Introduction

There has been considerable research interest in studies of diffusive ion transport through the saturated pore space of portland cement paste. Measurements of conductivity have provided a useful insight into characterizing relevant transport properties [1, 2, 3, 4, 5, 6, 7, 8]. This effort has been motivated by the direct relationship between the concrete conductivity and both the steel reinforcement corrosion rate and the ion diffusivity. The concrete conductivity is directly proportional to the reinforcement corrosion current, and it can also be related to the ion diffusivity through the Nernst-Einstein equation for electrical potential gradients or the Nernst-Plank equation for both electrical and chemical potential gradients (see reference [9], Chapter 4). Therefore, concrete conductivity can be an important transport property in determining concrete service life in corrosive environments.

Both AASHTO (T 277) and ASTM (C 1202) have a standardized test of electrical conduction through concrete, referred to here as the rapid chloride test (RCT). This test measures the cumulative electrical charge passing through a specimen subjected to a direct current (DC) potential of 60 volts over a period of six hours. However, changes in the pore fluid conductivity due to ohmic heating [10, 11], and changes in the microstructure due to electromigration [12] prevent the standardized six hour test from yielding a direct measure of specimen DC conductivity and diffusivity.

The work described herein explores whether specimen diffusivity can be determined from some other aspect of the RCT. Experiments have been performed elsewhere that demonstrate the direct relationship between specimen conductivity and diffusivity [13, 14, 15]. Experiments have also been performed which purport to demonstrate a causal relationship between measurements of diffusivity and either the total charge passed or the initial current using the RCT cell [16, 17, 18, 19, 20, 22]. However, no direct relationships between parameters of the RCT and diffusivity have been proven. Therefore, a means to accurately determine specimen conductivity from RCT data would be a crucial step towards establishing a relationship between RCT data and specimen diffusivity.

Accurate measurements of specimen conductivity can be made using the experimental techniques of impedance spectroscopy (IS). Using an alternating current (AC) that varies over a wide range of frequencies, effects such as electrode polarization and ion transfer can be eliminated, yielding the true specimen resistance. The specimen conductivity may then be calculated from the measured resistance and a knowledge of the specimen geometry. However, under certain conditions, these electrode effects may be negligible. There is a small range of frequencies over which the electrode effects have an insignificant contribution to the overall conductivity, leading a number of investigators to report specimen impedance at a fixed frequency. Further, it may be possible to estimate specimen conductivity in a situation where the electrode effects are relatively small compared to the applied potential. Experiments have shown that the combined electrode effects generate a voltage drop of less than two volts [3, 23]. Therefore, one may conjecture that if a specimen is subjected to a DC electrical potential of 60 volts, and if the combined electrode effects generate a negligible voltage drop, the specimen conductivity calculated from the resultant DC current may be sufficiently accurate for most purposes.

Reported herein are the results from IS measurements performed on specimens just prior to the application of the 60 volts specified by the RCT. A frequency spectrum of 10 Hz to 1 MHz was used to determine the true specimen resistance, from which the conductivity was calculated. The RCT setup used here recorded the current at 60 second intervals, including the instantaneous initial current, until the completion of the six hour test. Results show that the initial RCT current can be used to directly and accurately determine the specimen conductivity. Implications of this result for possible future "rapid" tests are discussed.