The combined use of electron microscopy and of selected area diffraction (SAD) has been applied to the study of montmorillonites from Wyoming and from Camp-Berteaux, of nontronite, and of hectorite. The study shows that only the elementary layers of the Wyoming montmorillonite and of the nontronite are single two-dimensional crystals. The elementary layers of the montmorillonite from Camp-Berteaux are formed by edge-to-edge associations of very small elements with mutual orientations of about 60° or multiples of 60°. The layers of hectorite are formed by edge-to-edge association of laths with fluctuations of orientation of the order of 10° rotation around adjoining edges. The SAD patterns of Wyoming montmorillonite show that the single layer plane symmetry group of this mineral is c1m1; nontronite layers belong to the symmetry group c2mm. The study of thick particles shows that in turbostratic smectites the layers are stacked with mutual rotations around the perpendicular to their plane. This mode of stacking explains the absence of hkl reflections. High resolution diffraction (HRD) patterns obtained with sample inclined to the electron beam show that only the structure of hectorite approaches the ideal model; nontronite and montmorillonite exhibit appreciable distortions with respect to the ideal model. HRD diagrams provide precise information for the refinement of the crystal structures of smectites

Spray drying dilute suspensions of bentonitic montmorillonite produces a powder that shows totally random orientation of the crystallites within a sample large enough to diffract X-rays. The powder is collected by an electrostatic precipitator and can be handled in the normal mounting processes without introducing preferred orientation. Electron micrographs show this powder to be composed on a small scale of thin, crumpled, and rolled films. The extremely small montmorillonite crystallites that make up the film are oriented with [001] directions perpendicular to the film surface. Orientation within the plane of the film is random as shown by selected area electron diffraction. Crumpling and rolling of the film is sufficient to make the orientation of [001] directions random in three dimensions in a large sample when X-ray diffraction is registered.

The X-ray diffraction patterns all show diffraction maxima (both hk and 00l), and their relative intensities with respect to each other can be determined. The line broadening of the 06 and the 003 peaks was studied. The average crystallite size as calculated from the line broadening varied from six to eleven unit layers thick for four bentonitic montmorillonites. The average lateral dimension of crystallites varied from 140 Å to 250 Å. Ratios of lateral dimensions to thickness varied from 2.3 to 3.4.

Volcanic ash of latitic composition in the Pliocene Bidahochi formation of northeastern Arizona has been altered in place to smectite clay. The Cheto clay, a 4-ft thick bed of high-quality activable smectite, offers an excellent opportunity for electron-optical study of the alteration processes in volcanic glass shards, spherulites, and aggregates in the tuff. No completely unaltered tuff remains in the exposed bed, but transitional materials showing progressive alteration from vitric tuff toward massive smectite have been collected and examined petrographically, electron-optically, and by X-ray diffraction.

The partly altered tuff shows several clearly identified structures relatable to the fresh material: spherical to subspherical spherulites, conchoidally fractured subtriangular to irregular vitric shards dominated by smoothly curved pitted surfaces, massive essentially structureless glass, and columnar fused aggregates which appear interstitial to the shards. Upon alteration these aggregates display an irregular comb-like structure from a planar base in cross-section, and a hackly polygonal to random texture normal to the comb structures. X-ray diffraction reveals only montmorillonite in the materials examined, although minor quartz and plagioclase feldspar have been reported elsewhere in the partially altered ash.

Argillization of the vitric tuff involves several phenomena: (1) the development of layering or banding in the glass; (2) the development of a braided surface aspect perhaps related to banding; (3) the inception of arcuate subparallel lineations on internal curved surfaces of the glass; (4) the growth of pseudohexagonal, weakly curled flakes whose boundaries appear threadlike on glass surfaces, and (5) the growth of grossly hexagonal matted granular networks in completely altered glass. Bladed, comb-like smectite develops along interfaces between glass and more resistant shards and from fused columnar aggregates in the original material. The mechanism appears to be one of essentially pervasive nucleation and growth of smectite in the devitrifying glass with the early development of preferred orientation of individual smectite crystallites. This orientation, manifested in layering and braiding, may reflect simple response to the space problem or response to cryptostruetures and polymerization in the original glass. Relic structures relatable to the unaltered vitric tuff are not obliterated and can be discerned even in the completely altered ash.

Previous studies of diffraction band profiles of montmorillonite are extended to the 13, 20 band which, for reasons discussed, is more sensitive to structural details than the diffraction bands considered previously. Hydration of Na-montmorillonite produces appreciable sharpening of this band, indicating, contrary to previous results, that the water layers have at least a partially ordered arrangement. For Cs-montmorillonite, the single water layer hydrate shows only a small sharpening of the band profile, possibly attributable to the water layer, but which clearly indicates that the Cs ions cannot be significantly displaced from their positions in the anhydrous material. For Ba-montmorillonite, the two-layer hydrate shows a small broadening of the diffraction band, which is possibly the result of a partial ordering of the water layers together with a movement of the Ba ions away from the hexagonal holes.

A number of otherwise poorly understood compositional-structural relationships in mixed-layer illite/montmorillonites appear to be explained by ordered interlayering of 2:1 units. The ordering appears to be between high and low charge 2:1 units. The relationship between per cent expandable layers and mean lattice charge in natural materials shows a scatter that is nicely bracketed by hypothetical curves constructed assuming random or completely ordered interlayering for several types of 2:1 charge distribution. Natural samples with 30–50% expandable layers plotting near the “ordered” curves have diffraction peaks at 25 Å for two water layers in the expandable layers and 27 Å with two glycol layers; samples plotting near the “random” curves do not have these peaks and have poorly developed 001/001 reflections.

Some tentative conclusions are: (1) illites are virtually non-expandable at lattice charges significantly less than that of ideal mica because of ordering of high-low charge 2:1 units, (2) illites and mixed-layer illite/montmorillonites with the same mean lattice charge show a range of expandability because of varying degrees of ordering, (3) the highly variable CEC of samples with the same expandability also results from a varying degree of ordering.

An interesting petrological implication results when it is realized that a sample with a given bulk composition can have a wide range of expandability, depending on ordering. Low expandability, achieved by ordering, should be promoted by formation at high pressures. Glauconites, known to form largely at the time of sedimentation, appear to be dominantly randomly interstratified; illites and illite/montmorillonites, which may form on deep burial of argillaceous sediments, are dominantly ordered.

Vermiculite (Libby, Mont.) was ground in a Waring blender in a 1 m NaCl solution and, after removal of excess electrolyte, the clay fraction was separated by sedimentation. The clay was predominantly vermiculite: X-ray diffraction patterns of Ca-saturated and oriented specimens showed an intense and sharp 15 Å and a weak 25 Å diffraction maxima and their integral orders. The intensity of the 25 Å reflection, attributed to regularly interstratified layers of vermiculite (15 Å) and mica (10 Å), was less than 20% of the 15 Å peak.

Additions of varying amounts of potassium or cesium, ranging from 10% to 100% of exchange capacity, to Ca-saturated clay showed that the collapse of the vermiculite lattice proceeds through a 1:1 regular interstratification of a 15 Å and a 10 Å lattice. Successive additions increased the 25 Å diffraction peak at the expense of the 15 Å reflection until the entire sample was interstratified. Further additions of K (or Cs) reduced the intensity of the 25 Å reflection and produced a 10 Å reflection until the entire sample was collapsed to 10 Å and no 25 Å reflection was recorded. These observations point out that under certain environmental conditions, the diagenetic formation of micas from vermiculite may proceed through an interstratification of the two in a manner analogous to weathering of biotite to vermiculite through an interstratified stage.

One-dimensional Fourier synthesis from the intensities of the 00ℓ diffraction maxima of the interstratified mixture was carried out. In addition, a mechanism for the formation of the interstratified mixture was postulated: the replacement of Ca by K (or Cs) in one layer reduces the effective negative charge on the adjacent layer. Consequently, the K cannot replace the Ca in this but replaces the Ca in the next layer forming the interstratified mixture.

Transmission and diffraction electron micrographs were made of illite and montmorillonite to study the changes that take place as the clays were heated continuously to 1200°C.

In both of the clay minerals studied, expulsion of a highly fluid material occurred subsequent to dehydroxylation and prior to the development of the first-formed high temperature phases. It was hypothesized that this material represented the expulsion of constituents in the clay in excess of those needed for the formation of the first high-temperature phase. Electron diffraction indicates that the high-temperature phases that formed all developed with some preferential orientation, and were strongly influenced in their development by the structure of the original clay.

The sorption of water, at 490°C and ca. 23–24 torr, on clay mineral surfaces has been studied using thermogravimetry and infrared absorption spectroscopy. Results from a series of synthetic hectorites, with and without lattice fluoride (see Granquist and Pollack, 1960), point to the presence of F− in the clay lattice as an important condition for the sorption. Similar studies on natural hectorite and montmorillonite, synthetic interstratified montmorillonite/mica with and without fluoride, and Al2F6 support this interpretation of the role of the F− ion and show that dioctahedral 2:1 clays sorb more water at 490°C than do trioctahedral 2:1 clays. The infrared spectra demonstrate that the sorbed water is for the most part hydrogen-bonded to the surface.

The relationships between spectrum and structure in layer-silicates are reviewed, and applied in the study of structural changes occurring during the heating of montmorillonites up to dehydroxylation, and their subsequent rehydroxylation. Information given by infrared spectroscopy on the binding of water in expanding layer silicates is presented, and the physical and chemical processes associated with entry of basic, neutral and acidic molecules into the interlayer space of these minerals are illustrated for ammonia, ethylamine, pyridine, nitrobenzene, and benzoic acid. Problems associated with the study of soil clays, which are often complex mixtures including poorly ordered and amorphous constituents, frequently firmly combined with organic matter, are discussed.

New evidence is presented concerning the environment of the two types of hydroxyl group in beidellite. The thermal stabilities of NH4+ and lattice OH in montmorillonite and beidellite, and the properties of their dehydroxylates, are contrasted. The nature of the collapsed phase formed in Li-, Mg-, and NH4-montmorillonite at 300–500°C is discussed. The presence of weak hydrogen bonds between lattice oxygens and interlayer water is established, although it is shown that the strength of hydrogen bonds formed between NH4+ and lattice oxygens is dependent on the sites of substitution in the layer lattice.

The 1–5 μ size fractions of different ground muscovites were treated for a 30-day period for K+ removal with a pH 7.5 solution containing sodium tetraphenylboron. The initial K+ contents ranged from 211 to 219 me/100 g. After the extraction period, the final K contents ranged from 21 to 37 me/100 g. The apparent structural charge on the expanded material decreased to values ranging from 156 to 184 me/100 g, which are within the charge range for natural vermiculites. The b-dimensions of these laboratory-produced vermiculites were found to vary with the exchange ion and with the hydration state of the ion. For hydrated samples, saturation with Cs+, Li+, and Mg2+ ions increased the observed b-dimension in comparison to that of the corresponding parent mica, whereas saturation with Sr2+ and La3+ ions had little effect on b. After dehydration at 350°C, only Cs+-saturated samples had a b-dimension greater than that of the parent mica. The observed b-dimension for the dehydrated samples was found to be a direct function of the crystal radius of the interlayer ion. Apparently, as the ion dehydrates, the surface oxygen triads rotate until some of the oxygens “lock” onto the ion, limiting the minimum b-dimension. Before dehydration, however, the water of hydration in the interlayer region is evidently held with sufficient energy to limit rotations of the oxygen triads that give rise to a decrease in b. When Li+ and Mg2+ ions occupy the interlayer region, their water network apparently even produces a slight increase in the observed b-dimension.

A vermiculite regolith sample from Llano, County, Texas, was divided into 16–32, 8–16, 4–8, 2–4, 1–2, ½–1, ¼–½, 1/10-1/4, and 5/100–1/10 micron equivalent spherical diameter size-classes. A well-crystallized vermiculite found in association with the regolith was ground to −400 mesh.

The samples were analyzed by X-ray diffractometry and differential thermal analysis, and cation exchange capacity measurements were made for selected size fractions. Chemical analyses by X-ray fluorescence revealed a systematic increase in silicon with decreasing particle size. An increase in the iron content and a decrease in magnesium were also directly correlated to a decrease in the crystallite size. Aluminum content was constant throughout the range of particle sizes.

X-ray analyses showed that the coarser particles were trioctahedral, whereas the smaller crystallites were dioctahedral. These data were supported by recast formulae based on the chemical analyses. The recast formulae were substantiated by DTA and cation exchange capacity data as being reasonable approximations of actual formulae.

These observations are interpreted as an expression of chemical alteration that clay minerals may undergo in an aqueous environment. The crystallographic, cation exchange capacity, DTA, and swelling properties are a function of the chemical composition of the crystallites. Chemical composition of the crystallites is shown to be directly correlated to particle size.