springer proceedings in physics 112

springer proceedings in physics91 The Dense Interstellar Medium

in GalaxiesEditors: S. Pfalzner, C. Kramer,C. Straubmeier, and A. Heithausen

92 Beyond the Standard Model 2003Editor: H.V. Klapdor-Kleingrothaus

93 ISSMGEExperimental StudiesEditor: T. Schanz

94 ISSMGENumerical and Theoretical ApproachesEditor: T. Schanz

95 Computer Simulation Studiesin Condensed-Matter Physics XVIEditors: D.P. Landau, S.P. Lewis,and H.-B. Schuttler

96 Electromagnetics in a Complex WorldEditors: I.M. Pinto, V. Galdi,and L.B. Felsen

97 Fields, Networks,Computational Methods and Systemsin Modern ElectrodynamicsA Tribute to Leopold B. FelsenEditors: P. Russer and M. Mongiardo

98 Particle Physics and the UniverseProceedings of the 9th Adriatic Meeting,Sept. 2003, DubrovnikEditors: J. Trampetic and J. Wess

99 Cosmic ExplosionsOn the 10th Anniversary of SN1993J(IAU Colloquium 192)Editors: J. M. Marcaide and K. W. Weiler

100 Lasers in the Conservation of ArtworksLACONA V Proceedings,Osnabruck, Germany, Sept. 15–18, 2003Editors: K. Dickmann, C. Fotakis,and J.F. Asmus

101 Progress in TurbulenceEditors: J. Peinke, A. Kittel, S. Barth,and M. Oberlack

102 Adaptive Opticsfor Industry and MedicineProceedingsof the 4th International WorkshopEditor: U. Wittrock

103 Computer Simulation Studiesin Condensed-Matter Physics XVIIEditors: D.P. Landau, S.P. Lewis,and H.-B. Schuttler

104 Complex Computing-NetworksBrain-like and Wave-orientedElectrodynamic AlgorithmsEditors: I.C. Goknar and L. Sevgi

105 Computer Simulation Studiesin Condensed-Matter Physics XVIIIEditors: D.P. Landau, S.P. Lewis,and H.-B. Schuttler

106 Modern Trends in GeomechanicsEditors: W. Wu and H.S. Yu

107 Microscopy of Semiconducting MaterialsProceedings of the 14th Conference,April 11–14, 2005, Oxford, UKEditors: A.G. Cullis and J.L. Hutchison

108 Hadron Collider Physics 2005Proceedings of the 1st HadronCollider Physics Symposium,Les Diablerets, Switzerland,July 4–9, 2005Editors: M. Campanelli, A. Clark,and X. Wu

109 Progress in Turbulence 2Proceedings of the iTi Conferencein Turbulence 2005Editors: M. Oberlack et al.

110 Nonequilibrium Carrier Dynamicsin SemiconductorsProceedingsof the 14th International Conference, July25-29, 2005, Chicago, USAEditors: M. Saraniti, U. Ravaioli

111 Vibration Problems ICOVP 2005Editors: E. Inan, A. Kiris

112 Experimental UnsaturatedSoil MechanicsEditor: T. Schanz

113 Theoretical and NumericalUnsaturated Soil MechanicsEditor: T. Schanz

Volumes 66–90 are listed at the end of the book.

Experimental UnsaturatedSoil Mechanics

123

T. Schanz (Ed.)

With 299 Figures and 67 Tables

ISSN 0930-8989

ISBN 978-3-540-69872-2 Springer Berlin Heidelberg New York

Library of Congress Control Number: 2007920611

This work is subject to copyright. All rights are reserved, whether the whole or part of the material isconcerned, specif ically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,reproduction on microf ilm or in any other way, and storage in data banks. Duplication of this publication orparts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in itscurrent version, and permission for use must always be obtained from Springer-Verlag. Violations are liableto prosecution under the German Copyright Law.

Springer is a part of Springer Science+Business Mediaspringer.com

© Springer-Verlag Berlin Heidelberg 2007

The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,even in the absence of a specif ic statement, that such names are exempt from the relevant protective laws andregulations and therefore free for general use.

Printed on acid-free paper SPIN: 11661139

Typesetting: Integra Software Services Pvt. Ltd., Pondicherry, IndiaCover design:

42/Integra 5 4 3 2 1 0

Professor Dr. Ing. habil. Tom SchanzBauhaus-Universitat WeimarLaboratory of Soil MechanicsCoudraystrasse 11c99421 Weimar, Germanytom.schanz@bauing.uni-weimar.de

WMX-Design, Heidelberg

Foreword

The event is a continuation of the series of International Conferences on Un-saturated Soils in Germany. The first International Conference was held dur-ing September 2003 in Bauhaus-University Weimar, Weimar, Germany. Thecurrent event is the second one in the series entitled “Mechanics of Unsat-urated Soils.” The primary objective of the Conference has been to discussand understand unsaturated soil behaviour such that engineered activities aremade better with times in terms of judgement and quality. We all realise bynow that in addition to the knowledge on the classical concepts, it becomesan enormous challenging task to adapt convincing new concepts and presentthem in such a way that it could be used in engineering practices. Duringthe last six years or so (2001–2007), scientific research works were exten-sively taken up by five scientific research teams from five German universities,whose scientific leaders are Wolfgang Ehlers (Universität Stuttgart), Jens En-gel (HTW Dresden), Rainer Helmig and Holger Claas (Universität Stuttgart),Tom Schanz (Bauhaus Universität Weimar), Christos Vrettos, Helmut Meiss-ner and Andreas Becker (Universität Kaiserslautern). The research studiesinvolved theoretical and numerical approaches along with experimental stud-ies on unsaturated soils. These two volumes present recent research findingsobtained within this collaboration by the above research groups along withexcellent contributions from several research groups throughout the World.

