Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/docs/documents/CAL-TN-0051-1-1.ps.gz Äàòà èçìåíåíèÿ: Thu Jun 24 17:17:16 2004 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:20:20 2012 Êîäèðîâêà:

XMM-Newton Calibration Technical Note
PN Optical Loading
XMM-SOC-CAL-TN-0051 Issue 1.1
M.J.S. Smith
April 13, 2004
1 Abstract
An investigation into PN optical loading has been performed through analysis of PN oset maps.
The extent of loading versus magnitude has been analysed and compared with the PHS Tools
model output. Also, the PN optical PSF has been analysed using optical loading in oset maps.
2 Introduction
A description of EPIC optical loading is given in [1].
The oset map analysis, which makes use of residual oset maps and quanties the extent of
loading in terms of residual oset, is described in [2].
Although an oset map is nominally in energy space, as each optically generated photoelectron
shifts the energy scale by 3.65 eV, the oset map can be considered as being in counts space for
optical photons. Assuming an oset created solely by optical photons and a gain of 5 eV per
ADU, the conversion from residual oset to optical counts per pixel is: 1 ADU  5=3:65 optical
photons.
3 Analysis
Over 350 oset maps of imaging, non (cal)closed exposures were investigated for the presence
of optical loading. Fig. 1 shows an example of optical loading cases in a PN residual oset map.
As far as possible, cases of X-ray loading were excluded by investigating the events image for
excessive (e.g. near pile-up) count rate. The remaining cases of, presumed, optical loading were
cross checked with catalogues. With very few exceptions a corresponding optical counterpart
could be found, together with, in many cases, information on magnitude and spectral type.
1

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-CAL-TN-0051
Issue/Rev.: Issue 1.1
Date: April 13, 2004
Page: 2
Figure 1: PN Residual oset map showing cases of optical loading. The bright object is
of magnitude V  6:7, the two fainter objects are of magnitudes V  9; the mode is Full Frame.
Together with the above-mentioned conversion of residual oset energy to number of detected
optical photons, this allows a comparison of detected optical ux with magnitude.
4 Results
As can be expected, optical loading cases were found mostly in thin lter exposures, with only
a few cases in medium lter exposures. No cases of optical loading were found with thick lter
in use. The faintest sources for which optical loading was detected is V  12 and V  6 for thin
and medium lters respectively.
Fig. 2 shows the relationship between peak residual oset (i.e. the maximum residual oset
found in the loaded region) and visual magnitude of optical loading cases. The data points
follow quite closely the theoretical ux versus magnitude relationship. The main deviations are
likely caused by the inclusion of various spectral types, uncertainty in visual magnitude due to
possible optical variability of some sources, and the radial variation of the optical PSF with
respect that on axis (discussed below).
For the optical loading cases for which the magnitude and spectral type is known, a comparison
was made between the measured loading and the expected loading based on the PHS Tools
model (see [1]). The model consists of a look-up table of the expected optical loading (expressed
in photons/pixel/sec) for dierent spectral types and lters, and for a visual magnitude of 0. For
PN, it assumes a fraction of 15% of light in the PSF central pixel. The results of the comparison
are shown in Fig. 3; data and model are expressed in units of ADU/pixel/0.073 sec.
There is clearly a strong correlation between data and model, but the model systematically over-

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-CAL-TN-0051
Issue/Rev.: Issue 1.1
Date: April 13, 2004
Page: 3
Figure 2: Peak residual oset versus visual magnitude. The peak residual osets of the
various modes have been normalised to the Full Frame integration time (0.073 s). Blue and red
data points represent thin lter and medium lter exposures respectively. The theoretical ux
versus magnitude relationship is shown with the dashed line. The drawn line represents a best
t exponential function.
Figure 3: Comparison of measured optical loading versus the PHS Tools prediction.
The data are shown in blue and red for thin and medium lter exposures respectively. A linear
t through the data is shown by the drawn line.

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-CAL-TN-0051
Issue/Rev.: Issue 1.1
Date: April 13, 2004
Page: 4
predicts the loading by a factor of  7 for the complete sample, with a minimum over-prediction
by a factor of 4. A similar conclusion follows from the analysis of MOS optical loading described
in [3].
The over-prediction could in part be explained by the fact that in the model the assumed fraction
of detected light at the central PSF pixel appears to be too large (indeed it was expected to be
a conservative estimate). This is illustrated in Fig. 4, which shows the PN optical and X-ray
PSFs at various o-axis angles (unfortunately there are no cases of suôciently strong optical
loading on axis, so the top panel is indicative only.) The encircled energy fraction in the central
pixel (i.e. radius up to  2 00
:5) decreases with increasing o-axis angle, and is always less than
 8% as compared to the assumed value of 15%. A similar nding for the MOS optical PSF is
presented in [4].
It may be noted that while the optical PSF is similar to the X-ray PSF at small oset angles,
it degrades much faster with increasing o-axis angles.
5 Impact on Science
The ndings mentioned above have motivated a corresponding fourfold increase in the PHS
Tools optical loading warning threshold from the original 25 photons/pixel/frame (based on a
maximum of 5 photons/pixel/frame r.m.s. [5]) to 100 photons/pixel/frame. This should help
avoid users unnecessarily opting for thicker lters or faster modes.
References
[1] PHS Tools - EPIC Optical Loading, XMM-PS-TN-40, D. Lumb, 2000
[2] PN X-Ray Loading, XMM-SOC-CAL-TN-0050, M.J.S. Smith, 2004
[3] MOS optical loading, XMM-SOC-CAL-TN-0043, B. Altieri, 2003
[4] EPIC optical PSF, XMM-SOC-CAL-TN-0040, B. Altieri, 2003
[5] Interpretation of PHS Tool Output, XMM-PS-TN-36, C. Erd, D. Lumb and R. Much, 2000

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-CAL-TN-0051
Issue/Rev.: Issue 1.1
Date: April 13, 2004
Page: 5
Figure 4: Comparison of PN optical and X-ray PSFs for various o-axis angles The
encircled energy fraction is plotted versus radius. The optical PSF is shown in red (including a
smoothing), the X-ray PSF is in blue.