Synthetic clays were prepared hydrothermally at 300° to 350°C and 15,000 psi from gels with compositions calculated to give a series of octahedrally and tetrahedrally substituted, 2:1-type layer silicates with charges of 0.5N, 1N, 1.5N, 2N, 2.5N, and 3N, where N equals the “normal” layer charge of montmorillonite. Besides montmorillonites and beidellites, the products also contained other components, kaolinite, saponite, paragonite and possibly amorphous materials, depending on the initial gel compositions. Cation exchange capacities of <2µ fractions treated for removal of amorphous constituents were within the narrow range of 0.95–1.35 meq per g rather than 0.45–2.70 meq per g as calculated from initial gel compositions.

Comparisons between the expanding properties of these materials showed that the beidellites expanded less readily than the montmorillonites upon solvation with glycerol vapor and tended to give single interlayer complexes. These differences were most pronounced when Mg was the saturating ion, although they were also apparent with Ca.

The K-saturated montmorillonites gave greater expansion than the beidellites on hydration at 100 per cent r. h. The montmorillonites also exhibited greater rehydration than the beidellites after heating at 400°–500°C.

Beidellite exhibits properties of expansion that are intermediate between those of montmorillonite and vermiculite. The results of this investigation are consistent with the existence of somewhat stronger ionic attractions in the beidellites than in the montmorillonites. The source of the charge thus has an effect upon the properties of expansion.

A sodium-saturated Texas bentonite was fractionated into six particle-size ranges. The expansion of each fraction was observed by X-ray diffraction under controlled humidity conditions ranging from 0–100 per cent r.h. At all humidities, the fine fractions expanded more than the coarse. A single size fraction expands through a series 10 Å, 12.4 Å and 15.5 Å. The expansion from 10–12.4 Å occurs at progressively lower relative humidities for smaller size fractions in the same way as the expansion from 12.4–15.5 Å. The gradation from high resistance to expansion in the coarse fractions to low resistance to expansion in the fine fractions appears to be continuous.

Interlayer cation density, which controls the resistance to expansion, is dependent upon the structural charge density on the silicate layer. The continuous variation of resistance to expansion that is a function of particle size is interpreted to indicate a corresponding continuous variation in structural charge with particle size.

The expansion of montmorillonite at various relative pressures of ammonia up to saturation was shown to be markedly affected by the kind of ion present on the exchange complex. A 12.3 Å spacing was obtained with Li-, Na-, K-, Mg- and Ca-montmorillonites at low relative pressures. This spacing was maintained up to saturation for the K and Mg systems and apparently represents one layer of ammonia in the interlamellar spaces. For Li-, Na- and Ca-clay, two layers of ammonia were formed at higher relative pressures. In the cases of Cu- and Co-montmorillonite, rational spacings at 12.8 and 15.1 Å respectively were obtained at low relative pressures and were maintained to saturation. The results are interpreted in terms of co-ordination complex formation.

Electrical conductance studies showed exponential increases in conductivity as the relative pressure of ammonia increased. The apparent activation energies showed decreases with increase in relative pressure for all systems except Co- and Mg-clay, which remained rather constant over the pressure range studied. The results indicated that both expansion and electrical conductivity were influenced by the formation of the ammonium ion and by the co-ordinating power of the exchangeable cation.

Muscovite samples were K-depleted with NaCl-NaTPB solutions. By varying the extraction period, samples that varied in total K from 219 to 15 meq per 100 g. were prepared. This treatment did not change the layer charge in the mineral, but it did increase the basal spacing of the mineral from 10 to 12.3 Å.

The changes in basal spacing that occurred when increasing amounts of interlayer K were replaced by Na and the effects of subsequent treatments for the removal of KTPB were determined with two size fractions. In each case, the interlayer K at the periphery of the muscovite particles was replaced by Na, and a fringe of K-depleted mineral with a basal spacing of 12.3 Å developed. In a <50µ sample the particles retained an inner core of 10 Å mineral as the weathered fringe increased. There was no evidence of interstratification. With this sample there was considerable Na trapping when the KTPB was removed with NH4Cl-acetone-water solutions. On the other hand, with a <2µ sample, mixed-layer structures developed as K was removed, and less Na trapping occurred. Na trapping was reduced by removing the KTPB with boiling NH4Cl.

