figure 1: Measured composition of the water sampled in the PAC and SUG boreholes in comparison withthe model porewater composition.

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

Na K Mg Ca Sr Cl SO4 TIC Iron Silicium

Concentration,mol/l

PAC460

PAC430

PAC475

PAC505

SUG

Reference model

P/GM/45


NATURAL ORGANIC MATTERFRACTIONATION ON MINERAL SURFACES:

A SPECTROSCOPIC APPROACH

F. Claret1,2, T. Schäfer3, P. Reiller1

1. CEA/DEN/DANS/DPC/SECR/Laboratoire de Spéciation des Radionucleides et des Molécules,Bâtiment 391, F-91191 Gif-sur-Yvette CEDEX, France ([email protected]).

2. Present address: BRGM – Bureau de Recherche Géologique et Minière, 3 Avenue ClaudeGuillemin, BP 36009, 45060 Orléans cedex 2, France ([email protected])

3. FZK-INE – Forschungzentrum Karlruhe-Institut für Nukleare Enstorgung, Karlsruhe, P.O.Box 3640, D-76201 Germany ([email protected])

INTRODUCTIONSorption to mineral surfaces and complexation by Natural Organic Matter (NOM) are two importantprocesses influencing the cycling of potentially toxic trace metals in the environment. The specificinteractions between metal cations and mineral surface absorbed NOM are still matter of debate. Indeed,comparison of results obtained using simple linear additive model (LAM), which essentially describes thesystem as a physical mixture of independent sorbent phases, with direct measurement in the ternary system(metal-NOM-mineral surface) showed significant deviations. Among hypotheses, the mineral surfaceinduced fractionation of NOM is advanced. Even thought results found in the literature might sometimesbe contradicting, carbon functional group composition is clearly affected by mineral surfaces (Reiller et al.,2006). Therefore, a detailed process understanding of the organic matter - mineral surface interaction isneeded. Parts of this work were conducted within the framework of the second research technologicaldevelopment component of the European integrated project FUNMIG, workpackage 2.2: “Formation,migration and transport processes of organic/humic colloids”.

EXPERIMENTALTwo Humic Acids (HA), the Gorleben humics acid-GoHy 573 and Purified Aldrich Humic Acid-PAHA,and one Fulvic Acid (FA), the Suwanee River Fulvic Acid-SRFA, were sorbed on alpha alumina (α-Al203).Carbonates were removed from α-Al203 following the protocol described by Alliot et al. (2006) to avoidany carbonate induced α-Al203 surface property changes. Batch sorption experiments are conducted at pH6 at room temperature and the suspension was shaken continuously for 24 hours. The pH was checkedfrequently until a constant pH value was reached. The solid phase (α-Al203 + sorbed organics) wasseparated from the dissolved organics via ultracentrifugation. HA and FA concentrations are then measuredspectrophotometrically at 254 nm using a spectrophotometer and the dissolved organic concentration(DOC) in solution was determined with a Shimadzu 5000 TOC analyser. Sample are acidified and purgedwith argon in order to outgas inorganic carbon as CO2. SRFA, GoHy573 and PAHA fractionation onα-Al203was studied using spectroscopic techniques (UV/VIS, Near Edge X-ray Fine Structure-NEXAFS,Time Resolved Fluorescence Laser Spectroscopy-TRFLS).

RESULTS AND INTERPRETATIONIn Figure 1, the percentage of sorbed SRFA onto the alumina surface is plotted as a function of the ratiobetween the SRFA and α−Al2O3 concentration RSRFA. If RSRFA is < 20, sorption values obtained by UV/Visare bigger than those determined by DOC analysis. This phenomena was previously observed by Gu et al.(1994) for Suwannee River NOM sorption on iron oxides and supports the idea of preferential sorption ofchromophoric functional groups on α−Al2O3 Indeed, in this case the measured absorbance in thesupernatant will be lower and therefore the calculated sorbed concentration higher than the one calculatedfrom DOC analysis. UV/Vis spectra where deconvoluted using a procedure published by Korshin et al.(1997). The ratio A0,Et/A0,Bz, where A is the relative band intensity (Electron transfer band versus Benzoid

P/GM/45



band), is plotted as a function of RSRFA (cf. Figure1). In the region of comparable UV/Vis and DOCanalysis (RSRFA>20) the A0,Et/A0,Bz ratio is almostconstant around 1, whereas for RSRFA values <20the band ratio is decreasing until 0.5. According toKorshin et al. (1997) the decrease of the A0,Et/A0,Bzratio indicates the loss of aromatic functionalgroups substituted with hydroxyls, carboxyl, estersand carbonyls groups.

C(1s) NEXAFS measurements made on thesupernatant after sorption indicates a decrease inaromatic, phenolic and aliphatic groups contentwith decreasing RSRFA and an increase in carboxylgroups. These results corroborate the diminutionof aromatic functionalities substituted by hydro-xyls groups seen in the UV/Vis measurements.However, for GoHy573 humic acid, fractionationphenomenon is not evidenced, unlike otherwiseevidenced on PAHA. It can be inferred that therelie some other physical and chemical differences

between this extract and other humic extracts. Using europium as a fluorescence probe, first measurementsmade by TRFLS on the fractionated PAHA seem to indicate different complexation behaviour incomparison to non sorbed PAHA. Indeed, the peak height ratio between the 7F2 and 7F1 transitions issmaller for fractionated PAHA than for initial PAHA. Therefore fractionation of natural organic matter canlead to different complexation properties and might explain the deviation between LAM calculation andexperiments. Further work is ongoing to confirm this hypothesis.

References:Alliot, C., Bion, L., Mercier, F., and Toulhoat, P. (2006). Effect of aqueous acetic, oxalic, and carbonicacids on the adsorption of europium(III) onto alpha-alumina: Journal of Colloid and Interface Science,298 (2), 573-581.

Gu, B. H., Schmitt, J., Chen, Z. H., Liang, L. Y., and Mccarthy, J. F. (1994). Adsorption and Desorption ofNatural Organic-Matter on Iron-Oxide - Mechanisms and Models: Environmental Science &Technology, 28 (1), 38-46.

