IGR J16328−4726: A NEW CANDIDATE SUPERGIANT FAST X-RAY TRANSIENT

M. Fiocchi; V. Sguera; A. Bazzano; L. Bassani; A. J. Bird; L. Natalucci; P. Ubertini

doi:10.1088/2041-8205/725/1/L68

1. INTRODUCTION

IGR J16328−4726 is a poorly studied X-ray source first reported by Bodaghee et al. (2007) as a transient unidentified source located in the Galactic plane. It is listed as a "blended" variable source in the 4th Imager on Board Integral Satellite (IBIS)/ISGRI Catalog (Bird et al. 2010). The blending is due to the vicinity of the bright transient X-ray binary 4U 1630−47 with a 20–100 keV flux of 4.41 × 10⁻¹¹ erg cm⁻² s⁻¹. This IGR J16328−4726 is present in the Swift BAT 39 month all-sky survey (Cusumano et al. 2010) as a gamma source with a flux of (4.8 ± 0.3) × 10⁻¹¹ erg cm⁻² s⁻¹ in the 14–70 keV energy band. Grupe et al. (2009) reported the detection of a flare observed with Swift/BAT (at 07:54:27 UT on 2009 June 10) which was followed up by a Swift/X-ray Telescope (XRT) observation, ∼385 s after the trigger. Using 2754 s of XRT photon counting (PC) mode data, they determined the best source position yet at R.A. = 248.1579 and decl. = −47.3951 with an uncertainty of 1.7 arcsec (90% confidence). The spectrum was modeled with an absorbed power law having Γ = 0.56^+0.75_−0.68 and an absorption column density N_H = 8.1^+5.7_−4.9 × 10²² cm⁻² in excess to the galactic value of ∼1.5 × 10²² cm⁻². Recently, Corbet et al. (2010) analyzed the Swift BAT 58 month survey (Baumgartner et al. 2010) light curve (from 2004 December 20 to 2009 September 30) in the 15–100 keV energy band and they reported a highly significant modulation at a period close to 10 days. This periodicity was interpreted as the orbital period of a high-mass X-ray binary, powered by accretion from the wind of a supergiant companion.

During the repeated Galactic plane scanning performed with INTEGRAL a new class of hard X-ray transient emitters have been discovered: the supergiant fast X-ray transients (SFXTs). These systems display short outbursts with very fast rise time (∼ tens of minutes) and typical duration of a few hours. The dynamical range observed in the 20–60 keV between quiescent (L_X ∼ 10³² erg s⁻¹) and peak (10³⁶–10³⁷ erg s⁻¹) luminosities is quite high (10²–10⁵) while the X-ray spectra measured are often hard, like those characterizing high-mass X-ray binaries (HMXBs) hosting an X-ray pulsar.

Here we provide a detailed analysis of the timing and spectral properties of IGR J16328−4726, showing that it is a highly variable source and a new candidate for the SFXT class.

2. OBSERVATIONS AND RESULTS

2.1. INTEGRAL/IBIS Results

In order to provide new information on IGR J16328−4726, we have analyzed the INTEGRAL/IBIS-JEM-X public database. We also reanalyzed XRT observations performed during a source flare.

The INTEGRAL (Winkler et al. 2003) observations are divided into uninterrupted ∼2000 s intervals, known as science windows (SCWs). Data are extracted from each individual SCW and processed using the Off-line Scientific Analysis (OSA v9.0) software released by the INTEGRAL Scientific Data Centre (Courvoisier et al. 2003). We analyzed all public IBIS/ISGRI (Ubertini et al. 2003) observations within 15° of the source, i.e., 3783 pointings performed between 2003 January 29 and 2009 March 20, yielding a total exposure time of about 9.8 Ms. These runs were performed with AVES cluster, designed to optimize performances and disk storage for the INTEGRAL data analysis by Federici et al. (2009). Images, spectra, and light curves were extracted for each Science Window in the 20–50 keV energy band.

2.2. The Image Analysis

For most of the observation time, IGR J16328−4726 was below the instrument sensitivity (∼0.5 mCrab for an exposure time of ∼9Ms). It was detected in several single SCWs, in groups of the SCWs, and also in a revolution without a clear periodicity, although this study is limited by the source coverage and by the fact that IGR J16328−4726 is only 0.2 degrees from the bright transient black hole 4U 1630−472 which often contaminates the source emission, the IBIS/ISGRI angular resolution being of the same order of the distance between the two objects. For this reason, we do not consider the activity periods associated with IGR J16328−4726 during those times when 4U 1630−472 is detected with a signal-to-noise ratio (S/N)>4σ.

We created an image for each SCW in the 20–50 keV energy band and searched for the detections of IGR J16328−4726 above a threshold of 4σ; we visually inspected all SCW images with a positive detection. This SCW image inspection was very important because it allowed to further check the presence of the near sources and so properly associate the hard X-ray emission of IGR J16328−4726 compared to the one from the black hole 4U 1630−472. This analysis has shown that IGR J16328−4726 was active in 27 SCWs with a significance >4σ, but only 14 detections can be assigned to IGR J16328−4726 for certain. In the remaining cases, the near black hole 4U 1630−47 is bright in the image and detected with an S/N higher than 4.

From the image analysis, two strong outbursts have been identified: the first in one SCW of revolution 287 and the second in four consecutive SCWs of revolution 768, occurring on 53420.65 MJD and 54859.99 MJD, respectively (see Table 1). We assumed the beginning of the first SCW during which the source was detected as the start time of the outburst and similarly the outburst stop time at the end of the last SCW during which the source was detected. Due to the sparse sampling of the observations, we cannot precisely determine the duration of these outbursts. The first outburst lasted for ∼1 hr (see details in Section 2.3), while for the second outburst the lower limit on the duration is ∼3.5 hr. Figure 1 (top panels) shows the image sequence of three consecutive SCWs (20–50 keV) during which the first outburst was observed and in Figure 1 (bottom panels) we report the image sequence of five consecutive SCWs (20–50 keV) of the second one. No data are available in the INTEGRAL archive to study the outburst behavior immediately after. Furthermore, IGR J16328−4726 was also detected in nine single SWCs (2–3 ks each) and in one of revolution (using its whole duration) with a significance 4σ–5σ and 7σ, respectively.

**Figure 1.** Top panel: ISGRI Science Window image sequence (20–50 keV) of the first outburst (No. 1 in Table 1) of IGR J16328−4726 (encircled). The duration of each ScW is ∼2000 s. Significance of the detection, left to right, was <2σ, 6σ, and <2σ. Bottom panel: ISGRI Science Window image sequence (20–50 keV) of the second outburst (No. 2 in Table 1) of IGR J16328−4726 (encircled). Significance of the detection, left to right, was <2σ, 11σ, 8σ, 4σ, and 7σ.
Download figure:
Standard image High-resolution image

Table 1. Summary of the Science Windows with Detections of IGR J16328−4726

Obs. ID	IJD	Exposure	ISGRI Rate^a
	(=MJD–51544)	(s)	(counts s⁻¹)
First outburst (2005 Feb 19 15:36:00.000 UTC): duration=∼ 1 hr
028700470010	1876.65	2576.0	3.8 ± 0.6

Second outburst (2009 Jan 28 23:40:33.880 UTC): duration=∼ 3.5 hr

076800590010	3315.99	2270.2	9.0 ± 0.8
076800600010	3316.03	2251.6	7.2 ± 0.8
076800610010	3316.07	2267.9	4.1 ± 0.9
076800620010	3316.11	2215.6	6.4 ± 0.9
Long time detection (2005:241:10:19:12.000 UTC)
rev 351	2067.43	129430	0.61 ± 0.08

Note. ^ain the 20–50 keV energy range.

Download table as: ASCII Typeset image

An image mosaic using data from five SCWs of two strong outbursts has been created. The scientific analysis of this mosaic shows that IGR J16328−4726 was detected with a significance level of ∼20σ at R.A. (J2000) = 248.150 and decl. (J2000) = −47.388, with an error circle of 1.5 arcmin (at 90% confidence level) and a flux of ∼3.2 × 10⁻¹⁰ erg cm⁻² s⁻¹ in the 20–50 keV energy band. The position, flux, and significance are slightly different from values reported by Bird et al. (2010). We expect that this analysis, tuned to the specific case of source, should yield a better position estimate than the standard analysis performed for the survey catalog production.

2.3. The Light Curve Analysis

ISGRI light curves (20–50 keV) are shown in Figure 2 for the first (top) and second (bottom) outbursts. The observed outbursts have complex structures characterized by several fast flares with both rise and decay times less than 1 hr or lasting a few tens of minutes. In the top panel, we show the time behavior around the first outburst: IGR J16328−4726 was below the IBIS detection limit for most of the time and becomes active in two consecutive SCWs (028700360010 and 028700370010) and then in the SCW 028700470010 i.e., later 20 ks. This last detection is the strong outburst of the revolution 287 that we are here describing as first outbursts. In the bottom panel, the time behavior around the second outburst was shown: IGR J16328−4726 was consistent with no emission for long time and then active in the consecutive SCWs 076800590010, 076800600010, 076800610010, and 076800620010. Figure 3 shows the light curve zoom at the time of the outbursts. In the left panel is reported the first outburst behavior: IGR J16328−4726 flux is initially consistent with zero, then suddenly turns on at 2005 February 19 15:32:46.91 UTC and quickly reaches the peak after ∼15 minutes (15:47:46.91 UTC). Then it drops to a lower flux level for ∼37 minutes and at 15:55:16.91 UTC the source completely disappears below the IBIS detection limit. Right panel shows the second outburst behavior: IGR J16328−4726 initially has a flux consistent with no emission, then suddenly turns on at 2009 January 29 00:30:36.00 and reaches the peak after ∼30 minutes and it remains active for more than 3 hr, after which unluckily there is not IBIS data coverage available in the INTEGRAL archive.

**Figure 3.** Zoom of the ISGRI light curve (20–50 keV) of the first (left) and second (right) outbursts of the IGR J16328−4726 (see Table 1). Time is in seconds since JD = 2453421 05:36:06.910 and 2454860 10:43:46.186, respectively, for the first and second outbursts.
Download figure:
Standard image High-resolution image

2.4. The Spectrum Analysis

We extracted the ISGRI spectra during the two strong outburst periods and during the rev. 351. The spectrum observed during the first outburst can be fitted with a single power law (χ²/dof = 4.1/3) having a photon index of 2.2 ± 0.9 and a 20–50 keV flux of ∼1.9 × 10⁻¹⁰ erg cm⁻² s⁻¹. The spectrum observed during the second and strongest outburst can be fitted with a single power-law (χ²/dof = 2.5/4) having a similar photon index (2.0 ± 0.4) and a higher 20–50 keV flux (∼3.3 × 10⁻¹⁰ erg cm⁻² s⁻¹). During revolution 351, the spectrum was well fitted with a simple power-law model (χ²/dof = 4.2/3), having the same spectral index (2.6 ± 1.2) and a flux of ∼4.4 × 10⁻¹¹ erg cm⁻² s⁻¹ in the 20–50 keV band.

These spectra are shown in Figure 4.

As is typical for these HMXB systems, we also fitted these spectra with a thermal bremsstrahlung model; the temperatures are 43⁺¹⁵⁶₋₂₄ keV, 30⁺¹⁷⁰₋₁₈ keV, and 51⁺⁵⁸₋₁₉ keV for the outbursts of revolution 287, 351, and 768, respectively.

The emerging picture is that this source shows strong variability on different timescales: two strong outbursts detected in single SCW or in a group of few SCWs, a weak detection for a long time (73 SCWs of rev. 351) and some very weak detections lasting only one SCW. The spectral analysis showed that the spectral shape is constant while the flux change by a factor up to ∼10.

In order to study the quiescent emission in the hard X-ray band, we obtained a mosaic image with exposure of ∼7 Ms using all available SCWs except for those in which both IGR J16328−4726 and 4U 1630−47 were detected. This analysis allows us to get a 3σ upper limit for the flux of 2.5 × 10⁻¹² erg cm⁻² s⁻¹, giving a ratio between the outburst and quiescent fluxes >165 in the 20–50 keV energy band.

2.5. ${\rm INTEGRAL}$ /JEM-X Results

We analyzed all public JEM-X pointings within 4 fdg 5 of the source, performed between 2003 January 29 and 2009 March 20, yielding a total exposure time of 372 ks. Unfortunately, because of the relatively small field of view of JEM-X, there is no JEM-X coverage during the outbursts detected in IBIS.

The JEM-X data analysis shows no detections for IGR J16328−4726, the mosaic image provides an upper limit of 9.6 × 10⁻¹³ erg cm⁻² s⁻¹ in the 3–10 keV energy range (2σ confidence level).

2.6. ${\rm Swift}$ /XRT Results

IGR J16328−4726 was observed with the XRT instrument ∼385 s after the BAT trigger on 2009 June 10 and during the following days. The source showed a clear decline trend during this period with the brightest state observed in the first observation and the lowest flux measured in the last pointing. Unfortunately, no IBIS data simultaneous with the XRT observation were available but we can still use these two observations to study the variability of the source. Observations considered in this work were performed on 2009 June 10 07:38:00 and 2009 June 14 00:10:00 with an exposure time of 8221.5 s and 2864.2 s, respectively. The XRT (Burrows et al. 2005) data reduction was performed using the XRTDAS standard data pipeline package (xrtpipeline version 0.12.2), in order to produce screened event files. All data were extracted only in the photon counting (PC) mode (Hill et al. 2004), adopting the standard grade filtering (0–12 for PC mode), and analyzed with XIMAGE version 4.4 in the 0.3–10 keV energy band. We extracted the image from the first 630 s of the first pointing where IGR J16328−4726 was at the peak of the X-ray outburst with a 0.3–10 keV flux of (2.4 ± 0.1) × 10⁻¹⁰ erg cm⁻² s⁻¹. During the second observation, the flux had dramatically decreased being the source not detected by XRT: the 3σ upper limit on the 0.3–10 keV flux was 1.4 × 10⁻¹² erg cm⁻² s⁻¹. This is consistent with that measured by the JEM-X instrument.

Using the XRT fluxes we can compute a dynamic range (ratio between the outburst and quiescent fluxes) >170 for IGR J16328−4726 which is a value typical of known SFXTs. This kind of X-ray transient activity is higher than that generally observed in other HMXBs, i.e., Be/X-ray transients. In fact in this type of emitters their ratio between outburst and quiescence fluxes is generally below 20.

The XRT restricted position at the 99% confidence level (R.A. = 248.1579, decl. = −47.3937 with an error box of 2.7 arcsec) allows us to pinpoint the infrared counterpart of this source as the Two Micron All Sky Survey object (2MASS J16323791−4723409), in the Spitzer source (GLIMPSEG 336.7492+00.4223⁴) and in the DENIS source (DENIS J163237.9-472341⁵) with R.A. (J2000) = 248.158 and decl. (J2000) = −47.395. This counterpart has magnitudes in the J, H, K bands of 14.631, 12.423, and 11.275, respectively. Due to this absorption, no optical counterpart was visible in UVOT images or listed in the USNO B1 catalog. An HMXB nature would be consistent with the high level of absorption found in the soft X-rays (Grupe et al. 2009).

3. DISCUSSION

Before the INTEGRAL era the known HMXBs in our Galaxy were mostly Be X-ray binary systems and only a few supergiant systems were known as persistent but variable X-ray emitters. The INTEGRAL satellite has changed our understanding of binary systems, showing the existence of a new population of compact objects orbiting around a massive supergiant star, exhibits unusual properties, being either extremely absorbed, or characterized by very short and intense flares.

In the scenarios proposed by Chaty 2008, for the classical supergiant X-ray binaries, the neutron star is orbiting at a few stellar radii only from the star. The absorbed supergiant X-ray binaries, such as IGRJ16318−4848 (Filliatre & Chaty 2004), are classical supergiant X-ray binaries hosting a neutron star constantly orbiting inside a cocoon made of dust and/or cold gas.

The SFXTs, consisting at the moment of ∼14 sources, have peculiar behavior with rapid outbursts, faint quiescent emission so far detected only in a few systems and hard X-ray spectra with outburst peak luminosity of 10³⁶ erg s⁻¹. With the increasing numbers discovered in the last 7 years it is now established that they can be divided into at least two different groups according to their L_min/L_max value, duration, and frequency of outburst: classical and intermediate SFXTs (for review, see Chaty 2010). In the classical SFXT, such as IGR J17544−2619 (Sidoli et al. 2009, 2010), the neutron star orbits on a large and eccentric orbit around the supergiant star, exhibits some recurrent and short transient X-ray flares, and accretes from clumps of matter coming from the wind of the supergiant. In intermediate SFXT systems, as IGR J18483−0311 (Sguera et al. 2007; Romano et al. 2010), the neutron star orbits on a small and circular/eccentric orbit, and the accretion takes place and the source is active in the X-ray band only when the neutron star is close to the supergiant star. Because it is passing through more diluted medium, the ratio between the outburst and quiescent luminosity is higher for classical SFXTs than for intermediate SFXTs. In particular, the typical ratio between quiescence and outburst flux is less than 20 both in standard HMXBs and in the absorbed systems, while is higher than 100 in SFXTs.

Although this scenario seems to describe quite well the characteristics currently seen in supergiant X-ray binaries, it very important to identify the nature of many more supergiant X-ray binaries to confirm this scenario and characterize their transient behavior to properly model their behavior.

IGR J16328−4726 was below the sensitivity of the IBIS instrument from 2003 to 2009 with a low quiescent emission (<2.5 × 10⁻¹² erg s⁻¹ cm⁻²), then occasionally it underwent fast transient activity on different timescale: from one hour to several hours and to several days. When the source becomes active its flux increased by a factor of more than ∼165 in both 0.3–10 keV and 20–50 keV bands. This is a typical behavior of the SFXT sources (Grebenev 2009; Sguera et al. 2005, 2006; Smith et al. 2006; Negueruela et al. 2006; Chaty 2010). Like many SFXTs, IGR J16328−4726 does not show any regularity in the outbursts, although the IBIS light curve discontinuity and the vicinity of the transient bright source 4U1630−472 prevent us from studying in detail the outburst periodicity.

Although the nature of IR counterparts of IGR J16328−4726 has not yet been identified, the transient and recurrent nature of the source, the spectral properties observed during the outbursts lasting only for few tens of minutes/hours, the Galactic plane location, and the faint quiescent emission with a dynamic range >170 in the 0.3–10 keV and >165 in the 20–50 keV all point to IGR J16328−4726 being a member of the class of SFXTs.

We note that the outburst durations for IGR J16328−4726 are shorter than the typical outburst times (∼1 day) observed for other SFXTs. Assuming a distance of 10 kpc, the luminosities are 2.3 × 10³⁶ erg s⁻¹, 4.0 × 10³⁶ erg s⁻¹, and 5.3 × 10³⁵ erg s⁻¹ for revolutions 287, 768, and 351, respectively. As reported in Grebenev & Sunyaev (2010) for the SFXT IGR J18462−0223, the short outburst duration could be a characteristic feature of the source IGR J16328−4726 or an observational effect if the source is more distant than other SFXTs.

It is very important to study this class revealed by INTEGRAL to define the dominant mechanism, explain the observed X-ray and hard X-ray behavior and confirm the scenario describing the normal SFXT and intermediate ones and finally to understand the formation and evolution of the binary systems.

The authors acknowledge the ASI financial support via ASI/INAF grants I/008/07/0/.

IGR J16328−4726: A NEW CANDIDATE SUPERGIANT FAST X-RAY TRANSIENT

Article metrics

Permissions

Author affiliations

Dates

ABSTRACT

1. INTRODUCTION