Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/docs/documents/CAL-TN-0005-1-0.ps.gz Äàòà èçìåíåíèÿ: Thu Jan 11 19:29:39 2001 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:15:35 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: íëó

Imaging performance of the XMM­Newton X­ray telecopes
B. Aschenbach a , U. Briel a , F. Haberl a , H. Br¨auninger a , W. Burkert a , A. Oppitz a ,
P. Gondoin b and D. Lumb b
a Max­Planck­Institut f¨ur extraterrestrische Physik,
Giessenbachstrasse, 85740 Garching, Germany
b European Space Research and Technology Center,
2200 AG Noordwijk, The Netherlands
ABSTRACT
The in­orbit imaging performance of the three X­ray telescopes on board of the X­ray astronomy observatory
XMM­Newton is presented and compared with the performance measured on ground at the MPE PANTER test
facility. The comparison shows an excellent agreement between the on ground and in­orbit performance.
Keywords: XMM­Newton, X­ray telescopes, X­ray optics testing, X­ray straylight
1. INTRODUCTION
On December 10 1999 the XMM spacecraft was placed in a 48h Earth orbit by the first commercial ARIANE V
launcher. XMM, or X­ray Multi­Mirror Mission, is the second cornerstone of the Horizon 2000 science program
of the European Space Agency ESA. XMM was designed as a facility class X­ray astronomy observatory to study
cosmic X­ray sources spectroscopically with the highest collecting area feasible in the 0.1 ­ 10 keV band. Over the
full band moderate spectral resolving power between 1 and 60 is required and medium resolving power of – 250
was aimed for in the 0.1 ­ 3 keV band. The high throughput is primarily achieved by a set of 3 imaging, highly
nested Wolter type I telescopes. The imaging performance of each of the grazing incidence mirror systems was
originally set to be better than 30 arcsec for the half energy width (HEW) of the point spread function (PSF)
with a goal of 10 arcsec. In a first attempt the scientific requirements have been put together at a workshop held
at Lyngby, Denmark, in June 1985 1 . A detailed analysis of the types of telescopes with which these requirements
could be achieved was carried out by the ESA established Telescope Working Group, which also arrived at a first
telescope design 2;3 . Because of launcher volume constraints which became apparent shortly after the publication of
the Telescope Working Report the baseline design was slightly modified in March 1987 4 . Since then the design was
not changed any more and the actual telescopes and their mirrors, now being in orbit, are of that design. Over ten
years in time the design was put into reality, the telescopes were built in industry including Carl Zeiss, Oberkochen
in Germany, Kayser­Threde, Munich in Germany and Media Lario, Bosisio Parini in Italy. The development and
building of the mirror shells up to the flight mirror modules was accompanied by X­ray tests in the 130 m long
X­test facility PANTER of the Max­Planck­Institut f¨ur extaterrestrische Physik 5 . The qualification of the mirror
modules was performed in EUV light at the Focal­X facility of the Centre Spatial de Li`ege in Belgium.
After a short description of the telescope design and its critera followed by an outline of the manufacturing
process of the mirrors the in­orbit imaging performance of each of the three X­ray telescopes is presented and
discussed and confronted with the full X­ray beam calibration measurements taken at the PANTER.
2. OPTICS DESIGN
The design of the optics was driven by the requirement of obtaining the highest possible effective collecting area
over a wide band of energies up to 10 keV and more with particular emphasis on the region around 7 keV, in which
the K lines of the astrophysically important iron appear. Thus, the mirror system has to utilize a very shallow
grazing angle of about half a degree in order to provide sufficient reflectivity at high energies. Both, Wolter type
and Kirkpatrick­Baez type telescopes can in principle be considered for the optics. The Kirkpatrick­Baez system
has however a focal ratio, defined by the ratio of the focal length over the diameter of the optics, which is two
times greater than thatfor a corresponding Wolter type I system. Taking the same focal length a Wolter type I
system can therefore make efficient use of a mirror system at the same grazing angle which is a factor of two wider

and an increase in collecting area by a factor of four. This factor of four is slightly reduced to a factor of ú since
the Kirkpatrick­Baez system builds rectangular whereas the Wolter system has circular mirrors. Because of this
major advantage the Telescope Working Group baselined the Wolter type I system for XMM.
