1: %\documentstyle[12pt,aasms4,pdflscape]{article} % per sottoporre
2: \documentclass[12pt,preprint]{aastex}
3: %\documentstyle{aastex}
4: %\usepackage{aastexug}
5: %\documentstyle[11pt,aaspp4,pdflscape]{article} % preprin 1 colonna
6: %\documentstyle[emulateapj]{article}
7: %\documentstyle[aas2pp4,pdflscape]{article} % preprint 2 colonne
8: %\received{4 August 1988}
9: %\accepted{23 September 1988}
10: %\journalid{337}{15 January 1989}
11: %\articleid{11}{14}
12: % \slugcomment{To appear in ...}
13: %\righthead {...}
14: %\usepackage{pdflscape.sty}
15:
16: \def\etal{{\it et~al.~}}
17: \def\nt{{nonthermal~}}
18: \def\bsax{{\it BeppoSAX~}}
19: \def\integral{{\it INTEGRAL~}}
20: \def\ginga{{\it Ginga~}}
21: \def\einstein{{\it Einstein~}}
22: \def\exosat{{\it EXOSAT~}}
23: \def\tenma{{\it Tenma~}}
24: \def\asca{{\it ASCA~}}
25: \def\rosat{{\it ROSAT~}}
26: \def\xmm{{\it XMM-Newton~}}
27: \def\chandra{{\it Chandra~}}
28: \def\rxte{{\it RXTE~}}
29: \def\kbeta{$K_{\beta}$~}
30: \def\kalpha{$K_{\alpha}$~}
31: \def\pho{~{\rm ph~cm}^{-2}~{\rm s}^{-1}~{\rm keV}^{-1}~}
32: \def\erg{~{\rm erg~ cm}^{-2}\ {\rm s}^{-1}~}
33: \def\ergs{~{\rm erg~s}^{-1}~}
34: \def\cts{~{\rm counts~s}$^{-1}$~}
35:
36: %\shorttitle{HXR excess in Abell 2256}
37: %\shortauthors{R.Fusco-Femiano \etal}
38:
39: %\doublespace
40: %\received{2003 November 21}
41: \begin{document}
42:
43: \newcommand{\lessim}{\ \raise -2.truept\hbox{\rlap{\hbox{$\sim$}}\raise5.truept
44: \hbox{$<$}\ }}
45:
46: \title{Nonthermal hard X-ray excess in the Coma cluster: resolving the discrepancy
47: between the results of different PDS data analyses}
48:
49: \author{Roberto Fusco-Femiano$^a$, Raffaella Landi$^b$, Mauro Orlandini$^b$
50: \footnote{$^a$Istituto di Astrofisica Spaziale e Fisica Cosmica
51: (IASF/Roma), INAF, via del Fosso del Cavaliere, I--00133 Roma,
52: Italy - roberto.fuscofemiano@iasf-roma.inaf.it; $^b$IASF/Bologna,
53: INAF, via Gobetti 101, I--40129 Bologna, Italy -
54: landi@iasfbo.inaf.it - orlandini@iasfbo.inaf.it}}
55:
56: %; $^c$Istituto di Radioastronomia , C.N.R., via Gobetti 101,
57: %I--40129 Bologna, Italy; $^d$Dip. di Astronomia, Univ. Bologna,
58: %via Ranzani 1, I--40127 Bologna, Italy}}
59:
60: \affil{}
61:
62: \begin{abstract}
63: The detection of a \nt excess in the Coma cluster spectrum by two
64: \bsax observations analyzed with the XAS package (Fusco-Femiano
65: \etal) has been disavowed by an analysis (Rossetti \& Molendi)
66: performed with a different software package (SAXDAS) for the
67: extraction of the spectrum. To resolve this discrepancy we
68: reanalyze the PDS data considering the same software used by
69: Rossetti \& Molendi. A correct selection of the data and the
70: exclusion of contaminating sources in the background determination
71: show that also the SAXDAS analysis reports a \nt excess with
72: respect to the thermal emission at about the same confidence level
73: of that obtained with the XAS package ($\sim 4.8\sigma$). Besides,
74: we report the lack of the systematic errors investigated by
75: Rossetti \& Molendi and Nevalainen \etal taking into account the
76: whole sample of the PDS observations off the Galactic plane, as
77: already shown in our data analysis of Abell 2256 (Fusco-Femiano,
78: Landi \& Orlandini). All this eliminates any ambiguity and
79: confirms the presence of a hard tail in the spectrum of the Coma
80: cluster.
81:
82: \end{abstract}
83:
84:
85: \keywords{cosmic microwave background --- galaxies: clusters:
86: individual (Coma) --- magnetic fields --- radiation mechanisms:
87: non-thermal --- X--rays: galaxies}
88:
89: \section{Introduction}
90:
91: Nonthermal hard X-ray (HXR) emission was predicted in the
92: seventies in clusters of galaxies showing extended radio regions,
93: radio halos or relics, since the same radio synchrotron electrons
94: can interact with the CMB photons to give inverse Compton (IC)
95: X-ray radiation (Perola \& Reinhardt 1972; Rephaeli 1979). Several
96: attempts to detect hard tails in the spectrum of a few clusters of
97: galaxies were performed with various experiments (Bazzano \etal
98: 1984,90; Rephaeli, Gruber \& Rothschild 1987; Rephaeli \& Gruber
99: 1988; Rephaeli, Ulmer \& Gruber 1994) that reported only upper
100: limits to the \nt flux. A significant breakthrough in the
101: measurement of \nt HXR emission was obtained thanks to the
102: improved sensitivity and wide spectral capabilities of the \bsax
103: and Rossi X-Ray Timing Explorer (\rxte) satellities. As pointed
104: out by Petrosian (2003), the discovery of \nt HXR radiation has
105: led to a remarkable increase of the theoretical investigations
106: regarding the possible acceleration mechanisms and origin of the
107: relativistic electrons responsible for the \nt emission, although
108: the presence of \nt phenomena in the intracluster medium (ICM) of
109: some clusters was established decades ago (Willson 1970).
110:
111: Nonthermal HXR radiation was detected in excess of the thermal
112: emission in the Coma cluster by a first \bsax observation
113: (Fusco-Femiano \etal 1999) using the Phoswich Detection System
114: (PDS) and confirmed by a second independent observation with a
115: time interval of about 3 yr (Fusco-Femiano \etal 2004; thereafter
116: FF04). The presence of a second component in the X-ray spectrum of
117: the cluster has been reported also by two \rxte observations
118: (Rephaeli, Gruber \& Blanco 1999; Rephaeli \& Gruber 2002). A2256
119: is the second cluster where a \nt excess has been measured by two
120: \bsax observations (Fusco-Femiano \etal 2000; Fusco-Femiano \etal
121: 2005) and by \rxte (Rephaeli \& Gruber 2003). At a lower
122: confidence level, with respect to Coma and A2256, \nt HXR
123: radiation has been detected by \bsax in A754 (Fusco-Femiano \etal
124: 2003). An upper limit to the \nt flux has been reported in A3667
125: (Fusco-Femiano \etal 2001), A119 (Fusco-Femiano \etal 2003) and
126: A2163 (Feretti \etal 2001). For the last cluster a \rxte
127: observation shows instead the presence of a HXR excess (Rephaeli,
128: Gruber, \& Arieli 2006). \rxte reports also some evidence of \nt
129: emission by the Bullet Cluster (Petrosian, Madejski, \& Luli
130: 2006).
131:
132: The PDS spectra of all the \bsax observations were extracted using
133: the XAS version 2.1 package (Chiappetti \& Dal Fiume 1997)
134: specifically created to handle the PDS peculiarities. However, a
135: PDS data analysis performed with a different software package
136: (SAXDAS) does not report evidence for the presence of a hard tail
137: in spectrum of the Coma cluster (Rossetti \& Molendi 2004;
138: thereafter RM04).
139:
140: To solve this contradiction we reanalyzed the PDS data of the two
141: \bsax observations using the SAXDAS 2.0.2 package and in this
142: letter we present our results and conclusions. In Sect. 2 we show
143: the results of this analysis that are compared with those reported
144: in FF04 using the XAS package and we examine the possible
145: systematic errors investigated by RM04 and Nevalainen \etal (2004,
146: thereafter NE04). In Sec. 3 we discuss the reasons that have led
147: to controversial results using different software packages.
148: Finally, Sect. 4 is devoted to the conclusions regarding the
149: presence of a hard tail in the spectrum of the Coma cluster.
150:
151: Throughout this Letter we assume a $\Lambda$CDM cosmology with
152: $H_o$ = 70~km~s$^{-1}$~Mpc$^{-1}$~$h_{70}$, $\Omega_m$ = 0.3 and
153: $\Omega_{\Lambda}$ = 0.7. An angular distance of $1^{\prime}$
154: corresponds to 29.0 kpc ($z_{Coma}$ = 0.0232). Quoted confidence
155: intervals are at $90\%$ level, if not otherwise specified.
156:
157: \section{PDS Data Reduction and Results}
158:
159: \begin{figure}
160: \centering \rotatebox{-90}{ \epsscale{0.8} \plotone{f1.ps}}
161: \caption{PDS combined spectrum of the Coma cluster obtained with
162: the XAS package (FF04). The continuous line represents the best
163: fit with a thermal component at the average cluster gas
164: temperature of 8.11 keV (David \etal 1993). The error bars are
165: quoted at the 1$\sigma$ level. The spectrum starts at 15 keV.
