Characterization of fiber optic sensor for corrosion monitoring

Abstract

This research was conducted to characterize the parameter that can be associated with corrosion activity of metals using fiber optics as sensor. Three samples of metal rods were selected namely iron, copper and aluminium. Unclad multimode fiber optics were strapped onto the metal and soaked in corrosive solution. The corrosion process was monitored by a spectrometer linked to a computer. Two measurement techniques were used in this research namely transmission and reflection. For transmission technique, the changes in the intensity spectrum of light from a white LED source after passing through the unclad region (corrosion activity region) at different exposure period were monitored. For reflection technique, a Y-probe fiber bundle was used to observe the reflected intensity from the corroded region. The changes in colour of the corrosion product had an effect on the reflected component from the light source. The changes in the intensity spectrum for both techniques were monitored at certain time intervals. Generally the result showed a reduction in the intensity level with exposure duration for both techniques but fluctuations also occurred at some exposures. This is thought to be due to the nature of the corrosion process itself. When the ratios of the intensities peaks were plotted with exposure, a very distinct drop in the value for iron occurred for both the reflection and transmission spectra from around days 37 to 44. This was the period where corrosion activity actually started. After that period, the value began to level off for reflection measurement but slowly increased for transmission measurement. For copper and aluminium, the intensity ratios from the reflected spectrum indicated a more gradual drop. These proved that they were more resistant to corrosion compared to iron. It seemed that the ratio of the peaks from the intensity spectrum was the parameter suitable to be associated with the onset of the corrosion activity of iron.