Core KM-3 at Searles Lake, California, comprises 693.4 m of lacustrine sediments deposited over the past 3.2 m.y. The lake water evolved from moderately saline, slightly alkaline, and dominated by Na, Ca, Cl, SO4, and HCO3 + CO3 to a highly alkaline brine dominated by Na and CO, ions. Sediments are chiefly muds and evaporites. Montmorillonite and illite are the principal detrital clay minerals supplied to Searles Lake at present and probably during the late Pliocene and Pleistocene.

The drill core is divided into three diagenetic zones on the basis of clay-mineral reactions. The upper zone (0-291.1 m) contains authigenic Fe-illite, Mg-smectite, K-feldspar, and analcime, which average 60–70% of the <2-μm silicate fraction of mud samples. The principal silicate reactants are detrital montmorillonite and kaolinite, which have been largely consumed. The middle zone (291.1–541.6 m) also contains authigenic Fe-illite, Mg-smectite, K-feldspar, and analcime, but they form only 20 to 30% of the <2-μm silicate fraction, and considerable detrital montmorillonite and kaolinite remain unaltered. In muds of the lower zone (541.6–693.4 m), a small amount of clinoptilolite is the only authigenic silicate mineral identified, although authigenic montmorillonite probably coexists here as well. Vitric ash is the silicate reactant, and detrital clay minerals apparently remain unaltered in the lower zone.

Diagenetic zoning reflects the pore-water chemistry, in which pH may have been the most important parameter. Sediments of the upper zone were deposited in highly alkaline lakes of variable salinity, and sinking brines with a pH of 9.0–10.0 have saturated all sediments. Sediments of the middle zone were deposited in lakes of moderate to high salinity. The pH was overall lower than in the upper zone, although it probably exceeded 9.0, at least locally, during silicate diagenesis. Moderate salinity and a slightly alkaline pH (~ 7.5–8.0) are inferred for both the lake and pore water of the lower zone.

Oxygen-isotope values of authigenic Fe-illite, Mg-smectite, and K-feldspar, and phillipsite in the upper diagenetic zone reflect a high degree of evaporative concentration and presumably of salinity. Equilibrium water values calculated for 22°C from the oxygen-isotopic composition of authigenic phyllosilicates range from −3.5 to +1.9%, averaging −1.2% (SMOW). Higher salinities are suggested for K-feldspar, for which water values range from +1.8 to +4.8%, averaging +3.4‰ The water value for phillipsite is +0.3‰ By comparison, rainfall at Searles Lake has an average δ18O value of about −9.8%, and a brine sample has a value of +4.0‰

Mud samples of the middle and upper diagenetic zones tend to be rich in Fe-illite or Mg-smectite, but not in both, indicating that the two minerals are favored to some extent by different environments. Fe-illite seems to be generally favored by oxidizing conditions and probably by a playa environment, and Mg-smectite seems to be favored by reducing conditions and an open-water environment. One strati-graphic unit of the upper zone does not fit this pattern and contains abundant Fe-illite in sediments of a deep perennial lake.

Formation of authigenic trioctahedral Mg-rich smectite is common in evaporative lake sediments, but was not described previously in modern marine evaporative environments. This study documents formation of authigenic K-rich, Mg-smectite during very early diagenesis in the dominantly siliciclastic Salina Ometepec (Baja California), a large supratidal evaporative sabkha complex near the mouth of the Colorado River. Here, sediment pore waters are exceptionally Mg2+-rich relative to other marine evaporative environments due to suppressed sulfate reduction which limits production of carbonate alkalinity and, hence, carbonate (particularly dolomite) precipitation. Sediment cores were obtained along a five km transect seaward across the hypersaline mud flat to evaluate how these atypical geochemical conditions would affect the clay mineral compositions.

Scanning transmission electron microscopy (STEM) observations show that the smectite from the marine Inlet, near the sediment source, consists of grains of irregular shape that give selected area diffraction (SAED) patterns reflecting dominant turbostratic stacking. Analytical electron microscopy (AEM) analyses indicate that K+ is the dominant interlayer cation; the mean composition is approximately K0.7(Al3.3Fe(III)0.3Mg0.5)(Al0.5Si7.5)O20(OH)4. Such smectite is implied to be detrital in part because it is similar to smectite known to be deposited by the Colorado River.