Undried Na-degraded muscovite with only 15 meq K per 100 g did not expand beyond 12.3 Å. The NH4-degraded muscovite samples had a basal spacing of 10.4 Å.

The frequencies, relative intensities and pleochroism of OH stretching bands in some micas and chlorites were studied, and the following conclusions are drawn

Trioctahedral 2:1 and 2:1:1 phyllosilicates, where all octahedral positions are filled with Mg, will have an absorption band about 3700 cm−1. These OH-groups have their axes normal to the mineral’s cleavage.

Substitution of Fe2+ for Mg results in a band about 3665 cm−1. Intensity ratios indicate that two Fe2+ most commonly substitute for two Mg in the same polyhedral group (pyramid).

Decreasing octahedral occupancy in the phlogopite-lepidomelane group, with increasing substitution of R3+ for Mg, causes bands with lower frequencies. These are not very sensitive to the incident angle for the infrared beam. In Li-micas where less than 50 per cent of the octahedral positions are filled with Li, and the rest of the positions are filled with Al and Fe2+, it seems probable that OH-groups residing in a pyramid with two Li and one Al cause a band about 3580 cm−1. Combinations of two Al and one Li or one Li, one Al and one Fe2+ cause absorption about 3480 cm−1.

Dioctahedral 2:1 phyllosilicates having Al or Fe3+ in octahedral positions cause absorption about 3620 cm−1.

Investigations are underway concerning inter-relations of X-ray analyses of chemically treated samples, infrared absorption and DTA for a group of 28 montmorillonite samples. Results to date generally support Greene-Kelly’s LiCl test for evaluation of the proportion of tetrahedral and octahedral layer charge in montmorillonite. Decrease in expandable characteristics of the montmorillonites owing to K+-saturation is much more closely related to the total layer charge than to charge distribution in tetrahedral or octahedral layers.

As the tetrahedral charge increases and the octahedral charge decreases, the infrared absorption band at 11.95µ is replaced by one at 11.35–11.5µ, and the absorption band at 9.2µ decreases or disappears.

Most DTA dehydroxylation peak temperatures are related to charge distribution as follows:

The nature of pyridine adsorption by acid- and base-saturated montmorillonite from Otay, California (A.P.I, no. H-24) was studied by infrared adsorption. The spectrum for chemisorbed pyridine was found to be distinctively different from physically adsorbed pyridine; the latter gave a spectrum very similar to liquid pyridine. Upon exposure of the clay films to pyridine at low vapor pressures, only the pyridinium ion spectrum was obtained; but when the vapor pressure was increased, both chemisorbed and physically adsorbed pyridine was detected. The intensity of the spectrum for physically adsorbed molecules could be reduced by a vacuum, whereas the intensities of the peaks for the chemisorbed pyridinium ions were unaffected.

Interlayer water was replaced by adsorption of pyridine. Water, in turn, replaced physically adsorbed pyridine from the clay surface but was unable to replace chemisorbed pyridine.

Expansion of kaolinite with potassium acetate reduces the intensity of the 3695 cm−1 absorption band and causes the appearance of an additional one at 3600 cm−1. Subsequent deuteration with D2O vapor shifts the 3695 cm−1 and 3600 cm−1 absorption bands to 2725 cm−1 and 2650 cm−1 respectively. The hydroxyls absorbing at 3620 cm−1 do not interact with the acetate anion and do not readily exchange with D2O vapor.

Deuteration of expanded hydrazine-kaolinite complex at room temperature shifts the infrared stretching frequencies 3695, 3670, 3650 and 3620 cm−1 to 2725, 2710, 2698 and 2675 cm−1 respectively. The OH-OD exchanges for the hydroxyls absorbing at 3695, 3670 and 3650 cm−1 are 67, 60 and 62 per cent respectively, and for the 3620 cm−1 only 22 per cent. The 3695, 3670 and 3650 cm−1 absorption bands are correlated predominantly to inner-surface hydroxyls and the 3620 cm−1 to inner hydroxyls located below the holes in the silica tetrahedral layer.

The ν(OH) absorption bands at 3695, 3670 and 3650 cm−1, and the ν(OD) at 2725, 2698, and 2675 cm−1 are pleochroic, whereas the 3620 cm−1 absorption band is non-pleochroic. The direction of the dipole moment change of OH groups absorbing at 3695 cm−1 and 3670 cm−1, and of OD groups absorbing at 2725 cm−1 and 2698 cm−1, is nearly at right angles to the basal plane (001); for the 3650 cm−1 and 2675 cm−1 bands, the angle is large, but less than 90°. The inner hydroxyls absorbing at 3620 cm−1 have their dipole moment change inclined at about 15° to the “ab” cleavage plane.

A comparison of the ν(OH) and ν(OD) absorption intensities for the film rotated 45° and normal to the infrared incident radiation shows that the pleochroism is more intense in the ν(OD) region than in the ν(OH) region. This indicates that partial deuteration perturbs the direction of dipole moment change of the “hydrogen bond” with respect to the initial orientation prior to deuteration; the net result is the formation of larger angles between the direction of the OD dipole moment changes and the basal plane (001) than existed for OH prior to deuteration.

Exploratory experiments on irradiation of mixtures of polyethylene and organo-montmorillonites with gamma rays indicated that the polyethylene became crosslinked to the organoclays. Subsequently, 264 samples of 1:1 clay-polyethylene compositions were irradiated with 2 MeV electrons and tested for tensile strength and weight fraction extractable in toluene at 110°C. From statistical analysis of the variables of clay lattice, clay surface, polyethylene type and radiation dose it is concluded that: (1) polyethylene can be bonded directly to the clay surface by ionizing radiation, and (2) at least one organoclay bearing a “polyethylenelike” chain, polyvinyl alcohol, is more effective than wholly inorganic clay for the purpose of radiatively linking polyethylene to clay surface. The data below show the markedly increased solvent resistance of polyethylene radiatively bonded to clay over that of equivalently irradiated pure polyethylene.

Results of the study of the kaolinite-sodium hydroxide interaction showed extensive dissolution of the kaolinite structure at particle edges, with some production of “silicate relicts” as a secondary effect. The method of specimen preparation for electron micro-scopy precluded electron-diffraction study of soluble reaction products. The study of the kaolinite-calcium hydroxide interaction revealed a similar attack on particle edges and formation of “silicate relics” and, in addition, formation of an insoluble reaction product that was tentatively identified by electron diffraction as prehnite, Ca2Al2Si3O10(OH)2. Both studies showed an ephemeral phase that, by selected area elec- tron diffraction, appeared to be a layer lattice silicate in (001) orientation. In the calcium hydroxide-treated kaolinite, formation of the reaction product was followed from nucleation along particle edges, after 24 hr, to growth of particles about 0.5 to 1 μ in size after 15 days. The combination of electron microscopy of surface replicas to detect changes in morphology with selected area electron diffraction of parallel pseudoreplicas for identification shows promise as a tool for study of the early stages in mineral- chemical interaction.

The exclusion of anions relative to tritiated water from the negatively charged interlamellar spacing of hydrobiotites that have a limited degree of lattice expansion permits measurement of the interlamellar volume. Measurement of the changes in the interlamellar volume and in the exchange capacity, when hydrobiotite lattices are collapsed by potassium treatment or expanded by treatment with sodium tetraphenyl boron, provides a means for estimating the distribution of the surface charge densities of these interstratified minerals.

In the hydrobiotite sample studied (Zonolite from Traveler’s Rest, South Carolina), the biotite fraction was estimated to have a surface charge density of 2.0 ± 0.14 × 10−7 meq cm−2 and the vermiculite fraction, 1.5 ± 0.51 × 10−7 meq cm−2. The charge density, and its relative distribution, was reflected in the tendency of these mineral fractions to collapse upon potassium treatment and to entrap trace concentrations of cesium in the interlayer spacing.

The objectives of this investigation were to measure the effects of moderate heat treatments, below the dehydroxylation temperature, on physical and chemical properties of a calcium-montmorillonite clay. Previous workers have noted the reductions in cation exchange capacity and swelling property after heating in the range 200°–400°C, and have suggested several possible explanations, such as hysteresis effect, increased interlayer attractions owing to removal of interlayer water, or changes in the disposition of interlayer or layer-surface ions.

The liquid limits of Ca-montmorillonite steadily decreased with increased treatment temperature, until they reached a constant level at 450° to 500°C. The plastic limit decreased slightly up to 350°C. Above this temperature samples could no longer be rolled into threads. The gradual change is in contrast with sudden major changes noted for weight loss (maximum rates of change at 100°C and 500°C), glycol retention surface area (maximum change at 520°C), and changes of peak intensity and breadth character after glycolation of the first diffraction maximum (530°C). Other properties showing more gradual reductions with heat treatment were: cation exchange capacity by NH4Ac method, d001 intensity (21 Å spacing) after storing at 100 per cent r.h. one month and re-wetting with water, and amount of calcium exchangeable without water soaking. Previous water soaking allowed much greater release of fixed Ca++ up to 450°C. Similar results were obtained with cation exchange capacities when samples were treated with In CaCl2 solution.

The 21.0 Å peak intensity curve showed close similarity to the liquid limit and plastic index curves in the low temperature range, and an explanation is suggested.

The effects of heat on the behavior of sodium and calcium bentonite in bonding foundry sands were studied.

Tests such as differential thermal analysis, viscosity, and melting or vitrification temperatures do not indicate the suitability of a bentonite for foundry use.

The deterioration temperatures of sodium and calcium bentonite are 1180° and 600°F respectively.

Reactivities of bentonites with cereal as affected by temperature and mulling were also studied. Flowability of sands, defined as mold densification in this paper, is better with heated calcium than sodium bentonite.

Heating bentonites alone does not simulate the conditions encountered in heating bentonite-sand mixtures in a mold.

Two kaolins were given five successive processing steps designed to remove a number of impurities: (1) removal of nonclay and soluble fractions; (2) removal of organic material; (3) removal of iron oxides; (4) removal of allophane; and (5) removal of three- layer lattice minerals. After each of these steps, consistency curves were obtained for each of the clays at both 6.5 and 9.0 pH on slurries containing up to 50 weight % solids.

Photomicrographs and electron micrographs showed that Ka-7 from Bath, South Carolina, contained generally smaller particles than Ka-2 from Macon, Georgia. Lowering the pH from 9.0 to 6.5 in Ka-2 produced no change in plastic viscosity at the lower concentrations and a decrease of about 30 per cent at higher concentrations after processing step 1. A comparison of the same evaluations of step-1-treated Ka-7 showed that the decrease in pH reduced the plastic viscosity by 50 per cent at 5 per cent clay concentration and increased the plastic viscosity by 100 per cent at 35 per cent clay concentration. After complete processing, lowering the pH from 9.0 to 6.5 increased the plastic viscosity at all concentrations for both kaolins by 5–20 per cent. At a clay concentration of 30 per cent, complete processing decreased the plastic viscosity of Ka-2 from 0.4 to 0.02 poise and of Ka-7 from 0.4 to 0.06 poise. Ka-7 showed progressive viscosity changes due to the processing, while Ka-2 showed only a marked change in viscosity for processing step 5. The rheological changes of Ka-2 and Ka-7 with processing are supported by the X-ray diffraction data at the various processing stages. X-ray diffraction data confirm that processing step 5 effectively removed a three-layer lattice expandable mineral from Ka-2, while processing step 3 effectively removed a different three-layer expandable mineral from Ka-7.

A series of experiments was set up, the purpose of which was to determine factors affecting long-term settling characteristics of some suspensions of clay-sized materials. Suspensions of approximately 8 g per 1. were prepared using Georgia kaolin, Mississippi bentonite, Wyoming bentonite, Iowa illite and finely ground quartz. In addition, 2 g and 16 g per 1. suspensions of kaolinite were prepared. All suspensions were observed for 2-week periods at constant temperatures of 19°C; kaolinite suspensions were also observed at 2°, 10°, 25° and 31°C.

After settling 3–4 days at 19°C, layering formed in all clay suspensions, and all the tops of the layers contained higher concentrations of suspensoid than bottoms. Tops and bottoms of the layers were parallel to liquid top surfaces.

At 2°C, no stratification developed in kaolinite suspensions. At 10°C, layering formed only in 2 g per 1. suspensions after 3 months of settling. At higher temperatures, layering formed after 3–4 days of undisturbed settling. Two grams per liter kaolinite suspensions formed five distinct layers; 8 g per 1., nine layers; and 16 g per 1. suspensions formed a many-layered, banded structure after approximately 2 weeks.

If layered suspensions were cooled, the layers were not disturbed. However, layering was very responsive to heating: they expanded and tilted downward away from the heat source. With prolonged heating, layering disappeared.

These experiments show that temperature and concentration markedly affect long-term settling characteristics of fine-grained suspensions. Layering forms most readily in water with relatively low viscosity and density. If such layered suspensions were formed in nature, such features as graded bedding, lamination, or varving probably would result after complete settling.

A fine-grained kaolinite-illite clay, in which planar particle orientation resulted from filtration or plastic deformation, swelled differentially when liquid water was imbibed by test pieces already in the plastic state. Swelling was greater perpendicular than parallel to the plane in which the clay platelets tended to be arranged. Filtration, in contrast to deformation, gave test pieces that swelled more rapidly and extensively. The slipbands, present only in plastically deformed test pieces, restricted the swelling. They were arranged diagonally to the plane in which the clay platelets tended to lie. The plastically deformed test pieces showed enhanced expansion perpendicular to this plane when they were made successively thinner by compression under a cylindrical roller. This was attributed to improved planar orientation of the clay platelets.

Quantitative X-ray diffraction analysis of minerals can most readily be performed on dry powdered samples. Current methods for determination of expandable clays require either wet oriented samples treated with glycerine or glycol or dry, randomly oriented samples treated with amines. Trimethylamine has been used to expand the montmorillonite structure to a 13.3 Å basal spacing. Unfortunately, this spacing is subject to considerable peak interferences.

We therefore undertook a review of amine-montmorillonite complexes for more suitable reagents. The formate and acetate salts of tri-n-hexylamine and of dimethyl-benzylamine meet our requirements. The peak positions are the same as for glycerine complexes, while the physical properties are the same as for trimethylamine.

Detailed studies of tri-n-hexylammoniura ion absorption suggest single layers of ammonium ions lying parallel to the crystal basal plane, double layers in similar orientation, and a single layer with molecules on end. The most useful orientation is the latter, which occurs in the presence of an excess of the ammonium salt and gives a montmorillonite spacing 18 Å in dry powdered samples.

Linear calibration curves with a standard deviation of about 5 per cent of the amount measured can be prepared using the conventional internal standard techniques. Minimum detectable amount of montmorillonite with rate-meter techniques is 0.4 per cent and 0.2 per cent with scalar techniques.

Several thousand analyses using tri-n-hexylammonium ion have demonstrated its effectiveness for expanding a wide variety of sediments.

Thermodynamic calculations of ΔG for all possible transformations of metakaolin at the temperature of the first DTA exothermic peak indicates that the most stable transformation is the one that yields mullite rather than γ-alumina.

The energy of crystallization of γ-alumina is quite small—36,513 cal per mol, compared with the energy of crystallization of mullite—336,180 cal per mol. Furthermore, the crystallization of γ-alumina is very slow, and the crystal growth never produces crystallite sizes much larger than the lower end of the colloidal region. Hence, the gradual release of the small amounts of energy liberated during the crystallization of γ-alumina would be extremely difficult to detect by DTA methods.