Semi-annual Progress Report Title: New Techniques in Nondestructive Testing by Acoustical and Exo-electron Emission Acorstical Emission Studifs

Deforming materials emit high-frequency acoustical waves. The mechanism or mechanisms responsible r^r the Initiation of these waves remains islatlvely unknown. Evidence has been produced to Indicate that the waves originate In the bulk of a body and are probably connected with the formation of mlcrocracks at stress concentrations and other crack nuclea-tlon sites. One of the most Important problems connected with making use of acoustical emission data Is the determination of the significance of the various sounds emitted. Studies so far have concentrated on recording the number of transient emissions occurring in a certain frequency range per unit time and correlating this Information with the mechanical state of the deforming specimen. The most popular frequency range seems to lie below 5 MHz. The purpose of the present investigation is to determine two points: (1) the upper range of the frequency spectrum of acoustical emissions and whether it is limited by the rapidly rising attenuation usually associated with rising frequency and (2) the existence of any recognizible peaks or valleys in the acoustical power spectrum Identifiable with a particular mechanism of acoustical emission, for example, with initial microcrack size.-2-A sample calculation shows that high-frequency stress waves should be expected from microcrack formation and propagation. Depending upon the nature of the stress concentration, frequencies as high as 100 MHz should be expected. Furthermore, as cracks extend, there should bo a corresponding decrease in the frequencies of the maximum in the power spectrum. In the present study, the acoustical output of quartz crystal transducers attached to deforming specimens has been analyzed by a Hewlett Packard spectrum analyzer. This method does not permit the frequency content of individual transient pulses (emission) to be measured. It can, however, measure the average frequency spectrum of several hundred pulses. At present, this experimental constraint seriously restricts the type of specimens that can be studied. More recent work, at present incomplete, has centered on a method of measuring the frequency spectrum of individual pulses. The Hewlett Packard spectrum analyzer is essentially a frequency swept, very high gain receiver. It has provisions for varying both the sweep frequency range and the local frequency of sweeping. Given the sweep range, only a limited number of sweeping rates are possible which will allow calibrated operation. Thus if a transient lasting only one millisecond Is Introduced into the receiver while sweeping at one millisecond per 1 MHz frequency range, then depending on …