3．2．XMM 1四レγZ701V 55
Angular Resolution
The EPIC MOS and PN cameras have pixels with sizes of 40 and 150μm， respectively For 七he focal length of the X−ray七elescopes（7．5 m），these pexel size corresponds to 1．1 arcsec and 4，1 arcsec on the sky Since they are smaller than the HEW of XRT（15 arcsec），
EPIC，s angular resolution is basically determined by the PSF of the mirror modules．
Energy Resolution
The resolving power of EPIC cameras is de七ermined by the intrinsic energy resolutioll of 七he individual pixels． Figure 3．26 and 3．26 show the energy resolution（FWHM）of MOS and PN， respectively The measured in一且ight FWHM of the Al Kα（1．5 keV）and Mn Kα
（5．9keV）， which are the on−board calibra七ion lines， are also plotted in Figure 3．26． It is well known that the energy resolution of MOS cameras has been gradually decrease due to the CTI（charge transfer inefHciency）efFect， which means the imperfect七ransfer of charge as it is transported through the CGD to七he output amplifiers． The latest calibration status is found at XMMlVeω舌oηScience Operation Centre．1 The accuracy of the energy determination is about 10 eV over the full energy range and for all modes except for MOS
timing mode．
8
凶
婁
冨 ニ エ
臣8
吊
o σrbit B4
● o由it 156
■ orbn 262 ロ o巾H361
■ o巾it 40ア ロ
冨
：
き
§
§
§
56 CHAPTER 3．1NSTRσMEN盟TION Quantum Ef且ciencies
The quantum e伍ciency of both七ypes of EPIC CCD chips as a func七ion of photon en−
ergy is displayed in Figure 3．27 and 3．27． These chips were calibra七ed using laboratory X−ray beams， synchrotron generated monochromatic X−ray beams， before launch， and ce−
les七ial X−ray source measurements． We can see the typical X−ray absorption fine structure
（XAFS）behavior around the silicon K edge a七1．838 keV． Ground calibration measure−
ments have shown that the quantum ef丘ciency of MOS CCDs is uniform above 400 eV．
Below this energy， spatial variations are seen as patches in the outer parts of the CCDs where the response is degraded． This inhomogeneity is currently not taken into account
by七he XMM−1Ve励oηscience analysis system（SAS）．
巴
♂巴
．蔓
トを5
ε 寸 ヨ oξ
8
2
o
1．o
0．8 ξ
：6∩．6 苫
…
盲。．イ ゴ
∩．儂
o．｛〕
／φ．r碗〆一一 、、 ・爆 ：凶一 才4 、 一
〆
翼鴨魁 上 寸一 ｝k．
1．850 Lε 10 LE5コ 1．63〔｝ t370 −
Fm・：yF・ ゴ］ 一 100 1000 104 0・1ど 1 1C 15
E．，，9ンζ，V） 三…gy［口V］
Fig．3．27：Left：Quantum efnciency of the EPIC MOS camera as a function of photon
energy Right：Quantum ef丘ciency of the EPIC PN camera as a function of pho七〇n energyLEPIC Filters
The EPIC CCDs are no七〇nly sensitive to X−ray pho七〇ns， but also to IR， visible and UV light． Therefore， if an astronomica1七arget has a high optical to X−ray flux ra七io， there is
apossibility七hat the X−ray signal becomes contaminated by those photons． To prevent
such a contribution， each EPIC camera is equipped with a set of 3 separate aluminised optical blocking filters， named批cん，ηzε∂九m and批η． The thick filter should be used for all point source七argets up to mγof 1−4（MOS）or O−3（PN）． The medium filter is about 1031ess ef丘cient than the thick filter，七herefore， it is useful for preventing optical contamination from point sources as bright as mγ＝8−10． The thin filter is about 105 1ess e伍cient than the thick filter， so七he use of this filter will be limited七〇point sourceswith optical magnitudes about 14 magnitudes fainter than the corresponding thick丘lter
limit ations．
Event Pattern
An absorbed sometimes deposits its energy over more than one pixels． This is called split event， and in this case the charges mus be summed up over the relevent pixels． This
3．2，XMM∫VEWτ0∧「 57
process is automatically done by analysis software． The split pattern is classi6ed i皿Figure 3．28．The patterns O12 fbr MOS and住4 fbr pn are considered to be X−ray events， while the others are false events induced by charged particles． Because of its much larger I）i）【el size than］㌧10S， the charge split occurs less丘equently in pn（Turner et al．2001），Any events which located at around an edge or bad pbζel are flagged by negative value．
These events have possibility that the ellergy of these events aエe not correct， If we make acondition that the events have flag＝0， we remove the events which are rocated around the edge ore bad pixeL
τm3gipg P分町甘n nh｜綬11s芭d in gn、11nd ca｜ihr角●ion alld in臼ighr
蝋 i藁鏑；
15 30 5．
i幽i鞘蝋i幽ii幽蹴蝋輯
9 10 ｝1 1ユ 25 2〔 28 〜9
i蝉牌…劉…禦酬蒔…
i囲…i圃i閣…聞…鰯醗縫…
i麗 蝋i酩曜
O ，＋ 15 1丙 17
Fig．3．28：Event pattems recognized by the MOS（pn）detectore． The red pixel is the
centre pixel， its signal is above threshold and is the largest signal in the 3×3 inner ma−
trix． The green pixels have signals above threshold． The white pixels have signal bdow threshold． The crosses indicate pixels no considered．
3．2．3 EPIC Background
The EPIC background can be divided into two parts：acosmic X−ray background（CXB）．
and an instrumental background． The latter compollent may be further divided illt o a detector noise component， which becomes important at low energies（i．e． below 200eV）
and a second component which is due to the particles interaction． This compo皿ent is characterized by a flat spectrum and is particularly important at high energies（i．巳above afew keV）．
The particle induced background can be divided into two components：an externaI
「flaring component， characterized by strong and rapid variability， which is often totally absent and a second more stable internal component． The aaring component is currentLy attributed to soft protons， which are presumably funneled towards the detectors by the X−ray mirrors． The stable component is due to the interaction of higしenergy−particle5
with the structure surrounding the detectors and possibly the detectors themselves．恥 summarize the all background component below．
In the following we describe some of the main properties of both components．
58 CHAPTER 3， JNS肥RσMEN題丁τON Back Ground
・Cosmic X−ray Background
・】hstru皿ental Backgrou皿d
−Detector noise component（below 200eV）
−Particle component（above a fヒw keV）
−naring component（a㎞buted soft photons）
−stable component（at㎞buted high−energy par6cles）
Temporal Pmperties
As shown in Figure 3．29， the EPIC background count rate often exhibits sudden increases by as large as two orders of magnitudes， called Hares1． Such phellom銀a are not observed in the、45乙4 SIS， This is mainly due to the difference in their orbits． A5α4 had an almost circular orbit with an altitude of 520−620 km， while XMM−Newton take highly eccentric orbits， with apogees of〜115，000 km and perigees of〜6，000 km． Therefore．
XMM−Newton且y mostly outside the Earth s magnet（テsphere． Now it is㎞own that the
background flares are caused by sof亡protons with ellergies below l MeV， re且ected and 長）cused by the X−ray mirrors． The spectra of soft proton flares are variable and no clear correlation is fOund between intensity and spectral shape． The current understanding is that soft protons are most likely organized in clouds populating the Earth s magneto−sphere， The number of such clouds encountered by XMM−Newton in its orbit depends
upon many factors， such as the altitude of the satellite， its position with respect to the magneto−sphere， and the amount of solar activity．30
■ξ