Altered basaltic tephra from Haleakala, Maui, are characterized using multiple techniques in order to identify the minerals formed under a variety of conditions and to understand the soil formation processes here. We collected samples that are representative of typical bulk weathered material in the crater, as well as solfataric alteration in a hydrothermal environment. For this study X-ray diffraction, electron probe microanalysis, scanning electron microscopy and transmission electron microscopy are coupled with spectroscopic techniques including Mössbauer, visible-infrared reflectance, transmission infrared, and thermal infrared emission spectroscopies to analyze these samples. The unaltered tephra are composed of feldspar, glass, pyroxene and olivine. Observed alteration products include Fe oxides, phyllosilicates and sulfates, as well as SAED amorphous Al-Si-bearing material. These samples are potential analogs for altered volcanic material on Mars as the pedogenic influences and contact with plants and animals are minimal. Results from this study may help to determine spectral signatures of these samples that could be used for identification on Mars of the minerals observed here.

The kinetics of cation exchange in phlogopite have been studied in situ by hydrothermal atomic force microscopy (HAFM). The exchange of interlayer K by octylammonium ions caused an increase in the interlayer distance and the formation of reaction fronts which can be locally resolved by AFM. The observed reaction fronts revealed substantial variations in their propagation rates — even within single interlayers. This observed variability in interlayer reactivity could mainly be attributed to chemical and structural inhomogeneities of the samples. A quantitative evaluation of the front propagation at representative sites yielded a diffusion coefficient of the K+ exchange by octylammonium of 1.2±0.6 × 10−11 cm2/s assuming negligible transport normal to the layers. The reverse reaction, i.e. the exchange of organic ions by K+, resulted in a retreat of the reaction fronts and a general restoration of the original morphological state. However, indications of structural alterations and areas with trapped octylammonium ions were found.

Polytypism in gümbelite and its relationship to the fibrous or ribbon-like morphology exhibited by this Mg-rich illite were investigated by powder X-ray diffraction (XRD), electron back-scattered diffraction (EBSD) and transmission electron microscopy (TEM). Comparison between the XRD pattern from oriented fibers using a conventional powder diffractometer and a randomly oriented pattern using a Gandolfi camera suggested that 2M2 is dominant but other polytypes belonging to subfamily A also exist, and that the fiber axis of gümbelite is parallel to <110> in 2M2, <110> in 2M1, and <100> in 1M. The EBSD analyses confirmed these crystallographic directions directly from individual crystals. Electron diffraction and high-resolution TEM showed that twinning and intergrowths of various polytypes including both subfamilies are common in a single crystal and that the two types of rotations [2n60° and (2n+1)60°] between adjacent layers are often randomly mixed at the monolayer level. The data suggest that high densities of twinning and intergrowths account for the origin of the fibrous morphology along <110> for 2M1 and 2M2 polytypes. Volume restriction in a confined vein space may also play a role.

Natural zeolites may represent one of the greatest under-utilized resources of the mineral world. Even with their unique character and properties, issues of consistency, homogeneity and purity preclude them from many premium applications. We report a simple method to upgrade mineral sedimentary Na-chabazite from the well known Bowie, Arizona, deposit to near synthetic purity and consistency. During this alkaline-silicate digestion process, initially soft chabazite ore granules gain substantial mechanical strength. This may allow direct employment in adsorption and purification processes without the need for binding and forming. These granules manifest significantly improved adsorption properties, including enhanced water and CO2 adsorptivity.

The octahedral composition of palygorskite in more than 300 samples from the Pefkaki deposit, W. Macedonia, Greece, has been studied by near-infrared (NIR) and X-ray diffraction (XRD), and evaluated according to the formula $y{\text{Mg}}_5{\text{Si}}_8{\text{O}}_{20}{\left(\text{OH}\right)}_2\cdot x{\text{Mg}}_2{\text{Fe}}_{\text{2}}^{\text{III}}{\text{Si}}_8{\text{O}}_{20}{\left(\text{OH}\right)}_2\cdot \left(1-x-y\right)$ Mg2Al2Si8O20(OH)2. Included in the study were PFl-1 and several commercial palygorskites. Our analysis of 2nd derivative NIR spectra shows that the dioctahedral composition is adequately described by three sharp overtone bands representing AlAlOH, AlFeIIIOH and FeIIIFeIIIOH in M2 dioctahedral sites, and that the summed intensity of these bands is proportional to the amount of dioctahedral component present (1−y). The samples show large variations in the degree of dioctahedral FeIII-for-Al substitution with FeIII occupying up to 70% of the dioctahedral M2 sites. Ternary analysis shows that the distribution of dioctahedral Al and FeIII is not random, but displays a tendency towards homoionic pairing. An overtone band at 7214 cm−1 and several combination bands are indicative of a trioctahedral Mg3OH component (y), and their appearance correlates with a distinct palygorskite signature in thermogravimetric analysis. Nevertheless, these bands cannot be used reliably for the quantification of a trioctahedral palygorskite component due to their close similarity to those of sepiolite. To circumvent this problem, we have evaluated y indirectly by calculating the difference between 1−y and the total concentration of palygorskite determined by the normalized intensity of the d110 XRD peak of palygorskite at 10.4 Å. Using this methodology, we have found that the samples conform to a trioctahedral limit of y ≈ 0.55, although within this limit they display large variations in octahedral character. Finally, we extend the above methodology to PLS chemometrics and show how NIR can be used to determine palygorskite content routinely in multimineralic geological samples.

The selectivities of clay minerals for larger organic cations over smaller ones have been attributed to favorable clay-organic interactions in clay interlayers and to hydrophobic effects resulting from (partial) dehydration of organic cations in the clay interlayers, but the magnitudes of these energy components have not been estimated. The objective of this study was to differentiate and quantify the contributions of clay-phase and aqueous-phase energy changes to the overall thermodynamics of cation exchange, and thereby to determine which forces control the general selectivity of smectites for organic cations. We compiled literature measurements and estimates for the free energies of overall cation exchange reactions and also for the free energies of organic cation hydration. Our study suggests that organic cation-exchange thermodynamics can be broken into three classes: (1) For two organic cations with identical head-groups, the difference in their cation exchange selectivities is driven almost quantitatively by the difference in their free energies of hydration. Here, the mechanism for organic cation selectivity is almost pure hydrophobic expulsion of the larger cation from water. The clay interlayer simply behaves like a subaqueous phase into which the least hydrophilic organic cations partition and the essentials of such cation exchange selectivity can be explained without any favorable clay-organic interactions. (2) For two organic cations with rather different head-groups, the difference in their cation exchange selectivities is just a small percentage of the difference in their free energies of hydration. This indicates that the clay phase interacts much more strongly with the cation having the smaller head-group, as might be expected on the basis of simple electrostatics. Here, the clay has an intrinsic strong preference for the cation with smaller head-group yet ‘selects’ for the cation with larger head-group because the aqueous-phase preference for the cation with smaller head-group is even stronger than the clay preference. (3) When the clay is already substantially loaded with organic cations, then van der Waals forces apparently can play a significant role in determining organic cation exchange selectivity differences.

The adsorption of Eu3+ from aqueous solution to natural Na+-montmorillonite (Kunipia F), synthetic saponite (Sumecton SA) and synthetic fluoro-tetrasilicic mica (Na+-TSM) clay samples was investigated. Adsorption capacities derived from the isotherms were 1.02, 0.71 and 1.00 meq/g of clay, respectively, for Kunipia F, Sumecton SA and Na+-TSM. The adsorption capacities were comparable to the cation exchange capacities of the clays, which were 1.19, 0.71 and 0.94 meq/g of clay, respectively. The greater slope of the TSM adsorption isotherm relative to the montmorillonite and saponite isotherms indicates a high affinity of Eu3+ for Na+-TSM. The high affinity of TSM for Eu3+ was thought to be related to the large electronegativity of the octahedral fluorine groups in TSM. Photoluminescence of adsorbed Eu3+ was observed for saponite and TSM, but not for montmorillonite. Quenching of Eu3+ luminescence by Fe in the montmorillonite structure is the probable reason for this phenomenon. The luminescence intensity varied with the amount of adsorbed Eu3+ for saponite and TSM as a result of self-quenching.

To provide structural and diffraction criteria for the identification of trans-vacant (tv) and cis-vacant (cv) mica varieties with different layer stackings, powder X-ray diffraction (XRD) patterns have been simulated for 1M, 2M1, 2M2, 3T and 2O structural models consisting of either tv or cv layers. The differences in the unit-cell parameters resulting from the specific structural distortions of tv and cv layers lead to the differences in the positions of reflections having the same indices in the XRD patterns for tv and cv 1M, 2M1 and 2M2 mica varieties. The tv 1M, 2M1 and 2M2 varieties of Al-rich micas can therefore be distinguished from the corresponding cv varieties using powder XRD diffraction provided that the d values are measured with high precision and accurately compared with those calculated from the unit-cell parameters for the corresponding hkl indices. The differences in reflection positions for these tv and cv varieties should decrease with increasing Mg and/or Fe contents, thus complicating their identification.

The peak positions and intensity distributions in the XRD pattern for the tv 3T variety are similar to those for the cv 3T structure with the vacancy in the right-hand cis site (3T-cv1), and both XRD patterns are similar to that for the 1M-cv mica. The simulated XRD pattern for the cv 3T structure with the vacancy in the left-hand cis site (3T-cv2) is similar to that for the 1M-tv variety. The similarities and dissimilarities in intensity distribution between the XRD patterns simulated for the 1M and 3T varieties in question may be associated with the differences in the mutual arrangement of cations and anions in successive layers.

Possible interstratification of tv and cv layers within the same structure should seriously complicate the identification of dioctahedral mica polytypes and polymorphs.

X-ray diffraction shows that ordered interlayer structures form when high-charged Llano vermiculite is reacted with HDTMA-Br or HDTMA-acetate, but the structures differ from those given by low-charged vermiculites. The differences arise in two ways; firstly when both HDTMA+ cations and HDTMA-Br (or -acetate) ion pairs are present as interlayer guest species, a superstructure forms which is less commensurate with the host structure of high-charge vermiculite than it is for lower-charged material. Secondly, although a commensurate 3a × b interlayer superstructure forms when HDTMA+ cations are the only interlayer guests, an increasing number of charge-balancing cations must be accommodated as the layer charge increases. This results in some structural disorder which is only observed in the diffraction pattern of HDTMA+-intercalated, high-charge vermiculite. Organic ion pairs tend to order along widely spaced rows corresponding to the intersections of two sets of scattering planes. For the acetate form, one set of planes has spacings of 3.52 Å, but for the bromide form the value is 3.67 Å. Scattering planes in the second set have spacings of 4.02 Å, independently of counter-anion size. These two sets of planes diverge from one another by ∼10°. Ordering is more apparent in the presence of acetate anions than it is for the smaller Br− anions. In the high-charged intercalates, from which organic ion pairs have been removed, HDTMA+ cations occupy positions on a centered two-dimensional superlattice and also randomly, some interstitial positions as necessitated by layer charge.

A field-mapping and crystal-chemical study of two alunite- and halloysite-rich deposits in the Turplu area, situated northwest of Balıkesir on the Biga Peninsula of northwest Turkey reveals a mineralogically diverse and a potentially economic clay deposit. The mineral assemblage along fault zones is dominated by halloysite and sometimes alunite. The alunite is nearly end-member in composition (a = 6.995 Å, c = 17.195 Å) often occurring with a minor Ca phosphate phase. Of the two deposits studied, the more northerly mine contains more alunite relative to halloysite. Geochemical alteration indices suggest that the northern mine has experienced a slightly greater degree of hydrothermal modification. Halloysite is found in both hydrated and dehydrated states and assumes a tubular morphology. Observations by transmission and scanning electron microscopy are consistent with a model of halloysite dehydration, where the shapes transform from an open-hole tubular morphology to a closed-hole unfurled morphology.

Mineral paragenesis includes the effects of initial deposition of volcanic tuffs and andesite on top of karstic terrain. The contact between altered volcanics and underlying limestones is irregular and appears to have provided a mechanism to flush both hydrothermal and meteoric waters through the volcanics. Periods of hydrothermal alteration (hypogene) contemporaneous with extensional and strike-slip faulting have resulted in alunite and halloysite deposits. Hydrothermal alteration is concentrated near the fault zones. Because of subsequent weathering (supergene) away from the fault zones, much of the andesitic volcanic rocks have been altered to a more smectite-rich and kaolinite-bearing assemblage. The deposits continue to be both plastically deformed in the alunite/halloysite regions and to undergo brittle deformation in the saprolitized volcanics. Tectonic deformation has mixed the contacts, such that limestone blocks are entrained into parts of the alteration zones. Gibbsite and gypsum are common weathering products associated with limestone block inclusions. Genetic models for the origins of alunite-halloysite deposits in NW Turkey should consider as possible influencing factors the underlying lithologies, the extent of hydrothermal alteration, and recent weathering by meteoric fluids. In the case of the Turplu deposits, karstic limestones, hydrothermal circulation of sulfate-rich waters, and a post-alteration history of meteoric weathering were all important factors in their formation.

