Next: Summary Up: Transmission Maps of the Previous: Transmission Maps
The ultimate goal of these calibrations is to obtain a transmission ``spectrum'' for each pixel, that is, an estimate of the filter transmission as a function of energy in the range 0.1-10 keV, the bandpass of the telescope. Below 0.1 keV the CCDs have virtually no response; the filter transmission goes to unity above 3 keV. Thus we present our estimates of the filter transmission profiles in this range. We performed non-linear least squares fits to the data using a model based on laboratory transmission data for common elements by Henke et al. (1993). The algorithm was supplied by George Chartas.
Example transmission profiles are given in Figure 5.15. The left panel shows the profile for one pixel in an imaging array filter; the right panel shows a similar single-pixel profile for a spectroscopy array filter. For all fits, we assumed that the filters were coated on each surface with 20 Å of aluminum oxide.
For all SRC data, the nominal energies are (in eV): [273, 522, 775, 1330, 1860] but the energies that yielded the best fit are a little different. This is not surprising, given that the beamline energy is set mechanically (not computer-controlled). The filters were measured one at a time, stepping through the energies. The energy repeatability is no better than (in eV, matching the energies given above): [10,10,3,10,10]. For imaging array filter 009, the best-fit energies are [283, 512, 778, 1320, 1850]. For spectroscopy array filter 003, the best-fit energies are [283, 512, 778, 1340, 1850].
These fits yield an estimate of the areal density of the filter material at each pixel position. Given a volume density for the material, one can calculate the filter thickness. We calculated the polyimide and aluminum thicknesses in this way for an example pixel on each flight filter to compare our results to those provided by Luxel Corporation. Luxel estimates the filter material thicknesses by a completely different technique.
Luxel measures the thickness of the polyimide film and the metallic coatings using a Tencor Profilometer which is calibrated at regular intervals. Several measurements are taken and averaged to arrive at the thickness measurement that is given on the Filter Certification.
The results are presented in Table 5.3. The volume densities assumed are 1.44 gm/cm3 for polyimide and 2.5 gm/cm3 for aluminum. As mentioned above, the Al2O3 thickness is fixed in the fit to be 20Å; its density is assumed to be 3.6 gm/cm3. We also assumed that the silicon thickness was 1% of Luxel's value for the aluminum thickness, since it is doped at 1%. The assumed Si density is 2.32 gm/cm3.