68 CHAPTER 4． OB8ERVATZON AND DA工4 REDσC質ON
Extended Spectral Analysis Software（XMM−ESAS）1．We utilized blank sky data to es−
timate background illtensities． Hayakawa 2006 carried out background subtraction with
blank sky data． However， they did not utilize exposure−corrected images for blank sky．In the outer region， we must care for background subtraction in slite photon counts． We carried out correction of background with exposure map．
In§3．2．3， we described the EPIC background properties， and found七ha七the back−
ground was divided into following components．
4．2．DA．ZX REDσσTION 69
（a）MOS1 （b）MOS2
DLコS句50301 田ヨ59503川
ロ ヂ《L垣辿：βh白 COumR■ e
z§ z§．
0 1 2 3 4 5 0 i 2 3 4 s Cm工1亀t6〔c㎜創3） CwびR寂〔c側戊！s｝
。竺
P ． ・V・・蝋 ． 1 。三 ・・⑭⑭
1：｝騨榊』｛酬く⌒亘1：榊酬く⌒脚｛岬⌒
． ≡一§一
り 0 5000 104 1．5刈r 2刈W Zづ厩lr 3xlO角 U O 50．0 194 1．5方10司 Z刈04 ユ．きx10冶 3刈04 T限1．〕 T暇ほ〕
ヤ
官苫r ． ．c輌L仙・（迦 言6 c㎜n幽c輌
ヨ 熱 〔 ．
∨ 0 5000 卵 1．5・1ぴ 2「】r 2．了・10冶 ヨ×1ぽ コ 0 ∫000 10■ 1．5ダ1ひ ユ×｜伊 2．撒×1 3qr τ血β㈲ τロ鰺「5）
ii幽くこ酬貞綱く、・｝，竃二酬｝酬綱酬｝
Fig．4．3：Light curve and its count rate histogram of AWM7 reprocessed data（MOS1（a）
and MOS2（b））．
names（EVENTS， ATTHK， ORBIT）for their type， and depending on the type， certain
attributes are checked to see whether the files are compatible with each other． In the at−titude and orbit cases， any 61es can be merged together（so long as they are both attitude or both orbit丘1es），whereas in the events case， some care should be taken，
mεηεcommand can correct difference between two d澁rent cloth pointing direction，
We processedπL町εcommand fbr all event files of A1413 and A2204． And net exposure
time is 122．4 and 12Lg ks fOr MOSI and MOS2 in the case of A1413， and 70．9 and 72．5ks in the case of A2204． We aIso merged attitude丘les of all observations of A1413 and A2204．Blank Sky Data as Background
We are飴rced to extract the background spectrum丘om the blank sky data sets of Read
＆Ponman（2005）in the same detector regiol1， in order to exclude the background events of（1）一（3）． In this case， the remaining problem is that the non X−ray background（2），
（3）is time variable， although the CXB（1）is Ilearly identical everywhere． Katayama et aL（2004）studied this problem and found that the count rate in the high energy balld
（above 10 keV）well represent the particle origin background． Therefore， the sourc合t（ト background count rate ratio was calculated丘om the count rates in the bands 10−12 keV and 12−14 keV for MOS and the backgroulld spectrum is scaled by this factor when it is subtracted．
Chapter 5
Background Analysis
Accurate estimation of七he background is particularly important when constraining the ICM surface brightness and tempera七ure in the outer region of clusters． We assumed that the background is comprised of three components：non−X−ray background（NXB），
cosmic X−ray background（CXB）and Galactic emission（GAL）， which itself is comprised of七wo components． In this section we describe how we estimate all these background components．
5．1 Point Source Analysis
We want to excise point sources because we are only interested in this paper in the ICM．
However， since the CXB is comprised of faint point sources， we then need to correct the background level for the resolved sources． This and the next section describe the procedure we used for these tasks．
In the case of A1413， We used the XMM−Newton image to detect point sources in the
XIS FOV because its spatial resolu七ion（14 half power diameter；HPD）is better than
Suzaku，s（2／HPD）． We detected 10 point sources usingψα掘e舌ec舌of GIAO， and extracted source and background spectra by setting the extraction radius of 33 and 33 −66 ，respectively First， we checked七hat the MOSI and MOS2 spectra of each source were
consistent． Then we summed the MOSI and MOS2 spectra to increase七he statistics， and fitted the spectrum of each source to evaluate individual spectral parameters． Finally we added the spectra of all the point sources to estimate how much of the CXB these sources resolve． We丘tted the spectra byωα6s×ρεgpωrIω． The best一丘t parameters for the individual point sources and their sum are shown in table 5，1， We obtained X2／dof ニ87．2／77for the power−law丘t to the combined spectrum（figUre 5．1（a））， iIldicating a reasonable spectral fit． The photon index is r＝1．92±0．09 and the flux is 3．23±8：2×10−13 erg cm−2 S−1．We also searched for point sources located ou七side of the XMM−Newton丘eld with Suzaku， finding an additional five sources by eye in A1413， We found ou七probabilities of point source detection by eye in table5．2．
71
72 CHAPTER 5． BA CKGROσND ANALYS阻
We can calculate probabilities of point source detection with following equa七ion，
q， 0・BS一雅（㌔9
δsr
@・・BS＋（AsrAbg）2α、
（5．1）
，in which（㌔r， OoBs， Obg are row counts of ICM， observed spectra， and background spec−
tra． δs， is fluctuationもof source counts． ノ4sr and／1bg are source and background area respectively We could detect poin七sources larger than 3．9σas shown in table 5．2． We performed spectral丘ts to all the point sources with Suzaku according to the following procedure． The source photons came from a circle of 40 radius with encircling anllular background region of 40， −100 radii． We selected the source regions so they did not overlap each other． These source and background areas could be slightly difFerent among the detectors and sources due to filtering by the cα1γηα5えregions and the presence of hot pixels． We added the FI spectra f士om XISO， XIS2， and XIS3 detec七〇rs， and summed the
BACKSCAL keyword in the FITS header， which correspond to the area of extraction
region，．4s， oL4bg． Then， we carried out spectral fi七s for the FI and BI spectra simul七a−neously using the same spectral model as before，丘rst for the individual sources and then the sum of all the point sources．
We show the best一丘t parameters for the individual point sources and their sum in table 5．1 excep七for sources O4，05， and O8， because they were fain七so that we could not estimate their background reasonably in A1413． Obtained fiuxes of the sources are slightly afFected by leaked photons of the target to the surrounding background regions・
To correct for this efFect， we calculated the ratio，五eak， of the leaked photons in each background region七〇七he detected photons in the source region using the 励582m FTOOL
（Ishisaki e七al．2007）． We corrected七he original source且ux by multiplying a factor 1／（1一 九eak．4sr／．4bg）＝1／（1−0．2力eak）in the五〜（columns of Suzaku in table 5．1． Figures 5．1（b）
and（c）show the combined spectra of all sources for FI and BI for A1413． In the case of A1413， we obtained）（2／dofニ113．1／117 for the power−1aw丘t to the combined spectrum，
indicating a reasonable spectral丘七， too． The photon index is I「＝1．82士0．12 and the且ux is 4．83±1：ll×10−13 erg cm−2 s−1（2−10 keV）．
The total flux of all the point sources agrees well between Suzaku and XMM−Newton within the error in A1413． The number of sources we found and their total flux are consistent wi七h that expected from the log N−10g S rela七ion summarized in丘gure 200f
Kushino et al．（2002）in A1413． The detected sources ranges from〜10−14 to〜10−13
erg cm−2 s−1． We excised all the point sources detected in either the Suzaku or XMM−Newton observations． Normally we excluded a region of 70 radius but used 125 radius for two sources（09 and 14 in table 5．1）in A1413．
5．2 Stray I、ight
We examined how many photons accumulated in the each region of A1413， A2204， and
AWM7 actually came from somewhere else on the sky because of the extended telescope
5．2．8TRAY L∫GHT 73
め
H
寸
H
＜ 甘
器 呂
8
の
甘
6
自 勺㊤
輌
o
㊤輌㊤ 廿
］
O
oo Φ
ち
日
』
輌
o
唱山
6
口 口 一①
卓
←
官
B
ら㊤
1
×
）
OQN O
十 OQ∋