In grazing incidence optics the effective area is generally increased by nesting a number of mirrors and thus
filling the front aperture as far as possible. The nesting efficiency is determined by the mirror shell thickness
and, in case of very low grazing angles for high energy optics, by the minimum radial mirror separation which
is required for integration and alignment. In soft X­ray telescopes of moderate grazing angles the separation is
of less importance since adjacent shells are ususally comfortably separated anyway to prevent blocking of on­axis
rays.
The thinner the mirror shells are and the narrower the shells are spaced the larger is the collecting area. In
the past very thin aluminum foils shaped to double cones have been used, for instance for the BBXRT telescopes,
the telescopes on ASCA and ASTRO­E. Due to the usage of foils and the very narrow foil placement this type of
telescope has the potential of achieving maximal collecting area. There are, however, two major drawbacks which
are inherent to foil telescopes. Firstly, the mirror sections are bent into cones, which produce an extended image
even for on­axis point­like sources, leaving significantly less tolerances for mirror surface irregularities and mirror
mounting. Secondly, foils have been reported to have intrinsic mid­frequency irregularities which limit the angular
resolution to something like one arcmin and more. Because of these disadvantages foil telescopes were considered
by the Telescope Working Group not to meet the XMM resolution requirements.
Thin­walled mirrors have also been produced by replication techniques and electroforming. Electroformed
mirrors have been built up mostly in solid nickel, which, however, suffer from too much mass, and they were
therefore initially discarded by the Telescope Working Group. Development work on the mirrors started by
producing carbon fibre reinforced plastics (CFRP) mirrors by epoxy replication from master mandrels polished
to X­ray imaging quality. Advantages of this approach are mirrors of high stiffness with thin walls, implying low
mass, and low thermal expansion. Disadvantages are the non­amorphous structure of the CFRP implying the
development of unacceptable surface structure (print­through) and sensitivity to humidity (out­gassing or even
''breathing'') thereby changing the surface with environment.
At the time of development start CFRP mirror carriers could be made as thin as 0.5 mm for the smallest
diameter shells, which is the thickness baselined for the optics design calculations. The shell thickness is assumed
to increase approximately linear with shell diameter to guarantee sufficient stiffness. A minimum radial shell
separation of 1 mm was considered to be feasible to handle the integration of the shells in a package. Both the
paraboloid section and the hyperboloid section were replicated as a single piece from a single mandrel, which limits
the axial length of the total mirror to 60 cm. The length is shared evenly by the paraboloid and the hyperboloid.
Although longer mirrors would have provided a larger collecting area they were discarded because they were
suspected to be very difficult to be removed from the mandrel. The production of paraboloid and hyperboloid
in one piece avoids the problems created by the additional tasks of aligning and mounting if paraboloid and
hyperboloid are separate. Given the length constraints imposed by the launcher the focal length was fixed at
7.5 m. The optimum design, which fulfils the collecting area constraints, was found by ray tracing. Gold was
considered as baseline for the mirror surface coating; attempts to produce replicated mirrors with an iridum
coating, which would have provided somewhat better high energy response, failed because the adhesive forces of
iridium on the mandrel surface turned out to be too strong so that patches of the iridum layer tended to stick on
the mandrel. The parameters of the optimum design are shown in Table 1. The mirror diameter of the outermost
shell is 700 mm. At low energies ! 2 keV the area could have been increased by adding larger shells. In contrast,
the area at 7 keV increases up a mirror diameter of 650 mm but drops for larger diameter shells. Adding more
shells to the inner region would increase the higher energy response but at rather low efficiency, which means
that the area per shell is significantly less than the area averaged over all shells. The optimum telescope design
is therefore defined by a maximum diameter mirror shell of 700 mm and 58 mirror shells in total. It is optimal
because it has a 7 keV response which is only marginally lower than maximal. On top it has very good low energy
response.
After several years of study, development, building and X­ray testing of CFRP mirrors the angular resolution
requirement of 30 arcsec could not be met. Mirrors of ! 1 arcmin were produced several times but their suscep­
tibility to humidity appeared to be an unsurmountable problem. Inspired by the progress which was made in the
production of good electroformed nickel shells in the context of the SAX and JET­X programs, it was decided in

Table 1. XMM­Newton Wolter Type I X­ray telescope characteristics.
Focal length 7500 mm
Outer mirror radius 350.0 mm
Inner mirror radius 153.1 mm
Axial mirror length (par. + hyp.) 600 mm
Outer mirror wall thickness 1.07 mm
Inner mirror wall thickness 0.47 mm
Minimum nesting distance 1 mm
Reflective surface material gold
Number of mirrors per telescope 58
Mirror surface per telescope 51.4 m 2
Number of telescopes 3
1991/92 by the ESA XMM Project Team in consultation with the XMM Telescope Advisory Group to terminate
the CFRP activities and to build the XMM mirrors from electroformed nickel. The original telescope design was
kept except that the mirror shells were reduced in wall thickness by about 25% to save mass as much as possible
(Table 1 shows the mirror wall thicknesses for the nickel mirrors). Because of the high mass involved with nickel
made mirror shells the number of X­ray telescopes was definitely limited to three. The high mass also contributed
to the discussion of switching launchers from ARIANE IV to ARIANE V.
The production of nickel shells involves three steps. As for any replication a master mandrel of negative
shape is produced for each of the 58 mirror shells. The mandrel is of solid aluminum covered with a thin layer
of Kanigen, which is ground and polished to the precise negative shape of the required mirror. The mandrel
surface is measured by metrology means, including profilometers of various trace lengths and interferometers. Ray
tracing of the metrology data reveals that the XMM mandrels have an angular resolution of Ÿ 5 arcsec HEW
and and an rms surface microroughness of Ÿ 0.5 nm. The mandrels are coated with a gold layer about 200 nm
thick. Each mandrel is put in an electrolytical bath in which a nickel layer is being built upon the gold surface
until the required thickness has been reached. By cooling the nickel shell separates from the mandrel, whereby
the gold surface sticks to the electroproduced nickel shell because of the higher adhesion forces. After separation
the nickel shells are exclusively supported by 16 strings threaded through 16 tiny holes which are equally spaced
around the circumference of the mirror closely below the rim of the paraboloid. After inspection of the surface
and measurement of a couple of meridional profiles each mirror is mounted to a spider like support structure by
glueing the parabola mirror into small grooves in each of the 16 arms of spider. This is done with the mirror
oriented vertically and under optical control. The spider support structure at the paraboloid front is the only
support structure for the entire mirror module. There is no counterpart at the aft of the hyperboloids; they are
not constrained or supported mechanically at all. Further details about the production of the XMM telescopes
are given by de Ref. 6.
From an X­ray optical point of view the telescope is complemented by an X­ray baffle, which does not eliminate
but significantly reduces X­ray straylight from sources outside the field of view. Sources outside the field of view
can reach the sensitive area of the focal plane detectors by single reflection from the rear end of the hyperbola.
