1: %\documentstyle[psfig,apjfonts,emulateapj5,multicol]{aastex}
2: %\documentstyle[psfig,aaspp4,12pt,multicol]{article}
3: %\documentstyle[emulateapj,apjfonts,psfig]{article}
4: \documentclass[12pt,preprint]{aastex}
5: %\input epsf
6:
7: \lefthead{M.~D.~Caballero-Garc\'{\i}a et~al.}
8: \righthead{GRO~J1655$-$40}
9:
10: %\received{Date}
11: %\revised{Date}
12: %\accepted{Date}
13:
14: %\journalid{vol}{date}
15: %\articleid{1}{4}
16: %\paperid{id}
17:
18: %\cpright{AAS}{1999}
19: %\ccc{x}
20:
21: \renewcommand{\markcite}[1]{#1}
22:
23:
24: \begin{document}
25:
26:
27: \title{The high energy emission of GRO~J1655$-$40 as revealed with INTEGRAL spectroscopy of the 2005 outburst}
28:
29: \author{M.~D.~Caballero-Garc\'{\i}a\altaffilmark{1},
30: J.~M.~Miller\altaffilmark{2},
31: E.~Kuulkers\altaffilmark{3},
32: M.~D\'{\i}az Trigo\altaffilmark{3},
33: J. Homan \altaffilmark{4},
34: W.~H.~G. Lewin \altaffilmark{4},
35: P.~Kretschmar\altaffilmark{3},
36: A.~Domingo\altaffilmark{1},
37: J.~M.~Mas-Hesse\altaffilmark{5},
38: R.~Wijnands\altaffilmark{6},
39: A.~C. Fabian\altaffilmark{7},
40: R.~P. Fender\altaffilmark{8,6},
41: M.~van der Klis\altaffilmark{6}
42: }
43:
44: \email{mcaballe@laeff.inta.es}
45:
46: \altaffiltext{1}{LAEFF-INTA, P.O. Box 50727, 28080 Madrid, Spain, mcaballe@laeff.inta.es}
47: \altaffiltext{2}{Department of Astronomy, University of Michigan, 500 Church Street, Ann Arbor, USA, MI 48109, jonmm@umich.edu}
48: \altaffiltext{3}{ESA/ESAC, Urb. Villafranca del Castillo, PO Box 50727, 28080 Madrid, Spain}
49: \altaffiltext{4}{MIT Kavli Institute for Astrophysics and Space Research, 77 Massachusetts Avenue, Cambridge, MA 02139, USA}
50: \altaffiltext{5}{CAB (CSIC-INTA) P.O. Box 50727, 28080 Madrid, Spain}
51: \altaffiltext{6}{Astronomical Institute ``Anton Pannekoek'', Kruislaan 403, University of Amsterdam, Amsterdam, 1098 SJ, The Netherlands}
52: \altaffiltext{7}{University of Cambridge, Institute of Astronomy, Cambridge CB3 OHA, UK }
53: \altaffiltext{8}{School of Physics and Astronomy, University of Southampton, Highfield, Southampton, SO17 1BJ, UK}
54:
55:
56:
57:
58:
59:
60: \label{firstpage}
61:
62: \begin{abstract}
63: %
64:
65: We present broadband (3$-500$\,keV) {\rm INTEGRAL} X-ray spectra and
66: X-ray/optical light curves of the luminous black hole X-ray transient
67: and relativistic jet source GRO~J1655$-$40. Our analysis covers four
68: Target of Opportunity observations of the outburst that started in
69: February 2005. We find that the high energy emission of GRO~J1655$-$40
70: can be modelled well with an {\it unbroken} power-law (with photon
71: indices of ${\Gamma}=1.72{\pm}0.03,2.21{\pm}0.04$ for the first and
72: the second observations, respectively). These correspond to hard and
73: thermal dominant states, respectively. In contrast to many other
74: black hole spectra, high energy complexity in the form of a break or
75: cut-off is not required for the hard state, contrary to previous
76: expectations for this state. We show for the first time that
77: Comptonization by non-thermal electrons is the dominant process for
78: the high energy emission in the hard state. We discuss our results in
79: terms of models for broad-band emission and accretion flows in
80: stellar-mass black holes.
81:
82: \end{abstract}
83:
84: \keywords{Black hole physics -- stars: binaries
85: (GRO~J1655$-$40) -- gamma rays: observations -- accretion, accretion disks
86: -- radiation mechanisms: non-thermal -- radiation mechanisms: thermal}
87:
88:
89: \section{Introduction} \label{introd}
90: %
91:
92: GRO~J1655$-$40 is a black hole X-ray binary whose parameters are well
93: known. Also called Nova Scorpii 1994, the source was discovered with
94: the Burst and Transient Source Experiment (BATSE) on board the Compton
95: Gamma-Ray Observatory (CGRO) on 1994 July 27 \citep{zhang94}. The
96: optical counterpart was discovered soon after by \citet{bailyn95a}
97: ($V{\sim}14.4$ mag). Subsequent optical studies regarding the
98: properties of the light curve during the outburst and quiescent period
99: showed that the system is an LMXB composed of a blue subgiant
100: (spectral type F4 IV) as the secondary and a black hole (hereafter BH)
101: as the primary ($m_{\rm BH}=7.02{\pm}0.22$$M_{\odot}$,
102: \citet{orosz97a}). The system is located at a distance of 3.2 kpc as
103: measured by \citet{tingay95}. Although \citet{foellmi06} have recently
104: suggested a smaller distance, their parameters imply that the
105: donor star would not fill its Roche lobe. \citet{bailyn95b}
106: established the orbital inclination of the system to be
107: ${\simeq}70^{\circ}$ (see also \citet{orosz97b} and \citet{hooft98});
108: an independent determination made by \citet{kuulkers98} based on X-ray
109: flux dips constrained the inclination of the system to be
110: $60^{\circ}-75^{\circ}$. The inclination of the inner disk may be as
111: high as $85^{\circ}$, indicating a slight mis-alignment with the
112: binary system \citep{hjellming95}.
113:
114: GRO~J1655$-$40 has displayed some of the most extreme behavior and
115: phenomena yet observed from any black hole X-ray transient.
116: \citet{strohmayer01} discovered a pair of high-frequency
117: quasi-periodic oscillations (QPOs) at $300$ and $450$ Hz in power
118: spectra from the 1996/1997 outburst of the source. If the higher
119: frequency is associated with the Keplerian frequency at the innermost
120: stable circular orbit (ISCO, \citet{shapiro83}, located at
121: $R_{ISCO}=6~R_{g}$ for a Schwarzschild black hole or at $R_{ISCO}=
122: 1.25~R_{g}$ in the case of a maximal Kerr BH of $a=0.998$; where $R_{g}=GM/c^{2}$
123: is the gravitational radius and $a=cJ/GM^{2}$; see \citet{bardeen72}
124: and \citet{thorne74}), the frequency observed indicates that
125: GRO~J1655$-$40 harbors a spinning black hole. This suggestion is
126: broadly consistent with spin estimates based on fits to skewed Fe~K
127: emission lines ($a \geq 0.9$ as per $r \leq 1.4~R_g$,
128: \citet{miller05}). GRO~J1655$-$40 has also ejected extremely
129: relativistic radio jets \citep{hjellming95}. Finally, unbroken
130: power-law emission (i.e.\ without a cutoff) from GRO~J1655$-$40 has
131: been detected out to 800\,keV \citep{tomsick99}, offering a crucial
132: insight on high energy processes in black hole systems.
133:
134: High energy processes and the periods of correlated behavior known as
135: ``states'' in black hole binaries are the focus of this paper. Three
136: active states are commonly recognized in the soft X-ray domain (see
137: \citet{remillard06}): the non-thermal dominant or hard state (formerly
138: called low/hard), the thermal-dominant state (formerly called
139: high/soft), the steep power law state (hereafter SPL). Additionally
140: there are transitions between these states which
141: are often referred to as intermediate states. In the hard state, the
142: soft X-ray emission is very weak and the spectrum is dominated by some
143: kind of non-thermal emission that is broadly consistent with a power
144: law (hereafter PL) at higher energies (${\gtrsim}20$\,keV). It has
145: been known for a long time that black hole spectra in the hard state
146: are exponentially cut off at ${\sim}100$\,keV (see \citet{grove98b}
147: and references therein). A radio jet is usually inferred in this state
148: from flat radio spectra (see \citet{fender04} for an unified model for
149: X-ray states and radio emission in black hole X-ray transients). In
150: the thermal-dominant state, the disk dominates the X-ray emission.
151: Although high energy emission (${\gtrsim}20$\,keV) also is seen in
152: this state, it is weaker, generally steeper, and extends up to 800\,keV
153: without any break \citep{grove98b}. In the SPL, these emission
154: components are combined -- both the disk and the non-thermal
155: power-law are strong, although the disk no longer follows the
156: $L{\propto}T^4$ relation that is
157: observed in the soft state \citep{kubota01}; \citep{saito07}.
158:
159: Several models have been proposed to explain the physical conditions
160: of the innermost accretion flow and the nature of hard X-ray emission
161: in the various states, but a clear picture has not yet been achieved.
162: At low mass accretion rates hard X-ray emission may arise in an inner
163: region filled by a hot ($kT_{e}{\sim}100$\,keV), radiatively
164: inefficient, advection-dominated accretion flow (ADAF)
165: (\citet{narayan96}, \citet{esin01}). A recent study made by
166: \citet{yuan04} shows that the ADAF scenario is not able to explain the
167: relatively high luminosities that have been observed in the hard state
168: of some black-hole X-ray binaries. Alternatively, some recent models
169: suggest that direct synchrotron emission, and/or synchrotron
170: self-Comptonization in a jet may dominate the hard X-ray emission
171: (\citet{markoff01}, \citet{markoff03}, \citet{markoff05}).
172: Both at low and high mass accretion rates thermal Comptonization
173: in a corona (fed by seed photons from the disk) may also be an
174: important source of hard X-ray emission (see, e.g., \citet{frontera03}).
175: An alternative source of Comptonization which is less reliant on
176: the disk is bulk motion Comptonizing (BMC, \citet{ebisawa96};
177: see also \citet{titarchuk02} for a co-moving Comptonizing medium).
178: In this case, the Comptonization is due to bulk motion of an
179: almost free falling (convergent accretion) flow close to the black hole.
180:
181: A thermal distribution of the Comptonizing particles (electrons)
182: necessarily leads to a turnover in the emitted spectrum around $kT$;
183: therefore, thermal Comptonization should lead to a turnover near to the electron
184: temperature of the corona, $kT_{e}$. Synchrotron emission and
185: non-thermal electron distribution do not necessarily predict
186: such a turn-over, however, and this difference provides an
187: observational tool to distinguish which processes dominate the hard
188: X-ray emission in different black hole states. The high energy
189: sensitivity of INTEGRAL is especially well-suited to this purpose.
190:
191: In this paper, we report on observations of GRO~J1655$-$40 made with
192: INTEGRAL during the outburst which began in February of 2005
193: \citep{markwardt05}. In Section \ref{observ} we describe our
194: observations and in Section \ref{curves} we show the light curves
195: obtained with JEM-X, ISGRI, SPI and OMC (instruments on-board
196: INTEGRAL) and RXTE and discuss the possible origin of their evolution.
197: In Section \ref{spectext} we present our spectral analysis made with
198: JEM-X, ISGRI and SPI. Finally, in Section \ref{discuss} we discuss our
199: results in the context of the different models and theories present in
200: the literature.
201:
202:
203: \section{Observations} \label{observ}
204:
205: \subsection{INTEGRAL observations}
206:
207: The data were obtained with INTEGRAL and cover the first part of the
208: 2005 outburst using the following instruments: the SPectrometer on
209: INTEGRAL (SPI; \citet{vedrenne03}), the INTEGRAL Soft Gamma-Ray Imager
210: (ISGRI; \citet{lebrun03}), the Joint European X-ray Monitor (JEM-X;
211: \citet{lund03} and the Optical Monitoring Camera (OMC;
212: \citet{mas-hesse03}). ISGRI is optimized for 15\,keV to 10 MeV imaging
213: and SPI is optimized for high-resolution spectroscopy in the 18\,keV
214: to 8 MeV band. The former provides an angular resolution of $12'$
215: full-width half maximum (FWHM) and an energy resolution, $E/{\Delta}E$
216: of ${\approx}12$ (FWHM) at 100\,keV. SPI provides an angular
217: resolution of $2.8^{\circ}$ (FWHM) and an $E/{\Delta}E$ of $430$ FWHM
218: at 1.3 MeV. JEM-X has a fully coded Field of View (FOV) of
219: $4.8^{\circ}$ diameter and an angular resolution of $3'$ FWHM. JEM-X
220: has medium resolution spectral capabilities in the energy range of
221: 3-35\,keV. The OMC is an optical monitor, with a FOV of
222: $5^{\circ}{\times}5^{\circ}$ and an astrometric resolution of $<1$
223: arcsec, and performs optical photometry in the V-band down to $18^{th}$
224: magnitude.