Korshin, G. V., Li, C.W., and Benjamin, M.M. (1997).Monitoring the properties of natural organic matterthrough UV spectroscopy: A consistent theory: Water Research, 31 (7), 1787-1795.

Reiller, P., Amekraz, B., and Moulin, C. (2006). Sorption of Aldrich humic acid onto hematite: Insightsinto fractionation phenomena by electrospray ionization with quadrupole time-of-flight massspectrometry: Environmental Science & Technology, 40 (7), 2235-2241.

Figure 1: RSRFA=[SRFA]ini (mgL-1)/[ α−Al2O3](gL-1)

dependent adsorption of SRFA on alumina in 0.1 MNaClO4. Circles and squares represent UV/Vismeasurement and DOC analysis, respectively.Triangles represent the evolution of the A0,Et and A0,Bzband ratio obtained after de-convolution of the UV/Vis spectra.

0 10 20 30 40 50 60 70 80 90 100

0

20

40

60

80

100

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

DOC measurement SRFA

UV measurement SRFA

%sorbedontoalphaalumina

RSRFA= [SRFA]ini(mgL-1)/[α-Al2O3](gL-1)

P/GM/46


EXAMINATION OF CLAY MINERALS-ORGANIC MATTER INTERACTIONS ATSUBMICRON-SCALE DURING CONFINEDPYROLYSIS OF A SIMPLIFIED SYSTEM

VOLCLAY BENTONITE - TYPE II KEROGENP. Michel1, Schäfer, Th.1, Claret, Fr.2, Elie, M.3

1. Forschungszentrum Karlsruhe, Institut fur Nukleare Entsorgung (INE), Postfach 3640, D-76021Karlsruhe, Germany ([email protected] and [email protected])

2. CEA Saclay, DEN/DCP/SECR/LSRM, 91191 Gif sur Yvette, France([email protected]).

Present address BRGM – Bureau de Recherche Géologique et Minière, 3 avenue ClaudeGuillemin, BP36009, 45060 Orléans cedex 2, France.

3. G2R, Nancy-Université, CNRS, BP239, 54506, Vandœuvre-lès-Nancy, France([email protected])

Understanding and modelling confinement properties over the long term of a deep geological repositoryfor nuclear waste is a key issue of performance assessments. In this context, specific experimental workshave been addressed on the impact of thermal or chemical perturbations on either the clay minerals ororganic matter (ANDRA, 2005). A string of investigations were carried out in the petroleum field toevidence the role of clays minerals in thermal alteration of organic matter. However, the full extent of clayminerals-organic matter interactions under thermal stress is still unknown because the findings were mainlybased on bulk and molecular parameters. Argillite of the host rock is a complex system formed of multiplemineral (i.e. various clays minerals, calcite, quartz, pyrite) and organic (i.e. origin) phases. Clues regardingthe clay minerals-organic matter interactions can be gained by observing at submicron-scale thetransformations occurring during pyrolysis of simplified systems in which the nature of the mineral phaseas well as the type of organic matter should be well defined. Here, microscopic observations andspectroscopic measurements have been conducted by using Electron Energy-Loss Spectroscopy coupled toTransmission Electron Microscopy (TEM-EELS) and Scanning Transmission X-Ray Microscope (STXM)in order to (1) observe at submicron-scale the nature of organic matter and the potential interactions withbentonite and (2) compare the informations provided by both methods.

MATERIALS AND METHODSConfined pyrolysis experiments in argon filled gold cells were performed (Monthioux et al., 1985) withmixtures of Volclay bentonite containing mainly smectite phase (Rassineux et al., 2003), and isolatedkerogen from the Toarcian shale (Paris Basin) and water. This type II kerogen was chosen for its abilityto generate high amounts of soluble organic matter during thermal maturation. The water/(bentonite +kerogen) ratio was 1:1, whereas the kerogen/bentonite ratio was fixed to 1:10. Experiments were conductedin the temperature range 200°C-365°C, under a total pressure of 300 bars during 72 hours. After each runand quenching, the gold cells were opened and the global products were analysed by using TEM-EELSand STXM.

RESULTS AND DISCUSSIONX-ray diffraction results revealed no organic matter in the smectite interlayer independent of the conditionsused for confined pyrolysis. Furthermore, test with ethylene glycol revealed that swelling properties ofsmectite were kept even in the presence of formed liquid organic matter. TEM-EDX measurements onpyrolysis products clearly show the presence of organic carbon predominantly as isolated particle oforganic matter. EELS spectra of such particles are similar to those obtained on the isolated kerogen

P/GM/46



(amorphous organic material) before pyrolysis (Figure 1a). Moreover, organic carbon was also found onsmectite surfaces (Fig. 1c). EELS spectra of such cloudy organic matter structures revealed an increase inoxygen functional groups by discriminated C K-edge bands (Figure 1b), in particular (1) a band around288/289 eV assigned to the 1s → π* transition of the C=O bond of carboxylic groups and (2) a shoulderat 286/287 eV corresponding to either the 1s → σ* transition of the C-H bonds or the 1s → π* transitionof the C=O bond of phenol type groups. The smectite could be identified by additional measurements atthe calcium L2,3 –edge around 350 eV.

Scanning Transmission X-Ray Microscopy (STXM) investigations confirmed TEM-EELS measurements.C K-edge spectra showed again distinctive different spectra for clay-associated organic matter andparticulate organic matter found.

In summary, the investigations show a clear consistency of the data obtained by both TEM-EELS andSTXM on the carbon K-edge. Furthermore, the spectromicroscopic observations demonstrate that thermalstress simulated by confined pyrolysis induces changes on clay associated organic matter, whereas pureparticulate organic matter seems to be structurally similar in experiment with or without clay. The initialswelling properties of the clay are conserved and not affected by the occurrence of clay associated organicmatter as cloudy/patchy structures. This implies that the association is restricted to the external surface.Further studies are underway to elucidate the involved organic matter association/reaction processes.

