The peculiar IGR J17091-3624 2011 outburst

# The peculiar 2011 outburst of the black hole candidate IGR J17091−3624, a GRS 1915+105 like source?

F. Capitanio, M. Del Santo, E. Bozzo, C. Ferrigno, G. De Cesare, A. Paizis
Istituto Nazionale di Astrofisica, IAPS, Via Fosso del Cavaliere 100, 00133 Rome, Italy
ISDC Data Centre for Astrophysics, Chemin d’Ecogia 16, 1290 Versoix, Switzerland
Istituto Nazionale di Astrofisica, IASF-Mi, Via Bassini 15, I-20133 Milano, Italy
E-mail: fiamma.capitanio@iasf-roma.inaf.it
Accepted 2012 February 28. Received 2012 February 24; in original form 2011 October 19
###### Abstract

We report on the long-term monitoring campaign of the black hole candidate IGR J170913624 performed with INTEGRAL and Swift during the peculiar outburst started on January 2011. We have studied the two month spectral evolution of the source in detail. Unlike the previous outbursts, the initial transition from the hard to the soft state in 2011 was not followed by the standard spectral evolution expected for a transient black hole binary. IGR J170913624 showed pseudo periodic flare-like events in the light curve, closely resembling those observed from GRS 1915105. We find evidence that these phenomena are due to the same physical instability process ascribed to GRS 1915105. Finally we speculate that the faintness of IGR J170913624 could be not only due to the high distance of the source but to the high inclination angle of the system as well.

###### keywords:
X-rays:binaries – accretion discs – Methods: observational
pagerange: The peculiar 2011 outburst of the black hole candidate IGR J170913624, a GRS 1915105 like source?LABEL:lastpagepubyear: 2011

## 1 Introduction

The Black Hole Candidate (BHC) IGR J170913624 was discovered by INTEGRAL/IBIS during a Galactic Centre observation on 2003 April 14–15 (Kuulkers et al., 2003). At the onset of the discovery outburst, the source showed a hard spectrum with a flux of about 20 mCrab in the 40–100 keV energy range. The analysis of IBIS, JEM-X, and RXTE/PCA data of the whole outburst  (Capitanio et al., 2005; Lutovinov & Revnivtsev, 2003; Lutovinov et al., 2005) revealed an indication of a hysteresis-like behaviour. The presence of a hot disc blackbody emission component during the softening of the X-ray emission of the source was also unveiled.

After the INTEGRAL discovery, IGR J170913624 was searched in the archival data of both TTM-KVANT (Revnivtsev et al., 2003) and BeppoSAX/WFC (in ’t Zand et al., 2003). In the former archive, one outburst was discovered dating back to 1994 and reaching a flux of 10 mCrab in the 3-30 keV energy band; the analysis of BeppoSAX/WFC data revealed that a second outburst had occurred in 2001, reaching a flux of 1420 mCrab (2-10 keV).

IGR J170913624 lies at 9.6 from another transient X-ray binary, IGR J17098-3628, discovered on 2005 March 24  (Grebenev et al., 2005) when it underwent a 4 year long outburst (Capitanio et al., 2009a). On 2006 August 29 and 2007 February 19, two XMM-Newton observations of the region around these two sources were performed. While IGR J17098-3624 was detected in a relatively bright state in both observations, IGR J170913624 was not detected and an X-ray upper limit of 710 erg s was obtained (assuming a distance of 8 kpc; Capitanio et al., 2009a).

The refined position of IGR J170913624 provided by Kennea & Capitanio (2009) ruled out the tentative radio counterpart previously proposed for the source (Rupen et al., 2003; Pandey et al., 2006). A re-analysis of the archival radio observations performed 9 days after the source discovery by IBIS in 2003, enabled the identification of a faint transient radio source (sub-mJy level at 5 GHz) that showed a flux increase in the subsequent two weeks and an inverted spectrum, a signature of a compact jet  (Capitanio et al., 2009a). This was consistent with the Low/Hard spectral state ( hereafter LHS) observed by INTEGRAL in the same period (Capitanio et al., 2005). The source behaviour during the 2007 observation campaign was typical of a BHC in outburst, even if the relatively low X-ray flux of the source ( 0.5–10 keV peak flux of 2 10 erg cm s) hindered a detailed spectral evolution study (Capitanio et al., 2009a).

At the end of January 2011 the Swift/BAT hard X-ray transient monitor reported a renewed activity from IGR J170913624. The source flux increased from 20 mCrab on January 28 up to 60 mCrab on February 3 in the energy range 15-50 keV (Krimm et al., 2011; Krimm & Kennea, 2011). The corresponding XRT spectrum obtained with a ToO observation was well described by an absorbed power law with a photon index of 1.730.29  (Krimm & Kennea, 2011). On 2011 February 7, the region around IGR J170913624 was also observed by the IBIS/ISGRI and JEM-X telescopes on board the INTEGRAL satellite. The estimated source flux in the 20-100 keV energy range was 120 mCrab. The combined ISGRI+JEM-X spectrum (5-200 keV) could be well described by an absorbed cut-off power law model with a photon index of 1.4 and a high energy cutoff of about 110 keV. This suggested that the source was in LHS  (Capitanio et al., 2011).

Follow-up radio observations carried out with the ATCA telescope measured a flat spectrum (Torres et al., 2011; Corbel et al., 2011; Rodriguez et al., 2011b) associated with self absorbed compact jets, as expected in accreting black holes in the LHS. Later on, Rodriguez et al. (2011b) reported also on the detection of a discrete jet ejection event usually observed when a BHC undergoes a transition from the Hard Intermediate State ( hereafter HIMS) to the Soft Intermediate State ( hereafter SIMS). A 0.1 Hz QPO, increasing in frequency with the source flux and spectral softening, was revealed by both  Rodriguez et al. (2011a) and Shaposhnikov (2011). These findings motivated a long monitoring campaign that was carried out with Swift/XRT, starting on February 28. The XRT observations were planned to be simultaneous with the INTEGRAL pointings already scheduled in the direction of the source, in order to ensure the broadest possible energy coverage (0.3-200 keV) during the entire outburst.

As reported by Del Santo et al. (2011), on February 28 the XRT+IBIS joint spectrum resulted in a typical High Soft State (HSS) shape, with a prominent disc black body component (kT1keV) and a power-law photon index of 2.20.2. No high-energy cut-off was present up to 200 keV. On 2011 March 14 (MJD 55634) a 10 mHz QPO was detected in a 3.5 ks RXTE observation (Altamirano et al., 2011b). One week later, RXTE/PCA showed a continuous progression of quasi-periodic flare-like events occurring at a rate between 25 and 30 mHz. This kind of variability resembles the “heartbeat” variation observed in the Black Hole (BH) binary GRS 1915105 (Altamirano et al., 2011c; Pahari et al., 2011a, b). Altamirano et al. (2011d) reported a detailed study of the behaviour of the flare-like events of IGR J170913624 during the first 180 days of the outburst. This study classified the different types of flares with the same scheme used by Belloni et al. (2000) for GRS 1915105.

In this paper we report on the Swift and INTEGRAL data analysis of the new outburst of IGR J170913624 started at the end of January 2011.

## 2 Data reduction and analysis

The XRT ToO follow-up observations were performed, when possible, simultaneously to the INTEGRAL ones (Capitanio et al., 2011). INTEGRAL data were collected in the framework of the Galactic bulge observations (public data) and the open time observation of the RX J1713.7-3946 field. Due to the long duration of the outburst, Swift/XRT data were collected also in the period in which the region around IGR J170913624 became unobservable by INTEGRAL. In this paper we made use of the whole available data set of INTEGRAL and Swift observations performed from 28 January to 14 August 2011.

The XRT observations were taken in window timing mode in order to avoid the pile-up effects. Each observation was composed of two or more segments. We reported only the analysis of the first segments of all XRT observations, since the other segments were always consistent with the first segments of each observation. For the XRT data analysis we followed standard procedures (Burrows et al., 2005) and the technique summarized in  Bozzo et al. (2009). XRT light curves and the hardness-intensity diagrams were obtained from the XRT data extracting two different energy ranges, 0.3-4 keV and 4-10 keV.

For the INTEGRAL data analysis, we used the latest release of the standard Offline Scientific Analysis, OSA version 9.0, distributed by the ISDC (Courvoisier et al., 2003) and the latest response matrices available. In particular, the IBIS response matrices were produced using the closest available Crab observations to the 2011 outburst of IGR J170913624. Our INTEGRAL analysis was focused on ISGRI (Lebrun et al., 2003), the low-energy detector of the -ray telescope IBIS  (Ubertini et al., 2003) and on the X-ray monitor JEM-X (Lund et al., 2003). Unfortunately, due to the INTEGRAL observing strategy combined to the small JEM-X field of view (FOV), IGR J170913624 was not in the JEM-X FOV in most of the observations. During the INTEGRAL observations both JEM-X modules were switched on. However, for the data analysis we used the second module (JEM-X2) and checked the consistency with module 1. The ISGRI and JEM-X spectra were extracted in 20-200 keV and 3-20 keV, respectively. A systematic error of 2% was taken into account for spectral analysis  (see also Jourdain et al., 2008).

Details on all the Swift and INTEGRAL data analysed in this paper are given in Table 1 (columns 1-4). The spectral and timing analysis have been performed with HEASOFT 6.9 package. In particular, the periods of the flare-like events were calculated with the FTOOL efsearch. The rms values were estimated from the source light-curves by using an ad hoc developed tool and the IDL Astronomy User’s Library procedures. For the rms calculation, we divided the light curves, extracted in 1 s bins, into 140 s chunks. For each segment we computed the fractional rms after subtracting the expected white noise. We then estimated the fractional rms of the light curves and its uncertainty from the average and standard deviation of the single determinations. The effective frequency range over which the rms is integrated is therefore 0.007-0.5.

## 3 Results

The 2011 outburst of IGR J17091-3624 can be divided in two main phases: during the first one, the source underwent the typical sequence of events of a transient BH in outburst (described in Section 3.1); during the second part, it exhibited “heartbeat” variability previously observed only in GRS 1915105 (Sections 3.2 and  3.3). Finally, a detailed study on the presence of a Compton reflection component and iron line upper limit are given in Section 3.4.

### 3.1 The initial phases of the outburst

The outburst of IGR J170913624 started on MJD55598 (Figure 1) and in about 12 days the X-ray flux of the source (2-10 keV) increased of about 70%. During this starting phase, the Swift and INTEGRAL simultaneous data, when available, could be well fit by an absorbed cutoff power-law model. The source showed a typical hard state spectrum and the photon index and high-energy cutoff remained consistent within the errors (1.5, E100 keV, see Table1 for details). The equivalent hydrogen column density value was consistent with the one reported by Krimm & Kennea (2011), N=(1.10.3)10cm.

Figure 2 shows the combined XRT-ISGRI unfolded LHS spectrum along with the residuals expressed in terms of sigmas (MJD=55603.2, observation n6 in Table 1).

On MJD55610.2, the source displayed evidence for a beginning of a spectral transition to the softer state. The flux continued to increase more rapidly: 100% from the observation n12 until the observation n15 (about 6 days). But, this time, a significant softening of the hard X-ray spectrum (see i. e. bottom panel of Figure 1) was observed, together with a drop in the hard X-ray flux.

During the transition the spectra became steeper and in about two days the fit required a multicolor disc blackbody component ( modeled with diskbb in XSPEC, Mitsuda et al., 1984, hereafter MDBB). Figure 3 shows two spectra extracted at the intermediate hardness values (HR0.2, observation n13 and n14). An acceptable fit to these spectra could be obtained by using an absorbed cutoff power law model). Adding the MDBB component, the F-test probability of a chance improvement is 7% and 0.4%, for the observations n13 and n14, respectively. Thus it is reasonable to add a MDBB component only to the second spectrum.
The obtained spectral parameters of the spectrum n14 are compatible with the intermediate spectral states of a BHC (see e. g. Fender et al. (2004) and Remillard & McClintock (2006) and references therein). During this transition from the hard to the soft state, the inner temperature of the MDBB component (kT) increased from 0.3 keV (observation n14) to 1 keV (observations n1516), while its normalization decreased significantly333 In the MDBB model  (Mitsuda et al., 1984) the square root of the normalization constant is proportional to the apparent inner radius of the truncated disc. However, when the high energy behaviour of the spectrum is modeled with a power law component, the evolution of the disc internal radius can be significantly underestimated (see e. g. Done et al., 2007, p. 28-29)..

At the end of the transition to the soft state (observation n16), the disc temperature reached a value of about 1 keV, while the power law photon index reached 2.1, with no cutoff detectable up to about 200 keV (see Table  1 for details). The fractional rms amplitude of the X-ray emission from IGR J170913624 as measured by XRT data decreased from previous values (2530%) up to about 45% (see Figure 5). Thus, as also reported by Del Santo et al. (2011), the source is probably in the HSS. In the following 65 days (until observation n42) the spectral characteristics of the source showed no significant variability. Figure 4 shows the unfolded spectrum of IGR J170913624 after the transition (observation n33). The fit to these data was obtained with an absorbed MDBB plus a simple power law component. No Compton-reflection from the disc surface and no iron line models were required by the data even though these components are usually expected to be very strong in the canonical soft state of BH binaries (Gierlinsky et al., 1999).

On MJD=55655.8 (observation n34) a short flare, reaching a peak flux of 310 erg cm s (2-10 keV) was detected. No significant changes in the spectral properties of the source were detected during this event.

### 3.2 The appearance of the “heartbeat”

Figure 5 shows the fractional rms amplitude as a function of the hardness ratio, hereafter HR444 We defined as the hardness ratio the ratio of the counts in the 4-10 keV energy band to the counts in the 0.3-4 keV energy band in each XRT observation.. As mentioned above, during the transition from the hard to the soft state, the fractional rms and the HR decreased as expected by a typical transient BH entering the HSS (Fender et al., 2004).

However, from observation n 26 the fractional rms amplitude moved away from the expected values and started to increase and decrease rapidly with a chaotic behaviour (see e. g. Figure 5). The rapid increases correspond to the observations in which the quasi-periodic flare-like events are detected in the light curves (the “heartbeat” in analogy with GRS 1915105, see also Section 1). As an example, Figure 6 shows a zoom of the light curve of one of the XRT observations in which the “heartbeat” is detected.

The “heartbeat” oscillations vary in intensity and in hardness; in some observations they are not detected at all (in these cases lower values of the fractional rms amplitude are measured). No significative variations can be observed in the spectra of each XRT observation with or without the presence of the “heartbeat”. We also observed that the flare-like events lose coherence and change their period with time. Figure 7 shows the evolution of the “heartbeat” period as a function of time. This behaviour is consistent with what observed with RXTE (Altamirano et al., 2011a, d). The two panels of Figure 8 show the “heartbeat” period as a function of hardness and XRT count rate, respectively. No evident correlation between the periods of the flare-like events with the count rate or the HR has been found. The only peculiarity is the presence of a sort of ”forbidden zone” in the possible period values (from 40 s to 65 s, Figures 78). For a detailed discussion of the different “heartbeat” states of IGR J170913624 see Altamirano et al. (2011d).

No significant detection of the “heartbeat” was found in the IBIS light curve because of the faintness of the source in the hard X-ray domain (20–200 keV) and the relatively poor statistics.

After MJD55690 (observations n 43-44), the “heartbeat” was no longer detected and at the same time the flux in the 15–50 keV energy band started to increase again (see the BAT light curve in Figure 1). The spectral analysis of the observations collected during this period showed that the inner temperature of the MDBB component decreased down to 1 keV and a power-law component was also required in order to have an acceptable fit of the XRT spectra. In the previous observations, a power law component additional to MDBB was required only when XRT and IBIS data were fitted simultaneously. Between observations n37-41 the INTEGRAL data were unavailable, and thus we could not constrain the properties of the source emission in the hard X-ray domain.

On MJD=55705.6 (observation n49), the 15–50 keV light curve started to decrease again. Correspondingly, the soft XRT light curve increase significantly (see Figure 1) and the XRT spectra reached again approximately the same shape observed during the previously detected soft state. On the same date, MJD=55705.6, a second group of recurrent flare-like events appeared again in the light curves. At this time the flux variation of the flare events was less pronounced and less coherent, while the periods scanned approximately the same range than in the previous group of events (see Figure 7).

As Figure 1 shows, from MJD55730 until 55770 there was an increase of the XRT flux together with a sharp hardening. The consequence of the hardening in the XRT spectra are an increase of the inner disc temperature and a decrease of the normalization constant of the MDBB model, NORM, that reached values of about 18 (see Table 1). In particular NORM is proportional to the square of the apparent inner disc radius and to cosi, where i is the angle between the disc and the observer (Mitsuda et al., 1984).555The connection between the apparent inner disc radius and the inner radius itself is reported by Kubota et al. (1998)). Thus in order to obtain an inner radius with a plausible length, cosi should be very small. Simultaneously to the spectral hardening, the XRT light curves and the corresponding power spectra clearly showed that a third group of recurrent flare-like events started with a remarkably decreased period (see Figure 7).

As an example, the four panels of Figure 9 show the XRT power spectra evolution, from MJD=55737.5 to MJD=55759.3 (observations n6064 ). This time interval corresponds to the reappearance of the flare-like events: at MJD=55737.3 (observation n60) there were no flare-like events and the power spectrum presented a power law like behaviour (panel a). On MJD=55741.6 the flare-like events started again and a prominent and broad feature appeared in the power spectrum shape (panels b and c). The frequency of this feature changed with time from 0.72 Hz to 0.22 Hz (panels d and e).

The “heartbeat” is always detected during the final part of the Swift campaign with periods that span from about 3 s until 30 s. The energy spectra of each single XRT observation were fitted with the same model than before (absorbed multicolor disc black body plus a power law component). However, after the observation n65, the inner temperature of the MDBB component decreased from 1.5 keV to 1.3 keV (see Table 1 for details). The INTEGRAL observations, performed during revolution 1078 (MJD=55785.0), showed that the fit of the hard part of the spectrum is consistent with a simple power law component with a photon index of =2.3 (see Figure 10).

We also note that during the periods in which IGR J170913624 displayed evidence for the “heartbeat” phenomena, its spectral evolution remained trapped in the top left corner of the hardness-intensity diagram (hereafter HID; see Figure 11) and outlined no more the canonical path through the different spectral states expected from a BHC in outburst (the so called q-track).

### 3.3 Spectra from the “heartbeat”

In order to investigate the origin of the changes in the hardness ratio during the “heartbeat”, we extracted XRT spectra in the time intervals corresponding to the highest (60 ct/s) and lowest (30 ct/s) count rates of the source spectra during the flaring activity. For these data, we performed a rate resolved analysis adding-up time intervals corresponding to the peaks and to the minima of the flare in each observation (note, however, that the hardening of the different peaks was not constant; see for example the HR behaviour in Figure 6). Because of the periodicity of the light curve, the rate resolved analysis overlaps with the phase resolved analysis.

A fit to the spectra was obtained by using an absorbed MDBB component. The spectral parameters at highest count rates indicated a higher inner disc temperature and a hint for a smaller inner disc radius (see Table 2 for details) than what measured at lower count rates. This behaviour is more evident in some observations of the first group of data showing recurrent flare like events (between MJD55630 and MJD55690), where the flux variation during the flares was more pronounced. In fact, unfortunately, due to the low data statistics, only in a few observations was it possible to constrain the MDBB normalization constant with enough confidence.

In the second group (from MJD55700 to MJD55730) the changes in the HR with the source count rates and the coherence of the “heartbeat” oscillation are less evident. We report in Table 2 the spectral parameters of three representative XRT observations selected at different time periods. The N is fixed to be the same for the different phases of the same observations.

The unfolded phase-resolved spectra obtained for the XRT observation n30 (MJD=55640.5) are shown in Figure 12 . We found evidence that the flares are due to an oscillation of the inner disc boundary (Table 2): at the peak of the flare the MDBB temperature (radius) is higher (smaller) with the disc approaching the BH event horizon. The opposite behaviour is observed during the minima of the flare. This is similar to what has been observed in the case of GRS 1915105 (Neilsen et al., 2011). The lower X-ray flux of IGR J170913624 with respect to GRS 1915105, however, does not allow us to study the “heartbeat” in the same details. Theoretical studies suggest that this phenomenon is due to the Lightman-Eardley instability, a limit cycle in the inner accretion disc dominated by the radiation-pressure (Lightman & Eardley, 1974; Nayakshin & Rappaport, 2000; Szuszkiewicz & Miller, 1998). According to this interpretation, the inner part of the disc empties and refills with a timescales of seconds (Belloni et al., 1997).

### 3.4 Reflection component

In order to investigate the presence of Compton-reflection component and the iron line in the spectra of IGR J170913624, we used the XRT, JEM-X and IBIS joint spectra showed in Figure 10. In this case the spectral parameters revealed that IGR J170913624 is in the soft state (observation n33) when the highest contribution from the reflection component is expected (see e. g. Ross & Fabian, 2007, and reference therein). The model used to fit the data is an absorbed MDBB plus an exponentially cut-off power-law spectrum reflected by neutral material (pexrav in XSPEC; Magdziarz & Zdziarski, 1995).

Considering the distance of the source estimated by  Pahari et al. (2011b) and  Rodriguez et al. (2011b), we took into account also the hypothesis that the source could belong to the Galactic halo and thus have a different metallicity with respect to the sources in the Galactic bulge, where normally LMXBs are concentrated (Grimm et al., 2002). No significant changes in the spectral fits were observed by leaving the metallicity of the reflecting medium free to vary. We thus assumed two different values of the metallicity, i.e. the solar one (the source belongs to Galactic bulge, Z/Z = 1) and Z/Z = 0.13 as reported by Frontera et al. (2001) for XTE J1118+480 which is a BH binary that lies at very high Galactic Latitudes. In both cases the estimated upper limit on the reflection component was of R=0.1, and the F-test probability indicated that there is not a clear evidence of a significant improvement in the by adding this component (the F-test probability in the two cases was of 7% and 2%, corresponding to a detection significance of 2.0 and 2.5, respectively).

We also estimated an upper limit on the normalization of the iron line fixing the centroid of the line at 6.7 keV. We assumed a broad line with =0.7 keV as in the case of GRS 1915+105 (see e. g. Martocchia et al., 2002, and references therein). The obtained upper limit on the equivalent width is EQ 0.9 keV.

## 4 Discussion

All the outbursts of IGR J170913624 observed before 2011 were fainter and poorly observed with respect to the last one. However the source, in the limit of the instruments capability, displayed the typical spectral and temporal evolution (Capitanio et al., 2005, 2009a) expected from a canonical BHC (for details on the transient BHC outburst evolution see e.g., Fender et al., 2004). The “heartbeat” phenomenon appeared only during the last 2011 outburst. Indeed, using all the available archival XRT observations in the direction of IGR J170913624, we verified that no “heartbeat” was visible during the previous outbursts of the source.

We summarize here the initial evolution phases of the outburst occurred on 2011. The source underwent to a transition from the LHS to the HSS moving from the bottom right corner of the HID to the top left corner (Figure 11, observations n115):

• during this transition, the source reached the intermediate states and the radio flare reported by Rodriguez et al. (2011b) should be the signature of the transition from the HIMS to the SIMS (Fender et al., 2004);

• the rms amplitude starting from values of about 30% in the LHS, decreased significantly reaching values that span from 6% to 2% (see Figure 5 and column n6 in Table 1);

• the spectrum became softer with the presence of a prominent disc blackbody component (starting from observation n15) with the high energy cutoff no longer detectable up to 200 keV.

The source remained in the HSS for about 10 days (from MJD=55623.5 till MJD=55633.3). Starting from MJD=55635 the source followed no more the standard evolution of a transient BHC in outburst: the properties of the X-ray spectra in each observation showed no significant variability, while the source displayed a sudden atypical timing variability in the form of flare-like events occurring at a 33 s period (“heartbeat”). The X-ray emission at the peak of these flares is typically harder than the average source emission (see the third panel of Figure 6).

Starting from MJD=55692 we measured a progressive decrease of the MDBB inner temperature with a corresponding hardening of the source emission. At this time the flare-like events were no longer visible in the light curve. The hardening continued uninterrupted for about two days, then the inner temperature of the disc started to increase again, leading to a clear increase in the soft X-ray flux and a decrease of the hard X-ray emission. At this epoch, the “heartbeat” became again visible.

The last part of the data analysed presented a short period oscillations (between 3 s and 30 s) and also a particularly hot inner disc temperature with a very small MDBB normalization constant that corresponds to a small apparent inner radius. Between MJD=55740 and MJD=55760, the 4-10 keV XRT flux increased significantly (a factor of 60%) . The peak in the 4-10 keV flux (see Figure 1) corresponds to a peak in the inner disc temperature (T1.7 keV on MJD=55759.3).

The period of the “heartbeat” changed with time (Figure 7) and it seems to have a decreasing trend.

### 4.1 Comparison with the BH binary GRS 1915+105

As reported by Altamirano et al. (2011d) and by Pahari et al. (2011b) the behaviour of the source resembles what observed from GRS 1915105 in the various flaring states. Thus the principal common characteristic between these two sources is just the presence of pseudo periodic flare-like events in the light curve i. e. the so called “heartbeat”. The HR (bottom panel of Figure 6) of IGR J170913624 is similar to the GRS 1915105 one, in the sense that in both sources the modulation of the light curve is projected also in the HR (Neilsen et al., 2011). However, in the GRS 1915105 case the hardness variation seems more pronounced (see for example Naik et al., 2002). Our phase resolved energy spectra of the XRT data revealed that the hardening of the source X-ray emission at the peak and at the lower part of each flare is similar to what measured in the case of GRS 1915105 (see e. g. Mineo et al., 2010; Belloni et al., 1997) and thus it is probably due to the same physical phenomenon (Lightman & Eardley, 1974).

The period of the “heartbeat” seems also to vary with time in the same range of values for the two sources, even though in GRS 1915105 the period amplitude gets larger for long time scales. This does not seem to be the case for IGR J170913624. Indeed, as showed in Figure 7, the period variation with time seems to decrease and, moreover, in the third group of observations (from MJD=55750 until the end) it reaches values of the order of few seconds (35 s). These values were not observed in GRS 1915105 (Neilsen et al., 2011).

Similarly to GRS 1915105, we measured also for IGR J170913624 particularly hot inner disc temperatures (in the case of GRS 1915105 the temperature can reach even higher values; see e.g., Belloni et al., 1997; Muno et al., 1999; Fender & Belloni, 2004) 666 The inner disc radius values reported for GRS 1915+105 by Muno et al. (1999) and related to inner temperatures greater than 1.6 keV, are too small to be associated with the ISCO for any reasonable black hole mass. Even if the hard part of the spectrum, modeled using a power law, could underestimate the inner disc radius (Done et al., 2007), it is not possible to exclude that, in these cases, the accretion geometry could be different from the one predicted by MDBB. However this should not be the case for IGR J17091-3624. In fact, the spectral parameters reported in our analysis are not as extreme as the ones reported by Muno et al. (1999) for GRS 1915+105.. This property, together with a small inner radius of the disc blackbody spectrum in X-ray binaries, has been directly associated with high values of the BH spin (Zhang et al., 1997; Devis et al., 2006).

Besides all these similarities between GRS 1915105 and IGR J170913624, a particularly striking difference is the X-ray flux intensity during the outbursts. This fact cannot be easily explained because, unlike GRS 1915105, for IGR J170913624 we do not have an estimation of the distance, the inclination angle, BH mass and spin, and the properties of the companion star. Some results on optical and NIR counter-part of IGR J170913624 have been reported by Torres et al. (2011).

Chaty et al. (2008), on the basis of optical and NIR photometric and spectroscopic studies of two possible counter-parts of the source, suggested that the source should belong to the Galactic bulge. However,  Rodriguez et al. (2011b) recently estimated a lower limit of the source distance from its hard to soft transition luminosity concluding that, if the transition occurred at luminosity that spans from 4% to 10% of the Eddington luminosity (assuming a BH mass of 10 M), IGR J170913624 is farther from the Galactic bulge, at a distance that spans from about 11 kpc up to 17 kpc. Moreover Pahari et al. (2011b), using a different method, based on QPO, estimated an even larger distance of 20 kpc and a mass range that spans from 8M to 11.4M.

Assuming a distance range of 11-17 kpc, the bolometric luminosities of IGR J170913624 estimated from the observation displaying “heartbeat” with the highest flux would be (3-7) 10 erg s which translates in to L(3-6)% L. However, considering the distance and the BH mass range supposed by Pahari et al. (2011b) these luminosities result in 1% and 8% L.

Since the flare-like events should be at Eddington limit regime (see e.g. Nayakshin & Rappaport, 2000; Neilsen et al., 2011, and reference therein), if we consider the values reported above, we conclude that the faintness of IGR J170913624 should not be only due to the source distance. For this reason  Altamirano et al. (2011d) supposed that the distance of the source could be even larger than 20 kpc, otherwise the BH mass should be extremely small (less than 3).

Other peculiar differences of IGR J170913624 with respect to GRS 1915105 are the lack of detection of the Compton-reflection component and the extremely low apparent inner disc radius (see Section 3.2).

Taking these results as a whole, we speculate that IGR J170913624 could be a highly inclined system and we suggest that the lower luminosity of IGR J170913624 could be also ascribed to the spectral deformation effects due to the high inclination angle as reported by  Cunningham (1998). Indeed, when a Kerr BH is seen at a high inclination angle (cosi 0.25, i75 degrees), the source appears significantly fainter (up to a factor that depends on the BH spin and mass but can reach about an order of magnitude less) with respect to system observed face-on.

At odds with this hypothesis is the lack of detection of eclipses. Although we do not have any information about the system, such as the orbital period or the companion star mass, we can speculate that the lack of eclipses could be related to a small ratio between the companion star and BH mass. Using the Eggleton approximation (Eggleton, 1983), we calculated the relation between the mass ratio, (M/M) and the Roche lobe radius. Then, the minimum inclination angle, , for which the Roche lobe does not cover the central engine along the observer line of sight, is extracted from simple geometrical considerations giving:

 RL/a

where R is the Roche Lobe radius calculated with the Eggleton approximation; is the distance between the BH and the companion star; is the inclination angle of the system. Plotting R versus  (Eggleton, 1983) and considering the equation 1, we found that for , the Roche lobe does not cover the central engine for inclination angles smaller than 75 degrees (Cunningham, 1998).

Moreover, the lack of information on the orbital period of the system hampers the search for the eventual presence of partial eclipses via the usual light curve folding techniques that increase the signal to noise ratio.

## 5 Conclusions

The outcome of the observational campaign presented here suggests that IGR J170913624 can be no longer considered as typical transient Black Hole (Fender et al., 2004). After the transition from hard to soft state in 2011 (Rodriguez et al., 2011b), the source did not follow the standard q-track in the HID diagram (see e. g. Homan et al., 2001; Homan & Belloni, 2005, and reference therein) and, since March 2011, it remained trapped in an oscillatory state, similarly to what observed during the flaring states of GRS 1915105 (Altamirano et al., 2011d).

As mentioned above (see section 4.1), the pseudo periodic bursts in the light curve of GRS 1915105 reach the Eddington luminosity and are believed to be related to disc oscillations. The physics that drives these inner disc oscillations is connected with both the local Eddington limit and the radiation pressure instability. If the “heartbeat” oscillations seen from IGR J170913624 are interpreted as being due to the same mechanism as in GRS 1915105  (as also supposed by Altamirano et al., 2011d), then the apparent ”faintness” of IGR J170913624 remains unexplained unless to suppose a huge distance or an extremely low BH mass (Altamirano et al., 2011d).

In Section 4.1 we noted that that a reduction of the apparent luminosity up to an order of magnitude can also be achieved if the system is seen nearly edge on (for inclination angles 75 degrees, Eggleton, 1983). According to this idea and considering also the L/L ratio calculated for the different distance values, we can speculate that the source, probably, not only lies far from the Galactic bulge, in agreement with  Rodriguez et al. (2011b), but it is observed at a high inclination angle as well. As also discussed in Section 4.1, this finding is not in contrast with the lack of eclipses in the source light curve. In fact, if the companion star is small, the eclipses can be undetected even for a high inclination angle, as for example in the case of the BHC XTE J1118+480 (see e.g. Wagner et al., 2001; McClintock et al., 2001).

We note that at present we cannot exclude that the faintness of IGR J170913624 is only due to a very large distance (20 kpc) or to the extremely low BH mass ( 3M), as suggested by Altamirano et al. (2011d). The large distance, unusual for low mass X-ray binaries generally concentrated in the Galactic bulge (Grimm et al., 2002), could agree with the hypothesis reported by Jonker & Nelemans (2004) that the distances of the LMXB could be affected by a systematic error due to misclassification of the companion star.

However recent results, reported by King et al. (2012), based on a Chandra observation campaign, support the hypothesis that IGR J170913624 is observed at high inclination angle. Future refined estimation of the distance and the BH mass of IGR J170913624 might help understand if GRS 1915105 and IGR J170913624 are very similar objects simply observed at very different distances or inclination angles. We point out that the high inclination of the system is a possible scenario to explain the low luminosity of the source without invoking very large distances or extremely low BH masses that may challenge the Rhoades & Ruffini limit (Rhoades & Ruffini, 1974).

Finally we suggest that, as in the case of GRS 1915105, also IGR J170913624 might show a ”quasi-persistent” outburst of the order of years. Thus the INTEGRAL and Swift observation campaign of the 2011 outburst probably caught the evolution of a transient BH in a persistent GRS 1915105-like phase.

## Acknowledgments

FC, MDS, AP and GDC acknowledge financial support from the agreement ASI-INAF I/009/10/0. FC thanks Giorgio Matt and Piergiorgio Casella for useful scientific discussions. MDS and GDC acknowledge contribution by the grant PRIN-INAF 2009. We would like to thank N. Gehrels and the Swift Team for making Swift observations possible. A special thanks goes to E. Kuulkers and the INTEGRAL Galactic bulge monitoring program. INTEGRAL is an ESA project with instruments and science data centre funded by ESA member states especially the PI countries: Denmark, France, Germany, Italy, Switzerland, Spain, Czech Republic and Poland, and with the participation of Russia and the USA.

## References

• Altamirano et al. (2011a) Altamirano D. et al., 2011a, Atel, 3299
• Altamirano et al. (2011b) Altamirano D. et al., 2011b, Atel, 3225
• Altamirano et al. (2011c) Altamirano D. et al., 2011c, ATel, 3230
• Altamirano et al. (2011d) Altamirano D. et al., 2011d, ApJL, 742, L17
• Belloni et al. (2000) Belloni T., Klein-Wolt M., Mendez M., van der Klis M., van Paradijs J., 2000, A&A, 355, 271
• Belloni et al. (1997) Belloni T., Mendez M., King A. R.van der Klis M., and van Paradijs J., 1997, ApJ, 479, L145
• Bozzo et al. (2009) Bozzo E., Giunta A., Stella L., Falanga M., Israel G., Campana S., 2009, A&A, 502, 21
• Burrows et al. (2005) Burrows D. et al., 2005, SSRV, 120, 165
• Capitanio et al. (2011) Capitanio F. et al., 2011, Atel, 3159
• Capitanio et al. (2005) Capitanio F. et al. 2005, ApJ, 622, 503
• Capitanio et al. (2009a) Capitanio F. et al., 2009a, ApJ, 690, 1621
• Capitanio et al. (2009b) Capitanio F., Belloni T., Del Santo M., Ubertini P., 2009b, MNRAS, 398, 1194
• Chaty et al. (2008) Chaty, S., Rahoui, F., Foellmi, C., Tomsick, J. A., Rodriguez, J., Walter, R., 2008, A&A, 484, 783
• Corbel et al. (2011) Corbel, S., Rodriguez, J., Tzioumis, T., Tomsick, J., 2011, Atel, 3167
• Courvoisier et al. (2003) Courvoisier T. J.-L. et al. 2003, A&A, 411, 53L
• Cunningham (1998) Cunningham C. T., 1998, ApJ, 202, 788
• Del Santo et al. (2011) Del Santo M. et al., 2011, Atel, 3203
• Devis et al. (2006) Devis W. D., Done C., Blaes O. M., 2006, ApJ, 647, 525
• Done et al. (2007) Done C., Gierlinski M., Kubota A., 2007, A&ARv, 15, 1
• Eggleton (1983) Eggleton P. P., 1983, ApJ, 268, 368
• Fender et al. (2004) Fender R. P., Belloni T. & Gallo E., 2004, MNRAS, 355, 1105
• Fender & Belloni (2004) Fender R. & Belloni T., 2004, ARA&A., 42, 317
• Frontera et al. (2001) Frontera F. et al., 2001, ApJ, 561, 1006
• Gierlinsky et al. (1999) Gierlinsky M., Zdziarski A. A., Putanen J., Coppi P. S., Ebisawa K., Johnson W. N., 1999, MNRAS, 309, 496
• Grebenev et al. (2005) Grebenev S.A., Molkov S.V., Sunyaev R.A., 2005, Atel, 444
• Grimm et al. (2002) Grimm H-J., Gilfanov M. & Suyaev R., 2002, A&A, 391, 923
• in ’t Zand et al. (2003) in ’t Zand J. J. M., Heise J., Lowes P., & Ubertini, P. 2003, ATel, 160
• Kennea & Capitanio (2009) Kennea J. A. & Capitanio F., 2009, Atel, 1140
• King et al. (2012) King A. L., et al. 2012, ApJL, 746, 20
• Krimm et al. (2011) Krimm H. A. et al., 2011, Atel, 3144
• Krimm & Kennea (2011) Krimm H. A. & Kennea J. A., 2011, Atel, 3148
• Kubota et al. (1998) Kubota A., Tanaka Y., Makishima K., Ueda Y., Dotani T., Inoue H., & Yamaoka K. 1998, PASJ, 50, 667
• Kuulkers et al. (2003) Kuulkers E., Lutovinov A., Parmar A., Capitanio F., Mowlavi N., Hermsen W. 2003, ATel 149
• Homan et al. (2001) Homan J., Wijnands R., van der Klis M., Belloni T., van Paradijs J., Klein-Wolt M., Fender R., Mendez M., 2001, ApJS, 132, 377
• Homan & Belloni (2005) Homan, J., Belloni, T., 2005, ApSS, 300, 107
• Jonker & Nelemans (2004) Jonker P. G. & Nelemans G, 2004, MNRAS, 354, 255
• Jourdain et al. (2008) Jourdain E., Gotz D., Westergaard N. J., Natalucci L., Roques J. P., 2008, 7th INTEGRAL Workshop, PoS, 144
• Lightman & Eardley (1974) Lightman A. P. & Eardley D. M., 1974, ApJ, 187, L1
• Lebrun et al. (2003) Lebrun F. et al., 2003, A&A, 411, 141
• Lund et al. (2003) Lund N., Budtz-Jorgensen C., Westergaard N. J, Brandt S. et al. 2003, A&A 411, 231L
• Lutovinov et al. (2005) Lutovinov A. A., Revnivtsev M., Molkov S., & Sunyaev R. 2005, A&A, 430, 997
• Lutovinov & Revnivtsev (2003) Lutovinov A. A., & Revnivtsev M. G., 2003, Astron. Lett., 29, 719
• Magdziarz & Zdziarski (1995) Magdziarz P. & Zdziarski A. A., 1995, MNRAS, 273, 837
• Martocchia et al. (2002) Martocchia A., Matt G., Karas V., Belloni T., Feroci M., 2002, A&A, 387, 215
• McClintock et al. (2001) McClintock J. E., Garcia M. R., Caldwell N., Falco E. E., Garnavich P. M., Zhao P., 2001, ApJL, 551, L147
• Mineo et al. (2010) Mineo T., Massaro E., Cusumano G., 2010, AIP Conf. Proc. 1248, 183
• Mitsuda et al. (1984) Mitsuda K. et al., 1984, PASJ, 36, 741
• Muno et al. (1999) Muno M. P., Morgan E. H., Remillard R., 1999, ApJ, 527, 321
• Munoz-Dariaz et al. (2011) Munoz-Darias T., Motta S., Belloni T. M., 2011, MNRAS, 410, 679
• Naik et al. (2002) Naik S., Agrawal P. C., Rao A. R., Paul B., 2002, MNRAS, 330, 487
• Nayakshin & Rappaport (2000) Nayakshin S. & Rappaport S., 2000, ApJ, 535, 798
• Neilsen et al. (2011) Neilsen J., Remillard R. A., Lee J. C., 2011, ApJ, 737, 69
• Pahari et al. (2011a) Pahari M., Yadav J. S., Bhattacharyya S., 2011a, Atel 3418
• Pahari et al. (2011b) Pahari M., Yadav J. S., Battacharyya S., 2011b, astro-ph/1105.4694, subitted to ApJL
• Pandey et al. (2006) Pandey M., Manchanda R., Rao A. P., Durouchoux P., Ishwara-Chandra C. H., 2006, A&A, 446, 471
• Remillard & McClintock (2006) Remillard, R. A., McClintock, J. E. 2006, ARA&A, 44, 49
• Revnivtsev et al. (2003) Revnivtsev, M., Gilfanov, M., Churazov, E., & Sunyaev R. 2003, ATel, 150
• Rodriguez et al. (2011a) Rodriguez J., Corbel S., Tomsick J. A., Paizis A., Kuulkers E., 2011a, Atel 3168
• Rodriguez et al. (2011b) Rodriguez J., Corbel S., Caballero I., Tomsick J. A., Tzioumis T., Paizis A., Cadolle Bel M., Kuulkers E., 2011b, A&A 533, L4
• Rhoades & Ruffini (1974) Rhoades C. & Ruffini R., 1974, PhRvL, 32, 324
• Ross & Fabian (2007) Ross R. R. & Fabian A. C., 2007, MNRAS, 381, 1697
• Rupen et al. (2003) Rupen M. P., Mioduszewski A. J., Dhawan V., 2003, Atel 150.
• Shaposhnikov (2011) Shaposhnikov N., 2001, Atel, 3179
• Szuszkiewicz & Miller (1998) Szuszkiewicz E. & Miller J. C., MNRAS, 1998, 298, 888
• Torres et al. (2011) Torres M. A. P., Jonker P. G., Steeghs D., Mulchaey J. S., 2011, Atel, 3150
• Ubertini et al. (2003) Ubertini, P. et al., 2003, A&A, 411, 131
• Wagner et al. (2001) Wagner R. M., Foltz C. B., Shahbaz T., Casares J., Charles P. A., Starrfield S. G., Hewett P., 2001, ApJ, 556, 42
• Zhang et al. (1997) Zhang S. N., Cui W., Chen W., 1997, ApJL, 482, L155
You are adding the first comment!
How to quickly get a good reply:
• Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
• Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
• Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters