The engineering problems of clay soils are closely connected with the “bulk properties,” i.e. shear strength and deformation properties as function of stress conditions and time. These properties are based upon factors in the atomic range and can never be understood without an understanding of the fundamental processes.

The bulk properties of virgin sediments differ from those of chemically or mechanically altered clays (called hypo-metamorphic).

The difference in view generally held by soil mechanics engineers and ceramic research people concerning soil-water systems as built up of minerals in contact with or separated by water, may only be apparent. The water at the mineral surfaces appears to have properties of a solid substance but not the atomic arrangement of ordinary ice. Within a small area, as at the points “of contact,” the H’s and O’s may take the arrangement of a perfect solid. The movement of a corner of one mineral along the plane of another can only take place if flow or creep in the condensed water is possible. In perfect crystals we have only elastic deformation or yielding. The secondary consolidation of clays and the permeability at low hydraulic gradients are discussed as functions of the physics of the water. The influence of temperature and electrolytes upon the properties is suggested as fundamentally important. A plea for further research in the physics and chemistry of surface processes on the various clay minerals is given.

Current knowledge of clay mineralogy and changing concepts of clay behavior have suggested a re-examination of the experimental data concerning adsorbed water. Data published between 1935 and 1959 have been studied and evaluated. In some instances this requires a reinterpretation of data that may lead to inferences or conclusions not intended by the original author. The intention of the re-evalution of data is to clarify understanding of the nature of adsorbed water and to suggest fruitful avenues for future research. While all literature on the subject has not been included, a concerted effort has been made to include representative data from all viewpoints and experimental methods. Data reviewed come under the following headings: X-ray and electron diffraction, density, dielectric and magnetic, thermodynamic, diffusion and fluid flow, freezing, and rigid water films.

Utilizing present knowledge of crystal and surface chemistry of clay leads to the following conclusions: (a) Positions of the oxygen atoms of the adsorbed water molecules have been established by X-ray diffraction of vermiculite. These positions preclude both the ice structure and the Hendricks-Jefferson net structure even after modification to accommodate the exchangeable ions, (b) Density of water sorbed on Na montmorillonite has a minimum value of about 0.97 g/cm3 at a water content of 0.7 g H2O/g clay (approximately the plastic limit). For water contents less than 0.7 the density rapidly rises to about 1.4, and for water contents greater than 0.7 the density gradually rises until at about 6.5 g H2O/g clay the density of the adsorbed water equals that of normal liquid water. (c) The differential entropy of water adsorbed on kaolinite has a minimum value approximately that of ice; however, (1) this minimum occurs at about 0.7 of a monolayer, and (2) the integral entropy is greater than that for normal liquid water up to at least two molecular layers. The apparently contradictory entropy of sorbed water on montmorillonitic clay has not been resolved but is believed to be at least partially associated with clay swelling. (d) Diffusion and fluid flow phenomena are shown to be extremely sensitive to clay fabric; therefore, it is the writer's opinion that diffusion and fluid flow data on loosely compacted clay are of little help in ascertaining the structure of the adsorbed water phase. (e) Adsorbed water is easily supercooled and an appreciable fraction of the adsorbed water remains unfrozen after ice has once formed.

The two major hypotheses indicate that the nature of the adsorbed water is. (1) a solidlike substance, o. (2) a two-dimensional fluid. In the writer’s opinion the only data that cannot be adequately explained by both of the hypotheses are the integral entropy data on kaolinite. The integral entropy data favor the two-dimensional fluid hypothesis; however, the paucity of data requires that this be a very tentative conclusion.

The major difficulty encountered in the re-examination of data for this review was that a rather poorly defined clay surface was employed by various investigators. If progress is to be made in unraveling the water-clay complex, it is deemed absolutely essential that experiments be carried out on very carefully defined and controlled clay surfaces. The nature of adsorbed water as interpreted from physico-chemical data on water-clay systems is no better than the purity of the clay surface regardless of the accuracy and precision of the measurements.