Less highly nested telescopes like those on the Einstein, ROSAT and Chandra observatories employ radial vanes
with one set inside and one set in front of each mirror to eliminate singly reflected rays. For XMM an external
multi­stage baffle has been conceived 7 , which basically consists of a set of concentric cylinders, one for each mirror
shell, placed in front of the mirror system. Each ring has an annular width slightly smaller than the corresponding
mirror wall thickness and a cylindrical height characteristic for each mirror by which the singly reflected rays are
blocked. The implementation of such a system would have eliminated the straylight completely. But because
of the mechanical complexity together with the limited space available the design was simplified to two parallel
annular plane sieve plates of equal thickness. Each plate consists of 58 rings and 16 radial struts. The plates are

mounted co­axial to and co­aligned with the front aperture cross section of the 58 mirror shells. The off­set of
the two sieve plates from the front of the mirror system is 385 mm and 439 mm, respectively 8 The efficiency of
the sieve plate system has been ray traced and demonstrated to reduce the straylight level by a factor of 5 to 10
depending on the position in the focal plane.
A complete overview of the telescope with all its subsystems including the mirror door, the entrance optical
baffle, the electron deflector and additional support structure is given in Ref. 6.
3. X­RAY CALIBRATION ON GROUND
During Phase C/D of the XMM program five X­ray mirror modules have been built three of which have been
selected for flight, i.e. FM2, FM3 and FM4. FM1 and FM5 have been kept as spares, numbering of the modules
is in chronological order of production. X­ray testing and final X­ray calibration including measurement of the
PSF both on­axis and off­axis for a few energies between 0.28 keV and 8.05 keV and the effective area as well as
the vignetting function for a series of energies covering the range from 0.28 up to 17 keV has been performed in
MPE's PANTER test facility between April 1997 (FM1) and July 1998 (FM4). FM5 was tested in 1999 but it
has not been completed to the extent to serve as flight spare, whereas FM1 actually is fully functional for flight.
MPE's PANTER originally built for the development and calibration of the ROSAT X­ray telescope was modified
and partially rebuilt in 1992/93 to cope with the size of the XMM telescopes and quality assurance requirements.
In simple terms, the facility consists of a 130 m long vacuum pipe of 1 m diameter. At the one end a micro­focus
characteristic X­ray line source is installed and the use of various anti­cathode materials provides the required
coverage of the energy band. The bremsstrahlung component can be used as a continuum source as well. The
X­ray source has a diameter of about 1 mm. At the other end a huge cylindrical vacuum tank 12 m long and 3.5
m in diameter houses the X­ray telescope to be tested. The typical distance between source and mirror module
midplane, i.e. the paraboloid hyperboloid intersection, is about 124 m. With a focal length of 7.5 m the image
of the X­ray source is produced at a distance of 7.97 m measured from the mirror module midplane. Given these
distances the angular size of the X­ray source is about 1 arcsec, which means that structures of the mirror surface
finer than this cannot be resolved.
Here we report about the measurements with the entire aperture of the mirror module illuminated, implying a
beam divergence between 8.5 arcmin and 19.5 arcmin full width between the innermost and outermost mirror shells.
The divergence of the illuminating beam implies that the rays reflected off the frontal one third of the parabolas
are not reflected from the hyperbolas. Therefore the surface of this section of the parabolas is not imaged and
does not contribute to the PSF. Furthermore the corresponding reflection loci on parabola and hyperbola differ
from those for parallel beam illumination. Images have been recorded with both a position sensitive proportional
counter (PSPC) which is a copy of the PSPCs flown on­board of ROSAT 9 and an X­ray CCD provided by the
X­ray astronomy group of the University of Leicester in the UK. Whereas the PSPC is well suited to measure
effective area and the wings of the telescope point spread function the CCD with it significantly better spatial
resolution has been used to assess the details of the PSF out to angular distances of about 125 arcsec limited by
the linear size of the CCD. The PSPC has a larger field of view and the PSF can be traced out to angular distances
of about 1000 arcsec. The comparision of the CCD and the PSPC encircled energy functions show that the CCD
images out to 125 arcsec contain 97% of the total flux at 1.5 keV and 93% at 8.05 keV, respectively. The results
have been reported in Ref. 10,11,12 and a summary is given in Ref. 6.
4. THE POINT SPREAD FUNCTION IN ORBIT
On January 19 2000 the X­ray telescope FM2 saw ''First Light'' of the X­ray sky 13 followed by FM3 and FM4
two days later. Two different X­ray cameras are used to record the X­ray images. In the focal plane of FM2 a
novel type of X­ray CCD based on pn­technology is used (FM2/pn) 14 whereas FM3 and FM4 each are equipped
with a more conventional CCD based on MOS­technology (FM3/MOS1 and FM4/MOS2). Unlike FM2, FM3 and
FM4 each carry a reflection grating spectrometer (RGS) 15 by which about half of the X­ray beam is diverted
from MOS1 and MOS2. At a press conference on February 9 2000 the ''First Light'' results were presented and
the new name of the mission XMM­Newton was announced. After ''First Light'' the imaging performance of
the three telescopes was of immediate interest and additional observations were made during the commissioning
phase of XMM­Newton until March 8. Sources studied include the sources of the ''First Light Field'' to the Large
Magellanic Cloud and the point sources EXO 0748­67, LMC X­3, PSR 0540­69 and PKS 0558­504. A first analysis

Figure 1. Surface brightness of a point source image taken with mirror module FM3 in orbit (left) and on ground
(right). The size of the frame is 44'' by 44'' for both the in­orbit and on­ground data. Contours are stepped by a
factor of two in brightness. In­orbit contours appear to be smoother because of much lower counting statistics.
shows that the observations, except for LMC X­3 at high energies, taken in full frame mode of the CCDs, i.e. the
entire field of view is read out from the CCD at once, are not useful for the assessment of the PSF because the
data suffer from ''pile­up'' 13;14 , i.e. there is more than one count per pixel per read­out frame, which happens
predominantly in the center of the image with the intensity underestimated there. In ''small window mode'' the
observational field is reduced and the CCD read­out time is shorter such that the effect of ''pile­up'' is reduced
accordingly. For this mode the PSF analyis could be done for the last two sources mentioned above. But the
radial extent of the PSF is limited accordingly. The PSFs taken on ground can therefore only be compared with
the in­orbit results out to the angular distance imposed by the ''small window'' size. The wings of the PSFs are
not accessible with these early measurements. Likewise, the encircled energy functions, i.e. the radially integrated
PSFs, have been normalized to the value measured on ground at the angular distance corresponding to the ''small
window'' size.
A major difference between the conditions on ground and in orbit is the type of detector used in the focal
plane, in particular its pixel size. The CCD used on ground had a pixel size of 25 ¯m equivalent to 0.65 arcsec.
The CCDs used in orbit have a pixel size of 40 ¯m for the MOSs and 150 ¯m for the pn, which relates to 1.1 arcsec
and 4.1 arcsec, respectively. In particular the pn­CCD pixel size appears too large to resolve the core of the PSF.
The in­orbit imaging performance of the telescopes is presented in terms of contour lines, describing the surface
brightness across the image (Figure 1), the PSF and the encircled energy function (Figure 2) for FM3/MOS1 at
1.5 keV. Figure 3 and Figure 4 show the corsseponding data for FM4/MOS2 at 1.5 keV as well. The in­orbit
results are compared with the equivalent set of data measured at PANTER. The agreement between on ground and
in­orbit measurements is striking. It is noted that the contour line plots have been produced using a smoothing
algorithm including the adjacent three pixels in either direction, which tends to increase the width of the core of
the PSF. For the determination of the FWHM of the PSF the unsmoothed contours have been used giving the
FWHM values listed in Table 2. Figure 5 shows the encircled energy function for FM2/pn at 1.5 keV for the source
PSR 0540­69 and the encircled energy function for LMC X­3 for an energy band of (8 \Sigma1) keV. The corresponding
ground calibration measurements are not shown but the results are identical within very small excursions with
the in­orbit performance. Because of the comparatively large pixel size of the pn­CCD details of the core of the
PSF are not accessible in orbit and the FWHM of the PSF of FM2 cannot be determined. The PSF of FM2 is
undersampled by the pn­CCD.