The structure and dynamics of alkylammonium-intercalated smectites were simulated using molecular dynamics employing the clayff-CVFF force field, the reliability of which was firstly validated for these systems. The layering behaviors of alkyl chains confirm the scenarios of the monolayer, transition and bilayer configurations for short, medium-length and long carbon tails, respectively. In the systems without water, the alkylammonium groups are all anchored firmly above the surface six-member rings through H bonds between ammonium hydrogen and surface oxygen, and the alkyl tails are a little more mobile. With water involved, some ammoniums are dragged out of the potential barriers of the six-member rings by water molecules through the strong H bonds between water oxygen and ammonium hydrogen. The intercalated water scarcely affects the basal spacing, alkyl chain layering or alkylammonium dynamics. It is also found that the systems with alkyl chains of 11 to 14 exhibit the greatest density, resulting in the extremely limited mobility of the intercalated species.

A calorimetric method for determining isothermal partial and integral heats of hydration reactions ($\text{Δ}{\overline{H}}_{\text{R,}T,P}$${\rm{\Delta }}{\bar H_{{\rm{R,}}T,\,P}}$ and $\text{Δ}{\stackrel{\sim }{H}}_{\text{R,}T,P}$${\rm{\Delta }}{\tilde H_{{\rm{R,}}T,\,P}}$, respectively) in zeolites and other mineral hydrates is presented. The method involves immersing a dehydrated sample in a humid gas stream under isothermal conditions within a thermal analysis device that records simultaneous differential scanning calorimetric (DSC) and thermogravimetric analysis (TGA) signals. Monitoring changes in sample mass (corresponding to extent of reaction progress) coincident with a quantitative measurement of heat flow allows for direct detection of $\text{Δ}{\overline{H}}_{\text{R,}T,P}$${\rm{\Delta }}{\bar H_{{\rm{R,}}T,\,P}}$ as a function of the extent of hydration, which can be integrated to determine $\text{Δ}{\stackrel{\sim }{H}}_{\text{R,}T,P}$${\rm{\Delta }}{\tilde H_{{\rm{R,}}T,\,P}}$. In addition, it eliminates uncertainties associated with imprecise knowledge of the starting and final states of a sample during hydration. Measurement under isothermal conditions removes uncertainties associated with heat capacity effects that complicate interpretations of DSC measurements of dehydration heats conducted under traditional scanning temperature conditions. Example experiments on the zeolites natrolite, analcime and chabazite are used to illustrate strategies for quantifying $\text{Δ}{\overline{H}}_{\text{R,}T,P}$${\rm{\Delta }}{\bar H_{{\rm{R,}}T,\,P}}$ and $\text{Δ}{\stackrel{\sim }{H}}_{\text{R,}T,P}$${\rm{\Delta }}{\tilde H_{{\rm{R,}}T,\,P}}$ and minimizing errors associated with baseline uncertainties. Results from this method agree well with previously published values determined by other calorimetric techniques and regression of phase equilibrium data. In the case of chabazite, the results allowed detailed measurements of the variation in $\text{Δ}{\overline{H}}_{\text{R,}T,P}$${\rm{\Delta }}{\bar H_{{\rm{R,}}T,\,P}}$ for energetically different water types encountered progressively as the sample absorbed water. This technique complements and in many cases improves the quality of thermodynamic data obtained through phase equilibrium observations and other calorimetric techniques.

Thermally induced lamellar structure changes due to phase transition and degradation in organoclays based on a synthetic ‘Somasif’ mineral and two organic surfactants, di-methyl dihydro-ditallow ammonia chloride (DMDTA) and tri-butyl-hexadecyl phosphonium bromide (HTBP) were investigated by differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), Fourier transform infrared (FTIR) spectroscopy, in situ simultaneous small-angle X-ray scattering (SAXS) and wide-angle X-ray diffraction (WAXD) over the temperature range 30–280°C. Results indicated that the surfactant layer in ‘Somasif’-based organoclays underwent thermally induced melting-like order-disorder transition followed by desorption of surfactant molecules, resulting in drastic changes in the character of the layer periodicity. The transition temperature (Ttr), determined from the endothermic transition in DSC, was found to depend strongly on the type and the content of surfactant incorporated. Temperature-resolved SAXS indicated complex intercalated layered structures, containing multiple lamellar stack populations of two different organic layer thicknesses. A weak scattering peak (s0), located at exactly the half angular position of the strong first scattering maximum s1 (s0 = 0.5s1), was found in all tested ‘Somasif’ clays. The presence of this peak can be attributed to a slight breaking of the translational symmetry in the layered structure, causing the 1D repeat period in real space to be doubled. In other words, some portions of layers are grouped into pairs and a single pair forms the new repeat unit. This arrangement is reminiscent of the Peierls-like distortion.

