Tests of backscatter coefficient measurement using broadband pulses.

An adaptation to a data reduction method is outlined for determining backscatter coefficients, eta, when broad-bandwidth pulses are employed. The accuracy of these eta values is assessed with well-characterized phantoms, for which backscatter coefficients based on their physical properties have been independently calculated. One phantom produces Rayleigh-like scattering, where the backscatter coefficient varies smoothly with frequency over the analysis bandwidth. A second phantom exhibits local maxima and minima in the scattering function versus frequency due to the presence of millimeter-sized graphite gel spheres in a gel background. The method was found to produce accurate results using time gate durations as small as 2 mus, although better accuracy is obtained for longer gate durations, particularly when the sample exhibits resonance peaks in backscatter versus frequency. Use of a Hamming window in place of a rectangular window extends the accuracy near the upper and lower limits of the frequency range.