166: \label{fig1}}
167: \end{figure}
168:
169:
170: The Coma cluster was observed for the first time by the PDS
171: instrument (Frontera \etal 1997) in December 1997 for $\sim$91 ks
172: and re-observed in December 2000 for $\sim$300 ks. The pointing
173: coordinates of \bsax are at J(2000): $\alpha:~12^h~ 58^m~ 52^s$;
174: $\delta:~ +27^{\circ}~ 58'~ 54''$. To resolve the discrepancy
175: between the results shown by FF04 and RM04 regarding the presence
176: of a \nt component in the HXR Coma spectrum, we have re-analyzed
177: the PDS data with the SAXDAS package used by RM04. However, we
178: start by reporting the procedure used in the XAS analysis and the
179: main results obtained. More details can be found in FF04.
180:
181: \subsection{XAS analysis}
182:
183: The PDS data used in the XAS analysis were selected by a first
184: automatic procedure followed by a visual check in order to
185: eliminate all the remaining spikes that can affect significantly
186: the analysis results. The effective exposure times in the two PDS
187: observations were 44.5 ks and 122.2 ks, respectively (thereafter
188: OBS1 and OBS2). Since the source is rather faint in the PDS band
189: ($\sim$5 mCrab in 15--100 keV) a careful check of the background
190: subtraction was performed, making use of the default rocking law
191: of the two PDS collimators that samples ON/+OFF, ON/--OFF fields
192: for each collimator with a dwell time of 96 sec (Frontera \etal
193: 1997). When one collimator is pointing ON source, the other
194: collimator is pointing toward one of the two OFF positions.
195: Initially, we used the standard procedure to obtain PDS spectra
196: (Dal~Fiume \etal 1997); this procedure consists of extracting one
197: accumulated spectrum for each unit for each collimator position.
198: We then checked the two independently accumulated background
199: spectra in the two different +/--OFF sky directions, offset by
200: $210'$ with respect to the on-axis pointing direction (+OFF
201: pointing: $\alpha:~12^h~58^m~ 57.8^s$; $\delta:~ +24^\circ~ 28'~
202: 55''.1$ --OFF pointing: $\alpha:~12^h~58^m~ 47.0^s$; $\delta:~
203: +31^\circ~ 28'~ 54''.7$). The comparison between the two
204: accumulated backgrounds (difference between the +OFF and --OFF
205: count rate spectra) showed that for OBS1 the difference was
206: compatible with zero ($0.044\pm 0.047$ \cts for a background level
207: of $21.66\pm 0.02$ \cts in 15--100 keV), while for the longer,
208: more sensitive OBS2, there was an excess of $0.064\pm 0.021$ \cts
209: (background $16.76\pm 0.01$ \cts)\footnote{The $\sim$20\%
210: variation in the PDS background is due to the \bsax orbital decay:
211: the lower orbit for OBS2 increased the shielding to ambient
212: particles, therefore lowering the diffuse background.}. As
213: reported in FF04, a careful check of possible variable sources in
214: the PDS offset fields led the attention to the BL~Lac source
215: 1ES~1255+244, present in the +OFF field, that was observed by
216: \bsax on May 1998 in the framework of a spectral survey of BL~Lacs
217: by Beckmann \etal (2002). Because of the very short exposure time
218: ($\sim$ 3~ksec), our analysis of the source has determined only a
219: 2$\sigma$ upper limit of 0.26 \cts in 15--100 keV, corresponding
220: to 1.6 mCrab, however compatible with the background excess
221: measured in OBS2. Moreover, just in the center of the +OFF field
222: is also present the extremely weak ROSAT source
223: RX~J125847.1+242741. The presence of these contaminating sources
224: justified the decision to exclude the +OFF field in the background
225: evaluation and consider only the --OFF field as the
226: "uncontaminated" background for both the Coma observations.
227:
228: The combined spectrum, obtained by summing the spectra of the two
229: observations (OBS1 \& OBS2), shows a \nt hard excess with respect
230: to the thermal component at the confidence level (c.l.) of $\sim
231: 4.8\sigma$ in the 20-80 keV energy range considering only the -OFF
232: background spectrum. The spectrum is reported in Fig.1. FF04 also
233: reported the confidence level of the \nt excess considering the
234: standard procedure, i.e. by using as background the average of the
235: spectra extracted from both the two offset fields. The c.l. is
236: lower ($\sim 3.9\sigma$) due to the contamination in the +OFF
237: background. The average gas temperature used in the analysis is
238: that measured by \ginga of 8.11$\pm$0.07 keV (David \etal 1993) in
239: a field of view comparable to that of the PDS (FWHM $\sim
240: 1.3^{\circ}$). This value of the average temperature is confirmed
241: by a determination of \rxte that reports a best-fit temperature of
242: 7.90$\pm$0.03 keV (Rephaeli \& Gruber 2002) in a field of view of
243: $\sim 1^{\circ}$ comparable to the field of view of \ginga and
244: PDS.
