Bergaya F. Handbook of Clay Science

Подождите немного. Документ загружается.

Sanz, J., 1990. Distribution of ions in phyllosilicates by NMR spectroscopy. In: Mottana, A.,

Burragato, F. (Eds.), Absorption Spectroscopy in Mineralogy. Elsevier, Amsterdam,

pp. 103–144.

Sanz, J., Herrero, C.P., Robert, J.L., 2003. Distribution of Si and Al in clintonites: a combined

MMR and Monte Carlo study. The Journal of Physical Chemistry 107, 8337–8342.

Sanz, J., Madani, A., Serratosa, J.M., Moya, J.S., Aza, S., 1988. Aluminum-27 and silicon-29

magic-angle spinning nuclear magnetic resonance study of the kaolinite-mullite transfor-

mation. Journal of the American Ceramic Society 71, C-418–C-421.

Sanz, J., Robert, J.L., 1992. Inﬂuence of structural factors on

Si and

Al NMR chemical

shifts of phyllosilicates 2:1. Physics and Chemistry of Minerals 19, 39–45.

Sanz, J., Serratosa, J.M., 1984.

Si and

Al high resolution MAS-NMR spectra of phyllo-

silicates. Journal of the American Chemical Society 106, 4790–4793.

Sanz, J., Serratosa, J.M., 2002. Nuclear magnetic resonance spectroscopy of organo–clay com-

plexes. In: Yariv, S., Cross, H. (Eds.), Organo–Clay Complexes and Interactions. Marcel

Dekker, New York, pp. 221–272.

Sanz, J., Stone, W.E.E., 1979. NMR study of micas. II. Distribution of Fe



and OH



the octahedral sheet of phlogopites. American Mineralogist 64, 119–126.

Sanz, J., Stone, W.E.E., 1983a. NMR study of micas. III. The distribution of Mg

and Fe

around the OH



groups in micas. Journal of Physical Chemistry Solid State Physics 16,

1271–1281.

Sanz, J., Stone, W.E.E., 1983b. NMR applied to minerals: IV. Local order in the octahedral

sheet of micas: Fe-F avoidance. Clay Minerals 18, 187–192.

Schultz, A., Stone, W.E.E., Poncelet, G., Fripiat, J.J., 1987. Preparation and characterization

of bidimensional zeolitic structures obtained from synthetic beidellite and hydroxy-alu-

minium solutions. Clays and Clay Minerals 35, 251–261.

Sherrington, D.C., 1980. Preparation, functionalization, and characteristics of polymer sup-

ports. In: Hodge, P., Sherrington, D.C. (Eds.), Polymer-supported Reactions in Organic

Synthesis. Wiley, Chichester, pp. 1–80.

Slichter, C.P., 1990. Principles of Magnetic Resonance, 3rd edition. Springer, Berlin.

Sugahara, Y., Satokawa, S., Kuroda, K., Kato, C., 1990. Preparation of a kaolinite–polya-

crylamide intercalation compound. Clays and Clay Minerals 38, 137–143.

Sullivan, D.J., Shore, J.S., Rice, J.A., 1998. Assessment of cation binding to clay minerals

using solid state NMR. Clays and Clay Mineral 46, 349–354.

Tennakoon, D.T.B., Jones, W., Thomas, J.M., 1986a. Structural aspects of metal-oxide-

pillared sheet silicates. Journal of the Chemical Society, Faraday Transactions I 82,

3081–3095.

Tennakoon, D.T.B., Schlo

gl, R., Rayment, T., Klinowski, J., Jones, W., Thomas, J.M., 1983.

The characterization of clay-organic systems. Clays and Clay Minerals 31, 357–371.

Tennakoon, D.T.B., Thomas, J.M., Jones, W., Carpenter, T.A., Ramdas, S., 1986b. Char-

acterization of clays and clay-organic systems. Cation diffusion and dehydroxylation.

Journal of the Chemical Society, Faraday Transactions I 82, 545–562.

Theng, B.K.G., Hayashi, S., Soma, M., Seyama, H., 1997. Nuclear magnetic resonance and

X-ray photoelectron spectroscopic investigation of lithium migration in montmorillonite.

Clays and Clay Minerals 45, 718–723.

Thompson, J.G., 1985. Interpretation of solid-state

C and

Si nuclear magnetic resonance

spectra of kaolinite intercalates. Clays and Clay Minerals 33, 173–180.

References 937

Thompson, J.G., Barron, P.F., 1987. Further consideration of

Si nuclear magnetic

resonance spectrum of kaolinite. Clays and Clay Minerals 35, 38–42.

Tunney, J.J., Detellier, C., 1993. Interlaminar covalent grafting of organic units on kaolinite.

Chemistry of Materials 5, 747–748.

Tunney, J.J., Detellier, C., 1996. Aluminosilicate nanocomposite materials. Poly(ethylengly-

col)-kaolinite intercalates. Chemistry of Materials 8, 927–935.

Wasylishen, R.E., Fyfe, C.A., 1982. High resolution NMR of solids. In: Webb, G.A. (Ed.),

Annual Report on NMR Spectroscopy. Academic Press, New York.

Weir, M.R., Kuang, W., Facey, G.A., Detellier, C., 2002. Solid-state nuclear magnetic res-

onance study of sepiolite and partially dehydrated sepiolite. Clays and Clay Minerals 50,

240–247.

Weiss, C.A., Kirkpatrick, R.J., Altaner, S.P., 1990. Variations in interlayer cation sites of clay

minerals as studied by

133

Cs MAS nuclear magnetic resonance spectroscopy. American

Mineralogist 75, 970–982.

Chapter 12.7: Nuclear Magnetic Resonance Spectroscopy938

Handbook of Clay Science

Edited by F. Bergaya, B.K.G. Theng and G. Lagaly

Developments in Clay Science, Vol. 1

939

Chapter 12.8

TRANSMISSION ELECTRON MICROSCOPY

F. ELSASS

INRA, Unite

de Science du Sol, Centre Versailles-Grignon, F-78026 Versailles, France

Transmission electron microscopy (TEM) techniques allow the following pro-

perties of clay minerals to be determined in a non-destructive way: (i) morphology;

(ii) structure (by selected area electron diffraction, SAED); (iii) lattice imaging (by

high-resolution transmission electron microscopy, HRTEM); and (iv) chemical

composition (by energy dispersive X-ray ﬂuorescence, EDXRF). The above infor-

mation may be obtained on individu al particles, domains of a selected area, or

aggregates depending on the nature of the material studied and the aim of the

investigation.

X-ray diffraction (XRD) was extensively used in the study of phyllosilicates with a

particle size of less than 2 mm(Decarreau, 1990; Drits and Tchoubar, 1990; Moore

and Reynolds, 1997; Parker and Rae, 1998; see also Chapter 12.2). Information

about clay mineral structures, provided by XRD, serves as a basis for the interpre-

tation of data obtained by TEM.

In surveying the literature using only ‘electron microscopy’ as the keyword, the

scientist can overlook a large number of articles where the fundamental properties of

clay minerals were interpreted on the basis of TEM data. Many case studi es would

indicate that TEM can provide insight into weathering and diagenetic phenomena.

In particular, SAED, HRTEM, and EDXRF, either singly or combined, can support

hypotheses concerning processes and mechanisms that occur in clay-rich materials,

such as soils or rocks.

This short review summarizes electron microscopy techniques that are most

commonly used in clay science, gives an update on operating modes, and presents

some major results. The aim here is to guide new users into following validated

approaches for sample preparation and data collection. This would help avoid cre-

ating artefacts that can lead to biased interpretation.

DOI: 10.1016/S1572-4352(05)01034-2

Present address: CNRS, Centre de Ge

ochimie de la Surface, 1 rule Blessig, F-67084 Strasbourg,

France.

12.8.1. TEM TECHN IQUES

A detailed description of the theoretical physics underlying TEM was given by

Drits (1987), Eberhart (1989), Buseck (1992), Reimer (1997), and Kogure (2002).

TEM methods are based on selected interaction phenomena occuring between

matter and the electron beam. Information about unit cells and cell dimensions may

be derived from imag es and diff raction diagrams arising from the diffusion and

diffraction of electron beams. Data on chemical composition may be extracted from

the X-ray ﬂuorescence spectrum of the sample irradiated by the electron beam. Here

we outline the basic principles of electron beam-matter interactio ns to help users

obtain images of clay minerals, and make an intuitive interpretation of these images.

Guidelines are given for setting operating conditions that are appropriate to the

study of clay minerals. However, the most appropriate settings do not necessarily, or

even rarely, correspond to the maximum capability of the apparatus.

A. Formation of Images

The image projected on the screen is produced in two steps. First, an image of the

sample is formed in the plane of the objective lens of the microscope, at low mag-

niﬁcation (around 30x). Second, this (ﬁrst) image is enlarged through a sequence of

magnifying lenses, involving magniﬁcations of 10

–10

. As a result, micrometric to

nanometric objects become visible to the human eye. Since images recorded on

microscope plates or as raster images can be further enlarged using common printers

or scanners, the level of magniﬁcation rarely needs to exceed 10

. Most routine work

is performed at magniﬁcations ranging between 10

and 5  10

. The imaging of a

wide window may be preferred to the imaging of a small detail as the former would

be more representative of the sample under investigation.

B. Principles

The basic principle of quantum physics, describing the propagation of an electron in

the microscope column, and its interaction with the sample, is the particle/wave

dualism (Willaime, 1987). It is expressed by a relation between the energy of the

electron and the associated wavelength of a ray. The kinetic energy, E, of an electron

having a charge, e, and submitted to an accelerating potential, V, is:

E ¼ eV

E can also be expressed in terms of the mass, m, and velocity, v, of the electron:

E ¼

Chapter 12.8: Transmission Electron Microscopy940

The wavelength, l, associated to the elect ron ray is expressed by:

l ¼

where h is Planck’s constant. Under an accelerating potential of 100 kV, the energy E

of an electron is 100 keV, and the wavelength l is equal to 0.00387 nm. For

E ¼ 120 keV, l ¼ 0:0037 nm.

A plane wave is associated with a moving group of electrons. The wave function is

the amplitude of the wave. The square of the wave function is the intensity of light at

a point. For an isolated electron, it represents the probability that the electron is

present at this point.

An electron passing in the neighbourhood of an atom (atomic number, Z)is

repelled by the negative electronic potential of the electron cloud of the atom. As the

electron is deviated without loss of energy, the process is referred to as elastic scat-

tering or diffusion. The scattering factor for the elastic diffusion of an electron (with

an energy eV) is proportional to Z

. As a result, clay minerals containing heavy

elements (high Z value) will produce diffraction patterns (and images) of relatively

high contrast. This is illustrated by comparing the patterns of Fe-rich clays with

those of their Mg- or Al-rich counterparts, irrespective of crystal structure. This

observation applies to micas as well as smectites.

The electron carries its energy in the form of a non-quantized kinetic energy. The

electron will lose this energy in random quantities during successive interactions

(inelastic scattering or diffusion). The scattering factor is markedly increased with

respect to X-rays that lose all their energy in one single interaction. Since the scat-

tering factor for inelastic diffusion is proportional to Z=V

, this parameter is less

sensitive to the atomic number (Z) than elastic diffusion is.

We should also mention that the probability of diffusion (both elastic and inelastic)

decreases as the accelerating voltage of the electron beam increases. Similarly, the

probability of diffusion increases progressively with the loss of energy of the electron

as it travels through matter. In other words, increasing the microscope voltage is not

an appropriate option for observing ‘light’ structures, such as layer silicates.

Wave Transf er in the Microscope

The power of resolution is the smallest distance between two objects giving a sep-

arate image in the apparatus. One can distinguish between point-to-point resolution

and line-to-line resolution, the latt er being easier to obtain in the image.

The wave issued from the object is affected by all interactions that happen be-

tween the electrons and the sample; i.e., it carries all the information about the solid

crossed by the electron beam (Fig. 12.8.1).

The transfer function of the apparatus, used to deﬁne the power of resolution, is

the highest spatial frequency transmitted by the microscope. In other words, the

transfer function deﬁnes the bandwidth of the microscope (Fig. 12.8.2).

12.8.1. TEM Techniques 941

A weak-p hase object has a weak scattering power either because it is extremely

thin, poorly ordered, or its atomic density is low. This applies to clays and clay

minerals with a particle size of o2 mm.

Using a high-energy electron beam and weak-phase objects, the intensity (i.e., the

square of the wave function) of the transmitted beam gives a good representation of

the atomic structure as projected along the optic axis.

Fig. 12.8.1. Image formation in HRTEM. The wave function Cs of the electrons after cross-

ing the object is transferred to the objective lens. The Fourier transform

Csðf Þ of the sample is

corrected by the transfer function of the microscope T (f). The wave function in the focal plane

Ciðf Þ forms the diffraction diagram. The intensity in the image is given by |Ci|2.

