1: \documentclass[12pt,preprint]{aastex}
2:
3: %%\usepackage{graphicx}
4:
5: \newcommand{\vdag}{(v)^\dagger}
6: \newcommand{\myemail}{Angelique.Joinet@cesr.fr}
7:
8:
9:
10: %\slugcomment{Not to appear in Nonlearned J., 45.}
11:
12:
13:
14: \shorttitle{Hard X-ray emission of the microquasar GRO~J1655$-$40.}
15: \shortauthors{Joinet et al.}
16:
17:
18: \begin{document}
19:
20: \title{Hard X-ray emission of the microquasar GRO~J1655$-$40 during the rise of its 2005 outburst.}
21: \author{A. Joinet\altaffilmark{1}, E. Kalemci\altaffilmark{2}, F. Senziani\altaffilmark{3,4,5}}
22:
23: \altaffiltext{1}{CESR, 9 Avenue du Colonel Roche, BP4346, 31028 Toulouse, France}
24: \altaffiltext{2}{Sabanc\i\ University, Orhanl\i\ - Tuzla, \.Istanbul, 34956, Turkey}
25: \altaffiltext{3}{INAF-IASF Milano, via Bassini 15, 20133 Milano, Italy}
26: \altaffiltext{4}{Universit\'e Paul Sabatier, 31062 Toulouse, France}
27: \altaffiltext{5}{Universit\'a degli Studi di Pavia, Dipartimento di Fisica Nucleare e
28: Teorica, via Bassi 6, 27100 Pavia, Italy}
29: %
30:
31: \begin{abstract}
32:
33: We present the analysis of the high energy emission of the Galactic black
34: hole GRO~J1655$-$40 at the beginning of its 2005 outburst.
35: The data from 458 ks of \textit{INTEGRAL} observations,
36: spread over 4 weeks, are analyzed, along with the existing simultaneous
37: \textit{RXTE} and \textit{Swift} data. The high energy data allow
38: us to detect the presence of a high energy cut-off and to study its
39: evolution during the outburst rise. This high energy feature is
40: generally related to thermal mechanisms in the framework of
41: Comptonization models from which we can estimate the plasma
42: parameters. We found an electron temperature of about 30-40 keV and
43: an optical depth around 1.8-2.1. The high energy cut-off decreased
44: along with the radio flux, and disappeared as the jet turned off.
45:
46: \end{abstract}
47: %
48: \keywords{stars: individual: GRO~J1655$-$40 - gamma rays : observations - black hole physics - accretion,
49: accretion disk - X-rays : Binaries}
50: %
51: %
52:
53:
54: \section{Introduction}
55: GRO~J1655$-$40 is a transient Galactic X-ray binary. Since its discovery on
56: July 27, 1994 with BATSE (Zhang et al. 1994) on-board the
57: Compton Gamma-ray Observatory (\textit{CGRO}), the source
58: underwent several outbursts (i.e. : in 1995 see Zhang et al. 1997, in 1996/1997
59: see Kuulkers et al. 2000 and M\'endez et al. 1998).
60: It is likely to be a low mass X-ray binary
61: and the compact object in this system is probably a black hole
62: (Bailyn et al. 1995) with a mass estimated at 6.3 M$_ \Sun$ (Greene
63: et al. 2001) and
64: a distance of 3.2 kpc (Hjellming \& Rupen, 1995) from which we based the
65: calculation of the Eddington Luminosity L$_{edd}$ of the source. A
66: recent work by Foellmi et al. (2006) places an upper limit of 1.7 kpc
67: to the distance, but this upper limit is still under debate.
68: Radio jets were discovered in the mid-1990s and revealed
69: apparent superluminal motion in opposite directions
70: (Tingay et al. 1995).
71:
72: A transient Galactic black hole usually exhibits
73: a complex spectro-temporal variability with the variation of the accretion
74: rate. The changes in their properties allow to characterise the state
75: of the source (see McClintock \& Remillard 2006, for a complete description).
76: In the High Soft State (hereafter HSS), the soft X-ray emission
77: dominates the spectrum in the form of a blackbody component which is due to the thermal
78: component from a standard accretion disk.
79: In the Low Hard State (hereafter LHS), the source is
80: characterised by a relatively low flux in the soft X-rays
81: ($\la$ 1 keV) and a high flux in the hard X-rays ($\sim$100 keV).
82: This is usually interpreted
83: (e.g. Shapiro et al. 1976, Narayan \& Yi 1995) as the Comptonization
84: of the soft X-ray photons emitted by a cold, geometrically thin disk
85: (Shakura \& Sunyaev 1973) by a
86: hot plasma surrounding this accretion disk. The hard component is generally
87: described by a power-law with a photon index of $\Gamma$=1.4-1.8, and an exponential
88: cutoff $E_c$ around 100-200 keV. In addition, a Fe K$\alpha$ line at $\sim$ 6.4 keV
89: and a Compton reflection bump peaking at $\sim$ 30 keV may exist.
90: These are signatures of the irradiation of the cold optically thick
91: disk by the hard X-rays from the corona (George \& Fabian 1991).
92: The LHS is characterized by radio emission which has been shown to be
93: consistent with a midly relativistic
94: (v $\simeq$ 0.6c) jet (Gallo et al. 2003).
95: Before entering into the HSS, the source may go through transitional
96: states also called intermediate states. Homan \& Belloni (2005)
97: divides the intermediate state into the hard intermediate state
98: (HIMS) and the soft intermediate state (SIMS) which depend on the
99: power density spectra and the spectral index. The jet starts to be
100: quenched in the intermediate state (e.g. Corbel et al. 2004).
101:
102: We present here an analysis of the spectral evolution of GRO~J1655$-$40
103: during the rising phase of the 2005 outburst. The analysis is based on
104: the data taken by the International Gamma-ray Laboratory
105: ({\it INTEGRAL}; Winkler et al. 2003). GRO~J1655$-$40 has been observed by
106: {\it INTEGRAL} starting from MJD 53425 (\textit{INTEGRAL} revolution 289
107: on 2005 February 24) to MJD 53448 (\textit{INTEGRAL} revolution 296 on
108: 2005 March 19). The spectral and timing evolution of the source from the
109: Proportional Counting Array (\textit PCA) on board the Rossi X-ray
110: Timing Explorer (\textit{RXTE}) in the 3-30 keV energy range
111: has been studied in detail in Shaposhnikov et al. 2007
112: from MJD 53419 up to MJD 53445. Brocksopp et al. 2006 presented
113: the data from the Burst Alert Telescope (BAT) on board the
114: \textit{Swift} observatory in the 14-150 keV energy range
115: from MJD 53430 up to MJD 53435. Caballero-Garcia et al. (2007) presented the
116: results of the analysis of 4 ToO observations from 27 February to 11
117: April of 2005 using JEM-X, ISGRI, and SPI onboard the \textit{INTEGRAL} observatory
118: and interpreted it in the framework of several physical models. They come
119: to the conclusion that no cut-off is necessary to describe the data in
120: the LHS.
