3 Coded mask imaging

A coded mask imager was developed for laboratory tests. Here the prototype detector with the RENA readout, a prototype coded mask ($23\times 23$MURA), and a small source of 41 keV X-rays, 0.75 mm diameter ¹⁵³Gd, were mounted in an optical bench, and their geometry was adjusted so the divergent X-ray beam magnified the mask's unit cells to 1300 microns at the detector. Since the shadowgram size is $60\times 60$ pixels and the detector is $22\times 22$ pixels, the mask/source combination was moved to several positions relative to the detector to record the full shadowgram. The shadow of the mask was cast upon the detector, and the counts versus position along one strip is shown in Fig. 3. The transitions from blocked to unblocked mask elements are at precisely the correct positions, showing the detector's high spatial accuracy, and the modulation transfer is 100%. Unblocked counts are consistent with a constant value, which demonstrates the uniformity of response across the detector.

  

Figure 3: A 1-dimensional cut of the shadow of the mask along one cathode strip. X-position corresponds to anode strip number. The mask pattern is indicated at the top, making clear the precise position resolution of the detector. Background is present at the level of $\sim$10 cts

Shadowgrams of the mask were measured by the detector and then deconvolved to obtain a correlation image. Figure 4 shows the resulting shadowgram and image. All the mask features were accurately recorded by the shadowgram. Systematic noise in the image is undetectably small and limited to <0.5% of the source peak. Thus, systematic errors can be expected to be minimal in the reconstruction and analysis of images. See Matteson et al. (1998b) for further information on imaging results.

  

Figure 4: Left: $60\times 60$ pixel shadowgram of one coded mask cycle measured by the prototype coded mask imager with 41 keV X-rays. One mask cell equals 2.6 pixels. High counts are indicated by white. Bright regions have $\sim$15 cts/pixel, dark regions $\sim$0.7 ct/pixel. Right: Deconvolved image from shadowgram. X and Y units correspond to detector pixels