Six kaolin samples from the Lower Tertiary Huber Formation near Wrens, Georgia were analyzed using transmission electron microscopy (TEM), electron diffraction (ED), powder X-ray diffraction (XRD), chemical analysis, and magnetic susceptibility to characterize the Ti-bearing phases. Selected samples were treated with 5 M NaOH to remove kaolinite and concentrate the Ti-bearing phases for additional analysis. TiO2 content in the bulk fraction ranges from 1.2 to 5.4 wt. %. There are at least three Ti-bearing phases, including anatase, rutile, and a poorly defined nanocrystalline form. Anatase is most abundant and is commonly found with {010} faces in association with kaolinite edge and basal faces. The nanocrystalline form occurs at 0–1 wt. %, and rutile occurs in trace amounts. Bulk XRD analysis correlates well with the bulk TiO2 chemical measurements. Average anatase unit-cell parameters are a = 0.37908 ± 0.0002 nm and c = 0.951 ± 0.001 nm. These parameters indicate an approximate chemical formula of Fe3+0.05Ti4+0.95O1.95(OH)0.05.

The distribution of TiO2 content as a function of depth may be useful to obtain original grain-size variations associated with relative sea-level changes responsible for the deposition of the Huber Formation. Evidence for original depositional sediment properties can be seen in the occurrence of pseudomorphic replacement of micas and fecal pellets by kaolinite. Additional evidence for post-depositional changes includes the sub-micrometer euhedral character and low Fe content of the anatase (relative to soil-derived anatase). These observations for the Huber Formation are consistent with a previously published theory for kaolin genesis that includes biomineralization of originally coarser-grained aluminosilicates into a kaolinite-rich ore body.

Clay mineralogy and whole-rock chemistry of the borate-bearing layers of the Hisarcik and Esbey mines were examined. The Hisarcik clays occur as laminated or unlaminated clay layers with sharp contacts. Unlaminated layers contain quartz derived from metamorphic rocks and carbonate fragments in a clay matrix, and are interpreted as reworked tuffs deposited in playa-lake environments. An important feature is that the unlaminated clays contain little MgO (3–15 wt. %) as compared with the laminated clays (15–30 wt. %). As previous studies have shown, the clay fraction of the studied profile contains predominantly Li-bearing saponite, and accounts for 60–90 wt. % of the clay fraction (<2 μm). Illite in the clay fraction varies from 0 to 67 wt. % and the average illite percentage never exceeds 40 wt. %. Chlorite is scarce (2–5 wt. %). Illite-smectite interstratified clays (illite at 70%, smectite at 30%) were only found in low concentrations in the laminated clay layers of the upper limestone unit (above the borate zone), where illite-2M of detrital origin is also present. The Esbey clays occur interstratified with colemanite layers and envelope colemanite nodules. Calcite is the major mineral of the clays whereas quartz, plagioclase, feldspar, colemanite, and cahnite are minor components. The MgO contents vary between 4.70–13.95 wt. % in the clays interstratified with colemanite layers, between 7.24–11.89 wt. % in the enveloping clays, and between 10.27–21.25 wt. % in clays located above the colemanite zone. The composition of the clay fraction (<2 μm) in all samples is similar. Smectite represents between 40–90 wt. % of the clay fraction in the upper portion of the stratigraphic profile and decreases towards the lower part of the stratigraphic profile. Smectite always occurs with illite which may vary from 20 to 90 wt. % of the clay fraction, and a small amount of kaolinite and chlorite. Illite-2M polytype is abundant. The d(060)-reflection position suggests that the smectite minerals from the Hisarcik and Esbey colemanite mines contain both dioctahedral and trioctahedral smectites to form a transitional zone. These smectites are a product of a magnesium-rich alkaline playa-lake environment.

As a consequence of treatments with glycine solutions, glycine molecules enter the interlayer of both Ca- and Cd-rich montmorillonite. Measurements of d value suggest that at low glycine concentration (0.01 and 0.1 M glycine solutions) a “flat” arrangement of the glycine molecules occurs in the interlayer. In contrast, intercalation of more than one monolayer of glycine molecules occurs for the montmorillonite treated with a higher concentration of glycine (1 M glycine solution).

Interlayer complexation of glycine occurs only for the Cd-rich form of montmorillonite, whereas no complexation is observed for Ca-rich montmorillonite. Both nuclear magnetic resonance (NMR) and Fourier-transform infrared (FTIR) results suggest that the adsorbed glycine, which fully protonates in the interlayer of montmorillonite to give the GlyH2− species, interacts with the interlayer Cd2+ to form the CdGlyx complex mainly through the carboxylate group. The interlayer cadmium, present as both Cd2+ and CdCl−, is complexed by the ligand glycine. In contrast, the cadmium adsorbed on the external surfaces of montmorillonite does not interact with the ligand. Complexation of CdCl+ only occurs for large amounts of adsorption of glycine (e.g., for samples treated with 1 M glycine solution).

Alumina-pillared montmorillonite is prepared by intercalation of polyoxyhydroxy aluminum cations (Al137+) of a natural montmorillonite from Dimitrovgrad, Bulgaria. Transmission electron microscopy, powder X-ray diffraction, energy dispersive X-ray spectroscopy, and surface area (BET) methods are used to study the untreated and pillared forms of the montmorillonite. A structural model involving deformed Al13 pillars is proposed. Four pillar types are derived and these pillars are uniformly distributed over the interlayer-cation positions of montmorillonite. Calculated electron diffraction patterns were simulated using the multi-slice method. The structural model explains the increased ordering along the c axis of the pillared form compared with the untreated montmorillonite. The model explains the structure of a pillared montmorillonite with different distributions of the pillars in the interlayer. The proposed model is consistent with the observed data.

The Mecca Quarry Shale Member from Velpen, Indiana contains abundant vanadium which occurs in solid solution within illite-rich illite-smectite (I-S) having an average content of 1.65 wt. % V, and an overall composition of K0.8(Al2.8Mg0.5Fe0.4V0.3)(Si7.2Al0.8 g)O20(OH)4, analogous to the V-rich dioctahedral mica, roscoelite. The illite contains more than twice as much V as the associated kerogen. Detrital mica has a composition typical of 2M1, muscovite and contains no vanadium. The V-rich illite has a structure and composition typical of formation during normal prograde diagenesis and probably is widespread in the Mecca Quarry Shale because the bed is enriched in V throughout the Midwest. The smectite-to-illite reaction can not be a result of passive burial metamorphism because the host strata were buried no deeper than ~0.5 km at Velpen. The formation of illite occurred in unlithified sediments at shallow depths under the influence of pervasive 80–110°C basinal brines, possibly the same fluids that were responsible for the Mississippi Valley-type lead-zinc mineralization common in the Midwest. The presence of two types of K-rich phyllosilicates may be part of the reason for the lack of correlation between bulk V concentrations and the intensities of X-ray diffraction peaks of illite reported by others.

Owing to their large and chemically active surface, hydroxy-intercalated and pillared clays can be potent sorbents for organic compounds. The sorption behavior of Al and Zr-hydroxy-intercalated bentonite (HAl-, HZr-MX80), Al and Zr-pillared bentonite (Al-MX80, Zr-MX80), and a commercial Al-pillared bentonite (EXM 534) for 3-chloroaniline (3-CA), atrazine (AT), and 3-chlorophenol (3-CP) was investigated. The results were compared with the sorption behavior of the untreated Na-rich bentonite (MX80) and granulated activated carbon (GAC). Also the influence of the salinity of the sorbate and the age of the sorbents was studied.

Al and Zr-hydroxy-intercalated and pillared bentonites sorbed higher amounts of 3-CA, AT, and 3-CP than the untreated bentonite. The quantities sorbed related to the electron-donating properties of the sorbate and the acidity of the sorbents. Sorbed quantities increased from the hydroxy-intercalated to the pillared species, and from the Al to the Zr forms. The organic bases, 3-CA and AT, were sorbed in higher quantities than the organic acid 3-CP. For AT, the sorbents exhibited a high affinity. Aging of the samples and a high ionic strength of the sorbate reduced the sorption of 3-CA, whereas the sorption of AT was not affected greatly. The sorption capacity of GAC for organic bases was generally higher than that of the hydroxy-intercalated and pillared bentonites.

The data suggest that at initial concentrations at a ppm level, 3-CA and AT can be entirely removed from aqueous solutions by Al and Zr-hydroxy-intercalated and pillared bentonites. These materials, especially Zr-pillared bentonites, represent potent alternative sorbents for atrazine, chloroanilines, and probably a wide range of other organic bases.

