28 November 2016 in Geophysical Journal International
We performed complex conductivity measurements on 28 core samples from the hole drilled for the Humu'ula Groundwater Research Project (Hawai'i Island, HI, USA). The complex conductivity measurements were performed at 4 different pore water conductivities (0.07, 0.5, 1.0 or 2.0, and 10 S m−1 prepared with NaCl) over the frequency range 1 mHz to 45 kHz at 22 ± 1 °C. The in-phase conductivity data are plotted against the pore water conductivity to determine, sample by sample, the intrinsic formation factor and the surface conductivity. The intrinsic formation factor is related to porosity by Archie's law with an average value of the cementation exponent m of 2.45, indicating that only a small fraction of the connected pore space controls the transport properties. Both the surface and quadrature conductivities are found to be linearly related to the cation exchange capacity of the material, which was measured with the cobalt hexamine chloride method. Surface and quadrature conductivities are found to be proportional to each other like for sedimentary siliclastic rocks. A Stern layer polarization model is used to explain these experimental results. Despite the fact that the samples contain some magnetite (up to 5 per cent wt.), we were not able to identify the effect of this mineral on the complex conductivity spectra. These results are very encouraging in showing that galvanometric induced polarization measurements can be used in volcanic areas to separate the bulk from the surface conductivity and therefore to define some alteration attributes. Such a goal cannot be achieved with resistivity alone.
18 February 2016 in Geophysics
Induced polarization can be used to estimate surface conductivity by assuming a universal linear relationship between the surface and quadrature conductivities of porous media. However, this assumption has not yet been justified for conditions covering a broad range of fluid conductivities. We have performed complex conductivity measurements on Portland sandstone, an illite- and kaolinite-rich sandstone, at 13 different water salinities (NaCl) over the frequency range of 0.1 Hz to 45 kHz. The conductivity of the pore water affected the complex surface conductivity mainly by changing the tortuosity of the conduction paths in the pore network from high to low salinities. As the fluid conductivity decreases, the magnitude of the surface conductivity and quadrature conductivity was observed to decrease. At relatively high salinities (), the ratio between the surface conductivity and quadrature conductivity was roughly constant. At low salinities (), the ratio decreased slightly with the decrease of the salinity. A Stern layer polarization model was combined with the differential effective medium (DEM) theory to describe this behavior. The tortuosity entering the complex surface conductivity was salinity dependent following the prediction of the DEM theory. At high salinity, it reached the value of the bulk tortuosity of the pore space given by the product of the intrinsic formation factor and the connected porosity. The relaxation time distributions were also obtained at different salinities by inverting the measured spectra using a Warburg decomposition. The mode of the relaxation time probability distribution found a small but clear dependence on the salinity. This salinity dependence can be explained by considering the ions exchange between Stern and diffuse layers during polarization of the former. The pore-size distribution obtained from the distribution of the relaxation time agreed with the pore-size distribution from nuclear magnetic resonance measurements.
1 March 2015 in Journal of Hydrology
1 December 2016 in Geophysical Journal International
We investigate the relationship between complex conductivity spectra and both permeability and pore mean size and distribution of 22 core samples (21 volcanic rocks and 1 clayey sandstone). The volcanic core samples were extracted from a wellbore drilled for the Humu‘ula Groundwater Research Project in the Humu‘ula saddle region between Mauna Kea and Mauna Loa volcanoes (Hawaii). The quadrature conductivity spectra of volcanic rocks exhibit a subtle, but generally detectable, relaxation frequency in the range 0.3 Hz to 45 kHz similar to the relaxation frequency observed for clayey sandstones. We find a fair relationship between this relaxation frequency and the pore size determined by mercury porosimetry. Combined with the intrinsic formation factor of the core samples, the relaxation frequency can be used as an indicator of the permeability of the material. The predicted values of the permeability are grossly consistent with the permeability values to air (in the range 0.001–100 mD) within two orders of magnitude. The measured permeability values are highly correlated to the peak of the pore size distribution determined from mercury porosimetry divided by the intrinsic formation factor. By fitting the complex conductivity spectra with the pore size distribution, we obtain the normalized chargeability of the core samples, which is, in turn, highly correlated to the measured cation exchange capacity.
15 April 2019 in Water Resources Research
In this study, we focus on the electrical tortuosity‐based permeability model k = reff2/8F (reff is an effective pore size, and F is the formation factor) and analyze its applicability to rocks experiencing mineral precipitation and dissolution. Two limiting cases of advection‐dominated water‐rock reactions are simulated, that is, the reaction‐limited and transport‐limited cases. At the pore scale, the two precipitation/dissolution patterns are simulated with a geometrical model and a phenomenological model. The fluid and electric flows in the rocks are simulated by directly solving the linear Stokes equation and Laplace equation on the representative elementary volume of the samples. The numerical results show that evolutions of k and F differ significantly in the two limiting cases. In general, the reaction‐limited precipitation/dissolution would result in a smooth variation of k and F, which can be roughly modeled with a power function of porosity ϕ with a constant exponent. In contrast, the transport‐limited precipitation/dissolution mostly occurs near the pore throats where the fluid velocity is high. This induces a sharp change in k and F despite a minor variation in ϕ. The commonly used power laws with constant exponents are not able to describe such variations. The results also reveal that the electrical tortuosity‐based permeability prediction generally works well for rocks experiencing precipitation/dissolution if reff can be appropriately estimated, for example, with the electrical field normalized pore size Λ. The associated prediction errors are mainly due to the use of electrical tortuosity, which might be considerably larger than the true hydraulic tortuosity.
8 December 2015 in Geophysics
The measured spectral-induced polarization (SIP) response of highly porous granular materials can be modeled by upscaling the Nernst-Planck equation for a single grain and performing a convolution product with the particle size distribution. The distribution of relaxation times associated with the polarization of consolidated porous media seems, intuitively, to be controlled by the pore size distribution because of the radius of curvature of the interface between the solid and the pore water. We have explored a mechanistic model connecting the SIP response of consolidated media to mercury injection capillary porosimetry (MICP) data. First, we developed a mechanistic electric double layer (EDL) polarization model for porous media, which considered the polarization of the Stern layer, which was possibly enhanced by the polarization of the diffuse layer and the effect of the pore size distribution. The new model revealed that the relaxation time of porous media such as sandstones could be overestimated if the enhancement of the overall polarization by the diffuse layer was not considered. Second, we have developed an empirical method to obtain the pore size distribution from MICP data by assuming a linear relation between the pore size and the pore throat size using a pore aspect ratio. Then, we compared the measured quadrature conductivities of nine samples with the theoretical quadrature conductivity obtained by convoluting the pore size distribution with the Warburg model. The Warburg model is considered to represent the quadrature conductivity response of a consolidated porous material characterized by an infinitively narrow unimodal pore size distribution coated by an EDL. The position of the peak of the phase was well-reproduced from the MICP data. In addition, the calculated broadness of the phase spectra was usually well-reproduced by convolving the Warburg function with the pore size distribution.