Spatial Distribution

While the spatial distribtuion of ExtCalSrc flux does not directly influence the gain calculation, it is intimately tied to the detected count rate of the K$\alpha$ lines. Two types of variation are observed. First, for a given source, there will be a certain pattern expected from its collimation beam and a reduction of intensity as the distance from the source axis increases. Second, since the photons for each flavor of K$\alpha$ lines come from a seperate source, we expect that intensity patterns will vary source to source. The housing for the ExtCalSrc contains locations for four individual sources (only three are occupied). Looking down on the focal plane such that S0 is in the lower left and I1 in the upper right (see Figures below), the Al source is in the upper left, the Ti source is in the upper right, and the Mn source is in lower left. Figures 4.115, 4.116, & 4.117 contain the intensity maps for each K$\alpha$ line. To account for chip-to-chip differences in quantum efficiency, all the chips have been normalized to the QE of S2 at each energy. For example, if S0 is only 90% as efficient as S2 at Al K$\alpha$, the observed Al K$\alpha$ intensity for S0 is multiplied by 1.0/0.9. Each 1024$\times$1024 CCD image was binned into a 64 $\times$ 64 array (16 pixels$\times$16 pixels/cell). For the typical 4800 sec exposures taken during ISIM-TV1, the total counts in a single cell is approximately equal to the count rates in Table 4.82. Since the Poisson noise was rather large (19%, 17%, and 10% for Al, Ti, and Mn) the images were smoothed with a 5$\times$5 box car function to reduce 1 $\sigma$ statistical fluctuations to the $\sim$5% level. Finally, the map was normalized with respect to the mean intensity of S2 (i.e. the mean intensity of S2 is unity for all the maps).

The most obvious of the macroscopic features is the patchy nature of the maps that results from Poisson fluctuations. The BI chips also contain significant banded structure along the quadrant boundaries and the top (row 1024) of the chip. This structure results from using a global value for the quantum efficiency, rather than accounting for the measured spatial variations in QE. This effect is most notable in chip S1 at Ti and Mn^4.11. Finally, the intensity distribution is discernable. For example, at Mn chips S4 and S5 have less flux, consistent with the placement of the bare ⁵⁵Fe source in the lower left of the ExtCalSrc holder. The Ti intensity dramatically dereases at the outer edge of S0, while the Al intensity drops dramatically at the egde of S5. Both of these structures also agree with placement of the Ti source in upper right holder position (farthest from S0) and the Al source in the upper left holder position (farthest from S5).
 

Figure 4.115: Relative intensity map of ExtCalSrc Al K-alpha. Quantum efficiencies have been normalized with respect to S2. Intensities were then normalized with resepct to S2.

 

Figure 4.116: Relative intensity map of ExtCalSrc Ti K-alpha. Quantum efficiencies have been normalized with respect to S2. Intensities were then normalized with resepct to S2.

 

Figure 4.117: Relative intensity map of ExtCalSrc Mn K-alpha. Quantum efficiencies have been normalized with respect to S2. Intensities were then normalized with resepct to S2.