Iron oxides in surface environments generally form at temperatures of 25 ± 10°C, but synthesis experiments are usually done at higher temperatures to increase the rate of crystallization. To more closely simulate natural environments, the transformation of 2-line ferrihydrite to hematite and goethite at 25°C in the presence of different Al concentrations and at pH values from 4 to 7 was studied in a long-term (16–20 y) experiment. Aluminum affects the hydrolysis and charging behavior of 2-line ferrihydrite and retards crystallization. Al also promotes the formation of hematite over goethite and leads to multidomainic discoidal and framboidal crystals instead of rhombohedral crystals. The strong hematite-promoting effect of Al appears to be the result of a lower solubility of the Al-containing ferrihydrite precursor relative to pure ferrihydrite. Hematite incorporates Al into its structure, as is shown by a decrease in the a and c-cell lengths and a decrease in magnetic hyperfine fields (Mössbauer spectroscopy). With hematite formed at low-temperature, these decreases were, however, smaller for the cell length and greater for the magnetic field than for hematite produced at higher temperatures. Both phenomena are removed by heating the hematite at 200°C. They are attributed to structural OH and/or structural defects. The relative content of Al in the structure is lower for hematite formed at 25°C than for hematites synthesized at higher temperatures (80 and 500°C). The maximum possible substitution of one sixth of the Fe positions was not achieved, similar to soil hematites. These results show that properties of widely distributed soil Al-containing hematites can reflect formation environment.

The effect of phosphate on the formation of Fe oxides from Fe(II) salts is important because phosphate is a ubiquitous anion in natural environments. For this reason, the products formed by oxidation of phosphate-containing Fe(II)SO4 solutions neutralized with bicarbonate were characterized. The rate of oxidation of Fe(II) increased with increasing P/Fe atomic ratio to 0.2 in the initial solution. Goethite (α-FeOOH) or lepidocrocite (γ-FeOOH) or both were produced and identified by powder X-ray diffraction (XRD). The ratio between lepidocrocite and goethite increased with increasing P/Fe. In the 5–8.5 pH range, the formation of goethite predominated at P/Fe < 0.005, but only lepidocrocite was detected by XRD for P/Fe > 0.02. Thus, phosphate favors lepidocrocite formation because lepidocrocite has (1) a layered structure (like its precursor green rust), and (2) a structure less dense than that of goethite, thereby requiring less complete removal of the green-rust interlayer phosphate to form. The lepidocrocite crystals were platy, with prominent {010} faces and the thickness of the plates decreased with increasing P/Fe from >25 nm for P/Fe < 0.005 to <5 nm for P/Fe > 0.1. The solubility of lepidocrocite in acid oxalate was nearly complete for P/Fe > 0.03. The lepidocrocite contained occluded phosphate, i.e., phosphate that could not be desorbed by alkali treatment. The decrease in the b unit-cell length with increasing P/Fe suggests that lepidocrocite may contain structural P.

KCl-, KBr-, and KI-kaolinite intercalation complexes were synthesized by gradually heating potassium-halide discs of the dimethylsulfoxide (DMSO)-kaolinite intermediate at temperatures to 330°C. Two types of complexes were identified by infrared spectroscopy: almost non-hydrous, obtained during thermal treatment of the DMSO complex; and hydrated, produced by regrinding the disc in air. The former showed basal spacings with integral series of 00l reflections indicating ordered stacking of parallel 1:1 layers. Grinding resulted in delamination and formation of a disordered “card-house” type structure. The frequencies of the kaolinite OH bands show that the strength of the hydrogen bond between the intercalated halide and the inner-surface hydroxyl group decreases as Cl > Br > I. The positions of the H2O bands imply that halide-H2O interaction decreases in the same order. Consequently, the strength of the hydrogen bond between H2O and the oxygen atom plane increases in the opposite sequence.

In the non-hydrous KCl-kaolinite complex the inner hydroxyl band of kaolinite at 3620 cm-1 is replaced by a new feature at 3562 cm-1, indicating that these OH groups are perturbed. It is suggested that Cl ions penetrate through the ditrigonal hole and form hydrogen bonds with the inner OH groups. In contrast, Br and I ions are too large to pass into the ditrigonal holes and do not form hydrogen bonds with the inner hydroxyls.

