sign in sign up
<<
Home , Celadonite, Illite

ClayMinerals (2001)36, 1–14

This paper is based on the first George Brown Lecture delivered at the Group Meeting held atthe University of Exeter on 6th April 2000.

Order-disorderinclaymineralstructures

A. PL A N C° O N *

ISTO,CNRS-Universite ´d ’Orle´ans,1A ruede la Fe´rollerie,45071 Orle ´ans,Cedex 2,France

(Received 6April 2000; revised 24 July 2000 )

ABSTRACT:Some recent works dealing with the concept of order-disorder in clay minerals are considered, including those aspects of order-disorder which appeared in the Brindley &Brown (1980) monograph, i.e.disorder in the distribution of cations, disorder in layer stacking, order- disorder in mixed-layer systems and finite crystal size as alattice disorder. Heterogeneity of samples and polymorphous transformations are also considered as other types of disorder. Most of these works emphasize that accurate structural characterization can only be obtained if several techniques are combined (e.g.XRDand IR, EXAFSand Mo¨ssbauer spectroscopies, etc.).Another conclusion is that accurate structural determination provides the key to the genesis of clays.

K EYWORDS:order,disorder,p hyllosilicates,clay minera ls,cat iond istribution,h eterogeneity,mixed layer.

Clayminerals can be described very simply by the structuralcharacterizati onof clay minerals, the stackingof two kinds of layers: 1:1 layers and 2:1 mechanismsby which they change when they are layers.In fact these layers can have a verywide subjectedto different physical or chemical stresses rangeof chemical compositions, of distributions of andthe correlation between structure and proper- isomorphouscations, and can stack in a lotof ties.The aim is rather to review the most recent differentways, including mixed layering. Their paperspublished on these topics, to show what the structuraltriperiodicity is the exception, in which ‘stateof the art ’ is. casede scriptionin volvestheorder-disorder Theclassical division of order-disorder which concept.On the other hand, these minerals have appearedinth eBrindley& Brown(1 980) theability to evolve, e.g. to change their layer or monographwill be retained here, i.e. disorder in interlayercations, to sustain polymorphous trans- distributionof cations, disorder in layer stacking, formationsby octahed ralcation migratio n,to long-rangeand short- rangeorder, order-disorder in modifythe stacking of the layers, and even to mixed-layersystems, finite crystal size as a lattice changethe type of the layers, from 1:1 into 2:1 for disorder,adding the polymorphous transformations example.The variety of clay minerals induces the andheterogeneity of the sample as two other types varietyin their properties. Accurate determination ofdisorder. ofthe structural characteristics leads to a clearer understandingof,and improv ementsin, th e THEDIFFERENTTYPESOF predictionof the properties of clays. ORDER-DISORDER Theaim of this paper is not to reconsider, in an exhaustiveway, what has been published about the Order-disorder in mixed-layersystems

Serpentineto chlorite transformation. It is known thatserpentine (S) transformsinto chlorite (C), and *E-mail: Alain.Plancon@univ -orleans.fr viceversa(Baile y,1 988),andthat random

# 2001The Mineralogical Society 2 A. Planc° on

serpentine-c hloritei nterstratificationsoccur. +a/3tetrahedralshift necessitates the repositioning Regularlyinterstra tifiedserpentine- chloritealso ofO- and OH-coordinati ngoctahedral cations,

existssuch as S 1C1 (dozyite),S 2C1, S3C1, S1C2, requiringmovement of the octahedral cations from S1C3 (Banfield et al.,1994;Bailey et al., 1995). It type-IIto type-I positions. Except for the 2:1 layers, wasnot known why these long- period,regularly thestacking and octahedral slants are inherited. interstratifiedvarieties crystallize instead of discrete Conversionof smectit eintokaolin ite. The serpentineor chlorite or a randominterstratificati on existenceof mixed- layerkaolinite- smectiteshas ofthese. Normal c oulombicintera ctionsare beenknown for some time (Sudo & Hayashi, unlikelyto allow structural communicati onon 1956).UsingX-ray diffraction (XRD) theywere suchlarge scales. This problem was clarified by describedas randomly interstratified layers (Schultz Banfield& Bailey(1996) who studied several et al.,1971;Wiewiora, 1971) and as 1- 1ordered specimensfromtheWoodsChr omeMin e, interstratification(Thomas, 1989 ).Althoughthe LancasterCounty, Pennsylvania, by high- resolution directobservat ionof the individu allayers is transmissionelectron microscopy (HRTEM) witha possiblewith HRTEM, nosuch study had been pointresolution of ~0.19 nm. They observed that performedfor interlayered -smecti teuntil regularlyinterstratified S -Carefrequently inti- recently,probably because the minerals formed matelyintergro wnwith serpent inesthat have duringa weatheringalteration process are poorly repeatdistances identical to those of the regular crystallizedand difficult to characterize. Amouric interstratification(Fig. 1). For example, dozyite &Olives(1998) reconsidered this problem, using a

S1C1 isintimatel yassociatedwiththree- layer low-lightcamera equipped with an yttrium alumi- octahedralorder I,I,II (whereII representan niumgarnet in order to minimize possible beam octahedralsheet of the serpentine layer rotated by damage.Two typesof smectite crystals were 1808 withrespect to I) .Longerperiod polymorphs observed:S, whichgives images showing only (S2C1, S1C2, S2C2, S3C2 and S1C4, all with b = 908) 1nmfringes and another, S ’ where1.25 nm fringes areeach accompanied by serpentines with equiva- arespaced evenly. In addition to these smectite lent c-axisperiodicities. They apparently form by crystals,mixed- layerkaolinite- smectitewas also theselective growth of Ib chloriteunits from I,II observed.Three samples from the Paris basin were octahedralsequences in long-periodserpentines. For studiedwith 10, 45 and 60 % kaolinlayers. The example.. IIIIIIIIII.. ? .. S S C S S C relativeabundance of S ’ layerscompared to the S ..(S2C1)or .. IIIIIIIIIII.. ? S C C S C ... layersincreases as thepercentage of K increases.In Becausepolytypes differ in their octahedral tilt the 10% and 45% kaolinitesamples the lateral sequencesthey almost certainly have different transitionsof one smectite layer into one kaolinite structuralenergies and should exhibit differential layerare visible in images, the 1 nmsmectite stability,which explains why I,II sequences are fringesprevailing in the 10 % kaolinitesample. In convertedselectively into chlorite. the 60% kaolinitesample, another type of lateral Microscopicstudies are consisten twiththe changeis also observed: the transition of a S ’ layer formationof regular interstratificati onby tetrahe- into a KS’ unit,which can be interpreted as the dralinversion within existing serpentine. Atomic partialintercal ationof one kaolinite layer in resolutionimages reveal that the tetrahedral sheet is smectite. displacedby a/3whereit inverts to form the 2:1 Two solid-statemechanisms seem to be respon- layer. A+a/3shiftis require dforhydroge n siblefor the formatio nofkaolinit e:(1) the bondingbetween OH ofthe newly- formedbrucite- transformationof 1 smectitelayer into 1 kaolinite likeinterlayer and O atomsof thenewly-formed 2:1 layer,by stripping a tetrahedralsheet and the layer.The sense of the shift is determined by the adjacentinterlayer region (Fig. 2); and (2) the stronginteraction between the octahedral cations in intercalationof 1 kaolinitelayer into smectite. thebrucite- likeinterlayer and the Si inthe 2:1 Determinationof -smectite structures. The layer,because direct superimposition is strongly determinationof the structure of illite- smectite(I- S) unfavourable.Distortion at the inversion point isimportant because these are common minerals probablylength ensSi Obondsin the next whosestructure changes through , weath- tetrahedron,facilitating the relocation of Si onthe eringand hydrothermal activity. The interstratifica- otherside of the basal O plane.The reversal of the tionis usually estimated from peak migration octahedralslant in the 2:1 layer occurs because the curvesin XRD patterns(Drits & Sakharov,1976; Order-disorder in clay structures 3

FIG.1. HRTEMstructural image down [100], illustrating serpentine 1:1 layers (top centre) passing smoothly into dozyite. Asterisks mark continuous layers in which the stacking sequences differ. Note that changes in stacking occur at the point of tetrahedral inversion, where two 1:1 layers pass into achlorite unit (from Banfield &Bailey, 1996).

S´rodon´,1980,1981, 1984; Watanabe, 1981, 1988; orderingfor R =1,2 or3, but not for segregated I Reynolds,1981, 1988; Tomita et al.,1988;Moore orS layersor with intermed iatedegrees of &Reynolds,1989;Drits et al.,1994).This ordering.Moreover these curves have mainly been techniquecan, however, be used for two- component usedfor glycolated I- Sassumingthat all smectite I-Swith random (Reichweite R =0),or maximum interlayerscontain two glycol layers and that their 4 A. Planc° on

FIG.2. Lattice fringe images of various interstratified kaolinite-smectites of the 10 % kaolinite sample.S=Stype smectite layer (1 nm thick).K=kaolinite layer (0.72 nm thick).In part c, the lateral transition S ? K is indicated by arrows (from Amouric &Olives, 1998). swellingproperties do not depend on the interlayer amountof non-swelling layers is 0.90 but 0.10 of cation.The most efficie nttechniq uefor the theselayers have d(001)= 10.33A Ê .Forall these determinationof the structural parameters is based modelsthe layer types are distributed at R =1.This onthe comparison between experimental XRD NorthSea sample is in fact a fourcomponent I-S-V- curvesa ndpatter nscalcul atedfor struct ural V’,whereS andV arelayers which, in the glycolated modelshaving different proportions and distribu- state,with each of the exchangeable cations, swell to tionsof I andS (Reynolds& Hower,1970; Drits & 16.6 17.2 AÊ and 12.9 13.6 AÊ ,respectivelyand Sakharov,197 6;Re ynolds,19 80;Moor e& where V’ (referredto as ‘low-chargevermiculite ’, Reynolds,1 989;Drits&Tchoubar,1 990). or ‘high-chargesmectite ’),has the swelling char - However,the simulation of XRD patternsrequires acteristicsof smectite or , depending on manystructural and instrumental parameters. The theexchangeable cation, which confirms the data effectof these numerous variables is that several publishedpreviously (Drits et al.,1997c).Forthis structuralmodels may fit the experimental data sample,the abundances of illite and vermiculite are equallywell, especially if the experimental and wI = 0.80 and wV =0.08respectively, while wS + wV’ calculatedpatterns are not superposed. =0.12.In Mg- or Ca-saturated specimens, both Sakharov et al.(1999b)showed that reliable glycolatedand air dried, V ’ swellslike smectite, but structuralmodels can be obtained if amulti-specimen behaveslike V inNa-saturated and glycolated profile-fittingprocedure is performed. For each specimens.Thus, the interpretation of XRD data sample,specimens saturated with different cations fromdioctahedral -smectite minerals requires the (Na,Mg and Ca) are analysed, both air-dried and analysisof the positions as well as of the intensities glycolated.One structural model fitting all the ofall the basal reflections in the observed XRD observedpatterns then provides the structure of the patternsand the interpretation of the XRD patternsof sample.For example, in caseof a NorthSea sample, glycolatedsamples may be erroneousif it is basedon thestructural models for the Ca-, Mg -andNa- thequalitative analysis of a singleXRD patternfrom saturated,air- driedspecimens have 14. 0A Ê and aNa-saturatedsample. 12.5 AÊ expandablelayers with the proportions 0. 12 Mechanismof the diagenetic transformation of and0. 08(Mg saturate d),0. 10and 0. 10(Ca theI-S-V phases. Illitizationis a phenomenonthat saturated),and 0. 06and 0. 14(Na saturate d) occursin various geological environments: benton- + (Fig.3 ).For the NH 4-saturatedspecimen, the ites,hydrothe rmally-alteredrocks, , etc. Order-disorder in structures 5

However,the structural mechanism of illitization, viainterstratified I-S may be different, depending oneach specific set of physicochemical conditions. Areviewof illitization processes was given by Altaner& Ylagan(1997), with a descriptionlimited tothe Reichweite and the abundances of illite and smectitelayers, which can be sufficient for benton- itesand hydrothermally-a lteredrocks, but not in the morecomplex case of shales, for which a multi- specimenapproach is required. Mixed- layerclay fractionsfrom the North Sea and Denmark were studiedusing XRD byDrits et al.(1997c).The patternswere computer simulated with the multi- componentprogram described above. Based on structuralcharacteristics and the degree of diage- netictransformation, the samples that were analysed canbe divided into three groups. The I-S- Vof groupone is predomi nantlydetrita landhas 0. 69 0.73illite,0.26 0.20smectiteand 0.04 0.07vermiculite interlayers, the illite, smec- titeand vermiculite interlayers being segregated. TheI-S-V of group two has been transformed diageneticallyand has 0.80 illite, 0.12 smectite and 0.08vermiculite interlayers, the vermiculite inter- layersbeing segregated whereas the illite and smectitehave the maximum ordering possible of R=1.The I- S-Vofgroup three has been transformedfurther during diagenesis and has 0.84 illite,0. 08smectiteand 0.08 vermiculite interlayers. Tounderstand the structural transformation, the

occurrenceprobabilities wij and wijk (i, j, k = I, S, V)ofpairs and triples have been calculated. If each memberof the series was formed from the precedingmember by a solid-statetransformation mechanism,a singleinterlayer transformation (SIT), thena decreasein amount of definite layer pairs, triples,etc. in one member should increase the amountof layer pairs, triples, etc. of other types in sucha waytha tanewset o foccurrence

probabilities wij, wijk,etc.will be equal to that of thenext member of the series. Because R =1,the probabilityof transformation of one definite type of FIG.3. Powder XRDpattern of North Sea Upper layerpair into another should be constant for a Jurassic sample from well 2/11-1, 4548 mdepth givensample, and does not depend on the preceding (sample 87).The observed and simulated patterns are andfollowing layer types. This is the case. For shown as solid and shaded lines, respectively, with the d-values of simulated patterns in brackets. The I-S-V example,forthe second stage of the I- S-V parameters are given above each pattern. Oriented transformation,smectite interlayers decrease from specimens: (a) Mg 2+-saturated and glycolated, Cu- Ka 0.20to 0. 12and are transfo rmedinto 0. 01 radiation; (b) Mg 2+-saturated and air dried, Co- Ka vermiculiteand 0. 07illite layers. A remarkable radiation; (c) Na +-saturated and glycolated, Co- Ka featureof this stage is that all the SS pairsare + radiation; (d) Na -saturated and air dried, Co- Ka transformed,77 % into IS, 14% intoSI andthe rest radiation (from Sakharov et al., 1999). intoVS andVI. The SV pairsare also totally 6 A. Planc° on transformedinto IV (80 %) and VI (20%). As a computersimulation of the distribution (Dainyak et resultthe tendency to segregation is changed into al.,1992),using experimental data obtained by themaximum possible alternation of illite and differentspectroscopic methods, with IR dataas the smectiteinterlayers. basis.A featureof the IR spectraof dioctahedral Thelack of further transformation of vermiculite micasin the OH-stretching region is that each intoillite is surprising because vermiculite inter- definitepair of cationsbonded to an OHgrouphas a layershave a higherlayer charge than smectite definiteposition for the corresponding OH band, interlayers,an dthenega tivecharg eofth e whereasthe content of a cationicpair is determined correspondingexpandable interlayers must increase bythe integrated optical (Slonimskaya et al., duringthe formation of illite interlayers. According 1986;Besson & Drits,1997a,b) .Thistechnique to Sudo et al.(1962),Drits (1966), Tettenhorst & providesinformation on cation distribution at the Johns(1966), Drits & Sakharov(1976), Gu ¨ven one-dimensionallevel only, i.e. for one crystallo- (1991)and Jakobsen et al.(1995),2:1 layers in graphicdirection, e. g.the b directionfor minerals dioctahedralinterstratified minerals have a polar made up of trans-vacant2:1 layers. Description of character.Accordingly a neoformationof illite or thetwo-dimensional distribution requires us: (1) to vermiculiteinterlayers should take place through a specifythe degree of preference for one of the two lateralreplacement of smectite interlayers including cis-positions(in a halfunit- cell)for every cation; anincrease of Al inthe tetrahedral sheets. The and(2) to provide additional information on the differentreactivity of S andV interlayersto a limitationsimposed on the cation pairs along b1 and transformationto illite may be due to the SIT b2 whichare the directions running at +1208 with mechanism.After the first stage of transformation, respectto the b direction.The simulation program thevermicul iteinterlay ersquickly absorb K + wasdeveloped to produce the multiple random cationsand collapse, thus being closed to further fillingof a honeycombpattern with cation pairs, reaction.The source of Al maybe adissolvedK-I-V, leadingto different NFe (numberof Fe cations containingmainly K, whichrepresents ~10 % of the aroundan Fe cation)contents. Those for which NFe clayfraction in the group I samplesand decreases valuesare close to that obtained using EXAFS are throughthe series to group III. It is also noticeable retained,andfor each of them a theoretical thatthe newly- formednon- expandingmica inter- Mo¨ssbauerspectrum is calculated and compared to layersare pure ammonium tobelite, the ammonium theexperimental one. The best fit of thisMo ¨ssbauer probablybeing supplied by the organic matter patternprovides the best cation distribution. Using duringoil occurrence (Drits et al.,1997a).So, thisapproach, Drits et al.(1997d)analysed different statisticalcalculations demonstrate that the I-S-V Fe-richsamples (, Fe- illiteand celadon- transformationin shales can be described as a SIT ites),and found that R2+ cationsprefer to occupy withincrystallites. oneof two symmetrically independent cis-sites in orderto maximize the local compensation of anion charges,and, in addition, R2+R2+ and/orFe- Al are Order-disorder in the distribution of cations prohibitedin the b1 and b2 directions.Under these Thedetermination of the actual octahedral cation conditions,the octahedral sheets of samples show distributionin dioctahedral is a complex domainstructures, i. e.they are represented by problem.X- raydiffraction provides only the average domainsof various sizes with almost pure Al or compositionof cations in the unit- cellsites. Fe3+ compositions,and by domains of mixed- cation Spectroscopicmethods are particularly useful in composition(Fig. 4) .Othergood examples of the thisinstanc e,since they probe local atomic usefulnessof the combination of several methods environmentsand have the potential to determine arefound in Muller et al. (1997),using EXAFS asa short-rangecation ordering, each with its own basisplus XRD andIR. They showed that in Camp- advantagesand limitation and therefore providing Berteauxmontmorillonite, Fe andMg cations are onlya partialsolution. A newmethodological segregatedin small clusters, and that the heating of approachhas combined XRD andIR spectroscopy samplesat temperatures >150 8Cisaccompanied by withother methods such as Mo ¨ssbauer,EXAFS or themi grationofNi cationsinto cis-vacant polarizedEXAFS. Thus,the two- dimensional octahedra.Using polarized EXAFS asa basis, distributionfor every type of cation in dioctahedral Manceau et al.(2000)determine dthecation 2:1layer silicates has been determined by the distributionin Swa- 1nontronite. Order-disorder in clay mineral structures 7

FIG.4. Two-dimensional cation distribution determined for sample 68/69 (from Drits et al., 1997).

Polymorphous transformations Tsipurski& Drits(1984) showed that the 2:1 layers ofmontmorillonite are cv asa rule,whereas Semiquantitativedetermination of trans-vacant normallyhave tv 2:1layers. Studying I-S from 28 andcis-vacant layers in illite-smectites and illite- bentonitesamples, Cuadros & Altaner(1998) also smectite-vermiculites. InterstratifiedI-S and I- S-V observedthat smectitic I- Stendto be cv and representingboth hydrother maland diagenet ic becomeincreasingly tv withincreasing illitization. transformationsand different degrees of structural Theevaluation of wcv shedlight on the mechanism orderwere investigated by Drits et al. (1998) for ofillitization in various geological environments. A cis-/trans-occupancyin the octahedral sheet using decrease in wcv duringillitization is probably due to XRD anddifferential thermal analysis (DTA) in adissolutionprecipitation,whereas an almost combinationwith evolved water analysis (EWA) constantvalue indicates a solid-statetransformation. usingan IR detector.By XRD, theamounts of cis- Byway of illustration, Deconinck & Chamley vacant (cv) and trans-vacant (tv)layerswere (1995)found two types of smectitic minerals in the determinedfor ordered samples both by the UpperCretaceous chalks of northern France. The simulationof the patterns and calculations based firsttype contains 65 95% smectitelayers, has a onthe position of 11 l reflections(Drits & McCarty, dehydroxylationpeak at ~500 8Candwas assumed 1996).If the EWA peaksbelow and above 600 8C tooriginate from continen talweatherin g.The areattributed to tv and cv octahedra,respectively, secondtype contains 90 100% smectitelayers, theproportions of cv layers (wcv),determinedby hasa dehydroxylationpeak at 600 7008C and was XRD andEWA arein agreement. For disordered interpretedas authigenic. From the results above, samples,having no diagnostic 11 l reflection,the thefirst type must have tv layerswhich confirms a EWA methodis the only one that can be applied. weatheringof mica or illite. The second type, 8 A. Planc° on assumedto be authigenic, effectively has the cv patterns,selected area electron diffraction (SAED), layersnormally found in smectite. andthermogr avimetricanalysis,Muller et al. Cationmigration during dehydroxylation -rehy- (1999a,2000a,b )showedthat the mechanism is droxylationprocess. Thestuctural modification of morecomplex for tv Fe-,Mg-richminerals (like the2:1 dioctah edrallayer silicat esduring a celadoniteand glaucon ites).In this case the dehydroxylation-rehydroxylationprocess depends dehydroxylationcauses a migrationof Fe andMg bothon the original occupancy of trans- and cis- cationswhich leads to a fairdistance between sitesand on the chemical composition of the cationsand a goodscreening of their repulsion. The octahedralsheet. The simplest case is that of tv migrationtakes place through the shared edges in Al-richoctahedral sheets (most illites). It is known the[010 ]and[310] directions. Depending on the afterUdagawa et al.(1974)that their dehydroxyla- dehydroxylationtemperature, primitive unit cells or tioninduces the replacement of two adjacent OH super-cellsare observed in SAED patterns.The groupsby a residualoxygen located in the plane of existenceof additional satellites in the patterns of thecations, midway between them. The rehydroxyla- somecrystals heated to 750 8Ccorrespondsto the tionof these minerals restores the initial structure. existenceof antiphase domains with a 3 b/2 width ForFe-rich layer silicates a cationmigration occurs andan antiphase shift of a/2.The rehydroxylation duringdehydroxylation (Tsipurski & Drits,1984). ofceladon itepreserv esthe octahed ralcation Usingmod elsge neratedfrom powde rXRD distributionformed after dehydroxylation (Fig. 5),

FIG.5.XRDpattern ( l-Ka Mo =0.70926 A Ê )of celadonite 69, in natural (N), dehydroxylated (D), rehydroxylated (R)and de-rehydroxylated (DR) states.Layers are trans-vacant in the Nspecimen, but cis-vacant in the others (from Muller et al., 1999). Order-disorder in clay mineral structures 9 whilethe rehydroxylation of (with a high thesame position as in kaolinite (Bookin et al., Al content)is accompanied by the reverse cation 1992).TheNMA moleculesare also trapped within migration,which restor esthe tv layers, and thesame environment, but due to the planar shape indicatesthat the migratio nprocessprobabl y ofthe molecule, the translation between adjacent dependson the cation composition. A lastcase is layers is 0.24a +0.20b (Fig.6b). Thus it is the that of cv Al-richminerals (like ). shapeof the molecule which seems to play the Thework of Muller et al. (1999b),based on DTA, prominentrole in the interlayer shift. showedthat the octahedral cations also migrate duringdehydroxylatio n,the layers becoming tv, and Finite sizeof particles as alattice disorder remaining tv afterrehydroxylation. The role of the chemicalcomposition seems to be prominent in the Claymineral particles usually contain a rather cv/tv transformationscaused by dehydroxylation- smallnumber of layers, the average values ranging rehydroxylationprocesses. fromfive to a fewtens. The distribution of these sizesexists and is usually well described by a lognormallaw (Drits et al.,1997b).Because the Order-disorder in layerstacking outersurfaces of particles comprise only a small Thestacking of the layers in 1:1 clay minerals partof the whole particle, it is generally assumed (kaolinsubgroup) can be modified, e.g. by the thattheir influence on spectroscopic and diffraction intercalationofwa terororgani cmolecules. resultsis rather small. Three different programs Intercalationwas studied in kaolinite by Thomson existfor the simulation of theXRD patternsof two- &Cuff(1985), Costanzo et al.(1984)and Bookin (andthree- )componentsystems: the widely used et al.(1992),in by Bookin et al.(1992),but NEWMOD program(Reynolds, 1985); that devel- notin the third member of the subgroup i.e. . opedby Sakharov (1976, not published), used for BenHaj Amara et al.(1998,1999) considered this theXRD intensitycalculations described above; mineralin order: (1) to determine the role of the andthePlanc¸on&Dritsp rogram(1999) polytypism,and then of the stacking of the layers, CALCMIX whichperforms the same calculation asa resultof the intercalation; and (2) to determine asthe Sakharov program but with a differentuser accuratelytheposition and orientatio nofthe interface.All these programs are based on the same intercalatedmolecules in the interlayer space, in Markovianstatistics for the description of the orderto understand the role of the shape of the stackingof the layers, but differ in the way these intercalatedmolecules in the layer stacking. This particlesend, the NEWMOD formalismalways studyhas mainly used the XRD onpowder and assuminga 2:1layer outer surface. The observation orientedsamples, and the IR spectroscopy.The ofnon- negligibledisc repanciesin calcul ated intercalatedmolecules were water, dimethylsulph- intensitiesbetween NEWMOD andCALCMIX oxide(DMSO) whichhas an isotropicshape, and n- foundtheir origin in the nature of the outer methyl-acetamide(NMA )whichis a planar surfaces.This led to the design of a newformalism molecule.The first information on the positions of andof th ecorrespondingcompu terprogram themolecule sandsite occupanc ieshas been (Sakharov et al.,1999a),whichall owsthe obtainedfrom mono- dimensionalFourier transform descriptionof the XRD intensitiesfor particles (MFT)analysisof XRD patternsfor oriented endingin any kind of conceivable surfaces. The specimens.Then, peak positions in powder XRD firstresults are that, as expected, the role of the patternshave provided the starting values of the outersurfaces is greater when the mean size of the layerstacking. The modelling of the powder XRD particlesdecreases, whatever the mineral under patternshas been performed with modifications of consideration.But even for particles composed of theintercalated molecule orientation, compatible severaltens of layers, the effect remains strong for withMFT andIR spectrato provide the best fit someminerals (e. g.chlorites) (Fig. 7) .Thisnew withexperimental data. It has been shown that domainof research, i. e.the determination of the waterand DMSO moleculesare trapped between outersurfaces by XRD, mayhelp to understand the theditrigonal cavities of the tetrahedral sheet and mechanismsof crystal growth, e. g.a layer-by-layer thevacant octahe dronof the adjace ntlayer orscrew dislocation growth. It may also be useful (Fig.6a), with a displacementbetween layers inunderstanding the surface reactivity of finely close to a/3 +b/3.The DMSO moleculesoccupy dispersedclay minerals. 10 A. Planc° on

a

b

FIG.6. Projection on the layer plane of the interlayer space of intercalated nacrite. Broken line: upper surface of the octahedral sheet of one nacrite layer; solid line: lower surface of the tetrahedral sheet of the adjacent nacrite layer; (a) DMSO;(b) NMA(from Ben Haj Amara et al., 1999).

Heterogeneity as adisorder exceptionsexist, e.g. the expert system for the descriptionof disordered (Planc ¸on& Itis generally assumed that each clay mineral Zacharie,1990) is based on the assumption of a bi- phasein a smallsize sample (a few grams) can be phasedsystem. Why does Birdwood kaolinite describedby one set of structural parameters. Some (SouthAustral ia)notinterc alateeasily with Order-disorder in clay mineral structures 11

FIG.7.Calculated XRDpatterns for oriented samples of chlorite.Uniform distribution of the thickness of particles between 3and 14 layers. All particles end with brucite layers on both outer surfaces (solid line) or with abrucite layer on one side and a2:1 layer on the other side (dotted line) (from Sakharov et al., 1999a). acetamideand formamide? It is well known that smallestfraction containing the most disordered kaoliniteswith few stacking defects intercalate kaolinitefor API#9 (Fig. 8), which is the opposite easilyand almost completely with acetamide and toKGa- 1.This remarkable dissimilarity suggests formamide,whereas high- defectkaolinites inter- thepossibility of a geneticrelationship. calateless fully. The Birdwood kaolinite was studiedby Frost et al. (1999).Based on the CONCLUSIONS Hinckleyindex (1. 35),this kaolinite is highly orderedbut intercalates with difficulty (18 days Recentwork shows that the structural characteriza- for a 20% intercalation,and extended treatment tionof clay minerals becomes, with time, more and neverproduces an intercalation over 60 %). Raman moreprecise and that the nature of the layers, their spectroscopyin the OH-stretching region showed chemicalcomposition and layer charge are now nochange in the intensity yet the kaolinites were insufficientto describeclay mineral particles. But is beingintercalated. X- raydiffraction showed the greaterprecisi onnecessar yforpractica luse? presenceof two phases of kaolinite, one highly Fortunately,the use of clay minerals has not orderedand one disordered. Minor dickite and two waitedfor their accurate structural characterization. amorphousphases were also found. Electro n However,the utility of a claymaterial can often be microscopyshows the presence of two particle predictedby a structurald etermination.For shapesattributed to the two kaolinites phases. The example,if thematerial is akaolinite,and structural difficultyof intercalation is attributed to the co- determinationreveals that it is disordered, it would existenceof these two phases, the high-defect bewise to avoid using it for paper coating or kaoliniteapparently coating the highly ordered intercalationof organic molecules. Nevertheless, a oneat the edges, thereby preventing and slowing simplestruct uraldeterm inationissometi mes intercalation.Bish & Chipera(1998) studied two insufficient.For example, the reactivity of clays standardkaolinites (KGa- 1andAPI#9) which could inan intercalation process can be unexpected, (e.g. easilyappear identical by XRD. Thestudy of thekaolinite from Birdwood, mentioned above), differentsize- fractionsreveals, however, that all the andonly an accurate structural study can show why, fractionsconsist of at least two kaolinites, the andtell if such problems are surmountable. In 12 A. Planc° on

generation.The I- Sinthese rocks contain tv 2:1 layersand tobelite interlayers are formed through solid-stateAl forSi substitutionin the tetrahedral sheetand by ammonium fixation in the corre- spondinginterlayers. This illitization deviates from theillitization in other rocks because of the supply ofammonium formed during oil generation and the fixationof this ammonium in the former smectite interlayers.This can be brought to the fore only by veryprecise structural study, as shown above. Two conclusionsconcerning methodology are worthemphasizing: (1) accurate structural char- acterizationcan be obtain edon lyif severa l complementarytechniques are used (e.g. XRD and IR,EXAFS andMo ¨ssbauerspectroscopies, etc. ); and(2) the limits of the experimental capabilities (e.g.HRTEM) ortheoretical developments (e.g. formalismtaking into account, in diffraction, the outersurface structures) have not been reached, and accuratestructural characterizatio nofclay minerals canbe improved further.

REFERENCES Altaner S.P.& Ylagan R.F.(1997) Comparison of structural models of mixed-layer illite/smectite and reaction mechanisms of smectite. Clays Clay Miner . FIG.8.X-ray diffraction patterns of six size-fractions 45, 517 533. of API#9 kaolinite (from Bish &Chipera, 1998). Amouric M. &Olives J. (1998) Transformation mechanisms and interstratification in conversion of smectite to kaolinite: an HRTEMstudy. Clays Clay anotherexample: a manufacturerruns out of the Miner. 46, 521 527. clayhe normally uses because the quarry is Bailey S.W., editor (1988) Hydrous Phyllosilicates exhausted,or it is inhomogeneous, and the clay (Exclusive of Micas) .Reviews in Mineralogy, 19. mineralextracted does not satisfy his requirements. Mineralogical Society of America, Washington, D.C. Inthat instance, a precisecharacterization of the Bailey S.W., Banfield J.F., Barker W.W. &Katchan G. materialis necessary to find an equivalent. It has (1995) Dozyite, a1:1 regular interstratification of alsobeen shown how an XRD studyof the serpentine and chlorite. Am. Miner. 80, 65 77. structuraloccupancy of trans and cis sites of a Banfield J.F. &Bailey S.W.(1996) Formation of material,or an HRTEM studyof its mixed- layering regularly interstratified serpentine-chlorite minerals canexplain its origin and its transformations. These by tetrahedral inversion in long-period serpentine resultsare guides in the search for equivalent polytypes. Am. Miner. 81, 79 91. materials. Banfield J.F., Bailey S.W. &Barker W.W. (1994) Thepractical benefits which may result from an Polysomatism, polytypism, microstructures and accuratestructura ldeterminationmay also be reaction mechanisms in serpentine and chlorite. Contrib.Mineral. Petrol. 117, 137 150. illustratedin the context of the occurrence of oil. Ben Haj Amara A., Ben Brahim J., Planc¸on A.& Ben Drits et al. (1997)suspected that in the diagenetic Rhaiem H.(1998) Etude de la structure d ’une nacrite transformationof the I-S- Vclaysof the North Sea hydrate´eet dehydrate´e. J. Appl. Crystallogr . 31, shales,ammonium was supplied during oil occur- 654 662. renceto form the new non-expanding interlayers, Ben Haj Amara A., Planc¸on A., Ben Brahim J. &Ben whichare pure ammonium tobelite. This assump- Rhaiem H.(1999) Intercalation of organic molecules tionis now validated (Lindgreen et al.,2000).The and water in nacrite. P.63 in: Abst. Conf. Eur. Clay illitizationin oil source rocks is linked to oil Groups Assoc., Krako´w . Order-disorder in clay mineral structures 13

Besson G.& Drits V.A.(1997a) Refined relationships mixed-layer illites-smectites and other interstratified between chemical composition of dioctahedral fine minerals. Clays Clay Miner . 42, 382 390. dispersed mica minerals and their infrared spectra in Drits V.A.,Lindgreen H.&Salyn A.(1997a) the OH-stretching region. Part I:Identification of the Determination of the content and distribution of stretching bands. Clays Clay Miner . 45, 158 169. fixed NH4 in illite-smectite by X-ray diffraction: Besson G.& Drits V.A.(1997b) Refined relationships Application to North Sea illite-smectite. Am. Miner. between chemical composition of dioctahedral fine 82, 79 87. dispersed mica minerals and their infrared spectra in Drits V.A., S´rodon´J.&Eberl D.D.(1997b) XRD the OH-stretching region.Part II: The main factors measurement of mean crystallite thickness of illite/ affecting OHvibration and quantitative analysis. smectite: reappraisal of the Kubler index and the Clays Clay Miner . 45, 170 183. Scherrer equation. Clays Clay Miner . 45, 461 475. Bish D.L.& Chipera S.J. (1998) Variation of kaolinite Drits V.A., Sakharov B.A., Lindgreen H.& Salyn A. defect structure with particle size. Abst. Ann. Meet. (1997c) Sequential structure transformation of illite- Clay Miner. Soc., Cleveland . smectite-vermiculite of Upper Jurassic shales from Bookin A.S., Drits V.A.& Tcherkachin V.I. (1992) the North Sea and Denmark. Clay Miner. 32, Structural modifications of kaolinite and dickite by 351 371. intercalation . Izv. Acad. Nauk, Ser.Geol . 4, 106 118 Drits V.A., Dainyak L.G., Muller F., Besson G.& (in Russian). Manceau A.(1997d) Isomorphous cation distribution Brindley G.W. &Brown G., editors (1980) Crystal in celadonites, glauconites and Fe-illites determined Structures of Clay Minerals and their X-ray by infrared, Mo¨ssbauer and EXAFSspectroscopies. Identification .Monograph, 5.Mineralogical Clay Miner. 32, 153 179. Society, London. Drits V.A., Lindgreen H., Salyn A., Ylagan R.& Costanzo P.M., Giese R.F.& Lipsicas M. (1984) Static McCarty D.K.(1998) Semiquantitative determina- and dynamic structure water in hydrated kaolinite. I tion of trans-vacant and cis-vacant 2:1 layers in –The static structure. Clays Clay Miner . 32, illites and illite-smectites by thermal analysis and 549 551. X-ray diffraction. Am. Miner. 83, 1188 1198. Cuadros J.&Altaner S.P.(1998) Compositional and Frost R.L., Van der Gaast S.J.,Paroz G.N., Zbik M. & structural features of the octahedral sheet in mixed- Kloproge J.T. (1999) Why Birdwood kaolinite layer illite/smectites from . Eur. J. (South Australia) does not intercalate with acetamide Mineral. 10, 111 124. and formamide? P.85 in: Abst. Conf. Eur. Clay Dainyak L.G., Drits V.A.& Heifits L.M.(1992) Groups Assoc., Krako´w . Computer simulation of cation distribution in Gu¨ven N.(1991) On the definition of illite/smectite dioctahedral 2:1 layer silicates using IR-data: mixed-layer. Clays Clay Miner . 39, 661 662. Application to Mo¨ssbauer spectroscopy of aglauco- Jakobsen H.J., Nielsen N.C.& Lindgreen H.(1995) nite sample. Clays Clay Miner . 40, 470 479. Sequences of charged sheets in rectorite. Am. Miner. Deconinck J.F. &Chamley H. (1995) Diversity of 80, 247 252. smectite origins in Late Cretaceous sediments: Lindgreen H., Drits V.A., Sakharov B.A., Salyn A.L., Examples of chalks from Northern France. Clay Wrang P.& Dainyak L.G.(2000) Illite-smectite Miner. 30, 365 379. structural changes during metamorphism in black Drits V.A.(1966) Some structural features of layer Cambrian Alum shales from the Baltic Sea. Am. silicates with long periodicity along caxis. Pp. Miner. 85, 1223 1238. 35 45 in: Physical Methods for Study of Manceau A., Bruno B., Drits V.A., Chategner D., Gates Minerals (A.G. Kossovskaya, W.P., Wu J.,Huo D.&Stucki J.W. (2000) Oxidation- editor).Nauka, Moscow (in Russian). reduction mechanism of in dioctahedral smec- Drits V.A.& McCarty D.K.(1996) The nature of tites: ICrystal chemistry of oxidized reference diffraction effects from Iand I-S consisting of . Am. Miner. 85, 133 152. interstratified tv and cv layers. Am. Miner. 81, Moore D.M. &Reynolds R.C.(1989) X-ray Diffraction 852 863. and the Identification and Analysis of Clay Minerals . Drits V.A.& Sakharov B.A.(1976) X-ray Analysis of Oxford University Press, Oxford. Mixed-Layer Minerals .Nauka, Moscow (in Muller F., Besson G., Manceau A.& Drits V.A.(1997) Russian). Distribution of isomorphous cations within octahe- Drits V.A.& Tchoubar C.(1990) X-ray Diffraction by dral sheets in from Camp-Berteaux. Disordered Lamellar Structures .Springer Verlag, Phys. Chem. Miner . 24, 159 166. Berlin. Muller F., Planc¸on A.& Drits V.A.(1999a) Structure of Drits V.A., Varaxina T.V., Sakharov B.A.& Planc¸on A. dehydroxylated and rehydroxylated dioctahedral 2:1 (1994) Asimple technique for identification of one- layer silicates. P.114 in: Abst. Conf. Eur. Clay dimensional powder X-ray diffraction patterns for Groups Assoc., Krako´w . 14 A. Planc° on

Muller F., Planc¸on A. &Drits V.A.(1999b) Application C.& Drits V.A.(1986) Interpretation of the IR of the structural models of heat-treated micas to the spectra of celadonites and glauconites in the region study of the dehydroxylation-rehydroxylation of of the OH–stretching frequencies . Clay Miner. 21, smectites. Abst.Colloque RX99, Strasbourg . 377 388. Muller F., Drits V.A., Planc¸on A.& Besson G.(2000a) S´rodon´J. (1980) Precise identification of illite/smectite Dehydroxylation of Fe 3+,Mg-rich dioctahedral in interstratifications by X-ray powder diffraction. micas: (I) Structural transformation. Clay Miner. Clays Clay Miner . 28, 401 411. 35, 491 504. S´rodon´J. (1981) X-ray identification of randomly Muller F., Drits V.A., Tsipurski S.I.&Planc¸on A. interstratified illite/smectite in mixtures with discrete (2000b) Dehydroxylation of Fe 3+,Mg-rich dioctahe- illite. Clay Miner. 16, 297 304. dral micas: (II) Cation migration. Clay Miner. 35, S´rodon´J. (1984) X-ray powder diffraction identification 505 514. of illitic materials. Clays Clay Miner . 32, 337 349. Planc¸on A.& Zacharie C.(1990) Anexpert system for Sudo T.& Hayashi H. (1956) Arandomly interstratified structural characterization of kaolinites. Clay Miner. kaolinite-montmorillonite in acid clay deposit in 25, 249 260. Japan. Nature 178, 1115 1116. Planc¸on A. &Drits, V.A.(1999) Software for X-ray Sudo T., Hayashi H.& Shimoda S.(1962) Mineralogical phase analysis of clays. P.121 in: Abst. Conf. Eur. problems of intermediate clay minerals. Clays Clay Clay Groups Assoc., Krako´w . Miner. 9, 378 392. Reynolds R.C.(1980) Interstratified clay minerals. Pp. Tettenhorst R.& Johns W.D. (1966) Interstratification 249 304 in: Crystal Structures of Clay Minerals and in montmorillonite. Clays Clay Miner . 13, 85 93. their X-ray Identification (G.W. Brindley &G. Thomas A.R.(1989) Anew mixed-layer clay mineral. Brown, editors). Monograph, 5.Mineralogical Regular 1:1 mixed-layer kaolinite-smectite.P.69 in: Society, London. Abst. Ann. Meet. Clay Miner. Soc. Sacramento, Reynolds R.C.(1985) NEWMODa computer program California. for calculation of one-dimensional diffraction pat- Thomson J.G. &Cuff C.(1985) Crystal structure of terns of mixed-layer clays .R.C.Reynolds, 9Brook kaolinite dimethylsulfoxide intercalate. Clays Clay Rd.,Hanover, NH03755, USA. Miner. 33, 490 500. Reynolds R.C.(1988) Mixed-layer chlorite minerals. Pp. Tomita K., Takahashi H.& Watanabe T.(1988) 601 630 in: Hydrous Phyllosilicates (Exclusive of Quantification curves for mica/smectite interstratifi- Micas ) (S.W.Bailey,editor).Reviewsin cations by X-ray powder diffraction. Clays Clay Mineralogy, 19.Mineralogical Society of America, Miner. 36, 258 262. Washington, D.C. Tsipurski S.I. &Drits V.A.(1984) The distribution of Reynolds R.C.& Hower J.(1970) The nature of octahedral cations in the 2:1 layers of dioctahedral interlayering in mixed-layer illite-montmorillonite. smectites studied by oblique-texture electron diffrac- Clays Clay Miner . 18, 25 36. tion. Clay Miner. 19, 177 193. Sakharov B.A., Planc¸on A. &Drits V.A.(1999a) Udagawa S., Urabe K.& Hasu H. (1974) The crystal Influence of outer surface structure of crystals on structure of dehydroxylate. Japan Ass. X-ray diffraction.P.129 in: Abst. Conf. Eur.Clay Miner. Petrol. Econ.Geol . 69, 381 389. Groups Assoc.,Krako´w . Watanabe T.(1981) Identification of interstratifications Sakharov B.A., Lindgreen H., Salyn A.& Drits V.A. of illite and montmorillonite by X-ray powder (1999b) Determination of illite-smectite structures diffraction. J.Miner. Soc.Japan, 15, 32 41 (in using multispecimen X-ray diffraction profile fitting. Japanese). Clays Clay Miner . 47, 555 566. Watanabe T.(1988) Structural model of illite/smectite Schultz L.G., Shepard A.O., Blackmon P.D.& Starkey interstratified minerals and the diagram for their H.C.(1971) Mixed-layer kaolinite-montmorillonite interstratification. Clay Sci. 7, 97 114. from the Yucatan Peninsula, Mexico. Clays Clay Wiewiora A.(1971) Amixed-layer kaolinite-smectite Miner. 19, 137 150. from Lower Silesia, Poland. Clays Clay Miner . 19, Slonimskaya M.V., Besson G., Dainyak L.G., Tchoubar 415 416.