The sample was a 1 cm square 2400Å thick YBa2Cu3O6.6 (001) film grown epitaxially on a MgO (001) substrate by pulsed laser deposition. The crystal was fully twinned in the ab plane and the c-axis mosaic spread was about FWHM. The x-ray polarization was perpendicular to both the scattering plane and the c-axis of the crystal. Simultaneous fluorescence measurements were used for the absorption correction and were also used as an reference to facilitate accurate comparisons between different Bragg peaks. The DAFS signals were measured across the Cu K-edge for the eight different (00L) Bragg peaks listed in Table i. As illustrated in Fig. 8, the measured contrast is very different for different Bragg peaks. When the peak is weak, the oscillating DAFS terms, and , are more important. The contrast is then larger and the DAFS signal can be measured more easily
for these weak, high contrast reflections than it can for the strong, low contrast reflections, or for the XAFS. This contrast enhancement is not possible in conventional XAFS because the XAFS contrast is fixed by the relative numbers of inequivalent absorbing atoms (with the same atomic type). The DAFS intensities were corrected for absorption using the measured fluorescence XAFS signal. The measured values were converted to absorption coefficient values by fitting the measured XAFS to the smooth absorption coefficient calculated using the values of McMaster . The properly normalized fluorescence XAFS data was then used to construct the absorption correction function, . The measured DAFS data was also corrected for instrument contributions.