The tropical weathering of sedimentary kaolin deposits from the plateaux surrounding Manaus (Alter do Chao formation, Amazon basin, Brazil) leads to the in situ formation of thick kaolinitic soils. The structural changes of kaolinite have been investigated quantitatively by infrared spectroscopy and electron paramagnetic resonance. Both techniques consistently show that each sample contains two types of kaolinite in various proportions. The progressive decrease in kaolinite order from the bottom to the top of the profile results from the gradual replacement of an old population of well-ordered kaolinite, typical of the underlying sedimentary kaolin, by a more recent generation of poorly ordered soil kaolinite. The vertical pattern of kaolinite replacement differs from that of the transformation of Fe oxides and oxyhydroxides previously observed in the same profile. The inherited fraction of well-ordered kaolinite ranges from 60% at a depth of 9 m to 30% in the upper levels of the soil. The persistence of sedimentary kaolinite in the upper horizons suggests that the rate of kaolinite transformation is relatively slow at the time scale of lateritic soil formation. Kaolinite inheritance unlocks the lateritic record of past weathering conditions.

Mixed-layer illite-smectite with high degree of ordering (rectorite-like clay) and with three types of interlayer cations — K+, Na+ and ${\text{NH}}_4^{+}$, was determined in the high-sulfide epithermal mineralization of the Western Carpathian Mountains. The tobelitic portion of the clay comprises 5–10%. Ammonium was detected both by chemical analysis and by Fourier transform infrared analysis. Some of the ammonium can be assigned to a poorly bound form, not fully fixed in the illitic interlayer. The finest size fraction separated from the samples behaves differently from the rest of the sample. It contains two layers of water molecules in the expandable interlayers as determined by X-ray diffraction, whereas coarser fractions have only one water layer in the interlayers.

Diagenetic kaolin minerals are very common in the Permo-Triassic succession from the SE Iberian Range, Spain. The morphology and crystal structure of kaolin minerals has been examined in four size fractions (<1 µm, <2 µm, <6.3 µm and <20 µm) of sandstone samples by means of scanning electron microscopy, X-ray diffraction, infrared spectroscopy, differential thermal analysis and thermogravimetry. Experimental data reveal that dickite is the dominant kaolin-type mineral in the entire range of size fractions, whereas small amounts of kaolinite coexists with dickite in all size fractions. Dickite appears typically as booklets of pseudo-hexagonal plates with blocky habit. The increase in size fraction is concomitant with the increase in the amount of dickite and the progressive improvement of its structural order. The extensive dickitization is attributed to the high paleogeothermal gradient recorded in the studied area and the increase in H+, presumably resulting from the flux of organic acids derived from the underlying Carboniferous rocks and/or the late Permian succession. These conditions are more likely to be associated with the late Cretaceous post-rift thermal stage of the eastern Iberian Basin. Lately, during the maximum burial depth, the fine crystalline kaolin minerals were slightly illitized. Given the very small feldspar content in the studied sequence, the results reflect the important contribution of mica alteration to the early diagenetic formation of kaolinite as well as the late conversion to dickite.

From leaching experiments with metallic uranium-aluminum research reactor fuel elements in repository-relevant MgCl2-rich salt brines, a Mg-Al layered double hydroxide (LDH) with chloride as the interlayer anion was identified as a crystalline secondary phase component. The incorporation behavior of europium into the structure of the Mg-Al-Cl LDH was investigated. Synthesis via co-precipitation was performed. The Mg-Al-Eu-Cl LDH obtained was treated with a concentrated ammonium carbonate solution. No release of Eu was detected; hence the molar stoichiometry of the LDH remained stable with respect to Mg, Al and Eu. This chemical behavior might be the first indication of the incorporation of Eu.

The material was further examined by powder X-ray diffraction. Structural parameters were obtained from comparisons of simulated and experimental diffraction patterns of a ${\text{CO}}_3^{2-}$${\rm{CO}}_3^{2 - }$-exchanged Mg-Al-Eu LDH and a Mg-Al LDH. The two materials showed different behaviors according to stacking order and lattice parameters. This is an indirect indication of the incorporation of Eu.

The weathering stages of K-depleted biotite, K-depleted phlogopite, and natural biotite were investigated using hydrothermal treatment with A1C13 solution at 200°C for 12 to 72 h. Although there were some differences in the degree of weathering of the two K-depleted micas, both first transformed to hydroxy-Al interlayered vermiculite (HIV), which then altered to kaolinite. In case of the natural biotite, the biotite first transformed to a K-depleted mica-like phase, which then altered to kaolinite. The natural biotite had resisted weathering to kaolinite more than did the K-depleted biotite, as expected. The K-depleted phlogopite had less resistance in weathering to kaolinite than the K-depleted biotite. The transformation process changed the color of the micas. The K-depleted biotite changed from greenish-black through yellow to pale gray whereas the natural biotite changed from greenish-black through beige to yellow. However, in the case of K-depleted phlogopite, there was no significant color change during the transformation process. The presence of the interlayer K+ ions and the structural Fe2+ ions in mica appear to have contributed to the differences in the degree of weathering to kaolinite among the micas investigated.

Somen-alkyldiamines with thegeneral formulae H2N(CH2)nNH2 (n = 2–5) were intercalated into the layered silicic acid magadiite, from aqueous solution, causing an increase in the original interlayer distance of 1172 pm. The synthetic magadiite and all intercalated compounds were characterized by elemental analysis, infrared vibrational spectroscopy, X-ray diffractometry, 29Si nuclear magnetic resonance in the solid state, thermogravimetry, scanning electron microscopy, surface area and porosity. The intercalation was followed through a batch-wise method at 298±1 K and gave the maximum amounts 3.70, 2.80, 1.75 and 1.18 mmol g−1, for n varying from 2 to 5, respectively. The well characterized magadiite was calorimetrically titrated in a heterogeneous medium, to obtain the thermodynamic data of intercalation at the solid/liquid interface. Linear correlations were obtained for the number of moles intercalated (Nf), th einterlamellar distance (d) and the specific enthalpy (Δinth) values of the interactive process as a function of the number of C atoms of the aliphatic organic chains (nC) for n-alkyldiamine: Nf = (5.36±0.25) − (0.86±0.07)nC, d = (1406.6±1.9) + (20.9±0.5)nC and Δinth = (5.96±0.25) + (0.06±0.01)nC. The basic N guest atom/silanol acidic center interactions inside the host nanospace gallery gave exothermic enthalpies, positive entropies and negative Gibbs free energy values. This set of data suggests the spontaneity of these intercalation reactions.

Thermogravimetric analysis combined with mass spectrometry was used to study H2O bound to samples of illite-1M, illite-2M2 and leucophyllite-1M. Samples were heated in a helium atmosphere at different temperatures and after heating at each given temperature were cooled to 35°C. Each cycle in the mass 18 spectrum of each illite sample contains a low-temperature peak at 60–80°C, a medium-temperature peak at 340–360°C, and a high-temperature peak at a temperature that is very close to the maximum temperature of sample heating of a given cycle. Within each heating-cooling cycle, the sample weight at the beginning of cooling is lower than that at the end of the same cooling stage because of H2O resorption. However, the number of H2O molecules released during each medium-temperature heating cycle is equal to the number of H2O molecules resorbed during the corresponding cooling stages.

The weight losses, under medium-temperature heating, of the illite samples are related to dehydration when H2O molecules located in K-free sites of the illite interlayers are removed. The medium-temperature peak is reproducible for each cycle because during each cooling stage the illite interlayers resorb the same number of H2O molecules that were lost during the preceding dehydration.

Two distinct features are characteristic of leucophyllite during heating-cooling treatments. First, the number of H2O molecules resorbed during cooling is significantly greater than the number of H2O molecules lost during dehydration. Second, the medium-temperature peaks in the spectrum appear only in the last five cycles and the maximum-peak temperature is 450–460°C. These data indicate that the heating-cooling treatments are accompanied by partial rehydroxylation. This rehydroxylation occurs during each coolingstage when a small number of resorbed H2O molecules are trapped in the interlayers, although most migrate into the octahedral sheet of the 2:1 layers and reform as OH groups. The crystal chemical factors responsible for the dehydration and rehydration as well as for the rehydroxylation reactions are discussed and speculation about the origin of the low- and medium-temperature H2O losses is presented.