Seven kaolins from Georgia (southeastern U.S.A.), ranging from high to low commercial grade, were characterized by X-ray powder diffraction and chemical techniques to establish that the variation in quality was caused by impurities. The Ca and Cs cation-exchange capacities (CEC) varied from 2.67 to 8.17 and from 3.29 to 8.77 meq/100g, respectively. Selective dissolution and correlation analyses strongly indicated that expandable 2:1 minerals, particularly smectite (1.2-5.9%), were responsible for most of the observed variations in Ca CEC (r = 0.85*). The external surface CEC of kaolinite ranged from 0 to 1 meq/ 100 g. The positive significant correlation (r = 0.90**) between the Ca CEC and the K-mica content (03.9%) suggested that Ca CEC may be related to the degree of mica weathering through an expandable mineral stage.

The Cs-retention capacity (0.19–1.14 meq/100 g) was closely related to Cs-measured vermiculite content (r = 0.80*), and this content plus specific surface (R = 0.93**) or mica content (R = 0.86*). The Cs retention appeared to be primarily related to the presence of interlayer wedges at mica/vermiculite XY interfaces.

Kaolinites of all kinds (fine, ‘fireclay,’ ‘type IV,’ etc.), some of which do not expand or expand incompletely with the usual intercalation methods used for comparison, are expanded completely by treatment of dry (110°C) clay with dry CsCl salt, followed by contact with hydrazine for 1 day at 65°C and then with DMSO overnight at 90°C. Comparison treatments were grinding in KOAc, soaking in hydrazine, and Li-DMSO, as well as combination of these. Following the Cs-hydrazine-DMSO treatment, the 7.2 Å spacing of 1:1 dioctahedral layer silicates shifts to 11.2 Å and the 11.2 Å/(7.2 + 11.2 Å) ratio ≃1.0. The trioctahedral 1:1 layer silicates and chlorite are not expanded by the Cs-hydrazine-DMSO procedure.

The influence of geomorphological site characteristics on soil clay mineral stability of montmorillonite-containing horizons of a southern Wisconsin soil catena was interpreted in terms of the solute activity function values of pSi(OH)4, pH-1/2pMg2+ and pH-1/3pAl3+ in suspensions of the separated clay fractions. Montmorillonite stability and/or formation vs that of kaolinite for the soil clays was evaluated by a plot of the solute activity functions on a three dimensional diagram derived for montmorillonite, kaolinite, and gibbsite at constant temperature (25°C) and constant pressure (one atm.). Although all the soil clays contained both montmorillonite and kaolinite, the position of the soil clay solute activity functions in the stability diagram clearly reflected the influence of the geomorphological—geochemical site conditions in which each soil horizon was developed, with corresponding differences in the SiO2/Al2O3 molar ratio of the reactive fraction. Montmorillonite stability positions of the solute activity functions were induced by soils (clays with reactive fractions with SiO2/Al2O3 molar ratios = 3–4) from calcareous or poorly drained horizons, while kaolinite stability positions of the functions were induced by soils (clays with reactive fractions of SiO2/Al2O3 molar ratios = 2) from acid, freely drained horizons.

Titanium in TiO2 minerals was differentiated from that isomorphously substituted into minerals by the use of dihydrogen hexafluorotitanate (hydrofluotitanic acid, H2TiF6), which selectively dissolved minerals containing substituted Ti4+, leaving free crystalline TiO2 minerals in the residue. Titanium analyses on the original samples and the residues remaining after H2TiF6 treatment, by both wet chemical (Tiron) and neutron activation methods, indicated that an average of 86 per cent of the titanium in seven kaolinite samples was present in the residual TiO2 form (largely anatase), whereas only 28 per cent in two bentonites was present in the TiO2 form. Residual Ti accounted for 100 per cent of the Ti in synthetic anatase and for 92 per cent of the Ti in coarse clay sized rutile, the latter value suggesting that about 8 per cent amorphous TiO2 was removed from the mechanically dry ground rutile by the H2TiF6 reagent. The Ti present as residual TiO2 in a variety of other samples ranged from 0 to 100 per cent.

A steady state reaction of apparent equilibrium of K mica + Ca2+ ⇄ Ca vermiculite + K+ was indicated by prolonged dissolution extractions from Blount soil clay (from northern Indiana) abundant in dioctahedral mica and vermiculite, with log Keq = 2.92 for the reaction when extrapolated to infinite time. From this and published free energies of formation of mica and kaolinite, a mineral phase stability diagram depicting the phase joins of Ca vermiculite, muscovite, and kaolinite was constructed with the solute activity functions pH-pK+, 2pH-pCa2+, and pSi(OH)4. These solute functions for 14-day reactions of calcareous (and dolomitic), poorly drained Harps soil (from central Iowa) fell near the calcite-dolomite-CO2-H2O phase join, suggesting equilibrium. These functions for Harps soil and the control minerals muscovite, biotite, and (or) vermiculite plus calcite were plotted on the mica-vermiculite stability diagram for various CO2 partial pressures. The points fell on the vermiculite-stable side of the mica-vermiculite plane at CO2 partial pressures of 0.15 and 0.20 atm (similar to soil air that would exist under frozen soil during winter and early spring; 2pH-pCa2+ ≃ 10.3). They fell on the muscovite-stable side of the muscovite-vermiculite plane at CO2 partial pressures of 0.0001 and 0.001 atm (similar to soil air under natural summer conditions; 2pH-pCa2+, 13.6 and 12.6, respectively) and therefore K+ (and 137Cs+ in rainfall) would be expected to be fixed.

The 2pH-pMg2+ values determined for Harps soil at the various CO2 partial pressures plotted either in the Mg montmorillonite stability field or on the Mg-montmorillonite-kaolinite phase join, in concordance with the abundance of montmorillonite and some kaolinite in the medium and fine clay fractions. The solute values for the nearby Clarion soil (upland, noncalcareous) plotted on the montmorillonite-kaolinite join, or with higher CO2 partial pressure, in the kaolinite stability area. The Gibbs free energy of formation (△Gf0) for a dioctahedral Ca vermiculite of −1303.7 kcal per 010 was determined from the Keq. The solute functions for the Blount soil showed kaolinite to be the thermodynamically stable phase with respect to dioctahedral mica and (or) vermiculite. The 14-day solute values for the Harps and upland Clarion soils were also on the kaolinite stability side of the kaolinite-vermiculite join. The kinetics of kaolinite formation in the upper midwestern U.S.A. are apparently slow on a scale of ~ 104 years.

The surface charge density of mica (001) cleavages was determined by counting the number of fission particle tracks in a given area of a 6-mm muscovite disc replica with optical and scanning electron microscopy after saturation of the layer charge by washing with 0.5 M UO2(NO3)2 solution, dilution of the excess salt by washing with 0.01 M UO2(NO3)2 in 0.005 M HNO3 (pH 2.4), blotting off the excess liquid, thermal neutron activation in contact with the muscovite disc, etching the muscovite, and counting the 235U fission tracks/cm2. In initial studies, the uranyl cations were found to hydrolyze from the cleavage surface continuously during the washings with water, ethanol or acetone to remove excess salts, but the uranyl cations in the interlayers near broken edges and crystallographical steps were strongly retained even against washings with 0.5 M CaCl2 solution. The hydrolysis of UO22 + from the smooth portions of the flake surfaces was avoided by the use of three 1-hr final washings with the 0.01 M UO2(NO3)2 in 0.005 M HNO3 solution. Each flake was pressed between filter papers three times to remove the excess solution. A negligible amount of excess salt remained on the cover glass controls. The UO22 + cations retained (mean, 3.6 ± 0.2 × 10−7 mequiv./cm2) on the cleavage surfaces of various micas were nearly equivalent to the theoretical surface charge (cation exchange capacity, 3.5 × 10−7 mequiv./cm2), showing that hydrolysis was prevented. The uranium on the unblemished mica planar surfaces increased with increasing uranyl concentrations in the final washing solution, indicating that the excess salt remaining on the surfaces had become significant. With a given UO22 + salt concentration, the uranium on the surface increased on increasing the solution pH from 2.5 to 3.5, attributable to the formation of polymeric ions such as U2O52 + and U3O82 + with higher uranium retention per unit positive charge equivalent to the fixed negative charge of the mineral surface. Uranyl cations replaced much of the interlayer cations from vermiculites even after K, Rb and Cs presaturation and drying from 110°C were employed. Strong adsorption of uranyl cations (in a form not replaced by washings with a neutral salt solution), which occurred in the defects of micaceous minerals, is important in the interpretation of actinide element retention in soils and sediments wherein these minerals are abundant.

The objective of this paper is to present flow sheets for a system of quantitative minera-logical analysis of clays of soils and sediments and to show representative results. Selective dissolution analysis by the Na2S2O7-HCl-NaOH procedure yields the quartz and feldspar contents (0 to 63%) and differentiates feldspar K from mica K. The NaOH-thermal system of selective dissolution yields the allophane plus gibbsite, kaolinite plus halloysite, and dickite contents (0 to 84% for the sediments; 1 to 25% for soil clays) Mica contents (0 to 92% for the rock specimens, 7 to 43% for soil clays) are determined by nonfeldspathic K (and Na). Vermiculite contents (1 to 97% of specimens; 3 to 21% for soil clays) are measured by blocking of interlayer CEC by drying at 110°C while K saturated and replacing with NH4Cl. Montmorillonite (and palygorskite) contents (0 to 85% of specimens; 3 to 36% of soil clays) are determined by the CEC not blocked by the K and NH4 sequence for vermiculite. Chlorite contents (0 to 85% for specimens; 0 to 37% for soil clays) are determined by thermal gravimetric analysis, after allocation of OH water lost between 300 and 950°C to other hydrous minerals determined.

The best evidence of the accuracy of the system of analysis lies in the consistent total recovery of 24 standard mineral samples averaging 100.4 ± 1.3 (± standard error of means) and of 22 soil clay samples averaging 99.5 ± 0.8. The different constituents were present in widely different proportions in the various samples, and were determined by independent methods. The complementary total of near 100% (maximum range 95 to 105% for specimens; 95 to 103% for soil clays) for the analyses is a significant measure of the specificity of the several determinations.

The layer structure of kaolinite from Twiggs, Georgia and fire-clay type kaolinite (Frantex B, from France), particle size separates 2–0·2 μm was studied by high resolution electron microscopy after embedment in Spurr low-viscosity Epoxy media and thin sectioning normal to the (001) planes by an ultramicrotome. Images of the (001) planes (viewed edge-on) of both kaolinites were spaced at 7 Å and generally aligned in parallel, with occasional bending into more widely spaced images of about 10 Å interval. Some of the 10 Å images converged to 7 Å at one or both ends, forming ellipse-shaped islands 80 to 130 Å thick and 300 to 500 Å long. The island areas and interleaved 10 Å layers between 7 Å layers may represent a residue of incomplete weathering of mica to kaolinite.

The proportions of micaccous occlusions were too small and the layer sequences too irregular to be detected by X-ray diffraction. The lateral continuity of the layers through the 7-10-7 Å sequence in a kaolinite particle would partially interrupt or prevent expansion in dimethyl sulfoxide (DMSO) and other kaolinite intercalating media. Discrete mica particles were also observed with parallel images at 10 Å, as impurities in both kaolinites. The small K content of the chemical analyses of the kaolinite samples is accounted for as interlayer K, not only in discrete mica particles but also in the micaceous occlusions.

Four Na2S2O4-reduced Na-vermiculites, each with some trioctahedral mica interstratified, were oxidized with H2O2 at pH 6·5 and again reduced with Na2S2O4 in suspensions at pH 7·5–8·0. The layer charge (CEC + K+), measured at pH 6·50, did not change significantly when octahedral Fe was oxidized (7–92 mmole 100g−1) or reduced (6–71 mmole 100 g−1). Electroneutrality was maintained within the octahedral sheet when Fe was oxidized or reduced. When Fe(II) was oxidized, electroneutrality was maintained by deprotonation of octahedral OH− groups,

(a)

${\left\{{\left[Fe\left(II\right)\right]}_{2}M{g}_4{O}_4{\left(OH\right)}_4\right\}}^{\pm 0}\leftrightarrows {\left\{{\left[Fe\left(III\right)\right]}_{2}M{g}_4{O}_4{\left(OH\right)}_2{O}_2^{\ast }\right\}}^{\pm 0}+2e^{-}+2H^{+}$${\left\{ {{{\left[ {Fe\left( {II} \right)} \right]}_2}M{g_4}{O_4}{{\left( {OH} \right)}_4}} \right\}^{ \pm 0}}\leftrightarrows{\left\{ {{{\left[ {Fe\left( {III} \right)} \right]}_2}M{g_4}{O_4}{{\left( {OH} \right)}_2}O_2^*} \right\}^ {\pm 0}} + 2{e^ - } + 2{H^ + }$

and by ejection of (dissolution of structural) octahedral metallic cations,

(b)

${\left\{{\left[Fe\left(II\right)\right]}_{5}Mg{O}_4{\left(OH\right)}_4\right\}}^{\pm 0}\to {\left\{{\left[Fe\left(III\right)\right]}_4{O}_4{\left(OH\right)}_4\right\}}^{\pm 0}+5e^{-}+F{e}^{3+}+M{g}^{2+}.$${\left\{ {{{\left[ {Fe\left( {II} \right)} \right]}_5}Mg{O_4}{{\left( {OH} \right)}_4}} \right\}^{ \pm 0}} \rightarrow {\left\{ {{{\left[ {Fe\left( {III} \right)} \right]}_4}{O_4}{{\left( {OH} \right)}_4}} \right\}^{ \pm 0}} + 5{e^ - } + F{e^{3 + }} + M{g^{2 + }}.$

When Fe(III) was reduced, electroneutrality was maintained by reprotonation of the deprotonated sites (O*, equation a). Reaction (b) was not reversible. Thus, reversibility of the reaction, Fe(II) ⇄ Fe(III) within the octahedral sheet decreased with increasing amount of ejected metallic cations. The amount of Fe(III) and Mg2+ ejected per Fe(II) oxidized was related to the degree of vermiculitization, being greatest with Na-degraded biotite [0·03 Fe3+ and 0·11 Mg2+ per Fe(II) oxidized] and lowest (nearly zero) with South African vermiculite. The number of deprotonated (O*) and reversible sites increased from 0·69 per Fe(II) oxidized with the K-depleted biotite to approximately 1·0 with South African vermiculite. The weathering increment was small since, of the total amount of Fe + Mg, less than 1·3 per cent was ejected from any of the four vermiculitic materials. When biotite was K-depleted, about 20 m-equiv of layer charge per 100g (300°C basis) was lost, while 51 mmole of Fe(II) per 100g was oxidized in the presence of Na2S2O4 and 82 mmoles in its absence in the aqueous suspensions. Since sequential reduction-oxidation-reduction treatments of K-depleted biotite and mica-containing vermiculites did not cause significant changes in layer charge (r2 = 0·04), the layer charge changes were concluded to be entirely independent of the oxidation or reduction of Fe in these minerals.

Quartz isolated from soils by the pyrosulfate-H2SiF6 method and chert samples of various origins were examined with the scanning electron microscope. Quartz isolates of the 20–50μm from the A2 and B21t horizons of the Baxter soil (AR), with quartz δ18O of 18·2 and 19·0‰, respectively, showed a mixture of detrital quartz particles and chert clusters of aggregates of fine euhedral quartz crystals. The 2–5 μm fractions of both horizons consisted mainly of euhedral quartz particles. The 20–50 μm fraction from the underlying chert, with a δ18O = 29·6‰, consisted of aggregates of euhedral quartz particles 1–10 μm dia and of subhedral particles 0·1–0·5 μm dia. In the soil fractions, the size and shape of the quartz particles as well as oxygen isotope data indicated that the aggregates were from the underlying chert but that irregular, unaggregated grains were detritally admixed loess, particularly in the medium and coarse silt fractions. This mixing of chert (of low temperature origin and heavy oxygen) with detrital quartz (of high temperature origin and light oxygen) gave rise to the intermediate δ18O values in the soil quartz. The SEM of cherts of different geological ages showed different morphologies. Prismatic, polyhedral microcrystalline quartz of 1–10 μm size as well as submicron, euhedral particles were observed in cavities. Submicron, subhedral particles and interlocking quartz grains were characteristic of Precambrian chert. Quartz grains more than 100 μm in size isolated by HCl from Ordovician dolomite (WI) had large (many microns) subhedral overgrowths and attached clusters of 0·1–0·5 μm microcrystalline quartz. A Danish flint formed in chalk had calcite-lined cavities (x-ray emission determined) in which spheroidal fibrous chalcedony occurred.

Deferration by reduction of free Fe2O3 with Na2S2O4 in the presence of Na citrate and NaHCO3 caused a change in valence state of 10 to 35 per cent of the total structural iron in micaceous vermiculites, soils, nontronite, and muscovite. An increase in Fe2+ on deferration was accompanied by an equivalent decrease in Fe3+. Subsequent treatment with H2O2 reoxidized the structural Fe2+ previously formed.

Sesquioxide coatings on micaceous vermiculites were examined electron microscopically. These coatings were composed predominantly of Fe2O3 with approximately 10 per cent by weight of Al2O3 and small amounts of SiO2, as determined by chemical analysis of the deferration extracts.

The cation exchange capacity (CEC) increased 10–60 per cent as a result of deferration of micaceous vermiculites and soils. Treatment of the deferrated sample with H2O2 restored the Fe3+ content to approximately the original value but the CEC was not affected. Consequently, the increase in CEC on deferration was attributed to the removal of the positively charged sesquioxide coating. The reversible change in valence of structural iron without an equivalent change in CEC was attributed to deprotonation-protonation of the structure (OH− ⇄ O2−) simultaneous with the oxidation-reduction of iron (Fe2+ ⇄ Fe3+) in the phyllosilicate layer.

Mafic chlorite from Benton, Arkansas was comminuted by rotary blending of a suspension, and the — 2 μm fraction separated by sedimentation in H2O. Droplets of suspension of the < 2 μm fraction were dried on a layer of Epoxy resin and then additional Epoxy was added and heat-cured at 48°C to form a resin sandwich. Cross-sections of 600–900 Å thickness were cut on a Reichert automated ultramicrotome. The sections were collected on standard electron microscope specimen screens, reinforced by vacuum evaporated C and examined by transmission electron microscopy (TEM). The Phillips EM 200 electron microscope was equipped with a “microgun” source to minimize heating of the specimen and to improve contrast and high resolution (HREM). Images of the (001) chlorite crystallographic planes spaced at 13·9Å intervals were visible on many of the particle sections. Imaging of the planes depended upon their being nearly parallel to the electron beam (within 0° 10’) and therefore, many particles which had other orientations did not show the 13·9Å image. Micrographs made before appreciable irradiation by the electron beam revealed images of fringes corresponding to the 7·22Å (002) spacing of chlorite. Loss of the 7·22 Å fringes and reinforcement of those at 13·9 A resulted from heating of the chlorite in the electron beam. This behavior is analogous to the well-known crystallographic effects of heating chlorite at 550–760°C.

The shrinkage of osmotically swollen natural and artificial blisters on vermiculite cleavages by exchange saturation with fixing cations such as Cs+, Rb+, NH4+, and K+ was investigated by replica electron microscopy. Incomplete collapse of either the natural or artificially produced blisters occurred with Cs+, Rb+, and NH4+ saturation, while K+ saturation completely collapsed the artificially produced blisters but not the natural blisters. The reason for incomplete collapse with Cs+, Rb+ and NH4+ was the incomplete replacement (trapping in the flakes) of interlayer hydrated cations such as Na+ shown by electron probe microanalysis. Much less trapping occurred with K+ saturation. Na+ entrapment increased with increasing size and decreasing hydration of cations, i.e. Cs+ >Rb+ >NH4+ >K+.

Semiquantitative determination of Na+, by electron probe microanalysis, in vermiculite flakes near the edge revealed that 1 N CsCl entrapped as much as 45·6 per cent while 1 N KCl entrapped only 7·5 per cent. In general, more Na+ was entrapped by 1 N solutions than by dilute solutions. With 0·01 N KCl solution, the Na+ entrapment was only 4·4 per cent. The amount of Na+ at the center of the macroflakes was less than at the edge, apparently as a result of more CEC at frayed edges and (or) because of incomplete diffusion of Na+ to the center. Shrinkage of artificial blisters by K+ could thus be attributed to its more effective removal of the interlayer hydrated cations, whereas the other fixing cations were less effective. Natural blisters on vermiculite from Libby, Montana were not completely collapsed even by K+, apparently because the layer charge density was too low in the blister areas.

Scanning electron microscopic (SEM) examination of the quartz isolated from chert of Transvaal and Swaziland, of 2000 and 3400 million year (MY) ages, respectively, in southern Africa revealed marked differences in quartz morphology. Well-defined individual euhedral quartz crystals, with polyhedral triple-point faces, were clearly evident on freshly fractured Transvaal chert surfaces as well as with the quartz isolates from the chert. The morphology and coarseness suggest crystal growth with little, if any, metamorphism; however, the δ18O values of 23.8–24.1‰ suggest crystallization temperatures of perhaps 40–45°C. In contrast, fracture surfaces of the older, strongly metamorphosed Swaziland cherts revealed a high degree of grain intergrowth which inhibited fracture between quartz grains. Quartz isolates from these showed strongly interlocked quartz crystal clusters and elongated chips of quartz with poorly defined irregular faces. Intercalation of mineral veins in the cherts on a mm scale and the intergrown character of the quartz grain boundaries provide evidence that the latter cherts have been strongly metamorphosed and recrystallized, in keeping with 14.6–15.1‰ δ18O values, corresponding to 80°C fractionation with water. The SEM micrographs of the fine quartz (1–10 µm) isolated from the Otavi dolomite formation of the 700-MY Damara System and from the 2000-MY Transvaal Dolomite Series revealed well-defined subhedral and euhedral quartz crystals of small size which, together with the 26.9–27.8‰ and 23.8‰ δ18O values, respectively, indicate that these dolomites have been affected little, if any, by post-depositional metamorphism; their crystallization temperatures fall in the range of 25–30 and 40–45°C, respectively.

Clay mineral associations in saprolite of two andesites from the Cascade Range of northeastern California were determined. Sesquioxidic allophane with a high CEC delta value dominates the clay fraction of the least weathered saprolite in each series (47% and 37% in hypersthene andesite and olivine andesite saprolites, respectively). With further weathering, the content of amorphous clay remains high (over 30% in all cases) but the CEC delta value of the clay drops markedly. The amorphous material in the more weathered saprolite has the properties of halloysitic allophane. Halloysite, present in all saprolites, is highest in concentration (over 30%) in the more strongly weathered members of each of the saprolite series. Formation of sesquioxidic allophane during early stages of weathering and its transformation to halloysitic allophane and halloysite during later stages of weathering are supported by X-ray diffraction, electron microscopic, DTA, elemental analysis, and CEC delta value data.

Three kinds of opal-cristobalite, differentiated by the sharpness of the 4·1 Å XRD peak, were isolated from the Helms (Texas) bentonite by selective chemical dissolution followed by specific gravity separation. The δ18O value (oxygen isotope abundance) for these cristobalite isolates ranged from approximately 26–30‰ (parts per thousand), increasing with increased breadth of the 4·1 Å XRD peak. Opal-cristobalite isolated from the Monterey diatomite had a δ18O value of 34‰. These δ18O values are in the range for Cretaceous cherts (approximately 32‰) and are unlike the values of 9–11‰ obtained for low-cristobalite (XRD peaks at 4·05, 3·13, 2·4, and 2·49) formed hydrothermally or isolated from the vesicles of obsidian. The morphology pseudomorphic after diatoms, observed with the scanning electron microscope, was more apparent in the opal-cristobalite from the Monterey diatomite of Miocene age (approximately 10 million yr old) than in the spongy textured opal-cristobalite from the Helms bentonite, reflecting the 40 million yr available for crystallization since Upper Eocene.

The oxygen isotope abundance of Helms montmorillonite (δ18O = 26‰) indicates that it was formed in sea water while the δ18O values of the associated opal-cristobalite indicate that this SiO2 polymorph probably formed at approximately 25°C in meteoric water. Although both cristobalite and mont-montmorillonite in the bentonite were authigenic, the crystallization of the SiO2 phase apparently required a considerably longer period and occurred mainly after tectonic uplift.

In contrast to the results for cristobalite, quartz from the Helms and Upton (Wyoming) bentonites had δ18O values of 15 and 21‰ respectively. Such intermediate values, similar to those of aerosolic dusts of the Northern Hemisphere, loess, and many fluvial sediments and shales of the North Central United States (U.S.A.), preclude either a completely authigenic or a completely igneous origin for the quartz. These values probably result from a mixing of quartz from high and low temperature sources, detritally added to the ash or bentonite bed.

Hexafluorotitanic acid (H2TiF6) selectively dissolves kaolinite and most other phyllosilicate minerals of soils and sediments, concentrating free crystalline (Ti,Fe)O2 minerals (partially substituted anatase and rutile) in the residue. A series of H2TiF6 reagents was standardized by analysis of the Ti content and by tests with pure anatase and commercial kaolinites. The Ti in the H2TiF6 solution selected (made from 49% HF + reagent TiO2) was 16·5% by weight as analyzed by the Tiron method. Treatment of pure anatase with the reagent H2TiF6 resulted in a 98% by weight recovery of TiO2 in the residue. The fraction of TiO2 recovered in the residue of commercial Georgia kaolinites was 88–101% after treatment with the selected H2TiF6 reagent. Isolates from nine Georgia kaolinite samples with varying amounts of TiO2 and Fe2O3 were examined by X-ray powder diffraction, scanning electron microscopy and elemental analysis. The main constituent of the (Ti,Fe)O2 isolates was anatase for all samples, with minor amounts of coarser rutile and mica from coarser kaolinite. The anatase and rutile isolates contained 74–93% (Ti,Fe)O2 with 0·5–3·1% Fe. The other constituents of the isolates were muscovite of mica (0·3–7%), quartz (0–9%) and amorphous relics of vermiculite and/or kaolinite (6–19%). Rutile, muscovite and quartz appear to be detrital but the anatase and relics are probably authigenic. Fine anatase appears to stick on the muscovite flakes as revealed by scanning electron microscopy and heavy liquid data for separation of these two minerals. The (Ti,Fe)O2 isolates from kaolinites which passed with the first magnetic concentrate of anatase were coarse, on the order of a few microns dia., as revealed by the scanning electron microscopy. Those passed with subsequent extensive magnetic concentrates from the same samples were finer. The anatase isolated from kaolinite purified by removal of as much of the impurities as possible by magnetic means was extremely fine, most of the particles being on the order of 0·1 µm dia. More than one third of the total Fe2O3 in kaolinites magnetically separated in the first pass was extracted by the citrate-bicarbonate-dithionite treatment after hot NaOH dissolution of 52–74% of the kaolinite, showing that the Fe2O3 had been mainly associated within the kaolinite. Only 2–6% of the total Fe2O3 was extracted from magnetically purified kaolinite after 40–50% of this kaolinite had been dissolved, indicating that most of the Fe is in the anatase and rutile fraction.

From 2 to 28% opal-cristobalite was isolated from the 2–20 µm fraction of rhyolitic and andesitic tuffaceous pyroclastics from the Island of Honshu, Japan, where it had been formed in hydrothermal springs at temperatures of ∼25–170°C as calculated from the oxygen isotopic ratios (18O/16O). Three of the isolates gave X-ray powder diffractograms with strong peaks at 4.07 Å. Two of these also had very weak peaks at 4.32 Å indicative of the presence of traces of tridymite. The fourth isolate had a strong 4.11 Å cristobalite peak and a very weak 4.32 Å peak. The morphology, determined by the scanning electron microscope, varied with the formation temperature indicated by the oxygen isotopic ratio (δ18O), from spheroidal and spongy for the opal-cristobalite formed at ∼25°C (δ18O = 26.0‰) in contrast to angular irregular plates and prisms for that formed at ∼115°C (11.9‰), ∼135°C (7.9 ‰) and ∼170°C (6.8 ‰). The differences in δ18O values are attributed to variation in hydrothermal temperature, but some variability in oxygen isotopic composition of the water is possible. The field-measured temperatures related roughly with the calculated fractionation temperatures except in one site, while the contrast in cristobalite morphology related well to calculated low and high fractionation temperatures. Low-cristobalite of hydrothermal origin in New Zealand (δ18O = 9‰) had characteristic rounded grains with some evidence of platiness. Co-existing quartz grains (δ18O = 10‰) showed more subhedral and irregular prismatic morphology.

Micaceous clays (“illites”) of a number of sediments, subjected to quantitative mineralogical analysis, ranged in different samples from 58 to 65% of a dioctahedral mica of muscovite type, from 3 to 8% vermiculite, from 0 to 10% montmorillonite, from 2 to 12% quartz, from 0 to 4% feldspars, from 1 to 2% rutile and anatase, from 0.3 to 1.2% free iron oxides, and from 16 to 30% chlorite. The soil clays contained less mica (13 to 35%) but more vermiculite (13 to 16%), montmorillonite (6 to 14%) and chlorite (30 to 33%), besides kaolinite (3 to 8%) and amorphous material (6 to 15%). The presence of the various minerals was verified by X-ray diffraction. When the interlayer K of these micaceous clays was replaced by hydrated Na ions through treatment with 0.2 n sodium tetraphenylboron in 1 n NaCl, the analytically determined vermiculite and montmorillonite contents increased. The freshly produced vermiculites had an interlayer charge of 177 to 198 me/100 g. During this replacement of interlayer K in a 1 N salt solution, the pH of the equilibrium matrix solution was observed to increase instead of decreasing as would be expected if oxonium (OH3+) ions were associated with K in the interlayer space, thus precluding the possibility of interlayer ONz+. The calculated stoichiometric equivalent of the pH rise established a significant uptake of protons by the lattice as K was released.

Thermal analyses of vermiculites show 0.5 to 0.7% higher OH water content than the parent micas. Oxidation of ferrous iron in the silicate structure, suggested as a possible mechanism for lowering of layer charge during weathering of mica to vermiculite, accounted for about 66% of this in samples high in iron (16.0% FeO) and only 30% of this in samples fairly low in iron. The rate of extraction of K from muscovite with no ferrous iron was found to be pH dependent, increasing at lower pH values. The data indicate proton incorporation with structural oxygens of the silicate sheet to form OH as a charge reduction mechanism in weathering of mica, independent of and additional to that involving the oxidation of ferrous iron.