Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0121-1-0.ps.gz Äàòà èçìåíåíèÿ: Fri Jul 26 19:27:39 2002 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 22:14:00 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ï ï ï ï ï ï ï ï ï ï ï ï ï

XMM­Newton CCF Release Note
XMM­CCF­REL­121
EPIC Spectral Response Distribution
R. D. Saxton, D. Lumb
12 July 2002
1 CCF components
Name of CCF VALDATE Blocks changed CAL VERSION XSCS flag
EMOS1 QUANTUMEF 0012.CCF 2000­01­01 QE TOTAL, QE CCD1 NO
EMOS2 QUANTUMEF 0012.CCF 2000­01­01 QE TOTAL, QE CCD1 NO
2 Changes
The MOS quantum efficiencies (QE) have been modified to give a better fit to ground
calibration data. The major change from the previous files is that the central CCD in
the MOS­2 camera now has greater quantum efficiency than the central CCD of MOS­1
at high energies. This was previously modelled as an empirical bump in the MOS­2 QE
curve in EMOS2 QUANTUMEF 0010.CCF but is now included as part of a semi­physical
model.
3 Scientific Impact of this Update
The difference in the QE and hence the ARF of an on­axis source in the MOS cameras
is shown in Figures 1--4. The effect of this change is to modify the QE of MOS­1 by
upto 5% between 0.1 and 10 keV (Figure 1). This change is relative to the q20 canned
response matrices, e.g. m1 thin1v9q20t5r6 all 15.rsp and to ARFs generated using the
file EMOS1 QUANTUMEF 0010.CCF (Figure 2).
The new MOS­2 QE has a large excess above 4 keV relative to the q20 canned ma­
trices, e.g. m2 thin1v9q20t5r6 all 15.rsp (Figure 3.) and this will be noticeable as a
reduction in high energy flux obtained from spectral fits of 5--10%. The change from
EMOS2 QUANTUMEF 0010.CCF to EMOS2 QUANTUMEF 0012.CCF is upto 5% be­
tween 0.1 and 10 keV (Figure 4).
1

The ratio of the QE curves for the central CCD of the MOS detectors is shown in Figure
5.
4 Estimated Scientific Quality
These revised data were obtained from a re­analysis of the quantum efficiencies measured
at the Orsay Synchrotron calibration campaigns. Individual QE vs. Energy data points
were subject to statistical errors of several %.
For each CCD (ignoring for the time being spatial variations), the high energy absorption
were well­characterised by an effective detection depth. This is not the same as a ``true''
depletion depth, as we are selecting (via. an energy­dependent mechanism) patterns 0­12,
however the fit of the whole QE curves for E–4--5 keV is very well matched to a single
detection depth. In addition the various fits cluster around two different depths, and these
are very well correlated to the known distribution of CCDs from two different batches of
silicon. The likely discrepancy in QE at highest (10keV) energies is of order 2--3%.
Once this effective depth is selected the rest of the QE is fit to a simple absorption
model for the electrode side. In fact the inferred stopping depth does not translate
well for the QE in the range 1.2­1.8keV, which is not unexpected because the event
selection is not characterised well for the same stopping depth at these lower energies.
The absorption in electrodes is characterised via. nominal widths of different electrode
structures, and thicknesses of nitride, gate oxide, polysilicon and thermally grown oxide
layers. Their thicknesses are leveraged by the depth of oxygen, silicon and nitrogen
absorption edges, and also the general shape of the measured curve. The best fit curve
with sensible absorbing structure depths is then merged to the high energy curve to
produce the desired QE data file. For the soft X­ray efficiency, the biggest discrepancies
are just below and above the Si edge, around the O edge, and at energies less than 200eV.
Discrepancies of 10% between the experimentally measured points and the predicted curve
occur at these limited regions.
5 Expected Updates
New curves have been generated for the central CCD in the MOS cameras. This work
should be extended to the outer CCDs in a future release.
6 Test procedures
Find the flux in various energy bands of a bright, non­piled­up, on­axis source using
canned matrices and the ARF produced by arfgen using these QE files. The results
should be consistent with Figures 1--4.
2

Table 1: Mos­1: fluxes derived from fits to an AGN spectrum
Response Inband flux a
0.5--2 2--3 2--5 5--10
Canned b 3.29 1.07 2.62 3.05
arfgen c 3.30 1.09 2.67 3.03
Change 0% 2% 2% ­1%
a The flux, in ergs s \Gamma1 cm \Gamma2 , returned by spectral fitting in the bands, 0.5--2 keV, 2--3 keV, 2--5 keV and
5--10 keV.
b The canned response file used was m1 medv9q20t5r6 all 15.rsp.
c The spectral fit used the canned response, m1 r6 all 15.rmf, in conjunction with an ARF produced
by arfgen V1.48.10 (SAS V5.3.3) using the CCF, EMOS1 QUANTUMEF 0012.CCF, with the encircled
energy correction turned off.
Table 2: Mos­2: fluxes derived from fits to an AGN spectrum
Response Inband flux a
0.5--2 2--3 2--5 5--10
Canned b 3.33 1.05 2.53 2.83
arfgen c 3.33 1.07 2.54 2.62
Change 0% 2% 0% ­8%
a The flux, in ergs s \Gamma1 cm \Gamma2 , returned by spectral fitting in the bands, 0.5--2 keV, 2--3 keV, 2--5 keV and
5--10 keV.
b The canned response file used was m2 medv9q20t5r6 all 15.rsp.
c The spectral fit used the canned response, m2 r6 all 15.rmf, in conjunction with an ARF produced
by arfgen V1.48.10 (SAS V5.3.3) using the CCF, EMOS2 QUANTUMEF 0012.CCF, with the encircled
energy correction turned off.
7 Test results
The flux found from spectral fits to an AGN spectrum for the MOS cameras using the
canned matrices and ARFs generated by arfgen using the new CCFs are shown in Tables
1 and 2.
The percentage changes for MOS­1 and MOS­2 agree qualitatively with those expected
from the QE differences seen in Figures 1 and 3.
NB: The differences seen in these energy bands do depend on the source spectrum but
the values quoted here are indicative of the magnitude of the change which will be seen
in a typical source.
References
3

Figure 1: Comparison of the canned matrix, m1 medv9q20t5r6 all 15.rsp, MOS­1 response to
a slope=1.5 power­law spectrum, with that of the SAS using the quantum efficiency file,
EMOS1 QUANTUMEF 0012.CCF
4

Figure 2: Comparison of the MOS­1 response to a slope=1.5 power­law spectrum, using the quantum
efficiency files, EMOS1 QUANTUMEF 0010.CCF and EMOS1 QUANTUMEF 0012.CCF
5

Figure 3: Comparison of the canned matrix, m2 medv9q20t5r6 all 15.rsp, MOS­2 response to
a slope=1.5 power­law spectrum, with that of the SAS using the quantum efficiency file,
EMOS2 QUANTUMEF 0012.CCF
6

Figure 4: Comparison of the MOS­2 response to a slope=1.5 power­law spectrum, using the quantum
efficiency files, EMOS2 QUANTUMEF 0010.CCF and EMOS2 QUANTUMEF 0012.CCF
7

Figure 5: The relative quantum efficiencies of the central CCD of the MOS cameras given by the CCFs
EMOS1 QUANTUMEF 0012.CCF and EMOS2 QUANTUMEF 0012.CCF
8