ACKNOWLEDGEMENTSTEM results are from the P. Michel's PhD thesis financially supported by ANDRA. We acknowledgeANDRA for this support and BNL/NSLS for beamtime allotment.

ReferencesAndra (2005): Dossier Argile – Evolution phénoménologique du stockage géologique, Collection « lesRapports », ANDRA, 523 p.

Monthioux, M., Landais, P. and Monin, J-C. (1985): Comparison between natural and artificial maturationseries of humic coals from the Mahakam delta, Indonesia. Organic Geochemistry, 8, 275-292.

Rassineux, F., Griffault, L., Meunier, A., Berger, G., Petit, S., Vieillard, P., Zellagui, R. and Munoz, M.(2001): Expandability-layer stacking relationship during experimental alteration of a Wyomingbentonite in pH 13.5 solutions at 35 and 60°C. Clay Minerals, 36, 197–210.

Figure 1: EELS spectra between 270 et 310 eV for the sample «bentonite + isolated kerogen + water» afterpyrolysis at 350°C with (α) isolated organic matter particle, (β) organic matter phase observed on thesurface of smectite particle and (c) the cloudy structures observed on clay found with TEM.

270 275 280 285 290 295 300 305 310

C=C

a

1s π*

1s σ*285 eV

293-300 eV

C=C, C -C

C=O, C-O

C-H

Energy Loss

b

C=C

1s π*285 eV

C-

1s σ*286 eV

C=O1s _ π*289 eV

1s σ*293-300 eV

C=C, C -C

C=O, C-O

C-H

270 275 280 285 290 295 300 305 310

Energy Loss

c

5 nm5 nm

P/GM/47


SORPTION OF NP(V), U(VI) AND CSONTO BENTONITE AND CLAYEY SOILS

M.N. Sabodina1,2, St.N. Kalmykov2, K.A. Artem’eva2, E.V. Zakharova1

1. Frumkin Institute of Physical Chemistry and Electrochemistry, Russian Academy of Science,Moscow, Russia

2.Chemistry dept., Lomonosov Moscow State University, Moscow, Russiae-mail: [email protected]

INTRODUCTION

Low-level nuclear wastes containing transuranium elements (TRU nuclides) such as neptunium anduranium are generated in spent nuclear fuel treatment procedure. The safe disposal of radioactive waste ingeological formations is based on what is known as multi-barrier concept. Bentonite is a natural claymineral that has high sorption capacity towards cations and is proposed in several countries as a maincomponent of isolation barriers at the nuclear waste repository sites. However, the possible use of sandysoils with high clay content is considered for this purpose as well.

The presence of colloid particles is essential for radionuclide transport and should be considered in designof such barriers. For bentonite pore waters that have reducing Eh values actinides could be present in theform of intrinsic hydroxocolloids or sorbed onto natural aquatic colloids forming so-called pseudocolloids.Presence of surface coatings on colloidal particles may change their coagulation properties (e.g. Fe(III)oxide coating or humic coating). The aim of this work was to study actinide speciation in bentonite porewaters and sorption of 237Np(V), 137Cs, 238U(VI) by clayey soils including possible role of surface coatingson it.

EXPERIMENT

Bentonite (Khakassiya deposit, Russia) was taken in Na-form for all experiments. The mineralogicalcomposition of the sample was characterized by powder X-ray diffraction, particle topology was studiedby SEM and TEM. The average particle size was 3 µm as determined by dynamic light scattering techniquehowever according to TEM the presence of clay nanocolloids was detected as well. Bentonite pore waterswere separated by ultrafiltration under inert (N2) atmosphere and analyzed by AAS, AES and ionchromatography. The pore wasters were found to be reductive with Eh= -100 mV due to presence of Fe(II)or organic matter. The major complexing ligand are CO3

2-/HCO3- and SO4

2-. Sorption experiments wereperformed under N2 atmosphere in plastic vials to avoid radionuclide adsorption onto walls. Bentonitesamples were left in the working solutions to swell for few days before sorption experiments wereperformed. After the desired concentration of radionuclide was added to the suspension, the required pHvalues were established and samples were left until the equilibrium was reached. Separation of suspendedmatter after the sorption was performed using microfiltration techniques.

For experiments with natural clayey sols, samples were fractionated granulometrically (0.5 − 0.25mm; 0.25− 50µm; 50µm − 10µm and <10µm). For each fraction specific surface area and porosity, were determined.All samples were characterized by SEM and TEM. The sorption of radionuclides was studied by initialclayey soil samples and by samples after extraction of Fe amorphous compounds (by Tamm methodsdescribed by Vorob’eva, 1998).


The sorption of U(VI) and Np(V) on bentonite was found to be pH dependant that indicate predominantsurface complexation mechanism of their sorption. For 137Cs the pH dependence of sorption was less

P/GM/47



pronounced that indicate the ion exchange as the major mechanism. The equilibrium constant of Na+/Cs+

exchange reaction calculated from sorption isotherms and pH dependences:

NaX + Cs+ ↔ CsX + Na+ log K = 1.7

For Cs+ sorption depended upon ionic strength as studied in the range from 0.001 to 0.1 M (NaClO4). Thesame effect is observed for Np(V) and U(VI) at pH<5. In case of neptunium and uranium ion exchangecontributes to their sorption at low pH values while at neutral and slightly alkaline solutions the sorptionis defined by surface complexation reactions.

The desorption of radionuclides by deionized water, 1M KCl and 1M HCl was studied. The largest fractionof Cs was desorbed by 1M KCl while for of Np and U the largest fraction were extracted by 1M HCl thatsupports the assumption on different sorption mechanisms. Only minor fraction of radionuclides wasdesorbed by deionized water.

For natural clayey soils the highest sorption of actinides (Np and U) correlated with Fe(III) content andwas highest for the largest fraction (0.5 − 0.25 mm) and the finest fraction (<10µm). The extraction ofFe(III) from the samples result in significant decrease of actinide sorption. Indicating, that sorption ofactinides is defined by presence of surface coatings of Fe(III) compounds. Earlier it was established bySEM, that Fe is present both in the form of separate hematite particles, and in the form of surface coatings.Presence of Fe(III) surface coatings did not effect on Cs sorption.