The experimental studies reported herein primarily focussed on the roleof microstructure and fabric for the complex coupled hydro-mechanical be-haviour of cohesive frictional materials. Several papers considered the rele-vance of temperature affecting the constitutive behaviour of clays. A carefulreader may recognise that in both the topics there is an ambiguity with re-gard to the conclusions derived. Common features of state of the art theoret-ical and numerical approaches, including TPM (theory of porous media) andmixture theory, intend to describe the complex multi-field problems of fullycoupled thermo-hydraulic-mechanical-chemical initial-boundary value prob-lems. Additional important field of research includes optimization of numer-ical schemes to gain better computational performance. Applications include

VI Foreword

highly toxic waste disposals, slope stability problems and contaminants trans-port in porous media. Some major significant contributions from the invitedand keynote speakers are also included.

I would like to extend my deep sense of appreciation as the editor andthe Head of the organizing committee, to many persons who have contributedeither directly or indirectly to organize the International conference and tofinalize these lecture notes. I would like to congratulate the authors for theirvery interesting presentations and the reported results and advances in thetopics of the conference. I would like to thank all of those who promoted theconference in their respective home countries. These two volumes would havebeen not possible without financial support by the German Research Founda-tion (DFG, Deutsche Forschungsgemeinschaft) through grant FOR 440/2. Wegratefully acknowledge the support of ISSMGE, especially TC6 “UnsaturatedSoils” with its chairman Eduardo Alonso. I appreciate the effort of the mem-bers of the Technical committee and reviewers, who have spent their time toselect the valuable contributions and to suggest the changes improving thepresentation of the submitted papers. Finally, I wish to convey my thanksto all the keynote and invited speakers, authors, and delegates attending theconference.

I would like to express my deep sense of gratitude for the outstandingwork performed by those involved in the technical and administrative organi-zation of these proceedings. Special thanks go to Yvonne Lins. Typesetting ofthe proceedings was done by Venelin Chernogorov (alias Wily, Sofia Univer-sity, Bulgaria, wily@fmi.uni-sofia.bg) in cooperation with Maria Datcheva.Last but not least we appreciate the fruitful cooperation with Springer pub-lishers, especially the guidance provided by Thomas Ditzinger.

Weimar, Tom SchanzMarch 2007

Contents

Part I Microstructure and Fabric

Influence of Relative Density and Clay Fraction on SoilsCollapseKhelifa Abbeche, Farid Hammoud, and Tahar Ayadat . . . . . . . . . . . . . . . . . 3

Microstructure Features in the Behaviour of EngineeredBarriers for Nuclear Waste DisposalPierre Delage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Microstructure of Gypsiferous Crust and Its Importance toUnsaturated Soil BehaviourGhazi Mokdad, Omran Alshihabi, and Leo Stroosnijder . . . . . . . . . . . . . . . . 33

Fabric Changes in Compacted London Clay Due to Variationsin Applied Stress and SuctionRafael Monroy, Lidija Zdravkovic, and Andrew Ridley . . . . . . . . . . . . . . . . 41

Microstructure of a Lime Stabilised Compacted SiltGiacomo Russo, Sebastiana Dal Vecchio, and Giuseppe Mascolo . . . . . . . . 49

Part II Measuring Suction

Errors in Total Suction MeasurementsSetianto Samingan Agus and Tom Schanz . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Application of a Dew Point Method to Obtain the Soil WaterCharacteristicGaylon S. Campbell, David M. Smith, and Brody L. Teare . . . . . . . . . . . . . 71

VIII Contents

A Comparative Study of Soil Suction Measurement UsingTwo Different High-Range PsychrometersRafaela Cardoso, Enrique Romero, Analice Lima, and Alessio Ferrari . . 79

Determination of the Soil Water Retention Curve withTensiometersSérgio Lourenço, Domenico Gallipoli, David Toll, Fred Evans, andGabriela Medero . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Tensiometer Development for High Suction Analysis inLaboratory LysimetersCláudio Fernando Mahler and Abdoul Aziz Diene . . . . . . . . . . . . . . . . . . . . 103

Part III Strength and Dilatancy

Dilatancy of Coarse Granular AggregatesEduardo E. Alonso, Enrique F. Ortega Iturralde, and EnriqueE. Romero . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

A Laboratory Investigation into the Effect of Water Contenton the CBR of a Subgrade soilSamuel Innocent Kofi Ampadu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

Shear Strength Affected by Suction Tension in UnsaturatedFine Grained Soils?Carola Bönsch and Christof Lempp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

Shear Strength Behaviour of Unsaturated Silty SoilAli R. Estabragh and Akbar A. Javadi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

Experimental Investigation on the Time Dependent Behaviourof a Multiphase ChalkGrégoire Priol, Vincenzo De Gennaro, Pierre Delage, and ThibautServant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

Testing Unsaturated Soil for Plane Strain Conditions: A NewDouble Wall Biaxial DeviceTom Schanz and Jamal Alabdullah . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

Influence of State Variables on the Shear Behaviour of anUnsaturated ClayViktoria Schwarz, Andreas Becker, and Christos Vrettos . . . . . . . . . . . . . . 179

Effect of Capillary and Cemented Bonds on the Strength ofUnsaturated SandsFabien Soulié, Moulay Saïd El Youssoufi, Jean-Yves Delenne, andChristian Saix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

Contents IX

Determining the Shear Strength of Unsaturated SiltShulin Sun and Huifang Xu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

Factors Affecting Tensile Strength Measurement and ModifiedTensile Strength Measuring Apparatus for SoilSurendra Bahadur Tamrakar, Toshiyuki Mitachi, and Yasuo Toyosawa . . 207

The Tensile Strength of Compacted Clays as Affected bySuction and Soil StructureRainer M. Zeh and Karl Josef Witt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

Part IV Temperature Effects

Modified Isochoric Cell for Temperature Controlled SwellingPressure TestsYulian Firmana Arifin and Tom Schanz . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

Some Aspects of the Effect of the Temperature on theBehaviour of Unsaturated Sandy ClayMoulay Smaine Ghembaza, Said Taïbi, and Jean-Marie Fleureau . . . . . . . 243

Influence of Temperature on the Water Retention Curve ofSoils. Modelling and ExperimentsSimon Salager, Moulay Saïd El Youssoufi, and Christian Saix . . . . . . . . . 251

Thermo-Hydro-Mechanical Behaviour of CompactedBentoniteAbbass Tavallali, Anh-Minh Tang, and Yu-Jun Cui . . . . . . . . . . . . . . . . . . . 259

Retention Curves of Two Bentonites at High TemperatureMaría Victoria Villar and Roberto Gómez-Espina . . . . . . . . . . . . . . . . . . . . 267

Part V Volumetric Behaviour – Expansive Materials

Experimental Study on Shrinkage Behaviour and Predictionof Shrinkage Magnitudes of Residual SoilsSarita Dhawan, Anil Kumar Mishra, and Sudhakar M Rao . . . . . . . . . . . . 277

Assessment of Swelling Deformation of Unsaturated KaoliniteClayMarkus Dobrowolsky and Christos Vrettos . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

Suction and Collapse of Lumpy Spoilheaps in NorthwesternBohemiaVladislava Herbstová, Jan Boháč, and Ivo Herle . . . . . . . . . . . . . . . . . . . . . 293

.

X Contents

Oedometer Creep Tests of a Partially Saturated KaoliniteClayPiotr Kierzkowski . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

Analysis of the Expansive Clay Hydration under LowHydraulic GradientMarcelo Sánchez, María Victoria Villar, Antonio Lloret,and Antonio Gens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

Moisture Effects on Argillaceous RocksChun-Liang Zhang and Tilmann Rothfuchs . . . . . . . . . . . . . . . . . . . . . . . . . . 319

Part VI Retention Behaviour

Results from Suction Controlled Laboratory Tests onUnsaturated Bentonite – Verification of a ModelAnn Dueck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329

Variation of Degree of Saturation in Unsaturated Silty SoilAli R. Estabragh and Akbar A. Javadi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

Mechanical Behaviour of Compacted Scaly Clay DuringCyclic Controlled-Suction TestingCamillo Airò Farulla, Alessio Ferrari, and Enrique Romero . . . . . . . . . . . 345

Prediction of Soil–Water Characteristic Curve Based on SoilIndex PropertiesNavid Ganjian, Yadollah Pashang Pisheh, and Seyed Majdeddin MirMohammad Hosseini . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355

Water Balance and Effectiveness of Mineral Landfill Covers –Results of Large Lysimeter Test-FieldsWolf Ulrich Henken-Mellies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369

A Retention Curve Prediction for Unsaturated Clay SoilsMehrez Jamei, H. Guiras, and N. Mokni . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377

Unsaturated-Zone Leaching and Saturated-Zone MixingModel in Heterogeneous LayersSamuel S. Lee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387

Prediction of SWCC for Coarse Soils Considering Pore SizeChangesXu Li and Limin Zhang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401

Contents XI

The Influence of the Pore Fluid on Desiccation of aDeformable Porous MaterialHervé Péron, Liangbo Hu, Tomasz Hueckel, and Lyesse Laloui . . . . . . . . . 413

Determination of the Soil Water Retention Curve and theUnsaturated Hydraulic Conductivity from the Particle SizeDistributionAlexander Scheuermann and Andreas Bieberstein . . . . . . . . . . . . . . . . . . . . 421

Part VII Field Applications

Earthquake-Induced Mudflow Mechanism from a Viewpointof Unsaturated Soil DynamicsMotoki Kazama and Toshiyasu Unno . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437

Plate-Load Tests on an Unsaturated Lean ClayJuan Carlos Rojas, Luis Mauricio Salinas, and Claudia Sejas . . . . . . . . . . 445

Selfsealing Barriers of Clay/Mineral Mixtures. The SBProject at the Mont Terri Rock LaboratoryTilmann Rothfuchs, Rüdiger Miehe, Norbert Jockwer, and Chun-LiangZhang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453

Preferential Water Movement in Homogeneous SoilsAlexander Scheuermann and Andreas Bieberstein . . . . . . . . . . . . . . . . . . . . 461

Compaction Properties of Agricultural SoilsAnh-Minh Tang, Yu-Jun Cui, Javad Eslami, and PaulineDéfossez-Berthoud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475

Bearing Capacity of Model Footings in Unsaturated SoilsSai K. Vanapalli and Fathi M.O. Mohamed . . . . . . . . . . . . . . . . . . . . . . . . . . 483

Influence of Soil Suction on Trench StabilityValerie Whenham, Monika De Vos, Christian Legrand, Robert Charlier,Jan Maertens, and Jean-Claude Verbrugge . . . . . . . . . . . . . . . . . . . . . . . . . . 495

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503

Part I

Microstructure and Fabric

Influence of Relative Density and Clay Fractionon Soils Collapse

Khelifa Abbeche1, Farid Hammoud1, and Tahar Ayadat2

1 Civil Engineering Department, LARHYA, University of Batna, Avenue ChahidBoukhlouf, Batna 05000, Algeria abbechek@yahoo.fr, hammoud_farid@yahoo.fr

2 Civil Engineering Department, University of Concordia, Sir George WilliamsCampus, 1515 St. Catherine West, EV002. 139, Montreal Quebec, CanadaH3G-2W1 tayadat@yahoo.com

Summary. Most of the works conducted to investigate the parameters likely togovern the behavior of collapsible soils, have been dedicated to initial dry density,water content, degree of saturation and applied load. On the other hand, relativelyfew works have been oriented towards the study of the influence of clay fraction andrelative density on the collapse of this type of soils. The experimental study, aboutreconstituted soils, presented in this paper, aims at highlighting the effect of relativedensity and clay fraction on the rate and magnitude of collapse.

Key words: collapsible soils, relative density, clay fraction, consolidation, oedome-ter test, inundation

1 Introduction

Rapid growth in arid and semiarid regions has brought increased focus onvolume change characteristics of moisture sensitive unsaturated soils. Onesuch type of concern is collapsible soil. The latter is likely to undergo a rear-rangement of its grains, and a loss of cementation, upon wetting, resulting insubstantial and rapid settlement under relatively low loads. Collapsible soilsoccur as naturally debris flows, rapid alluvial depositions, and wind-blowndeposits. These soils are typically silt and sand size with a small amount ofclay. The collapse phenomenon is also likely to occur in the case of compactedfills, used in man made structures such as earth dams, road subgrades, andembankments, which are also placed in an unsaturated state.

Some investigations were aimed at developing identification and stabiliza-tion methods (e.g. Jennings and Knight 1975, Ayadat and Belouahri 1996,Ayadat et al. 1998, Abbeche et al. 2005, etc). On the other hand, some otherresearchers (e.g. Ganeshan 1982, Ayadat et al. 1998, Cui and Magnan 2001,etc) have concentrated their works on the collapse mechanism.

4 K. Abbeche et al.

Most of the work carried out on the parameters that govern the collapsebehavior, has focused on the initial dry density, the water content, the degreeof saturation and the applied load. Few studies have been conducted regardingthe influence of the relative density and the clay fraction on the collapse ofsoils. The main interest of this paper lies in the study of the effect of theseparameters on the soil collapse.

2 Materials, Equipment and Testing Procedure

In the present study eleven reconstituted samples composed of sand and clayin different proportions were tested. The proportions were chosen in order toobtain samples which meet the collapse criteria proposed by Lutenegger andSaber (1988) which distinguish between collapsible soils and non collapsiblesoils.

The sand used is washed river sand, whose characteristics are summarizedas follows:

• sand equivalent, Es = 70%• grain size distribution situated between 0.08 mm and 2.0 mm, of which

3.8% < 0.08 mm• coefficient of uniformity, Cu = 2.5• coefficient of curvature, Cc = 0.56

The clay used has the following characteristics:

• liquidity limit, wL = 41.50%• plasticity limit, wP = 28.90%• specific gravity of soil solids, Gs = 2.7• percentage of particles finer than 2 μm (clay fraction), CF = 32%

The reconstituted soils have the geotechnical characteristics given inTable 1 and the grain size distribution curves presented in Fig. 1.

Table 1. Geotechnical characteristics of soils

Soil S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11

Sand (%) 100 90 80 70 60 50 40 30 20 10 0%PF < 80 μm 0 10 20 30 40 50 60 70 80 90 100emax 0.765 0.773 0.817 0.828 0.846 0.913 0.955 1.007 1.035 1.184 1.472emin 0.458 0.452 0.436 0.408 0.403 0.396 0.383 0.367 0.350 0.333 0.327wL (%) – 13.7 15.1 16.1 19.3 21.4 26.8 34.0 34.3 40.6 41.5wP (%) – 8.4 10.9 11.4 11.9 13.6 15.2 22.0 22.6 23.7 28.9wopt (%) 9.64 7.86 7.94 8.29 8.64 11.08 13.01 14.89 16.48 18.66 22.14

%PF < 80 μm percentage of particles finer than 80 μm

Influence of Relative Density and Clay Fraction on Soils Collapse 5

Fig. 1. Grain size distribution curves of reconstituted soils used

The soils studied was compacted at a given water content and a dry unitweight in a standard oedometric mold, in one layer, due to the small height ofthe ring (20 mm). The equipment used for the compaction procedure, whichwas made at the laboratory, is composed of a disk having a diameter slightlysmaller than that of the ring, which is fixed to a stem of guidance and a diskshaped weight. The weight, having a mass of 152 g, sliding along the stem andthat falls from a 15 cm height, comes to strike the disk, thus compacting thematerial in the oedometer ring.

The initial stage of sample preparation involves mixing the two compo-nents so as to obtain a well homogenized soil. Then it is absolutely necessarythat meticulous preparation of a sample should be carried out to ensure a nearperfect fit in the oedometer ring. This is achieved by first kneading the soilevenly in the oedometric mold and then compacting it with a certain energy(number of weight drops) following the procedure described above. The sam-ple was then struck off level with the top of the ring with a rigid steel blade,to get a plane surface. After weighed, the specimen was put back in positionin the oedometer so as to carry out the compressibility test described by Jen-nings and Knight (1975), for determining the collapse potential. The loadingstage of a sample, at an initial water content and an initial unit weight, tookplace incrementally up to a pressure of 0.2 MPa and then flooded with waterand left for 24 hours, and the consolidation test is carried on to its normalmaximum loading limit. The sample settlements, before inundation and after,at different time intervals had been measured. Figure2 shows the stages of thecollapse potential test. The collapse potential Cp is then defined as follows:

CP (%) =ΔeC

1 + e0× 100 (1)

6 K. Abbeche et al.

Fig. 2. Typical collapse potential one dimensional consolidation test

where ΔeC is the change in void ratio of the sample upon flooding and e0 isthe initial void ratio before loading.

The tests were conducted on soils at different water contents and relativedensities. The relative density is defined as

Dr =emax − e0emax − emin

(2)

The retained values of initial water content and relative density are:

• relative density, Dr: 10%, 30% and 50%• initial water content, w0: 2%, 4% and 6%

Each soil sample was tested at the values of Dr given above and for eachvalue, the initial water content was varied three times (i.e. 2%, 4% and 6%).As explained below, soils S2, S3, S4 and S5 were also tested at a relativedensity of 70%. Therefore, in all, 111 tests were carried out.

3 Results and Discussion

The results obtained clearly indicate that the collapses of the different soilsare in keeping with the classification of Jennings and Knight (1975) and agreewith the known properties of natural collapsible soils. For soils S1 to S11the collapse potential Cp was found to vary from 0.16% to 12.55% for a watercontent w0 = 2%, from 0.13% to 5.73% for a water content w0 = 4% and from0.045% to 3.63% for a water content w0 = 6%. These results correspond tothe categories ranging from “no problem” to “severe trouble” (Table 2). On theother hand, when the water content increases, the collapse potential decreasesor even can be equal to zero above a certain value. It is also noted that, ata given water content, the collapse potential decreases with the increase ofthe initial unit weight. These results are in agreement with those obtained

Influence of Relative Density and Clay Fraction on Soils Collapse 7

Table 2. Relation of collapse potential to the severity of foundation problem

CP (%) Severity of problem

0–1 No problem1–5 Moderate trouble5–10 Trouble10–20 Severe trouble> 20 Very severe trouble

by Lawton et al. (1989) and Ayadat et al. (1998). Consequently, it is noticedthat the artificially prepared soils possess a behavior analogous to those metin situ, which justifies the tests program adopted.

The influence of the relative density on the collapse potential of soils ispresented in Figs 3a, 3b and 3c. It can be seen that CP decreases whenthe relative density increases, whatever the soil type and the water contentconsidered. It is also noticed that for a given relative density, the collapsepotential decreases when the water content increases. For Dr = 50% andw0 = 6%, Figure 3c shows that except soils S2, S3, S4 and S5, all tested soilshave a value of CP < 1 and are therefore not susceptible to collapse. It is alsonoticed that the initial water content (i.e. 6%) is close to the values of theoptimum water content wopt of the collapsible soils studied (i.e. S2, S3, S4

(a) (b)

(c)

Fig. 3. Changes of collapse potential versus clay fraction, (a) for w0 = 2%, (b) forw0 = 4%, (c) for w0 = 6%

8 K. Abbeche et al.

(a) (b)

(c) (d)

Fig. 4. Collapse potential variation versus relative density, (a) for soil S2, (b) forsoil S3, (c) for soil S4, (d) for soil S5

and S5) given in Table 1. Consequently, it was judged useful to test these foursoils with a more important relative density (Dr = 70%). The results obtainedare presented in Figs 4a, 4b, 4c and 4d. It can be noticed from these figuresthat, for a water content w0 = 6%, the collapse potential is negligible whenthe relative density is superior to 65%. It is also noticed that the more thesoils are loose and have low initial water content and the more the increase ofthe collapse potential is important.

Figures 3a, 3b and 3c, which illustrate also the influence of the clay fractionon CP , show that the collapse of soils depends on the clay content present intheir structure, which confirms the observation made by Lawton et al. (1992).From these curves, it is clear that the collapse potential is negligible when theclay percentage is greater than 30%. Below 5%, a collapse settlement, whichremains small, is likely to take place, while maximum collapse is reached forabout 15%. This result is in keeping with the interval established by Lawtonet al. (1992) who indicated that maximum collapse potential, for the naturalsoils studied, is obtained when of the clay fraction is situated in the range10% and 40%. The classification proposed in Table 3 allows identifying thesoils that are subjected to the collapse phenomenon.

Influence of Relative Density and Clay Fraction on Soils Collapse 9

Table 3. Proposed classification of collapsible soils

Clay Fraction CF (% < 2 μm) Probability of collapse

5% < CF < 15% High probability15% < CF < 30% Probability of collapseCF > 30% No collapse

4 Conclusions

This paper has presented results of one-dimensional oedometer collapse testsusing the procedure described by Jennings and Knight (1975). The main con-clusions that can be drawn on the basis of the test results are as follows:

• It is possible to assess the soil collapsibility in the laboratory by means ofthe relative density, on the one hand, and the clay fraction, on the otherhand.

• It is estimated that a soil is not subjected to collapse if its relative densityis superior to 65% with initial water content close to the optimum watercontent.

• A soil is not susceptible to collapse if its clay fraction is greater than 30%.

References

Abbeche K, Mokrani L, Boumekik A (2005) Contribution à l’identification des solseffondrables, Revue Française de Géotechnique, 110:85–90

Ayadat T, Belouahri B (1996) Influence du coefficient d’uniformité sur l’amplitudeet le taux de l’affaissement des sols, Revue Française de Géotechnique, 76:25–34

Ayadat T, Belouahri B, Ait Ammar R (1998) La migration des particules finescomme approche d’explication du mécanisme de l’effondrement des sols, RevueFrançaise de Géotechnique, 83:1–9

Ayadat T, Ouali S (1999) Identification des sols affaissables basée sur les limitesd’Atterberg, Revue Française de Géotechnique, 86:53–56

Cui YJ, Magnan JP (2001) Affaissements locaux dus à l’infiltration d’eau en géomé-canique environnemental, Hermes, pp. 139–164

Ganeshan V (1982) Strength and collapse characteristics of compacted residual soils.Master Thesis, Asian Institute of Technology, Bangkok

Jennings JE, Knight K (1975) The additional settlement of foundation due to col-lapse of sandy soils on wetting, Proc 4th Inter Conf on Soil Mechanics andFoundation Engineering, 1:316–319

Lawton EC, Fragaszy RJ, James HH (1989) Collapse of compacted clayey sand, JGeotechnical Engineering, ASCE 155(9):1252–1267

Lawton EC, Fragaszy RJ, Hetherington MD (1992) Review of wetting-induced col-lapse in compacted soil, J Geotechnical Engineering 118(9):1376–1394

Lutenegger AJ, Saber RT (1988) Determination of collapse potential of soils,Geotechnical Testing J 11(3):173–178

Microstructure Features in the Behaviourof Engineered Barriers for Nuclear WasteDisposal

Pierre Delage

Ecole Nationale des Ponts et Chaussées, Paris (CERMES, Institut Navier), 6–8 av.B. Pascal, F–77455 Marne–la–Vallée cedex 2, Francepierre.delage@enpc.fr

Summary. Engineered barriers made up of bricks of compacted swelling clays areconsidered as potential barriers for the isolation of high activity nuclear waste atgreat depth. To better understand their coupled hydro-mechanical response, the mi-crostructure of engineered barriers has been studied by various authors by usingmost often scanning electron microscope observations and mercury intrusion poresize distribution measurements (MIP). These studies confirmed that the microstruc-ture of compacted bentonites was made up of aggregates with two classes of pores:inter-aggregates pores and intra-aggregate pores. Further investigation conductedby using X-Ray diffractometry at low angles conducted more recently provideddeeper insight into the hydration mechanisms that occur during hydration. Thispaper presents some results obtained by various authors by using these techniques.It shows how the hydration mechanisms occurring at the level of the clay particlesinside the aggregates help interpreting existing MIP data. Some conclusions aboutthe water retention properties of compacted bentonites with or without swelling al-lowed are given. Some consequences on the water transfer properties of compactedbentonites are also drawn. In both cases, the clogging of inter-aggregate pores dueto particle exfoliation has a significant effect.

Key words: engineered barrier, bentonite, hydration, microstructure, radioactivewaste, swelling, coupling

1 Introduction

Isolation of high activity nuclear waste at great depth is designed accordingthe concept of multibarrier system aimed at safely isolating the waste fromthe biosphere. In this system, the placement of engineered barriers made up ofcompacted swelling clay between the waste canister and the geological barrier(host rock) as shown in Fig. 1 is considered as a possible option.

In this option, bricks of clay are statically compacted at given water con-tent, often to unit masses as high as 1.8–2 Mg/m3. Bricks are placed all around

12 Pierre Delage

Fig. 1. Typical scheme of a deep geological repository for nuclear waste disposal(Gens and Olivella 2001)

the canister as shown in the figure, in order to fill the annular space betweenthe canister and the excavation. Engineered barriers are submitted to com-plex thermo-hydro-chemo-mechanical actions due to the heat emitted by thenuclear waste, to infiltration by pore water coming from the host rock, to themovement of radionuclides through the barrier and to the emission of gas dueto the anaerobic degradation of the canister, among other things. A typicalfeature that characterises the condition of engineered barriers is that they arehydrated with very small swelling permitted in the volume between canisterand host rock. We will see later on that this particular situation is interestingin terms of microstructure changes.

For these reasons, intense investigation has been carried out in the pastdecades on the behaviour of engineered barriers. Researches clearly showedthat microstructure features played an essential role. Some effects of mi-crostructure on the behaviour of engineered barriers are described in thispaper.

2 The Microstructure of Compacted Soils

2.1 Low Plasticity Soils

Following the initial Lambe’s model (1958) of the microstructure of compactedsoils that was based on the double layer theory, experimental observations ofthe microstructure of compacted soils based on scanning electron microscopy(SEM) and mercury intrusion porosimetry (MIP) afterwards provided a morerealistic description of the real microstructure. Pioneering works in this fieldwere provided by Barden and Sides (1970), Diamond (1970), Sridharan et al.

Microstructure Features in the Behaviour of Engineered Barriers 13

(1971). In mercury intrusion porosimetry, mercury, a non wetting fluid, is pro-gressively intruded in a porous medium by increasing its pressure. As shownby Laplace’s equation (1), the applied pressure can be related to an entrancepore radius as follows:

pHg = σ cos θ(

1r1

+1r2

)(1)

where r1 and r2 are the radius of curvature of the interface. Under the hy-pothesis of a cylindrical pore of radius r, the air-water interface is spheri-cal and r1 = r2 = r. Laplace’s relation between the mercury pressure andthe corresponding radius reduces to the following equation (sometimes calledWashburn’s equation, from Washburn 1921):

pHg =2σ cos θ

r(2)

The interfacial capillary parameters of the mercury/solid phase/vacuumsystem are the interfacial tension σ = 0.484 N/m and the contact angle θ =141.3◦ (Diamond 1970). Mercury intrusion tests are performed by running aprogressive mercury pressure increase from vacuum up to 200 MPa. Tests haveto be performed on dehydrated samples. Most often, dehydration is made byfreeze-drying, where the sample is firstly quickly frozen by rapid immersionin cooled liquid nitrogen and then sublimated in a freeze-dryer (Sridharanet al. 1971, Tovey and Wong 1973, Gillott 1973, Delage and Pellerin 1984). Toavoid any boiling when plunging samples, liquid nitrogen needs to be cooledto its freezing point (−210◦C) by applying vacuum. Also, samples have to besmall enough to allow high speed freezing, that leads to a cryptocrystallinestructure with no volumetric expansion.

Ahmed et al. (1974) presented the pore size distribution (PSD) curves ofthree samples of grundite compacted on the dry side, the wet side and atoptimum water content, as shown in Fig. 2.

The aggregate structure already suspected in soils compacted on the dryside of the optimum (Barden and Sides 1970, Diamond 1970) was confirmed bythe bimodal shape of the PSD curve. This curve defines two pore populations,namely the inter-aggregate and the intra-aggregate populations respectively.In the figure, the entrance pore diameter of the inter-aggregate pores is definedby the inflection point at 30 μm, which is quite large. Above the correspondingpressure, mercury fills all the inter-aggregate porosity, finally surrounding theaggregates themselves. Mercury starts to penetrate the aggregates at a pres-sure that corresponds to the entrance diameter of 0.05 μm. On each curve,the degree of saturation of the compacted sample is also plotted, showing theinflection point corresponds to porous volume that was full of water. This con-firms that the inter-aggregate pores in the dry sample are full of air. Resultssimilar to that of Ahmed et al. (1974) have been obtained by Delage et al.(1996) and Romero et al. (1999), among others.

14 Pierre Delage

Fig. 2. PSD curves of a grundite statically compacted at various points of theProctor curve (Ahmed et al. 1974)

A SEM photo of a dry compacted silt of low plasticity (Jossigny silt) ispresented in Fig. 3. To ensure a well defined observation surface, samples havebeen freeze-fractured just before being sublimated. In this technique (Delageet al. 1982), the observation plane is defined by the fracture ice that workslike an impregnation resin. It does not follow any possible weakness path inthe microstructure and directly cross the different microstructure levels.

