This paper describes the relationship between the micro-structure and hydraulic conductivity of simulated sand-bentonite mixtures (SSBMs) prepared with powdered and granular bentonite. Glass beads were used to simulate sand grains because of their superior optical properties. The micro-structure of SSBMs was observed using optical micrography and scanning electron microscopy. For mixtures prepared with powdered bentonite, the indications are that bentonite coats the particles. As the bentonite content increases, the thickness of bentonite coating increases and reduces the area available for flow. For mixtures containing granular bentonite, the dry bentonite granules occupy the space between the particles and then swell to fill the void space. As the bentonite content increases, the number of granules increases, leading to more void spaces being filled with bentonite. At higher bentonite content (>8%), flow paths devoid of bentonite are unlikely, and the hydraulic conductivity appears to be controlled by the hydraulic conductivity of bentonite. The changes in micro-structure that were observed are consistent with the decrease in hydraulic conductivity that occurs with increasing bentonite content. However, the relationship between hydraulic conductivity and bentonite content differs depending on whether a mixture contains powdered or granular bentonite.

Short- and long-term exposure to inorganic solutions can cause significant degradation of the hydraulic properties of bentonite clay used in geosynthetic clay liners (GCLs). In particular, the increase in hydraulic conductivity due to cation exchange when Na-montmorillonite is subjected to leachates rich in Ca and Mg has caused problems in incinerator ash landfill liners located in wet environments, where large quantities of leachates are generated. Experimental results are presented to evaluate the immediate change in hydraulic conductivity of seven types of GCL clays upon permeation with leachate generated from three ash landfills. The composition of the ash, which is a by-product of the incineration of municipal solid waste (MSW), in turn influences the composition of the resulting leachate. Falling head permeability tests were performed on flexible-wall permeameter specimens, with back-pressure saturation. Chemical analysis shows that the three leachate products contain high, medium, and low concentration Ca and Mg cations. The clay component of GCL materials tested in this study consists of regular and polymer-treated bentonite. Polymer treatment is believed to render the clay non-reactive to many organic and inorganic chemicals. The results of this study indicate that: (1) polymer treatment is generally more beneficial if the clay is first saturated with water and not directly with the leachate; (2) high swell potential of the bentonite is more advantageous than polymer treatment, especially when low hydraulic conductivity is required in the short term and if the clay is pre-hydrated. Experiment setup and special specimen preparation procedures are also discussed.

An experimental program was conducted to investigate the water-vapor sorption characteristics of smectite and kaolinite mixtures. End-member smectite and kaolinite were slurry-mixed together at mass-controlled ratios corresponding to 0%, 20%, 50%, 70%, 80%, 90% and 100% smectite. Vapor desorption isotherms for the mixtures were measured at 24°C for relative humidity (RH) ranging from ∼95% to 0%.

Results show that the amount of water adsorbed by the clay mixtures at a given RH increases systematically with increasing smectite content. Derivative analysis of the sorption isotherms shows evidence of transitions between the two-, one- and zero-layer hydrate-states for the smectite-rich mixtures. The transitions become less apparent as the smectite content decreases. Monolayer coverage, specific surface area, and heat of adsorption were estimated from the isotherms using BET theory. It is shown that monolayer coverage and specific surface for the clay mixtures can be reasonably approximated by weighted averaging of the end-member clay properties. General methodologies are presented for predicting the sorption behavior (i.e. soil-water characteristics) and effective specific surface area from measurements of the end-member sorption isotherms.

Two independent methods of crystal-size distribution analysis were compared: the Bertaut-Warren-Averbach XRD technique (MudMaster computer program) and high-resolution transmission electron microscopy (HRTEM). These techniques were used to measure thickness distributions of illite crystals (fundamental particles) from sets of illite-smectites from shales and bentonites that had expandabilities ranging from 86%S to 6%S. The illite-smectites were treated with a polymer (polyvinylopyrolidone, PVP) to separate them into fundamental particles for XRD and HRTEM investigations.

A systematic difference between XRD and HRTEM results was observed: XRD (area-weighted distributions) detected a larger fraction of thick (>4 nm) and a smaller fraction of thin crystals as compared to HRTEM (number-weighted distributions). As a result, XRD-determined distributions have larger mean thickness values and larger distribution parameters (α and β2). The measurements performed by the two techniques were verified by modeling XRD patterns of the PVP-illites, using the measured distributions as inputs. The modeling indicated that the XRD-determined distributions are very accurate. Selecting broader thickness distributions in MudMaster further improved the modeling results. The HRTEM measurements underestimate the proportion of coarse particles, in particular in shale samples, and this inaccuracy is attributed to the effect of using number-weighted (rather than area-weighted) distributions and to inaccurate counting statistics for thick crystals.

