粘土沉积

www.elsevier.com/locate/clay

Impactofclaymicrostructureandmassabsorptioncoefficienton

thequantitativemineralidentificationbyXRDanalysis

V.R.Ouhadia,*,R.N.YongbbFacultyofEngineering,Bu-AliSinaUniversity,Hamedan,Iran

GeoenvironmentalResearchCentre(GRC),CardiffUniversity,Cardiff,UK

aAbstract

Quantitativemineralevaluationofclayeysoilsisanessentialsteptoprovidethereliablesoilrecognitionandbehaviour.InthispapertheapplicationofthefourcurrentXRDmethodsformineralquantificationpurposesisexperimentallyreviewedforclayeysoils.Theproblemsinvolvedinclaymineralevaluationbasedonthedirectuseofpeakintensityorareasunderthepeakareaddressed.Inaddition,severalmistakesinducedbytheapplicationofonlythemajorreflectionlinesofclaymineralsforthispurposeshavebeencited.Finally,amethodforXRDquantitativemineralevaluationwaspresentedinthispaper.Byvalidatingtheproposedmethod,averygoodagreementwasobservedbetweenthecomputedandtherealquantityofeachmineralpresentintheseriesofartificialsamples.

D2003ElsevierScienceB.V.Allrightsreserved.

Keywords:Quantitativemineralevaluation;XRD;Palygorskite;Sepiolite;Claymicrostructure;Peakintensity

1.Introduction

Inordertoinvestigateonthebehaviourofclayeysoilsandtoevaluatetheroleoftheirdifferentmineralsontheirbehaviour,itisnecessarytohaveaquantitativeXRDevaluation.CurrentXRDtechni-quesforthisanalysisarebasedonfourmethodsasfollows:(1)analysingbasedonpeakareas;(2)usingtheidenticalmassabsorptioncoefficientmethod;(3)mineraldiagnosingbasedonaninternalstandard;and(4)quantitativemineralevaluationusinganexternal

*Correspondingauthor.Tel.:+98-811-8272031;fax:+98-811-8272046.

E-mailaddress:vahido@basu.ac.ir(V.R.Ouhadi).

standard.TheintensitiesoftheXRDpatternofanindividualmineralareknowntobeproportionaltotheconcentrationsofthedifferentmineralspresent.Therefore,bymeasuringtheintensitiesofpatterns,someideaoftherelativeamountsofeachphasecanbeachieved.Thediffractedintensityofanyhklreflectionfromanycrystallinematerialcanberelatedtoitscomposition,matrix,andinstrumentalcharacter-istics(AlexanderandKlug,1948;NorrishandTaylor,1962;YongandWarkentin,1975;YongandOuhadi,1997).Basedondifferentevaluationsofachieveddensities,anumberofquantitativemethodshavebeendevised.Thesemethodsareexperimentallyreviewedinthispaperandtheproblemsassociatedwiththemareaddressed.Basedonatheoreticalderivation,aquantitativeXRDanalysisforclayeysoilsisproposedandvalidated.

0169-1317/03/$-seefrontmatterD2003ElsevierScienceB.V.Allrightsreserved.doi:10.1016/S0169-1317(03)00096-6

142V.R.Ouhadi,R.N.Yong/AppliedClayScience23(2003)141–148

2.Materialsandmethods

Thisstudywasperformedusingcalciteandseveralpureclaymineralsincludingkaolinite,illite,palygor-skiteandsepiolite.Thesemineralsarechosenbecausetheyareknowntobethemajorfractionsofmarlysoils(Davis,1967;YongandOuhadi,1997).Theartificialsoilswerekaolinite(hydritePXobtainedfromtheGeorgiaKaolin)andillite(Domtarsealbond)obtainedfromDomtarConstructionMaterials,whichcomesfrompulverizingCanadianoldmarineshale.ThepropertiesofthesehavedescribedbyYong(2002).Thepurepalygorskiteisextractedfromthenaturalmarl(OuhadiandYong,2001).Inaddition,thepuresepiolitesamplewasobtainedfromTolsa,Spain.ThepropertiesofthesetwosoilsampleshavebeenaddressedbyOuhadiandYong(2001).XRDexperi-mentwasperformedfollowingproceduresandmeth-odsgivenintheliterature(MooreandReynolds,1989).Forsamplepreparation,ineachcase,0.2gofair-driedsoilwasmixedwith10mldistilledwaterinasmallvolumetricflaskandshakenbyhandfor2–3min.ThenthevolumetricflaskwasplacedonanultrasoundtoadequatelydispersethesoilbeforeXRDtesting.Finally,usingamicropipette,4dropsofthepreparedsolutionwaspositionedonaglassslide.Thecoatingonglassslidesforallsampleshadsimilarthickness.TheXRDspectrawereobtainedbyscanninginthe2hrangeof5–17jand5–70j.APhillipsapparatus‘‘PW1710’’withCu-KaradiationwasusedtoobtaintheX-raydiffractionpatternsofsoilsamples.ThesepatternswereidentifiedbycomparisonwithfilestandardX-raypowderdiffractionpatternsofthemostcommonlyfoundminerals(InternationalCentreforDiffractionData(ICDD),MineralPowderDiffractionFileSearchManual1988).Bymaking19XRDanalyses,twomajorsourcesoferrorinquanti-tativeXRD,includingsamplepreparationandinstru-mentalerrors,wereevaluatedatthefirststep.

3.Resultsanddiscussion

3.1.DeficiencyofXRDanalysisusingareasunderpeaks

ThismethodisbasedonthedirectcomparisonofintensitiesorpeakareasofmineralsidentifiedbyXRD

testing.Duetoitssimplicityofmineralevaluationofsoils,thismethodisverycommonamongresearchers.Eventhoughthistechniquecangiveanestimationofthequantitiesoftheindividualminerals,duetothedifferencemassabsorptioncoefficientsofminerals,thismethodmightleadtoawrongquantitativemineralestimation(Carroll,1970).Toquantitativelyevaluatetheaccuracyofthismethod,thepeakintensityofseveralartificialmixturesofpalygorskiteandcalciteareanalysedbyXRD.Fig.1indicatesthevariationofpeakintensityofthemajorreflectionlineofpalygor-skiteandcalciteinthesesamples.Undersimilarconditions,calciteindicateshigherreflectionlineintensityincomparisontopalygorskite.Inotherwords,ascanbeseeninFig.1,quantitativeanalysisbasedonpeakareaorpeakheightofthemineralpresentinthemixturemightstronglyleadtoanunderestimationofthequantityofpalygorskite.ThisisshowninFig.2,whichillustrateshowextensivelythemineralevaluationbasedonpeakheightmightunderestimatethepresenceofpalygorskiteandover-estimatethepresenceofcalciteinatwo-phasesystem.Thecorrelationfactorbetweenthepredictedquantitybasedonareasunderthepeakandtherealquantitypresentinthesampleforpalygorskiteis0.22.Thesamedeficiencywasobservedinthetwo-phasesystemofsepioliteandkaolinite,whileinthelattercase,analysisbasedonpeakheightorareaswillleadtoanoverestimationofthepresenceofsepiolite.Therefore,itseemsthatanalysisbasedononlypeak’sareaorheightwillnotbereliableintermsofquantitative

Fig.1.Peakintensity,artificialsamples.

V.R.Ouhadi,R.N.Yong/AppliedClayScience23(2003)141–148143

Fig.2.Realandestimatedquantity.

evaluation.Furthermore,althoughitmaygiveanoverallestimationofthepresenceofmineralsinthesoilmixture,stillintermsofmineralidentificationbasedondescendingorascendingorderofexistingmineralsinthesoil,thismethodwillnotbereliable.3.2.XRDanalysisusingidenticalmassabsorptionSinceintensityisproportionaltoconcentration,thenforamixtureoftwoelements,thefollowingproportioncanbewrittenas(Wilson,1987):WA=WB;

IA=IB

ð1Þ

inwhichAandBarerepresentativeoftwoelementspresentinthemixtures;WAandWBaretheweightpercentagesoftwoelements.Inaddition,IAandIBaretheintensitiesoftwoselecteddiffractionsofthesetwoelements,respectively.Theproportionalitysigncanbereplacedbyproportionalitycoefficient,K,asfollows:WA=WB¼KðIA=IBÞ

ð2Þ

TodetermineK,itisrequiredtoprepareamixtureofthesementionedminerals(i.e.AandB)ofknownweightpercentages.Themainprobleminvolvedwiththismethodincludesthelimitationofderivationequa-tionswhenmorethantwomineralsareused.Inthismethodtherelationshipbetweenintensityandconcen-trationofthemajorreflectionlineisassumedtobeconstant.

3.2.1.Palygorskiteandsepiolitequantification

Toinvestigatetheaccuracyofthisaboveassump-tion,aseriesofmixturesofpalygorskiteandsepioliteatdifferentconcentrationswaspreparedandrunbyXRD.Fig.3indicatesthevariationofXRDpeakintensityofsepioliteversusitspercentageinaseriesofartificialsamples.Ascanbeseeninthisfigure,inatwo-phasesystemofsepioliteandkaolinite,byincreasingthepercentageofsepiolite,theintensityofsepiolite’smajorpeakincreases.Threesamplesinamultiple-phasesystemhavingacontrolledamountofsepiolitewereanalysedfortheintensityofsepiolite’smajorpeak.Thesearepresentedonthesamegraph.This,infact,showstheeffectofmineralinterferenceontheX-raydiffractionintensitiesofsepiolite’smajorpeak.Therelativedeviationfromtherealquantitypresentintheartificialsampleforsepiolitedetermi-nationinamulti-phasesystemincomparisontothestandardgraphwillbe45%,3%and4%,where25%,50%and100%sepiolitearepresentinthetestedmixtures,respectively.Infact,theseinterferenceeffectsareconcentration-dependent.Aswillalsobeconfirmedlater,atalowconcentrationofmineral,intensityshowsahigherdeviationfromthemulti-phasesystem.Fig.4comparesthevariationofpro-portionalitycoefficientofsepioliteandpalygorskiteversustheconcentrationoftheseminerals.Twodiffer-entsetsofexperimentsusingdifferentmixturesofsepiolite–kaoliniteandpalygorskite–calcitehavebeenused.Thisfigureillustratestheconcentratedependencyofproportionalitycoefficientsofthetestedminerals.Ahighproportionalitycoefficientofsepioliteinitslowconcentrationindicatesthateveninlowconcentrations,sepiolitemightshowhighdiffrac-tionintensity.Inadditiontotheabovediscussion,the

Fig.3.Intensityandconcentration.

144V.R.Ouhadi,R.N.Yong/AppliedClayScience23(2003)141–148

Fig.4.Proportionalitycoefficientvariation.

calculatedproportionalitycoefficientobtainedforthesemixturesisnotconsistentwhendifferentreflec-tionlinesforKmeasurementareused.

3.3.MineraldiagnosingusinginternalstandardsThismethodisbasedontheuseofaninternalstandardtocalculatehowpeakintensitiesvarywithvariationofconcentration.Byaddingacertainamountofaparticularcomponenttothetestedsampleandusingdifferentartificialsamples,havingdifferentcon-centrationsofthedesiredmineral,theestablishmentofacalibrationgraphispossible(Mossmanetal.,1967;Quakernaat,1970;Burtner,1974).Thismethod,inspiteofitsadvantageofnotassumingalinearrelation-shipbetweentheintensityofacertaindiffractionpeakandconcentrationduringtheadditionofstandardtotheunknownsample,assumesalinearrelationbetweentheintensityofconsumedstandardandmineralconcen-tration.Therefore,itfinallyignorestheinterferenceeffectsofdifferentmineralsinmakingandusingthecalibrationgraphs.Inotherwords,aswaspreviouslyaddressed,despiteofproblemsinvolvedwiththeuseofonlythemajorreflectionline,inthismethodtheroleofotherpeaksofacertainmineralwillnotbeconsidered.Duetothesimilarityoftheprinciplesassociatedwiththeinternalandexternalmethodsandtheaboveexplanations,onlytheexternalmethod,aswillbediscussedinthefollowing,hasbeenexaminedinthisresearch.