Self-diffusion coefficients of H2O molecules in Na-rich hectorite gel were measured by 'H nuclear magnetic resonance (NMR), Spin-echo pulse sequences with magnetic field gradient pulses for the translational diffusion measurement were applied to the hectorite gel at the Larmor frequency of 20 MHz. Effects of clay fraction (0-51.2 wt. %) and temperature (20.0-60.3°C) were studied. The results show: (1) Phenomenologically, the self-diffusion coefficient, D, of 1H2O in the clay gel is expressed by the normalized diffusivity, D/D0 = exp(−0.0257w), where D0 is the water self-diffusivity in bulk water at temperature and w is the weight fraction of the hectorite (wt. %). (2) The activation energy of H2O diffusivity in the hectorite gel is nearly equal to that in bulk water. Hence, the normalized diffusivity, D/D0, obeys a temperature-independent curve. (3) The exponential dependence of D/D0 on w for w <30 wt. % is explained by a random-walk model, in which free or unbound H2O molecules migrate in the geometrically complex and tortuous pore structure of randomly scattered clay-mineral grains.

A commercial bentonite (primarily smectite) from Fischer Scientific Company (F bentonite) and a natural bentonite from Peru (P bentonite) were used in the preparation of pillared clays with polyoxymetal cations of Al that were subsequently modified with Ce and La. Several Al/metal ratios (5 and 9) were used to investigate the effects on the thermal and hydrothermal stability of these synthetic clays. The structure of these materials was studied by X-ray diffraction. Isotherms were determined by N2 adsorption. Thermal stability was determined using thermogravimetric (TG) measurements and ara-monia-TPD (temperature programmed desorption) was used to obtain acidity data. These materials exhibited basal spacings from 16 to 20 Å, with surface areas from 239 to 347 m2g−1, with microporosity contributing from 50 to 80% of the total surface area. Pillared clays prepared from F bentonite generally showed larger basal spacings and surface areas than those prepared from P bentonite. Pillared clays modified with Ce or La did not show any apparent structural changes relative to the Al-pillared clays. Pillared clays modified with Ce and La had similar acid properties as Al-pillared clays. In contrast, the thermal and hydrothermal stabilities of these materials were greater than Al-pillared clays. However, Ce-pillared clay appears to be more effective than La-pillared clay in delaying the dehydroxylation of pillared clays with increasing temperature. The intercalation of Ce and La into Al-pillared clays improved the thermal stability, which may increase the utility of these materials as catalysts.

Trioctahedral smectite is a constituent of Mg-rich carbonate crusts and moonmilks (pasty deposits) in caves of the Guadalupe Mountains of southeastern New Mexico. Energy dispersive X-ray microanalysis of individual crystallites and their aggregates along with the X-ray diffraction analysis indicates that the smectite is probably stevensite. Saponite is likely present in some samples also. The smectite is intimately associated with dolomite crusts and huntite moonmilks in Carlsbad Cavern, Lechu-guilla Cave, and other dolostone caves. Clay particles appear as fibers and films, with aggregates comprising decimicron-sized filamentous masses that envelop crystals of dolomite, huntite, and magnesite. The occurrence of smectite is related to the genesis of the Mg-rich carbonate minerals. In water films, progressive evaporation and carbon dioxide loss results in the sequential precipitation of Mg-rich calcite, aragonite, dolomite, huntite, and magnesite. This sequence of carbonate precipitation removes Ca and greatly increases the Mg/Ca ratio in the solutions. Silica is commonly available probably because of high pH conditions, and consequently, smectite forms in the Mg-rich alkaline environment. Along with the Mg-rich carbonate minerals, opal, quartz, and uranyl vanadates may precipitate with the smectite.

Microbial structures in the form of banded zebra patterns have been found as periodic iron-manganese layers in living biomats on the coast of Satsuma-Iwo Jima, a small volcanic island near southern Kyushu, Japan. Electron microscopic observation shows that coccus, fibrous, and bacillus-type bacterial communities construct zebra architecture Fe-Mn layers through biomineralization on and within cells. A living microbial fumarolic ferro-manganese precipitation growing in seawater around an active volcanic island explains one mechanism of banded formation. Biological processes form the elemental zebra pattern, with periodic distribution of bacterial cells with Fe-Mn in each layer of the architecture. Fibrous bacteria are sometimes mineralized with goethite, ferrihydrite, and buserite microcrystals, coated with granular mucoid substances. The biomineralization may then mature to form a recent stratified banded-iron formation. The Satsuma-Iwo Jima zebra architecture is unusual in that it forms under aerobic conditions in a warm shallow-water environment, in contrast to the intermittent oxidizing and reducing conditions in which deep-sea analogues develop.

