TRI-OCTAHEDRAL DOMAINS IN SYNTHETIC CLAYS: IMPLICATIONS FOR LACUSTRINE PALEOENVIRONMENTAL RECONSTRUCTION

Sepiolite (Mg₄Si₆O₁₅(OH)₂·6H₂O) occurs as an authigenic precipitate in a number of lake and wetland settings, such as in the Amboseli Basin in Kenya and Tanzania (Hay et al., 1995). The proportion of authigenic to detrital clay minerals (e.g. Al-rich smectite) in terrestrial sediments is variable, and it has previously been hypothesized that pure Mg-silicates in regions such as Amboseli occur due to the absence of Al-rich detritus. We are examining this issue by preparing Mg-silicate saturated solutions containing no detrital material and then containing variable amounts of Al-rich smectite (Clay Minerals Society Source Clay SWy-2).

Preliminary results show the effects of adding Al-rich clay to Mg-silicate supersaturated solution. We replicated Holm-Denoma and Das’ (1991) sepiolite synthesis and prepared a Mg-silicate gel. Examination of the precipitate was performed using XRF and XRD and the molar Si:Mg ratio of the gel was determined to be 2.3 : 1. A broad diffraction hump from ~3.0 – 5.0 Å was produced, but the 060 peak associated with clay octahedral sheets was not apparent. The Mg-silicate gel was then seeded with a purified sub-micron sepiolite standard (Sep-sp-1), sonicated, and heated to 200° C for 50 hours in a Teflon lined Parr bomb. This produced a precipitate with no key differences from Sep-sp-1 by oriented XRD, but a slight shift of the Sep-sp-1 060 peak was observed from ~ 1.549 to 1.516 Å in randomly oriented mounts. The experiment was then repeated with the addition of Na-montmorillonite (SWy-2) in place of Sep-sp-1. The oriented pattern produced two broad humps at 4.51 Å and 2.56 Å, and a new peak at ~1.526 Å was produced, contrasting with the original SWy-2 dioctahedral 060 peak (1.492 – 1.504 Å). This suggests that we synthesized trioctahedral domains in a smectite structure.

Investigation is continuing to characterize these newly synthesized phases and determine the compositional and textural relationship to the Al-rich detritus. We are also altering the amount of aqueous Mg-silicate gel present in the experiments to determine whether the presence of Al-rich detritus is more favorable for the precipitation of Mg-silicates. Future precipitates will be characterized by XRD, XRF, SEM, TEM, AA and ICP-AES.