References:Vorob’eva L.A. Chemical analysis of sandy soils. Moscow State University, 1998, 272 P.

This research was supported by the RFBR (grant 05-03-33028).

P/GM/48


THE PHREEQC/TRACES COUPLING TOOLWITHIN THE ALLIANCES PLATFORM

A. Dimier*, J. Gaombalet*, N. Leterrier1

* Agence Nationale pour la gestion des déchets radioactifs

1. Commissariat à l’énergie atomique

The safety assessment of nuclear waste disposals requires an accurate prediction of the radionuclides andchemical species migration through engineered barriers and geological media. It is therefore necessary todevelop and assess qualified and validated tools which couple the transport mechanisms through geologicalmedia to chemical phenomena governing the mobility of radionuclides. Such a reactive transport tool hasbeen developed in the context of the numerical software platform Alliances. Among available codes, wewill focus here on two open source tools: PHREEQC for the chemical part, developed at the USGS, andTRACES for the transport part, developed at the IMFS of Strasbourg. These two tools have been coupledwithin Alliances with an iterative scheme, the affordable phenomenology being: aqueous speciation,dissolution- precipitation, sorption and surface complexation.

We will focus here on two simulation examples to illustrate the new affordable phenomenology inAlliances 2.1.

SEALING

We have studied here the evolution of porosity in a column initially filled with quartz and calcite. Theexperiment has been described by V. Lagneau in his thesis, see [1]. Alteration of the matrix by sulphateions conducts to gypsum and Smithsonite formation. In volumes, 37 ml of calcite theoretically conductsto 107 ml of secondary minerals; the final porosity value is in agreement with the experimental one. In thecase of varying permeability, K is expressed as:

We can state on figure 1 the dependency of perturbation length on permeability variation.

Figure 1: Dependency of porosity alteration with permeability variation.

K K0ω3

1 ω–( )3

ω03

1 ω0–( )3-----------------------------=

P/GM/48



The second illustration of the potentiality of the coupling between phreeqC and Traces is presented via thestudy of the 3D simulation of a bentonite / cement interface. Taking a 3D meshing allows to take massinteraction correctly into account, that wasn’t available with previously conducted 2D simulations.

References:[1] Vincent Lagneau : Influence des processus géochimiques sur le transport en milieu poreux; applicationau colmatage de barrières de confinement potentielles dans un stockage en formation géologique

[2] Gaucher E. C., Blanc P., Matray J.-M., et Michau N. (2004) Modeling diffusion of an alkaline plume ina clay barrier. App. Geochem. 19 pp.1505–1515

Figure 2: 3D Meshing and Montmorillonite alteration after 90000 years.

P/GM/49


IS PORE WATER CLOSE TO THE SURFACESOF A CLAY ROCK ENRICHED IN HEAVY

WATER ISOTOPES AS COMPAREDTO FREE PORE WATER?

Th. Gimmi1,2 and H.N. Waber1

1. Rock-Water Interaction, Institute of Geological Sciences, University of Bern,Baltzerstrasse 1–3, CH-3012 Bern, Switzerland ([email protected], [email protected])

2. Laboratory for Waste Management, NES, Paul Scherrer Institut, CH-5232 VilligenSwitzerland.([email protected])

INTRODUCTIONArgillaceous rocks like Opalinus Clay are envisaged as potential host rock formations for the disposal ofradioactive waste (Nagra, 2002). The large-scale transport properties of such argillaceous rocks are oftenevaluated by studying the distribution of stable water isotopes in the pore water across the low-permeabilityzones (e.g., Patriarche et al., 2004; Gimmi et al., 2007). In such argillaceous rocks, water is typically finelydispersed and resides in narrow pores with pore sizes in the order of nanometers to a few micrometers.Because of its close vicinity to the mineral surfaces, a part of this pore water is believed to be restrictedin its rotational and translational mobility and is denoted as bound pore water. This bound pore water maydiffer in its thermodynamic properties from free bulk water. Accordingly, water molecules containingheavy isotopes could have the tendency to be enriched within the bound pore water as compared to freepore or bulk water.

Such enrichment is of importance in many respects. For instance, one has to account for it (provided it issignificant) when stable isotope data are compared that were obtained with different methods likesqueezing or diffusive equilibration. During squeezing, for instance, typically only small fractions of thepore water can be extracted, which probably belong to the more mobile and thus the potentially depletedfraction. Diffusive equilibration, on the other hand, would need a correction if all pore water is bound andenriched with heavy isotopes compared to free bulk water. An enrichment of heavy isotopes in bound porewater may also be relevant for the interpretation of profiles of stable water isotopes. Notably, theheterogeneity of properties of the rock like mineralogy, clay content, porosity, or pore sizes may lead to aspatial variation of the fraction of bound pore water. As a consequence, even in an equilibrium situation,where no net transfer of stable water isotopes occurs, the stable isotope contents could vary as a functionof space. Thus, knowledge about the enrichment of heavy water isotopes within bound pore water helpsto interpret stable isotope signatures.

METHODSIn order to investigate the possible fractionation of heavy water isotopes within the pore water of a lowpermeability rock, we interpreted the stable water isotope data of the borehole Benken (northeasternSwitzerland) that were obtained by different methods. In one case, the diffusive isotope equilibrationmethod (Rübel et al., 2002) was applied. It is believed to give reliable data for the isotope contents andthe water-content porosities of the clay rock samples. With the other method, vacuum distillation, it is ingeneral not possible to remove the pore water from a clay rock quantitatively, and a Rayleigh-typefractionation occurs. In case of homogeneous properties of all pore water fractions, including the boundwater, the fractionation factor would remain constant throughout the distillation. However, if bound porewater has different properties than free water, the fractionation factor would vary as a function of theamount of water evaporated or, accordingly, of the remaining water content.We used the Benken data setsto estimate the pore water-vapour fractionation factors during the distillation. On the basis of these factors,

P/GM/49



one can then calculate bound pore water-free water fractionation factors, or also the enrichment of heavywater isotopes within different pore water fractions.

