Transmission Maps

The reduced data consist of images (transmission maps) of the filters in 273 eV, 522 eV, and 775 eV light and estimates of the filter thicknesses for all pixels that were mapped at all five energies, based on fits to the transmissions at these energies. The 273 eV images usually show small (1-2%) gradients in the aluminum deposition across the filter. Subtle striations in the polyimide are visible at the 0.5-1% level in some filters (see the 522 and 775 eV images). The 522 eV images were obtained mainly to verify the results at 273 eV and 775 eV. They should show a combination of the properties evident at 273 eV and 775 eV, and indeed they always do. Since the filters are so thin, the features revealed by the transmission maps never correlate at all with macroscopic wrinkles or other visible features on the filters.

  

Figure 5.13: Imaging Array Filters 009 (left) and 019 (right). The top images show the filter transmission at 273 eV (Filter 009: $T \sim 24.8\%$; Filter 019: $T \sim 27.8\%$); the middle images are 522 eV maps (Filter 009: $T \sim 50.5\%$; Filter 019: $T \sim 53.1\%$); the bottom images are 775 eV maps (Filter 009: $T \sim 75.3\%$; Filter 019: $T \sim 77.1\%$). The left filter (009) was chosen as the imaging array flight unit. It is quite uniform at all energies mapped, implying that both the polyimide substrate and the aluminum coatings have uniform thickness and will contribute minimally to artifacts in ACIS images. The 273 eV map of filter 019 shows smooth but substantial variation in the aluminum thickness, although some ``pinwheel'' striations due to spinning out the polyimide are detectable in this image as well as in the higher-energy maps. These complex curved features correlate well row-to-row within a given image and between images, confirming that they are not artifacts of the data collection.

The transmission maps are presented in Figures 5.13 and 5.14. All maps are displayed in a range of $\pm2\%$ of the median value in the map, around that median. See the figure captions for specific comments about each filter. Imperfect data reduction is apparent in the transmission variations that show up as horizontal lines. These features are not real; they result because the data are acquired by scanning across the filter one row at a time. If the unattenuated intensity measured for that row is not exactly correct or if the beam intensity changes across a row in an unexpected manner, the calculated transmission reflects these errors and yields horizontal features.

  

Figure 5.14: Spectroscopy Array Filter 003, with the 273 eV map ($T \sim 34.9\%$) at the left, the 522 eV map ($T \sim 56.5\%$) in the middle, and the 775 eV map ($T \sim 79.0\%$) at the right. The smooth transmission gradient at low energies is presumably due to slight nonuniformities in the aluminum coatings. The diverging striations visible in the 522 and 775 eV maps are from the polyimide substrate. This is the most uniform of the polyimide-based spectroscopy array filters.

To test the reproducibility of the data, a single row of pixels was scanned multiple times and the results compared pixel-wise. This was repeated for several isolated rows throughout the mapping at each energy. The transmission values for each pixel are usually reproducible to approximately 0.1%, based on a comparison of these scans. The dominant source of error in the system appears to be beam instability. Our pixel size is closely matched to the spatial extent of the synchrotron beam for our three lower energies, thus very slight beam motion produces measureable transmission fluctuations.