0811.3381/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[letter]{aa}
3: %\usepackage{graphicx}
4: %\usepackage{txfonts}
5: %\usepackage{natbib}
6: 
7: \newcommand{\integral}{{\it INTEGRAL}}
8: \newcommand{\rxte}{{\it RXTE}}
9: \newcommand{\granat}{{\it GRANAT}}
10: \newcommand{\src}{{1E~1740.7--2942}}
11: \def\nh{N${\rm _H}$}
12: \def\ergsec{erg s$^{-1}$}
13: \def\ferg{erg cm$^{-2}$ s$^{-1}$}
14: \def\phs{ph~cm$^{-2}$~s$^{-1}$}
15: \def\chisq{$\chi^{2}$}
16: \def\rchisq{$\chi_{\nu}^{2}$}
17: 
18: \shorttitle{\src\ at high energy}
19: \shortauthors{Bouchet et al.}
20: 
21: 
22: 
23: 
24: 
25: \begin{document}
26: 
27: 
28: \title{Unveiling the high energy tail of 1E~1740.7--2942
29: with {\it INTEGRAL}\altaffilmark{*} }
30: 
31: \author{
32: L. Bouchet\altaffilmark{1}, M. Del Santo\altaffilmark{2}, E. Jourdain\altaffilmark{1}, J. P. Roques\altaffilmark{1}, A. Bazzano\altaffilmark{2}, 
33: G. De Cesare\altaffilmark{2,3,1}
34:         }
35: %\footnote{
36: \altaffiltext{*}{INTEGRAL is an ESA project with instruments and science data centre
37: funded by ESA member states (especially the PI countries: Denmark, France, Germany, Italy, 
38: Spain, and Switzerland), Czech Republic and Poland with participation of Russia and USA.}
39: \altaffiltext{1}{CESR--Universite de Toulouse/CNRS, 9 Av. du Colonel Roche, 31028 Toulouse Cedex~04, France; jourdain@cesr.fr}
40: \altaffiltext{2}{INAF/Istituto di Astrofisica Spaziale e Fisica cosmica - Roma, via del Cavaliere 100, 00133 Roma, Italy} 
41:  \altaffiltext{3}{Dipartimento di Astronomia, Universita' degli Studi di Bologna, Via
42: Ranzani 1, I40127 Bologna, Italy} 
43: 
44:   
45: \begin{abstract}
46: The microquasar \src\  is observed with \integral\ since Spring 2003. Here, we report on the source high energy
47: behaviour by using the first three years  of data collected with SPI and IBIS telescopes,
48: taking advantage of the instruments complementarity.
49: Light curves analysis showed 
50: two main states for \src: 
51: the canonical low/hard state of black-hole candidates and a ``dim'' state, characterised 
52: by a  $\sim$ 20 times fainter emission, detected only below 50 keV and when summing more than 1Ms of data.
53: For the first time the continuum of the low/hard state has been measured up to $\sim 600$ keV with a spectrum 
54: that is well represented by a thermal Comptonization plus an additional component necessary to fit the data above 200 keV. 
55: This high energy component could be related to non-thermal processes as already observed in other black-hole 
56: candidates.
57: Alternatively, we show that a model composed by two thermal Comptonizations provides an equally representative
58: description of the data: the temperature of the first population of electrons 
59: results as (kT$_{e}$)$_{1} \sim$ 30 keV while the second, (kT$_{e}$)$_{2}$, is fixed at 100 keV.
60: 
61: Finally, searching for 511 keV line showed no feature, either narrow or broad,
62:  transient or persistent. 
63: \end{abstract} 
64: 
65: \keywords{black hole physics -- gamma rays: observations --  radiation mechanisms: general --X-rays: binaries -- X-rays: individual: \src}
66: 
67: 
68: %\maketitle
69: 
70: 
71: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
72: \section{Introduction}
73: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
74: \src\ is a bright hard  X-ray source located at less than one 
75: degree off the Galactic Centre (Hertz \& Grindlay 1984; Cook et al. 1991; Roques et al. 1991),
76: classified as Black Hole Candidate (BHC; e.g. Sunyaev et al. 1991).
77: When Mirabel et al. (1992) discovered a double sided radio jet reaching large angular
78: distances from the core ($\sim$ 1'), the "microquasar" class was born with \src\
79: as its first member.
80: \\
81: All observations performed so far revealed that \src\ spends most of the time in the canonical
82: Low/Hard (LH) state of BHCs (Smith et al. 2002 and ref. therein). In this state, the
83:  X/$\gamma$ ray spectrum is empirically described 
84: by a power-law with a photon index of 1.4-1.5 plus a roll-over around 100 keV \cite{zdz00}. 
85: In a few occasions, soft spectral states have been observed during \src\ low flux levels \cite{smith02}.
86: Moreover, simultaneous \integral\ and \rxte\ broad-band spectral study
87: performed in 2003 report on an intermediate/soft spectral state
88: occurred just before the source quenching \cite{delsanto05}.
89: 
90: In 1990, the SIGMA telescope on-board \granat\ detected a broad line  
91: around the electron-positron annihilation energy (Bouchet et al. 1991; Sunyaev et al. 1991). This 
92: transient   feature appeared clearly during a 13 hours observation 
93: and then possibly in two further occasions but at a less significant level.  Numerous works dedicated to
94: similar line searches have followed and all led to negative conclusions (see for example Cheng et al. 1998 
95: and references therein). In this context, it is  interesting   
96: to perform a deep analysis of  SPI data in this energy domain,
97: and to seek for  any feature around 511 keV associated with \src.
98: \\
99: The superior energy resolution of the SPI telescope allows for a specific dedicated study 
100: of this topic. Indeed, for the first time, an instrument is capable to look for
101: a narrow feature in this particular source. During the first year  of observations, no evidence for point source
102: emission at 511 keV has been detected with SPI.  The upper limit at 3.5 $\sigma$ level is $1.6 \times 10^{-4}$ \phs\ for
103: a narrow line (Teegarden \& Watanabe 2006) while
104: the IBIS  data set a 2 $\sigma$ upper limit 
105: of $1.6 \times 10^{-4}$ \phs\ in the 535-585 keV energy band for 
106: an exposure time equal to 1.5 Ms (De Cesare et al. 2006).
107: 
108: 
109: We report here on the high-energy spectral properties as revealed with 
110: the \integral\ high-energy instruments.
111: The sensitivity and imaging capabilities of IBIS/ISGRI allow to determine the contribution of all 
112: the emitting sources in large fields of view, while the SPI telescope brings some 
113: additional spectral informations above 150-200 keV with a deep investigation of the 511 keV line status.
114:  
115: 
116:  
117: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
118: \section{\integral\ Observations}
119: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
120: 
121: Since its launch on October 17th 2002, \integral\ \cite{wink03} observed the Galactic
122: Center region two times per year, in the Spring and Fall visibility windows. 
123: Observations are performed in dither pattern with each pointing (named science window, SCW) lasting beetwen 1700 and 3600 seconds.
124: We have analysed all public data collected 
125: between Spring 2003 and Fall 2005 by the spectrometer
126: SPI \cite{vedrenne03} and the imager IBIS \cite{ube03}.
127: 
128: After image analysis and cleaning, the useful data set consists in about 3500 exposures
129: for a total useful time of 8 Ms divided in 6 periods (Spring and Fall, 2003, 2004 and 2005,
130: see Table \ref{log}).
131: 
132: 
133: 
134: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
135: \section{Data analysis}
136: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
137: 
138: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
139: \subsection{IBIS}
140: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
141: 
142: 
143: The unprecedented IBIS \cite{ube03} angular resolution combined with sensitivity ($<$1 mCrab
144: for 1 Ms ; Bird et al. 2007) 
145: allow us to resolve sources lying in crowded field, as the Galactic Centre. 
146: The IBIS Partially Coded Field Of View is 29$^\circ \times$29$^\circ$ 
147: at zero response, but the full instrument sensitivity 
148: is achieved in the 9$^\circ \times$9$^\circ$ Fully Coded Field of View.
149: For our aims, we selected  IBIS  observations 
150: including \src\  in the FOV up to 50\% coding (19$^\circ \times$19$^\circ$; see Tab. \ref{log}). 
151: In this paper, we refer to data collected with the IBIS low energy detector, ISGRI \cite{lebrun03}, 
152: covering the 15-1000 keV energy band.
153: 
154: The IBIS  scientific analysis has been performed using the
155: \integral\ off-line analysis software, OSA  \cite{goldwurm03}. 
156: The IBIS/ISGRI images have been extracted SCW by SCW in three energy bands, i.e. 
157: 20-40~keV, 40-100~keV and 100-300~keV.
158: Mosaic images by revolution  have been used
159: to measure fluxes of all sources within 2 degree off \src\ used as input for SPI analysis (see
160: Section \ref{spibis}).
161: \\
162: Spectra have been extracted SCW by SCW in 35
163: logarithmic bins spanning from 20 keV to 600 keV.
164: The response matrices (RMF and ARF) used for spectral fitting 
165: are those delivered with OSA 5.1 distribution. To take into account
166: the improvements included in the matrices delivered in OSA-7, we modified
167: the ISGRI spectra by the factors corresponding to ratios between the 
168: Crab spectra measured respectively with OSA-5 and OSA-7 packages.
169: 
170: 
171: 
172: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
173: \subsection{SPI}
174: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
175:  In addition to its spectroscopic capability, SPI can 
176: image the sky with a spatial resolution of $2.6^\circ$ (FWHM ) over a field 
177: of view of $30^\circ$ (Roques et al. 2003). \\
178: The signal recorded by SPI camera consists of the contributions from sources in the field of view plus 
179: background. A system of equations is to be solved to determine sources and background intensities.
180: In order to reduce the number of unknowns necessary to describe the data, we introduce some known
181:  information on both components. For the background, 
182: the relative count rates of the 19 Ge detectors (uniformity maps) 
183: are very stable and can be kept constant within each considered period (see Table \ref{log}) 
184: while the global normalisation factor is determined by 6 hours intervals.
185: Concerning the sources, timescales are chosen in function of the source intensity and temporal behaviour, 
186: the faintest ones being  considered as constant.
187: Detailed description of the data analysis algorithms and methods, using matrices available in the OSA package,
188:  can be found in Bouchet et al. (2005; 2008).\\
189: Exposures were selected on the basis of their pointing direction 
190: which is here required to be less than $12^\circ$ from \src. 
191: This ensures to keep the maximum sensitivity for \src\ and reduces the
192: total field of view spanned by the observations, leading to
193: a simpler description of the sky.
194: 
195: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
196: \subsubsection{SPI correction from IBIS inputs}\label{spibis}
197: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
198: The region around \src\ is particularly crowded (Bird et al. 2007; Belanger et al. 2006).
199: Due to  the modest SPI angular resolution ($\sim 2.6^\circ$), the spectrum directly 
200: extracted at \src\ position may contain contributions from other weak/close/``not seen'' 
201: sources. Nevertheless, it is possible to obtain the emission spectrum of 
202: \src\ from SPI data thanks to the information provided by IBIS/ISGRI.\\
203: 
204: 
205: For that, we need to determine the fraction of the flux extracted at the \src\ position
206: that  actually originates from the source itself. This has been determined by a set of simulations. 
207: The first step consists to extract the flux
208: of all emitting sources in its neighbourhood measured by IBIS. We then  
209: simulate the counts projected by them on the SPI detector plane, taking into 
210: account 
211: the complete (angular dependent) SPI response. Applying the 
212: standard analysis method to these simulated data  gives us a "\src\ region" flux, 
213: that we can compare to the \src\ flux injected as input in the simulation.
214: The ratio between these two fluxes corresponds to the factor we  
215: have to apply to correct the flux measured by SPI at the \src\ position in the 
216: observed data to obtain the flux attributable to the source itself.
217: This procedure has been repeated in a few broad bands and for each observational period.
218: This cleaning procedure takes into account a global contribution of all
219: potential contaminations, and the corresponding IBIS error bars can be considered as
220: very small  compared to the SPI ones. 
221: However, the contamination effect which is important in the low-energy  domain,
222: becomes negligible when going up to higher energies. In  fact, as can be seen 
223: in Fig. \ref{image}, above 100 keV only  \src\ is detected with  a significant flux within $2^\circ$  
224: off the source.
225: 
226:  
227: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
228: \subsubsection{Modelling the diffuse background : e+e- annihilation line and positronium emission}
229: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
230: 
231: The SPI design makes it sensitive to both source and diffuse emissions. On the other hand, 
232: the Galactic Centre (GC) region  is 
233: dominated in the $\sim$ 300 keV up to 
234: 511 keV domain by the Galactic diffuse annihilation radiation. 
235: The annihilation line emission 
236: is detected with SPI with a flux of 
237: $\sim 1 \times 10^{-3}$ \phs\  
238: and a $\sim 8^\circ$  axisymmetric Gaussian spatial distribution   centered 
239: at the Galactic Centre \cite{kn05, bouchet08}. \src\ continuum emission around 511 keV
240: is expected to be (in the ``hard state'') 
241: of the order of a few percents of the 
242: galactic background line intensity. It is thus crucial for our work to determine this latter accurately.
243: This task has been performed using a larger data set (see Bouchet et al. 2008), which includes observations
244:  at larger latitudes and longitudes.
245:  
246: The  diffuse emission has been described in this process by two gaussians 
247: while eight known sources has been introduced as potential emitters (including \src).
248: Thus, the fitting algorithm (based on a \chisq\ minimization method, see Bouchet et al. 2008
249: for more details) is able to adjust simultaneously point sources fluxes and the diffuse component
250: contribution in the 511 keV line domain. The energy centroid and width of the positron annihilation
251: line were 
252: fixed at 511 keV and 2.5 keV FWHM respectively \cite{chura05}. The  fit procedure results in a
253:  model consisting in  two Gaussians with FWHM of 3.2$^\circ$  and 10.8$^\circ$  and fluxes of 2.3 and 7.0  
254: $\times 10^{-4}$ \phs\, respectively, as the best description of the 
255: annihilation line spatial distribution and flux, without any significant emission
256: from the point sources (see Weidenspointner et al.  2008,
257:  supplementary material, for independent analysis). \\
258: Concerning the positronium we assumed it to follow  the same two Gaussians
259: spatial distribution as the 511 keV line and determined its flux by the same fitting
260: procedure. This results in a positronium  fraction of 0.98, a value that is compatible with 
261: all SPI measurements \cite{bouchet08}.\\
262: Finally, the contributions of  these Galactic diffuse  components on the detector plane
263: are subtracted from the data in the counts space.\\
264: 
265: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
266: \section{Results}
267: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
268: 
269: 
270: 
271: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
272: \subsection{Temporal analysis}
273: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
274: 
275: Table \ref{flux} 
276: gives the \src\ averaged fluxes for the different periods in two broad bands.
277: The source mean  hardnesses in 2003 and 2005 are similar indicating that \src\ was in the 
278: LH state. Its intensity is rather stable on the revolution timescale, within 40 and 60 mCrab 
279: in the 20-40 keV energy band, except in the fall 2003 period, during which a continous decrease,
280:  from 85 to 27 mCrab, preceded the quenching observed in 2004 \cite{delsanto05}.
281: Indeed, in 2004, the source  was weaker with no detection above 5 $\sigma$ within individual revolutions. 
282: The data accumulation by periods allows to determine a mean flux of  a few mCrab.
283: 
284: A dedicated study in a narrow (10 keV) band around  511 keV has been used to search for 
285: any transient emission from the annihilation process. We have tested 0.5 day  
286: and 1 day timescales  without detecting 
287: any significant emission. The actual
288: durations of each temporal bin  depends on the observational
289: planning and is thus variable. The 2 $\sigma$ upper limits range from $4.2 \times 10^{-4}$ \phs\ 
290: to $3.6 \times 10^{-3}$ \phs\ with an averaged value   $\sim 8 \times 10^{-4}$ \phs\, 
291: for a 0.5 day timescale, and  from $3.1 \times 10^{-4}$ \phs\ 
292: to $2.3 \times 10^{-3}$ \phs\ with an averaged value of $6.8 \times 10^{-4}$ \phs\,
293: for a day timescale. These results are illustrated by the distribution of the measurements in 
294: $\sigma$ unit for the 12 hours timescale (Fig \ref{histo511}, solid line) while
295: a 2 $\sigma$ upper limit of
296: $4.8 \times 10^{-5}$ \phs\ is deduced for the total
297: duration (see Table \ref{flux} for upper limits by periods).\\ 
298: Finally, a study has been performed for a broad feature, based on the 240 keV width (FWHM) reported in SIGMA data (Bouchet et al. 1991,
299: Sunyaev et al. 1991). The continuum emission (not negligible in such a broad band) 
300:  has been estimated by the mean flux over the considered period  and subtracted from the data. 
301: Here too, no  significant excess above the expected continuum emission  
302: can be claimed over
303: the 2003-2005 periods (Fig. \ref{histo511}, dashed line). 
304: The 2 $\sigma$ upper limits span from $1.7 \times 10^{-3}$ \phs\ 
305: to $1.3 \times 10^{-2}$  \phs\ with an averaged value close to $3 \times 10^{-3}$ \phs,  
306: for a 0.5 day timescale, and  from $1.2 \times 10^{-3}$ \phs\ 
307: to $8.8 \times 10^{-3}$ \phs\ with an averaged value of $2.5 \times 10^{-3}$ \phs, 
308: for a day timescale. Note that the line flux reported by SIGMA was $1.3 \times 10^{-2}$ \phs.
309:  
310: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
311: \subsection{Spectral analysis}
312: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
313: After correction of the SPI data (Section \ref{spibis}), 
314:  SPI and IBIS/ISGRI spectra have been fitted simultaneously. 
315: We have first built averaged spectra for 2003 and 2005 separately. 
316: In a second step, being the source  in a similar state 
317: during these two periods, we achieved an averaged  LH state spectrum 
318: in order to obtain a better statistics at high energy.\\
319: Spectral fitting of these 3 data sets have been performed with the standard XSPEC v.11.3.1 tools.
320: We have included  a normalisation
321: factor during each fitting procedure and noticed that it remains between 0.94 and 1.0 
322: (ISGRI factor fixed to 1.0). Indeed, SPI spectra are 
323: very similar to the ISGRI ones (as illustrated in fig. \ref{spectrum}), even if they present  
324: some fluctuations at low energy, easily understandable in terms of
325: residual  cross-talk between neighbouring sources. However, this effect is limited 
326: and even  negligible above $\sim $ 100 keV.\\
327: During both periods, the source
328: emission extends up to 500 keV with a spectral shape presenting a clear cutoff around
329: $\sim$ 140 keV.
330: This cutoff is undoubtedly required by the \chisq\ statistics:
331: we obtain \rchisq\ of 10 and 6.6 (for 71 dof) with a power-law model, while 
332: adding a high-energy cutoff these values result as 0.9 and 1.14 (70 dof).
333: \\
334: Then we used a Comptonization model ({\sc{comptt}}, Titarchuk 1994) as this mechanism is expected 
335: to play the major role in our energy domain and to produce  such a cutoff.
336: We obtain electron temperatures (kT$_{e}$) of roughly 50 keV with optical depthes ($\tau$) close to 1 (see 
337: Tab. \ref{fit}), that are quite canonical values for this class of objects. 
338: However, the 2005 and total (2003+2005) spectra give 
339: high \rchisq\ values (1.35 and 1.8 for 69 dof). These, combined  with residuals at high energy,
340: suggest the presence of a supplementary 
341: component explaining data points above 200 keV. We have studied this hypothesis
342: in the total spectrum since its statistics allows us to better constrain 
343: the spectral parameters.
344: In order to model the high-energy data, we added a power law component  and obtain a 
345: \rchisq\ close to 1 with a photon index of 1.9 $\pm$ 0.1. The plasma temperature is thus 
346: decreased as  27 $\pm$ 2.2 keV with  $\tau$ = 1.9 $\pm$ 0.25.\\
347: Even though this two components model is quite satisfying, with a classical explanation of a non-thermal process
348: responsible of the power-law emission, as in High Soft States,  we also made an attempt to test
349: whether alternative scenario with only
350: thermal mecanisms were excluded. \\
351: Indeed,  a two (thermal) Comptonization model gives a similarly acceptable description
352: of the spectrum. 
353: The constraints on the parameters are very poor, so we impose a
354:  second population temperature at 100 keV and consider that it
355: comptonizes photons coming from the first Comptonizing region, i. e. (kT$_{seed}$)$_2$ = (kT$_{e}$)$_1 \sim$ 30 keV.
356: The optical depthes of both regions are found similar and compatible (1.6 $\pm$ 0.1 and 2.2 $\pm$ 0.8) 
357: while the \rchisq\ = 1.07 (67 dof) leads to an F-test probability of $\sim 10^{-8}$ for the existence of such 
358: a second component.  
359: Results corresponding to this scenario are displayed in Fig. \ref{spectrum} and  Tab. \ref{fit}. 
360: Since  this analysis is strongly based on ISGRI data at low energy, we performed a fit  using
361: SPI data above 100 keV only (no contamination effect) and found that the parameters are unchanged.
362: 
363: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
364: \section{Discussion and conclusions}
365: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
366: 
367: Even though the \src\  region is particularly difficult to analyse with the SPI telescope, we have demonstrated
368: that the use of the IBIS and SPI complementarity allows us to get a common spectrum of \src\  itself. Thanks
369:  to the inputs from the ISGRI detector, we have estimated the relative
370: contribution of all sources active in a $1^\circ$ circle around \src\  and showed that we can reconstruct
371: the  SPI \src\  spectrum with a precision better than 5 \% relatively to the ISGRI one.
372: 
373: 
374: Two main states have been observed: the canonical LH state with a 
375:  flux of $\sim$ 50  mCrab   and 60 mCrab  in the 20-40 keV and 40-100 keV bands respectively,
376: and a "dim" state during which the flux of \src\ is below the IBIS/ISGRI detection limit on the revolution timescale
377: and detected at a level of 2.6 -2.7 mCrab (14 $\sigma$) when integrated over $\sim$ 3 months periods.
378: \\
379: Spectra have been built for periods when the source was clearly detected (2003, 2005 and the sum of both).
380: The problematic of the SPI analysis leads to rather large error  bars but in all cases, the emission extends  
381: up to $\sim$ 500 keV, even though a high energy cutoff appears clearly in the data.
382:  When adjusted with a single Comptonization model,   an additional component is strongly required to fit
383: the data above 200 keV, particularly in the total spectrum because of the very significant emission
384: at these energies.\\  
385: This high energy component has been observed in several BHCs (e.g. McConnell et al. 2000; Zdziarski
386: et al. 2001), usually during high soft spectral states, and  explained
387: as  Compton up-scattering by a non-thermal electrons population \cite{zdz-gie04}. 
388: As alternative scenario, jets can easily produce hard X-ray emission via synchrotron radiation in addition to the inverse Compton scattering \cite{mark03}.
389: However, by computing radio-to-gamma ray Spectral Energy Distribution, Bosch-Ramon et al. (2006) 
390: ruled out the jet emission for the hard-X ray spectrum of \src,
391: favoring rather the corona origin. \\     
392: Recently, high energy excesses have been observed in transient BHCs when in LH state (i. e.  Del Santo et al. 2008; Joinet et al. 2007),
393: that could be the result of spatial/temporal variations in plasma parameters \cite{mal-jou00}.
394: We have demonstrated with our data that,  for the LH state spectrum of \src, a model consisting of two thermal
395:  Comptonization components, with a second hotter 
396: population ((kT$_{e}$)$_2$  fixed to 100 keV)  interacting with the photons produced by the first one,
397:   provides an interesting alternative to the non-thermal scenarios. This two 
398:    temperature model could  either correspond to  two distinct heating mecanisms/regions or
399: reflect the presence of a gradient of temperature in the Comptonising plasma.
400:  
401: Finally, even though the complexity of the considered region makes
402:  it difficult to attribute firmly the 
403: detected emission to \src, 
404: the presence of photons with energy greater than several hundreds of keV
405:  in a more or less persistent way (something as half or 2/3 of the time),
406:   together with previously reported annihilation emission,  support  a scenario in which 
407: \src\ is a source of positrons. Indeed, as proposed by van Oss \& Belyanin (1995),
408:  a plasma detected with a temperature much lower than 1 MeV is able to produce
409: positrons throught photon-photon absorption. The basic argument is that the 
410: hard X-ray emission comes from the regions close to the central black hole,
411: where the gravity field is very strong. The high local temperature is thus 
412: lowered, leading to an observed value far from  the relativistic domain, while pairs 
413: are created in the innermost disk and driven away. Annihilation outbursts could occur 
414: when the accretion flow intercepts the pair wind (van Oss \& Belyanin 1995). \\ 
415: Concerning the 511 keV line itself, no feature, broad or narrow, transient or persistent,
416: has been found, confirming the rare occurence of such a phenomenon in line with numerous
417: different studies already performed on this topics. It is worth to note, however,
418:  that SPI/INTEGRAL gives the first
419: opportunity to investigate it in terms of narrow feature (a few keV) with actual constraints on 
420: its parameters. 
421: Unfortunately, any strong conclusion would require more information on the 
422: expected duration (and width) of the emission.  
423: 
424:  
425: 
426: 
427: \acknowledgements
428: We are grateful to ASI (contract I/080/07/0) and CNES for support.
429: As one of the thousands Italian researchers with a long-term temporary
430: position, MDS acknowledges the support of Nature (455, 835-836) and thanks
431: the Editors for increasing the international awareness of the current
432: critical situation of the Italian Research.
433: 
434: 
435: 
436: 
437: 
438: 
439: 
440: 
441: 
442: 
443: \begin{thebibliography}{}
444: \bibitem[Belanger et al. 2006]{Belanger2006}
445: Belanger. G., Goldwurm, A., Renaud, M., et al. 2006, ApJ, 636, 275
446: 
447: \bibitem[Bird et al. 2007]{Bird07}
448: Bird, A.J., Malizia, A., Bazzano, A., et al. 2007, ApJS, 170, 175
449: 
450: 
451: \bibitem[Bosch-Ramon et al. 2006]{bosch06} 
452: Bosch-Ramon, V., Romero, G. E., Paredes, J. M., Bazzano, A., Del Santo, M.
453: \& Bassani, L. 2006, A\&A, 457, 1011
454: 
455: \bibitem[Bouchet et al. 1991]{Bouchet91} 
456: Bouchet, L.  et al. 1991, ApJ, 383, L45 
457: 
458: 
459: \bibitem[Bouchet et al. 2005]{Bouchet05}
460: Bouchet, L., Roques, J. P., Mandrou, P., Strong, A., Diehl, R., Lebrun, F. \& Terrier, R.
461:  2005, ApJ, 635, 1115
462: 
463: 
464: \bibitem[Bouchet et al. 2008]{bouchet08}
465: Bouchet, L., Jourdain, E., Roques, J. P., Strong, A., Diehl, R., Lebrun, F. \& Terrier, R. 2008, ApJ, 679, 1315
466: 
467: \bibitem[Cheng et al. 1998]{cheng98}
468: Cheng, L. X.,  Leventhal, M., Smith, D., Gehrels, N., Tueller, J. and Fishman, G., 1998, ApJ, 503, 809  
469: 
470: \bibitem[Churazov et al. 2005]{chura05}
471: Churazov, E., Sunyaev, R., Sazonov, S., Revnivtsev, M. \& Varshalovich, D.   2005, MNRAS, 357, 1377
472: 
473: 
474: 
475: \bibitem[Cook et al.\ 1991]{cook1991}
476: Cook, W.R., Grunsfeld, J. M., Heindl, W. A., Palmer, D. M., Prince, T. A.,
477:  Schindler, S. M. \& Stone, E. 1991, ApJL, 372, L75
478: 
479: 
480: \bibitem[De Cesare et al. 2006]{decesare06}
481: De Cesare, G., Bazzano, A., Capitanio, F., Del Santo, M., Lonjou, V., 
482: Natalucci, L., Ubertini, P. \& von Ballmoos, P. 2006, AdvSpR, 38, 1457
483: 
484: \bibitem[Del Santo et al. 2005]{delsanto05}
485: Del Santo, M.  et al. 2005, A\&A, 433, 613
486: 
487: \bibitem[Del Santo et al. 2008]{delsanto08}
488: Del Santo, M., Malzac, J., Jourdain, E., Belloni, T. \& Ubertini, P. 2008, 
489: MNRAS, in press. 
490: astro-ph 0807.1018
491: 
492: \bibitem[Goldwurm et al.\ 2003]{goldwurm03}
493:  Goldwurm, A. et al. 2003, A\&A, 411, L223
494: 
495: 
496: \bibitem[Hertz \& Grindlay 1984]{her84} Hertz, P. \& Grindlay, J. E. 1984, ApJ, 278, 137
497: 
498: \bibitem[Joinet et al. 2007]{joi07} 
499: Joinet, A., Jourdain, E., Malzac, J., Roques, J. P., Corbel, S., Rodriguez, J.,
500:  \& Kalemci, E. 2007, ApJ, 657, 400
501: 
502: %\bibitem[Jung et al. 1995]{jung95} 
503: %Jung, G. V., et al. 1995, A\&A, 295, L23
504: 
505: \bibitem[Kn\"odlseder et al. 2005]{kn05}
506: Kn\"odlseder J. et ~al. 2005, A\&A, 441, 513
507:  
508: \bibitem[Lebrun et al. 2003]{lebrun03} 
509: Lebrun, F. et al. 2003, A\&A, 411, L141
510: 
511: \bibitem[Malzac \& Jourdain 2000]{mal-jou00} Malzac, J., \& Jourdain, E. 2000, A\&A, 359, 843
512: 
513: \bibitem[Markoff et al. 2003]{mark03}
514: Markoff, S., Nowak, M., Corbel, S., Fender, R., Falcke, H.  2003, A\&A, 397, 645
515: 
516: \bibitem[McConnell et al. 2000]{McCo00}
517:  McConnell, M. L.  et al. 2000, ApJ, 543, 928 
518: 
519: \bibitem[Mirabel et al. 1992]{Mirabel1992}
520: Mirabel, F., Rodriguez, L. F., Cordier, B., Paul, J. \& Lebrun, F. 1992, Nature, 358, 215
521: 
522: %\bibitem[Mirabel et al. 1993]{mira93} 
523: %Mirabel, F., Rodriguez, L. F., Cordier, B., Paul, J. \& Lebrun, F.  1993, A\&AS, 97, 193
524: 
525: \bibitem[van Oss \& Belyanin 1995]{oss1995}
526: van Oss, R. F. \& Belyanin, A. A., 1991, A\&A., 302, 154
527: 
528: \bibitem[Roques et al. 1991]{Roques1991}
529: Roques, J.P.  et al.  1991, AdSpR, 11, 869
530: 
531: \bibitem[Roques(2003)]{Roques03}
532: Roques J.P.  et~al  2003, A\&A, 411, L91
533: 
534: \bibitem[Smith et al. 2002]{smith02} 
535: Smith, D. M., Heindl, W. A. \& Swank, J. H. 2002, ApJ, 569, 362
536: 
537: 
538: \bibitem[teegarden \& Watanabe 2006]{Teegar06}
539: 
540: Teegarden, B. J. \& Watanabe, 2006, ApJ, 646, 965
541: 
542: \bibitem[Sunyaev et al. 1991]{susu91} 
543: Sunyaev, R.  et al.  1991, ApJ, 383, L49
544: 
545: \bibitem[Titarchuk 1994]{tita94}
546: Titarchuk, L. 1994, ApJ, 434, 313
547: 
548: 
549: \bibitem[Ubertini et al. 2003]{ube03} 
550: Ubertini, P. et al. 2003, A\&A, 411, L131
551: 
552: \bibitem[Vedrenne et al. 2003]{vedrenne03}
553: Vedrenne, G. et al. 2003, A\&A, 411, L63
554: 
555: 
556:    
557: \bibitem[Weidenspointner et al. 2008]{weden08}
558: Weidenspointner, G. et al. 2008, Nature, 451, 159 
559: 
560: 
561: \bibitem[Winkler et al. 2003]{wink03}
562: Winkler, C. et al. 2003, A\&A, 411, L1
563:   
564:  \bibitem[Zdziarski 2000]{zdz00}
565:  Zdziarski, A. A. 2000, IAUS, 195, 153
566: 
567: 
568:  \bibitem[Zdziarski et al. 2001]{zdz01}
569:  Zdziarski, A. A., Grove, E., Poutanen, J., Rao, A. R. \& Vadawale, S. V.  2001, ApJL, 554, L45
570: 
571:  \bibitem[Zdziarski \& Gierli\'nski 2004]{zdz-gie04}
572:  Zdziarski, A. A., Gierli\'nski 2004, Progress of theoretical Physics, 155, 99
573: 
574: \end{thebibliography}
575: 
576: \clearpage
577: 
578: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
579: %%% Table 1 observation log
580: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
581: \begin{table} 
582: %\begin{center}
583: \caption{Observations log.\label{log}}
584: \begin{tabular}{ccccc}
585: \tableline\tableline
586:  Period & Start Time & End Time & Pointings (or Scw) & Exposure (Ms)    \\
587:         & (UT)        & UT       & IBIS  /  SPI            & IBIS  /  SPI \\           
588: \tableline
589: 2003 spring  &  2003 Feb 28  &  2003 Apr 22  & 298 / 434  & 0.62 / 0.86  \\
590: 2003 autumn    &  2003 Aug 19  &  2003 Oct 9   & 717 / 709 & 2.26 /  2.14  \\
591: 2004 spring  &  2004 Feb 17  &  2004 Apr 20  & 547 / 544 & 1.29 / 1.18      \\
592: 2004 autumn    &  2004 Aug 21  &  2004 Oct 28  & 747 / 610 & 1.78 / 1.35    \\
593: 2005 spring  &  2005 Feb 16  &  2005 Apr 20  & 802 / 697 & 1.48 / 1.27    \\
594: 2005 autumn    &  2005 Aug 16  &  2005 Oct 5   & 353 / 407 & 0.70 /  0.93    \\
595: \tableline
596: 
597: \end{tabular}
598: %\end{center}
599: \end{table}
600: 
601: 
602: 
603: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
604: %%% Table 4 511 keV detection
605: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
606: \begin{table} 
607: \scriptsize
608: \begin{center}
609: \caption{Averaged \src\ flux by periods. For the 511 keV feature, 2$\sigma$ upper limits are given 
610: for a narrow line ($\Delta $ E = 10 keV).\label{flux}}
611: \begin{tabular}{ccccccccccc}
612: \tableline\tableline
613:  Period &  2003 spring & 2003 fall  & 2004 spring & 2004 fall & 2005 spring &  2005 fall & 2003 \& 2005 & 2004   &   total period    &  \\
614:  \tableline
615: $F_{20-40 keV}$&51&75&2.6&2.7&53&53&62\\
616:  mCrab & & & & & & & & & &\\
617: \tableline
618: $F_{40-100 keV}$&60&95&-&-&56&54& 68 \\
619:  mCrab & & & & & & & & & &\\
620: \tableline 
621:  $F_{511keV}\times 10^{-4} $ &   $<$ 1.22  &  $<$ 0.66 &   $<$ 1.02 &  $<$ 1.10  & $<$ 1.10 &   $<$ 1.32 & $<$ 0.48 & $<$ 0.74 &  $<$ 0.40\\
622:  \phs\ & & & & & & & & & &\\
623:  \tableline
624: \end{tabular}
625: \end{center}
626: \end{table}
627: 
628: %%%% Table 3 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
629: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
630: \begin{table}
631:  
632: \caption{The best fit  parameters for the combined ISGRI and SPI \src\ spectra with Comptonization
633: model ({\sc{comptt}}, Titarchuk 1994). T0 is fixed to 0.3 keV.    
634: For the 2003+2005 spectrum, the second line corresponds to a 2 Comptonization model ((kT$_{e}$)$_{1}$ , 
635: $\tau_{1}$, (kT$_{e}$)$_{2}$ , $\tau_{2}$) 
636: with the 2nd seed photon temperature fixed to (kT$_{e}$)$_{1}$  and (kT$_{e}$)$_{2}$  fixed to 100 keV (see text). \label{fit}}
637: \begin{tabular}{lcccccccccc}
638: \tableline\tableline
639: period & (kT$_{e}$)$_{1}$&$ \tau_{1}$ &  (kT$_{e}$)$_{2}$&$ \tau_{2}$ & \rchisq\ (dof) &   F-test $ ^{(1)}$\\
640:        &     keV    &    & keV     &                & &                 &          \\
641: \tableline 
642: 2003 & 50.1 $\pm$ 2.6 &  1.00  $\pm$ 0.06 &&& 1.1 (69) & \\
643: 2005 & 45.4 $\pm$ 2.7  & 1.10 $\pm$ 0.08  &&& 1.35 (69)& \\
644: 2003+2005& 48.1 $\pm$ 2.0  & 1.06 $\pm$ 0.05 &&& 1.8 (69)&  \\
645: 2003+2005& 29.4 $\pm$ 3.1 & 1.6 $\pm$0.1 & 100 & 2.2 $\pm$ 0.8 & 1.07 (67) & 10$^{-8}$  \\ 
646: \tableline
647: \end{tabular}
648:  \tablenotetext{(1)} { for the presence of a second Comptonization component.} 
649:   
650: \end{table}
651: 
652: 
653: \clearpage
654: 
655: 
656: 
657: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
658: %%%%%%%%%% Figures 
659: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
660:  
661: \begin{figure}
662:  \plotone{f1.eps} 
663: \caption{IBIS images of the 1E region in 2003 (left) and 2005 (right) in the 100-300 keV
664: energy range. The cyan circles are $2^\circ$ in radius.\label{image}} 
665: \end{figure}
666:  
667:  
668: \begin{figure}
669:  \plotone{f2.eps} 
670: \caption{Distributions of the fluxes measured for the 511 keV line in sigma units, 
671:  for two hypotheses (a narrow feature: solid line, a broad feature: dashed line).\label{histo511}} 
672: \end{figure}
673:  
674: 
675: \begin{figure}
676:  \plotone{f3.eps}
677: \caption{1E~1740.7--2942 spectrum observed by IBIS and SPI for 2003 + 2005 observations.
678: Solid line is a two comptonisation model (2* comptt), with hotter temperature fixed to 100 keV.
679: The other parameters are  described in the text.\label{spectrum}} 
680: \end{figure}
681: 
682: \end{document}