3.4.AnalysingbasedontheexternalstandardThismethodisbasedontheuseofpureminerals,asastandard,whentheyareknowntobepresentinthe

testingsample.Forpreparingcalibrationgraphs,bymakingartificialsampleshavingdifferentpercentagesofaknownpuremineral(i.e.Y),thepeakintensityofthemajorreflectinglineofthisaddedpuremineralwillbemonitoredinallartificialsamples,puremineralandunknownsampleaswell.WiththecalculationoftheratioofpeakintensityofYinallsamplesoveritsintensityinthepuremineral,thepreparationofacalibrationgraphwillbepossible.Therefore,havingtheratioofpeakintensityofthismineralintheunknownsampleoverthepeakintensityofpuremineralandusingcalibrationgraph,itispossibletodeterminetheestimatedpercentageofmineralintheunknownsample.

3.4.1.Multi-phasesystemofsoilsandtheexternalmethod

Fig.5indicatestheuseofdifferentmixturesofkaoliniteandsepioliteasanexternalstandardforquantitativemineralanalysisinaseriesofartificialsamples.Forinstance,applyinglinearregressionfordatapresentedinFig.5givesthecorrelationfactorof0.94.Followingtheestablishmentofastandardgraph,sepiolitequantificationoffiveartificialmulti-phasemixtureswasexaminedbytheexternalmethodwherethepercentageofsepiolitewasfixedat25%,50%and75%.Ascanbeseen,atalowpercentageofsepiolite,thereisnoticeabledeviationamongtheachieveddatawithrespecttothecalibrationgraph.Forinstance,formulti-phasesampleshaving75%,50%and25%sepiolite,theuseoftheseestablishedstandardgraphs

Fig.5.Useofexternalmethod.

V.R.Ouhadi,R.N.Yong/AppliedClayScience23(2003)141–148145

givestherelativedeviationofestimationequalto4%,6%and52%,respectively.

Infact,aswasaddressedinthispaperandbypreviousresearchers(i.e.CarrollandHathaway,1963),intheuseofthismethod,oneshouldalwaysnoticethatlowconcentrationsofmineralcouldleadtoawrongquantitativeevaluation.Problemsaddressedintheuseoftheinternalmethodexistintheapplica-tionoftheexternalmethod.Oneoftheadvantagesoftheexternalmethodwillbethepossibilityofdiffer-entiatingbetweenthemagnitudesofthemassabsorp-tioncoefficientofmineralsasitisshowninFig.6.3.5.ProposedXRDanalysis

ThetheoreticalbasisandderivationoftheproposedquantitativeXRDanalysisisbasedontherelationshipestablishedbetweentheintensityandtheabsorptioncoefficientofthesoilsampleassuggestedbyKlugandAlexander(1954),andBrindley(1961).Accordingtothem,theintensityoftheX-rayreflectedbyacompo-nentCinamixtureisknowntoberelatedtothevolumeproportionVcasfollows:Ic¼ðPcVcÞ=llm

ð3Þ

inwhichllmisthemassabsorptioncoefficientof

mixtures,andPcisaconstantforanycertainreflectionfromthecomponentC.Toavoidthepreviouslyaddresseddifficultiesinvolvedwiththeuseofjustmajorreflectionlinesforquantitativeanalysis,theotherreflectionlines,especiallythosewhichhavearelativelyhighintensityweightingfactor,shouldbe

Fig.6.Peakintensity,artificialsample.

takenintoconsideration(Ouhadi,2002).Thiswillhavethefollowingadvantagesinclayidentification:First,thepossibilityofanymistakeofclayidentifica-tionwhenconsideringonlythemajorreflectionlinewillbeeliminated.Second,theconceptofthecrystal-lizationdegree,whichcausesthedisappearanceofsomereflectionlines,willbetakenintoaccount.Third,allthereflectionlinesofclaymineralshavingmorethanonemajorlinewillbeconsidered.Therefore,theproportionalitycoefficient,whichisusuallycalculatedbasedonthemajorreflectionlines(Cullity,1976;MooreandReynolds,1989),canbeestimatedbasedonatleastthosereflectionlineswitharelativelyhighintensityweightingfactor.Thereflectionlineswithanintensityweightingfactorofequaltoormorethan30areusedinthisresearchtodeterminetheproportion-alitycoefficientsofminerals.Todoso,forproportion-alitycoefficientmeasurement,theaverageweightedintensityiscalculatedasfollows:

Iawi¼Xn,

ðIiWiÞX

nWið4Þ

i

i

inwhichIiistheintensityofthereflectionlinehavingaweightingfactorequaltoormorethan30,Wiisthestandardweightingfactorofthatreflectionline,nisthenumberofreflectionlineshavingweightingfactorsequaltoormorethan30,andIawiistheaverageweightedintensityforproportionalitycoefficientmeasurementpurposes.Replacingthelatterequationintheformerequation,onewillget:

Iawi¼Xn,

ðIiWiÞX

nWi¼ðKcVc=llmÞð5Þ

i

i

inwhichKcisaconstantfortheexistingreflectionsfromthecomponentC.Ontheotherhand,volumeproportionVccanbedefinedasfollows(Brindley,1961):

Vc¼Wcðqm=qcÞ

ð6Þ

inwhichqmistheaveragedensityofthemixture,andqcandWcaretheweightproportionanddensityofcomponentC,respectively.ReplacingVcintheformerequationwillresult:

Xn,

ðIiWiÞXnWi¼KcðWcqmÞ=ðqcllmÞð7Þ

i

i

146V.R.Ouhadi,R.N.Yong/AppliedClayScience23(2003)141–148

BasedontheapplicationofLambert’slawtoX-rayanalysis,themassabsorptioncoefficientofamixturecanbedefinedasfollows:lm¼llm=qm

ð8Þ

Byreplacingthemixture’smassabsorptionratiointhepreviousequation,thefollowingcanbewritten:

X

n,

ðIiWiÞX

nWi¼kcðWcÞ=ðlmÞ

ð9Þ

i

i

Inanymulti-componentmixture,thelatterequa-tioncanbewrittenfortwocomponentstoeliminate

themassabsorptioncoefficientofthemixtureasfollows:(

X

n,

,(,

ðIciWciÞX

n)WX

nci

ðIdiWdiÞX

n)Wdii

i

i

i

¼kcðWcÞ=kdðWdÞ

ð10Þ

Thelatterequation,inturn,indicatesthatthe

intensityratioisdirectlyproportionaltotheweightratioofcomponents.Theapplicationofthislatterequationwillbeapplicabletoquantifythepercentofmineralsintheunknownsample.

3.5.1.PracticalgraphsforXRDmineralquantifica-tionanditsvalidation

Toincludetheseimportantfactors,theabovemen-tionedconsiderationsareusedinmineralidentifica-

Fig.7.Applicationofproposedmethod.Fig.8.Applicationofproposedmethod.

tionandquantitativemineralevaluationofasetofcontrolledsoilsamples.Verygoodcorrespondencewasobservedasdescribedinthefollowing.Aswaspresentedbefore,themainelementsofmarlysoilsconsistofpalygorskite/sepiolite,calcite,quartz,illiteandkaolinite.Sincethebehaviourofmarliscon-trolledbyacombinationofthesemajorelements,thequickquantitativeanalysistoquantifythemagnitudeofeachfractionisofinterest.Thequartzfractioncanbeeasilyseparatedbythedifferentsoilwashingsteps(OuhadiandYong,2001).This,infact,willeliminatetheinterferenceeffectofquartzandclaysinthequantitativeanalysis.Therefore,themajorelementstobequantifiedwillincludesepioliteorpalygorskite,calcite,kaoliniteandillite.Basedonthesuggestedquantitativeevaluationtechnique,aseriesofbinarymixturesofstandardmineralswaspreparedindiffer-

Fig.9.Applicationofproposedmethod.

V.R.Ouhadi,R.N.Yong/AppliedClayScience23(2003)141–148147

Fig.10.Applicationofproposedmethod.

entconcentrations.Figs.7–10showtheresultsobtainedforpalygorskite–calcite,sepiolite–kaolinite,palygorskite–sepioliteandkaolinite–illiteinthebinarymixtures.Inthesegraphs,thehorizontalaxisindicatestheweightratiooftwomineralsandtheverticalaxisrepresentstheratiooftheaverageweightedintensityoftwomineralsinthemixtureaswasdefinedbefore.Thelinearpatternischosenforregressionanalysispurposesbasedonthepriorassumptionofarelationshipbetweenintensityandconcentration(AlexanderandKlug,1948).Asisshowninthesegraphs,thecorrelationfactorforeachseriesofdataismorethan0.94,whichindicatestheacceptablelinearfittingoftherelationshipbetweenweightratioofmineralsandtheaverageweightedintensity.

3.5.2.ValidationofestablishedgraphsforXRDquantification

Tovalidatetheextentofaccuracyofthesegraphsforaquantitativeanalysisofmineralsinmarlysoils,asetofartificialmulti-componentmixtureswaspre-pared.Theartificialmixturesconsistofdifferentpercentagesofpalygorskite,sepiolite,calciteandkao-linite.Twodifferentseriesofartificialsampleswereused.Thefirstseriesofartificialsamples,whichispresentedinthispaper,consistsofseveraldifferentmixturesofcalcite,sepioliteandkaoliniteasrepre-sentativeofmarlsoils.Chemicalanalysiswaschosenforcalcitedeterminationthatgaveaccuracywithin1%(YongandOuhadi,1997).TheresultsobtainedareplottedinFig.11,inwhichtherealquantityofmineralinthemixtureisplottedagainstthepercentageachievedfromXRDanalysis.Forcomparativepurpo-ses,thebisectorofthegraphisdrawnandcalledtheidealcaseinwhichthequantitativeXRDevaluationgivesthesameresultsasintherealcase.Thecorre-lationfactorbetweenthesepointsandtheidealcase(bisector)isfoundtobe0.90.

4.Concludingremarks

Theresultsofthisresearchprovidethefollowingconclusions:(a)ThedirectquantitativeXRDanalysisbasedonpeakintensityorareasunderthepeaksmightstronglycauseanoverestimationorunderestimationof

Fig.11.Validationofproposedmethodbypreparationofseveralartificialsoilsamples.

148V.R.Ouhadi,R.N.Yong/AppliedClayScience23(2003)141–148

thequantityofclayminerals.(b)Quantitativeevalua-tionbasedononlythemajorreflectionlinemightstronglyinducesomemistakesinmineralidentifica-tionofclaysoils.(c)TheproposedquantitativeX-rayanalysisusingdifferentreflectionlinesofmineralsthattakesintoaccounttheimpactofclaymicrostructureshowedanacceptablevalidationevaluation.

AppendixA

C=Cal.=CalciteCoef=Coefficienti=Illite

I=Majorreflectionline’speakintensityofmineralIawi=Ratiooftheaverageweightedintensityoftwomineralsinthemixture

Imix=Intensityratioofthemajorbasalspacingofmineralinthemixture

Ipure=Intensityratioofthemajorbasalspacingofpuremineral

K=Kaolin.=Kaol=KaoliniteMix=Mixture

P=Paly.=Pal.=PalygorskiteR=r=CorrelationfactorS=Sep.=Sepiolite

References

Alexander,L.E.,Klug,H.P.,1948.BasicaspectofX-rayabsorptioninquantitativediffractionanalysisofpowdermixtures.Ann.Chem.20,886–889.

Brindley,G.W.,1961.Quantitativeanalysisofclaymixtures.In:Brown,G.(Ed.),X-rayIdentificationandCrystalStructureofClayMinerals.MineralogySoc.,ClayMineralsGroup,London,pp.489–514.

Burtner,R.L.,1974.QuantitativeX-raymineralogy,sampleprepa-rationandanalysiswithanaluminuminternalstandard.23rdAnn.ClayMineralsConf.,Cleveland,OH.

Carroll,D.,1970.ClayMinerals:AGuidetoTheirX-rayIdenti-fication.TheGeologicalSocietyofAmerica,Boulder,CO.Carroll,D.,Hathaway,J.C.,1963.MineralogyofselectedsoilsonGuam.WithasectiononDescriptionofsoilprofilesbyC.H.Stensland.U.S.Geol.Surv.Prof.Pap.403-F,403–418.

Cullity,B.D.,1976.X-rayDiffraction.Addison-WesleyPublishing,Reading,MA.3rdprinting.

Davis,A.G.,1967.Mineralogyandphaseequilibriumofkeupermarl.Q.J.Eng.Geol.1,25–38.

Klug,H.P.,Alexander,L.E.,1954.X-rayDiffractionProcedures.JohnWileyandSons,NewYork.

Moore,D.M.,Reynolds,R.C.,1989.X-rayDiffractionandIdenti-ficationandAnalysisofClayMinerals.OxfordUniv.Press,NewYork.

Mossman,M.H.,Freas,D.H.,Bailey,S.W.,1967.OrientinginternalstandardmethodforclaymineralX-rayanalyses.ClaysClayMiner.15,441–453.

Norrish,K.,Taylor,R.M.,1962.QuantitativeanalysisbyX-raydiffraction.ClayMiner.Bull.5(28),98–109.

Ouhadi,V.R.,2002.StudyoftransformationofclaymineralsintheinteractionprocesswithadditivesbyuseofscanningelectronmicrosopeandXRDanditsrelationtomechanicalbehaviour.Iran.J.Crystallogr.Mineral.10(1),87–97.

Ouhadi,V.R.,Yong,R.N.,2001.Theroleandinfluenceofclayfractionofmarlysoilsontheirgeotechnicalandgeoenvi-ronmentalperformance.In:Yong,R.N.,Thomas,H.R.(Eds.),ProceedingsofthethirdBritishGeotechnicalAssociationGeo-environmentalEng.Conf.CardiffSchoolofEngineering,Tho-masTelford,pp.216–223.

Quakernaat,J.,1970.Directdiffractometricquantitativeanalysisofsyntheticclaymineralmixtureswithmolybdeniteasorientationindicator.J.Sediment.Petrol.40,506–513.

Wilson,M.J.,1987.AHandbookofDeterminativeMethodsinClayMineralogy.Blackie,Glasgow.

Yong,R.N.,2002.Influenceofmicrostructuralfeaturesonwater,iondiffusionandtransportinclaysoils.Proc.,ClayMicrostructureanditsImportancetoClayBehaviour,Lund,Sweden.Geodevel-opmentAB,Sweden,pp.1–14.

Yong,R.N.,Ouhadi,V.R.,1997.Reactionfactorsimpactingoninstabilityofbasesonnaturalandlime-stabilizedmarls.SpecialLecture,KeynotePaper,Int.Conf.onFoundationFailures,Sin-gapore,pp.87–97.

Yong,R.N.,Warkentin,B.P.,1975.SoilPropertiesandBehaviour.Elsevier,NewYork.