121:
122: We studied the broadband spectral evolution using the
123: Spectrometer on board the \textit{INTEGRAL} (SPI, Vedrenne et al. 2003)
124: in the 23-600 keV energy range combined with all the publicly
125: available data from the PCA and HEXTE detectors on board the \textit{RXTE} observatory,
126: as well as the BAT detector in order to cover the 3-200 keV energy range.
127: One of the main interest in studying such a source during the LHS to the
128: HSS transition is to follow the evolution of the high energy cutoff.
129: We propose to quantify its value as well as its significance during the
130: rising phase of the outburst.
131:
132: We briefly describe the data analysis methods for
133: \textit{INTEGRAL} (SPI and IBIS/ISGRI), \textit{RXTE} (PCA and HEXTE), and \textit{SWIFT} (BAT)
134: in Section 2. The scientific results from the spectral modelling
135: are presented in Section 3.
136:
137: \section{Observations and data reduction}
138:
139: GRO~J1655$-$40 started a new outburst on February 17, 2005 (MJD 53418)
140: observed by Markwardt \& Swank (2005) with the
141: PCA detector onboard the \textit{RXTE} observatory. This reactivation was
142: confirmed by Torres et al. (2005) who reported the near infrared activity of the
143: source on Februay 21, 2005. \textit{INTEGRAL} performed follow-up observations
144: on GRO~J1655$-$40 starting at MJD 53425. Table~\ref{tab:tab1} gives the
145: details of each \textit{INTEGRAL} revolution and corresponding
146: quasi simultaneous \textit{RXTE} observations used in this analysis.
147:
148: \subsection{\textit{INTEGRAL} data reduction}
149:
150: The data from the SPI detector were reduced as explained in Joinet
151: et al. (2005) except that only 17 detectors were active
152: compared to 19 in the beginning of the mission. The SPIROS
153: V6 algorithm has been used in order to derive the position of
154: sources detected in the field of view of GRO~J1655-40. We refer to
155: Joinet et al. 2005 (section 2.1.1) for the method of the spectral
156: reconstruction. Only pointings for which GRO~J1655$-$40 was at a
157: distance less than 12 degrees from the central axis were taken into
158: account for the analysis. We excluded pointings affected by a solar
159: flare or by exit/entry into the radiation belts. We obtained 458 ks
160: of useful data during the observation period covered by the
161: revolutions 289 up to 296 (see Table~\ref{tab:tab1}).
162: Revolutions 295 \& 296 have been merged and will be named
163: 295+296. The variability time scale of each source was estimated (see
164: section 3.1) on the basis of both their intensity and their known
165: temporal behavior. The background flux is stable within each of the
166: considered revolution except for revolutions 292 and 293 for which a
167: variability time scale of one pointing (with a duration of 30-40
168: minutes) was used. We consider only one normalization parameter per
169: orbit with an uniformity map determined for the 17 detectors
170: configuration.
171:
172: We limited the energy range from 23 to 600 keV for the SPI data and added a
173: 3 \% systematic error to all spectral channels.
174: The IBIS/ISGRI data corresponding to revolutions 290, 295 \& 296
175: which were published by Cabarello-Garcia et al. (2007), have also been included
176: in the analysis. We used the high-level results from them who used the Standard
177: Off-Line Science Analysis (OSA) 5.1 software package.
178:
179: \subsection{\textit{RXTE} data reduction}
180:
181: We also analyzed the public data from the PCA and HEXTE detectors on
182: board the \textit{RXTE} observatory (Bradt et al. 1993, Rothschild
183: et al. 1998) data. Table~\ref{tab:tab1} summarizes the set of
184: \textit{RXTE} observations performed contemporaneously with the
185: \textit{INTEGRAL} data observation periods except for the revolution
186: 294 for which there are no \textit{INTEGRAL} data available. This
187: observation period has been divided into three. Several \textit{RXTE}
188: observations have been merged in order to correspond to one
189: integrated \textit{INTEGRAL} revolution.
190:
191: For both instruments the data reduction were performed using the FTOOLS
192: routines in the HEAsoft software package distributed by NASA's HEASARC
193: (version 6.0.4).
194:
195: All available PCUs of the PCA (Bradt et al. 1993) detector have been
196: used for the data extraction. We added 0.8\% up to 7 keV and 0.4\%
197: above 7 keV as systematic error (Tomsick et al. 2001)
198:
199: For HEXTE, we used the response matrix created by the FTOOLS, and applied
200: the necessary dead time correction (Rothschild et al. 1998). The HEXTE
201: background is measured throughout the observation by alternating between the
202: source and background fields every 32 s. The data from the background regions
203: are merged.
204:
205: We limited the energy range from 3 to 25 keV, and 16 to 227 keV for the PCA
206: and HEXTE data, respectively. \textit{HEXTE} channels were grouped by 2
207: for channels 16-31, by 4 for channels 32-59, by 10 for channels 60-99
208: and by 64 for channels 100-227.
209:
210: \subsection{\textit{Swift} data reduction}
211:
212: All the publicly available data from the BAT detector on board the
213: \textit{Swift} observatory (Gehrels et al. 2004 ,
214: Barthelmy et al. 2005) covering the rising phase of the 2005
215: outburst were also analysed (see Table~\ref{tab:tab2}). Some of them were
216: simultaneous with the {\em INTEGRAL} revolutions (292, 295+296).
217: The revolution 292 has been divided into two datasets composed of 7
218: BAT pointings with a duration of 10.09 ks for 292$_{sw}$-A, and
219: 7.66 ks for 292$_{sw}$-B. We also used the \textit{Swift} data
220: covering the observation period between the \textit{INTEGRAL} revolutions 292
221: and 293 ($293_{sw}$) and between revolutions 293 and 295-B
222: ($294_{sw}-A$ and $294_{sw}-B$). We reduced the data using the
223: standard \textit{Swift} software\footnote{{\em
224: http://swift.gsfc.nasa.gov/docs/software/lheasoft}} (version 2.4). A
225: standard filtering was applied in order to discard the data affected
226: by high background rate and source occultations. For each interval
227: over which the BAT pointings were unchanged, we extract a background
228: subtracted spectra together with the response matrix using the
229: mask-weighting technique. Systematic errors were applied to the
230: spectra using {\em batphasyserr} BAT FTOOL. For the spectral
231: analysis, we limited the energy range from 16 to 150 keV.
232:
233: For both the \textit{RXTE} instruments and the BAT, the normalization factor
234: was set free with respect to SPI normalization for all fits.
235:
236: \section{Results}
237: \subsection{Light curve}
238:
239: The flux extraction of GRO~J1655$-$40 from SPI observations was
240: performed taking into account the hard X-ray sources detected in the
241: field of view of the source. We used a timescale of one pointing
242: (whose duration is about 3600 s) for 4U1700-377 and OAO~1657-415 and
243: two pointings for GX~340+0 and 4U~1705-322. We also considered other
244: sources (GRO~J1655$-$40, 1E~1740-2942, 4U~1630-47, GX337+00 and
245: GX349+2) with a constant flux within each revolution.
246: As the significance of sources decreases above 150 keV, we
247: considered only two sources with a timescale resolution of 3600 s
248: for 4U1700-77, and one revolution for GRO~J1655$-$40, to extract
249: fluxes in the 150-600 keV energy range. The light curve of the
250: source from revolution 289 up to revolution 296 in the 23-51 keV
251: energy range is shown in Figure~\ref{fig:fig1}. As the source was in
252: the border of the SPI field of view between \textit{INTEGRAL} revolutions 293
253: and 295-B, we used the \emph{RXTE} and the
254: BAT observations ($294_{sw}$-A, $294_{sw}$-B) to cover this period.
255: Table~\ref{tab:tab4} gives the SPI and BAT fluxes
256: in different energy bands.
257:
258: We also extracted the light curve obtained by the \textit{RXTE}
259: all-sky monitor ASM in the 1.5-12 keV energy range in the same
260: period (data taken from the public RXTE
261: database\footnote{http://xte.mit.edu/lcextrct/asmsel.html}). The
262: different states harbored by the source were determined by
263: Shaposhnikov et al. (2007) on the basis of the X-ray properties from
264: the PCA observations and are summarized in Figure~\ref{fig:fig1}.
265: The source was in the LHS from revolution 289 up to revolution 292,
266: it then entered the HIMS from revolution 293 (also observed by the
267: BAT detector : datasets $293_{sw}$ and $294_{sw}$-A), the SIMS
268: during the revolution 294-B and was in the HSS from revolution
269: 295-A. Revolution 294 has been divided into two \textit{RXTE} datasets
270: (294-B and 294-C) in order to follow the evolution of the
271: high energy cutoff during the SIMS.
272:
273: As seen from the SPI light curve (Figure~\ref{fig:fig1}), the 23-51
274: keV flux increased by a factor of 2.9 between MJD 53425 ($\Phi$=42 $\pm$ 2
275: mCrab) and MJD 53433 ($\Phi$=121 $\pm$ 7 mCrab) while the source was in the
276: LHS. A radio ejection was observed between MJD 53429-53433
277: with a radio peak at 5 GHz (Shaposhnikov et al. 2007).
278: After this ejection, the X and $\gamma$-ray flux increased
279: exponentially up to \textit{INTEGRAL} revolution 293 or MJD 53437-53438
280: ($\Phi$=280 $\pm$ 7 mCrab) during which the source was in the HIMS.
281: As the flux in the 3-25 keV energy range measured from PCA data
282: increased by a factor 3.9, the flux in the 23-51 keV energy range
283: decreased by a factor 3.5 between revolution 293 and revolution (295+296)$_{sw}$
284: indicating the LHS to the HSS transition.
285:
286:
287: \subsection{Spectral Modeling of the X and $\gamma$-ray data}
288:
289: The spectra corresponding to each set of data in Table~\ref{tab:tab1} and
290: \ref{tab:tab2} have been fitted with various
291: models available in the standard XSPEC 11.3.1 fitting package (Arnaud 1996).
292: In all fits, the iron emission line was modelled by a narrow Gaussian line
293: fixed at an energy of 6.4 keV with a free width.
294: For all models, the inner disk inclination was fixed at $70^\circ$
295: (Van de Hooft et al. 1998). We also consider the emission from a multicolor
296: disk blackbody (DISKBB in XSPEC, Mitsuda et al. 1984). We account for the
297: interstellar absorption (PHABS in XSPEC) using a hydrogen
298: column density $N_H = 7 \times 10^{21} \rm \, cm^{-2}$ for most of the observations (see
299: \S~3.2.1 for exceptions) which is based on the values constrained
300: by XMM-Newton observation during the 2005 outburst (Diaz Trigo et
301: al. 2007).
302:
303: \subsubsection{Power law with a cutoff}
304:
305: First, a power-law component is added to the base model described above.
306: The best fit parameters are presented in Table~\ref{tab:tab3}. During the LHS
307: (revolutions 289, 290, 291, 292), we found a constant spectral index
308: of $\Gamma$ $\simeq$ 1.47 (with a reduced $\chi^2$ ranging from 1.7 up to 2.0).
309:
310: During revolutions 289, 290 and 291, the fits are
311: significantly improved, as seen from the reduced $\chi ^2$ values
312: (with an F-test probability less than $10^{-9}$),
313: by adding a high energy cutoff component (see Table~\ref{tab:tab3}).
314: A constant spectral index $\Gamma$ of about 1.33-1.36 has been
315: obtained, which is almost the same value found by
316: Shaposhnikov et al. (2007) ($1.35 \pm 0.03$). The high energy cutoff
317: of $E_c = 231^{+94}_{-50}$ keV is consistent with the range value of
318: 163-214 keV presented in Shaposhnikov et al. (2007) for revolution
319: 289. The SPI data that we analyzed for all the outburst allow to
320: constrain its value and to precisely describe its evolution which
321: has not been done in Shaposhnikov et al. (2007). However we
322: reach a conclusion contrary to Caballero-Garcia et al. (2007) which
323: show that no cutoff was required to describe the data corresponding
324: to revolution 290. We fit the data set corresponding to
325: revolution 290 using
326: PCA, ISGRI and SPI data with the following
327: model : PHABS(DISKBB+POWERLAW)HIGHECUT in order to compare
328: our results with the ones derived by Caballero-Garcia et al. (2007).
329: We found a high
330: energy cut-off powerlaw constrained to a value
331: of 22$^{+6}_{-8}$ keV and a folding energy of
332: $E_{fold}$=193$^{+14}_{-21}$ keV with a reduced $\chi ^2$ of 0.83 (87 degrees of freedom or
333: dof).
334: %By fitting ISGRI and PCA then SPI and PCA, we found a folding energy
335: %of 193$^{+27}_{-22}$ keV (with a reduced $\chi ^2$ of 0.79 (71 dof)) and
336: %201$^{+91}_{-92}$ keV (with a reduced $\chi ^2$ of 0.65 (56 dof)), respectively.
337: By fitting ISGRI, HEXTE and PCA data, we found a folding energy
338: of $E_{fold}$=253$^{+34}_{-31}$ keV and a photon index of
339: $\Gamma$ =1.40$^{+0.02}_{-0.02}$ (with a reduced $\chi ^2$ of 1.43 (90 dof)).
340:
341: For revolution 292, the high energy cutoff is constrained to a value of
342: $E_c$=$187^{+22}_{-20}$ keV. BAT data were combined with the
343: simultaneous PCA, HEXTE and SPI data corresponding to the
344: \textit{INTEGRAL} revolution 292 in order to check the cross
345: calibration between all instruments.
346: %The
347: %addition of BAT parameters decreases the reduced $\chi ^2$ value
348: %from 0.85 with 76 degrees of freedom (dof) (with no BAT data) down
349: %to 0.68 (223 dof) (with BAT data, see Table~\ref{tab:tab5}). The
350: %normalisation factors of BAT with respect to
351: % SPI are 0.90 and 1.07 for dataset 292$_{sw}$-A and 292$_{sw}$-B, respectively.
352: During revolutions 289-292, the disc component was very weak.
353: We then notice a decrease of the high energy cutoff down to $E_c$=$87^{+4}_{-5}$ keV during
354: revolution 294a as the source changes into the HIMS (Shaposhnikov et al.
355: 2007). The disc component also got stronger, while the power law
356: index remained constant.
357:
358: By fitting simultaneously BAT, PCA and HEXTE data for revolution 294
359: (datasets 294-B and 294$_{sw}$-B), when the source has entered the
360: SIMS, a pure powerlaw results in a reduced $\chi ^2$ of
361: 1.16 (118 dof). Adding a high-energy cutoff results in significant
362: reduction in reduced $\chi ^2$ (0.99, 117 dof) but the folding
363: energy is not well constrained ($E_{fold}$=$302^{+225}_{-81}$ keV), and strongly depends on
364: the power-law index. The disk component also starts to dominate the energy
365: spectrum as shown from the clear changes in the $\phi _p / \phi _b$ ratio
366: (Table~\ref{tab:tab3}) (where $\phi _p $ is the powerlaw flux and
367: $\phi _b$ the blackbody flux in the 2-20 keV energy range). This
368: ratio decreases by a factor 13 between revolution 293 and 294-B
369: due to a large increase of the disk component $\phi
370: _b$ by a factor of 21. The contribution of the hard component with respect to the disk
371: component is lower than 50\% during revolution 294-B. Moreover, the
372: photon index becomes steeper with $\Gamma \simeq 2.1$, indicating a
373: spectral transition.
374:
375:
376: %The decrease of the hydrogen column density $N_H$
377: %from $0.7 \times 10^{22}$ cm$^{-2}$ down to $0.5 \times 10^{22}$ cm$^{-2}$
378: %allows to decrease the $\chi ^2$ from values above 2 down to values lower than
379: %1.8 from the dataset 294-B.}
380: %THE LAST SENTENCE IS IRRELEVANT AND DISRUPTING THE FLOW.
381:
382: From the dataset 294-C for which only PCA and HEXTE data are
383: available, there is no evidence of a high-energy cutoff up to 286
384: keV.
385: The reduced $\chi ^2$ decreases from 2.00(61 dof) down to
386: 1.81(61 dof) and from 1.66(62 dof) down to 1.40(62 dof) for the datasets 294-C
387: and 295-A, respectively, using $N_H = 5 \times 10^{21} \rm \, cm^{-2}$. The high reduced $\chi ^2$ is due to poor modeling of
388: the low energy part of the spectra. A complex spectral feature
389: around 7 keV has been fitted by Diaz Trigo et al.
390: (2007), using XMM-Newton and \textit{INTEGRAL} data.
391: The spectra corresponding to these fit parameters (Table~\ref{tab:tab3})
392: are shown in Figure~\ref{fig:fig4} and~\ref{fig:fig5}.
393:
394: %During the rising phase of the outburst, the equivalent width $W_{Fe}$
395: %increased from 75 eV (revolution 289) up to $\simeq$ 811 eV
396: %(revolution 294-A+294$_{sw}$-B) and decreased below 400 eV when the
397: % source entered the HSS (revolution 295-A). The value of the photon
398: %index $\Gamma$ is correlated with the normalisation of the disk blackbody
399: %component (Shaponiskov et al. 2007). The thermal disc
400: %component appeared in revolution 293, as the source was into the HIMS.
401: %The power law photon index did not change significantly compared to the
402: %values obtained during the LHS, but there is a decrease of the high energy
403: %cutoff. After the source entered SIMS (294$_{sw}$-B), the
404: %thermal disc component clearly dominated the spectra.
405:
406:
407:
408: \subsubsection{Reflection model PEXRAV}
409:
410: We fitted all data with a reflection model, PEXRAV in XSPEC
411: (Magdziarz \& Zdziarski 1995), consisting of a power-law with a high
412: energy cut-off and reflection from neutral medium (see Table~\ref{tab:tab5}).
413: During the LHS, as the luminosity in the 3-600 keV
414: energy range increased by a factor of $\simeq$ 2.8 from revolution 289 up to
415: revolution 292, the spectral index is constant ($\Gamma$ $\simeq$ 1.3-1.4).
416: Similar values of the spectral index and of the high
417: energy cutoff are found using the CUTOFFPL and the PEXRAV models:
418: it is explained by the fact that the reflection fraction is so low
419: that an additional reflection component is not required in the fit.
420: The range of values for the high energy cutoff (180-380 keV)
421: are consistent with the results of Shaposhnikov et al. (2007) ($196
422: \pm 48$ keV) for the revolution 289. We determine an upper limit on
423: the energy cutoff when the source is into the HSS : $E_c >2000$ keV
424: for the datasets (295+296)$_{all}$ (see Table 5).
425:
426: During all observations (from the LHS to the HSS),
427: the reflection component
428: is not constrained and does not exceed $\Omega = 0.2$, which is
429: lower than the value of $\Omega \simeq 0.5$ found by Shaposhnikov et al. (2007).
430:
431: %\subsubsection{Evolution of the soft and hard components}
432:
433: %The source followed a continuous spectral softening observed from
434: % the hardness of ASM/RXTE data in the 1.5-12 keV (see Figure~\ref{fig:fig1}).
435: %From the revolution 294-B (SIMS), there is a clear evolution of the hardness
436: %combined with an increase of the powerlaw slope $\Gamma$ obtained from
437: %CUTOFFPL model (Table~\ref{tab:tab3}) as well as from PEXRAV model
438: %(Table~\ref{tab:tab5}). With both models, during the LHS and IS, the disk
439: %blackbody component is negligible. The disk component starts to dominate
440: %the energy spectrum as shown from the clear changes in the
441: %$\Phi _p / \Phi _b$ ratio (Table~\ref{tab:tab3})
442: %(where $\phi _p $ is the powerlaw flux in the 2-20 keV energy range
443: %and $\phi _b$ the blackbody flux). This ratio decreases by a factor
444: %12.93 between revolution 293 and 294-B due to a large increase (by a
445: %factor 21) of the disk component $\phi _b$. The contribution of the hard
446: %component with respect to the disk component is lower than 50\% during
447: %revolution 294-B, indicating a transition.
448: % During this dataset, the high energy cutoff is determined as
449: %$E_c > 286$ keV. Moreover, the photon index becomes steeper
450: %($\Gamma \simeq 2.10$). It is accompanied by a sharp rise of the X-ray
451: %luminosity in the 2-200 keV energy range which increases by a factor 2.1
452: %between revolution 293 (L = 0.017 L$_{edd}$) and revolution 294-B
453: %(L = 0.04 L$_{edd}$) which are similar to the values obtained in Shaposhnikov
454: %et al. 2007. It then reaches a luminosity of 0.05 L$_{edd}$ during revolution
455: % 295-A, which correspond to the first maximum of the X-ray luminosity during
456: %the 2005 outburst. The Eddington luminosity L$_{edd}$
457: %has been determined assuming a distance of 3.5 kpc and a mass of 6.3 M$\sun$.
458: %If the upper limit value of 1.7 kpc is assumed (Foellmi et al. 2006),
459: %the luminosity would be 0.005 and 0.01 L$_{edd}$ for the dataset 293 and
460: %294-A respectively. We can also derived the value of the disk inner radius
461: %from the normalization of the blackbody component determined from
462: %the spectral modelling presented in Table 4 and 5. With a gravitational
463: %radius $R_g=GM/c^2$= 10 km, the inner disc radius would be
464: %1.01 $R_g$ with a distance of 1.7 kpc during the observation corresponding
465: %to the INTEGRAL revolution 296.
466: %THIS IS VERY DANGEROUS, IS IT SPINNING MAXIMALLY?
467:
468: \subsubsection{Comptonization model COMPTT}
469:
470: The hard power law plus cut-off model of the LHS is usually
471: interpreted as thermal Comptonisation in a hot ($kT_e \sim 100 \rm \, keV$)
472: optically thin plasma (the corona). We used the COMPTT model (Titarchuk
473: 1994) in order to describe the high
474: energy spectrum. The inner temperature of the disk (T$_{in}$) of
475: the multicolor
476: disk blackbody model is forced to be equal to the soft photon temperature
477: (T${_0}$) of the Comptonization model. We see in Table~\ref{tab:tab6} that the
478: optical depth is $\tau$ $\simeq 1.8-2.0$ while the temperature decreases
479: from $kT_e$ $\simeq 40$ down to $kT_e$ $\simeq 30$ keV from the LHS to the HIMS.
480: By freezing the temperature to $kT_e=$ 37 keV which is the value obtained by
481: Shaposhnikov et al. (2007) for the dataset corresponding to the
482: \textit{INTEGRAL} revolution 289, we determined an optical depth of
483: $\tau=1.98^{+0.05}_{-0.05}$ which is lower by a factor of 2.2 compared to
484: the value found in this reference. Moreover the inner disk T$_{in}$
485: temperature is higher : T$_{in} =$ 0.83 keV instead of T$_{in} =$ 0.60 keV.
486: This could be explained by the fact that the continuum above 100 keV
487: constrained by the \textit{INTEGRAL} data is different in this
488: reference.
489:
490: Regardless of the fit models, we can describe the spectral evolution
491: in terms of an evolution of the geometry based on the Compton
492: parameter $y = (4 kT / m_e c^2) {\rm max} (\tau, \tau^2)$ (where the electronic
493: temperature kT and the optical depth $\tau$ are derived from a
494: Comptonization model, Sunyaev \& Titarchuk, 1980). During
495: the LHS, the Compton parameter is not varying significantly ($y \simeq$ 0.6). It starts to decrease from the
496: HIMS (revolution 293), and is correlated with the increase of the
497: normalization of the disk component. It decreases (see value in
498: Table~\ref{tab:tab6}) by a factor of 4.3 between revolution 289 and
499: revolution 294-B when the source changes into the SIMS. The LHS to the
500: Intermediate state (both HIMS and SIMS) transition
501: corresponds to a gradual decrease of the inner radius of the cold
502: accretion disk, associated with either the cold disk penetrating
503: the hot inner flow, or the latter collapsing into an optically thick
504: accretion disk with small active regions of hot plasma on top of it
505: (Zdziarski et al. 2002). The enhanced soft photon flux from the disk
506: tends to cool down the hot phase, leading to softer spectra. This can be
507: directly observed from the evolution of the blackbody component with
508: respect to the powerlaw component
509: (see the ratio $\phi _{p} / \phi _{b}$ given in Table~\ref{tab:tab3}).
510:
511: %\subsubsection{High energy cutoff}
512: \section{Discussion}
513:
514: \subsection{High energy cutoff and jet quenching}
515:
516: We determined that all datasets in the low hard state require a high
517: energy cut-off. In addition to F-tests that indicate that the cut-offs
518: are required, we have produced contour plots of cut-off versus
519: power-law index for each dataset and those also
520: substantiated the existence of the cut-off in single revolutions. In
521: Figure 5, the contour plot corresponding to the revolution
522: 290 clearly shows the requirement of the high energy cutoff during this
523: observation period which has also been studied by Caballero-Garcia et al. (2007)
524: who claimed that no cut-off is required for this dataset.
525: The main difference between our analysis and theirs is the fact that
526: we have used the PCA to constrain the low energy part of the spectrum
527: whereas in Cabarello Garcia et al. (2007) JEM-X has been used.
528:
529: For PEXRAV and CUTOFF models, the data from datasets 294-B and 294-C
530: indicate that either the high energy cutoff
531: increases significantly, or the cutoff vanishes completely.
532: Such evolution was observed for GX339-4
533: in the HIMS (Belloni et al. 2005), and now appeared in the SIMS
534: in the case of GRO~J1655$-$40.
535:
536: We attempt to describe this high energy feature by adding a powerlaw
537: component to the Comptonization model COMPTT. We found that the
538: addition of a power-law component leads to an improvement of the fit
539: significant at the $99.9998 \%$ level for both datasets according to
540: a F-test, with a best fit photon index of
541: $\Gamma$=$3.86^{+0.12}_{-0.18}$ and $\Gamma$=$3.95^{+0.40}_{-0.35}$ for the data sets 294-B and 294-C, respectively.
542: The powerlaw component contributes at $23\%$ and $20\%$ of
543: the 2-20 keV flux for the revolution 294-B and 294-C, respectively.
544: The extension of a powerlaw at high energy has already been observed in
545: the Steep Power Law State of GRO~J1655$-$40 during its 1996 outburst
546: (see Remillard \& McClintock 2006). It is interpreted as inverse
547: Compton scattering that occurs in a non-thermal corona, which may be
548: a simple slab operating on seed photons from the underlying disk (
549: Zdziarski et al. 2005). The origin of the Comptonizing electrons is
550: still a subject of debate. Poutanen \& Fabian (1999)
551: suggested that flare regions erupting from magnetic
552: instabilities in the accretion disk could explain such a
553: Comptonizing media. The association of a non-thermal process with a
554: state transition has also been discussed for Cyg~X-1 (Malzac et al.
555: 2006), in this case interpreting the power-law tail as the
556: Comptonization of soft photons by accelerated electrons (see also
557: Titarchuk 1997). Finally, such a high energy emission could also
558: come from an optically thick electron-positron outflow covering the
559: whole inner region of the accretion disk (Beloborodov 1999b).
560:
561: %The nonthermal spectrum which we modelled with the addition of a
562: %powerlaw component increases in flux as well as the equivalent width
563: %(or the flux) of the iron line. This correlation is also mentioned
564: %in Shaposhnikov et al. 2007. In the HSS, this non thermal component
565: %is negligible and contributes to $<0.2\%$.
566:
567:
568: As seen from the Figure~\ref{fig:fig7}, there is a correlation
569: between the radio intensity and the exponential energy cutoff. The
570: maximum flux in radio is achieved when the source is in the LHS.
571: When the source enters the HIMS, the high energy cut-off
572: decreases (or the plasma cools down) as well as the radio flux.
573: Quenching of the jet in the soft states of black hole binaries is
574: known for many sources (Fender et al. 1999, Corbel et al. 2000).
575: Here we show a clear evolutionary path to this quenching in terms of
576: plasma temperature, perhaps associated with the cooling and
577: shrinking of the corona.
578:
579: \subsection{The jet ejection event}
580:
581: The jet ejections usually occur at or near state
582: transitions prior to the source entering the Intermediate State (Corbel et al.
583: 2004). A radio flare ejection is observed for GRO~J1655$-$40 during
584: the revolution 291, just before the X- and $\gamma$-ray flux peak.
585: %is accompanied by the maximum of luminosity in the X- and
586: %$\gamma$-ray domain few days later (during revolution 293, see
587: %Figure 1 and the details
588: % about the high energy evolution in different energy bands in Table 3).
589: Such a behavior has also been observed for the black hole candidate
590: H1743-322 and GX339-4 (Joinet et al. 2005, Corbel et al. 2000,
591: Fender et al. 1999) and could be linked to a coronal ejection event.
592: The decrease of the Compton y-parameter (see Table 6), and the
593: appearance of the non thermal component ( addition of the powerlaw
594: component) indicate large changes in the corona geometry.
595: % Such small
596: %ejection events, related to relativistic plasma events have
597: %also been observed during state transitions of
598: % GX339-4 (Corbel et al. 2000, Fender et al. 1999).
599:
600:
601: Moreover, the small reflection fraction combined with
602: a high coronal temperature (or high energy cutoff) can be
603: explained by the model of
604: Beloborodov (1999a)
605: which argues that the corona above the accretion
606: disk is fed by magnetic flares or by a non-static coronae
607: (Malzac et al. 2001). GRO~J1655$-$40 started to enter the IS from the end of
608: the
609: revolution 292. This state corresponds to the decay of the radio emission
610: (Figure 4) which is usually dominated by the decaying optically thin synchrotron
611: emission from the jet ejections. During this state, such an emission
612: is decoupled from the black hole system as the emitting electrons are far from
613: it. The radio emission is completely quenched starting from the SIMS (revolution
614: 294-B).
615:
616:
617: \section{Conclusion}
618:
619: We clearly observed a transition from the LHS to the HSS during the
620: rising phase of the 2005 outburst of GRO~J1655$-$40 between MJD 53425 and
621: MJD 53445. The outburst was covered with SPI, HEXTE, PCA and BAT
622: which allowed us to determine the high energy cutoff with a high precision.
623: During revolution 294, only data from the \textit{RXTE} and
624: \textit{Swift} observatory were available to constrain the high energy
625: cutoff during an intermediate state. An evolution of the high energy feature
626: was noticed during the LHS : it decreased from a value of about 200 keV down to
627: 130 keV when the source reached the maximum of luminosity in the $\gamma$-ray
628: domain (above 23-600 keV). For the HSS, a lower limit for the
629: high energy cutoff has been determined. We deduced that
630: there is no cut-off detectable in this spectral state.
631: %natural to interpret the non-detection as disappearance of the cutoff, rather
632: %than a much higher cutoff energy than the detectors could detect.
633: This decrease
634: corresponds to a decrease in the radio flux, and the cut-off disappeared
635: along with the radio jet. Finally, we also discussed on the relation
636: between state transition and the emission of jet in X-ray binaries
637: which is a subjet of significant importance in order to have a complete
638: view of the X-ray geometry. It would be interesting to perform a
639: broadband spectral fitting (from radio up to MeV) in order to
640: discriminate the different non thermal processes suggested in
641: this study.
642:
643: \acknowledgments
644:
645: The SPI project has been completed under the responsibility and leadership
646: of the CNES. We are grateful to ASI, CEA, DLR, ESA, INTA, NASA and OSTC
647: for support. \\
648: Specific softwares used for this work have been developed by L.
649: Bouchet. A. J. thanks N. Shaposhnikov and M. Rupen for all
650: informations concerning the radio detection and M.D.
651: Caballero-Garcia who provided the ISGRI data used in this paper. E.
652: K. acknowledges support of TUB\.ITAK Career Program 106T570, Turkish
653: National Academy of Sciences Young and Successful Scientist Award
654: and Marie Curie International Reintegration Grant
655: MIRG-CT-2005-017203. Finally the authors are grateful to the anonymous referee for the very fruitful comments.
656: %!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
657: \begin{thebibliography}{b}
658: \bibitem[Arnaud (1996)]{Arn}
659: Arnaud K. A. 1996, in ASP Conf. Ser., Vol. 101, Astronomical Data Analysis Software and Systems V, ed. G. H. Jacoby, J. Barnes J. (San Francisco: ASP), 17
660: \bibitem[Bailyn et al. (1995)]{Bai}
661: Bailyn, C. et al., 1995, Nature, 374, 701
662: \bibitem[Barthelmy et~al. (2005)]{Bar}
663: Barthelmy, S., et al., 2005, Space Sci. Rev., 120, 143
664: \bibitem[Belloni et~al. (2005)]{Bel}
665: Belloni, T., Homan, J., Casella, P., van der Klis, M.,
666: Nespoli, E., Lewin, W. H. G., Miller, J. M., M\'endez, M., 2005, A\&A, 440, 207
667: \bibitem[Beloborodov (1999)]{Bela}
668: Beloborodov, A.M., 1999a, ApJ, 510, L123
669: \bibitem[Beloborodov (1999)]{Belb}
670: Beloborodov, A.M., 1999b, MNRAS, 305, 181
671: \bibitem[Bradt et~al. (1993)]{Bra}
672: Bradt, H.V., Rothschild, R.E., \& Swank, J.H. 1993, A\&AS, 97, 355
673: \bibitem[Brocksopp et al. (2006)]{Bro}
674: Brocksopp, C., et al., 2006, MNRAS, 365, 1203
675: \bibitem[Caballero-Garcia et al. (2007)]{Cab}
676: Caballero-Garcia, M. D., et al., 2007, ApJ, 669, 534
677: \bibitem[Corbel et al. (2000)]{Cor}
678: Corbel, S., Fender, R. P., Tzioumis, A. K., Nowak, M.,
679: McIntyre, V., Durouchoux, P., Sood, R., 2000, A\&A, 359, 251,
680: \bibitem[Corbel et al. (2004)]{Corb}
681: Corbel, S., Fender, R. P., Tomsick, J. A., Tzioumis, A. K., Tingay, S.,
682: 2004, ApJ, 617, 1272
683: \bibitem[Diaz (2006)]{Dia}
684: Diaz Trigo, M., Parmar, A. N. , Miller, J.,Kuulkers, E., \& Caballero-Garcia, M. D.,2007, A\&A, 462, 657
685: \bibitem[Gallo et al. 2003]{Gal_01}
686: Gallo, E., Fender, R. P., \& Pooley, G. G., 2003, MNRAS, 344, 60, 72
687: \bibitem[Gehrels (2004)]{Geh}
688: Gehrels, N., et al., 2004 , ApJ, 611, 1005
689: \bibitem[George et al. (1991)]{Geo}
690: George, I. M., Fabian, A. C., 1991 , MNRAS, 249, 352
691: \bibitem[Greene et al. (2001)]{Gre01}
692: Greene, J., Bailyn, C. D., Orosz, J. A., 2001, ApJ
693: \bibitem[Fender et al. (1999)]{Fen}
694: Fender, R., et al., 1999 , ApJ, 519, 165
695: \bibitem[Foellmi (2006)]{Foe}
696: Foellmi, C., Depagne, E., Dall, T. H., \& Mirabel, I. F., 2006 , A\&A, 457, 249
697: \bibitem[Gallo et al. 2003]{Gal_02}
698: Gallo, E., Fender, R. P., \& Pooley, G. G., 2003, MNRAS, 344, 60, 72
699: \bibitem[Greene et al. 2001]{Gre}
700: Greene, J., Bailyn, C. D., Orosz, J. A., 2001, ApJ, 554, 1290
701: \bibitem[Homan (2005)]{Hom}
702: Homan J., \& Belloni T., 2005, Ap\&SS, 300, 107
703: \bibitem[Hjellming (1995)]{Hje}
704: Hjellming R. M., \& Rupen, M. P., 1995, Nature, 375, 464
705: \bibitem[Joinet A., et~al.(2005)]{Joi}
706: Joinet A., et al. 2005, ApJ, 629, 1008
707: \bibitem[Kuulkers (2000)]{Kuu}
708: Kuulkers, E., et al., 2000, A\&A, 358, 993
709: \bibitem[Magdziarz \& Zdziarski 1995]{Mag95}
710: Magdziarz, P., \& Zdziarski, A. A. 1995, MNRAS, 273, 837
711: \bibitem[Malzac et al. (2001)]{JM01}
712: Malzac, J., Beloborodov, A. M., Poutanen, J., 2001, MNRAS, 326, 417
713: \bibitem[Malzac et al. (2006)]{JM06}
714: Malzac, J., et al. 2006, A\&A, 448, 1125
715: \bibitem[Markwardt et al. (2005)]{Tor_01}
716: Markwardt, C. B.,Smith,E., \& Swank, J., 2005, Atel, 415, 1
717: \bibitem[McClintock \& Remillard (2006)]{Mcl}
718: McClintock, J.E., \& Remillard, R.A. 2006, in Compact Stellar X-ray Sources,
719: ed. W.H.G. Lewin \& M. van der Klis (Cambridge: Cambridge University Press), p157
720: \bibitem[Mendez (1998)]{Men}
721: M\'endez, M., Belloni, T. \&, van der Klis, M., 1998, ApJ, 499, 187
722: \bibitem[Mitsuda et~al.(1984)]{Mit}
723: Mitsuda, K., Inoue, H., Koyama, K., et al., 1984, \pasj, 36, 741
724: \bibitem[Narayan \& Yi (1995)]{nar}
725: Narayan, R., \& Yi, I., 1995, ApJ, 452, 710
726: \bibitem[Poutanen \& Fabian (1999)]{Po}
727: Poutanen, J., Fabian, A. C. 1999, MNRAS, 306, 31
728: \bibitem[Remillard (2006)]{Rem}
729: Remillard , R. A.,\& McClintock, J.E., 2006, ARAA, 44,49
730: \bibitem[Rothschild et al. (1998)]{Rot}
731: Rothschild, R. E., et al. 1998, ApJ, 496, 538
732: \bibitem[Shakura \& Sunyaev (1973)]{Shak}
733: Shakura, N. I, Sunyaev, R. A., 1973, A\&A, 24, 337
734: \bibitem[Shapiro et al.(2003)]{shap}
735: Shapiro, S.L., Lightman A.P., \& Eardley, D. M., 1976, ApJ, 204, 187
736: \bibitem[Shaposhnikov et al. (2007)]{Sha}
737: Shaposhnikov, N. , Swank, J., Shrader, C.R., Rupen, M., Beckmann, V.,
738: Markwardt, C.B., Smith, D.A., 2007, ApJ, 655, 434
739: \bibitem[Sunyaev et~al.(1980)]{Sun}
740: Sunyaev, R., \& Titarchuk, L., 1980, A\&A, 86, 121
741: \bibitem[Tingay et al. (1995)]{Tin}
742: Tingay, S. J., et al., 1995, Nature, 374, 141
743: \bibitem[Titarchuk et al.(1994)]{tit}
744: Titarchuk, L. 1994, ApJ, 434, 570
745: \bibitem[Titarchuk et al. (1997)]{Tit}
746: Titarchuk, L., 1997, ApJ, 487, 834
747: \bibitem[Tomsick et al. (2001)]{Tom}
748: Tomsick, J. A., Corbel, S. \& Kaaret, P. 2001, ApJ, 563, 229
749: \bibitem[Torres et al. (2005)]{Tor_02}
750: Torres, M. A. P., Steeghs, D., Jonker, P. \& Martini, P., 2005, Atel, 417, 1
751: \bibitem[Van de Hooft et al. (1998)]{Hoo}
752: Van de Hooft, F., Heemskerk, M. H. M., Alberts, F., \& van Paradijs, J., 1998
753: , A\& A, 329, 538
754: \bibitem[Vedrenne et~al.(2003)]{Ved_01}
755: Vedrenne, G., et al. 2003, \aap, 411, L63
756: \bibitem[Winkler et~al.(2003)]{Ved_02}
757: Winkler, C., et al. 2003, A\&A, 411,1
758: \bibitem[Zhang et al. (1994)]{Zha}
759: Zhang, S. H., et al., 1994, IAUC, 6106
760: \bibitem[Zhang et al. (1997)]{Zhab}
761: Zhang, S. N., Ebisawa, K., Sunyaev, R., Ueda, Y.,
762: Harmon, B. A., Sazonov, S., Fishman, G. J., Inoue, H., Paciesas, W. S.,
763: \& Takahashi, T., 1997, ApJ, 479, 381
764: \bibitem[Zdziarski et~al.(2002)]{Zdza}
765: Zdziarski, A. A., Poutanen, J., Paciesas, W, S., Paciesas, W.S.,\& Wen, L., 2002, \apj, 578, 357
766: \bibitem[Zdziarski et~al.(2005)]{Zdzb}
767: Zdziarski, A. A., et al., 2005, MNRAS, 360, 825
768: \end{thebibliography}
769:
770:
771: \clearpage
772: \thispagestyle{empty}
773: \setlength{\voffset}{-15mm}
774: \input{tab1}
775: \clearpage
776: \setlength{\voffset}{0mm}
777: \input{tab2}
778: \clearpage
779: \input{tab3}
780: \clearpage
781: \input{tab4}
782: \clearpage
783: \input{tab5}
784: \clearpage
785: \input{tab6}
786:
787: \clearpage
788: \begin{figure}
789: \begin{center}
790: %\includegraphics[scale=0.8]{light-gro.ps}
791: \includegraphics[scale=0.4]{f1.eps}
792: \end{center}
793: %\hspace{2cm}
794: %\vspace{-0.5cm}
795: %\vspace{-0.4cm}
796: \caption{Upper panel : ASM (1.5-12 keV), SPI (23-51 keV)
797: light curves of GRO~J1655$-$40
798: during the rising phase of the 2005 outburst. The legend of letters
799: are: a=293$_{sw}$, b=294$_{sw}$-A,c=294$_{sw}$-B, d=295$_{sw}$.
800: The flux of points a,b,c,d has been extracted from BAT observations. The
801: different states harbored by the source are summarized on the graph
802: (see text for the definition): LHS=Low Hard State, HIMS= Hard
803: Intermediate State, SIMS = Soft Intermediate State, HSS=High Soft State. Two arrows
804: indicate (a) the period associated to a radio flare event (see text).
805: Lower panel : the evolution of the hardness from ASM
806: in the 3-12 keV energy range which is
807: defined as the ratio of the ASM fluxes $\Phi$
808: in two energy ranges $\Phi$ [5.-12. keV]/$\Phi$ [3.-5. keV].}
809: \label{fig:fig1}
810: \end{figure}
811:
812:
813:
814:
815: \begin{figure}
816: \begin{center}
817: \includegraphics[scale=0.4]{f2.eps}
818: \end{center}
819: %\hspace{2cm}
820: %\vspace{-0.5cm}
821: %\vspace{-0.4cm}
822: \caption{Spectra of GRO~J1655$-$40 from
823: PCA, SPI and HEXTE data fitted simultaneously
824: with the model described in Table 4.
825: PCA data for revolutions 289 (black), 292(indigo), 293 (brown);
826: HEXTE data for revolutions 289 (red), 292(violet), 293 (grey);
827: SPI data for revolutions 289 (yellow), 292(blue), 293 (green).}
828: \label{fig:fig4}
829: \end{figure}
830:
831: \begin{figure}
832: \begin{center}
833: %\includegraphics[scale=0.8]{clgxb.ps}
834: %\includegraphics[scale=0.8]{asmb.eps}
835: \includegraphics[scale=0.4]{f3.eps}
836: \end{center}
837: %\hspace{2cm}
838: %\vspace{-0.5cm}
839: %\vspace{-0.4cm}
840: \caption{Spectra of GRO~J1655$-$40 from
841: PCA, SPI and HEXTE data fitted simultaneously
842: with the model described in Table 4.
843: PCA data for revolutions 293 (brown), 294-B (black), 295+296 (indigo);
844: HEXTE data for revolutions 293 (gray), 294-B (red), 295+296 (violet);
845: SPI data for revolutions 293 (green), 295+296 (blue)
846: and BAT data for revolution 294$_{sw}$-B (yellow).}
847: \label{fig:fig5}
848: \end{figure}
849:
850: \begin{figure}
851: \begin{center}
852: \includegraphics[scale=0.3,angle=270]{f4.eps}
853: \end{center}
854: %\hspace{2cm}
855: %\vspace{-0.5cm}
856: %\vspace{-0.4cm}
857: \caption{Evolution of the high energy cutoff ($E_c$ expressed in keV)
858: derived from the model given in Table 4, as a function of the
859: radio flux measured by VLA at 8.460 GHz (Shaposhnikov et al. 2007). The
860: \textit{INTEGRAL} revolution numbers have been mentioned on the graph. There are no radio
861: data available after revolution 293.}
862: \label{fig:fig7}
863: \end{figure}
864:
865:
866: \begin{figure}
867: \begin{center}
868: %\includegraphics[scale=0.3,angle=270]{f5.eps}
869: \includegraphics[scale=0.3,angle=270]{f5.eps}
870: \end{center}
871: %\hspace{2cm}
872: %\vspace{-0.5cm}
873: %\vspace{-0.4cm}
874: \caption{$\Gamma$-$E_c$ contour plot (revolution 290) for the simultaneous PCA, HEXTE, ISGRI and SPI
875: observations using the model described in Table 4 . The curves refer to
876: $\delta \chi ^2$ = 2.30, 4.61, 9.21, corresponding to confidence
877: levels of
878: 68, 90 and 99\% for two interesting parameters.}
879: %\label{fig:fig7}
880: \end{figure}
881:
882:
883: \end{document}
884: