Bergaya F. Handbook of Clay Science

Подождите немного. Документ загружается.

Mackinnon and Kaser, 1987). Romero et al. (1992a) tested different analytical conditions

(time, count rate) on reference kaolinite samples. As TEM conditions with weak electron

beam deﬁnitely appear to induce only limited mass loss, elemental microanalysis under

these conditions is recommended. Elemental microanalysis of rather large domains

(at least 100 nm in diameter) under low electron dose provides sufﬁcient count rates for

satisfactory statistical analysis to be performed, and even permits the transformation of

gels into clay precursors and halloysite to be followed (Romero et al., 1992b).

Recent improvements in signal treatment, leading to higher count rates, can

reduce counting time down to 10–30 s. By limiting the exposure of the clay miner al

sample to the electron beam, irradiation damage is minimized.

C. Data Treatment and Inter pretation

When dealing with layer silicates, the acquisition and treatment of EDXRF data are

based on the quantiﬁcation of the concentration of structural cations relative to that of

silicon. As oxygen and hydrogen are generally not quantiﬁed, such analyses are semi-

quantitative, i.e., OH and H

O contents are not quantiﬁable using this technique.

Irrespective of irradiation damage, all major and minor elements (Na, Mg, Al, K, Ca,

Ti, Fe) can be detected and quantiﬁed. The analytical error and detection limits can be

evaluated by classical methods as done for any other analytical measurement.

12.8.3. PREPARATION TECHNIQUES

Sample preparation is critical to investigating clay minerals by electron micros-

copy. As clay minerals are commonly hydrated, it is important to preserve their

natural layer arrangement. Drying should be avoided because it causes severe mod-

iﬁcations in texture and irreversible layer collapse (Tessier, 1987, 1991). The samples

processed by the embedding techniques are concentrates of clay fractions. Concen-

trates must be in a gel or paste state for proper curing.

A. Agar Coating

Agar was ﬁrst used by Shomer and Mingelgrin (1978) to determine the number of

layers in smectite tactoids by TEM. To this end, several drops of a 2% aqueous

solution of agar was poured on a drop of clay suspension mounted on a glass slide in

order to ﬁx the organization of the clay mineral layers. A 2 mm

portion of the

boundary region of the solidiﬁed agar in which the clay particles were ﬁxed was

transferred for embedment in resin. Similarly, Chenu and Jaunet (1990) prepared a

suspension of smectite in agar to study textural modiﬁcations following the adsorp-

tion of agar. However, they did not give the concentration of the agar in the mixture.

In the following procedure, agar is used as an external coater for hydrated pellets of

clay gels or pastes. Agar powder is placed in a container and dissolved in pure water to

12.8.3. Preparation Techniques 947

give a 5% concentration. The container is put in a water bath, heating the agar

suspension close to the boiling point for 1 h. Glass cups are ﬁlled with liquid agar.

After a few minutes the clay samples are dropped in the liquid. At this stage the agar is

about to gellify, the solidiﬁcation point being around 40 1C. It is recommended to limit

the size of the clay pellets to 1 mm

in order to ensure further rapid and complete

dehydration-impregnation of the sample as it becomes embedded. Such a size is also

suitable for optimal slicing with a diamond knife to give ultra-thin sections.

B. Water Potential Control

The apparatus used for controlling the water potential of clays was ﬁrst described by

Tessier and Berrier (1979). The ﬁltration cells are plunged in a beaker containing a

solution in contact with the sample through a membrane ﬁlter. This allows the

samples to be submitted to a range of gas pressures between 3.2 and 100 kPa

corresponding to the range between the gel and paste states.

The agar cakes are put in ﬁltration cells under a selected gas pressure to equil-

ibrate the water potential at a constant suction. Equilibrium is reached overnight

when the samples are in a state close to saturation. The agar coating ensures perfect

contact with the porous membrane of the ﬁltration cell, and homogeneous hydration

of the clay pellet. The agar coating also protects the clay against dehydration by

direct contact with dry air, and preserves the clay organization from mechanical

stress when handled with tweezers. After equilibration at the selected water potential,

the agar cakes are reduced to cubes of 2 mm

, thus keeping the clay pellet coated with

a thin shell of agar for embedding.

The ﬁltration cells can also be used to concentrate clay suspensions into gels or

pastes, following the same law of correspondence between water potential and

applied gas pressure.

C. Resin Embedding

Foster and De (1971) were the ﬁrst to investigate in detail the techniques for im-

pregnating hydrate d mineral samples, using a seri es of liquids to replace the water

initially present. The embedding technique, originally developed by Spurr (1969) for

curing biological specimens, was modiﬁed by Tessier (1984) who studied changes in

the organization of clay minerals that had been submitted to desiccation–rehydra-

tion cycles. As the microstructure of the clay is preserved (Tessier and Pedro, 1982;

Tessier, 1987), the modiﬁed method was extensively used by clay mineralogists.

Further modiﬁcations and reﬁnements were introduced by soil scientists who use the

method for curing soil samples that are more difﬁcult to handle than pure smectites

rodon

et al., 1990; Robert et al., 1991; Kim et al., 1995). Using a transmission

XRD device, Elsass et al. (1998) were able to follow the change in interlayer distance

during the replacement of the saturating liquid by solvent, and resin. Na

- and

-exchanged smectite samples were equilibrated at 3.2 kPa (pF ¼ 1) and 100 kPa

Chapter 12.8: Transmission Electron Microscopy948

(pF ¼ 3) gas pressures. Intercalation of alcohol and resin is pervasive, and XRD

indicates a signiﬁcant collapse of basal spacings (1.9 nm in water to 1.7 nm in resin).

The embedding process consists of 10 successive exchanges (Table 12.8.1).

For samples of o1 mm in diameter, and pastes equilibrated under a low gas

pressure (3.2 kPa), the time for exchanges can be reduced down to 10 min if

leaching of adsorbed elements or mobilization of colloids is suspected. For com-

pact samples with ﬁne pores, such as rock chips of argillites, micro drill cores of

clayey matrix in thin sections, and clays equilibrated under a relatively high

gas pressure (>100 kPa), the time for exchange must be at least 20 min. This can

be extended to one hour with no major risk of dispersion since the sample is

moulded in agar.

The cubes of agar containing clay pellets saturated with water must be embedded

in resin, thus maintaining liquid saturation throughout the whole procedure. Sam-

ples containing organic matter that is soluble in solvents must be treated so as to ﬁx

the orga nic components before embedding (Thierry, 1967; Glauert, 1975). Further

staining using contrast enhancers for organic matter is carried out on the ultra-thin

sections obtained by slicing the hardened blocks of polymerized resin.

D. Ultra Thin Sectioning

The hardened blocks of resin are allowed to cool at room temperature overnight

before ultra-thin sectioning. The blocks are then sliced up at the appropriate thick-

ness (McKee and Brown, 1977).

Table 12.8.1. Procedure of impregnation in Spurr resin

Exchange step Water

(%)

Methanol

(%)

Epoxy-

propane

(%)

Resin (%) Time Temperature

(1C)

Initial state 100 25

Exch. 1 75 25 10–60 min 25

Exch. 2 50 50 10–60 min 25

Exch. 3 25 75 10–60 min 25

Exch. 4 100 10–60 min 25

Exch. 5 75 25 10–60 min 25

Exch. 6 50 50 10–60 min 25

Exch. 7 25 75 10–60 min 25

Exch. 8 100 10–60 min 25

Exch. 9 50 50 16 h 25

Exch. 10 100 3 h 25

Polymerization 100 20 h 70

Cooling 100 24 h 25

12.8.3. Preparation Techniques 949

12.8.4. CHARACTERIZATION OF CLAY PHASES

The concepts of ‘layer’, ‘particle’, ‘aggregate’, and ‘association’ in clay material

(see Chapter 1) are based upon criteria of unit-cell size, 3-dimensional arrangement,

and chemical composition.

A. Morphological Criteria and Crystal Size Measurement

Lateral extensions in the (a, b) plane are measured on images obtained for objects

lying normal to the optic axis (with care to eucentric condition for magniﬁcation

calibration; see Section 12.8.1). The thickness, perpendicular to the (a, b) plane, is

measured on images obtained for objects oriented parallel to the optic axis of the

microscope (for orientation of strongly anisomorphic objects, see Section 12.8.3).

The layer thickness is characteristic of the type of clay mineral (kaolinite, mica,

smectite, chlori te). HRTEM images can give basal spacings (S

rodon

et al., 1990)

corresponding to the stacking periodicity measured by XRD (0.7 nm for 1:1 phyllo-

silicates, 1.0 nm for 2:1 non-swelling phyllosilicates and up to 1.4 nm for 2:1 swelling

phyllosilicates and chlorite). Intermediate values are obtained for mixed-layer clay

minerals ( S

rodon

et al., 1990, 1992; Veblen et al., 1990; Veblen, 1992).

The aggregate thickness is best expressed by the total thickness, indicating the

number of particles superimposed in the aggregate normal to their (a, b) plane.

B. Structural Criteria and Lattice Parameter Calculation

Structural criteria imply three characteristics that must be veriﬁed together: crys-

tallinity, stacking orientation, and stacking periodicity.

Crystalline Versus Amorphous

The crystalline status of an object can be assessed when the diffraction plane is

normal to the optical axis of the microscope. This conﬁguration is easily obtainable

by placing a suspension of the clay mineral on the microscope grid, and allowing the

water to evaporate, inducing the clay mineral particles to lie with their (a, b) plane

horizontal.

To verify the cryst alline status of an object, a doted diffraction diagram must be

obtained, at least for one orientation of the sample. For clay minerals a single silicate

layer, made up of ordered tetrahedral and octahedral sheets, is sufﬁcient for

obtaining a symmetrical diffraction diagram. This allows the a and b parameters to

be derived from the six dots of the (a, b) plane.

Ordered Versus Disordered Stacking Orientation in the (a, b) Plane

When several layers are stacked on top of each other, the electron diffraction shows

the classic symmetrical diffraction diagram of six dots when all layers are oriented in

the (a, b) plane in exactly the same way (coherent orientation of the a and b axes).

Chapter 12.8: Transmission Electron Microscopy950

The stacking of layers is then called ‘ordered’. A similar diffraction diagram is

obtained for layers parallely stacked with rotation angles of 601 modulo n in

the (a, b) plane (‘semi-ordered stacking’). When rings are obtaine d instead of dots,

the layers are disordered in the (a, b) plane with respect to the a and b axes. The

stacking is then referred to as ‘disordered’ .

Regular Versus Random Periodicity along the Stacking Direction

Particles that are oriented with their layers parallel to the optic axis can show either

regular or random stacking periodicity. In the case of regular stacking periodicity

along the c axis, the electron diffraction pattern shows a single line of dots corre-

sponding to basal reﬂections d(0 0 l). In the case of random stacking periodicity, no

dot is clearly visible but only a blurry line is seen along the (0 0 1) direction. When

several particles are superposed, and all of them are oriented with their layers

parallel to the optic axis but with angular disorientation between particles, the

diffraction diagram is composed of several lines of dots, each particle producing one

single line of dots.

C. Chemical Criteria and Structural Formul ae

Derivation of Structural Formulae

The chemical composition of a sample can be used to derive its structural formula.

Once a structural model was determined, the last step is the identiﬁcation of the

mineral according to the international nomenclature. The structural model is generic

and deﬁnes the group of phyllosilicates such as kaolinite, mica, chlori te, smectite

(see Chapter 1). The chemical composition determines the species of clay mineral.

Chemical data can sometimes complicate the identiﬁcation of minerals that are

commonly found in the earth crust, because their compositions differ from those of

the pure end-members used for classiﬁcation (Caille

re and Henin, 1963 ; Deer et al.,

1965; Velde, 1977).

Structural hypotheses have to be used in interpreting the physical chemistry of clay

minerals that occur either as pure phyllosilicate phases, or mixed with other non-

phyllosilicate minerals. The non-phyllosilicate minerals such as metal (hydr)oxides,

phosphates, sulphates, and carbonates are then considered as impurities. In order

to avoid confusion and misunderstanding, it is necessary to deﬁne the basic layer

structures of common phyllosilicates, 1:1 (TO) or 2:1 (TOT) (2:1) (see Chapter 2),

and the chemical occupancies of the different layers. Each chemical analysis has then

to be transformed into a structural formula. The classical method, proposed by

Foster (1953, 1962, 1963) and Ko

ster (1977), is still applicable. The negative charges of

the oxygen and hydroxyl ions balance the positive charges of the structural cations.

The cation content is then normalized to a number of oxygen ions of the corre-

sponding hypothetical structure. Hypotheses on pure phyllosilicate phases are re-

stricted to three main structural models: O

(OH)

-7O for a 1:1 kaolinite/serpentine

12.8.4. Characterization of Clay Phases 951

structure; O

(OH)

-22O for a 2:1 mica/illite/smectite structure; O

(OH)

-14O for

a chlorite structure.

An initial difﬁculty arises from the fact that EDXRF can measure the concen-

tration of atoms but not their oxidation state. This may be a real problem in the case

of iron. The possibility of signiﬁcant mineral oxidation has to be considered and

evaluated in relation with other bulk measurements and considerations. Mo

ssbauer

(see Chapter 12.1) and IR spectroscopy (see Chapter 12.6) can be used to assess the

oxidation state and structural location of iron (Petit et al., 2002, see also Chapter 8).

Another limitation is that the proposed structural scheme for establishing struc -

tural formula is not usable when non-detectable cations, such as H

or NH

, are

present as is common in natural environments.

Representation of Chemical Compositions

Different basic representations of clay mineral compositions can be chosen to interpret

chemical analysis obtained by EDXRF (Velde, 1977; Peacor, 1992). Despite the prob-

lem (and limitation) of determining the structural formula of phyllosilicates, semi-

quantitative analyses can be directly treated using the relative cation content of the clay

minerals. Ternary diagrams can be constructed showing domains distinguishable by

the amount of layer charge deﬁcit, charge location in the structure, the di-trioctahedral

character, and the composition of the interlayer cations. One can select any of the

constituting cations of the phyllosilicate layer (Si

,Al

,Fe

,Mg

,Mn

) and the interlayer cations (K

,Na

,Ca

), or combinations of them

¼ K

+Na

¼ Mg

+Fe

+Mn

¼ Al

+Fe

)to

plot the chemical data. Ternary diagrams can display the derived chemical compo-

sition of individual particles, the composition of selected groups of them, or the mean

composition of a given population. Reference compositions of pure end-members may

be plotted together with experimental data to assist the determination of clay minerals

based on chemical analyses obtained by EDXRF in TEM.

12.8.5. CHARACTERIZATION OF CLAY MATERIAL AT DIFFERENT

STATES

A. Gel State: Highly Hydrated Smectites

Smectite reaches its highest hydration state when fully saturated with water under

low partial pressure. Observation by HRTEM of ultra-thin sections after resin im-

pregnation displays loose individual layers or groups of 2–3 layers intercalated with

resin. This state and the measured interlayer distances are characteristic of a gel.

Chenu and Jaunet (1990) reported interlayer distances of X2 nm, while Al-Mukhtar

et al. (1996) studied the fabric of clay gels under controlled mechanical and

hydraulic stress.

Chapter 12.8: Transmission Electron Microscopy952

B. Paste State: Pillared Clays, Organo-Clays, Mixed-Layer Clay Minerals

Pillared Clays

When smectite is pillared with aluminium polycations, it forms particles of 20–30

layers with basal spacings of 1.75 nm (Suquet et al., 1994, see also Chapter 7.5). When

local concentrations of excess Al occurs, this excess can be observed on particle edges

as wedge-shaped ﬁllings in the interlayer spaces or as thick coatings on planar surfaces.

Organo-Clays

Vali and Ko

ster (1986) were the ﬁrst to present TEM images of clay minerals after

intercalation of alkylammonium ions. TEM is a valuable technique for investigating

such (intercalated) organo-clays which were widely used in the synthesis of clay

mineral-polymer nanocomposites (see Chapters 7.3 and 10.3)

Mixed-Layer Clay Minerals

The smectite content of mixed-layer clay minerals can be evaluated by measuring the

thickness of fundamental particles (Chapters 5 and 12.2), expressed as the number of

collapsed layers per particle, and calculating the overall particle thickness distribu-

tion and mean number of collapsed layers in the sample (S

rodon

et al., 1990, 1992;

Veblen et al., 1990; Mystkowski et al., 2000).

ucha et al. (1996) used this calcul ation method after embedment of the water-

saturated clay fraction. Even though the number of measured particles is low (o50),

the deduced expandability results are consistent with those obtained by other meth-

ods (XRD, Pt-shadowing, K content ).

By dispersing Na

-saturated mixed-layer illite–smectite in polyvinylpyrrolidone

(PVP-10), Uhlik et al. (2000) were able to increase the interlayer distance between

expanded layers, and hence count the number of collapsed layers. They found good

agreement between Pt-shadowing and HRTEM for PVP-10-intercalated illite–smectite

in terms of the thickness distribution of illite particles, calculated mean crystal thick-

ness, and total expandability of the material. The bimodal thickness distribution of

illite particles suggests two origins: (i) coarse illite of detrital micromica type and (ii)

authigenic (diagenetic) illite from burial of shales and clay stones. The thickness dis-

tribution of authigenic illite particles has a characteristic log-normal shape.

Dudek et al. (2002) applied the same protocol of preparat ion, measurement, and

calculation in comparing TEM with XRD data. Their aim was to interpret crystal

thickness distributions in terms of illitization mechanisms in shales.

Since its introduction by Eberl et al. (1998), using 33 illite samples, the PVP-10

dispersion method was extensively used and documented. Measurements by various

techniques, including the Bertaut-Waren-Averbach and integral peak width by

XRD, ﬁxed cation content by chemical analys is, and Pt-shadowing (in TEM) give

comparable results on fundamental illite particle thicknesses. To assess the size of

fundamental particles by TEM and XRD swelling effects must ﬁrst be eliminated. As

these techniques reﬂect interparticle diffraction by aggregates of fundamental

12.8.5. Characterization of Clay Material at Different States 953

particles ( Nadeau et al., 1984 ; Uhlik et al., 2000), the aggregation of particles and

their ability to form mixed-layer clay minerals can be evaluated (Robert et al., 1991).

Righi and Elsass (1996) characterized multi-phase clay mineral assemblages, using

a curve decomposition program of XRD patterns and direct measurements on

HRTEM photographs obtained after impregnation of moist or wet samples. When

performed on K

-saturated samples that had been subjected to wetting–drying

cycles, HRTEM observations were consistent with XR D results. Vermiculite and

high-charge smectite preserve their initial hydration state on impregnation but with

low-charge smectites the resin molecules can penetrate into the interlayer space.

C. Dry State: Strongly Indurated Materials; Bricks of Consolidated Clay

Deep diagenesis and low-grade metamorphism give rise to samples containing non-

dispersible clays. Because of their low porosity and strong hardness, ion-milling is

the only way to obtain samples suitable for TEM examination. Most bricks of

consolidated clay, especially after heating at temperatures >150 1C are irreversibly

dehydrated and must be treated like solids of low porosity. Increasing metamorphic

grade leads to the formation of thick particles (N>20) as Warr and Nieto (1998)

observed with strongly indurated materials, using the same method of measur ing

particle thickness distributions.

12.8.6. IMAGE ANALYSIS

A. Morphometry

Morphometry involves measuring an object in the image, which thus becomes a

source of quantitative data. The quantitative aspect of microscope images is often

neglected when they are used to illustrate a talk. The real power of electron mi-

croscopy lies in its capablity of allowing scale changes over at least 4 orders of

magnitude to be followed. To perform image analysis, while ensuring representa-

tiveness, it is important to go from a general scale (large area at small magniﬁcation)

to a detail scale (small area at high magniﬁcation). Data must be recorded at pro-

gressive scales without forgetting intermediate scales. The sampling of data on a

microscope grid must be done as for any other sampling method where data on a set

of aliquots are acquired. Several distinct areas (3 is minimum, 10 is now normally

required for publications) must be visited all over the grid in order to avoid bias due

to sample heterogeneity, segregation effects, or locally produced artefacts.

The measured parameters generally qualify the size (diameter, area) of layers,

particles or aggregates and are equivalent to laser granulometry (Dur et al., 2004).

Image analysis is also concerned with the organization of the material in the wide

sense, such as distances betw een objects, arrangement of organo-mineral associa-

tions, and density of objects, as a function of physical or chemical variables. The

Chapter 12.8: Transmission Electron Microscopy954

measurement of parameters that refer to individual objects may allow objects to be

separated into classes (e.g., grain size) or selection criteria (e.g., loose or aggregated).

Measuring Size

For particles having a regular shape (circle, square, rectangle), size parame ters such

as length and width are easily measured. The particles of many clay minerals, how-

ever, are non-rigid, and hence their linear (size) dimensions can only be roughly

characterized. Nevertheless, size measurements give a fair approximation of the clay

mineral surface. Moreover, this is the only way to estimate the real geometrical

surface the reactivity of which is affected by association with other compounds. This

information is important for complex materials such as soils and industrial clays.

Measuring Shape

Shape parameters are non-dimensional, and derive from size parameters (Coster and

Chermant, 1989; Paciornik, 2000; Paciornik and Mauricio, 2004). For clay minerals

with layer structures the most important parameters are the aspect (length/width)

ratio, the form (area/perimeter

) factor, and convexity (convex perimeter /external

perimeter). These parameters are very sensitive to the aggregation mode of funda-

mental units and reﬂect the texture of polydisperse objects.

B. Quantitative Analysis

The quantity of a given object in TEM images may be expressed by its frequency of

occurrence or relative surface. In combination with size and shape measurements,

quantity is of great interest for the practitioner who is concerned with the frequency

of a phenomenon or the relative abundance of different objects.

Many additio nal features and parameters set up for image analysis were described

by Russ (1991, 1995), Gonzalez and Woo ds (1992), and Castleman (1996). The

Journal of Computer Assisted Micros copy should be consulted for advanced ana-

lysis. Numerous softwares are available on the international market; some of which

offer a free-share version having alrea dy the basic functions to perform measure-

ments and classiﬁcations.

By linking morphometry with chemical element analysis and additional features

(e.g., certain types of defects), a physical speciﬁcation for the elements or group

of elements can be obtained. The use of cross-linked data opens the possibility of

establishing a quantitative typology of the studied objects.

12.8.7. CLAYS IN PEDOGENETIC ALTERATION

The combination of TEM with analytical electron microscopy can be used to

assess the processes involved in the formation of clays in weathering proﬁles, and the

underlying mechanisms.

12.8.7. Clays in Pedogenetic Alteration 955

Soils develop from all types of rocks that occur at the earth surface, and almost

always contain clay minerals. The genetic printout of the clay is the key to tracing its

origin, the physicochemical factors affecting its formation, and the progressive

alteration with time (S

ucha et al., 2001).

When the age of soils in a chronosequence is approximately known, the clay

mineral transformations can be used to evaluate the effect of time on the genesis of

expandable clays as in podzolization (Gillot et al., 2000). Such transformations can

also support the resul ts of geological studies on tectonic movements and paleo-

weathering at a large regional scale. For example, the Alpine orogeny of the Iberian

Hercynian Massif is associated with general deep weathering before fossilization

occurs under a sedimentary cover (Vicente et al., 1997).

Wilson (1987) described a podzol smectite that is morphologically similar to

vermiculite, yielding a ‘single spot’ type of electron diffraction pattern. Based on

these assessments, Gillot et al. (2000) ascribed the production of clay particles of

various shape and size to dissolution and physical breakdown. The processes of

dissolution and fragmentation induce modulated structural changes.

According to Aoudjit et al. (1996) and Hardy et al. (1999) physical breakdown,

i.e., fragmentation and exfoliation without major chemical alteration, typically

produce short and thick particles (of biotite and chlorite). Being less susceptible to

weathering than biotite and chlorite, phengitic mica gives rise to relatively thin and

ﬂexible particles (Gillot et al., 2000). The signature of the parent material indicates

evolution of phengite into beidellite through an alteration process which can be

described as a 2:1 solid-phase transformation without extensive dissolution or

chemical modiﬁcation of the original structure (Vicente et al., 1997). The distinction

between beidellite and montmorillonite is based on the structural formula, using the

chemical composition deduced from AEM. The beidellite, formed by transformation

of phengite, is characterized by: (i) replacement of Al by Fe–Mg from the octahedral

phengitic sheet; (ii) a high layer charge, still partially saturated by K

from the mica

structure; and (iii) a speciﬁc arrangement of dense and rigid particles inherited from

the typical morphology of micas.

The smectite in close proximity to muscovi te pa rticles was classiﬁed as neoformed

(Vicente et al., 1997) but it could also arise from a similar transformati on process,

involving only slight local structural reorganization of the 2:1 network. Smectite is

also found in microsites, formed by precipitation of dissolution products of alu-

minosilicates such as plagioclase (Aoudjit et al., 1995).

In studying buried paleosols, Elsass et al. (1997) observed aggregates of

illite–smectite particles with two distinct morphologies, containing different propor-

tions of illite layers. HRTEM observations are not decisive regarding the origin of

variation within the illitic material. This may arise from burial diagenetic illitization

of smectitic material, or represent telogenic alteration of illitic clay by acid waters

penetrating down from the co al bed. In either case, it is an aggregate-by-aggregate

and not a fundamental particle-by-particle process, and cannot be explained by a

simple opening or collapse of interlayer spaces.

Chapter 12.8: Transmission Electron Microscopy956