To assign far infrared (FIR) absorption bands of K+ in muscovite, dichroic experiments were performed. For a muscovite crystal rotated about a crystallographic axis, c*, a, or b, two bands corresponding to vibration modes of K+ appear, respectively, at 107 and 110 cm−1 (rotation about c*), 107 and 143 cm−1 (rotation about a), and 110 and 143 cm−1 (rotation about b). Two in-plane modes at 107 and 110 cm−1 and one out-of-plane mode at 143 cm−1 are identified for the vibrations of K+ in muscovite. Each of these transition moments are near the crystallographic axes b, a, and c, respectively. These observations match well predictions based on the approximate C3i symmetric environment of K+, although the site symmetry in the space group of muscovite is only C2.

A hydrothermally altered rhyolitic hyaloclastite from Ponza island, Italy, has four alteration zones with unique clay assemblages: (1) a non-pervasive argillic zone characterized by smectite; (2) a propylitic zone with interstratified illite-smectite (I-S) containing 10–85% illite (I); (3) a silicic zone composed of I-S with ≥90% I and pure illite; and (4) a sericitic zone with I-S ranging from 66% I to pure illite. Atomic force microscopy reveals abrupt changes in particle morphology with illitization, including initial changes from anhedral plates to laths and then to euhedral plates and hexagonal plates. I-S particles progressively thicken with illitization and mean particle area (basal plane) remains constant from pure smectite to I-S with 80% I. However, particle area increases from 90 to 100% illite. Computer modeling of I-S structural forms indicates octahedral cation ordering progressively changes from cis- vacant smectite to interstratified cis- and tnuis-vacant I-S, and then to trans-vacant illite. In addition, polytypes progressively change from 1 Md to 1M, and then to 2M, illite. Electron-microprobe and X-ray fluoresence analyses show that I-S chemistry progressively changes during illitization, evolving toward a phengitic composition with ∼0.89 fixed interlayer K+ per O10(OH)2. Octahedral Mg2+ shows little change with illitization, varying from 0.3 to 0.5 cations per O,10(OH)2. The layer charge of smectite is ~0.38 equivalents per O10(OH)2.

On the basis of abrupt changes in morphology and progressive changes in polytype and chemistry, smectite illitization on Ponza involved a dissolution and recrystallization mechanism with multiple stages of nucléation and crystal growth. In this multi-step model, temperature of alteration provided the major control for the layer composition, polytype, and morphology of I-S crystallites. Other factors that may play a secondary role include: K+ availability, water-rock ratio, and permeability. Alternatively, the mechanism of I-S and illite formation at Ponza and other hydrothermal environments may occur by direct precipitation of I-S crystallites from rhyolite glass and may not involve progressive reactions of smectite precursors.

Selective-dissolution techniques by ammonium oxalate (OX), dithionite-citrate-bicarbonate (DCB), and dithionite-ethylenediaminetetraacetic acid (D-EDTA), and X-ray diffraction and Mössbauer spectroscopy were used to identify and characterize iron oxides and oxyhydroxides in the <2-mm, <50-μm, and <2-μm size fractions of a Mollisol from Bahia Bianca, Argentina. Iron compounds are present at low concentrations in mixtures with quartz, Na-rich feldspar, illite, interstratified illite-montmorillonite, and traces of kaolinite. Total Fe and Al content increases as soil particle size decreases, from 4.3 and 13.3 wt. % in the <2-mm size fraction to 8.5 and 22.8 wt. % in the clay fraction (<2 μm), respectively. No more than 25–30% of the total Fe is associated with the crystalline and the amorphous Fe oxides. Weakly ferromagnetic hematite and goethite were identified in the different fractions. These phases have small particle sizes and/or low crystallinity. They may also have Al for Fe substitutions. Crystalline magnetite or maghemite is rare. These Fe-rich phases are probably coating coarser particles.

The efficiency of Fe removal is highest for the D-EDTA treatment and least efficient for the OX method, for all fractions. The opposite is true for Al removal. Poorly crystalline hematite and goethite, which are soluble in oxalate, are only present in the coarser fractions. Poorly crystalline and crystalline hematite and goethite, which are soluble in DCB and EDTA, are present in coarser fractions, but do not occur in the clay fraction. DCB treatment probably dissolves Al in the 2:1 type phyllosilicates occurring in this soil, whereas D-EDTA dissolves Fe in the hydroxy interlayers of the smectite minerals or in the silicate phases.

Hydrothermal syntheses were performed at various pH values and temperatures to induce variability in kaolinite defect density. Temperature of synthesis ranged from 200 to 240°C, for 21 d. Initial pH at room temperature ranged from 0.5 to 14. The starting material was a hydrothermally treated gel, with an atomic Si/Al ratio of 0.93, partly transformed into kaolinite.

Kaolinite was obtained for a wide range of pH. Although no influence of temperature on “crystallinity” (i.e., defect density) was observed, the effect of pH was important. A continuous series was obtained from a low-defect kaolinite, with high thermal stability and a hexagonal morphology for the most acidic final pH, to a high-defect kaolinite, with low thermal stability and lath shape for the most basic final pH. These variations of kaolinite properties appear related to the pH dependence of kaolinite surface speciation. Increasing pH value results in increased cation adsorption on the kaolinite external surfaces and increases in the elongation of particles.

The adsorption and degradation of the herbicide triasulfuron [2-(2-chloroethoxy)-N-[[(4-methoxy-6-methyl-1,3,5-triazin-2-yl)amino]carbonyl]benzenesulfonamide] (CMMT) on homoionic Fe3+-, Al3+-, Ca2+-, or Na+-exchanged montmorillonite in aqueous medium were studied. Ca- and Na-exchanged montmorillonite were ineffective in the adsorption and degradation of triasulfuron. The adsorption on Fe-and Al-exchanged montmorillonite was rapid, and equilibrium was attained after 5 min. Degradation of the herbicide was slow and the type of the degradation products depended on the nature of the exchangeable cations. In the presence of Fe3+-rich montmorillonite, the metabolites 2-(2-chloro-ethoxy)benzenesulfonamide (CBSA), 2-(2-chloroethoxy)-N-[[(4-hydroxy-6-methyl-1,3,5-triazin-2-yl)amino]carbonyl]benzenesulfonamide (CHMT), and l-[2-(2-chloroethoxy)benzene-l-sulfonyl]-7-acetyl-triuret (CBAT) were the only identified products, whereas 2-amino-4-methoxy-6-methyltriazine (AMMT), CBSA, CHMT, and CBAT were the primary metabolites for the Al3+-rich montmorillonite. A Fourier transform infrared (FT-IR) study of montmorillonite samples after the interaction with triasulfuron in organic solution suggests that the hydrolysis mechanism involves the adsorption of the herbicide on the 2:1 layers.

A method is proposed to measure the absolute concentration of paramagnetic Fe3+ ions in kaolinite from various geochemical environments using powder X-band electron paramagnetic resonance (EPR) data. An Fe3+-doped corundum sample is used as a concentration standard. The Fe3+ signal is calibrated by calculating the powder EPR spectra of Fe3+ ions in corundum and low-defect kaolinite. The paramagnetic Fe3+ concentration in other samples is obtained by an extrapolation procedure. This study provides a direct assessment of the iron distribution between isolated structural Fe3+ ions and other iron species, such as Fe3+ concentrated phases and Fe2+ ions. The concentration of isolated structural Fe3+ ranges between 200–3000 ppm and represents less than half of the total iron within kaolinite crystals.

A model to compensate the 2:1 layer having excess negative charge owing to the reduction of Fe3+ to Fe2+ by sodium dithionite buffered with citrate-bicarbonate in nontronite, beidellite, and montmorillonite is proposed. This model is based on reassessing published experimental data for Fe-containing smectites and on a recently published structural model for reduced Garfield nontronite. In the reduced state, Fe2+ cations remain six-fold coordinated, and increases of negative charge in the 2:1 layer are compensated by the sorption of Na+ and H+ from solution. Some of the incorporated protons react with structural OH groups to cause dehydroxylation. Also, some protons bond with undersaturated oxygen atoms of the octahedral sheet. The amount of Na+ (p) and H+ (ni) cations incorporated in the structure as a function of the amount of Fe reduction can be described quantitatively by two equations: p = m/(1 + K0mrel) and ni = K0mrel/(1 + K0mrel); with K0 = CEC (9.32 − 1.06mtot + 0.02mtot2), where mtot is the total Fe content in the smectite, m is the Fe2+ content, mrel is the reduction level (m/mtot), CEC is the cation-exchange capacity, and K0 is a constant specific to the smectite. The model can predict, from the chemical composition of a smectite, the modifications of its properties as a function of reduction level. Based on this model, the structural mechanism of Fe3+ reduction in montmorillonite differs from that determined in nontronite and beidellite owing to differences in the distribution of cations over trans- and cis-octahedral sites.