225:
226: Our program consisted of 4 ToO observations of 100 ks each, spread
227: from 27 February to 11 April of 2005 (we will refer to these as epochs
228: 1--4 below; see Table \ref{tobserv} for more details). The difference
229: in the exposure times between the INTEGRAL instruments given in Table
230: \ref{tobserv} (JEM-X, ISGRI and SPI) are due to the difference in the
231: dead times and variation in the efficiency along the fields of view.
232: The dithering pattern used during the observations was $5{\times}5$
233: (square of 25 pointings separated by 2.17 degrees centered on the main
234: target of the observation); this is the best pattern in order to
235: minimize background effects for the SPI and ISGRI instruments in
236: crowded fields. Data reduction (in the case of JEM-X, ISGRI and SPI)
237: was performed using the standard Off-line Science Analysis (OSA) 5.1
238: software package available from the INTEGRAL Science Data Centre
239: (ISDC; \citet{courvoisier03}). In the case of SPI, because of the
240: lower angular resolution and crowded field of view in ${\gamma}$-rays
241: at this position of the sky
242: ($l$,$b$)=($344.98^{\circ}$,$+2.46^{\circ}$), (see Figure
243: \ref{mosaic}), we used a non-standard procedure in the analysis of the
244: data, described in \citet{deluit05} and \citet{roques05}. For the same
245: reason, in the case of OMC, we used a standard pipeline available in
246: OSA 6.0 (recently delivered) for extraction of fluxes. For OMC all
247: the public data available from ISDC were downloaded (this resulted
248: only in a slight increase in the ammount of data). Because of the
249: steep fall in response in the case of JEM-X, and because of its
250: reduced FOV (${\approx}5^{\circ}$ of diameter), we limited the
251: pointing radius with respect to the GRO~J1655$-$40 position to be
252: within $4^{\circ}$. In the case of SPI and ISGRI, with large Fully
253: Coded Fields of Views (FCFOV) ($16^{\circ}{\times}16^{\circ}$ for SPI
254: and $8.3^{\circ}{\times}8^{\circ}$ for ISGRI), pointing selections
255: were not necessary. In total, 199 individual pointing (or Science
256: Window -- each having exposure times lasting from 1800 to 3600 s and
257: following a $5{\times}5$ dithering pattern on the plane of the sky
258: \citet{courvoisier03}) data were used for both SPI and ISGRI, 96
259: pointings for JEM-X and 66 pointings for OMC.
260:
261:
262: \subsubsection{Extraction of light curves}
263:
264: GRO~J1655$-$40 was covered with INTEGRAL as part of the Galactic
265: Bulge Monitoring Program \citep{kuulkers07}. Precisely at the start of
266: this program GRO~J1655-40 was reported to become active
267: \citep{markwardt05}. The subsequent outburst of GRO~J1655-40 was also
268: followed with the RXTE/ASM and with a dense program of pointed
269: RXTE/PCA observations (Homan 2005). In Figure \ref{lcurves} we show
270: the ISGRI (20-60 and 60-150\,keV), RXTE/ASM (2-12\,keV) and OMC
271: (optical) light curves. The light curve derived from the IBIS/ISGRI in
272: the 20-60 and 60-150 keV energy bands had 1\,800 sec exposures (about
273: 150 counts/s correspond to 1 Crab on-axis in the 20-60\,keV energy
274: band, see Appedix A in \citet{kuulkers07}). GRO J1655-40 was observed
275: at a large off-axis angle (${\approx}15^{\circ}$ from the center of
276: the field of view; so is in the partially coded field of view of
277: IBIS/ISGRI and not visible with X-ray monitor JEM-X).
278:
279: Since GRO~J1655$-$40 is located in a very crowded FOV for the OMC, in
280: the flux extraction process we force the photometric aperture to be
281: centered at the source coordinates, which are taken from the OMC Input
282: Catalogue \citep{domingo03}.
283:
284: The typical limiting magnitude of the OMC in the Galactic bulge is
285: between $V=15$ and $V=16$ ($3{\sigma}$). This value depends strongly
286: on sky background and source contamination. In our case
287: ($l$,$b$)=($344.98^{\circ}$,$+2.46^{\circ}$) we can confidently get
288: limiting magnitudes down to $V=16$ ($3{\sigma}$). For each INTEGRAL
289: pointing, the OMC monitors the sources in its FOV by means of shots of
290: variable integration time. Typical values in the range 10 to 200\,s
291: (currently 10, 50 200\,s) are used to optimize sensitivity and to
292: minimize read-out noise and cosmic-ray effects. For the faintest
293: objects, several 200\,s exposures in the same pointing can be combined
294: during data analysis on the ground. We obtained one photometric point
295: by combining several 200\,s OMC shots. To increase the signal to
296: noise ratio, we combined the individual photometric points of every
297: pointing and calculated the final photometric points by making running
298: averages in order to minimize the dispersion.
299:
300: \subsubsection{Extraction of spectra}
301:
302: For JEM-X and ISGRI, individual spectra were obtained for each
303: pointing. The spectra were then combined to obtain an averaged
304: spectrum per epoch using the $spe\_ pick$ OSA tool and standard
305: procedures (see \citet{jemx05} and \citet{ibis05}) to rebin the
306: response matrices. In the case of SPI, we derived directly one
307: spectrum per revolution. The SPI spectra were extracted over an energy
308: range of (23-800)\,keV (with 26 logarithmic bins) using the SPIROS
309: package within OSA, applying maximum likelihood optimization
310: statistics \citep{skinner03}. Due to the crowdedness of the field and
311: the low spatial resolution of the instrument (i.e. sources separated
312: less than $2.8^{\circ}$ can not be resolved) it was necessary to apply
313: special techniques to extract spectra (see \citet{roques05} and
314: \citet{deluit05}). This included accounting for variability of the
315: sources present in the FOV. The background \footnote{Due to the fact
316: that JEM-X, ISGRI and SPI are detectors based in coded mask optics,
317: the detection of sources is made in basis of a deconvolution process,
318: taking also into account background, in an iterative process called
319: IROS (Iterative Removal Of Sources). In OSA 5.1 the spectra obtained
320: are always background subtracted and it is not necessary to apply in
321: XSPEC any background correction to the obtained spectra. The light
322: curves obtained are also background subtracted. However, in JEM-X
323: analysis it is possible to skip the level of background
324: subtraction. We refer to the \citet{osaspi}, \citet{goldwurm03},
325: \citet{westergaard03} and \citet{dubath05} for more details.} was
326: determined by using flat fields from particular INTEGRAL revolutions
327: (revolution 220 in our case -- i.e. the time closest publicly available
328: flat field to our period of observations), and the use of background method $3$
329: (determination of the background based on some specific flat field
330: INTEGRAL observations). Due to the fact that SPI is a high resolution
331: spectrograph in its energy range (20-8\,000\,keV) with a reduced
332: number of detectors (namely, 19) it is not optimized in the detection
333: of sources without taking into account previous information about
334: the spatial distribution of them. Thus, we used as an input catalog
335: of sources those obtained from the ISGRI in the (20-40)\,keV mosaic images
336: (see Figure \ref{mosaic}).
337:
338: The signal from GRO~J1655$-$40 was too soft to detect any emission in
339: the last two epochs with SPI. We combined single revolution SPI
340: spectra from each epoch, since there was no significant evolution and
341: in order to increase the signal to noise ratio. This resulted in one
342: SPI spectrum in revolution 290 and another one combining the
343: revolutions 295 and 296. The same was done for the low energy
344: instruments (JEM-X and ISGRI), thus obtaining four spectra, namely one
345: for each epoch, as can be seen in Table \ref{tobserv}. We applied
346: 2$\%$ systematic errors to the JEM-X, ISGRI and SPI spectra. We
347: restricted our analysis in the energy ranges of $5-30$, $23-600$ and
348: $23-800$\,keV for the JEM-X, ISGRI and SPI spectral analysis, as
349: recommended in \citet{lubinski05}. The SPI and ISGRI spectra were
350: rebinned at high energies (${\gtrsim}200$\,keV) with the FTOOL
351: $grppha$ procedure to reach the detection level of $3{\sigma}$.
352:
353:
354: \section{Analysis of light curves} \label{curves}
355:
356: Figure \ref{lcurves} shows the GRO~J1655$-$40 light curves obtained
357: with ISGRI (from the INTEGRAL Galactic Bulge Monitoring Program;
358: \citet{kuulkers07}) in two energy bands (60-150\,keV and 20-60\,keV)
359: together with the OMC (optical). The public RXTE/ASM (2-12\,keV)
360: light curve in the same period of time is shown in the same Figure for
361: comparison.
362:
363: The light curve in Figure \ref{lcurves} does only show the first month
364: of outburst of GRO~J1655$-$40. For a full outburst light curve, refer
365: to \citet{brocksopp06}. \citet{brocksopp06} show the X-ray light
366: curve obtained by SWIFT, jointly with that obtained in the optical
367: and ultraviolet band using the Ultraviolet/Optical telescope on-board
368: SWIFT (UVOT). In Figure \ref{lcurves2} we show the RXTE/PCA light
369: curve in the range (2-60)\,keV plus a hardness ratio (calculated as
370: the ratio of the 9.4-18.2\,keV and 2.8-5.7\,keV count rates) evolution
371: of the source during the entire outburst, taken from \citet{homan07}.
372: The light curves were made from 520 RXTE observations, with one
373: (averaged) data point for each observation. The horizontal lines shown
374: in Figures \ref{lcurves} and \ref{lcurves2} indicate the time
375: intervals (100 ks each; in 4 observations) over which our average
376: spectra were obtained. Note that black hole transients usually begin
377: and end their outbursts in the hard state (see \citet{nowak95},
378: \citet{fender04}, \citet{homan05}) and the 2005 outburst of
379: GRO~J1655--40 is no exception. As can be seen in Figure
380: \ref{lcurves}, the beginning of the outburst first started at high
381: energies ($20-150$\,keV) rather than at softer X-rays ($2-10$\,keV).
382: Moreover, as can also be seen from the hardness ratios in Figure
383: \ref{lcurves2}, the observations of epoch 1 correspond to the hard
384: state of black holes (see Section \ref{discuss}). The other three
385: epochs were done during a period when the source spectrum was much
386: softer. However, it should be noted that it was only later in the
387: outburst that the softest spectra (corresponding to the
388: thermal-dominant state) were observed (see Figure \ref{lcurves2}).
389:
390: The first indications of an impending outburst of GRO~J1655-40 came
391: from RXTE/PCA bulge-scan observations on 2005 February 19
392: \citep{markwardt05}. On February 20, observations made with the PANIC
393: camera on the 6.5\,m Magellan-Baade telescope at Las Campanas
394: Observatory revealed a J-band (near-infrared) magnitude of
395: J=13.2${\pm}$0.1 \citep{torres05}, indicating that GRO~J1655-40 was
396: brighter by ${\approx}$0.5 mag in J relative to its magnitude
397: in quiescence (J=13.7-14) \citep{greene01}. On the same date a radio
398: detection of the source was reported \citep{rupen05}.
399:
400: As can be seen from \citet{buxton05} and \citet{brocksopp06} the
401: optical light curve behaves differently from the X-rays. The optical
402: behaviour consists of an increase in the flux until a constant
403: level at MJD=53\,455. This behaviour can also be seen with the
404: OMC, because the flux in the optical light curve was increasing,
405: reaching a constant and detectable value ($V{\sim}15$ mag) in
406: MJD=53\,455. Then the optical flux became constant. It is also
407: interesting to note the rapid increase of the radio emission
408: at MJD${\approx}$53\,450, coinciding with our epoch 3 of
409: observations and also with the beginning of the plateau in both the
410: optical and infrared light curve.
411:
412:
413: \section{Spectral analysis of INTEGRAL data} \label{spectext}
414:
415: We performed fits to the combined JEM-X, ISGRI and SPI spectra, for
416: each of the four epochs (see Table \ref{tobserv}) using XSPEC
417: \citep{arnaud96} v.11.3. All errors quoted in this work are 90\%
418: confidence errors, obtained by allowing all variable parameters to
419: float during the error scan. In all fits, we fixed the value of the
420: column density to $N_{H}=8.0{\times}10^{21}$ atoms$/cm^{2}$ as
421: obtained by \citet{diaz06} using XMM-Newton data of GRO~J1655$-$40
422: during the same outburst.
423:
424: Our main aim in these fits is to characterize the broad continuum as
425: seen with INTEGRAL and its broad energy coverage. To account for
426: uncertainties in relative instrument calibrations, we fixed JEM-X
427: multiplicative calibration constant to 1 and let that of ISGRI and SPI
428: free to vary in the fit for the different data sets as shown in
429: Table \ref{param_spec}.
430:
431: In Figures \ref{fspec1}, \ref{fspec2}, \ref{fspec3}, \ref{fspec4} and
432: \ref{fspec5}
433: we show spectra from each epoch because these provide a
434: convenient way to see the evolution of
435: the source during our observations. We fitted the spectra with
436: several models and the derived parameters are presented in Table
437: \ref{param_spec}. We modelled the spectra with simple and
438: phenomenological disk plus power-law models and common Comptonization
439: models described in Section \ref{introd}. We find that the former
440: are very successful (see Section \ref{powerlaw}), and there is no evidence for
441: any spectral break in the data up to ${\approx}600$\,keV
442: (see Section \ref{compttmod}). Although a simple power-law provided
443: marginally acceptable fits in the first two epochs, it was necessary to include
444: an iron line and edges components in the fits of the last two epochs
445: (see Section \ref{edges}).
446:
447: \subsection{Fits of the former two epochs}
448:
449: \subsubsection{Fits with a pure power-law} \label{powerlaw}
450:
451: We initially performed fits with a phenomenological power-law model
452: ($powerlaw$ model in XSPEC) in the spectra that extend up to
453: $600$\,keV.
454:
455: In the first epoch, the source spectra extend in the energy range of
456: $5-20$\,keV for JEM-X and $23-600$\,keV for both ISGRI and SPI,
457: respectively. In fitting with the $power-law$ model, we obtained a
458: reduced chi-square of ${\chi}^{2}_{\nu}=1.26$ with ${\nu}=63$ (${\nu}$
459: meaning the number of the degrees of freedom of our fit). The presence
460: of a multicolor disk component was not significantly required in our
461: fit. The value obtained for the photon index was
462: ${\Gamma}=1.72{\pm}0.03$, common for black holes in a hard spectral
463: state.
464:
465: For the second epoch of our observations, the source spectra extend in
466: the energy range of $5-30$\,keV, $23-500$\,keV and $23-400$\,keV for
467: JEM-X, ISGRI and SPI, respectively. This time the presence of emission
468: coming from a disk was significant, and we added an absorbed multicolor
469: disk component \citep{mitsuda84} to the power-law ($phabs(diskbb + powerlaw)$
470: model in XSPEC). We obtained ${\chi}^{2}_{\nu}=1.38$ with ${\nu}=88$.
471: The value obtained for the inner disk temperature was
472: $kT_{in}=1.25{\pm}0.01$\,keV. The spectrum is softer,
473: with a value for the photon index of ${\Gamma}=2.21{\pm}0.04$.
474:
475: For the last two epochs, in fitting with a phenomenological power-law
476: (plus a multicolor disk black body component), the spectra showed
477: large negative residuals at ${\approx}10$\,keV, compatible with
478: the presence of $Fe$ edges. The presence of
479: positive residuals in the range $6-8$\,keV is consistent with the
480: presence of a broad $Fe$ emission line. In Section \ref{edges} we go
481: into more detail on the fits of both these epochs.
482:
483:
484: \subsubsection{Fits with thermal Comptonization models} \label{compttmod}
485:
486:
487: As explained in Section \ref{introd}, all models involving thermal
488: Comptonization processes as the origin of the high energy emission
489: seen in black hole transient systems predict an energy turn-over at
490: the electron temperature of the corona. So, in order to assess the
491: role of Comptonization during the time of our observations, we fitted
492: with the thermal Comptonizing model of \citet{titarchuk94} (called
493: $CompTT$ in XSPEC), which deals with the special case of high
494: temperatures and/or small opacities (so the relativistic effects are
495: taken into account). We also made fits using the BMC model of
496: \citet{ebisawa96} (see also also \citet{titarchuk02} for a co-moving
497: Comptonizing medium), which deals with the Comptonization due to bulk
498: motion of the almost free falling (convergent accretion) flow close to
499: the black hole. Moreover, we fitted with a phenomenological
500: and multiplicative cutoff (called $highecut$ in XSPEC) model in
501: order to find any turn-over energy in our data which could be in
502: agreement with Comptonization models playing the major role in the
503: high energy emission of GRO~J1655$-$40.
504:
505: We thus fitted the spectra of the first two epochs \footnote{As we
506: explained in Section \ref{powerlaw} the presence of significant
507: residuals compatible with broad $Fe$ emission line, edges and likely
508: reflection could distort our undertanding of the continuum. Thus, we
509: do not use last two epochs in our study of the continuum
510: models. Moreover, the spectra in these epochs do not show significant
511: emission at high (${\gtrsim}150$\,keV) energies.} with following
512: models: $phabs(diskbb+powerlaw)highecut$ (hereafter called model 1),
513: and $phabs(diskbb+compTT)$ (hereafter called model 2) and $phabs(bmc)highecut$
514: (hereafter called model 3). In the case of the first epoch, the
515: presence of a multicolor disk component was not significantly required
516: in our fit. We assumed a spherical geometry for the Comptonizing medium
517: for the model 2 in all our fits.
518:
519: For epoch 1 of our observations, we obtained the following statistics for all
520: three models, i.e. ${\chi}^{2}_{\nu}=1.13,1.17,1.13$
521: (${\nu}=61,60,59$), respectively. We find that when fitting the data
522: with a cut-off power-law (model 1), while showing slightly better
523: statistics, the high energy cut-off can not be well constrained
524: (values between 5 and 54 keV are compatible with the data). The
525: obtained folding energy is $>253$\,keV. We made fits to separate {\rm
526: JEM-X+ISGRI} and {\rm JEM-X+SPI} to see if in one of the two data
527: sets a break is possible. In the former, we obtained uncostrained
528: values for the cut-off and the folding energies, being in the ranges
529: of 5-38\,keV and $>238$\,keV, respectively. In the second data set
530: there was not any break in the data as well. We conclude that the
531: cut-off features found are of instrumental origin and are not
532: physically meaningful. Thus, there is not any real cut-off in our
533: data. This issue is also supported by the fact that the parameters
534: optical depth and temperature of the electrons (${\tau}$ and $kT_{e}$)
535: when fitting with model 2 are unconstrained as well. This model
536: clearly is not a good description of the high energy spectrum
537: (${\gtrsim}20$\,keV). The model 3, while giving the lowest residuals,
538: does not allow to constrain the values for the parameters (neither of
539: $bmc$ and $highecut$ components). Thus, this model does not represent
540: a physical description of the data. The fact that model 3 provides
541: slightly better statistics than the power-law model described in
542: Section \ref{powerlaw} would be due to the fact that this model is a
543: convolution of both soft and hard emission components taking into
544: account the physical condition in the inner region of the accretion
545: disk. We derive from fitting with all these models that high energies
546: can not be reproduced by thermally upscattering photons by a
547: Comptonizing corona alone. In order to take this issue into account,
548: we fitted the spectrum of this epoch with the hybrid Comptonization
549: model EQPAIR \citep{coppi99} proper for very hot
550: plasmas which describes the physics and emission properties of hybrid
551: plasmas, where the particle distribution energy is approximately a
552: Maxwellian plus a power-law. This model deals a hot plasma cloud, mainly
553: modeled as a spherical corona around the compact object, illuminated
554: by soft thermal (Maxwellian) and
555: non-thermal (either power-law or monoenergetic distributed) electrons
556: that lose energy by Compton, Coulomb and bremsstrahlung interactions.
557: This model was shown to be successful in accounting for the high energy
558: spectra of Cygnus X-1 and other black hole candidates in different
559: spectral states and over a broad energy band ranging from soft X-rays
560: to gamma-rays (see e.g. \citet{mcconnell00}, \citet{mcconnell02},
561: \citet{cadolle06}, \citet{malzac06}).
562:
563: For epoch 2 of our observations, we obtained the following statistics for all
564: three models, i.e. ${\chi}^{2}_{\nu}=1.47,1.45,2.07$
565: (${\nu}=86,87,86$), respectively. Again, model 1 does not show a
566: break or cut-off below $500$\,keV (i.e. values of both cut-off and
567: folding energies can not be costrained) and model 3 is not a proper
568: fit to the data due to the bad statistics. In this epoch, the values
569: obtained for the optical depth of the Comptonizing medium and the
570: temperature of the electrons (${\tau}$ and $kT_{e}$) can not be
571: constrained. Thus, the data show that thermal Comptonization is not
572: the main process that generates the emission at high energies
573: (${\gtrsim}20$\,keV). As in epoch 1, we tested the model of
574: \citet{coppi99}, optimized for very hot coronae (see description in
575: paragraph above), this time coupled with $diskbb$ component to
576: describe the soft emission from the accretion disk.
577:
578:
579: \subsubsection{Fits with the {\rm EQPAIR} non-thermal Comptonization model} \label{comptteqpair}
580:
581: This model (EQPAIR) takes into account angle dependence, Compton
582: scattering (up to multiple orders), photon pair production, pair
583: anihilation, bremsstrahlung as well as reflection from a cold disk.
584: As noticed by \citet{coppi99}, if a spectrum extends up
585: to $500$\,keV high energy emission coming from a non-thermal
586: population of electrons is clearly present.
587:
588: As can be seen in Figure \ref{fspecpo1} and \ref{fspecpo2}, in which we
589: fitted both 1 and 2 epochs with a single power-law (see description
590: in Section \ref{powerlaw}), an small deficit of counts of the ISGRI spectra
591: above 200\,keV with respect to SPI is clearly present. Moreover, the
592: large bins of ISGRI spectra in this range of the spectra, would indicate
593: that the source is not detected above 200\,keV. This is a comprehensive
594: issue since SPI instrument is optimized for doing spectroscopy
595: up to 1\,MeV while ISGRI has poorer sensitivity above 200\,keV.
596: Taking these considerations into account, we limited our analysis of
597: ISGRI data in the energy range of $23-200$\,keV.
598:
599: The $eqpair$ model allows to inject a non-thermal electron distribution
600: with Lorentz factors between ${\gamma}_{min}$ and ${\gamma}_{max}$ and
601: a power-law spectral index ${\Gamma}_{p}$. The cloud is illuminated
602: by soft thermal photons emitted by an accretion disk. These photons
603: serve as seed for Compton scattering by both thermal and non-thermal
604: electrons. The systems is characterized by the power (i.e. luminosity)
605: $L_{i}$ supplied by its different components. We express each of them
606: dimensionlessly as a compactness parameter,
607: ${\rm l}_{i}={\rm L}_{i}{\sigma}_{\rm T}/({\rm R}{\rm m}_{e}{\rm c}^{3})$,
608: where ${\rm R}$ is the characteristic dimension and ${\sigma}_{\rm T}$
609: the Thompson cross-section of the plasma. Thus, ${\rm l}_{s}$, ${\rm l}_{th}$,
610: ${\rm l}_{nth}$ and ${\rm l}_{h}={\rm l}_{th}+{\rm l}_{nth}$ correspond to
611: the power in soft disk entering the plasma, thermal electron heating,
612: electron acceleration and the total power supplied to the plasma.
613: The total number of electrons (not including ${\rm e}^{+}$
614: and ${\rm e}_{-}$ pairs) is determined by ${\tau}$,
615: the corresponding Thompson optical depth, measured from the center
616: to the surface of the scattering region. We considered the source to
617: be moderately compact and fixed ${\rm l}_{s}=10$, as broadly reported
618: for other sources with similar characteristics.
619:
620: The disk spectrum incident on the plasma is modelled with a multicolor
621: disk blackbody as given by the $diskbb$ model in XSPEC of \citet{mitsuda84}.
622: The temperature of the inner edge of the accretion disk was fixed to
623: ${\rm kT}_{e}=0.5$\,keV. The limits of the accretion disk were fixed
624: to ${\rm R}_{max}=100{\rm R}_{g}$ and ${\rm R}_{min}=6{\rm R}_{g}$. We
625: attempted to fit both spectra of epochs 1 and 2 fixing the reflection
626: covering factor to zero.
627:
628: For epoch 1 of our observations we first performed a fit with non-thermal
629: electrons injected with a power-law distribution of Lorentz factors
630: from ${\gamma}_{min}=1.3$ to ${\gamma}_{max}=1000$. The upper and the
631: lower limits ${\gamma}_{min}$ and ${\gamma}_{max}$ were kept fixed
632: while fitting the power-law index ${\Gamma}_{p}$. This resulted in
633: an acceptable fit with a reduced chi-square of ${\chi}_{\nu}=1.21$
634: (${\nu}=57$). The unfolded broadband spectrum and residuals are shown
635: in Figure \ref{fspec1}. The best fit paramaters are presented in Table
636: \ref{param_spec}. ${\rm l}_{h}/{\rm l}_{s}$ is about unity, i.e.
637: intermediate between what is found in the hard state (4-10)
638: and the thermal dominant state (${\leq}0.4$) \citep{ibragimov05}. The
639: heating of the plasma is dominated by the non-thermal acceleration
640: (${\rm l}_{nth}/{\rm l}_{h}{\approx}1$). We also fitted the spectra
641: of this epoch considering a mono-energetically distributed population
642: of non-thermal injected electrons instead. The obtained reduced chi-square
643: is slightly better (${\chi}_{\nu}=1.19$, ${\nu}=57$). The values obtained
644: for ${\rm l}_{h}/{\rm l}_{s}$ and ${\rm l}_{nth}/{\rm l}_{h}$ remained
645: unchanged, except for an increasing of the Thompson scattering depth
646: (from $1.0{\pm}0.6$ to $2.7{\pm}0.4$, for the former and the second fit,
647: respectively). The increasing of the optical depth gives support to
648: the scenario of Comptonization through injected non-thermal electrons
649: as being the dominant mechansim in the sense that if this is done by
650: electrons with a mono-energetic distribution then a denser cloud would
651: produce the same effect as taking into account a broader distribution
652: (both in energy and spatially). Thus, we conclude that in order to
653: reproduce the spectrum of first epoch, we have to consider an almost
654: purely distribution of non-thermal accelerating particles.
655:
656: As in epoch 1, we tested the model of \citet{coppi99},
657: optimized for very hot coronae in observations of epoch 2, this time coupled
658: with $diskbb$ component to describe the soft emission from
659: the accretion disk. We performed a fit with non-thermal
660: electrons injected with a power-law distribution of Lorentz factors
661: from ${\gamma}_{min}=1.3$ to ${\gamma}_{max}=1000$. The upper and the
662: lower limits ${\gamma}_{min}$ and ${\gamma}_{max}$ were kept fixed
663: while fitting the power-law index ${\Gamma}_{p}$. This resulted in
664: an acceptable fit with a reduced chi-square of ${\chi}_{\nu}=1.44$
665: (${\nu}=83$). The unfolded broadband spectrum and residuals are shown
666: in Figure \ref{fspec2}. The best fit paramaters are presented in Table
667: \ref{param_spec}. ${\rm l}_{h}/{\rm l}_{s}$ again is about unity.
668: However, the heating of the plasma by non-thermal particles is practically zero
669: (${\rm l}_{nth}/{\rm l}_{h}{\approx}0$). The situation does not
670: improve by considering a mono-energetic distribution of the non-thermal
671: accelerated electrons. This issue is not surprising, since, as claimed
672: by \citet{coppi99}, a very good spectrum extending above 200\,keV is mandatory
673: in order to disentangle a likely population of non-thermal particles
674: in the source. In the case of our observations, the large bin at
675: ${\approx}200$\,keV shows that the source was not detected above
676: this energies. We conclude that while in epoch 2 of our observations
677: a break is not observable in the high energy data ${\gtrsim}20$ the
678: energy coverage is not great enough in order to test the presence
679: of non-thermal processes. This is due to the fact that in this period
680: the spectrum became to be very soft, compared to epoch 1, with
681: radiation detected to ${\approx}200$\,keV as maximum.
682:
683:
684: \subsection{Fits to epochs 3 and 4} \label{edges}
685:
686: \subsection{Fits with $Fe$ emission line and absorption edges}
687:
688: As shown in the shape of the residuals (see
689: Figure \ref{fspec3}), fits with simple multicolor disk model of
690: \citet{mitsuda84} and power-law did not give formally acceptable fits
691: to the last two epochs. The presence of large residuals at $6-8$\,keV
692: and around 10\,keV require us to take into account iron emission line,
693: iron absorption edges and likely disk reflection components. These
694: features are theoretically required in very ionized mediums as the
695: close vicinity of the black holes (see \citet{ueda98} for a former
696: detection of $Fe$ absorption lines and edges in GRO~J1655$-$40 in the
697: context of observations of the 1996 outburst and \citet{george91} and
698: \citet{laor91} for a description of the $Fe$ line profile produced in
699: accretion disks around black holes). The reflection component takes
700: into account the physical condition of the Compton-reflected continuum
701: of a source of hard X-rays by an accretion disk
702: \citep{gilfanov00}.
703:
704: We fitted our spectra with the following model: $phabs(diskbb +
705: powerlaw + gaussian)edge*edge$ and the results of the fits are shown
706: in Table \ref{param_spec}. Regarding the Gaussian $Fe$ emission line
707: component, we constrained the line center in the range of
708: $6.4-6.97$\,keV, which is the allowed range due to the different
709: ionization states of $Fe$. Also, we constrained the width
710: (${\sigma}$) of the iron emission line to be 1\,keV as a fiducial
711: maximum value in order to get convergence of our fits. Regarding the
712: two iron absorption edges, one was fixed at 9.278~keV, which corresponds
713: to Fe XXVI and is expected to appear given the likely range of
714: temperatures and ionization.
715:
716: Fitting this model to the spectra for epochs 3 and 4, we obtained
717: better results than fitting with a power-law model (plus an absorbed
718: multicolor disk), i.e. ${\chi}^{2}_{\nu}=9.99$, with ${\nu}=21$, and
719: ${\chi}^{2}_{\nu}=2.73$, with ${\nu}=22$, for the third and fourth
720: epoch, respectively. Although the fit was unacceptable for the third
721: epoch, because of the large residuals at around ${\approx}10-20$\,keV,
722: which could be due to reflection \footnote{It is important to notice
723: that \citet{diaz06} did not find reflection signatures in the analysis
724: of joint XMM and INTEGRAL data corresponding to our fourth epoch. The
725: disagreement could be due to the $20-30$\,keV removed bin in their
726: study, made in order to improve the statistics for the joint spectrum
727: obtained in this period.}, the fit was reasonable for fourth
728: epoch. The multicolor disk component gave an inner disk temperatures
729: of $kT_{in}=1.27{\pm}0.17$\,keV for the fourth epoch, in which we also
730: obtained a very soft photon index (${\Gamma}= 4.7{\pm}0.6$).
731:
732: We substituted the $power-law$ component by the $pexriv$ reflection
733: model in the third epoch. This component is a power-law with an
734: exponential cut-off in order to take into account reflection effects
735: in the spectrum. In fitting with the pexriv model, we imposed an
736: overabundance of $Fe$ of 2.8 with respect to the solar value and
737: $cos(i)=0.45$, the last implying an inclination of $63^{\circ}$ for
738: the inclination of the reflection medium, namely the disk (as found by
739: \citet{diaz06} based on XMM and INTEGRAL data for the same period of
740: observations). The temperature of the disk that gave the smallest
741: residuals was $T_{disk}=1.2{\times}10^{+07}$\,K so we fixed this value
742: in our fits. This value is consistent with having very ionized
743: material in the accretion disk. Also, in order to properly fit the high energy part of
744: the spectra (${\gtrsim}20$\,keV) it was necessary to fix the $pexriv$
745: e-folding energy to a very high value, namely $1000$\,keV, implying a
746: non-detection of any cut-off up to ${\approx}200$\,keV.
747:
748: As a result of this fit for third epoch, we obtained a reflection
749: covering factor of $R{\leq}0.3$. Actually, this value ($R$)
750: approximately measures the solid angle subtended by the reflecting
751: medium as seen from the source of the primary radiation,
752: $R{\approx}{\Omega}_{refl}/2{\pi}$, so that $R=1$ for an isotropic
753: point source above an infinite optically thick slab. We fixed the
754: ionization parameter to a very high value, consistent with a very high
755: ionization medium, $x_{i}=4{\pi}F_{ion}/n$=$5000$\,erg${\times}$cm$/$s
756: (where $F_{ion}$ is the 5-20\,keV irradiating flux and $n$ is the
757: density of the reflector; see \citet{done01}).
758:
759: In order to estimate the source luminosity in each of our
760: observations, we used the power-law plus an absorbed multicolor disk
761: model for all the epochs. The contribution of the iron emission and
762: absorption effects together contribute only by $<10\%$ of our data in
763: the last two epochs. The un-absorbed flux in the range of $5-100$\,keV
764: energy range is $F(5-100)_{X}=1.4{\times}10^{-9}, 1.2{\times}10^{-8},
765: 1.0{\times}10^{-8}, 1.0{\times}10^{-8}$ erg\,cm$^{-2}$ s$^{-1}$ for
766: the first to fourth epoch, respectively. For a distance of 3.2 kpc
767: \citep{tingay95}, these fluxes give luminosities in the $(5-100)$\,keV
768: energy range of $L(5-100)_{X}=1.7{\times}10^{36}, 1.5{\times}10^{37},
769: 1.2{\times}10^{37}, 1.2{\times}10^{37}$ erg$/$s. The un-absorbed
770: fluxes in the 5-600 \,keV energy range are
771: $F(5-500)_{X}=3.7{\times}10^{-9}, 1.2{\times}10^{-8}$,
772: $1.0{\times}10^{-8},1.0{\times}10^{-8}$ erg\,s$^{-1}$ cm$^{-2}$,
773: giving luminosities of $L(5-600)_{X}=4.6{\times}10^{36},
774: 1.5{\times}10^{37}$, $1.2{\times}10^{37}, 1.2{\times}10^{37}$ erg$/$s
775: for the first to fourth epochs, respectively. Even if we take the
776: broader energy range as more indicative of the true source luminosity,
777: these values represent only 0.5\%, 1.5\%, 1.3\%, and 1.3\% of the
778: Eddington limit ($L_{\rm EDD}=9.4{\times}10^{38}$ erg$/$s for a black
779: hole of 7 $M_{\odot}$), respectively.
780:
781:
782:
783: \section{Discussion and Conclusions} \label{discuss}
784:
785:
786: We have made fits to four epochs of {\rm INTEGRAL} broad-band spectra
787: of the stellar-mass black hole GRO J1655$-$40 during its 2005
788: outburst. We find that GRO~J1655$-$40 was in the hard state in the
789: first epoch, based on the low photon index (i.e. $1.72{\pm}0.03$) and
790: the absence of a strong thermal disk component. The source evolved to
791: a state that resembles a thermal dominant state in the classification
792: scheme of \citet{remillard06}. These statements are in agreement with
793: the previous study of this outburst by \citet{brocksopp06}. However,
794: \citet{saito07}, on the basis of their analysis of RXTE/PCA data of
795: this source, showed that the luminosity of the accretion disk deviates
796: from the the $L{\propto}T^{4}$ relation typical of thermal-dominant
797: states during our other 3 observations (epochs 2, 3 and 4). So, the
798: state observed may not be a true thermal-dominant state, and/or
799: non-thermal effects may need to be modeled to describe the accretion
800: disk emission \citep{kubota01}.
801:
802: For the two latest epochs, we found that our data is best fitted by
803: adding an iron emission line and edges in the model consisting of an
804: absorbed multicolor disk \citep{mitsuda84} plus a power-law component.
805: Although the obtained fits were not formally acceptable
806: (${\chi}^{2}_{red}=9.99, 2.73$ for the third to fourth epochs,
807: respectively), the fit in the fourth epoch was improved by
808: ${\Delta} {\chi}^{2}_{red}=4.8$. Also, the shape of the residuals
809: using the model consisting of a
810: multicolor disk black body plus a power-law showed a clear excess in
811: the $6-8$\,keV range and a drop at around 10\,keV, features that can
812: only be explained by the presence of iron emission line and the
813: presence of $Fe$ edges. The spectrum in the third epoch showed an
814: excess of absorption in the $10-20$\,keV energy range with
815: respect of that expected taking into account the edges obtained in
816: \citet{diaz06} based on XMM and INTEGRAL spectra (see Section \ref{edges}).
817: This excess could be explained by the
818: presence of reflection, but the situation is still unclear. We
819: attempted to deal with this feature in third epoch by fitting with a
820: broken power-law model ($pexriv$ in XSPEC) but there is no clear
821: evidence for a break in this spectrum. We conclude that for
822: GRO~J1655$-$40, it is difficult to study reflection features. This may
823: be due to the high inclination of the source, which could alter the
824: shape of the reflection features through scattering. The presence of
825: these features in the spectra of epochs 3 and 4 may reveal differences
826: in the disk outflow properties with respect to any outflow in epochs 1
827: and 2. In fact, our spectrum obtained in epoch 3 precedes a {\it
828: Chandra} observation revealing a line-rich spectrum\citep{miller06}~
829: by only 3 days.
830:
831: Assuming a value of $63^{\circ}$ \citep{diaz06} for the inclination and
832: a distance of 3.2\,kpc \citep{tingay95}, the disk
833: normalization factor (in fits made only with a power-law model plus a
834: multicolor absorbed disk) gives inner disk radii \footnote{This radii
835: are measured from infinity, using the formula: \\
836: $R_{in}=D{\times}(diskbb_{norm}/cos(i))^{0.5}$\\ Where $R_{in}$ is in
837: km and D is in units of 10\,kpc. This formula is inferred without
838: taking into account gravitational effects from the General Relativity
839: Theory. If gravitational corrections would be taken into account, then
840: smaller co-moving radius would be obtained.} of $16.5{\pm}0.2$\,km for
841: the second epoch (the first epoch was in a hard state and had no
842: significant contribution from the disk, as shown in Section
843: \ref{powerlaw}). Since $R_{g}=10.4$\,km is the value of the
844: gravitational radius for a $7M_{\odot}$ black hole, we find that the
845: matter arrives at inner radii of ${\approx}1.6R_{g}$ in the second
846: epoch. This value is consistent with the value predicted for a
847: maximally rotating black hole as explained in Section \ref{introd}
848: ($R_{ISCO}=1.25~R_{g}$) and is similar to that obtained by
849: \citet{tomsick99}. Analyzing {\it RXTE} data, \citet{tomsick99} found
850: inner radii values of
851: $R_{in}=10.9^{+2.6}_{-2.6}-21.9^{+5.2}_{-5.2}$\,km, depending on the
852: value adopted for the disk inclination. Of course, values of the inner
853: disk radius inferred from any continuum fits are suspect, and must be
854: viewed with caution. A number of effects \citep{merloni00} can serve
855: to distort the observed inner disk parameters (see \citet{saito07}
856: for a determination of more realistic values for the radii, since
857: these are not affected by a strong power-law component). According to
858: \citet{merloni00}, the dominant effect seems to be that the opacity is
859: dominated by electron scattering rather than free-free absorption.
860: The net result is that the derived temperature given by the ${\rm
861: kT}_{in}$ parameter overestimates the effective inner temperature by a
862: factor of 1.7 or more \citep{shimura95}.
863:
864: The most interesting issue regarding our INTEGRAL observations of
865: GRO~J1655$-$40 is undoubtely the presence of very significant and
866: unbroken high energy emission up to $500$\,keV in the hard state
867: of GRO~J1655-40, as noticed from first
868: epoch. \citet{grove98b} made a comparative measurement of a number of
869: systems with CGRO/OSSE and found that these systems showed a cut-off
870: at high energies at around ${\sim}100$\,keV in the hard state
871: of many systems \footnote{CGRO/OSSE
872: integration times were very long (of order of weeks), so different
873: states would be mixed. This is not an issue for INTEGRAL, since
874: exposures for each obtained spectrum are around two days and it is not
875: hoped to have noticeable high energy evolution with this timing, as
876: infrerred from the light curves.}. Since then several studies have
877: tried to measure the cut-off in the spectra that could manifest the
878: validity of Comptonization processes by thermal electrons playing
879: the major role in the high energy emission of black holes in the
880: hard state. In our work we did not find such a break and the
881: presence of a non-thermal population of relativistic electrons was infrerred
882: from the fitting of our spectrum in the hard state.
883:
884: Our finding is that non-thermal processes are the most important in order to
885: explain the high energy emission of the hard state of
886: GRO~J1655-40. Additionally, there is not any break in the data indicating
887: that the high energy emission is mainly produced by thermal
888: Comptonization as previously claimed. Moreover, non-thermal Comptonization is the
889: main source of the high energy emission in the second epoch
890: (thermal dominant state). This condition
891: for the thermal dominant state has been broadly
892: reported in the literature since \citet{grove98b}.
893:
894: In Table \ref{comp_previous} we summarize a list of references
895: showing breaks in the high energy emission of several black hole
896: candidates in the hard and the intermediate state. One
897: contemporaneous study by \citet{shaposhnikov06}
898: in the hard state of GRO~J1655-40 used
899: INTEGRAL observations covering a period of time slightly prior to that
900: of our first epoch. They pointed out the presence of a cut-off in
901: their data at around 200\,keV in their ISGRI and SPI spectra,
902: as a manifestation of the role of thermal Comptonization being the
903: main source of the high energy emission. However, their sensitivity at
904: energies ${\gtrsim}200$\,keV is not high enough in order to
905: disentangle the energy emission of a non-thermal population of
906: electrons. As also noticed by \citet{coppi99}, if a spectrum extends up
907: to $500$\,keV high energy emission coming from a non-thermal
908: population of electrons is clearly present. The finding of
909: the contribution of the non-thermal population of electrons in the
910: high energy emission was reported before for the Cyg X-1 system
911: (\citet{malzac06}, \citet{cadolle06} both in the intermediate state)
912: and may be for the GRS~1915+105 (\citet{zdziarski01},
913: \citet{rodriguez04} both in the thermal dominant state). With our study, we
914: extend the list of sources showing high energy emission coming from
915: non-thermal electrons with GRO~J1655-40, this time in the hard
916: state. \citet{joinet06} also reported the detection of a non-thermal
917: population of electrons in the hard state of GX 339-4. However,
918: they indicated also the presence of a cut-off in the spectra.
919:
920: In Table \ref{comp_previous} it can be seen that while some cut-offs
921: appear to be close to the upper boundaries of the high energy instruments
922: used, others appear to be physically meaningful. The last correspond to
923: the systems GX 339-4 and Cyg X-1, both systems with known low inclination
924: angle of ${\rm i}{\approx}45^{\circ}$ and ${\rm i}{\approx}35^{\circ}$,
925: respectively. GRO~J1655-40 is a very high inclination system with an inner
926: disk that could have an inclination of ${\rm i}=85^{\circ}$ (see Section
927: \ref{introd}). Comptonization processes of soft photons through thermal
928: and/or non-thermal electrons seems to be highly isotropical from an
929: observer external to the system, if the electrons do not acquire outflow
930: velocities $v/c{\gtrsim}0.2$, \citep{beloborodov98}, so in principle it would
931: not depend on the inclination angle of the system. But other processes,
932: like reflection, could depend on the view angle, being almost unobserved for
933: high inclination systems like GRO~J1655 -40 (as pointed out above with our fitting of epoch 3).
934: So, GRO~J1655-40 in the hard state would be an excellent source for the studying of
935: Comptonization processes, without the apparent disturbance of any other
936: high energy mechanisms (including soft emission from a disk). If these hypothesis
937: are correct, we would have discovered that Comptonization of black
938: hole transients in the hard
939: state occurs by mainly Comptonization through non-thermal relativistic electrons.
940: This conclusion is also supported by the studies of \citet{malzac06} of
941: Cyg X-1 in the intermediate state because while showing a high energy cut-off at
942: the energy around 100\,keV, they also point out to the presence of high
943: energy emission from a non-thermal population of electrons in an state close
944: to the hard (in the case of \citet{joinet06} in the hard state of
945: GX 339-4).
946:
947:
948: \acknowledgments
949:
950: We thank Danny Steeghs for helpful discussions. We also thank the
951: RXTE instrument teams at MIT and NASA/GSFC for providing the ASM light
952: curve. MDCG is a MEC funded PhD student supported under grants
953: PNE2003-04352+ESP2005-07714-C03-03. This work is based on
954: observations made with INTEGRAL, an ESA science mission with
955: instruments and science data centre funded by ESA member states and
956: with the participation of Russia and the USA. We thank to K. Arnaud
957: for providing part of the eqpair code used in this work. We finally thank the
958: anonymous referee for the careful reading of this manuscript.
959:
960:
961: \begin{thebibliography}{}
962:
963: \bibitem[Arnaud (1996)]{arnaud96}
964: Arnaud, K. A., 1996, ADASS, 5, 17A
965: \bibitem[Bailyn et~al.(1995)]{bailyn95a}
966: Bailyn, C. D., Orosz, J. A., Girard, T. M., Shardha, J., della Valle, M. et~al., 1995, Nature, 374, 701
967: \bibitem[Bailyn et~al.(1995b)]{bailyn95b}
968: Bailyn, C. D., Orosz, J. A., McClintock, J. E. \& Remillard, R. A., 1995, Nature, 378, 157
969: \bibitem[Bardeen et~al.(1972)]{bardeen72}
970: Bardeen, J.~M., Press, W.~H. \& Teukolsky, S.~A., 1972, ApJ, 178 ,347B
971: \bibitem[Beloborodov (1998)]{beloborodov98}
972: Beloborodov, A., M., 1998, ApJL, 510, 123
973: \bibitem[Borozdin et~al.(1999)]{borozdin99}
974: Borozdin, K., Revnivtsev, M., Trudolyubov, S., Shrader, C. \& Titarchuk, L., 1999, ApJ, 517, 367
975: %\bibitem[Brocksopp et~al.(2002)]{brocksopp02}
976: %Brocksopp, C., Fender, R. P. \& Pooley, G. G., 2002, MNRAS, 336, 699
977: \bibitem[Brocksopp et~al.(2006)]{brocksopp06}
978: Brocksopp, C. , McGowan, K. E., Krimm, H., Godet, O., Roming, P., et~al. 2006, MNRAS, 365, 1203B
979: \bibitem[Buxton et~al.(2005)]{buxton05b}
980: Michelle Buxton, Charles Bailyn, Dipankar Maitra, 2005, ATel 418
981: \bibitem[Buxton \& Bailyn (2005)]{buxton05}
982: Buxton, M. \& Bailyn, C., 2005, ATel 485
983: %\bibitem[Cadolle et~al.(2004)]{cadolle04}
984: %Cadolle Bel, M., Rodr\'{\i}guez, J., Sizun, P., Farinelli, R., et~al., 2004, A\&A, 426, 659
985: \bibitem[Cadolle et~al.(2006)]{cadolle06}
986: Cadolle Bel, M., Sizun, P., Goldwurm, A., Rodriguez, J., Laurent, P., Zdziarski, A.~A., 2006, A\&A, 446, 591C
987: %\bibitem[Cadolle et~al.(2006)]{cadolle06b}
988: %Cadolle Bel, M., Ribo, M., Rodriguez, J., Chaty, S., and Corbel, S., 2006, ApJ, astro-ph/0612575
989: \bibitem[Coppi (1999)]{coppi99}
990: Coppi, P. S., 2000, Bulletin of the American Astronomical Society, 32, 1217
991: \bibitem[Chernyakova (2005)]{jemx05}
992: Chernyakova, M., 2005, ISDC/OSA-UM-IBIS
993: \bibitem[Chernyakova et~al.(2005)]{ibis05}
994: Chernyakova, M., Kretschmar, P., JEM-X team, Neronov, A., 2005, ISDC/OSA-UM-JEMX
995: \bibitem[Courvoisier et~al.(2003)]{courvoisier03}
996: Courvoisier, T. J.-L., Walter, R., Beckmann, V., et~al., 2003, A\&A, 411, L53
997: \bibitem[Deluit (2005)]{deluit05}
998: Deluit, S., (2005) SPI-NS-0-4307-CESR
999: \bibitem[D\'{\i}az Trigo et~al.(2006)]{diaz06}
1000: D\'{\i}az Trigo, M., Parmar, A., N., Miller, J. M., Kuulkers, E., Caballero-Garc\'{\i}a, 2006, astro-ph/0610873
1001: \bibitem[Domingo et~al.(2003)]{domingo03}
1002: Domingo, A., Caballero, M.~D., Figueras, F., Jordi, C., Torra, J., Mas-Hesse, J.~M., Gim{\'e}nez, A., Hudcova, V. \& Hudec, R., 2003, A\&A, 411, L281
1003: \bibitem[Done \& Nayakshin (2001)]{done01}
1004: Done, C., \& Nayakshin, S., 2001, MNRAS 328, 616
1005: \bibitem[Dubath et~al.(2005)]{dubath05}
1006: Dubath, P., Kn{\"o}dlseder, J., Skinner, G.~K., Connell, P., Kreykenbohm, I., Strong, A., Sizun, P., Atti{\'e}, D., Schanne, S., Cordier, B., Bouchet, L., \& von Kienlin, A., 2005, MNRAS, 357, 420
1007: \bibitem[Ebisawa et~al.(1996)]{ebisawa96}
1008: Ebisawa, K., Titarchuk, L. \& Chakrabarti, S. K., 1996, PASJ, 48, 59
1009: \bibitem[Esin et~al.(2001)]{esin01}
1010: Esin, A. A., McClintock, J. E., Drake, J. J. et~al., 2001, ApJ, 555, 483
1011: \bibitem[Fender et~al.(2004)]{fender04}
1012: Fender, R. P. and Belloni, T. M. \& Gallo, E., 2004, MNRAS, 355, 1105
1013: \bibitem[Foellmi et~al.(2006)]{foellmi06}
1014: Foellmi, C., Depagne, E., Dall, T. H. \& Mirabel, I. F., 2006, astro-ph/0606269
1015: \bibitem[Frontera et~al.(2003)]{frontera03}
1016: Frontera, F., Amati, L., Zdziarski, A. A. et~al., 2003, ApJ, 592, 1110
1017: \bibitem[George \& Fabian (1991)]{george91}
1018: George, I. M., Fabian, A. C., 1991, MNRAS, 249, 352
1019: \bibitem[Ghisellini et~al.(1993)]{ghisellini93}
1020: Ghisellini, G., Haardt, F. and Fabian, A.~C., 1993, MNRAS, 263, L9
1021: \bibitem[Gierli\'nski et~al.(1999)]{gierlinski99}
1022: Gierli\'nski, M., Zdziarski, A., Poutanen, J. et~al., 1999, MNRAS, 309, 496
1023: \bibitem[Gilfanov et~al.(2000)]{gilfanov00}
1024: Gilfanov, M., Churazov, E. \& Revnivtsev, M., 2000, Proceedings of 5-th Sino-German workshop on Astrohpysics, 1999, Eds. Gang Zhao, Jun-Jie Wang, Hong Mei Qiu and Gerhard Boerner, SGSC Conference Series, vol.1, pp.114-123, astro-ph/0002415
1025: \bibitem[Goldwurm et~al.(2003)]{goldwurm03}
1026: Goldwurm, A., David, P. \& Foschini et~al., 2003, A\&A, 411L, 223G
1027: \bibitem[Greene et~al.(2001)]{greene01}
1028: Greene, Bailyn \& Orosz 2001, ApJ, 554, 1290
1029: \bibitem[Grenier (1996)]{grenier96}
1030: Grenier, J., 1996, A\&A, 120, 239
1031: \bibitem[Grenier et~al.(1995)]{grenier95}
1032: Grenier, J., Predehl, P. \& Pohl, M., 1995, A\&A, 297, L67
1033: \bibitem[Grove et~al.(1998)]{grove98a}
1034: Grove, J.~E., Strickman, M.~S., Matz, S.~M., Hua, X.-M., Kazanas, D. \& Titarchuk, L., 1998, ApJ, 502L, 45G
1035: \bibitem[Grove et~al.(1998)]{grove98b}
1036: Grove, J. E., Johnson, W. N., Kroeger, R. A. et~al., 1998, ApJ, 500, 899
1037: \bibitem[Hameury et~al.(1997)]{hameury97}
1038: Hameury, J.-M., Lasota, J.-P., McClintock, J.~E. \& Narayan, R., 1997, ApJ, 489, 234
1039: \bibitem[Hjellming \& Rupen (1995)]{hjellming95}
1040: Hjellming, R. M., \& Rupen, M. P., 1995, Nature, 375, 464
1041: \bibitem[Homan \& Belloni (2005)]{homan05}
1042: Homan, J. \& Belloni, T. M., 2005, in From X-ray Binaries to Quasars: Black Hole Accretion on All Mass Scales, ed. T. J. Maccarone, R. P., Fender, L. C., Ho (Dordrecht: Kluwer) (astro-ph/0412597)
1043: \bibitem[Homan (2005)]{homan05b}
1044: Homan, J., 2005, ATel 440
1045: \bibitem[Homan et~al.(2007)]{homan07}
1046: Homan et~al., 2007, (in prep.)
1047: \bibitem[Ibragimov et~al.(2005)]{ibragimov05}
1048: Ibragimov, A. Poutanen, J. Gilfanov, M., et~al., 2005, MNRAS, 362, 1435I
1049: \bibitem[Joinet et~al.(2006)]{joinet06}
1050: Joinet, A. Jourdain, E., Malzac, J., Roques, J.~P., et~al., 2006, astro-ph/0611064
1051: \bibitem[Kubota et~al.(2001)]{kubota01}
1052: Kubota, A., Makishima, K., \& Ebisawa, K., 2001, ApJ, 560, L147
1053: \bibitem[Kuulkers et~al.(1998)]{kuulkers98}
1054: Kuulkers, E., Wijnands, R., Belloni, T., Mendez, M., van der Klis, M. \& van Paradijs, J., 1998, ApJ, 794, 753
1055: \bibitem[Kuulkers et~al.(2007)]{kuulkers07}
1056: Kuulkers, E., Shaw, S., E., Paizis, A., Chevenez, J. et~al., 2007, astro-ph/0701244
1057: \bibitem[Laor (1991)]{laor91}
1058: Laor, A., 1991, ApJ, 376, 90
1059: \bibitem[Lasota (2001)]{lasota01}
1060: Lasota, J.-P., 2001, New Astr Rev, 45, 449
1061: \bibitem[Lebrun et~al.(2003)]{lebrun03}
1062: Lebrun, F., Leray, J. P., Lavocat, P., et~al., 2003, A\&A, 411, L141
1063: \bibitem[Lubi\'nski et~al.(2005)]{lubinski05}
1064: Lubi\'nski, P., Dubath, P., Paltani, S. et~al., 2005, INTEGRAL cross-calibration status for OSA 5.1, ISDC/CCR 1.0
1065: \bibitem[Lund et~al.(2003)]{lund03}
1066: Lund, N., Budtz-Jorgensen, C., Westergaard, N. J., et~al., 2003, A\&A, 411, L231
1067: \bibitem[Malzac et~al(2006)]{malzac06}
1068: Malzac, J., Petrucci, P.O. Jourdain, E., Cadolle Bel, M., Sizun, P., Pooley, G. et~al., 2006, A\&A, 448, 1125M
1069: \bibitem[McConnell et~al. (2000)]{mcconnell00}
1070: McConnell, M.~L. Ryan, J.~M. Collmar, W. et~al., 2000, ApJ, 543, 928M
1071: \bibitem[McConnell et~al. (2002)]{mcconnell02}
1072: McConnell, M.~L. Zdziarski, A.~A. Bennett, K. et~al., 2002, ApJ, 572, 984M
1073: \bibitem[Miller et~al(2006)]{miller06}
1074: Miller,~J,~M., Raymond,~J, Fabian,~A., Steeghs,~D., et~al., 2006, Nat, 441, 953
1075: \bibitem[Magdziarz \& Zdziarski (1995)]{magdziarz95}
1076: Magdziarz, P. \& Zdziarski, A. A., 1995, MNRAS, 273, 837
1077: \bibitem[Markwardt \& Swank (2005)]{markwardt05}
1078: Markwardt, C. \& Swank, J., (2005), ATel 414
1079: \bibitem[Mas-Hesse et~al.(2003)]{mas-hesse03}
1080: Mas-Hesse, J. M., Gim\'enez, A., Culhane, J. L., et~al., 2003, A\&A, 411, L261-L268
1081: \bibitem[Markoff et~al.(2001)]{markoff01}
1082: Markoff, S., Falcke, H. \& Fender, R., 2001, A\&A, 372, 25
1083: \bibitem[Markoff et~al.(2003)]{markoff03}
1084: Markoff, S., Nowak, M., Corbel, S., Fender, R. \& and Falcke, H., 2003, A\&A, 397, 645
1085: \bibitem[Markoff et~al.(2005)]{markoff05}
1086: Markoff, S., Nowak, M.~A. Wilms, J., 2005, ApJ, 635, 1203
1087: \bibitem[Merloni et~al.(2000)]{merloni00}
1088: Merloni, A., Fabian, A.~C. \& and Ross, R.~R., 2000, MNRAS, 313, 193
1089: \bibitem[Miller et~al.(2005)]{miller05}
1090: Miller, J. M., Fabian, A. C., Nowak, M. A., \& Lewin, W. H. G., 2005, in the Proceedings of the Tenth Marcel Grossman Meeting, Rio de Janiero, Brazil, eds. M. novello, S. Perez Bergliaffa, \& R. Ruffini, World Scientific Publishing, Singapore
1091: \bibitem[Mitsuda et~al.(1984)]{mitsuda84}
1092: Mitsuda, K., Inoue, H., Koyama, K., et~al., 1984, PASJ 36, 741-759
1093: \bibitem[Narayan (1996)]{narayan96}
1094: Narayan, R. (1996), ApJ 462, 136-141
1095: \bibitem[Nowak (1995)]{nowak95}
1096: Nowak, M., A., 1995, PASP, 107, 1207
1097: %\bibitem[Natalucci et~al.(2005)]{osaisgri}
1098: %IBIS team/IASF Rome, 2005, 2005, ISGRI/OSA 5, Spectral analysis and Response Matrix Status (version 1.0)
1099: \bibitem[Orosz \& Bailyn (1997)]{orosz97a}
1100: Orosz, J. A., Bailyn, C. D., 1997, ApJ, 477, 876
1101: \bibitem[Orosz et~al.(1997)]{orosz97b}
1102: Orosz, J. A., Remillard, R. A., Bailyn, C. D., \& McClintock, J. E., 1997, ApJ, 478, L83
1103: %\bibitem[Poutanen \& Svensson (1996)]{poutanen96}
1104: %Poutanen, J., \& Svensson, R., 1996, ApJ, 470, 249
1105: \bibitem[Pottschmidt et~al.(2003)]{pottschmidt03}
1106: Pottschmidt, K. Wilms, J. Chernyakova, M. Nowak, M.~A. \& Rodriguez, J., 2003, A\&A, 411, 383
1107: \bibitem[Remillard \& McClintock (2006)]{remillard06}
1108: Remillard, R. A. \& McClintock, J. E, 2006, \araa, 44, 49
1109: \bibitem[Rodriguez et~al.(2004)]{rodriguez04}
1110: Rodriguez, J., Fuchs, Y., Hannikainen, D., Vilhu, O., Shaw, S., 2004, ESA SP-552, 377
1111: \bibitem[Roming et~al.(2005)]{roming05}
1112: Roming, P. W. A. et~al., 2005, Space Reviews in press (astro-ph/0507413)
1113: \bibitem[Roques \& Jourdain (2005)]{roques05}
1114: Roques, J-P., Jourdain, E., 2005, SPI-NS-0-4305-CESR
1115: \bibitem[Roques et~al.(2005b)]{osaspi}
1116: Roques, J-P., Jourdain, E. \& Shrader, C., 2005, SPI-NS-0-4308-CESR
1117: \bibitem[Rupen et~al.(2005)]{rupen05}
1118: M.P. Rupen (NRAO/GSFC), V. Dhawan, A.J. Mioduszewski (NRAO), 2005, ATel 419
1119: \bibitem[Rupen et~al.(2005)]{rupen05b}
1120: M.P. Rupen (NRAO/GSFC), V. Dhawan, A.J. Mioduszewski (NRAO), 2005, ATel 489
1121: \bibitem[Saito et~al.(2007)]{saito07}
1122: Saito, K., Homan, J., Yamaoka, K., Fukuyama, M., 2007, astro-ph/0702022
1123: \bibitem[Shakura \& Sunyaev (1973)]{shakura73}
1124: Shakura, N. I. and Sunyaev, R. A., 1973, A\&A 24, 337-366
1125: \bibitem[Shapiro \& Teukolsky (1983)]{shapiro83}
1126: Shapiro, S. L. and Teukolsky, S. A., 1983, Black Holes, White Dwarfs and Neutron Stars: The Physics of Compact Objects (Wiley, New York)
1127: \bibitem[Shaposhnikov et~al.(2006)]{shaposhnikov06}
1128: Shaposhnikov, N., Swank, J. H., shrader, C. R. et~al., 2006, astro-ph/0609757
1129: \bibitem[Shimura \& Takahara (1995)]{shimura95}
1130: Shimura, T. \& Takahara, F., 1995, ApJ, 445, 780
1131: \bibitem[Skinner \& Connell (2003)]{skinner03}
1132: Skinner, G. \& Connell, P., 2003, A\&A, 411, L123
1133: \bibitem[Sobczak et~al.(1999)]{sobczak99}
1134: Sobczak, G., J., McClintock, J., E., Remillard, R. A. et~al., 1999, ApJ, 520, 776
1135: \bibitem[Strohmayer (2001)]{strohmayer01}
1136: Strohmayer, T.~E., 2001, ApJ, 552, L49
1137: %\bibitem[Sunyaev \& Tr\''umper (1979)]{sunyaev79}
1138: %Sunyaev, R.~A., \& Tr\''umper, J., 1979, Nature, 279,506
1139: \bibitem[Thorne (1974)]{thorne74}
1140: Thorne, K.~S., 1974, ApJ, 191, 507T
1141: \bibitem[Tingay et~al.(1995)]{tingay95}
1142: Tingay, S. J., et~al., 1995, Nature, 374, 141
1143: \bibitem[Titarchuk (1994)]{titarchuk94}
1144: Titarchuk, L., 1994, ApJ, 434, 570
1145: \bibitem[Titarchuk \& Shrader (2002)]{titarchuk02}
1146: Titarchuk, L. \& Shrader, C. R., 2002, ApJ, ApJ, 567, 1057
1147: \bibitem[Tomsick et~al.(1999)]{tomsick99}
1148: Tomsick, J.~A., Kaaret, P., Kroeger, R.~A. \& Remillard, R.~A., 1999, ApJ, 512, 892
1149: \bibitem[Torres et~al.(2005)]{torres05}
1150: M.A.P. Torres (CfA), D. Steeghs (CfA), P. Jonker (SRON), P. Martini (CfA), 2005, ATel 417
1151: \bibitem[Ueda et~al.(1998)]{ueda98}
1152: Ueda, Y., Inoue, H., Tanaka, Y., et~al., 1998, ApJ, 492, 782
1153: \bibitem[Ubertini et~al.(2005)]{ubertini05}
1154: Ubertini, P., Bazzano, A., Capitanio, F., De Cesare, G., 2005, ScienceDirect, 38, 1369
1155: \bibitem[van der Hooft et~al.(1998)]{hooft98}
1156: van der Hooft, F., Heemskerk, M. H. M., Alberts, F., \& van Paradijs, J., 1998, A\&A, 329, 538
1157: \bibitem[van der Hooft et~al.(1999)]{vanderhooft99}
1158: van der Hooft, F., Kouveliotou, C., van Paradijs, J., Paciesas, W.~S., Lewin, W.~H.~G., van der Klis, M., et~al., 1999, ApJ, 513, 477V
1159: \bibitem[Vedrenne et~al.(2003)]{vedrenne03}
1160: Vedrenne, G., Roques, J. P., Sch\"onfelder, V., 2003, A\&A, 411, L63
1161: \bibitem[Westergaard et~al.(2003)]{westergaard03}
1162: Westergaard, N.~J., Kretschmar, P. \& Oxborrow et~al., 2003, 411L, 257W
1163: %\bibitem[Westergaard et~al.(2005)]{osajemx}
1164: %Westergaard, N. J., Oxborrow, C. A., Chebenez, J., Lund, N., Brandt, S., et~al., 2005, JEM-X Scientific Analysis: Scientific Validation Report (version 5.0)
1165: \bibitem[Yuan \& Zdziarski (2004)]{yuan04}
1166: Yuan, F. \& Zdziarski, A. A., 2004, MNRAS, 354, 953
1167: \bibitem[Zdziarski et~al.(2001)]{zdziarski01}
1168: Zdziarski, A.~A., Grove, J.~E., Poutanen, J., Rao, A.~R., 2001, ApJ, 554, L45
1169: \bibitem[Zdziarski et~al.(2005)]{zdziarski05}
1170: Zdziarski, A.~A., Gierli{\'n}ski, M., Rao, A.~R., Vadawale, S.~V. \& Miko{\l}ajewska, J., 2005, MNRAS, 360, 825
1171: \bibitem[Zdziarski et~al.(2004)]{zdziarski04}
1172: Zdziarski, A. A., Gierli{\'n}ski, M., et~al. 2004, MNRAS, 351, 791
1173: \bibitem[Zhang et~al.(1994)]{zhang94}
1174: Zhang, S. N., Wilson, C. A., Harmon, B. A., Fishman, G. J., Wilson, R. B., Paciesas, W. S., Scott, M., \& Rubin, B. C., 1994, IAU Circ. 6209
1175:
1176:
1177: \end{thebibliography}
1178:
1179: %\clearpage
1180:
1181: %_____________________________________________________________________________________________________________________________
1182:
1183: %________________________TABLES_______________________________________________________________________________________________
1184:
1185: \clearpage
1186:
1187: \begin{deluxetable}{cccccc}
1188: %\tabletypesize{\scriptsize}
1189: \rotate
1190: \tablecaption{INTEGRAL observing ToO during which GRO J1655-40 was observed in the 2005 outburst, giving the exposure times of the summed spectra analyzed for each ToO and instrument.\label{tobserv}}
1191: \tablewidth{0pt}
1192: \tablehead{\colhead{Epoch number} & \colhead{INTEGRAL revolution \& MJD} & \colhead{Start \& End Date} & \colhead{JEM-X} & \colhead{ISGRI} & \colhead{SPI}
1193: }
1194: \startdata
1195: & (days) & (yyyy/mm/dd) & [s] & [s] & [s] \\
1196: & & & & & \\
1197: \hline
1198: 1 & 290 & 2005/02/27-28 & 44504.83 & 69212.87 & 90849.58 \\
1199: & 53428.20-53429.50 & & & & \\
1200: 2 & 295-296 & 2005/03/16-18 & 43641.93 & 71827.16 & 92584.05 \\
1201: & 53445.10-53447.80 & & & & \\
1202: 3 & 299 & 2005/03/26-28 & 44673.42 & 69848.25 & $-$ \\
1203: & 53455.80-53457.00 & & & & \\
1204: 4 & 304 & 2005/04/10-11 & 46037.85 & 67883.58 & $-$ \\
1205: & 53470.00-53471.33 & & & & \\
1206: \enddata
1207: \end{deluxetable}
1208:
1209:
1210: \begin{deluxetable}{ccccc}
1211: \tabletypesize{\scriptsize}
1212: \rotate
1213: \tablecaption{Parameters obtained for the best fits of the joint JEM-X, ISGRI and SPI spectra (see text for details), using the model $constant{\times}eqpair$
1214: in the first epoch, $constant{\times}phabs(diskbb+eqpair)$ in the second epoch,
1215: $constant{\times}phabs(diskbb + gaussian + pexriv)edge{\times}edge$ in the third epoch and $constant{\times}phabs(diskbb + gaussian + powerlaw)edge{\times}edge$
1216: in the fourth epoch.\label{param_spec}}
1217: \tablewidth{0pt}
1218: \tablehead{
1219: \colhead{} & \colhead{{\rm Epoch 1}} & \colhead{{\rm Epoch 2}} & \colhead{{\rm Epoch 3}} & \colhead{{\rm Epoch 4}}
1220: }
1221: \startdata
1222: %& & & & \\
1223: Parameter & & & & \\
1224: & & & & \\
1225: & & Powerlaw & & \\
1226: & & & & \\
1227: ${\Gamma}$ & $-$ & $-$ & $-$ & $4.7{\pm}0.6$ \\
1228: ${\rm N}_{\rm pow}$ [${\rm ph}/{\rm keV}/{\rm s}/{\rm cm}^{2}$] at 1 keV & $-$ & $-$ & $-$ & $580{\pm}60$ \\
1229: & & & & \\
1230: & & eqpair & & \\
1231: & & & & \\
1232: ${\Gamma}_{\rm p}$ & $0.6{\pm}0.3$ & ${\approx}0$ & $-$ & $-$ \\
1233: ${\gamma}_{\rm min}$ & $1.3$ (f) & $1.3$ (f) & $-$ & $-$ \\
1234: ${\gamma}_{\rm max}$ & $1000$ (f) & $1000$ (f) & $-$ & $-$ \\
1235: ${\rm l}_{s}$ & $10$ (f) & $10$ (f) & $-$ & $-$ \\
1236: ${\rm l}_{h}/{\rm l}_{s}$ & $0.8{\pm}0.3$ & $1.2{\pm}0.6$ & $-$ & $-$ \\
1237: ${\rm l}_{nth}/{\rm l}_{h}$ & $0.8{\pm}0.1$ & $0.2{\pm}0.3$ & $-$ & $-$ \\
1238: ${\tau}$ & $1{\pm}0.5$ & $4{\pm}1$ & $-$ & $-$ \\
1239: ${\rm Refl}$ $[{\Omega}/2{\pi}]$ & 0(f) & $0$ (f) & $-$ & $-$ \\
1240: ${\rm kT}_{\rm in}$\,(keV) & $0.5$(f) & $1.25{\pm}0.02$ & $-$ & $-$ \\
1241: ${\rm R}_{\rm in}$\,(keV) [${\rm GM/c}^{2}$] & $6$ (f) & $6$ (f) & $-$ & $-$ \\
1242: ${\rm R}_{\rm out}$\,(keV) [${\rm GM}/c^{2}$] & $100$ (f) & $100$ (f) & $-$ & $-$ \\
1243: & & & & \\
1244: & & diskbb & & \\
1245: & & & & \\
1246: ${\rm kT}_{\rm in}$ [keV] & $-$ & $1.25{\pm}0.02$ & $1.28{\pm}0.02$ & $1.27{\pm}0.17$ \\
1247: ${\rm N}_{\rm bb}$ [$({\rm R}_{\rm in}[{\rm km}]/{\rm D}[10{\rm kpc}])^{2}{\times}{\rm cos}{\theta}$] & $-$ & $1321^{+72}_{-67}$ & $710{\pm}40$ & $500{\pm}60$ \\
1248: & & & & \\
1249: & & gaussian & & \\
1250: ${\rm E}_{\rm gauss}$ [keV] & $-$ & $-$ & $6.7{\pm}0.3$ & $6.7{\pm}0.9$ \\
1251: ${\sigma}$ [keV] & $-$ & $-$ & $0.63{\pm}0.15$ & $0.8{\pm}0.5$ \\
1252: ${\rm N}_{\rm gauss}$ [${\rm ph}/{\rm cm}^{2}/{\rm s}$] & $-$ & $-$ & $0.017{\pm}0.008$ & $0.05{\pm}0.03$ \\
1253: & & edge (Fe XXV) & & \\
1254: ${\rm E}_{\rm edge}$ [keV] & $-$ & $-$ & $8.64{\pm}0.20$ & $8.6{\pm}0.9$ \\
1255: ${\tau}$ & $-$ & $-$ & $0.20{\pm}0.05$ & $0.20{\pm}0.10$ \\
1256: & & edge (Fe XXVI) & & \\
1257: ${\rm E}_{\rm edge}$ [keV] & $-$ & $-$ & 9.278 (f) & 9.278 (f) \\
1258: ${\tau}$ & $-$ & $-$ & ${\leq}0.02$ (f) & ${\leq}0.02$ (f) \\
1259: & & pexriv & & \\
1260: Photon index & $-$ & $-$ & $2.50{\pm}0.23$ & $-$ \\
1261: ${\rm E}_{\rm f}$ [keV] & $-$ & $-$ & 1000.0 (f)& $-$ \\
1262: ${\rm R}$ $[{\Omega}/2{\pi}]$ & $-$ & $-$ & ${\leq}0.30$ & $-$ \\
1263: ${\rm Fe}$ abundance & $-$ & $-$ & 2.8 (f) & $-$ \\
1264: ${\rm cos(i)}$ & $-$ & $-$ & 0.45 (f) & $-$ \\
1265: ${\rm T}_{\rm disk}$ [K] & $-$ & $-$ & 1.2E+07 (f) & $-$ \\
1266: ${\rm xi}$ [$4{\pi}{\rm F}_{\rm irr}/{\rm n}$] & $-$ & $-$ & 5000 (f) & $-$ \\
1267: ${\rm N}_{\rm pexriv}$ [${\rm ph}/{\rm keV}/{\rm cm}^{2}/{\rm s}$] & $-$ & $-$ & $3.8{\pm}1.5$ & $-$ \\
1268: & & & & \\
1269: & & & & \\
1270: & & & & \\
1271: & & & & \\
1272: & & & & \\
1273: & & Intrumental normalization factors & & \\
1274: & & & & \\
1275: ${\rm C}_{\rm JEM-X}$ & 1.0 (f) & 1.0 (f) & 1.0 (f) & 1.0 (f) \\
1276: ${\rm C}_{\rm ISGRI}$ & $1.1{\pm}0.1$ & $1.0{\pm}0.1$ & $0.46^{+0.11}_{-0.15}$ & $0.39{\pm}0.14$ \\
1277: ${\rm C}_{\rm SPI}$ & $1.4{\pm}0.1$ & $1.1{\pm}0.1$ & $-$ & $-$ \\
1278: & & & & \\
1279: %\hline
1280: % & & & & \\
1281: ${\chi}_{\nu}^{2}$ & 1.21 & 1.44 & 2.62 & 2.73 \\
1282: ${\nu}$ & 57 & 83 & 19 & 22 \\
1283: & & & & \\
1284: %\hline
1285: \enddata
1286: \tablecomments{Parameters fixed in the fits are denoted by 'f'. We fixed the value of the column density to $N_{H}=8.0{\times}10^{21}$ atoms$/cm^{2}$ as obtained by \citet{diaz06} using XMM-Newton data of GRO~J1655$-$40 during the same outburst.}
1287:
1288: \end{deluxetable}
1289:
1290: \clearpage
1291: \thispagestyle{empty}
1292: \setlength{\voffset}{20mm}
1293: \begin{deluxetable}{cccccccc}
1294: %\tabletypesize{\scriptsize}
1295: \rotate
1296: \tablecaption{Values for the components of Comptonization models in previous studies of several sources in several states close to the hard (LH means hard
1297: state and IS means intermediate state). Also, values for the high-energy cut-offs and/or
1298: break energies (if present) are reported.\label{comp_previous}}
1299: \tablewidth{0pt}
1300: \tablehead{\colhead{${\Gamma}$} & \colhead{${\rm State}$} & \colhead{$E_{cut-off}$(\,keV)} & \colhead{$E_{max}$(\,keV)} & $T_{e}$ (\,keV) & ${\tau}$ & \colhead{Source} & \colhead{Reference}
1301: }
1302: \startdata
1303: $1.67{\pm}0.06$ & LH & $195{\pm}50$ & ${\approx}200$ & $35^{+200}_{-9}$ & $1.45^{+0.5}_{-1.4}$ & IGR~J17497$-$2821 & Walter et al. 2007, A\&A, 461, L17 \\
1304: $1.4-1.6$ & LH & $50-200$ & ${\approx}200$ & $-$ & $-$ & GX~339$-$4 & Miyakawa et al. 2007, astro-ph/0702087 \\
1305: $1.70{\pm}0.01$ & LH & $115{\pm}5.6$ & & $-$ & $-$ & XTE~J1550$-$564 & Yuan et al. 2006, astro-ph/0608552 \\
1306: $1.53{\pm}01$ & LH & $460{\pm}300$ & & $-$ & $-$ & XTE~J1550$-$564 & Yuan et al. 2006, astro-ph/0608552 \\
1307: $1.92{\pm}0.05$ & IS & $72{\pm}8$ & ${\approx}200$ & $-$ & $-$ & GX~339$-$4 & Belloni et al. 2006, MNRAS, 367, 1113 \\
1308: $1.9{\pm}0.1$ & LH &${\approx}150$ & ${\approx}600$ & $67^{+8}_{-6}$ & $1.98{\pm}0.22$ & Cyg X-1 & Cadolle Bell et al. 2006, A\&A, 446, 591 \\
1309: $2.1$ & IS &${\approx}100$ & ${\approx}1000$ & $20-65$ & $0.55-1.36$ & Cyg X-1 & Malzac et al., 2006, A\&A, 448, 1125 \\
1310: $1.65-2.0$ & LS & $130-250$ & ${\approx}150$ & $-$ & $-$ & Cyg~X$-$1 & Wilms et al. 2006, A\&A, 447, 245 \\
1311: $1.35{\pm}0.03$ & LH & $100-200$ & ${\approx}600$ & & & GRO~J1655$-$40 & Shaposhnikov et al.,2006, astro-ph/0609757 \\
1312: $1.72{\pm}0.03$ & LH & $-$ & ${\approx}500$ & $-$ & & GRO~J1655$-$40 & This work \\
1313: \enddata
1314: \end{deluxetable}
1315: \clearpage
1316: \setlength{\voffset}{0mm}
1317:
1318:
1319:
1320:
1321:
1322: %_____________________________________________________________________________________________________________________________
1323:
1324: %\clearpage
1325:
1326: %______________________FIGURES_________________________________
1327:
1328: \setcounter{figure}{0}
1329:
1330: \begin{figure}
1331: \centering
1332: \includegraphics[width=1.0\linewidth]{f1.eps}
1333: \caption{Mosaic significance image (obtained in revolution 295) of the GRO~J1655-40 region as seen with ISGRI in the $20-40$\,keV energy range. Besides the target source, several other high energy sources are visible.}
1334: \label{mosaic}
1335: \end{figure}
1336:
1337: \clearpage
1338: \begin{figure}
1339: \centering
1340: \includegraphics[width=0.6\linewidth]{f2.eps}
1341: \caption{Light curves obtained with ISGRI from the INTEGRAL Galactic Monitoring Program in two energy
1342: bands (60-150\,keV and 20-60\,keV), together with OMC (optical). The horizontal lines in the OMC panel show the equivalence in magnitudes of the fluxes. In the third panel, ASM/RXTE (2-12\,keV) light curve is shown in the same period of time. The horizontal
1343: lines indicate the time intervals (one revolution each) over which INTEGRAL spectra were obtained.}
1344: \label{lcurves}
1345: \end{figure}
1346:
1347: \clearpage
1348: \begin{figure}
1349: \centering
1350: \includegraphics[width=0.6\linewidth]{f3.eps}
1351: \caption{RXTE/PCA light curves in the 2-60\,keV band and hardness ratio calculated as the ratio of
1352: the 9.4-18.2\, keV and 2.8-5.7\,keV count rates. The curves were made from 520
1353: RXTE observations, with one (averaged) data point from each observation. Data taken from \citet{homan07}.
1354: }
1355: \label{lcurves2}
1356: \end{figure}
1357:
1358:
1359:
1360: \begin{figure*}
1361: \centering
1362: \includegraphics[angle=270,width=1.0\linewidth]{f4.eps}
1363: \caption{Fitted INTEGRAL spectra, corresponding to epoch 1 with a simple phenomonological power-law model. Refer to the text in Section \ref{powerlaw} and \ref{comptteqpair} in order to get detailed information about this fit.}
1364: \label{fspecpo1}
1365: \end{figure*}
1366:
1367: \begin{figure*}
1368: \centering
1369: \includegraphics[angle=270,width=1.0\linewidth]{f5.eps}
1370: \caption{Fitted INTEGRAL spectra, corresponding to epoch 2 with a simple phenomonological power-law model. Refer to the text in Section \ref{powerlaw} and \ref{comptteqpair} in order to get detailed information about this fit.}
1371: \label{fspecpo2}
1372: \end{figure*}
1373:
1374:
1375: \clearpage
1376: \begin{figure}
1377: \centering
1378: \includegraphics[angle=270,width=1.0\linewidth]{f6.eps}
1379: \caption{Fitted INTEGRAL spectra, corresponding to epoch 1 (hard state). These spectra were fitted with the EQPAIR Comptonization model of \citet{coppi99} ($eqpair$
1380: model in XSPEC). Details about the fitting and the parameters obtained in Section \ref{compttmod} and in Table \ref{param_spec}. JEM-X (single line -black-), ISGRI
1381: (star line -red-) and SPI spectra (square line -green-), are shown, respectively.}
1382: \label{fspec1}
1383: \end{figure}
1384:
1385:
1386: \clearpage
1387: \begin{figure}
1388: \centering
1389: \includegraphics[angle=270,width=1.0\linewidth]{f7.eps}
1390: \caption{Fitted INTEGRAL spectra, corresponding to epoch 2 (thermal dominant state). These spectra were fitted with the EQPAIR Comptonization model of \citet{coppi99} ($eqpair$
1391: model in XSPEC) considering also the soft emission from an accretion disk ($diskbb$ model in XSPEC). Details about the fitting and the parameters obtained in
1392: Section \ref{compttmod} and in Table \ref{param_spec}. JEM-X (single line -black-), ISGRI (star line -red-) and SPI spectra (square line -green-), are shown, respectively.}
1393: \label{fspec2}
1394: \end{figure}
1395:
1396:
1397: \clearpage
1398: \begin{figure}
1399: \centering
1400: \includegraphics[angle=270,width=1.0\linewidth]{f8.eps}
1401: \caption{Fitted INTEGRAL spectra, corresponding to epoch 4 (thermal dominant state). The model used was one consisted of a pure power-law plus emission from an absorbed
1402: multicolor accretion disk of \citet{mitsuda84}. In this period we noticed a change in the properties of the accretion outflow with respect to epochs 1 and 2 (see
1403: Section \ref{powerlaw} and \ref{discuss} for details).}
1404: \label{fspec3}
1405: \end{figure}
1406:
1407:
1408: \clearpage
1409: \begin{figure}
1410: \centering
1411: \includegraphics[angle=270,width=0.8\linewidth]{f9a.eps}
1412: \includegraphics[angle=270,width=0.8\linewidth]{f9b.eps}
1413: \caption{Fitted INTEGRAL spectra, corresponding to epochs from 3 and 4 (upper to lower panel). These spectra were fitted with an absorbed multicolor disk
1414: \citep{mitsuda84} plus a simple power-law iron emission line and edges for both two epochs (and reflection in the case of epoch 3). Details about the fitting
1415: and the values of the parameters obtained in Section \ref{edges} and Table \ref{param_spec}. JEM-X (single line -black-) and ISGRI (star line -red-) are shown,
1416: respectively.}
1417: \label{fspec4}
1418: \end{figure}
1419:
1420: \clearpage
1421: \begin{figure}
1422: \centering
1423: \includegraphics[angle=270,width=0.4\linewidth]{f10a.eps}
1424: \includegraphics[angle=270,width=0.4\linewidth]{f10b.eps}
1425: \includegraphics[angle=270,width=0.4\linewidth]{f10c.eps}
1426: \includegraphics[angle=270,width=0.4\linewidth]{f10d.eps}
1427: \caption{Unfolded spectra from epochs 1 to 4 (upper-left to lower-right). The continuum line shows the total model (see the text and Table \ref{param_spec} for
1428: details), the dashed-dotted line shows the iron K${\alpha}$ emission line, the long dashed line the accretion disk component and the short dashed line the
1429: power-law ($pexriv$ in the third epoch).}
1430: \label{fspec5}
1431: \end{figure}
1432:
1433:
1434:
1435:
1436: \end{document}
1437:
1438: