Thickness and conductivity of metallic layers from pulsed eddy-current measurements

We describe a time-domain (pulsed) eddy-current technique for determining the thickness and conductivity of conductive coatings on metal plates. The pulsed eddy-current instrument records the transient current induced in an absolute, air-cored coil placed next to a layered sample and excited with a step-function change in voltage. Signals are digitized with 16-bit resolution at a sampling rate of 1 megasamples per second, and the excitation is repeated at a rate of 1 kHz. The instrument displays the difference in the transient current measured on the substrate and on the substrate plus coating. We measured pulsed eddy-current signals for a series of metal foils of varying thickness placed over 1 cm thick metal plates. Seven combinations of foil and substrate metals were studied including pure aluminum, copper, and titanium foils over substrates of aluminum, titanium alloy, and stainless steel. We report results for three types of samples: aluminum foils on Ti-6Al-4V substrate, titanium foils on 7075 aluminum alloys, and aluminum foils on AISI 304 stainless steel. Foil thickness ranged from 0.04-1.00 mm. We found that three features of the signal - the peak height, the time of occurrence of the first peak, and a characteristic zero-crossing time - depend sensitively upon the thickness of the layers and the relative electrical conductivity of coating and substrate. Theoretical calculations were compared to the measurements. Absolute agreement between calculated and measured signals was, in most cases, within 3%. No calibration with respect to artifact standards was used. Finally, a feature-based rapid inversion method was developed and used to infer the thickness and conductivity of the layers. The accuracy of the inversion depends upon the thickness of the layer and the contrast in conductivity between layer and substrate. For the materials studied the thickness could be determined within 13%, while the error in determining conductivity was 20%-30%. The time-domain method is much simpler and hundreds of times faster than the frequency-domain method previously reported by Moulder et al.