Phase-contrast radiography (PCR) generates an image from gradients in the phase of the probing X-radiation induced by
the radiographic object, and can therefore make visible features difficult or impossible to see with conventional,
absorption-contrast (ACR) radiography. For any particular object, variations in either the real or imaginary parts of the
index of refraction could be greater. Most practical difficulties of PCR arise from the very small deviation from unity
(~10<sup>-5</sup>-10<sup>-6</sup>, depending of material and energy) of the real part of the index of refraction. In principal, straightforward
shadowgraphy would provide a phase-contrast image, but in practice this is usually overwhelmed by the zero-order
(bright field) signal. Eliminating this sets the phase-contrast signal against a dark field (as in Schlieren photography with
visible light). One way to do this with X-rays is with a grating that produces a Talbot interference pattern. Minute
variations in optical path lengths through the radiographic object can significantly shift the Talbot fringes, and these
shifts constitute a dark-field signal separate from the zero-order wave. This technique has recently been investigated up
to ~20keV [1-3]; this work addresses what sets the practical upper limit, and where that limit is. These appear to be
grating fabrication, and ~60keV, respectively.
High-current pulsed bremsstrahlung X-ray sources with endpoint energies in the 100 keV to 1 MeV regime are commonly used to radiograph dynamic events. Knowledge of the output spectra can assist in both improving these sources and in analyzing the imagery. Consequently, we have begun developing a spectrometer for this regime, which we refer to as the Collimated Step-wedge Spectrometer (CSS). It is based on a set of severl input channels, each comprising a collimator, X-ray filter, and scintillator. The scintillation light from all channels is recorded in a single from of a high-S/N CCD camera. Knowledge of the filters' attentuation and the scintillator's spectral response should allow unfolding an X-ray spectrum. An Initial test was performed with a bremstrahlung source of 300 keV endpoint energy, but with a non-optimal filter set.
For a sufficiently short gamma pulse, pile-up prohibits conventional PHA spectroscopy, which depends on temporal separation to distinguish individual counts. For such pulse applications, we are developing a spectrometer which detects and records individual gamma photon signals simultaneously in independent parallel channels. This is a two-dimensional array of BGO scintillation elements, each of which is a high-aspect ratio, square-section cylinder. The elements are closely packed in a regular rectangular array, and separated by reflective walls which optically isolate the elements and channel the visible light to the end. The array is oriented with the elements end-on to the gamma source. The downstream array face is imaged, and a histogram of the light levels of the individual elements corresponds to a gamma spectrum. We report on tests performed on the array component intended for such a system.