view article
Figure 4
Example of a sample (5 × 5 hex array) where the background has been intentionally increased (using the diffuse scattering of 3 × 75 µm of polyimide material). The top row shows data for the sample alone. The sample's scattering peaks (which should appear along the cyan circle) are not discernible (the blue square shows a zoomed-in region). The intensity along this azimuthal arc (black curve, right) similarly shows no hint of the sample peaks. Performing a correlation analysis (magenta curve) does not recover the sample symmetry (which should match the model curve, shown in red). The bottom row shows an equivalent sample in the presence of an amplifier. Although the sample peaks still cannot be seen, distinct interference fringes are clearly visible. The angular intensity (black curve) can be analyzed to extract the second-order correlation function (magenta), which closely matches the expected sample symmetry (red). Thus, the interference fringes encode the sample's structural information. By exploiting interference between the sample and an amplifier, the sample's structure can be reconstructed, even though the sample by itself cannot be measured.

IUCrJ
Volume 4| Part 5| September 2017| Pages 604-613
ISSN: 2052-2525
https://doi.org/10.1107/S2052252517008107