Introduction

Recently the REV2 reprocessing of PSPC data was completed. These new data are a major improvement to previous processings. In addition to revised corrections also the raw event information is now available to guest observers.

However, present spectral fits still yield an apparently systematic and significant behaviour of residuals, especially at channels 30-40 and . Additionally observations of some sources where we believe to know the spectral shape (e.g. Mkn421), have shown an unexpected excess at higher energies. For several soft sources (e.g. white dwarfs) a higher PSPC response was reported than expected from present calibration data. This has led to several attempts to improve the PSPC calibration data set, i.e. effective area (EFFAR) and detector response matrix (DRM).

Systematic residuals can also originate from the source model itself, like different abundances for the absorbing ISM or non-LTE models instead of a Raymond-Smith plasma model. Also a clumpy ISM may lead to a soft excess which can not be satisfactorily modelled by standard spectra. Note, that the galactic neutral hydrogen column density, , need not to represent the measured X-ray absorption of the PSPC spectrum using standard cross sections and abundances. In the following we assume that nothing peculiar occurs in the intrinsic source spectra.

Trials and Findings

1

Previously applied improper spatial gain corrections (x-y-error) lead to a somewhat broader photon distribution. This error was corrected with SASS-version 7.2.

2

All on-board Al-K calibration measurements have been re-analysed using raw PSPC data. The three peaks of the Al-K source have been studied independently. Over the whole field-of-view of the PSPC residual spatial gain variations have been identified, i.e. lower gain in the detector center and higher gain at some outer parts. The size of these variations also changes with time and is of the order of .

3

New values for the optical constants for gold have been used to create a new effective area for the X-ray telescope (XRT) by ray-tracing programs. These XRT areas have to be multiplied with the PSPC window transmission (including support structure) as well as with the quantum efficiency to obtain XRT/PSPC effective areas (like in EXSAS). The new areas show higher response below 200 eV (+10cm ) and above 1.06 keV (+30cm ), reduced response between 0.2 and 1.06 keV (-10cm ), and several small-scale differences.

4

The PSPC window transmission has not been calibrated below energies of 0.188 keV leading to uncertainties in the soft PSPC response. However, trying various compositions of the window constituents does not lead to significant improvements. The carbon edge of the PSPC window was slightly shifted to account for chemical effects. Atomic physics allow about 5eV but even 10eV were not sufficient to remove the trends. A spare PSPC window will be analysed at BESSY by the Leicester group to obtain additional information on the low-energy window transmission.

5

The present DRM has been fudged to data with improper spatial gain correction. A new DRM is under development.

Summarizing, each additional improvement alone does not neccessarily lead to an improved significance of the spectral fit. Combining all, however, leads to at least the same with highly reduced systematic residuals at 0.9keV. The low-energy residuals can be reduced but are still significant. Globally, residuals appear more randomly distributed than before (this has not yet been quantified). The changes of the residua of individual sources do strongly depend on the source spectrum as well as on the detector position, but no general trend can be predicted.

New Corrections in EXSAS

New corrections will be distributed in the next EXSAS release (March 1997) and will be available on a patch level before. They concern the spatial gain variation in the detector. PROCESS/CT will apply these corrections to the raw data if desired (see item 2 above). This command can also be used to correct REV2 data that have been reprocessed with a SASS-version before 7.2 (see item 1 of the previous section). Additionally, several tools for determination and correction of the PSPC gain will be included in the new EXSAS release. The additional corrections will also be incorporated in a future SASS reprocessing (REV3). Gain images of the PSPC will be placed in the calibration area. A new calibration data set (EFFAR, DRM) for REV2 data will also be provided as an EXSAS patch as soon as a consistent picture of the remaining residuals is available (items 3 - 5 of previous section).

Suggestions

At the moment you are strongly encouraged to use the present REV2 data (instead of REV0 or REV1) with the present calibration effective area table (EFFAR) and detector response matrix (DRM). In some cases spectral fits may give a slightly worse . There are only a few cases of bright sources (> 20cts/s) which have been observed in the detector center later than May 1992 that show significant changes of the spectral parameters.

Upto the time where new calibration data sets (EFFAR, DRM, gain images) are available you can already make use of the new corrections (PROCESS/CT), e.g., if you have to select your background from a region with quite different effective gain relative to your source position. Then you should correct the background photons as well as the source photons with the corresponding effective gain value, use the corrected background spectrum for the correction of the source spectrum, and use the present calibration set for spectral fits.

Über dieses Dokument ...

This document was generated using the LaTeX2HTML translator Version 96.1-f (May 31, 1996) Copyright © 1993, 1994, 1995, 1996, Nikos Drakos, Computer Based Learning Unit, University of Leeds.

The translation was initiated by Ortwin Schwentker on Mo 27 Jan 14:14:40 1997

© Roentgengruppe des MPE (group)
letzte Änderung:1999/09/08 13:57:38, Editor dieser Seite:Ortwin Schwentker