a) The Ground CalibrationFor all filters, the calibration analysis had been conducted at the Space Research Laboratory, Utrecht. For each filter the mass absorption coefficient, the absolute thickness and the thickness as a function of position were determined. A number of measurements at different (monochromatic) energies were used as a starting point.
Analysis of the Boron filter data is quoted as "the most accurate available", however the higher intrinsic complexity of the boron filter should be noted.
Most of the filters are a single layer of a pure material. The Al/Par filter is two separate layers plus a support grid and can be described in a relatively simple manner (ref. FOT Handbook, section 126.96.36.199). On the other hand, the Boron filter is made of Boron in an organic solvent, sprayed on a polypropylene support, and description by separate components was therefore not possible. Modelling was necessary to describe it as a single effective component. The model was based on a number of measurements, one at an energy just below the Boron edge and six at higher energies.
b) In-Flight VerificationNote the Boron effective area features (e.g. see Fig Ec, in the FOT Handbook, section 8.1): there is a significant window below the Boron edge, then the oxygen and carbon edges are clearly visible with the bulk of the transmission around 1 keV.
A comparison of calibration measurements of the Crab Nebula with predicted count rates leads to a good agreement (see FOT Handbook, section 188.8.131.52). However, since the column to the Crab is quite high (3 x l02l cm-2), this simply verifies the higher energy part of the effective area.
Some more calibration measurements were made on a soft X-ray source, the hot white dwarf HZ43. It is possible to obtain a spectral solution consistent with the grating measurement carried out during the performance verification phase. The spectrum derived from the grating observation and the CCF area gives a predicted Boron count rate of around a factor 10 higher than the measured one.
The same problem (predicted count rate higher than observed) has occurred in some observations of soft sources like catalysmic variables, or AGN's with very low column density.
A sample of AGN's has been analysed with 1.7 power law spectrum and column density derived from two additional LE filters, usually 3000Å Lexan and Al/P. The discrepancy between the predicted and observed count rates in the Boron filter is dependent on NH, ranging from a factor ~2.5 at 5 x 1019 cm-2 to 1.3 at 3 x 1020 cm-2.
Checks of the filter ratios on HZ43 with the known grating spectrum and for a few other soft sources have been carried out.
One current area of investigation is the Boron filter point spread function which is found to be broader than the CCF values, possibly resulting from scattering within the filter. Checks are in progress to assess the influence on count rate estimates.
HEASARC Home | Observatories | Archive | Calibration | Software | Tools | Students/Teachers/Public
Last modified: Thursday, 26-Jun-2003 13:48:32 EDT