RESULTSThe pore water-vapour fractionation factors obtained from the data are shown in Figure 1. The resultsshowed that there is a tendency of increasing fractionation factors with decreasing amount of residual porewater for both, 18O and 2H. The factors are larger (up to about 1.07) for 2H as compared to 18O (up to about1.02). At room temperature, the water-vapour fractionation factors are 1.0793 for 2H and 1.0094 for 18O(Majzoub, 1971). Fractionation factors decrease with increasing temperature, and values estimated for100˚C or 105˚C, which is the temperature used during the distillation, are about 1.0250 for 2H and 1.0049for 18O. For both isotopes, the estimated pore water-vapour fractionation factors seem to approach thesevalues for larger pore water fractions.

On the basis of these observations, we conclude that stable water isotopes are indeed enriched within thepore water fraction that is closest to the surfaces. Because only average and not local values over theremaining pore water fraction were obtained, it is not possible to define exactly how much of the porewater is really affected by the enrichment. It seems, however, that in Opalinus Clay from Benken, withtypical porosities on the order of 0.12 m3 m–3, less than about one third of the pore water has higher porewater-vapour fractionation factors and is thus enriched with heavy isotopes compared to free pore water.

References:Gimmi, Th., H. N. Waber, A. Gautschi, and A. Rübel (2007): Stable water isotopes in pore water ofJurassic argillaceous rocks as tracers for solute transport over large spatial and temporal scales, WaterResour. Res., in print.

Majzoub, M., (1971): Fractionnement en oxygen-18 et deuterium entre l’eau et vapeur, J. Chem. Phys. 68,563–568.

Nagra (2002): Projekt Opalinuston: Synthese der geowissenschaftlichen Untersuchungsergebnisse, NagraTechnical Report NTB 02-03, Nagra, Wettingen, Switzerland.

Patriarche, D., E. Ledoux, J.-L. Michelot, R. Simon-Coinçon, and S. Savoye (2004): Diffusion as the mainprocess for mass transport in very low water content argillites: 2. fluid flow and mass transportmodeling, Water Resour. Res., 40, W01517, doi10.1029/2003WR002700.

Rübel, A., C. Sonntag, J. Lippmann, A. Gautschi, and F. J. Pearson (2002): Solute transport in formationsof very low permeability: Profiles of stable isotope and dissolved gas contents of the pore water in theOpalinus Clay, Mont Terri, Switzerland, Geochim. Cosmochim. Acta, 66, 1311-1321.

Figure 1: Pore water-vapour fractionationfactors as a function of the residual porewater fraction estimated for clay samplesfrom Benken.

1.00

1.01

1.02

1.03

1.04

1.05

1.06

1.07

1.08

0 0.05 0.1 0.15 0.2 0.25 0.3

2H

18O

Fraction of residual pore water, fp [–]

1.025

1.0049

P/GM/50


AGEING EFFECT ON THE CHEMICAL ANDISOTOPIC (SR AND AR) CHARACTERISTICSOF CORED MONT TERRI OPALINUS CLAYS

I. Techer1, N. Clauer2, N. Liewig3

1. Laboratoire GIS / CEREGE / UMR 6635, 150 rue Georges Besse, 30035 Nîmes cedex 1, France([email protected])

2. Centre de Géochimie de la Surface / UMR 7517 / ULP-CNRS, 1 rue Blessig, 67084 Strasbourgcedex, France ([email protected])

3. Institut Pluridisciplinaire Hubert Curien / UMR 7178 / ULP-CNRS, 23 rue Becquerel, 67087Strasbourg cedex, France ([email protected])

INTRODUCTION

The study of deeply sitting geological formations requires core drilling and shaft excavation to access them.These underground openings induce perturbations and favor the development of so-called ExcavatedDamaged Zones (EDZ) and Excavated disturbed Zones (EdZ) that evolve in time and space around thegalleries, niches (Bossart et al., 2002; Mayor et al., 2006; Matray et al., 2007), and drill holes to a lesserextent. Detailed studies have outlined the formation of fracture networks in the EDZ, and tiny butsignificant changes resulting from mineral/organic reactions monitored by desaturation, dehydration/rehydration, decompression, changes in temperature, pH, redox,… in the EdZ. On the other hand, whensamples are cored or excavated and brought to surface, they undergo also changes generally considered tobe very discreet. Thus, the most difficult in the evaluation of the host-rock properties is to ensure that insitu alteration effects of EDZ and EdZ are avoided and that the analyses and observations made on coresin the laboratory after extraction and varied periods of storage are best representative of the originalcharacteristics of the rock formation.

The goal of this study is the characterization of the geochemical behavior of core samples from OpalinusClays of the Mont-Terri Underground Rock Laboratory after air-drilling and storage in the laboratoryduring several months, either as core fragments or as crushed powders. Geochemical (major and tracecontents, and Sr and Ar isotopic systematics) and morphologic (SEM observations) data were examined(1) to quantify the changes relative to excavation and yearly storage, and (2) to possibly identify thereactions and the kinetics of the processes that induced these reactions.

SAMPLING AND STORAGE CONDITIONS

In December 2003, an 8-m long core was drilled perpendicularly to the MI niche of the Mont TerriUnderground Rock Laboratory excavated in 1998, to collect samples away from the EDZ and desaturatedzone induced by this construction, that are limited respectively to 1 and 1.5m behind the niche wall(Wileveau and Rothfuchs, 2003). Height samples were directly collected on the core and were stored inAl-bags under slight vacuum. A first set of 5 samples was analyzed about 4 months after drilling andconditioning. The powders were stored and analyzed again 21 months later. A second set of 3 samples wastreated and analyzed 25 and 32 months after drilling and conditioning.

All samples were leached with dilute HCl in order to identify the elements that became mobile, and madeavailable, either to migrate or to recombine with others. The contents of the major and trace elements ofthe whole rocks and of some of the separated clay fractions (<2 µm) were determined and comparedrelative to the storage duration, as well as the Sr isotope composition of the soluble mineral phases of therocks. The K-Ar data of the rocks and some of the clay fractions were also determined for tracing the tinychanges, especially on the K-rich mineral phases.

P/GM/50



RESULTS AND INTERPRETATIONThe aim of this study was, in a first step, to identify the changes in the elemental composition of the rocksrelative to their ageing. However, we also observed significant chemical variations along the studied coreaccording to the distance to the MI niche. These changes can be explained in part by a variation of thelithology of the studied rocks, but need for some of them to call for damages induced by the previous nicheopening. SEM observations showing pyrite alteration as well as chemical measurements pointing to anincrease of the whole rock ignition loss support this interpretation and suggest an opening damage zone(EdZ) extending at least over 2 to 4-m from the niche. The elementary evolution of the rock soluble phasein this damaged zone is similar to that measured on the samples after storage.

Analyses performed on the same samples before and after storage, either as ships or as powders, outlinesignificant chemical and isotopic changes that are definitely not related to variations in the rock lithology.These changes concern especially the soluble phase of the rock, that is to say those of its pore system, asit is expected to be the most labile phase. The leachates are significantly different before and after storagewith notably lower K, Al, Si and at a lesser extent Fe contents and higher Na and Mg contents relative tostorage time. It seems that the changes induced by lasting exchanges with air alter mainly the salts, knownto occur commonly in the pore system. The same changes are observed close to the MI niche wall over a4-m thick zone. About the behavior of the leached Ca and Sr, it is interesting to notice that they remainquite similar for all analyzed samples whatever the storage duration and whatever the location along thestudied core. Alternatively, the 87Sr/86Sr ratio of this soluble phase decreases towards the niche wall, toreach a minimum value identical for all stored samples, whatever its initial ratio and distance to the wall.These disturbances are, at least partly, monitored by pyrite alteration and crystallization of gypsum (SEMobservations), and probably some oxi-hydroxides. However, other processes have to be considered toaccount for all the data.

INTERPRETATION AND CONCLUSIONSIt appears from a detailed study of samples from Mont-Terri Opalinus Clays, that dry-air drilling andstorage both introduce a (chemical and physical) disturbance in the pore system of the cored rocks thatmodifies the behavior of elements such as Na, K andMg, and of the Sr and Ar isotope compositions. Thesechanges seem to be identical in the in situ EdZ and in the stored samples. They are in part related toalteration of pyrite into gypsum, probably due to a dehydration/rehydration process such as that describedat the surface of building rocks. But for other parameters, including the 87Sr/86Sr ratio, the explanation isyet less straightforward, even if interference of metastable oxi-hydroxides in the pore system is possible.Processes like (1) overpressure induced by drilling and/or (2) decompression and desaturation afterextraction and storage are discussed. This study raises the difficulty to evaluate the chemical properties ofunderground rock in situ and when cored and brought rapidly to the surface, and to qualify therepresentational degree of such extracted samples.

ACKNOWLEDGMENTSThis study was completed in the frame of the GdR FORPRO (CNRS-Andra) and the NF-PRO IntegratedProject (VI PCRD). We thank these organizations for financial support.

ReferencesBossart, P., Meier, P.M., Moeri, A., et al. (2002) Geological and hydraulic characterisation of the excavationdisturbed zone in the Opalinus Clay of theMont Terri rock laboratory. Eng. Geol., 66, 19-38.

Matray, J.M., Savoye, S., Cabrera, J. (2007) Desaturation and structure relationships around driftsexcavated in the well-compacted Tournemire’s argillite (Aveyron, France). Eng. Geol., 90, 1-16.

Mayor, J.C., Velasco, M., Garcia-Sineriz, J.L. (2006) Ventilation experiment in the Mont Terriunderground laboratory. J. Phys. Chem. Earth, doi: 10.1016/j.pce.2006.04.030.

Wileveau, Y., Rothfuchs, T. (2003) HE-D Experiment: test plan. Mont Terri Project Technical Note 2004-20. 65 pp.

P/GM/51


EFFECT OF CONCENTRATION OF THECOUPLES OF CATION (CU2+, PB2+) ON THENATURAL SELECTIVITY PHENOMENA INTHE CASE OF NA-MONTMORILLONITE:

QUALITATIVE XRD ANALYSIS

W. Oueslati 1, M S. Karmous1, H. Ben Rhaiem1, B. Lanson2, A. Ben Haj Amara1

1. PMLNMH : Laboratoire de Physique des Matériaux Lamellaires et Nano Matériaux Hybrides,Faculté des Sciences de Bizerte, Tunisia ([email protected])

2. LGIT : Equipe Géochimie de l’Environnement, Maison des Géosciences, Université JosephFourier BP 53, 38041 Grenoble Cedex 9, France.

INTRODUCTION

Contamination of soils and sediments with heavy metal ions is an important concern in many parts of theworld. Several remediation technologies have been developed and implemented in recent years to clean upthe enormous number of heavy metal contaminated sites that threaten the health of ecosystems (Conner &Hoeffner, 1998). In the grounds, the ETM are distributed between the solid phase and the liquid phases.Generally, the quantity existing in the soil solution represents only one negligible percentage of the totalityof the pollutant (_ 10-1M to 10-4M). Metals thus concentrate in the solid fraction, distributed in the variousorganic and especially clay mineral fractions.

Montmorillonite, is a clay material mostly constituted of smectite, is considered a promising material asan engineered barrier in the context of municipal waste disposal sites because of its low permeability whencompacted and its cation retention ability (Ferrage et al, 2005). Most studies on the swelling of bentonitehave focused on the change in microscopic swelling related to the basal spacing of smectite (Sato et al,1992) Previous tasks were dealing with the interaction between charged cation and this clay were mainlybased on monovalent cation. Indeed, the solutions coming from the household trash contain more thanmono ionic, bi ionic and even more cation which have variable concentrations. Thus we opt for studyingthe dioctahedral clay behaviour and her selectivity in presence of bivalent cations: (Cu2+, Pb2+) byqualitative XRD analysis.