The partial molar free energies, entropies and heat contents of the water in suspensions of Li-, Na-, and K-clay were measured relative to pure water by means of a tensiometer. This method of obtaining these thermodynamic properties is far more sensitive than the more common method utilizing adsorption isotherms and a form of the Clausius-Clapeyron equation. The partial molar free energy of the water decreased with time after stirring and increased rapidly upon stirring. This was taken as evidence that a gradual increase in the degree of orientation of water molecules associated with the clay particles occurred as the particles fell into the more random edge-to-surface thixotropic structure. The partial molar entropies and heat contents of the water in the suspensions were found to be less than those for pure water. Hence, this evidence also supported the postulate that an ordered water structure exists in the thixotropic suspensions.

An account is given of the soil thermal resistivity problem faced by the electric power industry and its co-operative engineering solution. Data obtained in this activity and collected from other sources are employed to check the validity of various concepts that hare been developed to explain the phenomena observed when soils, composed of solid, liquid and gaseous phases, are placed in a thermal energy field. It is shown that formulas based on simple parallel or series arrangement of the component phases cannot account for the actual thermal transmission properties of moist soils. Nor can any one of the moisture transfer mechanisms, that have been proposed in explanation of the water movement associated with heat conduction in moist soils, by itself account for the observed magnitudes of thermal conductivity. Analysis of pertinent data renders it very probable that both evaporation-condensation and oriented film mechanisms must be involved in the coupled heat and moisture transport. What really takes place is an intricate interaction of a number of different events that can be understood qualitatively on the basis of the best available knowledge of soil-water interaction and its temperature dependence. Much work, however, remains to be done before a quantitative understanding can even be approached. The nature and magnitude of the remaining scientific problem can be appraised from the following pertinent statement by Szent-György. (1958) “The complexity of the situation with its very subtle equilibria makes conditions most colorful and begins to resemble the subtle complexity which characterizes life.”

Illite and montmorillonite have been compressed into cores under a pressure of 1200 atm. The structural status of these clays has been examined by means of detailed nitrogen adsorption-desorption isotherms and by means of electron micrographs of fracture surfaces of the clay cores. The subsequent swelling behavior of sodium and calcium clay cores has been analyzed in terms of the model derived from these measurements and the specific surface area of the clay. A generalized theory for the condensation of plate-shaped particles into domains has been proposed.

Measurements of the effect of electrolyte concentration and hydrostatic suction on the swelling of sodium- and calcium-saturated clay cores indicate that diffuse double layer theory may play a significant part in determining the swelling of monovalent clay systems but not divalent clay systems.

The magnitude of the film thicknesses apparently developed in solutions as concentrated as molar and 4 M, where the condition of ideal solutions required by diffuse double layer theory is certainly not fulfilled, seems to indicate that solution uptake is increased by a relaxation of structural strains within the clay matrix on lubrication by solutions. These strains may arise from the distortion of crystals in packing during the drying process. Such relaxations which appear to occur between p F 4 and 3 could initiate the development of a gel structure.

The factors. (1) possible violations of Darcy’s law. (2) electrokinetic coupling. (3) high viscosity. (4) tortuous flow paths, an. (5) unequal pore sizes have been suggested as possible explanations for the differences between hydraulic flow rates in liquid-saturated clays and sands. The effects of these factors on hydraulic flow rates through saturated clays were investigated.

Hydraulic flow rates, electrical conductivities, and streaming potentials were measured on natural, sodium, and calcium samples of kaolinite, illite and Boston blue clay. Data were taken after increments of one-dimensional consolidation and rebound over the pressure range from one-sixteenth to 256 atm.

The influences of electrokinetic coupling on the hydraulic flow rates were calculated from irreversible thermodynamic relationships together with the hydraulic and electrical data. The other factors were studied by examining the extent to which each factor explains the discrepancies between measured flow rates and those predicted from Darcy’s law and the Kozeny-Carman equation.

The results show that. (1) the possible violations of Darcy’s law and electrokinetic coupling are insignificant. (2) high viscosity and/or tortuous flow paths fail completely to account for the discrepancies between measured and predicted flow rates in clays, an. (3) unequal pore sizes can explain all the discrepancies.

