Electrochemical noise data (ECN) have been collected for three active systems (mild steel/NaCl, A1 6061/NaCI and A1 2024/NAC1) and a passive system (Ti-6A1-4V/Ringer's solution). The ECN data have been analyzed in the time and frequency domains. Parameters such as the noise resistance R n and the skewness and kurtosis of potential and current fluctuations have been determined from the analysis in the time domain. It has been concluded that the localization index (LI) is not related to a particular corrosion mechanism, but reflects the degree of similarity of the corrosion kinetics of the two electrodes used for the electrochemical noise measurement (ECNM). Comparisons of noise impedance spectra with traditional impedance spectra have been made. Good agreement has been observed for the passive system, but not for the active systems. Comparisons have also been made between R n and the polarization resistance Rp obtained from analysis of the impedance spectra. It has been observed that R n generally is a function of the bandwidth Af of the ECNM. For the passive system R n was much smaller than Rp. The relationship between potential and current fluctuations has been evaluated for all four systems. For the two A1 alloys random fluctuations occurred, while for mild steel a systematic trend of potential vs. current was observed. For the Ti alloy noise from the measurement system seemed to dominate the ECN data.
The ECN technique is becoming increasingly popular due to some of its perceived advantages over other electrochemical techniques. These advantages include the low cost of the equipment, which consists mainly of a zero resistance ammeter (ZRA) and digital voltmeters, and the ease of data collection. It is assumed sometimes that valid ECN measurements can be carried out in a much shorter time than that required to measure an impedance spectrum. Furthermore it has been suggested that certain parameters derived from the statistical analysis of ECN data such as the localization index LI provide mechanistic information. It is becoming clear however that some of these advantages do not exist. For example, previous work in this laboratory has shown that the noise resistance R n depends on the sampling frequency fs and the bandwidth Af of the electrochemical noise measurement (ECNM) [1-3]. For passive systems such as stainless steel and a Ti alloy R n was found to be much smaller than the polarization resistance Rp obtained from analysis of the impedance spectra [41. Rn increased with decreasing fs. Similar results were obtained for polymer coated steel exposed to seawater [1,2]. In order to obtain better agreement between R n and Rp it is often necessary to collect ECN data for long time periods. In general R n is expected to agree with Rp only for systems with high corrosion rates for which the impedance spectrum had reached the de limit within zXf [ 1-41. Analysis of ECN data for mild steel and Ti-6A1-4V has shown that LI is not related to the prevailing corrosion mechanism, but reflects instead the differences between the corrosion kinetics of the two electrodes used for the ECNM [5]. In this presentation additional ECN data and their analysis will be presented for several active systems and for a passive system. Comparisons will be made with impedance spectra collected for these systems under the same exposure conditions. The relationship between R n and Rp will be evaluated and the significance of LI as a tool for identifying corrosion mechanisms will be tested further. In addition, the effect of trends in the ECN data on data analysis will be discussed.
EXPERIMENTAL APPROACH
Materials and Methods
Materials. Mild steel, AI 2024-T3 and A16061 were exposed to 0.5