Chapter 12.8: Transmission Electron Microscopy942

C. Transfer in the Real Microscope

Optical Settings of the Illumination Stage

If microscopes were perfect and without focusing faults, the wave function of the

image would perfectly reproduce the wave function of the object. If the ﬁlms and

Transfer function

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

1.90

1.75

1.60

1.45

1.30

1.15

1.00

0.85

0.70

0.55

0.40

0.25

0.10

periodicity (nm)

intensity

def=63 (200kV) def=100 (200kV) def=210 (200kV)

Transfer function

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

2.00

1.85

1.70

1.55

1.40

1.25

1.10

0.95

0.80

0.65

0.50

0.35

0.20

0.05

periodicity (nm)

intensity

def=210 (200kV) def=210 (120kV) def=210 (100kV)

Fig. 12.8.2. Simulation of the transfer function in periodicity: (a) for different defocus in a

200 kV microscope and (b) for a 210 nm defocus in microscopes operated at 200, 120, and

100 kV.

12.8.1. TEM Techniques 943

cameras were perfect, the registered intensity would accurately reproduce the

distribution of electrons exiting the object. But microscopes are far from perfect.

Aberrations of the lenses strongly disturb the transfer of the wave function. The

spherical and chromatic aberrations are the most important. These can be minimiz ed

by selecting the app ropriate operating conditions.

The spherical aberration of the objective lens is inherent to the geometry of lens

construction. Beams issued from a same point of the sample and diffracted at different

angles do not converge to the same point in the focal plane of the lens. The man-

ufacturer gives a value of the spherical aberration coefﬁcient (Cs) for the objective lens.

The more elevated the spatial frequency relative to the image (small distances), the

more important is the inﬂuence of the aberration. In some microscopes, the spherical

aberration is reduced by inserting a diaphragm between the objective lens and the focal

plane, thus limiting the divergence angle, a, of the electron beam trajectory.

The radius, r, of the spherical aberration circle is given by 2r

¼ 1=2Csa

Chromatic aberration is due to the instability of the power supply, giving rise to a

spotty image and a loss of spatial resolution. Chromatic aberration can be reduced

by improving the stability (DV =V ) of the accelerating voltage. The coefﬁcient of

chromatic aberration (Cc) can also be minimized by using parallel illumination,

e.g., by increasing the intensity of the magnetic ﬁeld in the last condenser lens. The

radius, r, of the chromatic aberration circle is given by 2r

¼ CcaDE=E.

A small part of the axial astigmatism is due to imperfections of the objective lens.

More common and damaging is the presence of dirt on the beam trajectory. Regular

cleaning of the diaphragms, or preferably replacement with new ones, is recommended.

Optical Settings of the Objective Stage

A diaphragm is placed in the focal plane of the objective so as to limit the number of

waves available to form the image. Although the information contained in this image

will be reduced, the structural information selected by the operator will be enhanced .

This selection can be done by observing the diffraction diagram and choosing a

diaphragm of the proper size which lets through only the main beams and discards

the superﬂuous diffracted beams. The ﬁrst order of diffraction is very important as it

shows the structure (stacking periodicity of the layers in the case of cross-sections of

layer silicates), and should always pass through the diaphragm. Higher orders of

diffraction, indicative of long-r ange order, are only important to keep when dealing

with micas. As a simple rule, only diffracted beams displaying clear spots or circles in

the diffraction diagram should be kept for image formation. Diffused diffraction will

only add ‘noise’ to the image.

D. High Resolution Imaging

High-resolution electron microscopy deﬁnes the operating conditions to be used in

order to obtain images of spatial frequencies (f) of atomic planes for weak-phase

objects wi thout major distortion.

Chapter 12.8: Transmission Electron Microscopy944

The transfer function of frequencies is used to characterize the power of the

microscope in high-resolution imaging. It is used to deﬁne the defocus settings

required to optimize the intens ity in the transferred image. T(f) is proportional to 2

sin g(f) with:

g fðÞ¼2P Csl

þ Dzl



As an illustration, the term 2 sin g(f) was calculated for three microscopes, having

different spherical aberration coefﬁcients and operated at different acceleration

voltages (Fig. 12.8.2 ). The assumed conditions correspond to the most common

microscopes currently used in laboratories. The right transfer corresponds to neg-

ative values of 2 sin g(f). It enables a direct interpretation of the position of atoms.

The positive values of 2 sin g(f) result in reversed intensities in images. Objective

diaphragms must be adapted to the transfer curve in order to eliminate improper

transfer.

12.8.2. ENERGY DISPERSIVE X-RAY FLU ORESCENCE (EDXRF)

A. Chemical Analysis by EDXRF

Clay mineral layers, particles, and aggregates have speciﬁc chemical compositions

that can be established using the characteristic X-ray ﬂuorescence of the constituting

ions under the electron beam. The energy of the emitted X-rays is characteristic of

the ions in question. Since the intensity of the peaks is proportional to the emission

by the atoms, the intensity of the emitted X-rays reﬂects the concentration of ions in

the irradiated domain, according to the Cliff-Lorimer law (Cliff and Lorimer, 1972,

1975). Quantitative X-r ay microanalysis can be performed if the specimen satisﬁes

the thin-ﬁlm criterion (Metha et al., 1979). This means that the emitted X-rays are

neither absorbed by interaction with neighbouring atoms nor cause X-ray ﬂuores-

cence. The ratio technique uses the expression:

¼ k

where I

and I

are the simultaneously measured characteristic X-ray intensities, and

and C

are the fractions of two elem ents A and B in the thin ﬁlm. The constant

is independent of sample thickness and composition as long as the thin-ﬁlm

criterion is satisﬁed (Golds tein, 1979). However, k

varies with operating voltage

and is sensitive to absorption by (i) the window of the EDS detector; (ii) the carbon

that builds up on the window; and (iii) the carbon-contaminated surfa ce layer. It is

therefore recommended that k values be calibrated for a speciﬁc instrument and

12.8.2. Energy Dispersive X-Ray Fluorescence (EDXRF) 945

under speciﬁc operating conditions. Different instruments may therefore yield

different values of k

. The effect of absorption is especially important for X-ray

energies of o1.6 keV, such as the characteristic emmissions of Na, Mg, and Al

(Lorimer et al., 1977).

The detector is a Si/Li diode mono-crystal maintained under vacuum and at a low

temperature (using liquid nitrogen). The complete emission spectrum is analysed by

energy dispersion, the detected X-rays being stored in a multi-channel analyser (Adda

et al., 1993). An energy spectrum is displayed on the screen. The detector is sensitive to

X-rays with an energy between 1 and 20 keV. This allows almost all elements to be

detected by at least one of their characteristic X-rays (K, L, M series). However, light

elements with Z o11 (e.g., sodium) may be difﬁcult to detect using standard detectors

with a beryllium window, while elements with Z o6 (e.g., carbon) require detectors

with an ultra-thin window. The current energy resolution is around 133 eV, permitting

all characteristic X-rays of the Ka series to be separated. X-rays of the L and M series

are more difﬁcult to separate. As the concentrations of the detected elements are

normalized to 100, the procedure is semi-quantitative. The structural formulae of

minerals can be derived but not the complete OH and H

O contents.

B. Calibration for Quantitative Chemical Analyses of Clay Minerals

The calibration factors, also called ‘k-ratios’, are given in the software package of

modern analytical systems. They are set for a given conﬁguration of detection, the

type of detector, and the accelerating voltage. The calibration factors must be

checked experimentally by the analyst using his/her own standards and corrected in

the software when necessary (and if possible). Analysis of clay minerals requires the

use of clay mineral standards and the same operating conditions. This is because

irradiation damage, leading to a loss of matter, is unavoidable. The level of damage

is strongly related to (i) the structure of the matter and (ii) the operating conditions,

notably beam intensity. Hydroxyl groups are decomposed by heating under the

electron beam, and atoms are volatilized in the microscope column (Eberhardt,

1989). Ahn et al. (1986) observed numerous lenticular ﬁssures in electron micro-

graphs of paragonite, due to irradiation damage. AEM analysis shows a signiﬁcant

loss of sodium from the beam- damaged paragonite, the degree of damage being a

function of exposure to the beam. The authors suggested that the ﬁssures are directly

associated with the interlayer Na

ion, an d that beam damage effects in micas are

related to interlayer cations. The less well crystallized the material, the greater is the

loss of matter. Clay minerals with weakly bound interlayer cations, as is the case for

hydrated cations, are dramatically sensitive to volatilization in the microscope

column.

Since clay minerals can easily be damaged by irradiation, mild analytical con-

ditions should be used to limit mass loss. Scanning transmission electron microscopy

(STEM), using a small probe and intense electron ﬂux, is known to be highly

damaging to structures such as plagioclase and kaolinite (Mackinnon et al., 1986 ;

Chapter 12.8: Transmission Electron Microscopy946