245:
246: \subsection{SAXDAS analysis}
247:
248: The first combined spectrum was extracted considering the PDS data
249: resulting by the automatic selection operated by the SAXDAS
250: package \footnote{See
251: http://www.asdc.asi.it/bepposax/software/saxdas/}. The confidence
252: level of the excess, in the range 20-80 keV and for an average gas
253: temperature of 8.11 keV, is $\sim$ 2.9$\sigma$ (observed count
254: rate = 0.1717$\pm$0.0146 \cts, model predicted rate = 0.1295 \cts)
255: taking into account only the uncontaminated -OFF position. The
256: effective exposure time is 42.8 ks in OBS1 and 119.3 ks in OBS2
257: for a total exposure time of 162.1 ks. To resolve the discrepancy
258: between these results and those obtained by FF04 with a XAS
259: analysis (see previous section), a new combined spectrum was
260: extracted considering the \textit{same} time windows used in the
261: XAS analysis. These time windows are different from those used by
262: RM04 in their SAXDAS analysis. The reason is in the different
263: criterion for eliminating the spurious spikes produced by the
264: charged particles that interact with the PDS detectors: a
265: semi-automated procedure (i.e. an automated procedure followed by
266: a visual check) for XAS and a completely automated procedure for
267: SAXDAS (Guainazzi \& Matteuzzi 1997). With the same time windows
268: used by XAS, the c.l. of the excess in the SAXDAS spectrum raises
269: at $\sim$ 4.2$\sigma$ (observed count rate = 0.1902$\pm$0.0148
270: \cts, model predicted rate = 0.1280 \cts), while the total
271: exposure time reduces to 160.9 ks (it was 162.1 ks).
272:
273: \begin{figure}
274: \centering \rotatebox{-90}{ \epsscale{0.8} \plotone{f2.ps}}
275: \caption{PDS combined spectrum of the Coma cluster obtained with
276: the SAXDAS package (see text). The continuous line represents the
277: best fit with a thermal component at the same average cluster gas
278: temperature of Fig. 1. The error bars are quoted at the 1$\sigma$
279: level. The spectrum starts at 12 keV.\label{fig2}}
280: \end{figure}
281:
282:
283: The two packages, starting from the same time windows, produce a
284: difference in the total effective exposure time of 5.8 ks
285: ($t_{XAS}$ = 166.7 ks, $t_{SAXDAS}$ = 160.9) essentially due to:
286: a) the different earth angle above which data are selected (the
287: earth angle is defined as the angle between the earth limb and the
288: \bsax pointing direction). The earth angle is 5$^{\circ}$ for XAS
289: and 10$^{\circ}$ for SAXDAS; b) the removal of observational time
290: after any South Atlantic Geomagnetic Anomaly (SAGA) passage. The
291: SAXDAS package removes 5 minutes after any passage, while XAS
292: eliminates the time necessary to reach the correct voltage of the
293: instruments after their shut-down during the SAGA passage.
294:
295:
296: In conclusion, we have a lower exposure time for the SAXDAS
297: analysis which may imply a lower c.l. of the \nt excess. To
298: quantify this difference we have extracted a further combined
299: spectrum using the same time windows selected in the XAS analysis
300: but imposing an earth angle of 5$^{\circ}$ in the SAXDAS package.
301: The total effective time exposure of the two observations is now
302: $t_{SAXDAS}$ = 169.1 ks instead of 160.9 ks. The excess results to
303: be at the c.l. of $\sim$ 4.6$\sigma$ (observed count rate =
304: 0.1944$\pm$0.0144 \cts, model predicted rate = 0.1280 \cts). This
305: HXR spectrum is reported in Fig. 2. The difference in the initial
306: energy of the SAXDAS and XAS spectra (SAXDAS spectrum starts at 12
307: keV, while XAS at 15 keV) gives insignificant variations in the
308: c.l. values of the \nt excess.
309:
310: \subsection{Possible systematic errors}
311:
312: The systematic errors examined by RM04 and NE04 have already been
313: discussed by us in the PDS analysis of A2256 (Fusco-Femiano, Landi
314: \& Orlandini 2005), but considering their importance in the
315: analysis results we intend to report here again the main parts of
316: the discussion.
317:
318: 1) RM04 claim the presence of an "Instrumental background
319: residual" (see sect. 2.1 of their paper) derived from the analysis
320: of 15 "blank fields", i.e. fields which do not contain sources
321: showing significant emission in the PDS energy range. By summing
322: the spectra from these observations they find that the spectrum
323: differs from zero: the count rate is ($1.45\pm 0.77)\times
324: 10^{-2}$ \cts in the 12-100 keV energy range. This seems to
325: indicate that the background in the ON position is larger than
326: that in the $\pm$OFF positions producing an instrumental
327: contribution not removed by the background subtraction procedure.
328: This effect has been studied by Landi \etal (2005) considering the
329: complete sample of 868 PDS pointings with galactic latitude
330: $|b|>15^\circ$, and selecting the 15--100 keV net count spectra
331: for which there is source detection below 1$\sigma$ (that is,
332: ``blank fields''). These spectra have been summed imposing a net
333: exposure greater than 20~ksec. A net count rate of $(1.67\pm
334: 5.30)\times 10^{-3}$ \cts has been derived, consistent with the
335: definition of ``blank field''. Also NE04 do not report evidence
336: for an instrumental residual.
337:
338: 2) The other effect evaluated by RM04 regards the systematic
339: differences between the background fields. They analyze a sample
340: of 69 observations whose target is outside the galactic plane and
341: with a long exposure time (see Appendix of their paper). RM04 find
342: that the mean value of the difference between ON and the two -OFF
343: and +OFF sky positions is significantly different from zero and
344: positive. Also this effect has been investigated by Landi \etal
345: (2005) on the {\em whole} sample of PDS observations. The obtained
346: value of $(5.3\pm 6.3)\times 10^{-3}$ \cts is consistent with no
347: contamination at all. We presume that the value found by RM04
348: could be due to the small sample of observations they considered.
349: NE04, analyzing a larger sample of data with respect to that used
350: by RM04 (164 PDS observations), found a systematic difference
351: between ON and the two offset pointings that cancels out in the
352: standard usage of both offsets.
353:
354: 3) In addition, NE04 introduce a systematic error in the net
355: source count rates due to unresolved and not significantly
356: detected point sources present in the PDS field of view. They find
357: that an excess of 0.019 counts~s$^{-1}$ has to be added to the net
358: count rate spectra (no errors are given on this measurement), when
359: the standard method of background evaluation is used, and 0.027
360: counts~s$^{-1}$ has to be added when the background is evaluated
361: from only one offset field. We performed this same analysis on a
362: set of 868 observations (NE04 use 164 fields) and find that the
363: contribution of background fluctuations due to unresolved and not
364: significantly detected sources in the offset fields (in other
365: words, the PDS confusion limit) introduces a variance
366: $\sigma^2_{fluc} = (9.5\pm 10.3)\times 10^{-4}$
367: (counts~s$^{-1}$)$^2$ consistent with zero (Fusco-Femiano, Landi
368: \& Orlandini 2005). Therefore PDS data are not affected by this
369: systematic effect.
370:
371: \section{Discussion}
372:
373: Fig. 2 shows that also the PDS combined spectrum extracted with
374: the SAXDAS package evidences the presence of a \nt excess with
375: respect to the thermal emission at about the same c.l. reported by
376: the XAS analysis (FF04) ($\sim$ 4.8$\sigma$ for XAS and $\sim$
377: 4.6$\sigma$ for SAXDAS). The discrepancy between the results
378: reported in FF04 and RM04 is mainly due: a) to an accurate
379: selection of the events. In fact, we have shown that the SAXDAS
380: analysis of the same PDS time windows used with the XAS software
381: leads to a significant increase of the c.l. of the excess (from
382: $\sim 2.9\sigma$ to $\sim 4.2\sigma$); b) to a correct
383: determination of the background. The XAS and SAXDAS spectra shown
384: in Fig.1 and Fig. 2, respectively, are obtained considering only
385: the -OFF background. The exclusion of the +OFF position in the
386: background determination implies a more pronounced detection of
387: \nt radiation from the Coma cluster. The check regarding the
388: presence of contaminating sources in the +OFF field was not
389: operated by RM04.
390:
391: An increase of $\sim$ 0.4$\sigma$ in the c.l. value is obtained
392: adopting in the SAXDAS analysis the same earth angle of
393: $5^{\circ}$ used in the XAS procedure. The lower earth angle
394: implies a greater exposure time that justify an improvement in the
395: c.l. value of the \nt excess. Considering in the SAXDAS analysis
396: the same time windows and the same earth angle used in the XAS
397: analysis the difference in the total exposure time is 2.4 ks in
398: favour of the SAXDAS package ($t_{XAS}$ = 166.7 ks, $t_{SAXDAS}$ =
399: 169.1 ks) due to the different removal of the time operated by the
400: two packages after any SAGA passage. The XAS package results to be
401: more conservative waiting more time than SAXDAS to allow the PDS
402: high voltage to reach the correct levels. So, some spurious events
403: could be present in the SAXDAS analysis, in particular for the
404: longer OBS2 observation.
405:
406: In order to reproduce the results obtained by RM04 we have
407: computed the c.l. of the excess for OBS1 and OBS2 spectra (RM04 do
408: not report the c.l. value for the combined spectrum) using the
409: automatic selection of the events operated by the SAXDAS package
410: and the standard procedure for the background determination. The
411: c.l. values are 2.90$\sigma$ and 1.34$\sigma$ for OBS1 and OBS2,
412: respectively, in the energy range 25-80 keV and for an average gas
413: temperature of 8.21 keV. These values are not much distant from
414: those reported in Table 2 of RM04 (2.84$\sigma$ for OBS1 and
415: 1.11$\sigma$ for OBS2)\footnote{These values of the c.l. of the
416: excess refer to the line "No Subtraction" (no subtraction of the
417: instrumental residual that Landi \etal (2005) find consistent with
418: zero; see Sect. 2.3).} confirming that the selection procedure of
419: the PDS events and the exclusion of the +OFF position for the
420: background determination are the main reasons of the different
421: results reported in FF04 and RM04. The PDS data analysis of very
422: weak sources like Coma in the HXR band requires a rigorous
423: selection of the events in order to eliminate the presence of any
424: spikes able to introduce \textit{noise} that hides the presence of
425: a \nt excess with respect to the thermal radiation. The presence
426: of contaminating sources in the offset fields does not allow the
427: use of the standard procedure for the background evaluation.
428:
429:
430: \section{Conclusions}
431:
432: We have shown that the presence of a \nt excess with respect to
433: the thermal emission in the spectrum of the Coma cluster does not
434: depend on the used software package (XAS or SAXDAS) for the PDS
435: data analysis. The spectra extracted with XAS and SAXDAS, reported
436: in Fig. 1 and 2, respectively, show both a \nt excess at about the
437: same c.l. value when in the SAXDAS analysis the same time windows
438: used in the XAS analysis (FF04) are adopted and the +OFF sky
439: direction is not taken into account in the background
440: determination for the presence of contaminating sources. The
441: systematic effects claimed by RM04 and NE04 can be excluded
442: considering the whole sample of the PDS observations off the
443: Galactic plane.
444:
445: This re-analysis of the PDS data, using the SAXDAS package,
446: explains the different results reported in FF04 and RM04
447: confirming the presence of a \nt component in the Coma cluster
448: spectrum.
449:
450:
451: \section{Acknowledgments}
452: We wish to thank L.Feretti and G.Giovannini for a critical reading
453: of the paper and the referee for his helpful comments on the
454: manuscript.
455:
456: \newpage
457: \begin{references}
458:
459: \reference{}Bazzano, A., Fusco-Femiano, R., La Padula, C.,
460: Polcaro, V.F., \& Ubertini, P. 1984, ApJ, 279, 515
461:
462: \reference{}Bazzano, A. \etal 1990, ApJ, 362, L51
463:
464: \reference{}Beckmann, V. \etal 2002, A\&A 383, 410
465:
466: \reference{}Chiappetti, L., \& Dal Fiume, 1997, in Data Analysis
467: in Astronomy, ed. V.Di Gesu` \etal (Singapore: World Scientific),
468: 101
469:
470: \reference{}Dal Fiume, D. \etal 1997, Proc. of the Fifth Workshop
471: {\it "Data Analysis in Astronomy"}, eds.: V. Di Gesu' \etal , p.
472: 111
473:
474: \reference{}David, L.P., Slyz, A., Jones, C., Forman, W., \&
475: Vrtilek, S.D. 1993, ApJ, 412, 479
476:
477: \reference{}Feretti, L., Fusco-Femiano, R., Giovannini, G., \&
478: Govoni, F. 2001, A\&A, 373, 106
479:
480: \reference{}Frontera, F., Costa, E., Dal~Fiume, D., Feroci, M.,
481: Nicastro, L., Orlandini, M., Palazzi, E., \& Zavattini, G. 1997,
482: A\&AS., 122, 357
483:
484: \reference{}Fusco-Femiano, R., Dal Fiume, D., Feretti, L.,
485: Giovannini, G., Grandi, P., Matt, G., Molendi, S., \& Santangelo,
486: A. 1999, ApJ, 513, L21
487:
488: \reference{}Fusco-Femiano, R., Dal Fiume, D., De Grandi, S.,
489: Feretti, L., Giovannini, G., Grandi, P., Malizia, A., Matt, G., \&
490: Molendi, S. 2000, ApJ, 534, L7
491:
492: \reference{}Fusco-Femiano, R., Dal Fiume, D., Orlandini, M.,
493: Brunetti, G., Feretti, L., \& Giovannini, G. 2001, ApJ., 552, L97
494:
495: \reference{}Fusco-Femiano, R. \etal 2003, A\&A, 398, 441
496:
497: \reference{}Fusco-Femiano, R. \etal 2003, Proc. of "Matter and
498: Energy in Clusters of Galaxies", Taiwan, ASP Conf.Ser., eds.:
499: S.Bowyer and C.-Y. Hwang; p. 109
500:
501: \reference{}Fusco-Femiano, R., Orlandini, M., Brunetti, G.,
502: Feretti, L., Giovannini, G., Grandi, P., \& Setti, G. 2004, ApJ,
503: 602, L73 (FF04)
504:
505: \reference{}Fusco-Femiano, R., Landi, R., \& Orlandini, M. 2005,
506: ApJ, 624, L69
507:
508: \reference{}Guainazzi, M. \& Matteuzzi, A. 1997, SDC Technical
509: Report - TR-
510: 011\\ftp://www.asdc.asi.it/pub/sax/doc/reports/sdc-tr14.ps.gz
511:
512: \reference{}Landi, R. 2005, Ph.D. thesis, Bologna Univ. (paper in
