Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0084-1-0.ps.gz Äàòà èçìåíåíèÿ: Tue Aug 7 13:10:46 2001 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 22:01:05 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ngc 6992

XMM­Newton CCF Release Note
XMM­CCF­REL­84
EPIC Spectral Response Distribution
D Lumb
August 6, 2001
1 CCF components
Name of CCF VALDATE List of Blocks
changed
CAL VERSION XSCS flag
EMOS1 REDIST 0011.CCF 2000­01­01 EBINS NO
EMOS2 REDIST 0011.CCF 2000­01­01 EBINS NO
XRT1 XAREAEF 0008.CCF 2000­01­01 ONAXISAREAEF, NO
VIGNETTING NO
XRT2 XAREAEF 0009.CCF 2000­01­01 ONAXISAREAEF, NO
VIGNETTING NO
XRT3 XAREAEF 0008.CCF 2000­01­01 ONAXISAREAEF, NO
VIGNETTING NO
XRT1 XENCIREN 0003.CCF 2000­01­01 XENCIREN NO
XRT2 XENCIREN 0003.CCF 2000­01­01 XENCIREN NO
XRT3 XENCIREN 0003.CCF 2000­01­01 XENCIREN NO
EMOS1 QUANTUMEF 0007.CCF 2000­01­01 QE CCDn, EBINS NO
EMOS2 QUANTUMEF 0007.CCF 2000­01­01 QE CCDn, EBINS NO
EPN QUANTUMEF 0009.CCF 2000­01­01 QE CCDn, EBINS NO
2 Changes
These changes have been made to support the release of the rmfgen and arfgen tasks for SAS
v5.1.
2.1 Redistribution
The EMOSn REDIST files have an increased density of energy channel spacing which especially at
low energies seems necessary to match the response available from the existing fixed on­axis response
1

XMM­Newton CCF Release XMM­CCF­REL­84 Page: 2
matrices (e.g. m1 thin1v9q19t5r5 all 15.rsp) previously distributed.
2.2 Mirror Area
The XRT XAREA files have an increased density of energy points in the VIGNETTING extension,
in order to overcome limitations with interpolation in energy which resulted in an underprediction
in off­axis area loss with energy.
The XRT XAREA files have a modified on­axis area that includes a modification to the gold M
edge region, which reduces the residuals in fitting MOS and PN spectra simultaneously.
2.3 The encircled energy CCF
These have been created from a PSF parameterisation based on a King profile developed in IFCTR­
Milan. The Milan calibration team have analysed MOS observations of point sources and have
formulated the PSF as a function of energy and off­axis angle [1]. This work has been separately
extended to the EPIC­PN [2]. The formulation differs in detail from previous parameterisations
based on ground calibration data but gives broadly similar FWHM and HEW values.
The King profile provides an excellent fit to the in­orbit PSF (Figure 1). It is slightly energy
dependent and at high energies has a narrower core and flatter wings than at the lower energies.
This leads to the encircled energy fraction being noticeably greater for higher energies for extraction
radii smaller than ¸ 1 arcminute (Figures 2--4).
A paucity of good off­axis data has delayed the determination of the relationship of the PSF
with off­axis angle. The work has been completed for MOS­1 and there it shows a gradual decrease
of encircled energy with distance from the optical axis (Figure 5). The relationship to off­axis
angle of the MOS­2 and PN encircled energy functions (EEF) has still to be calculated. For these
instruments the on­axis function is used everywhere. If these telescopes follow the patten of XRT­1
then this will introduce a normalisation error for sources at large off­axis angles but the spectral
effect will be small. Due to the format of these CCF files the extrapolation of the EEF to off­axis
angles is done independently of energy. As we have seen the effect on the spectrum will be small.
2.4 Quantum Efficiency
The MOS detection efficiencies were changed to match those used in the existing Leicester­supplied
on­axis response matrices. Changes for the off­axis efficiencies were introduced based on measure­
ments taken on the ground at Orsay synchrotron, where we believe a batch­dependent detection
depth scales the high energy detection efficiency a few % from CCD to CCD.
The new knowledge of the PN CCD pattern ratios and better estimates of overall QE are ingested
to provide a better match to the arfgen generation to the supplied instrument matrices. Note there

XMM­Newton CCF Release XMM­CCF­REL­84 Page: 3
Figure 1: The radial profile of Zeta­Puppis modelled with a King profile
Figure 2: The encircled energy for MOS­1

XMM­Newton CCF Release XMM­CCF­REL­84 Page: 4
Figure 3: The encircled energy for MOS­2
Figure 4: The encircled energy for EPIC­PN

XMM­Newton CCF Release XMM­CCF­REL­84 Page: 5
Figure 5: The relationship between encircled energy fraction and off­axis angle for MOS­1
are still problems at energies Ÿ500eV that will need to be fixed by working in PHA space and not
PI domain. This is still to be addressed in future releases.
3 Scientific Impact of this Update
These updates should allow the SAS v5.1 tasks to return response matrices which are consistent to
within 1 ­ 2 % of the response matrices supplied by the instrument teams for on­axis. However the
tasks allow now the generation of responses for arbitrary off­axis angles and for arbitrary inclusion
radii.
A more detailed description of the application is given in the reference [3]
4 Estimated Scientific Quality
There are remaining spectral residuals seen around the instrumental absorption edges such as O
(0.5keV), Al (1.49keV), Si (1.84keV) and Au (2.1 keV) . Adjusting systematic errors in the region
of these features to ¸5% in the spectrum files would be prudent. In addition at the extremes of the
energy range (Ÿ250eV and –8keV for example), the deviations between MOS and PN are noticeable

XMM­Newton CCF Release XMM­CCF­REL­84 Page: 6
at a few percent level.
5 Expected Updates
Future changes are expected to be real improvements in the physical data representing improved
knowledge of the instrument
For example we continue to investigate the complex relationship between redistribution and
CCD Quantum Efficiency, the off­axis encircled energy and CCD spatial efficiency variations all will
contribute to the off­axis response, which thus needs further development.
We are carrying out detailed cross­calibrations between RGS and EPIC, and hope to use this
work to obtain a more consistent set of data in time for the update of SAS in the Autumn 2001.
References
[1] EPIC­MCT­TN­008, Ghizzardi, S, Feb 8, 2001.
[2] Griffiths, G., Saxton, R., in prep.
[3] R Saxton, Status of the SAS spectral response generation tasks for XMM­EPIC, XMM­SOC­
PS­TN­43