Smectite from the hypersaline mud flat occurs as aggregates of small subhedral pseudohexagonal plate or lath-shaped crystals ≤250 nm in diameter, with thicknesses varying between three and ten layers. The SAED patterns reflect substantial turbostratic stacking, but with a greater frequency of interlayer coherency as compared with detrital smectite. Crystals from greater sediment depths are larger and more nearly euhedral. This smectite is dominantly trioctahedral, with mean composition approximately K0.7(Al0.7Fe(III)0.5Mg4.45)(Al1.2Si6.8)O20(OH)4 (saponitic). This smectite is inferred to be dominantly authigenic in origin.

The X-ray diffraction (XRD) and STEM/AEM data collectively imply that detrital aluminous dioctahedral smectite reacts to form authigenic Mg-rich trioctahedral smectite, driven in part by the high Mg2+/ Ca2+ ratio of pore waters. Such early-formed Mg-rich smectite may be the precursor for the trioctahedral mixed-layer smectite, corrensite, and chlorite assemblages found in ancient marine evaporative sequences. These results also add to the accumulating evidence that interlayer K+ in marine smectite is fixed during the earliest stages of marine diagenesis near the sediment water interface.

Smectite formation in alkaline-saline environments has been attributed to direct precipitation from solution and/or transformation from precursor minerals, but these mechanisms are not universally agreed upon in the literature. The objective of this work was to investigate the mineralogy of smectites in the soils surrounding a representative alkaline-saline lake of Nhecolândia, a sub-region of the Pantanal wetland, Brazil, and then to identify the mechanisms of their formation.

Soils were sampled along a toposequence and analyzed by X-ray diffraction, transmission electron microscopy-energy dispersive X-ray analysis, and inductively coupled plasma-mass spectrometry. Water was collected along a transect involving the studied toposequence and equilibrium diagrams were calculated using the databases PHREEQC and AQUA.

The fine-clay fraction is dominated by smectite, mica, and kaolinite. Smectites are concentrated at two places in the toposequence: an upper zone, which includes the soil horizons rarely reached by the lake-level variation; and a lower zone, which includes the surface horizon within the area of seasonal lake-level variation. Within the upper zone, the smectite is dioctahedral, rich in Al and Fe, and is classified as ferribeidellite. This phase is interstratified with mica and vermiculite and has an Fe content similar to that of the mica identified. These characteristics suggest that the ferribeidellite originates from transformation of micas and that vermiculite is an intermediate phase in this transformation. Within the lower zone, smectites are dominantly trioctahedral, Mg-rich, and are saponitic and stevensitic minerals. In addition, samples enriched in these minerals have much smaller rare-earth element (REE) contents than other soil samples. The water chemistry shows a geochemical control of Mg and saturation with respect to Mg-smectites in the more saline waters. The REE contents, water chemistry, and the presence of Mg-smectite where maximum evaporation is expected, suggest that saponitic and stevensitic minerals originate by chemical precipitation from the water column of the alkaline-saline lake.

Experimental alteration of a rhyolitic obsidian by solutions containing variable concentrations of Mg was carried out at 100, 150 and 200°C, for 30, 60 and 90 days, to investigate the mechanism of smectite formation. The smectite exhibits two distinct morphologies: (1) small flakes (aggregates of a few crystals); and (2) large flakes (massive groups). The small flakes are the earlier alteration products. Both morphological types have distinct chemical compositions: the smallest laminae are dioctahedral and contain more Fe, whereas the particles from the rose-shaped aggregates contain Mg and are trioctahedral. Intermediate compositions have been observed between the extreme compositions.

It is suggested that the two morphologies are the result of two distinct genetic processes: (1) the small flakes (Fe-rich smectite) form by solid state rearrangements of the hydrated external layer of glass and/or via the formation of domains within the glass; (2) the rose-shaped aggregates (Mg-rich smectite) form by precipitation from solution.