Pore structure of volcanic clasts: Measurements of permeability and electrical conductivity

The pore structure of volcanic clasts is examined using measurements of porosity, permeability, and electrical properties. Permeability varies by several orders of magnitude among volcanic clasts and does not depend solely upon porosity. Electrical property measurements of saturated volcanic samples illustrate the influence of pathway tortuosity and pore shape on permeability. For equivalent eruption conditions, silicic samples show higher tortuosities, smaller vesicle sizes, and lower permeabilities than mafic samples. These differences are largely due to variations in vesiculation and crystallization history. Differences between explosive and effusive samples reflect the relative ability of bubbles to form and maintain connected pathways during bubble expansion and collapse. Isotropic samples (variably expanded breadcrust bombs and most pumice fall samples) have pore pathways that simplify with increasing porosity. Highly vesicular anisotropic samples (e.g., tube pumice) have high permeabilities and low tortuosities parallel to pore elongation and low permeabilities and high tortuosities perpendicular to elongation. These pathways simplify with increasing deformation (i.e. tortuosity decreases as porosity decreases), until pore geometries collapse sufficiently to form intersecting cracks. More generally, Archieʹs Law (power law) relationships between electrical conductivity formation factor (F) and porosity (ϕ) have an Archieʹs exponent, m, between 1 and 4 (where F = ϕ− m) for low porosity volcanic clasts. However, samples with higher connected porosities (> 20% for silicic samples and > 50% for mafic samples) have m values that increase with increasing porosity, reaching up to 15. We also find that a single Archieʹs Law fit to a suite of samples is not appropriate either for sample suites with widely varying porosities or for anisotropic samples with a directional variation in measured properties. These measurements caution against simple application of cross-property relationships derived from sedimentary rocks to models of permeability in volcanic samples.