Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0229-1-1.ps.gz Äàòà èçìåíåíèÿ: Wed Jul 11 18:08:48 2007 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 22:47:51 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: horizon

XMM­Newton CCF Release Note
XMM­CCF­REL­229
RGS Background Spectra Templates
R. Gonz’alez­Riestra
June 19, 2007
1 CCF components
Name of CCF VALDATE EVALDATE Blocks changed XSCS flag
RGS1 TEMPLATEBCKGND 0004 2000­01­01T00:00:00 -- X100 P090 1 0.00... NO
X100 P095 2 8.00
RGS2 TEMPLATEBCKGND 0004 2000­01­01T00:00:00 -- X100 P090 1 0.00... NO
X100 P095 2 8.00
2 Changes
This is the third release of the RGS background spectra templates CCFs to be used by the SAS
task rgsbkgmodel.
The structure of the calibration files is as described in the Release Note for the previous version
of the CCF (RGS[12] TEMPLATEBCKGND 0003): 64 extensions, 32 for first­order and 32 for
second­order background spectra. The first 32 extensions in the file correspond to an extraction
region of 90% in PI (16 for first order, 16 for second order). The last 32 extensions correspond to
an extraction region of 95% in PI (16 for first order, 16 for second order).
These new CCFs di#er from the previous ones in the way the template spectra have been derived.
In previous releases, all empty fields were merged in a single event file per RGS, and the spectra
of the 16 levels were extracted from it. This procedure has been found to handle incorrectly the
REJPIX extensions, and leads to too low values of BACKSCAL in CCDs 1 and 2, when a large number
of pixels are rejected during the processing due to the enhanced fixed­pattern noise.
In this new release, the template corresponding to each of the 16 background levels is the average
of the spectra extracted for that given level from each of the 29 blank fields.
1

XMM­Newton CCF Release XMM­CCF­REL­229 Page: 2
Figure 1: Left: Comparison of the BACKSCAL values derived from the old (blue) and the new (red)
CCFs. Right: Results for Obsid. 0073740301, Supernova Remnant 3C 391, no source spectrum).
The black line shows the total spectrum, as extracted with rgsspectrum without background sub­
traction. The blue and the red lines shows the resulting model background computed using CCFs
003 and 004, respectively. The improvement in the estimation of the background, is clearly visible.
As in the previous release, these new background template spectra have been extracted with
version 2.6 of the SAS task rgsspectrum.
3 Scientific Impact of this Update
The use of the background spectra templates is an alternative to the conventional analysis, in which
the background is derived from the o#set region in the spatial image. The method is fully described
in in XMM­SOC­CAL­TN­0058.
The procedure used to derive the spectra used in previous releases of these CCFs did not handle
correctly the REJPIX extensions of the event files. This produced too low values of BACKSCAL in the
templates which, in turn, gave rise to too high model background spectra once they were scaled to
the extraction area used for the source.
This problem has been solved in this new version, as shown in Fig. 1.

XMM­Newton CCF Release XMM­CCF­REL­229 Page: 3
4 Estimated Scientific Quality
The model background derived from these new CCFs shows a substantial improvement, in particular
at the longest wavelengths, where the previous CCFs overestimated the background. The imple­
mentation of the present CCFs will allow more accurate measurements of relevant emission lines as
e.g., the the C VI line near 34 š A.
With the exception of pathological cases (less than a few percent) the establishment of the overall
background spectrum in an observation using the background spectra templates (as done by the
SAS task rgsbkgmodel) works very well, constituting a valid alternative for any observation, and
the only one for largely extended sources. A very good reproduction of the background obtained
with the conventional analysis could be achieved in all the control cases tested so far, using di#erent
types of object with di#erent levels of background.
5 Expected Updates
Further updates can be made in the future to improve the signal­to­noise ratio of the templates by
adding more blank fields and/or to cope with software updates.
6 Test procedures
General checks:
. use fv (or another FITS viewer) for file inspection. It should contain 64 binary extensions,
each with five columns: CHANNEL, RATE, QUALITY, BACKSCAL and AREASCAL.
. use the SAS task cifbuild to see if the CAL digests and creates correctly the calibration
index file.
7 Summary of the test results
A full description of the method and results obtained using the new SAS task rgsbkgmodel is given
in the Calibration Technical Note XMM­SOC­CAL­TN­0058 ''Templates for the RGS Background''
[1].
General checks:
The fits viewer fv was used to inspect the two CCF files, wrt their structure and validity dates.
Everything was OK.

XMM­Newton CCF Release XMM­CCF­REL­229 Page: 4
The SAS task cifbuild was run successfully in order to check the ingestion of the files into the
calibration index file.
References
[1] ``Templates for the RGS Background'', R. Gonz’alez­Riestra, XMM­SOC­CAL­TN­0058, October
2004
[2] ``RGS Background Spectra Templates'' R. Gonz’alez­Riestra, XMM­CCF­REL­217, June 2006