Figure 2. Point spread function (left) and encircled energy function (right) of mirror module FM3 (solid line: in
orbit, dashed line: on ground).
Figure 3. Surface brightness of a point source image taken with mirror module FM4 in orbit (left) and on ground
(right). The size of the frame is 44'' by 44'' for both the in­orbit and on­ground data. Contours are stepped by a
factor of two in brightness. In­orbit contours appear to be smoother because of much lower counting statistics.

Figure 4. Point spread function (left) and encircled energy function (right) of mirror module FM4 (solid line: in
orbit, dashed line: on ground).
Figure 5. Encircled energy function at 1.5 keV (left) and at 8 keV (right) of mirror module FM2 in orbit.

Table 2 summarizes the results in terms of characteristics numbers like FHMW and HEW (half energy width
of the encircled energy function).
Table 2. Summary of the imaging performance of the XMM­Newton X­ray telescopes.
1.5 keV 1.5 keV 1.5 keV 1.5 keV 8 keV 8 keV 8 keV 8 keV
orbit ground orbit ground orbit ground orbit ground
telescope FWHM FWHM HEW HEW FWHM FWHM HEW HEW
('') ('') ('') ('') ('') ('') ('') ('')
FM2/pn ! 12.5 6.6 15.2 15.1 !12.5 6.6 15.0 14.8
FM3/MOS1 4.3 6.0 13.8 13.6 ­ 5.1 ­ 12.5
FM4/MOS2 4.4 4.5 13.0 12.8 ­ 4.2 ­ 12.2
5. CONCLUSIONS
Providing that the PSFs of the three X­ray telescopes on board of XMM­Newton have not developed significantly
bright wings outside of about 1 arcmin radius from ground to early orbit the following conclusions can be drawn.
ffl From the contour lines and the FWHM of the PSFs and the agreement with the ground calibration mea­
surements it follows that all three focal plane cameras are in focus.
ffl The PSFs in orbit are basically identical with those measured on ground; the core of FM2 cannot be resolved
in orbit because of the large pn­CCD pixel size.
ffl The encircled energy functions of FM2, FM3 and FM4 measured in orbit at 1.5 keV agree with those
measured on ground, at least out to an angular radius of about 40 arcsec. At 8 keV the FM2 shows the same
encircled energy function in orbit as measured on ground.
Beyond this direct comparison further conclusions can be drawn which are related to the mechanical integrity
of the telescopes and the X­ray calibration test set­up.
ffl The X­ray telescopes have not changed their shape during storage on ground or by launch loads.
ffl The focal plane instrumentation was correctly installed at the required position and no significant movement
occurred during launch.
ffl The mounting and fixture of the mirrors in just one spider support structure at the parabolas' front is
sufficient to maintain the mirrors' performance.
ffl Effects of gravity which would potentially distort the mirrors during horizontal X­ray testing in the MPE
PANTER test facility are not apparent concluded from the comparison of the ground measurements with
the in­orbit results of the PSF.
ffl The divergence of the X­ray beam in the PANTER test geometry does not limit the assessment of the PSF.
Interestingly the non­illumination of the first third of the parabolas' sections does not change the PSF of the
XMM­Newton telescopes. This means that the angular distortions along the meridian lines of the parabola
mirrors imposed by the mounting at their front ends are not significantly larger than along the remainder of
the meridian lines towards the hyperbolas' exits.
ACKNOWLEDGMENTS
The provision of the pn­CCD and MOS­CCD in orbit data of XMM­Newton by the EPIC team led by the Principal
Investigator Martin Turner of the Leicester University, UK is acknowledged.

REFERENCES
1. ''A Cosmic X­ray Spectroscopy Mission'', Proc. of a Workshop held in Lyngby, Denmark on 24­26 June
1985, Proc. ESA SP­239, 1985.
2. B. Aschenbach, O. Citterio, J.M. Ellwood, P. Jensen, P. de Korte, A. Peacock and R. Willingale, ''The High­
Throughput X­ray Spectroscopy Mission'', Report of the Telescope Working Group, Proc. ESA SP­1084,
1987.
3. B. Aschenbach and Heinrich Br¨auninger, '' Grazing Incidence Telescopes for ESA's X­ray astronomy mission
XMM'', in X­ray Instrumentation in Astronomy II, L. Golub, ed., Proc. SPIE 982, pp. 10­15, 1988.
4. B. Aschenbach, ''Re­Design of the XMM Optics'', Techn. Note XMM/O/MPE, 1987.
5. B. Aschenbach, H. Br¨auninger, K.­H. Stephan and Joachim Tr¨umper, ''X­ray test facilities at Max­Planck­
Institut Garching'', in Space Optics ­ Imaging X­ray Optics Workshop, M. Weisskopf, ed., Proc. SPIE 184,
pp. 234­238, 1979.
6. D. de Chambure, R. Lain`e, K. van Katwijk, A. Valenzuela, G. Grisoni et al., ''Lessons learnt from the
development of the XMM optics'', Intern. Symposium on the Optical Design and Production, Berlin 1999,
EUROPTO Series 1999, in press.
7. B. Aschenbach, ''Design of an X­Ray Baffle­System'', Techn. Note XMM­TS­FMP004, 1996.
8. D. de Chambure, R. Lain`e, K. van Katwijk, W. R¨uhe, D. Schink et al., ''The X­ray Baffle of the XMM
telescope: Development and Results'', Intern. Symposium on the Optical Design and Production, Berlin
1999, EUROPTO Series 1999, paper no. 3737­53, in press.
9. E. Pfeffermann, U.G. Briel, H. Hippmann, G. Kettenring, G. Metzner et al., ''The focal plane instrumentation
of the ROSAT telescope'', Proc. SPIE 733, pp. 519­532, 1986.
10. R. Egger, B. Aschenbach, H. Br¨auninger, W. Burkert, T. D¨ohring and A. Oppitz, ''X­ray calibration of the
mirror module FM2 at PANTER test facility ­ test results'', Techn. Note XMM­TS­PA063/970829, 1997.
11. R. Egger, B. Aschenbach, H. Br¨auninger, W. Burkert, T. D¨ohring and A. Oppitz, ''X­ray calibration of the
mirror module FM3 at PANTER test facility ­ test results'', Techn. Note XMM­TS­PA070/980115, 1998.
12. R. Egger, B. Aschenbach, H. Br¨auninger, W. Burkert, K. Molthagen and A. Oppitz, ''X­ray calibration of
the mirror module FM4 at PANTER test facility ­ test results'', Techn. Note XMM­TS­PA077/980723,
1998.
13. U.G. Briel, B. Aschenbach, M. Balasini, H. Br¨auninger, W. Burkert, K. Dennerl et al., ''The in­orbit Perfor­
mance of the EPIC­PN CCD­Camera on Bord of XMM­Newton'', in X­ray Optics, Instruments and Missions,
J. Tr¨umper and B. Aschenbach, eds., Proc. SPIE this volume, 2000.
14. L. Str¨uder, N. Meidinger, E. Pfefferman, R. Hartmann, H. Br¨auninger et al., ''X­ray pn­CCDs on the XMM­
Newton Observatory'', in X­ray Optics, Instruments and Missions, J. Tr¨umper and B. Aschenbach, eds.,
Proc. SPIE this volume, 2000.
15. J.W. den Herder, A.J. den Boggende, G. Branduardi­Raymont, A.C. Brinkman, J. Cottam et al., ''Descrip­
tion and performance of the reflection grating spectrometer on board of XMM­Newton'', in X­ray Optics,
Instruments and Missions, J. Tr¨umper and B. Aschenbach, eds., Proc. SPIE this volume, 2000.