Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0139-1-0.ps.gz Äàòà èçìåíåíèÿ: Mon Jan 20 12:57:29 2003 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 22:19:32 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ï ð ï ð ï ï ð ï ð ï ï ð ï ð ï ï ï ð ï

XMM­Newton CCF Release Note
XMM­CCF­REL­139
EPIC pn quantum efficiency
R. D. Saxton
12 Dec 2002
1 CCF components
Name of CCF VALDATE Blocks changed XSCS flag
EPN QUANTUMEF 0012.CCF 2000­01­01 QE TOTAL, FRACTION ENERGY, NO
QE CCDn,
FRACTION CHANNEL, CHBINS FRACTION
2 Changes
The total quantum efficiency (QE) for the Epic­pn detector has been revised following a
re­analysis of the ground calibration data. The effect is rather small giving a reduction
in the depth of the Oxygen edge of about 2% and an increase in the QE at high energies,
rising from ¸ 1% at 8 keV to ¸ 4% at 14 keV. This is applied to the QE TOTAL
extension.
Parallel adjustments have been made to the CTI correction , particularly for Small Win­
dow mode observations, and to the low­energy redistribution function. These changes
have led to small adjustments being needed to the pattern fraction ratios both in energy
and channel space. These have been applied to the FRACTION ENERGY and FRAC­
TION CHANNEL extensions.
The arrays QE 0 and QE 1 in the QE CCDn extensions have been updated correspond­
ingly.
This release includes calibrated pattern fractions for Timing and Burst modes for the first
time. Previously these were simply copies of the FullFrame mode fractions.
An additional extension, CHBINS FRACTION, has been added to quantify the PI chan­
nels over which the FRACTION CHANNEL array is defined. This extension has the
column PI CHAN.
1

3 Scientific Impact of this Update
This update, in conjunction with changes to the CTI correction for PN data, results
in a marked improvement to spectral fits around the Oxygen edge. With the previous
calibration, there appeared to be too much flux between 0.5 and 1.0 keV giving rise to
unfeasibly low values for the galactic absorption in spectral fits. This is now much better;
fits to isolated neutron stars, expected to have a simple thermal spectrum, show little
excess flux now around the Oxygen edge.
Timing and Burst mode data can now be fit with realistic effective area files (ARF)
generated by the task arfgen.
4 Estimated Scientific Quality
This update of the quantum efficiency is intimately linked to the CTI changes described in
XMM­CCF­REL­128 . That document gives two examples of the improvement in spectral
fits due to the combination of QE and CTI changes. These examples also show that the
previous significant mismatch between the PN and MOS from 0.5 to 1.0 keV has largely
been resolved.
5 Expected Updates
6 Test procedures
The changes introduced here are primarily used within arfgen, rmfgen and calview. The
integrity of the release can be verified by:
1. Compare a PN timing mode RMF, generated by rmfgen using this CCF element,
against the canned RMF epn ti40 sdY9.rmf.
2. Compare a timing mode RMF and ARF generated by the SAS using this CCF file
against the canned response file, epn ti40 sdY9 medium.rsp.
7 Test results
Figure 1 shows that the Timing mode RMF produced by the SAS reproduces, almost
perfectly, the canned response. Similarly, the ratios of the SAS and canned effective areas
are shown to agree to well within 1% below 10 keV (Fig 2.). At the highest energies the
2

Figure 1: Comparison of the SAS and canned redistribution functions for a narrow line at 0.6 keV
observed in Timing mode
discrepancy is higher due to a known difference in interpolation techniques over sharp
edges which is seen in all observing modes.
3

Figure 2: The ratio of the SAS generated response to the canned response epn ti40 sdY9 medium.rsp,
for a power­law spectrum of slope 1.5, observed in timing mode.
4