Volcanic tuff from the Cabo de Gata region in Almería, southeastern Spain, was altered under hydrothermal conditions at different temperatures (60 to 180°C), reaction times (60 to 360 days), and reacting solutions (deionized water and NaCl and KCl solutions with Na/K ratios from 0.01 to 100, and a total salt concentration of 0.1 to 1 M).

X-ray diffraction (XRD) patterns of the reacted samples revealed a very weak, broad peak at ∼13 Å that migrated to 17 Å upon glycolation. Comparison between simulated (NEWMOD) and experimental XRD patterns indicated that the neoformed phase is a random mixed-layer illite-smectite (I-S) with 75% expandable layers. Fourier transform infrared (FTIR) spectroscopy showed that I-S formation was most extensive for high pH (8–9) solutions, corresponding to dilute solutions and, especially, to deionized water. Analytical electron microscopy (AEM) analyses of isolated I-S particles showed that most of them are smectite-rich I-S regardless of the experimental conditions, in agreement with XRD results. The I-S particles had a wide range of octahedral Mg contents. The pH and Na, K, Ca and Mg concentrations in the final solutions suggested cation (including H+) exchange as a major process in the alteration experiments. Analysis of aqueous activity diagrams (log aK/aHvs. log $a_{{\text{SiO}}_{\text{2}}}$) showed that some solution compositions are consistent and some are inconsistent with I-S formation. These results, combined with complementary electron microscopy analyses (de la Fuente et al., 2000a), are interpreted to be due to direct transformation of the glass into I-S in a process controlled by glass composition.

The hydrothermal stabilities of a low-Fe clinochlore and a high-Mg chamosite, in the presence of kaolinite, were investigated recently at T ⩽ 200°C and $P_{\text{v}}=P_{{\text{H}}_{\text{2}}\text{O}}$ (Aja and Small, 1999; Aja and Dyar, 2002). Standard state thermodynamic properties ($S_{298}^0$, ${\Delta }_f{H}_{1,298}^0$ and ${\Delta }_f{G}_{1,298}^0$) have been obtained for the two chlorites whose structural formulae are (${\text{Al}}_{2.33}{\text{Fe}}_{1.00}^{2+}{\text{Fe}}_{0.14}^{3+}{\text{Ca}}_{0.02}{\text{Mn}}_{0.01}{\text{Ni}}_{0.02}{\text{Cr}}_{0.01}{\text{Mg}}_{8.40}{\alpha }_{0.07}$) (Si5.66Al2.34)O20(OH)16 and (${\text{Fe}}_{0.60}^{3+}{\text{Fe}}_{5.43}^{2+}{\text{Mg}}_{2.30}{\text{Al}}_{2.98}{\text{Mn}}_{0.05}{\text{Ca}}_{0.03}{\text{Zn}}_{0.01}{\alpha }_{0.60}$)(Si5.63Al2.37)O20(OH)16. For the low-Fe clinochlore, the respective thermochemical properties are 430 J mol−1 K−1, −8770.64±35.24 kJ mol−1, and −8120.54±32.63 kJ mol−1. ${\Delta }_f{H}_{1,298}^0$, ${\Delta }_f{G}_{1,298}^0$ and ${\text{S}}_{298}^0$, similarly obtained for the Windsor chamosite are −7851.29±23.14 kJ mol−1, −7271.01±21.43 kJ mol−1 and 668±5 J mol−1K−1, respectively. Ideal site-mixing models of chlorite composition, along the chamosite-clinochlore binary, fail to model satisfactorily these chlorite-fluid equilibria only at lower temperatures (T <175°C). The magnitudes of the excess thermodynamic properties calculated for these chlorites, within the ternary clinochlore-daphnite-sudoite system, suggest significant deviations from ideality.

The crystal structure of cronstedtite-2H2 was refined in a hexagonal cell, space group P63, Z = 2, using two acicular crystals from Wheal Maudlin, Cornwall, England, and from Pribram, Czech Republic. The Wheal Maudlin sample has the chemical composition $\left({\text{Fe}}_{2.291}^{2+}{\text{Fe}}_{0.709}^{3+}\right)\left({\text{Si}}_{1.298}{\text{Fe}}_{0.707}^{3+}{\text{Al}}_{0.004}\right){\text{O}}_5{\left(\text{OH}\right)}_4$ and the Příbram sample has the composition $\left({\text{Fe}}_{2.269}^{2+}{\text{Fe}}_{0.731}^{3+}\right)\left({\text{Si}}_{1.271}{\text{Fe}}_{0.724}^{3+}{\text{Al}}_{0.005}\right){\text{O}}_5{\left(\text{OH}\right)}_4$. The results of refinements are as follows: a = 5.500(1), c = 14.163(2) Å, V = 371.08(8) Å3, R = 3.83%, from 381 independent reflections, and a = 5.4927(1), c = 14.1481(2) Å, V = 369.70(4) Å3, R = 4.77%, from 1088 independent reflections for the Wheal Maudlin and Příbram samples, respectively. The best Fovs.Fc agreement was achieved when the structure was interpreted as merohedral twin; several possible twinning laws are discussed. The cronstedtite layer consists of one tetrahedral sheet and one octahedral sheet. There is one octahedral (M1) position, occupied by Fe only, and two tetrahedral (T1, T2) positions in the structure. Refinement of occupancy of tetrahedral sites led to values Si:Fe = 0.45:0.55(1) (Wheal Maudlin) and 0.432:0.568(8) (Příbram) in T1, and Si: Fe = 0.99:0.01(1) (Wheal Maudlin) and 0.888:0.112(7) (Příbram) in 72. Whereas the size of T1 is reasonable (average dT1-O = 1.693 Å (Wheal Maudlin), 1.691 Å (Příbram)), T2 is unusually large: (dT2-O= 1.740 Å (Wheal Maudlin), 1.737 Å (Příbram)) with respect to the small or almost zero Fe content. As an explanation, an alternative structure model comprising a certain amount of vacancies in T2 is presented. The tetrahedral rotation angle α is highly positive (+12.1° and +12.5° for the Wheal Maudlin and Příbram samples, respectively), and the layer belongs to the Franzini type A. Distortion parameters of octahedra and tetrahedra are given for both samples. One hydrogen atom engaged in the hydrogen bond was located in the Wheal Maudlin sample.

The adsorption of metals by clay minerals is a complex process involving different mechanisms, and is controlled by different variables which can interact. The aim of this work was to study the retention mechanisms of Cd on illite. We obtained Cd adsorption isotherms at constant pH, adsorption edges as a function of pH, adsorption isotherms at 5, 25 and 45°C (pH = 7), and response surfaces of the simultaneous effect of pH, initial concentration, ionic strength, and temperature on the retention of Cd on illite. Below pH 6, adsorption of Cd on illite is via ion exchange with H3O+ and Na+ ions which saturate the exchange sites, the exchange with Na+ being the main mechanism between pH 4.5 and 6.0. For pH values >6, the effect of ionic strength on the amount of Cd2+ retained decreased with pH, being negligible at pH 8; the proton stoichiometry was greater than for pH values <6 and an increase in the temperature favored the retention of Cd. These facts are compatible with a more specific process involving hydrolyzed species, in which Cd can associate with illite as an inner sphere complex.

Porous silica products obtained by selective leaching of phlogopite using an acid solution were investigated by XRD, MAS NMR, SEM, TEM, DTA/TG, and N2 and Ar gas adsorptions. The phlogopite powder was leached by a nitric acid solution at various concentrations (0.01–10 M) at 5–150°C for 10 min–480 h. Selective leaching of the phlogopite powder became extensive when the concentration of nitric acid was >1 M. Only SiO2 remained after the treatment, the other components (MgO, Al2O3, K2O and Fe2O3) being selectively leached from the product. At higher leaching temperatures, the leaching rate became faster and the resulting maximum specific surface area of the porous silica product became larger at each leaching temperature. The porous silica products were found by SEM and TEM, to maintain their original platy particle shape even after the selective leaching. The 29Si MAS NMR spectra of the products, however, revealed that the linkage structure of SiO4 tetrahedra converted to a framework type from a layered type in the original phlogopite. The porous silica product with the maximum specific surface area (532 m2/g) was obtained by leaching in a nitric acid solution with concentration of 5 M at 90°C for 15 min. The pore-size distribution of the porous silica product was bimodal with micropores of ∼0.7 nm and mesopores of ∼4 nm. The pore size of the products changed from 0.7 nm to 4 nm and further to 6 nm with increased leaching time. The present results are discussed in relation to those reported for phlogopite by other workers.

The impact of alkaline solutions (pH = 13.2) on the clay mineralogy of the Callovo-Oxfordian formation hosting the French underground laboratory for nuclear waste disposal investigation (Meuse-Haute Marne site) has been studied experimentally. Initially, each of the four samples selected as representative of the mineralogical transition in this Callovo-Oxfordian formation consists of a mixture of three main clay phases: discrete illite, discrete smectite and a randomly interstratified mixed-layered mineral (MLM) containing ∼65% of non-expandable layers. Clay separates were altered in batch reactors at 60°C using high solution:solid ratios. The mineralogy of this clay fraction and solution chemistry were monitored as a function of reaction time. In addition, the interactions between organic matter and clay particles were investigated using scanning transmission X-ray microscopy (STXM).

The clay mineralogy is little affected even though the pH is still high after 1 y reaction time. The only significant mineralogical evolution is the partial dissolution of the discrete smectite component leading to the formation of a new randomly interstratified illite-expandable MLM. Additional mineralogical transformations lead, for one sample, to the dissolution of micro-crystalline quartz and, for another sample, to the crystallization of a tobermorite-like phase. The low reactivity of clay minerals may be attributed to the presence of organic matter in the samples. In their initial state, all outer surfaces of clay particles are indeed covered with organic matter. After 1 y reaction time, STXM studies showed the basal surfaces of clay particles to be devoid of organic matter, but their edges, which are the most reactive sites, were still protected.

Bentonite, biotite, illite, kaolin, vermiculite and zeolite were treated with 0.1, 1.0 and 5.0 mol. dm−3 HCl or NaOH. Suspensions of Na homoionic forms of the initial and the treated minerals were back titrated with 0.1 mol. dm−3 NaOH. From back-titration data variable surface charge, QV, vs. pH dependencies and apparent surface dissociation constant distribution functions were estimated. Variable charge vs. pH curves were scaled against cation exchange capacity at pH 7.2 to obtain actual charge, QA, vs. pH plots. In general both treatments led to an increase of QV, while the QA value increased and decreased depending on the mineral and the treatment. Products of 5 mol dm−3 NaOH action on zeolite and of 1 and 5 mol. dm−3 NaOH on kaolin apparently developed a positive surface charge below pH ∼3.7 and below ∼4.3, respectively.

The heterogeneity of charge-generating surface groups was observed in natural minerals. During acid treatment, the number of weakly-acidic surface functional groups increased while the number of groups of stronger acidic character decreased. The opposite was found for alkaline treatment: the number of surface groups of intermediate acidity increased and that of low acidity decreased.

New experimental data with respect to the solubility of natural and synthesized magadiite at elevated temperatures and in alkaline solutions are presented. The results show that the solubility of magadiite increases according to the expression ln(Kmag) = −8146·T(K)−1 − 5.71 from 20 to 100°C ($(\Delta G_{\text{R}}^0({\text{K}}_{\text{mag}})=19.6\pm 0.3$ and $\Delta H_{\text{R}}^0({\text{K}}_{\text{mag}})=16.2\pm 0.4\text{kcal}{\text{mol}}^{-1}$). The experimental results and the related data from the literature suggest that the formation of magadiite may be favored by a decrease of temperature and pH (pH >9) as well by large amounts of Na+ ions and low ionic strength. These effects are related to the value of $\Delta H_{\text{R}}^0({\text{K}}_{\text{mag}})$, the distribution of dissolved silica species, the stoichiometry of magadiite, and the occurrence of negatively-charged species, respectively.

Lattice dynamic calculations for the sepiolite and palygorskite structures using polarized Raman and FTIR spectra provide a fundamental basis for interpreting spectral features by assigning vibrational modes. The Si-O stretch and O-Si-O bond bending force constants determined for palygorskite are similar to equivalent values calculated previously for other phyllosilicates. The Mg-O bond stretch values, on the other hand, are about half of those determined for the equivalent Al-O and Mg-O bond stretch environments in other phyllosilicates, suggesting that the bonding within the octahedral ribbons in palygorskite and sepiolite is weaker than that in the continuous octahedral sheets in micas. The weaker bonding allows more flexible octahedral environments in palygorskite and sepiolite, giving rise to higher probabilities for cation substitutions and vacancies relative to the micas. Above ∼700 cm−1 in the IR and 750 cm−1 in the Raman spectra, the eigenmodes are dominated by atomic displacements within the silicate sheets. Below 700 cm−1 the eigenmodes become mixed with motions among the Mg octahedra and the silicate sheets; the eigenmodes assigned to the most prominent peaks in the Raman spectra (near 700 cm−1) belong to this group. As mode frequencies decrease, the corresponding eigenmodes evolve from more localized Mg-O stretch, O-Mg-O bend and O-Si-O bend motions to longer-range motions such as silicate sheet deformations caused by silicate tetrahedra rotation and silicate sheet shearing around the Mg-octahedral sheets.