Assessing the high frequency behavior of non-polarizable electrodes for spectral induced polarization measurements

Abstract

During the last decades, the usage of spectral induced polarization (SIP) measurements in hydrogeology and detecting environmental problems has been extensively increased. However, the physical mechanisms which are responsible for the induced polarization response over the usual frequency range (typically 1mHz to 10–20kHz) require better understanding. The phase shift observed at high frequencies is sometimes attributed to the so-called Maxwell–Wagner polarization which takes place when charges cross an interface. However, SIP measurements of tap water show a phase shift at frequencies higher than 1kHz, where no Maxwell–Wagner polarization may occur. In this paper, we enlighten the possible origin of this phase shift and deduce its likely relationship with the types of the measuring electrodes. SIP Laboratory measurements of tap water using different types of measuring electrodes (polarizable and non-polarizable electrodes) are carried out to detect the origin of the phase shift at high frequencies and the influence of the measuring electrodes types on the observed complex resistivity. Sodium chloride is used to change the conductivity of the medium in order to quantify the solution conductivity role. The results of these measurements are clearly showing the impact of the measuring electrodes type on the measured phase spectrum while the influence on the amplitude spectrum is negligible. The phenomenon appearing on the phase spectrum at high frequency (>1kHz) whatever the electrode type is, the phase shows an increase compared to the theoretical response, and the discrepancy (at least in absolute value) increases with frequency, but it is less severe when medium conductivity is larger. Additionally, the frequency corner is shifted upward in frequency. The dependence of this phenomenon on the conductivity and the measuring electrodes type (electrode–electrolyte interface) seems to be due to some dielectric effects (as an electrical double layer of small relaxation time formed at the electrodes interface). Therefore, this dielectric response should be taken into account at high frequency to better analytically separate the medium own response from that linked to the measuring electrodes used. We modeled this effect by adding a capacitance connected in parallel with the traditional equivalent electric circuit used to describe the dielectric response of medium.