EXPERIMENTAL CONCEPT

The clay fractions were prepared according to the classic protocol of extraction. The <2_m fraction of aCrook montmorillonite (Wyoming, USA). Its half cell structural formula, as obtained by electronmicroprobe is (Ben Rhaiem et al, 1998):

The Na saturated clay minerals were prepared by washing suitable amounts of the clay suspension fivetimes in 1N NaCl. This step aims to disperse as well as possible the clay fraction. The samples were thenwashed several times with a 1N respectively leads chloride and cooper chloride. This step aims to preparea references sample named respectivelyWy-Na, Wy-Cu andWy-Pb. The excess of chloride was eliminatedby washing the clay in distilled water and by subsequent dialysis. An oriented preparation was preparedby depositing a clay suspension on to a glass slide. For qualitative identification, XRD patterns wererecorded on air-dried using D8 advanced Bruker installation, by reflection setting, using Cu-K_ radiation.

( ) + +Si Al Al Ti Fe Fe3 923 0 077 1 459 0 0184

0 0393

, , , , , 00 0452

0 3822

10 2 0 1772

0 027, , , ,+ + +( ) ( )Mg O OH Ca Na++( )

P/GM/51



RESULTS AND INTERPRETATIONHomogeneous sample: The XRD patterns produced respectively by Wy-Na, Wy-Cu show two basalreflections respectively at 2θ=7, 31° (d001 =12,29Å) and 2_=7, 21° (d001=12,36Å). The rationality of thereflexion positions for these samples indicates a homogeneous character. The XRD patterns produced byWy-Pb complex shows a basal spacing d001=13.15Å at 2θ=6.8° for this sample we note an irrationalreflexion position and exhibits little asymmetry of the 001 reflection to lower angles indicating aninterstratified character. We can deduce similar values of d001 basal spacing for Wy-Na and Wy-Cucomplexes but a small difference with Wy-Pb complexes related to her ionic radius.

STRUCTURAL EVOLUTION AND SELECTIVITY OF SAMPLESVERSUS THE CONCENTRATION OF SOLUTION IN THE CASEOF MIXTURE CONTAINING 50% CU (II), 50% PB (II)We present in (Fig.1) the different XRD patterns obtained for decreasing normality (i.e.10-2N→10-4N) inthe case of mixture containing (50% Cu (II), 50% Pb(II)). In the case of normality ≈ 10-2 N, we note thepresence of an overlapping in the001 reflection characterized by twobasal spacing distance respectivelyd00l=14,59Å indicating evolutiontowards 2W layer hydration state (i.e.two water layer in the interlamellarspace) and d00l=12,47Å attributed to Pbcation characterized by 1W layerhydration state (i.e. one water layer inthe interlamellar space). We can inter-pret this result by a minor contributionof two water layer characterised by ad001 ≈ 15.4 Å and a major contributionof one water layers with d001 ≈ 12.4Å(Sato et al.,1992).by exploiting the 2θrange diffraction we note an irrational-ity of the reflexion position for thesesamples. there after we can deduce aninterstratified character and tendancy to fixes the Pb (II) cation wich have high hydration spher. Bydecreasing normality to 10-3N, XRD patterns showed apparition of reflection at 2_=7.11°(d001=12,40Å)indicating 1W hydration state related to Cu(II) cation and an irrational position reflection indicating theinterstratified character. At low normality ≈ 10-4N, XRD patterns showed apparition of reflection at2θ=7.06° (d001=12,36Å) signifying presence of Cu (II) cation with 1 W layer hydration state. The irratio-nality of position reflection indicated interstratified character. This means that at low concentration the clayCEC is saturated with Cu2+cation which is characterised by a weak ionic radius, for high concentration, theclay fixes the cation with high ionic radius (i.e. Pb2+) which are characterised by high hydration state.

ReferencesJR Conner, SL Hoeffner (1998). Critical Reviews in Environmental Science & Technology 28(4),397–462.

Ferrage, E., Lanson, B., Sakharov, B. A. (2005), American Mineralogist 90, 1358-1374.

Sato T., Watanabe T., Otsuka R. (1992). Effects of layer charge, charge location, and energy change onexpansion properties of dioctahedral smectites. Clays & Clay Minerals, 40, 103-113.

H. Ben Rhaiem, D. Tessier, C. H. Pons, A. Ben Haj Amara (1998). Evolution of the microstructure ofinterstratified Ca-saturated clays during dehydration; SAXS and HRTEM analysis Clay Minerals 33/4,619-628.

Figure 1: XRD patterns obtained for decreasing normality(i.e.10-2N_10-4N) for a mixture containing (50% Cu (II), 50% Pb(II)).

10 20 30

0

20000

40000

10-4N

10-3N

10-2N

12.36

12.40

14.59≈

a.u

°2θ (Cu-Kα)

12,47A°

A°

A°

P/GM/52


QUALITATIVE STUDY OF SYNTHETICNA-HECTORITE SATURATED BY HEAVY

METALS CATIONS

M.S. Karmous1*, W. Oueslati1, M. Meftah1, H. Ben Rhaïem1, J. L. Robert2, A. Ben Haj Amara1.

1. Laboratoire des Physiques des Matériaux et Nanomatériaux Hybrides (LPMNH), Facultédes Sciences de Bizerte, Tunisia, ([email protected]).

2. Institut de Minéralogie et de Physique des Milieux Condensés (IMPMC), Campus Boucicaut.140 Rue de Lourmel, 75015 Paris. France.

