Original paper

The formation of agate structures: models for silica transport, agate layer accretion, and for flow patterns and flow regimes in infiltration channels

Walger, Eckart; Mattheß, Georgvon; Seckendorff, Volker; Liebau, Friedrich


In agate structures three substructures can be discerned: common agate banding, infiltration channel banding, and horizontal layering. Infiltration channel banding has so many specific features that it cannot be created by a process involving a high degree of randomness, such as deformation. Infiltration channels are indeed inflow structures that formed on the interior side of the very first chalcedony layer. Both, the common agate banding and the infiltration channel banding are generated by two different processes of silica transport, that depend upon the same general mechanisms: A silica concentration gradient is generated and maintained, directed from the intergranular solution of the cavity-surrounding rock into the cavity, in which the agate structure is building up. The driving force is the osmotic pressure difference caused by the semipermeability of the currently youngest, innermost silica gel membrane. The very first gel layer (membrane) that immediately lines the inner surface of the cavity has a special role. (i) If the very first gel layer remains intact, common agate banding is formed: silica is transported by diffusion down the concentration gradient directly into the cavity (direct process). (ii) If the very first gel layer is penetrated by one or more capillary fissures, infiltration channel banding is formed: silica is transported by injection of solution through capillary fissures into the cavity (injection process). Both these processes contain a relaxation mechanism ('internal rhythm' reflected by the agate banding), caused by the incommensurable time scales of deposition and ageing of the silica gel. The depleted solvent is recycled back to the intergranular space of the surrounding rock through the total outer surface of the cavity (continuity condition). To test the hydrodynamic plausibility of the model, flow patterns were computed, (A) based on Schlichting's model of the 'plane laminar jet' and (B) for Hagen-Poiseuille flow supposed to be fully developed in the fissure. In the vicinity of the entrance point the strongly accelerated flow diminishes the probability of accretion continuously towards the entrance point of the jet, where it is zero, reflected by the striking elegance with which the agate bands generally thin out in this direction. The combined model predicts entrance velocities, which are so low that the flow is dissipated to Brownian motion within an order of magnitude of the cavity dimensions. The calculated transport rate of silica into the cavity yields durations of formation between decades and thousands of years, in accordance with geological field experience. As full agate (completely silicified) is the normal case and porous biscuit agate occurs rarely, the model proposed contains the idea that during the transition from biscuit agate to full agate the densification of the structure is caused by the incorporation of moganite. The densification of each single layer is finished, before the deposition of the next layer starts.