Pore fluid pressures that develop within soil masses as a result of both mechanical and physico-chemical effects influence the magnitude of the intergranular or effective stresses. The intergranular stresses control, in many cases, soil behavior in shear and compression. The physical significance of pore water pressure in a cohesive soil is examined in terms of several components which combine to give the total pressure. An analysis of fluid pressures at various points within a soil mass based on a condition of no flow at equilibrium shows that changes in any one component of the total pressure from point to point are offset by changes in other components.

The analysis shows that a pore pressure measurement reflects interparticle repulsive pressures and water adsorptive forces as well as purely hydrostatic pressures arising from mechanical effects. The individual components of total pore pressure are not directly measurable except in special systems. Data are presented which indicate that a significant portion of the swell of a compacted soil is attributable to water pressure deficiencies caused by mechanical and capillary components which act in addition to osmotic pressure components.

Component water pressures are related to stresses between particles. It is shown that intergranular pressures are dependent on osmotic and adsorptive components of the total water pressure.

In recent soil mechanics literature attempts have been made to explain the engineering properties of clays in terms of surface chemical theories, particularly the Gouy—Chapman theory. Experimental confirmation of these theoretical predictions have been restricted almost exclusively to studies performed on fine grain size fractions of montmorillonite. It seemed desirable to determine experimentally the effect of physico-chemical variables on the engineering properties of a more typical engineering material. The < 2 μ fraction of Fithian illite was selected.

A study was made of the effect of variations in the pore water electrolyte concentration on the engineering properties of homoionic Ca- and Mg-illite. Pore water electrolyte concentrations ranged from 1 N down to about 10-4 N. Engineering tests included Atterberg limits, one-dimensional consolidation tests, and effective stress triaxial tests.

The electrolyte concentration was found to have only a small effect on the Atterberg limits, the maximum limits occurring at electrolyte concentrations between 0.01 N and 0.1 N. For virgin consolidation the soil seemed strongest at about the same range in electrolyte concentrations but the effect was not large. Electrolyte concentration had almost no effect on the position of the rebound (swelling) curves. The most important effect of electrolyte concentration seemed to be its effect on the geometric arrangement of the particles for samples that were sedimented from dilute suspension. The geometric arrangement of particles seems to be a more significant variable than the osmotic repulsion between particles.

There still appears to be merit in the old mechanical approach where the compression characteristics are explained by elastic bending (reversible), slippage (partially reversible), and rupture (irreversible) of particles. Double layer phenomena seem to exert a much smaller influence on the engineering properites of nonexpanding lattice clay minerals than is commonly inferred in the literature.

Shear strength of saturated, remolded clay mineral samples was measured in an attempt to relate strength to forces acting between clay particles. The fine fractions of the clays, purified when necessary, were consolidated to different void ratios and subjected to quick, translatory shear in a direct shear device. Relative values for shear strength were obtained in this way. Interparticle forces in the clays were altered and the effect on shear strength was observed.

For a high-swelling, Na-saturated montmorillonite, strength at constant void ratio decreased with increasing salt concentration. To explain the decrease in strength with decrease in force of repulsion between particles a model is suggested where strength results from the force of repulsion resisting displacement of particles in the shear plane. Addition of small amounts of dispersing and of soil-conditioning chemicals had little effect on shear strength.

Calcium montmorillonite had a lower strength than sodium montmorillonite when compared at the same void ratios. This is related to the lower surface area of interaction and lower repulsion found with divalent ions.

Water content at the liquid limit showed the same dependence upon interparticle forces in the montmorillonite clays as did shear strength.

In kaolinite, interparticle forces of attraction result in a structure or particle arrangement which has the major influence on shear strength. A clay at low pH which was flocculated in a network, edge-to-face structure owing to presence of positive edge-charge had higher strength than a dispersed clay, but a clay flocculated by electrolyte which produced a lamellar, face-to-face structure had a lower strength.

The concept of activation energy for ion movement is discussed. It is shown that the activation energy for exchangeable ion movement is directly correlated with the specific volume of the adsorbed water, with the resistance to freezing of this water and with the activation energy for deuterium oxide diffusion in this water. Electrical interaction between ion and clay appears to have little effect on the activation energy. Consequently, it is concluded that the most important factor governing exchangeable ion movement in the pores of a clay-water system is the structure of the adsorbed water.

A special integral form of the solution of Fick's law was applied to study the self-diffusion of Na, Cs, and Ca ions in Wyoming bentonite plugs by means of a radioactive tracer technique. A simple procedure for measuring diffusion coefficients was developed by depositing a thin radioactive tracer film on the surface of clay mineral plugs at controlled temperatures. Thin sections of the clay plugs were removed and radioactivity measurements made on the portion of the plugs remaining. The self-diffusion coefficients of the ions were evaluated by a graphical method from the solution of Fick's law at special boundary conditions. Under oven-dry conditions at 105°C diffusion of Na ion in a bentonite film was observed, while in the same experiment, SO42− ion did not diffuse. In a clay-water system, however, the anion diffused much more rapidly than the cation. Based on the experimental results of observing Na22- and S35-tagged Na2SO4 diffusion, different diffusion paths were proposed using a picture of the clay matrix as a structure with “channels.” Diffusion coefficients of exchangeable cations decreased with increasing clay contents. From the results of activation energy and self-diffusion coefficients of Na and Ca ions at various clay-water concentrations, the relative “hopping distances” of diffusing ions were calculated and discussed. It is suggested that in the self-diffusion of exchangeable Na ion in bentonite the ion could be hopping from one individual clay plate to another, and the Ca ion from one “package” of plates to another “package” of plates.

Experimental evidence of the salt filtering ability of compacted bentonite and clay material is presented. Apparatus has been developed in which samples of semi-dry powdered bentonite and shale are compacted under pressures of 10,000 psi and subjected to flow with sodium chloride solutions under fluid pressures up to 10,000 psi. Porosities of 34–41 percent have been obtained with the bentonite and 24 percent with the shale. Permeabilities of 10−5 to 10−6 millidarcies have been measured. A ratio of input to output solution concentration as high as 8 has been found with the bentonite samples and dilute (0.164 N) NaCl solutions. The ratio falls off with increased input concentration and with increased porosity. The shale sample has given a ratio of 1.7 with an input concentration of 0.129 N NaCl.

Streaming potential differentials of approximately 1 mV per 100 psi have been measured during the flow experiments. The potential differential is highest with the shale material and shows a decrease with increasing salt concentration.

Addition of HCl, H2SO4, and H3PO4 to montmorillonite clay and a montmorillonitic soil had little effect on the liquid limit of these materials over a wide range of concentrations. The liquid limit of a kaolinitic soil clay was reduced markedly by both HCl and H2SO4, whereas H3PO4 decreased the liquid limit value at lower concentrations and increased it at higher ones. These results are discussed in light of the charge variability and of anion retention by the kaolinitic soil.

The flocculation behavior of clay specimens from altered basalt, altered granodiorite, and a transported (?) deposit in Latah County, Idaho, was investigated in order to determine if flocculation behavior would be an aid in characterizing clay samples in a way that could be related to the origin and history of the clays. The clay minerals are of the kaolinite group.

As flocculants, various chlorides differ one hundred-fold and sodium salts twenty-fold in the concentrations necessary to flocculate a clay to a fixed settling rate. The concentrations of individual chlorides necessary to flocculate different clays vary three-fold and the concentrations of individual sodium salts vary ten-fold. In all cases the higher the valence of the cation the lower is the concentration required for flocculation; the relative effectiveness, however, varies with the clay. The relative effectiveness of the anions as flocculants varies greatly with the clay and no definite order of effectiveness is apparent.

Halloysite requires less electrolyte for flocculation and has a larger settling volume than kaolinite. Well crystallized kaolinite has a smaller settling volume than poorly crystallized kaolinite.

Wide differences of flocculation behavior may provide a valuable, rapid method of classifying the clays of a district.

An investigation of the behavior of clay systems in drilling fluids was begun in 1958 to determine the cause of gelation in lime-treated muds with increased temperatures incurred during drilling operations. Mud systems were evaluated by physical property measurements, cation exchange data, X-ray diffraction analysis, and electron microscopy before and after hydrothermal treatment. Selected samples were rerun after two years' storage at laboratory temperatures.

High-temperature gelation of lime-caustic muds is caused by the synthesis of calcium silicate hydrates, hereafter referred to as CSH, in general and tobermorite (XCaO, SiO2, H2O) specifically. These materials are considered the principal cementing agents of Portland cement. Spheroidal poorly crystallized tobermorite was produced from interlayered illite-montmorillonite muds. Laths and needles of crystalline tobermorite (11.3–11.6 Å type) were synthesized from the pure montmorillonite systems. Zeolite minerals were synthesized concurrently and were the dominant alteration product when Ca(OH)2 was not added to the system.

Synthesized aged (2 years) mud systems were studied after 8 weeks of dialysis with distilled water. Corundum was observed in samples where tobermorite formed, and was the dominant insoluble residue after HCl treatment. Much sharper diffraction maxima were obtained of the CSH compounds after dialysis or glycerine washing.

“Al-chlorite,” “Al-vermiculite,” and “Al-micas” were synthesized when treatments included aluminum ions. The products formed were of varying stabilities after washing with distilled water and HCl. Glycerol expanded the montmorillonite which had not reacted, but no expansion from the 10 Å or 14 Å d(001) spacing of the new product occurrred.

The flocculation and subsequent subsidence of colloidal suspensions are markedly increased by the addition of flocculating agents; for example various starches and recently developed synthetic polymers.

The rate of flocculation depends not only upon the extent of adsorption of the agent (flocculant) but also upon the order of mixing and the character and extent of the subsequent agitation which determines the character of the floc formed. Electrolytes flocculate by reducing the electrostatic repulsions between charged particles, but long chain polymers produce flocs by a bridging mechanism which overcomes electrostatic repulsions.

Quantitative relationships are developed between the optimum concentration of flocculant and the rates of flocculation, subsidence, and particularly rate of filtration through the filter cake.

Desilication and removal of bases from the basalitic rocks of the Hawaiian Islands produces large amounts of halloysite and gibbsite, together with variable quantities of allophane and nontronite, iron and titanium oxides, and amorphous mineral and gel material.

The nature and relative abundance of intermediate and end products of weathering depend primarily upon the amount of rainfall, angle of slope, and texture of the rock. The formation of clay minerals is the common step in the transformation of primary silicates to oxides and hydroxides, but the amount and relative importance of the clay “stage” depend on the intensity of the weathering process.

Halloysite forms from plagioclase usually by alteration of first the core and subsequently the rim of the feldspar laths. No kaolinite was observed, its absence being attributed to the absence of fine-grained mica as an intermediate weathering product.

Gibbsite is produced by (1) removal of silica from halloysite, (2) dehydration of Al-gel, and (3) precipitation from solution. Although it is possible that the mineral may form directly from feldspar, halloysite is the common crystalline intermediate on both the megascopie and microscopic scales.

An amorphous transition state, probably ranging in composition from allophane to Al-gel, exists as part of the change from halloysite to gibbsite as evidenced by electron microscope and diffraction work on pseudomorphs after halloysite tubes found in certain samples studied in more detail than others.

Volcanic glass is the apparent source of most Al-Fe-Si gel material which upon dehydration becomes allophane, cliachite, or gibbsite depending on the kind and relative proportion of the cations present.

X-ray analyses of samples from thirteen soil profiles that were formed on glacial till, loess, and Mississippian limestones of Indiana indicate that (1) in some soil profiles chlorite of the parent material is changed completely to montmorillonite with intermediate stages of random mixed layers of chlorite-vermiculite-montmorillonite; (2) some of the illite of the parent material produces montmorillonite through random mixed layering of illite-montmorillonite, but the illite is not entirely altered in any profile; and (3) kaolinite may be produced as a weathering product, or it may remain unchanged from the parent material.