INTRODUCTIONClays, or layered silicates, have been widely studied and discussed as geological materials, andmodern industrial uses of clays are based on this history and knowledge (Hummel,1998. Velde, 1992).Clays are layered solids, which can be classified into three groups according to the rigidity of thelayers with respect to distortions involving atomic displacement transverse to the layer plane (Lee etal ., 1991). This classification divides the layered solids into three classes. Class I describes layeredsolids that are composed of monatomic thin planes of atoms and can easily sustain long-wavelengthdeformations transverse to the layer planes. Examples include graphite and boron nitride. Class IImaterials are typically composed of three distinct planes of strongly bonded atoms and are moreresistant to transverse distortions than the previous class. Some such materials are layerdichalcogenides and metal chlorides. Clays are class III materials, where the layers can be quite thick.They are the stiffest of lamellar solids, and quite rigid against transverse layer distortions. Most of thenatural clays are very heterogeneous mixtures of minerals, and this can cause both experimental andtheoretical difficulties. However, pure clays can now be made synthetically. Hectorite is a 2:1phyllosilicate, meaning that the platelets are formed by two inverted silicate tetrahedral sheets, sharingtheir apical oxygens with one octahedral sheet sandwiched in between. It is classified as atrioctahedral smectite since Li1+ substitutes for Mg2+ in the octahedral sheet sites, which are fullyoccupied. In one hand, one of the important issues in mineral clay surface analysis is the retention ofmetals on clay surfaces. It was widely studied on illite and other clays (Echeverria at al., 2002; Alvarez–Puebela et al., 2002). In other hand the chemical forms of metals newly incorporated in soil are wellknown. For example, lead emitted from metallurgical industry is mainly in the form of particulatesulphate (Sobanska et al., 1999), (Bataillard et al., 2003) studies the behaviour and fate of trace metals,in particular lead and cadmium.

In this present paper, we will determine the structural properties of the hectorite exchanged with thedifferent heavy metals cations namely Cobalt, Nickel, Lead, Cadmium, Magnesium and Zinc and find outthe abundance and the probability of each phase if the complexes are not totally homogenous, that meansthey present an interstratification character.

EXPERIMENTAL CONCEPTSynthesis: The synthetic hectorite sample was prepared by hydrothermal treatment of hydrolyzed gelsprepared by coprecipitation of Na, Mg, Al, and Si hydroxides at pH = 14, according to a slightly modifiedversion of the gelling method of Hamilton and Henderson, 1968. The source of Na was sodium carbonate,the sources of Al and Mg were titrated solutions of their nitrates and the source of Si was (C2H5O)4Si(TEOS). This resulting gel is slowly dried up to 200°C. It is then calcined at 600°C by further temperatureincrease. . It is then introduced in Morey type externally heated pressure vessels in which the samples areinsulated from the vessel wall by a silver coating. The hydrothermal reactor was then heated at 400 °Cunder a 1000 bar water pressure. Samples were recovered after four weeks. The started synthetic materialhas a structural formulae: [Na0.4]

inter[Mg2.6Li0.4]oct[Si4]

tetO10(OH)2, x H2O (x is the number of water molecule

P/GM/52



per cation). Hectorite Cd2+, Pb2+, Ni2+,Zn2+ , Mg2+and Co2+ complexes were prepared by conventional ionsexchanges reactions using respectively aqueous solutions of 0.1M of CoCl2 , CdCl2, PbCl2 , MgCl2 ,ZnCl2and NiCl2. Removal of excess chloride was performed by washing in distilled water until a negativeAgNO3 test was obtained; the solids were deposed on glass slide to obtain an oriented aggregate; thesamples are referred as H-Cd and H-Co, H-Ni, H-Zn, H-Mg and H-Pb.

X-ray: XRD patterns were recorded using a Bruker D8-advance using Cu-Ká radiation (1.5406Å). Datawere recorded in the range 5-50° 2θ with a step of 0.02°2θ and 80 s per step.

RESULTS AND INTERPRETATIONFig.1 shows the evolution of the d(001) values measured on the experimental XRD patterns.

The qualitative survey of diffractions patterns shows a dissymmetry in most of these diffractograms withregard to the first order and it is very remarkable for the two complex H-Ni and H-Pb (Fig.1- d & e), forall the complexes the basal distances are all inferior to 15 Å except the H-Zn complex (Figure1-f) wherethe d(001) is equal to 15.58 Å. For superior orders, it is clear the presence of the 002-003-004-005reflections (H-Zn), the 002 reflection decreases for the H-Cd, H-Co and H-Mg (figure1-a, b &c) and thisreflection disappears for the H-Ni and H-Pb (Fig.1- d & e), the 003-004 and 005 reflections decreases forall the complexes H-Cd, H-Co (fig.1-a, b) and these orders had a weak intensity for the H-Mg, H-Ni andH-Pb (Fig.1-c, d & e). We can conclude that all the complexes presents a bilayer water in the interlayerspace; now we will try to determine the number and the position of the water molecules surrounding thedifferent cations and show if our complexes are homogenous or not.

References:Hummel, R.E. (1998): Understanding Materials Science .Springer- Verlag, New York.

Velde, B. (1992): Introduction to Clay Minerals. Chapman and Hall, London.

Lee . S and. Solin, S.A (1991) .Phys. Rev. B, 43, 12 012.

Echeverria, .J. C. Churio, E. Garrido, J. J. 2002, Clays Clay Miner, 50. 614.

Alvarez –Puebela, R. A. Aisa, C. Blasco, J. Echeverria, J.C. Mosquera, B. Garrido, J. J. (2004): Appl.Clay Sci., 25,103.

Sobanska, S., Ricq, N., Laboudigue, A., Guillermo, R.., Brémard, C, Laureyns, J (1999): EnvironmentalScience and Technology, 33, 1334-1339.

Bataillard. P, cambier. P, Picot. C (2003): European journal of soil science, 54, 365-376.

Hamilton, D. L.; Henderson, C. M. B, (1968): Mineralogy Mag., 36, 832.

Figure 1: XRD patterns of the Hectorite saturated: a) H-Cd, b) H-Co, c) H-Mg, d) H-Ni, e) H-Pb and f)H-Zn.

295 296 Poster GM.qxd - Andra session_GM.pdf · p/gm/2 international meeting, september 17...>...18, 2007, lille, france page 299 clays in natural & engineered barriers for radioactive

Documents