513: preparation by Orlandini \etal)
514:
515: \reference{}Nevalainen, J., Oosterbroek, T., Bonamente, M., \&
516: Colafrancesco, S. 2004, ApJ, 608, 166 (NE04)
517:
518: \reference{}Perola, G.C., \& Reinhardt, M. 1972, A\&A, 17, 432
519:
520: \reference{}Petrosian, V. 2003, ASPC, 301, 337
521:
522: \reference{}Petrosian, V., Madejski, G., \& Luli, K. 2006,
523: accepted by ApJ (astro-ph/0608455)
524:
525: \reference{}Rephaeli, Y. 1979, ApJ, 227, 364
526:
527: \reference{}Rephaeli, Y., Gruber, D.E., \& Rothschild, R.E. 1987,
528: ApJ, 320, 139
529:
530: \reference{}Rephaeli, Y. \& Gruber, D.E. 1988, ApJ, 333, 133
531:
532: \reference{}Rephaeli, Y., Ulmer, M., \& Gruber, D.E. 1994, ApJ,
533: 429, 554
534:
535: \reference{}Rephaeli, Y., Gruber, D.E., \& Blanco, P. 1999, ApJ,
536: 511, L21
537:
538: \reference{}Rephaeli, Y., \& Gruber 2002, ApJ, 579, 587
539:
540: \reference{}Rephaeli, Y., \& Gruber 2003, ApJ, 595, 137
541:
542: \reference{}Rephaeli, Y., Gruber, D., \& Arieli, Y. 2006, accepted
543: by ApJ (astro-ph/0606097)
544:
545: \reference{}Rossetti, M., \& Molendi, S. 2004, A\&A, 414, L41
546: (RM04)
547:
548: \reference{}Willson, M.A.G. 1970, MNRAS, 151, 1
549:
550: \end{references}
551:
552: \clearpage
553:
554: %\textbf{Figure Captions}
555:
556: %\bigskip\par\noindent
557:
558:
559: %\begin{landscape}
560: %\begin{deluxetable}{ccccccc}
561: %\tablecolumns{7} \tablewidth{0pc} \tablecaption{Non-thermal HXR
562: %excess in the 20--80 keV energy
563: %range\tablenotemark{a}\label{tab1}}
564: %\tablehead{%
565: %\colhead{} Observation & \colhead{Epoch} & \colhead{PDS exposure}
566: %& \colhead{Observed rate} & \colhead{Predicted rate} &
567: %\colhead{Excess} &
568: % \colhead{Flux\tablenotemark{b}} \\
569: %\colhead{} & \colhead{} & \colhead{(ks)} &
570: %\colhead{(\cts)} & \colhead{(\cts)} &
571: %\colhead{(c.l.)} &
572: % \colhead{($10^{-11}\erg$)}
573: %}
574:
575: %\startdata
576: %OBS1/XAS\tablenotemark{c} & Dec 1997 & 44.5 & $0.3900\pm 0.0330$ & 0.2780 & 3.4$\sigma$ & $2.3\pm 1.0$ \\
577: %OBS2/XAS\tablenotemark{c} & Dec 2000 & 122.2 & $0.3330\pm 0.0170$ & 0.2750 & 3.4$\sigma$ & $1.3_{-0.6}^{+0.5}$ \\
578: %Combined/XAS\tablenotemark{c} & & 166.7 & $0.3490\pm 0.0150$ & 0.2760 & 4.8$\sigma$ & $1.5\pm 0.5$ \\
579: %OBS1/SAXDAS & Dec 1997 & & & & & \\
580: %OBS2/SAXDAS & Dec 2000 & & & & & \\
581: %Combined/SAXDAS & & & & & & \\
582: %\enddata
583:
584: %\tablecomments{Background evaluated considering only the -OFF
585: %field. Quoted errors for the reported count rates are at
586: %1$\sigma$.} \tablenotetext{a}{\ Excess with respect to a 8.11 keV
587: %(David \etal 1993) thermal bremsstrahlung component (see text for
588: %details).} \tablenotetext{b}{\ A photon index of 2 was used to
589: %derive the flux (see text for details).} \tablenotetext{c}{\ see
590: %FF04.}
591:
592: %\end{deluxetable}
593: %\end{landscape}
594:
595:
596: %\clearpage
597:
598: %\begin{landscape}
599: %\begin{deluxetable}{ccccccc}
600: %\tablecolumns{7} \tablewidth{0pc} \tablecaption{Non-thermal HXR
601: %excess in the 20--80 keV energy
602: %range\tablenotemark{a}\label{tab2}}
603: %\tablehead{%
604: %\colhead{} Observation & \colhead{Epoch} & \colhead{PDS exposure}
605: %& \colhead{Observed rate} & \colhead{Predicted rate} &
606: %\colhead{Excess} &
607: % \colhead{Flux\tablenotemark{b}} \\
608: %\colhead{} & \colhead{} & \colhead{(ks)} &
609: %\colhead{(\cts)} & \colhead{(\cts)} &
610: %\colhead{(c.l.)} &
611: % \colhead{($10^{-11}\erg$)}
612: %}
613:
614: %\startdata
615: %OBS1/XAS & Dec 1997 & 44.5 & 0.3656$\pm$ 0.0268 & 0.2731 & 3.4$\sigma$ & $x.x\pm x.x$ \\
616: %OBS2/XAS & Dec 2000 & 122.2 & 0.3053$\pm$ 0.0140 & 0.2725 & 2.3$\sigma$ & $x.x_{-x.x}^{+x.x}$ \\
617: %Combined/XAS & & 166.7 & 0.3214$\pm$ 0.0125 & 0.2727 & 3.9$\sigma$ & $x.x\pm x.x$ \\
618: %OBS1/SAXDAS & Dec 1997 & & & & & \\
619: %OBS2/SAXDAS & Dec 2000 & & & & & \\
620: %Combined/SAXDAS & & & & & & \\
621: %OBS1/SAXDAS/ASDC & Dec 1997 & 42.8 & 0.4019$\pm$0.0279 & 0.3086 & 3.3$\sigma$ & \\
622: %OBS2/SAXDAS/ASDC & Dec 2000 & 124.9 & 0.4077$\pm$0.0144 & 0.3742 & 2.3$\sigma$ & \\
623: %Combined/SAXDAS/ASDC & & 167.7 & 0.3593$\pm$0.0127 & 0.3112 & 3.8$\sigma$ & \\
624: %\enddata
625:
626: %\tablecomments{Background evaluated considering the average of the
627: %$\pm$OFF fields. Quoted errors for the reported count rates are at
628: %1$\sigma$.} \tablenotetext{a}{\ Excess with respect to a 8.11 keV
629: %(David \etal 1993) thermal bremsstrahlung component (see text for
630: %details).} \tablenotetext{b}{\ A photon index of 2 was used to
631: %derive the flux (see text for details).}
632:
633: %\end{deluxetable}
634: %\end{landscape}
635:
636: %\clearpage
637:
638: %\begin{landscape}
639: %\begin{deluxetable}{cccccc}
640: %\tablecolumns{6} \tablewidth{0pc} \tablecaption{Non-thermal HXR
641: %excess in the 25--80 keV energy
642: %range\tablenotemark{a}\label{tab3}}
643: %\tablehead{%
644: %\colhead{} Observation & \colhead{Epoch} & \colhead{PDS exposure}
645: %& \colhead{Observed rate} & \colhead{Predicted rate} &
646: %\colhead{Excess} \\
647: %\colhead{} & \colhead{} & \colhead{(ks)} &
648: %\colhead{(\cts)} & \colhead{(\cts)} &
649: %\colhead{(c.l.)} }
650:
651: %\startdata
652:
653: %OBS1/XAS & Dec 1997 & 44.5 & 0.2510$\pm$0.0259 & 0.1697 &
654: %3.14$\sigma$ \\
655: %OBS2/XAS & Dec 2000 & 122.2 & 0.1983$\pm$0.0135
656: %& 0.1693 & 2.15$\sigma$ \\
657: %Combined/XAS & & 166.7 &
658: %0.2124$\pm$0.0121 & 0.1694 & 3.55$\sigma$ \\
659: %OBS1/SAXDAS & Dec 1997 & & & & \\
660: %OBS2/SAXDAS & Dec 2000 & & & & \\
661: %Combined/SAXDAS & & & & & \\
662: %OBS1/SAXDAS/ASDC & Dec 1997 & 42.8 & 0.2235$\pm$0.0264 & 0.1440 & 3.01$\sigma$ \\
663: %OBS2/SAXDAS/ASDC & Dec 2000 & 124.9 & 0.2426$\pm$0.0138 & 0.2136 & 2.10$\sigma$ \\
664: %Combined/SAXDAS/ASDC & & 167.7 & 0.1877$\pm$0.0120 & 0.1474 & 3.36$\sigma$ \\
665: %OBS1/SAXDAS/RM04 & Dec 1997 & -- & -- & -- & 2.84$\sigma$ \\
666: %OBS2/SAXDAS/RM04 & Dec 2000 & -- & -- & -- & 1.11$\sigma$ \\
667: %Combined/SAXDAS/RM04 & & -- & -- & -- & -- \\
668: %\enddata
669:
670: %\tablecomments{Background evaluated considering the average of the
671: %$\pm$OFF fields. Quoted errors for the reported count rates are at
672: %1$\sigma$.} \tablenotetext{a}{\ Excess with respect to a 8.21 keV
673: %thermal bremsstrahlung component (see text for details).}
674:
675: %\end{deluxetable}
676: %\end{landscape}
677:
678:
679: %\clearpage
680:
681: %\begin{figure}
682: %\rotatebox{-90}{ \epsscale{0.7} \plotone{a2256_mecs_pds_new.ps}}
683: %\vspace{2cm} \caption{Abell 2256
684: %--- MECS and PDS data. The continuous line represents a
685: %thermal component at the average cluster gas temperature of
686: %7.65$\pm$0.17 keV in the central circular region of 8$'$ in
687: %radius. The errors bars are quoted at the 1$\sigma$
688: %level.\label{fig1}}
689: %\end{figure}
690:
691: \end{document}
692: