Artificial stacking faults can be created within a well-crystallized kaolinite by intercalating and removing hydrazine. X-ray powder patterns with electron microscopy show that the created defects are ±b/3 translations with a proportion 0.30. The infrared spectrum of the treated kaolinite is not modified with respect to the starting one. On the other hand, a natural kaolinite containing defects by displacement of Al vacancies in a similar proportion shows an infrared spectrum significantly different from that of a well-crystallized kaolinite. The modification of the infrared spectra of natural disordered kaolinites is then related to the presence of defects by change of Al vacancy positions

Until now, the different attempts to describe the defects of kaolinites were based on the ideas that (i) the hkl reflections with k = 3n are Bragg reflections, while (ii) the hkl reflections with k ≠ 3n are affected by ±b/3 translations or ±2π/3 rotations. With regard to this conception, this work provides several important precisions: (i) The h, 3n, l′ reflections are true continuous diffraction bands, more or less modulated, and disturbed by the existence, in the stacking, of random shifts parallel to the layer plane. (ii) The major defect in natural kaolinites is not the ±b/3 translation, but the displacement from one layer to the other (or from one domain to another in the same layer) of the Al vacancies. (iii) The model containing true rotation of layers should be rejected because it does not allow us to interpret all the different parts of the experimental diagrams. Such a concept of defects in kaolinites is in agreement with the existence of polytypes of kaolinite, with the presence of twins, and allows us to interpret some physico-chemical properties such as the infrared spectra.

This paper describes the calculation of the diffracted intensity for models of powdered minerals with lamellar structure, in order to compare experimental and theoretical X-ray diagrams. The calculation takes into account at the same time (i) the possibility of coexistence of different kinds of layers in the stackings, (ii) the possibility of stacking faults between the layers (translative or rotative ones) and (iii) the unavoidable orientation of particles in the powder sample (the method of determination of the orientation is described).

Previous studies of the defect structure of kaolinite have examined samples having a restricted level of defects. This study examined nine kaolinite samples having a large diversity of defect contents, as indicated by Hinckley indexes ranging from 1.44 to 0.18. The samples were chosen so as to cover this range in as regular a manner as possible. The types and abundances of the defects were determined by examining the X-ray powder diffraction profiles for the 02,11 and 20,13 bands. The diffraction intensities were measured by counting for a fixed time in steps of 0.01°2θ. Analysis of these diffraction profiles indicated that (1) the major defect is the existence of a translation between adjacent layers, which is not the usual (approximately a/3), but is related to that translation by the pseudo-mirror plane coincident with the long diagonal of the unit cell; (2) the existence of a few C layers among the B layer stacking is a minor defect; (3) many of the samples could be accurately modeled only by assuming the existence of two kaolinite phases; (4) the existence of only a few C layers in some samples does not support the idea of a continuous series from kaolinite to dickite through disordered intermediates; and (5) the Hinckley indexes of several samples depend on the relative proportions of the two types of kaolinite in the mixture.

The nine kaolinite samples fall into three groups: those having a low to moderate abundance of defects (Hinckley index > 0.43) are mixtures of two types of kaolinite (one almost free of defects, the other richer in defects); those having low Hinckley indexes (0.43 to 0.18) are single phases with different proportions of defects; and those which contain a single type of kaolinite, unlike the others in the nature of the interlayer translations and the greater abundance of C layers. The agreement between calculated and observed X-ray diffraction profiles is excellent for all specimens, except one sample (from Charentes) for which the fit is acceptable but not perfect.

A comparison of the structural characteristics of the kaolin-group minerals, mainly kaolinite and dickite, shows that they differ in both the two-dimensional periodicity in the 1:1 layers and the rotation angles of the polyhedra. Distortions in a real 1:1 layer, compared with an idealized layer, do not allow such stacking faults as ± 120° layer rotations and vacancy displacements, because the second layer is incommensurable with the first. The 1:1 layer structure and the fact that the unit cell is symmetrical with respect to the plane passing through the long diagonal of the unit cell suggest the possibility of defects resulting from the two stacking sequences for the same layers. For a regular alternation of translations, a halloysite-like structure should be the end-member of such a series of defect kaolinite types.

The formation of layers having vacant octahedral C-sites is another possible type of fault. Because of the minor dilference between γ and 90°, dickite-like layers should exist. A regular alternation of B and C layers yields dickite as the end-member structure. In materials containing few defects, stacking faults of both types lead to similar X-ray powder diffraction patterns. Thus, the nature of the stacking faults is difficult to determine experimentally. In materials containing many defects, however, the two models lead to different calculated diffraction patterns. Therefore, only a study of defect-rich types of kaolinite can determine which types of defects exist in natural kaolinite samples.

A very simple technique is proposed for a quantitative or semiquantitative interpretation of X-ray diffraction (XRD) patterns for two-component mixed-layer structures. It is suitable for a determination of the Reichweite (R) values and proportions of component layers from graphical simulations of basal peak positions for mixed-layer structures with definite layer types. This technique can be successfully used for illite-smectites, but the accuracy of the results obtained for other mixed-layer structures is somewhat lower. In addition to the graphical technique, simple linear relationships are proposed for the calculation of layer proportions. Such relationships can be easily obtained for any mixed-layer structure with any R and any thicknesses of interstratified layers.

Structural parameters reported in the literature for mixed-layer illite-smectites, kaolinite-smectites, etc., were used to check the reliability of the method presented. It is concluded that the technique works well and produces parameters that are in agreement with those published.

X-ray phase analysis of clays is difficult because these materials generally consist of a mixture of different phases, i.e., mixed-layer minerals, individual clay minerals (non mixed-layer), and associated minerals, such as calcite and quartz. The analysis requires knowledge that presently is incorporated in a computer-based expert system. This expert system is capable of a) identification of associated minerals; b) identification of individual clay minerals; c) identification of the nature of the mixed-layer minerals; d) approximate structural characterization of the mixed-layer minerals; and e) precise structural determination of the mixed-layer minerals by comparison of experimental X-ray diffraction (XRD) patterns with calculated patterns for different models. Accuracy of the conclusions drawn by the expert system has been verified with literature data. Programs for the structural characterization of mixed-layer minerals must allow a) modification of the structural characteristics, abundances, and order-disorder distribution of the layers; b) modification of the distribution of the sizes of coherent scattering domains; and c) consideration of mixed-layer clays with more than two components. Two programs were written to calculate the XRD patterns of two- and three-component mixed-layer minerals consisting of any layer type and without any limitation in the order-disorder relationships.

The structural transformation of dioctahedral 2:1 layer silicates (illite, montmorillonite, glauconite, and celadonite) during a dehydoxylation-rehydroxylation process has been studied by X-ray diffraction. thermal analysis, and infrared spectroscopy. The layers of the samples differ in the distribution of the octahedral cations over the cis- and trans-sites as determined by the analysis of the positions and intensities of the 11l, 02l reflections, and that of the relative displacements of adjacent layers along the a axis (c cos ß/a), as well as by dehydroxylation-temperature values. One illite, glauconite, and celadonite consist of trans-vacant (tv) layers; Wyoming montmorillonite is composed of cis-vacant (cv) layers, whereas in the other illite sample tv and cv layers are interstratified. The results obtained show that the rehydroxylated Al-rich minerals (montmorillonite, illites) consist of tv layers whatever the distribution of octahedral cations over cis- and trans-sites in the original structure. The reason for this is that in the dehydroxylated state, both tv and cv layers are transformed into the same layer structure where the former trans-sites are vacant.

The dehydroxylation of glauconite and celadonite is accompanied by a migration of the octahedral cations from former cis-octahedra to empty trans-sites. The structural transformation of these minerals during rehydroxylation depends probably on their cation composition. The rehydroxylation of celadonite preserves the octahedral-cation distribution formed after dehydroxylation. Therefore, most 2:1 layers of celadonite that rehydroxylate (~75%) have cis-vacant octahedra and, only in a minor part of the layers, a reverse cation migration from former trans-sites to empty octahedra occurred. In contrast, for a glauconite sample with a high content in IVA1 and VIAl the rehydroxylation is accompanied by the reverse cation migration and most of the 2:1 layers are transformed into tv layers.

Interaction between metal Fe and a variety of natural and synthetic smectite samples with contrasting crystal chemistry was studied by scanning electron microscopy and X-ray diffraction from experiments conducted at 80°C. These experiments demonstrate an important reactivity contrast as a function of smectite crystal chemistry. An XRD method involving the use of an internal standard allowed quantification of the relative proportion of smectite destabilized as a function of initial pH conditions as well as of smectite structural parameters. In mildly acidic to neutral pH conditions, a significant proportion of metal Fe is corroded to form magnetite without smectite destabilization. Under basic pH conditions, smectite and metal Fe are partly destabilized to form magnetite and newly-formed 1:1 phyllosilicate phases (odinite and crondstedtite). More specifically, systematic destabilization of both metal Fe and smectite is observed for dioctahedral smectites while trioctahedral smectites are essentially unaffected under similar experimental conditions. In addition, smectite reactivity is enhanced with increasing Fe3+ content and with the presence of Na+ cations in smectite interlayers. A conceptual model for smectite destabilization is proposed. This model involves first the release of protons from smectite structure, MeFe3+OH groups being deprotonated preferentially and metal Fe acting as proton acceptor. Corrosion of metal Fe results from its interaction with these protons. The Fe2+ cations resulting from this corrosion process sorb on the edges of smectite particles to induce the reduction of structural Fe3+ and migrate into smectite interlayers to compensate for the increased layer-charge deficit. Interlayer Fe2+ cations subsequently migrate to the octahedral sheet of smectite because of the extremely large layer-charge deficit. At low temperature, this migration is favored by the reaction time and by the absence of protons within the di-trigonal cavity. Smectite destabilization results from the inability of the tetrahedral sheets to accommodate the larger dimensions of the newly formed trioctahedral domains resulting from the migration of Fe2+ cations.

The Markovian model of mixed-layering that has been used until now for the modeling of X-ray diffraction (XRD) patterns of 2:1 phyllosilicates describes the mixed-layer crystals as stacks of ‘layer units’ (LUs) that associate a 2:1 layer with an interlayer content. This model is not consistent when it is applied to a mixed-layer mineral (MLM) involving electrically charged layers, i.e. most of the mixed-layer phyllosilicates with 2:1 layers. Two consistent models can be proposed for these MLMs, which, instead of LUs, stack ‘interlayer units’ (IUs), composed of an interlayer content sandwiched between two half-layers. The IU-NPL model (interlayer units, non-polar layers) imposes non-polarity on the 2:1 layers resulting from the stack of these IUs, which implies restrictions on the succession of the IUs. In the other consistent model, the IU-PL one (interlayer units, polar layers), these restrictions are not imposed. These two models reproduce the Non-Polar 2:1 Layer Model and the Polar 2:1 Layer Model described by Altaner and Ylagan for illite-smectite mixed layering, and the second model corresponds to the stacks of O0.5 T I′ T O0.5 units described by Olives et al.

The present work points out the similarities and discrepancies of the calculated XRD patterns by using computer programs designed for the three models of mixed-layering (LU, IU-NPL and IU-PL) and for two-and three-component MLMs. Illustrations are provided for some I-S and I-S-V MLMs.

The IU-PL model leads to XRD patterns similar to those of the LU model. It is in agreement with lattice-energy calculations, expandability measurement, HRTEM imaging and NMR spectroscopy.

This study presents a mathematical formalism describing diffraction effects from periodic and mixed-layer minerals in which the outer surface layers of crystals possibly differ from layers forming the core of the crystals. The X-ray diffraction (XRD) patterns calculated for structure models of chlorite and irregular chlorite-smectites terminated on both sides of the crystals by either brucite-like sheets or 2:1 layers show the strong influence that different outer surface layers have on the distribution of basal reflection intensities. Simulation of the experimental XRD patterns from two chlorite samples having different Fe contents shows that in these two samples the chlorite crystals were terminated by brucite-like sheets on both sides. In contrast, crystals in a corrensite sample were terminated by water molecules and exchangeable cations. The nature of diffraction effects due to outer surface layers is discussed.

The general approach to the problem of the real structure of smectites requires an analysis based on the relationship between structural characteristics and diffraction patterns. This paper, which considers only the models corresponding to dioctahedral smectites saturated by K-cations and collapsed, includes: 1. Successive consideration of all models which are crystallochemically possible. These models may differ in (i) the structure and chemical composition of layers and interlayer spaces; (ii) the azimuthal orientations, translations and the mode of alternation of the layers; (iii) independent parameters which describe quantitatively the models (e.g. abundance of each type of layer, probability parameters defining the succession of layers, …). 2. Calculation, in all accessible domains of reciprocal space, of the distribution of intensities and profile variations, obtained by changing only one parameter at a time, that defines one type of structural feature (e.g. cation distribution in individual layers, stacking of the layers, nature of stacking faults, …). 3. A systematical analysis of the calculated diffraction patterns to establish the diffraction criteria which will help to interpret the experimental data explicitly.

Until recently, the determination of the defect structures (previously referred to incorrectly as “crystallinity”) of kaolinites has been obtained in one of two ways: (1) measurement of the Hinckley index, or (2) by comparing calculated X-ray diffraction patterns based on a model of the defect structure (including types of defects and abundances) with experimental diffraction profiles. The Hinckley method is simple and easy to perform but contains no real information about the defect structure. Calculated XRD patterns are based on real defects but these calculations are time consuming and require some skill in application. Another approach is proposed: an expert system which will accurately describe the defect structure of kaolinites based on a few measurements taken from a normal powder diffraction profile. This system has been verified for nine kaolinite samples for which the defect structure was previously determined by comparison of calculated and observed diffraction profiles. The expert system reproduced the correct defect structure for each of the samples.

The expert system described in the first part of this paper has been applied to the identification of mixed-layer phyllosilicates (mica-smectite, mica-vermiculite, chlorite-smectite, chlorite-vermiculite, chlorite-swelling chlorite, chlorite-mica, chlorite-talc, kaolinite-smectite, talc-smectite), and to the determination of their structural parameters (Reichweite, R, and proportions of constituting layers, Wi). The expert system has been run utilizing the data extracted from (1) experimental XRD patterns for which structural parameters had been evaluated by comparison with calculated patterns, or (2) patterns calculated using pre-selected values of the structural parameters. In all cases examined, the expert system provided correct conclusions concerning the identification of a mixed-layer phyllosilicate and the value of the Reichweite, while the abundances of the component layers were evaluated with a margin of error usually <5%.

The (02,11) X-ray diffraction band from a low-defect kaolinite from Cornwall (Hinckley index HI = 1·22) was examined to determine the defect structure. No combination of interlayer translations and admixing of dickite layers accurately modelled the observed diffraction pattern. Calculated diffraction patterns which gave a good agreement with the shape, position, and intensity of the observed peaks, uniformly had inter-peak intensities which were too weak. By treating the kaolinite as a mixture of low-defect (HI = 1·76) and moderate-defect (HI = 0·29) kaolinites, the agreement between the observed and calculated patterns was improved substantially. The existence of a mixture of two kaolinites was also found for a number of low-defect samples (HI > 0·4) from Georgia and Cornwall, and may be of even wider occurrence. The HI, which is very sensitive to the inner-peak intensities, does not estimate the types or abundances of various structural defects (the classical “crystallinity”), but is related directly, in a non-linear fashion, to the proportions of the two kinds of kaolinite which are present in the sample.

Comme la plupart des vermiculites, l'hydrate à une couche de la vermiculite-Na (d001 = 11·85 Å) présente une structure semi-ordonnée, avec des défauts d'empilement (glissements selon l'axe b). Le plan (h0l)* du réseau réciproque est entièrement discret, tandis que les rangées n'appartenant pas à ce plan sont plus ou moins diffuses dans la direction c*. La projection x0z est obtenue par une classique analyse de Fourier bidimensionelle. Dans les plans (0kl)* et (1kl)*, certaines rangées semblent presque discrètes: ceci permet d'essayer de déterminer la translation selon l'axe 0y entre deux feuillets adjacents. Le calcul des intensités des diffusions le long des rangées des plans (0kl)* et (1kl)* conduit à ne retenir qu'un seul modèle. Le désordre résulte d'une distribution aléatoire de trois translations parallèles à l'axe Oy: l'une est nulle et les deux autres sont de +0·307b et −0·307b. Le principal intérêt de ce travail est de présenter une nouvelle méthode d'approche pour l'étude des structures semi-ordonnées.

Celadonite and glauconite samples heated at different temperatures were studied by X-ray and electron diffraction. For dioctahedral micas the in-plane component of the translation between layers (ccosβ/a), which is strongly dependent on the position of the vacant octahedral site, significantly decreases at temperatures greater than the temperature of maximum dehydroxylation. The simulation of XRD patterns from different structural models reveals the actual crystal structure of dehydroxylated samples as well as the dynamics of the structural transformations. In the nonheated state the samples consist of tv (trans-vacant) 2:1 layers. During dehydroxylation, cations migrate from cis- into trans-octahedra and have 5-fold coordination. In the averaged unit-cell the ‘residual’ anions formed after the dehydroxylation reaction occupy the former OH sites with probability equal to 0.5. The migration of octahedral cations is accompanied by the transformation of the C-centred layer unit-cells into primitive ones. In contrast to Fe, Al and Mg cations have a greater ability to migrate.

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, orderdisorder in mixed-layer systems and finite crystal size as a lattice 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. XRD and IR, EXAFS and Mo¨ssbauer spectroscopies, etc.). Another conclusion is that accurate structural determination provides the key to the genesis of clays.

An expert system has been designed for the structural characterization of disordered phyllosilicates as well as the mixed-layer minerals and the minerals containing stacking faults or isomorphic substitutions. X-ray diffraction being the simplest tool for structural characterization, the expert system uses such data but can account for data provided by any characterization technique (infrared spectroscopy, EXAFS, etc.). The expert system aims to summarize in a unique set the different approaches of the structural characterization using diagrams or tables which have been proposed, avoiding the requirement of the knowledge of their application field and excluding calculations. It runs in an interactive way, to mimic the way a human expert would draw its conclusions.

The quantitative interpretation of the reflection intensities of the SAED patterns of glauconites reveals the mechanism of migration of the octahedral sheet cations during heating up to 750°C. It confirms that Mg2+ has a greater ability than Fe3+ to migrate from cis-to trans-sites, as previously found by means of XRD pattern modelling. For samples heated to 650°C, the two formerly vacant trans-sites in the base-centred unit-cell become occupied, but differently, which leads to a primitive unit-cell, and the cis-sites remain filled almost entirely by Fe3+ only. The cation migration occurs through the nearest shared edges in the [010] and [310] directions. The samples heated to 750°C reveal a base-centred super-cell with A = 3a and B = b. All Mg cations leave the cis-sites to occupy completely four of the six available trans-sites of the super-cell. Migration also occurs through the shared edges in the [010] and [310] directions. The primitive unit-cell is not an intermediate step in the migration process leading to the super-cell. The existence of additional satellites in the SAED patterns of some crystals heated to 750°C corresponds to the existence of antiphase domains with a 3b/2 width and an antiphase shift of a/2.