In Figure 3, the aggregates made up silt grains (10 to 20 μm in diame-ters) are clearly apparent together with the inter-aggregate pores. In the drystate, the clay fraction of Jossigny silt is not very apparent because plateletsare closely stuck to the silt grains. As commented previously, large inter-aggregates pores are full of air and hence physico-chemical effects involvingclay water interactions take place inside the aggregates. Based on this photo,the question of the location of air water menisci is interesting and indeed can

Fig. 3. SEM micrograph of a compacted Jossigny silt, dry side (Delage et al. 1996)

Microstructure Features in the Behaviour of Engineered Barriers 15

Fig. 4. SEM micrograph of a compacted Jossigny silt, wet side (Delage et al. 1996)

be further elucidated by using an environmental scanning electron microscopein which observation is carried out on wet samples and where the ambient rela-tive humidity can be increased. It is probable in Fig. 3 that menisci are locatedinside the aggregates and that the connecting links between aggregates, thatgovern the overall macroscopical resistance, are made up of hydrated claylinks.

A significantly different microstructure is observed on the wet side, asshown in Fig. 4. The microstructure is more of a matrix type with dominantappearance of clay platelets that constitute a matrix containing the silt grains.Neither aggregates nor the corresponding pore population can be observed.In other words, even in low plasticity soils, hydration makes the clay particlesbecome much more voluminous and visually apparent. The unimodal PSDcurve of Fig. 2 (also observed in the Jossigny silt) is compatible with thematrix microstructure.

Going back to MIP investigation, an interesting feature relating mi-crostructure changes when volume is decreased has been obtained from theresults of Sridharan et al. (1971), as presented in Fig. 5 (Delage and Graham1995). The figure, that compares samples statically compacted at w = 21% atvarious void ratios included between 0.85 and 0.56, shows that the decrease involume at constant water content only concerns the largest pores that existprior to compression. With the pore volume saturated by water in smallestpores at w = 21% being represented on the figure, the figure also shows thatcompression affected the pores full of air. The PSD of saturated pores smallerthan 0.2 μm in diameter is remarkably unaffected by the decrease in volume.Note that a similar mechanism had been observed in a natural sensitive clayof moderate plasticity by Delage and Lefebvre (1984). It appears here thatcompression at void ratio smaller than 0.56 would concern a quasi-saturatedmicrostructure with build-up and subsequent dissipation of pore pressures.

16 Pierre Delage

Fig. 5. Effect of increased compaction on the pore size distribution of a compactedsoil (Delage and Graham 1995, after Sridharan et al. 1971)

2.2 Compacted Plastic Clays, Engineered Barriers

In common geotechnical engineering, compaction studies rather concerned lowplasticity soils able to constitute earthworks like embankments or earthdams,with little attention given to plastic soils. This tendency has changed in thepast decade with increased attention given to the behaviour of engineeredbarriers made up of heavily compacted swelling clays considered as possi-ble barrier to be used in the multibarrier concept of isolating at great depthhigh activity nuclear waste. Investigations carried out on engineered barri-ers progressively showed the importance of microstructure effects for a soundunderstanding of the coupled hydro-mechanical behaviour of engineered bar-riers. In this context, the pioneering works conducted on low plasticity claydescribed above have been extended to high plasticity compacted clays.

Microstructure units

Swelling clays (that include bentonites, a term used in practice) are madeup a significant amount of smectites, the active clay mineral responsible forswelling when hydrated. As compared to less plastic clays like kaolinite, thespecificity of smectites (or montmorillonites) is related to the mineralogicalcomposition of the elementary layer, depicted in Fig. 6. Whereas kaoliniteis made up of the juxtaposition of one tetrahedral and of one octahedrallayer (see the figure where the composition and structure of these layers arerepresented), a octahedral layer is located between two tetrahedral layer insmectites. Whereas two adjacent kaolinite particles will be strongly linkedtogether because of possible hydrogen bridges acting between the oxygen andhydroxyls ions on top and bottom side of the layer, no such strong link existsbetween the two montmorillonite planes where oxygen atoms are located. As

Microstructure Features in the Behaviour of Engineered Barriers 17

Fig. 6. Structure of clay minerals (from Mitchell and Soga 2005). Left: kaolinite;Right: montmorillonite

shown schematically in the figure, exchangeable cations with their hydratingwater molecules shells can penetrate this space due to the electrical attractioncaused by the electrical deficiency that is typical of clay minerals.

Note that in the case of illite that has the same structure as montmoril-lonite, the stability of the stacking of various elementary layers is stronglyensured by potassium ions K+. In other words, the main difference betweenkaolinite and illite on one side and montmorillonite on the other side is re-lated to the stability of the clay platelet made up of stacks of elementarylayers stuck together. The various terms used to describe clay microstructureare now recalled:

• The elementary layer, represented in Fig. 6, is defined for montmorilloniteby a 9.6 Å (96 nm) thickness in the direction perpendicular to the plane(called d(001)). This level can be observed indirectly by standard X-Raydiffraction analysis. Direct observation can also be conducted by transmis-sion electron microscope (see for example Ben Rhaïem et al. 1985).

• The clay platelet or particle (crystal or quasicrystal are somewhat foundin the literature, see Aylmore and Quirk 1962), made up the stacking ofelementary layers, with a number comprised between several (in the case ofclay suspensions) to several hundreds in dry states. This important featurewill be described further on in more detail. Clay particles can be observedusing scanning electron microscope techniques.

• The aggregate made up of clays particles aggregated together, and some-what similar with those previously observed. Aggregates can be observed

18 Pierre Delage

using scanning electron microscope and detected by MIP, as discussedabove.

The microstructure of compacted bentonite

Figure 7 presents a SEM photo of a compacted Kunigel clay (Cui et al. 2002)obtained on a freeze fractured and freeze-dried sample. Kunigel clay (Komineand Ogata 1994) is a Japanese clay made up of 64% Na smectite with aliquid limit wL equal to 474 (wP = 27) and a specific surface of 687 m2/g.The sample observed in the figure was statically compacted at a unit mass of2 Mg/m3 and a water content of 8%, giving rise to an as-compacted suctionof 57 MPa.

Fig. 7. SEM photo of a compacted Kunigel clay (Cui et al. 2002)

In spite of the high compaction density, aggregates made up of aggregatedparticles of clays are clearly apparent. Aggregates are separated by inter-aggregates pores that appear to be significantly large with a diameter upto 2–5 μm. Aggregates are composed of clay platelets made up of stacks ofelementary layers as described in Fig. 6. Advanced investigation based on lowangle X-Ray diffractometry on similar compacted bentonites by Saiyouri et al.(2000) that will be discussed further on provide interesting further precisionson the structures of the aggregates and on their changes during hydration andswelling. Observations carried out using an ESEM have also been publishedby Lloret et al. (2003) and Agus and Schanz (2005), among others.

MIP investigation conducted by various researchers (including Romero etal. 1999, Cui et al. 2002, Lloret et al. 2003, Agus and Schanz 2005 and De-lage et al. 2006) showed PSD curves similar to that of Ahmed et al. (1974)showed in Fig. 2, evidencing the inter-aggregate and the intra-aggregate pore

Microstructure Features in the Behaviour of Engineered Barriers 19

Fig. 8. Intruded pore volume in compacted Almeria clay (Lloret et al. 2003)

populations (see Fig. 8). By comparing the intruded mercury volume at themaximum mercury pressure to the total pore volume, as suggested in De-lage and Lefebvre (1984), it is possible to check whether or not the sampleporosity has been entirely filled by mercury. This appeared to be the case inlow plasticity soils as for instance in the grundite of Ahmed et al. (1974) orthe silt of Delage et al. (1996). This is no longer true in the case of plasticsoils because of the presence of montmorillonite minerals and of the very thinpores that exist inside the aggregates. As shown in Fig. 6, the clay plateletsare not stable with changes in water content and they may become quite thinat higher water contents. The fineness of the platelets gives rise to thin poresthat are much thinner than the smaller entrance diameter accessible in MIP,close to 6 nm (60 Å).

This point is illustrated in the PSD curve of compacted Almeria clay pre-sented in Fig. 8 (Lloret et al. 2003). In this figure, the total intra-aggregatevoid ratio appears to be much larger than the intruded one, due to an impor-tant amount of pores of very small entrance diameter. Interestingly, the valueof this unintruded porosity is the same at the two densities considered in thefigure (dry unit mass of 1.55 and 1.73 Mg/m3 respectively, water content of13%).

Observation of Fig. 8 also shows that the entrance inter-aggregate diameteris in the order of 18 μm whereas the intra-aggregate entrance diameter is closeto 0.012 μm. Note also that the two curves are similar in shape for the poresthat are smaller than 1 μm. A presentation of the same curves as done inFig. 5 would also show more clearly that compression from a void ratio equalto 0.57 down to 0.47 only affected the largest existing inter-aggregate poreswith no effect on the intra-aggregate porosity. In other word, no compressionof the aggregates occurred between these two void ratios.

20 Pierre Delage

3 Effect of Changes in Water Content

Swelling soils are known to be sensitive to changes in water content withswelling when hydrated and shrinkage when dried. A common explanation isgenerally based on the ability of water molecules that hydrate exchangeablecations to penetrate in the interlayer planar spaces inside the platelets, asmentioned in Fig. 6. The most commonly used theory of soil swelling is basedon the diffuse double layer theory and is based on the hypothesis of regularlyscattered individual clay layers, as suggested by Bolt (1956). Although hy-dration mechanisms appear not to respect this hypothesis, this theory gavesome interesting results, mainly in the case of Na+ montmorillonite (see forinstance Sridharan and Jayadeva 1982, Tripathy et al. 2004, Schanz and Tri-pathy 2005). The calculations of repulsion due to electrical interaction is verydependent of the valence of the cations, with significant differences betweenNa+ and Ca++ for example. Indeed, the behaviour of swelling soils also dependof the valence of the dominant exchangeable cation with swelling propertiesenhanced with Na+, but in a different fashion from what the DDL theory pre-dicts. No more considerations on DDL approaches will be made in this paper,and interested readers are invited to read the papers mentioned previously.

Figure 9 shows the PSD curve of two compacted MX 80 bentonites (MX 80is a Wyoming montmorillonite with wL = 520 and S = 700 m2/g) compactedat the same moderate density (1.32 Mg/m3) at two different water contents(12.5 and 28.5%) with suction respectively comprised between 2 and 7 MPaat 12.5% and higher than 50 MPa at 28.5% (Delage et al. 2006). The figure

Fig. 9. Effect of initial water content on two compacted MX 80 bentonite samplesat the same density (Delage et al. 2006)

Microstructure Features in the Behaviour of Engineered Barriers 21

shows that, whereas the two PSD curves are similar in shape – with however asmaller entrance pore radius in the wettest sample – the unintruded porosityis twice larger for the wettest sample at low suction.

The observation of the effect of drying and swelling on a heavily compacted(ρd = 2 Mg/m3) sand bentonite mixture (50–50) made up of Calcigel bentonite(50–60% in montmorillonite) at a water content of 11% has been carried out byAgus and Schanz (2005). Sample was afterwards equilibrated using the vapourequilibrium technique under a suction of 22.7 MPa that gave an equilibratedvalue of water content equal to 9%. The swollen sample was obtained by leftinga compacted specimen above distilled water in a desiccator for 180 days anda final water content of 19% was reached. Results are presented in Fig. 10in terms of PSD curves. As compared to the bimodal PSD curve of the as-compacted sample that defines an inter-aggregate entrance pore diameter closeto 10 μm and an intra-aggregate entrance pore diameter close to 0.12 μm, thePSD curve after swelling shows two points:

• A new pore population around 1 μm appears, at a diameter smaller thanthat of the inter-aggregate porosity and larger than that of the intra-aggregate porosity. The corresponding pore volume corresponds to 44% ofthe total pore volume.

• A large amount of porosity is created below the limit of MIP, defining anunintruded porosity approximately equal to the intruded one. This featureis compatible with that observed on samples at higher water content inFig. 9, showing that samples under low suction have a large very smallporosity.

Fig. 10. Effect of wetting and drying on a heavily compacted Calcigel – sand mixture(Agus and Schanz 2005)