Accession Number:



Nonlinear Electromagnetic Scattering Techniques for the Detection and Characterization of Closed Cracks.

Descriptive Note:

Final rept. 1 Oct 80-30 Sep 81,

Corporate Author:


Report Date:


Pagination or Media Count:



This study seeks to develop an improved nondestructive evaluation technique for detecting and characterizing tightly closed cracks on surfaces of metal parts. This technique is based on exploiting the nonlinear eddy-current interaction that would occur if a nonlinear junction existed between the faces of a closed crack. Both second- and third-harmonic signals in the 1- to 3-GHz range have been detected emanating from cracks in flat samples of aluminum and copper. The third-harmonic signals were typically four orders of magntidue larger than the second-harmonic signals. These results suggest that the dominant physical mechanism involved is electron tunneling in a metal-oxide-metal junction. This crack detection technique is not as reliable as desired because it is difficult to keep spurious background levels from masking the desired crack signals, and because nonlinear junctions required for detection do not always form in the crack. However, experiments conducted using modified aluminum compact-tension specimens suggest that monitoring of nonlinear crack signals can yield information on the rate of crack growth. A simple model for a nonlinear crack is described. This model shows that the nonlinear signal produced by a crack is strongly dependent on the impedances of the crack at the measurement frequencies, and particularly on the area and thickness of the nonlinear junction. Calculations using this model indicate that the use of frequencies in the low microwave range should produce the largest nonlinear signals. However, quantitative agreement between these calculations and the experimental results is not very good, and so a better model probably stochastic is needed.

Subject Categories:

  • Metallurgy and Metallography
  • Test Facilities, Equipment and Methods
  • Optics

Distribution Statement: