Open Access
Issue
A&A
Volume 674, June 2023
Article Number A100
Number of page(s) 10
Section Astrophysical processes
DOI https://doi.org/10.1051/0004-6361/202346284
Published online 08 June 2023

© The Authors 2023

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication.

1. Introduction

IGR J17407−2808 (hereafter: J17407) is an X-ray binary discovered by the International Gamma-Ray Astronomy Laboratory (INTEGRAL) in 2004 during a period of enhanced emission activity that ended with a bright and fast flare that achieved a peak flux of ∼9.5 × 10−9 erg cm−2 s−1 (20−60 keV) and lasted for about one minute (Kretschmar et al. 2004; Götz et al. 2004; Sguera et al. 2006).

J17407 has been observed so far by several past and currently operating X-ray facilities, including ROSAT, INTEGRAL, Chandra, Swift, and XMM-Newton (Sidoli et al. 2001; Kretschmar et al. 2004; Ducci et al. 2010; Mereminskiy et al. 2020; Tomsick et al. 2008; Heinke et al. 2009; Romano et al. 2011, 2016).

The object is frequently observed in a low emission state characterised by a 2−10 keV flux that varies from ∼1.7 × 10−13 erg cm−2 s−1 to ∼2 × 10−12 erg cm−2 s−1. In almost all pointed X-ray observations performed during the low emission state (endowed with typical exposure times of ∼10−20 ks), at least one short (50–400 s) and weak (FX[2 − 10 keV]≈1 − 8 × 10−12 erg cm−2 s−1) flare has been detected. Brighter flares are rare, but in some cases peak fluxes as high as FX[20 − 60 keV]≈10−9 − 10−8 erg cm−2 s−1 have been measured. The ‘activity duty cycle’ (fraction of time spent by the source above ∼2 × 10−10 erg cm−2 s−1 in 20−40 keV) is therefore estimated to be ≈0.05% (Ducci et al. 2010; Romano et al. 2014).

In the soft X-ray domain (0.3−10 keV), the X-ray spectrum of J17407 is usually well fit with a simple absorbed power law model. The hint of a curvature in the hard part of the source spectrum was reported by Romano et al. (2016) based on simultaneous Swift/X-ray Telescope (XRT) and Burst Alert Telescope (BAT) data (∼0.3 − 70 keV) collected during a bright flaring event (the measured unabsorbed flux was of FX[0.5 − 100 keV]≈3.6 × 10−9 erg cm−2 s−1). The curvature was modelled using an exponential cutoff with parameters E f = 14 + 8 4 $ E_{\mathrm{f}} = 14{+8\atop -4} $ keV, E c = 20 + 6 20 $ E_{\mathrm{c}}=20{+6 \atop -20} $ keV (90% c.l.).

The X-ray observations of J17407 showed that absorption column density (NH) in its direction displays significant variability, ranging from ∼1022 cm−2 to ∼4.8 × 1023 cm−2, and uncorrelated with the flux.

Shortly after the discovery of J17407, Sguera et al. (2006) proposed that this source is one of the so-called supergiant fast X-ray transient (SFXT) based mainly on the spectral properties of the hard X-ray emission and the high flux variability.

SFXTs are a subclass of high-mass X-ray binaries (HMXBs) hosting a compact object (likely a neutron star) that is accreting from the stellar wind of a massive OB supergiant. These objects show typical dynamic ranges of ∼104 − 105 (up to 106, Romano et al. 2015) and sporadic flares with durations of ∼103 s. The recorded peak X-ray luminosities during these events are usually of 1036 − 1037 erg s−1 (Romano 2015; Sidoli 2017; Romano et al. 2023). Several accretion mechanisms have been proposed to explain the fast and strong variability of SFXTs. These include gating mechanisms, settling accretion regimes, and accretion of inhomogeneous winds from the donor star (e.g., in’t Zand 2005; Grebenev & Sunyaev 2007; Bozzo et al. 2008; Ducci et al. 2010; Shakura et al. 2014).

The identification of CXOU J174042.0−280724 as the Chandra counterpart to J17407 (Romano et al. 2011; Heinke et al. 2009) enabled the association of the X-ray source with an infrared object (Greiss et al. 2011; Kaur et al. 2011). This was studied in detail by Greiss et al. (2011) using the 4 m Visible and Infrared Survey Telescope for Astronomy (VISTA) at Paranal Observatory. The J2000 coordinates of the optical counterpart are RA = 17:40:42.0168, Dec = −28:07:25.050, with an astrometric fit rms of 0.1 arcsec. The International Astronomical Union (IAU) identifier of this object is VVV J174042.01−280725.05. Assuming extinction values based on the properties of red clump giants and the VISTA Variables in the Via Lactea (VVV) survey data (Gonzalez et al. 2011), Greiss et al. (2011) classified this star as a late type-F dwarf at d ≈ 3.8 kpc. By exploiting archival data collected by the instrument Son OF ISAAC (SOFI) mounted on the ESO-New Technology Telescope (NTT), Kaur et al. (2011) was able to prove that VVV J174042.01−280725.05 underwent a high luminosity event (of about one magnitude brighter than in the normal luminosity state) about four days after the detection of X-ray flares caught by Swift. This was interpreted as irradiation of the optical star by the X-rays from the compact object (Romano et al. 2016). The optical data were clearly pointing to J17407 being a low-mass X-ray binary (LMXB) rather than an HMXB, thus excluding the previously proposed association of the source with the SFXT class.

LMXBs are binary systems in which a neutron star (NS) or a black hole (BH) accretes matter from a low-mass donor star with a typical mass of ≲1 M. The accretion process occurs more commonly via an accretion disc but there are a number of objects where the accretion onto the compact object takes place directly from a stellar wind (see e.g., Done et al. 2007; Bahramian & Degenaar 2023, and references therein). LMXBs typically have X-ray luminosities in the range of 1033 − 1038 erg s−1, with some sources exhibiting outbursts that can last weeks or months and bursts in addition to these that can increase the luminosity by up to two orders of magnitude for short periods of time (∼1 − 100 s). Most LMXBs also show a dramatic X-ray spectral variability that is often associated with changes in the mass accretion rate (see e.g. van der Klis 2006; van Paradijs et al. 1998). Type-I bursts, displayed by some LMXBs, are rapid increases in X-ray luminosity caused by thermonuclear explosions that occur on the surface of an accreting NS in a LMXB. Romano et al. (2016) ruled out the possibility that J17407 belongs to the subclass of LMXBs called burst-only sources because the properties of the flares observed from J17407 substantially differ from the type-I bursts typical of the members of this subclass. We remind that burst-only sources show type-I bursts during X-ray luminosity states that do not exceed ∼1035 − 1036 erg s−1. This is at variance with the typical behaviour of other LMXBs, where bursts show up during persistent emission or outbursts brighter than 1036 erg s−1 (see e.g., Cornelisse et al. 2004; Campana 2009). Regarding J17407, the classification as a burst-only source was previously proposed by Sguera et al. (2006) as an alternative to the SFXT hypothesis. Based on the X-ray spectral and variability properties of J17407, Romano et al. (2016) also excluded that this object could be a very faint X-ray transient (VFXT), a subclass of LMXBs whose members show faint outbursts with peak X-ray luminosities two-three orders of magnitude lower than those of other LMXBs (in the range of Lx ≈ 1034 − 1036 erg s−1). About 30% of VFXTs are also known to display type-I bursts (Shaw et al. 2017; King & Wijnands 2006; Del Santo et al. 2010).

In an attempt to clarify the nature of this object, in this work we report the remarkable flaring activity detected for the first time from J17407 by NuSTAR during an observation performed in 2022. We also analyse the first source broadband spectrum (∼0.2–60 keV) obtained by combining the NuSTAR data with a quasi-simultaneous observation carried out with XMM-Newton during the persistent low-luminosity state of the source. We discuss the results obtained from the X-ray data in light of the revised classification of J17407 and complement this discussion with a comprehensive investigation of the different spectral type possibilities for the donor star hosted in this system, exploiting archival photometric data.

2. Data analysis

2.1. NuSTAR

The Nuclear Spectroscopic Telescope Array (NuSTAR) satellite launched on June 13, 2012, hosts two identical co-aligned telescopes equipped with the focal plane modules FPMA and FPMB, and operates in the 3−79 keV energy band (Harrison et al. 2013). NuSTAR observed J17407 on 13 and 14 September 2022, for a net exposure time of about 55 ks (see Table 1). We reduced the data using NUSTARDAS v2.1.2 included in HEASOFT v6.30.1 and the calibration files distributed with the CALDB v20221019 (Madsen et al. 2022). For the source, we extracted events from a circular region centred on the source, with a radius of 24″ for both the FPMA and FPMB. We tested the effect of different extraction regions for the background on the final spectral results. The final prescription adopted to maximise the signal-to-noise ratio (S/N) of the data involved regions located on the same detector of the source (Det-0) but in a zone of the focal plane free from the emission of J17407. As the observation was affected by stray-light contamination, we visually verified that all considered regions for the background extraction were characterised by a level of stray-light contamination similar to that affecting the source. In particular, for the FPMA we selected an elliptical extraction region centred on RA: 17:40:52.49, Dec: −28:06:25.6 (J2000), with a major (minor) axis of rmax = 91.283″ (rmin = 33.178″) and a rotation angle of ϕ = 346.79° relative to the RA axis. For the FPMB, we instead adopted a circular extraction region centred on RA: 17:40:36.61, Dec: −28:05:25.0, and characterised by a radius of r = 65″ (see Fig. 1).

thumbnail Fig. 1.

NuSTAR images (3–79 keV) of the J17407 field. Top panels: images for the modules A (left) and B (right) obtained using the entire observation. Bottom panels: images obtained during the first part of the observation (t < 59836.372534 MJD), where J17407 is faint and its flux constant (see Fig. 2). The red circle shows the extraction regions centred on the target. Magenta ellipses and circles show the extraction regions used for the background. Colorbars in the bottom show the count rate for each pixel.

Table 1.

Summary of the X-ray observations.

2.2. XMM-Newton

The X-ray Multi-Mirror Mission (XMM-Newton) was launched on December 10, 1999 (Jansen et al. 2001). It hosts the European Photon Imaging Camera (EPIC) that comprises the pn, Metal Oxide Semi-conductor 1 and 2 (MOS1 and MOS2) CCDs, operating in the 0.2−12 keV energy band (Strüder et al. 2001; Turner et al. 2001). XMM-Newton observed J17407 on 13 and 14 September 2022 (see Table 1). We reduced the data using the XMM-Newton Science Analysis System (SAS v20.0.0), with the latest calibration files available in the XMM-Newton calibration database (CCF). Calibrated event lists for the pn, MOS1, and MOS2 were obtained from the raw data files exploiting the SAS tasks epproc and emproc. For the pn, we used single – and double-pixel events, while for the MOS data we used single – to quadruple-pixel events. We excluded time intervals where the background was too high to perform a meaningful spectral analysis using standard procedures and criteria1. The net exposure time obtained for the XMM-Newton observation is reported in Table 1. For each of the pn, MOS1, and MOS2 cameras, we extracted the source events using circular regions centred on the best known position of J17407. The radii of these extraction regions were rpn = 11″, rMOS1 = 14″, and rMOS2 = 17″, respectively. These were derived by the SAS task eregionanalyse to maximise the S/N. Background events were accumulated for each of the three cameras using extraction regions not contaminated by the emission from J17407. The extraction regions were a circle for pn (centred on RA: 17:40:36.5703, Dec: −28:07:32.127 and with a radius of r = 38.5″) and an annulus for both the MOS1 and MOS2 (the annuli were characterised by r in MOS 1 = 28 $ r_{\mathrm{in}}^{\mathrm{MOS1}}=28^{\prime\prime} $, r out MOS 1 = 48 $ r_{\mathrm{out}}^{\mathrm{MOS1}}=48^{\prime\prime} $ for the MOS1 and r in MOS 2 = 34 $ r_{\mathrm{in}}^{\mathrm{MOS2}}=34^{\prime\prime} $, r out MOS 2 = 55 $ r_{\mathrm{out}}^{\mathrm{MOS2}}=55^{\prime\prime} $ for the MOS2). For a better alignment with the NuSTAR spectra, we applied corrections to the effective area of pn, MOS1, and MOS2 spectra in accordance with the CCF Release Note XMM-CCF-REL-3882.

3. Results

3.1. Variability and timing analysis

Figure 2 shows the background-subtracted NuSTAR light curve in the energy range 3−60 keV. In the first part of the light curve (t < 59836.372534 MJD), J17407 is barely detected in FPMB because of stray light contamination (see Fig. 1). To increase the S/N, for the first part of the light curve we considered only data from FPMA, which are less affected by the stray light. To bring out the high dynamic range on a short scale experienced by J17407, to reduce as much as possible the bins with upper limits, and to avoid smearing out the variability of the fast flares, the light curve is binned using the optimal segmentation technique based on the Bayesian block representation described in Scargle et al. (2013). The algorithm proposed by Scargle et al. (2013) splits a list of photon arrival times into an optimal maximum number of blocks such that within each block the arrival times of the photons can be described by a Poisson distribution from a constant rate, and adjacent blocks are statistically different. We adopted a relatively large false-positive rate probability (the probability of erroneously reporting the presence of a change point in the data; it is used to compute the prior on the number of blocks) of p = 0.1 so as to be sensitive to the fast rate fluctuations displayed by the flares. Once the optimal segmentation of the data was obtained, we calculated, for each bin, the rate and its error using the standard tool for the analysis of NuSTAR data nuproducts. The flux conversion to obtain the right y-axis of Fig. 2 was obtained adopting the spectral model of the low flux state (see Sect. 3.2). Figure 2 shows numerous flares clustered in a time interval of ∼40 ks, with a dynamic range of up to ∼2 × 103. The durations of the flares are ∼1 − 100 s, and they can show single or multi-peak structures. The absorbed X-ray luminosity (3−60 keV) during the first part of the observation, where flares were absent (black points in Fig. 3), is L X 1.7× 10 33 d 4 2 $ L_{\rm X} \approx 1.7\times 10^{33}d^2_4 $ erg s−1 (d4 is the distance in units of 4 kpc). During the second part (blue points in Fig. 3), the average inter-flare luminosity slightly increases to L X 5× 10 33 d 4 2 $ L_{\rm X} \approx 5\times 10^{33}d^2_4 $ erg s−1. The maximum luminosity reached by J17407 during the flaring activity (red points in Fig. 3) is L X 2.6× 10 36 d 4 2 $ L_{\rm X} \approx 2.6\times 10^{36}d^2_4 $ erg s−1.

thumbnail Fig. 2.

NuSTAR light curve in the energy range 3−60 keV obtained through the Bayesian block segmentation method. Top panel: light curve over the entire observation period. Bottom panels: three zoomed-in sections of the light curve (corresponding to the three grey-shaded areas in the top panel) to better show the typical time duration and structures of the flares. The ‘holes’ among bins, especially noticeable in the low-luminosity state of the top panel, arise from the passage of NuSTAR through the South Atlantic Anomaly regions.

thumbnail Fig. 3.

XMM-Newton and NuSTAR light curves and HRs. Panel a: XMM-Newton/pn light curve (0.2−12 keV). Panel b: NuSTAR light curve (3−60 keV). Panel c: NuSTAR light curve in the energy band 3−9 keV. Panel d: NuSTAR light curve in the energy band 9 − 60 keV. Panel e: NuSTAR HRs as a function of time. Panel f: NuSTAR HRs as a function of rate. The first part of the light curve, where the flares are absent, is displayed in all panels with black points (diamonds for XMM-Newton, squares for NuSTAR). Flares: Red triangles. Inter-flares: Blue circles.

Figure 3 shows the NuSTAR hardness ratios (HRs) as a function of time and flux (panel e). The HRs are calculated as HR = (H − S)/(H + S), where S is the rate in the soft energy band 3 − 9 keV, and H is the rate in the hard energy band 9 − 60 keV. The boundary at 9 keV is chosen to ensure a similar number of average counts in S and H. In general, Fig. 3 does not show dramatic HR variability, although a first visual inspection suggests a possible small variability in panel f (i.e., HRs vs. flux). A more thorough exploration of the spectral variability is presented in Sect. 3.2.

Panel a of Fig. 3 shows the background-subtracted XMM-Newton/pn light curve in the energy range 0.2−12 keV, with a bin size of 1 ks. The HRs based on the XMM-Newton data, in the energy bands 0.2−3 keV and 3−12 keV, do not show significant variability in hardness.

We searched for periodic signals in the 0.3−12 keV pn events corrected to the Solar System barycentre with the barycen SAS task and in the 3−60 keV NuSTAR events corrected to the Solar System barycentre with the barycorr task from ftools, using a Rayleigh test Z2 (see e.g., Buccheri et al. 1983) from 1 to 3 harmonics. No statistically significant pulsations were detected in the frequency range 0.001 − 175.44 Hz for XMM-Newton/pn and 0.001 − 1000 Hz for NuSTAR. We used the approach described in Brazier (1994) to calculate the 3σ upper limit on the pulsed fraction pf of a sinusoidal signal. We found, for XMM-Newton/pn, pf = 50%, and for NuSTAR, pf = 40%.

3.2. Spectral analysis

We divided the data into three subsets to search for weak spectral variability not detectable with the hardness ratios. We divided the data using the same scheme shown in Fig. 3:

  • 1) first part (t < 59836.372534 MJD), where the flux is low and flares are absent; these data were fitted simultaneously with the XMM-Newton data (black points in Fig. 3);

  • 2) low: inter-flare emission (rate < 0.15) observed by NuSTAR after t > 59836.372534 MJD (blue points in Fig. 3);

  • 3) high: flares (rate > 0.15) observed by NuSTAR after t > 59836.372534 MJD (red points in Fig. 3).

NuSTAR and XMM-Newton spectra were rebinned so as to have at least 25 counts per bin to enable the use of the χ2 statistic as a fit statistic. Renormalisation constant factors were included in the spectral fitting to account for intercalibration uncertainties between instruments. We used the tbabs model and the interstellar medium abundances wilm in XSPEC3 to model the photoelectric absorption (Wilms et al. 2000). Errors on spectral fit parameters indicate 1σ confidence level throughout the paper. We fitted the spectra with different models. In the following, we concentrate on the simplest phenomenological models that provide a reasonably good fit. The best-fit models and parameters are reported in Table 2. The corresponding spectra and residuals are shown in Figs. 4, 5, and 6. For the first part, we obtained a good fit with an absorbed power law ( χ ν 2 $ \chi^2_\nu $ = 0.963, 46 d.o.f.). For the low state, the same model gives an acceptable fit ( χ ν 2 $ \chi^2_\nu $ = 1.125, 39 d.o.f.), although we obtained a slightly better fit with an absorbed power law with high-energy cutoff ( χ ν 2 $ \chi^2_\nu $ = 0.944, 37 d.o.f.). For the high state, it is necessary to employ a more complex model to describe the residuals at ≈8.5 keV (see bottom panel of Fig. 6). We obtained the best fits with an absorbed power law with a Gaussian in absorption E gabs = 8.6 + 0.6 0.5 $ E_{\mathrm{gabs}}=8.6{+0.6\atop -0.5} $ keV ( χ ν 2 $ \chi^2_\nu $ = 1.240, 86 d.o.f.), or a power law with a high-energy cutoff plus a steep power law, both components absorbed ( χ ν 2 $ \chi^2_\nu $ = 1.221, 85 d.o.f.). For the high state, we evaluated the chance probability of improvement of the fit by adding the high-energy cutoff, simulating 104 data sets with the simftest routine of XSPEC. We find that the probability that data are consistent with a model without the high-energy cutoff component is 0.30%. The best-fit model obtained for the high state data applied to the low state4 gives χ ν 2 $ \chi^2_\nu $ = 0.773, 41 d.o.f. Therefore, the relatively low statistic of the low state spectrum does not allow us to determine whether or not there is spectral variability between these two flux states. A similar fit applied to the first part of the data gives an unacceptable fit, even if the column density is free ( χ ν 2 $ \chi^2_\nu $ = 8.568, 48 d.o.f.). Vice versa, the best-fit model obtained for the first part of the data applied to the low state5 gives an acceptable fit only if the column density is free ( χ ν 2 $ \chi^2_\nu $ = 1.248, 40 d.o.f.). However, the absorption column density we obtain, NH ≈ 9 × 1023 cm−2, seems too high compared to the typical values in this source and other typical X-ray binaries. A similar fit applied to the high state gives an unacceptable χ ν 2 $ \chi^2_\nu $ = 2.629, 90 d.o.f., and a very high NH ≈ 1024 cm−2. Therefore, it seems that there is significant spectral variability, especially driven by the increase in the column density between the first and second (low and high) parts.

thumbnail Fig. 4.

XMM-Newton (black: pn; red: MOS1; green: MOS2) and NuSTAR (blue: module A) spectra of J17407 during the first part of the light curve (black points in Fig. 3), fitted with an absorbed power law (see Table 2). The lower panel shows the residuals of the fit.

thumbnail Fig. 5.

NuSTAR (blue circles: module A; red triangles: module B) spectra of J17407 during the low luminosity state between flares (blue circles in Fig. 3). Top panel: spectra are fitted with an absorbed power law with a high-energy cutoff. Middle panel: residuals of the same fit of the top panel. Bottom panel: residuals of the fit of the spectra with an absorbed power law. See Table 2 for the best-fit parameters.

thumbnail Fig. 6.

NuSTAR (blue circles: module A; red triangles: module B) spectra of J17407 during the high luminosity state (i.e., flares; red triangles in Fig. 3). Top panel: spectra are fitted with two power law, both absorbed, one of them with a high-energy cutoff. Middle panel: residuals of the same fit of the top panel. Bottom panel: residuals of the fit of the spectra with an absorbed power law. The purpose of this panel is to highlight the presence of a feature in the residuals at ∼8.6 keV, which requires the use of a more complex model, as described in Sect. 3.2. See Table 2 for the best-fit parameters, also with other spectral models.

Table 2.

Best-fit spectral parameters to describe the first part (black points in Fig. 3), low luminosity state (blue points in Fig. 3), and high luminosity state (i.e., flares; red points in Fig. 3).

The low-state spectrum from module B shows an emission feature at ∼6.4 keV, which could be an iron line (Fig. 5). To understand whether or not it is necessary to add a Gaussian component to model it, we rebinned the spectra of modules A and B to have a minimum of 1 count per bin and used the W statistic (Wachter et al. 1979; Cash 1979) to find the best fit. We then evaluated the chance probability of improvement of the fit by adding this component by simulating 104 data sets with the simftest routine of XSPEC. We find that the probability that the data are consistent with a model without the Gaussian component is 0.13%. Therefore, the hypothesis of the absence of an iron line cannot be rejected at a 4σ confidence level.

4. Optical counterpart

We assessed the nature of the donor star in J17407 using photometric measurements of VVV J174042.01−280725.05, based on the Panoramic Survey Telescope and Rapid Response System (Pan-STARRS) catalogue (Chambers et al. 2016) and on the VVV Survey. The survey consists of images in five bands obtained by the VISTA InfraRed CAMera (VIRCAM) that equips the VISTA, a 4 m telescope designed for use in wide field surveys. The photometric information used to build the spectral energy distribution (SED) of VVV J174042.01−280725.05 is reported in Table 3. The star is very faint and no Gaia DR3 counterpart is catalogued within the VISTA 3σ error region (Greiss et al. 2011; Gaia Collaboration 2022). Given the small number of points available in the SED, we fitted it with a simple absorbed blackbody model, which is in first approximation a reasonable model to describe the optical/near-infrared spectrum from a star:

λ F ( λ ) = 2 π h c 2 λ 4 [ 1 e h c / ( λ k T ) 1 ] R 2 d 2 10 0.4 A λ , $$ \begin{aligned} \lambda F(\lambda ) = \frac{2\pi hc^2}{\lambda ^4}\left[ \frac{1}{e^{hc/(\lambda kT)}-1} \right] \frac{R^2}{d^2} 10^{-0.4A_\lambda }, \end{aligned} $$(1)

Table 3.

Energy fluxes of the optical counterpart of J17407.

where T is the effective temperature of the star, R its radius, d is the distance from the Sun, and Aλ is the absorption for the wavelength λ. We calculated Aλ for each λ reported6 in Table 3 using the analytical expression of Cardelli et al. (1989) and O’Donnell (1994), and assuming Rv = 3.1. A close look at Eq. (1) shows a degeneracy between parameters R and d. In addition, due to the limited number of points of our SED, there is also some degeneracy between Av and T. Therefore, only two parameters out of four can be constrained simultaneously.

We used two approaches to overcome these limitations. First, we fitted the SED with Eq. (1) for assumed distances in the range 1–30 kpc, with steps of 1 kpc, and we also fixed the extinction Av assuming the lowest absorption column density value; this was measured in a XMM-Newton observation carried out on 6 March 2016 (obsid: 0764191301). The value of NH reported by Romano et al. (2016; N H = 0.77 + 0.70 0.48 × 10 22 $ N_{\mathrm{H}}=0.77{+0.70\atop-0.48}\times 10^{22} $ cm−2, 90% c.l. uncertainties) was based on the use of phabs absorption model in XSPEC, with the abundance of elements by Anders & Grevesse (1989). In environments with high column density, these choices for the absorption model and abundances could give significantly different results compared to using the tbabs model with wilm abundances. Therefore, we reanalysed these data using SAS 20.0 and tbabs with wilm abundances, which is coherent with the analysis method described in Sects. 2.2 and 3.2. We obtain NH = 1.2  ±  0.4  ×  1022 cm−2 (1σ c.l.). We converted NH to Av using the relation given in Foight et al. (2016). The best-fit parameters T and R are displayed in Fig. 7: they consist of three horizontal lines, that is, for the mean, upper, and lower limits of Av (given the uncertainties on NH). A grid of distances in black is overplotted. The colours of these horizontal lines follow the scheme of the Av colour bar. The best-fit values of T and R obtained with this method have χ red 2 $ \chi^2_{\rm red} $ ranging from 1.19 to 2.18, with χ red 2 $ \chi^2_{\rm red} $ = 1.63 (8 d.o.f.) for the solutions corresponding to Av = 4.2 (NH = 1.2 × 1022 cm−2; horizontal line with orange colour in Fig. 7). This method does not take into account that Av increases with d and, consequently, solutions for many of the assumed distances are obviously wrong. Therefore, we adopted a second approach, in which Av is a function of d. We used the directional 3D maps of interstellar dust reddening and extinction bayestar2019 based on Gaia, Pan-STARRS 1, and Two Micron All Sky Survey (2MASS; Green et al. 2019). Reddening was converted to Av using the appropriate conversion described in Green et al. (2019)7, and adopting the 5%–95% percentile boundaries. The range of reliable distances of these maps in the direction of our target is 1.59 − 9.40 kpc8. The best-fit parameters T and R for different values of d and Av(d) are shown in Fig. 7 (“Γ” shape). The distances are overplotted in red. Figure 7 also shows the most relevant spectral classes for our analysis, which are taken from de Jager & Nieuwenhuijzen (1987) and Pecaut & Mamajek (2013). All the best-fit parameters obtained with the second method and shown in Fig. 7 are from fits that give χ red 2 $ \chi^2_{\rm red} $ < 2. We obtain the best fit ( χ red 2 $ \chi^2_{\rm red} $ = 1.06) for a star at a distance of ∼1.66 kpc, with temperature T ≈ 4 × 103 K and radius R ≈ 9 × 1010 cm. These values of T and R and those obtained assuming Av = 4.2 ± 1.4 (based on the lowest NH = (1.2 ± 0.4)×1022 cm−2 from X-ray observations) correspond to a relatively peculiar ∼K4-M7 sub-subgiant type star.

thumbnail Fig. 7.

Best-fit values of effective temperature, radius, and extinction obtained from the fit of the photometric SED of J17407 (see Table 3) with an absorbed blackbody model, with two different methods. Method one: best-fit values of T and R from the fit of the photometric SED, assuming distances in the range 1–30 kpc and for three values of Av (see main text for more details). These solutions show up as three horizontal lines whose colours reflect the values assumed for Av (see vertical colour bar on the right). A grid of distances (in units of kpc) in black is overplotted. Method two: “Γ-shape” coloured area shows the best-fit parameters ( χ red 2 $ \chi^2_{\rm red} $ < 2) T and R for different values of d, where the dependency of Av on the distance is based on Green et al. (2019). For this method, the distances (in units of kpc) from 1.6 to 9.4 kpc are overplotted in red. The most relevant spectral classes for our analysis are shown in solid (main sequence), dashed (sub-giant), and dot-dashed (giant) red lines. The inset figure shows an example of a best-fitting SED (dashed red line). Blue points are the photometric measurements (see Table 3). The fit residuals (observed-model)/error are shown in the lower panel.

The coloured area in Fig. 7 shows that other classifications are possible, indicating that the spectral type of the donor star of J17407 is poorly constrained.

It is important to point out that the spectral classification study presented here is to be taken with due caution. We recall that the near-infrared variability observed by Kaur et al. (2011) could indicate that at least on some occasions (when the source is brighter) there is a significant contribution to the observed near-infrared emission produced by the reprocessing of the X-ray emission from the compact object on an accretion disc (whose presence is not certain) or on the surface of the companion star (Romano et al. 2016). If this is the case, considering the magnitude values when the system is faint should have minimised the impact of reprocessing on the optical near-infrared emission. Nevertheless, should this still be present, it would mean that the donor star of J17407 is even weaker than has been measured, further supporting the LMXB scenario with a particularly light and small donor star. Due to the lack of measurements of the fundamental parameters of the system, in particular its distance from the Sun and the orbital separation, it is not possible to unambiguously quantify the contribution of reprocessing to the observed emission, and therefore it is not possible to determine whether or not this significantly affects the result reported in this paper.

5. Discussion

NuSTAR detected J17407 in a flaring state characterised by a variability as large as three orders of magnitude on timescales of a few tens of seconds. If we consider the lowest and highest fluxes measured for this source since its discovery, they span more than four orders of magnitude.

In Sect. 3 we show that one of the possible spectral models adopted to describe the source emission during the flaring period is a power law subjected to a large extinction at the lower energies (≲3 keV) due to a high column density and featuring a Gaussian absorption feature centred at Egabs ≈ 8.6 keV. Similar absorption features are often detected in the X-ray spectra of accreting NSs and widely interpreted as cyclotron resonant scattering features (CRSFs; also collectively termed ‘cyclotron lines’). CRSFs provide a direct measurement of the NS magnetic field strength close to the surface of the compact object through the simple relation Ecyc ≈ 11.6B12 keV (see e.g., Staubert et al. 2019, for a recent review). For the case of J17407, we would therefore obtain a magnetic field of B ≈ 7 × 1011 G. If confirmed by future observations, this would imply that J17407 hosts a NS with a relatively strong magnetic field, which is uncommon in LMXBs (but see the cases of GRO J1744−28 and 4U 1822−371 for examples of strongly magnetised NSs in LMXBs, as well as the cases of IGR J17329−2731 and 4U 1700+24 for strongly magnetised NSs in SyXBs; D’Aì et al. 2015; Anitra et al. 2021; Jonker et al. 2003; Bozzo et al. 2018, 2022b). We note that, as discussed in Sect. 3.2, the available NuSTAR data on J17407 can also be successfully fit with alternative models that do not comprise a CRSF. Therefore, the presence of this absorption component and the derivation of a putative NS magnetic field estimate have to be taken with caution (and possibly confirmed by future observations). The spectral analysis also shows a weak hardening and increase of NH between the first part of the observation, where the flaring activity is absent, and the second part. The increase in the absorption might reflect an increase in the gravitationally captured mass by the compact object. The hardening could be ascribed to a more efficient inverse Compton scattering of the soft X-ray photons emitted in the vicinity of the compact object by the increasing number of accreting electrons.

Our limited knowledge on the fundamental characteristics of the stellar components hosted in J17407, such as the orbital period, distance between the two stars, eventual spin period, and magnetic field strength of the compact object, prevents us from adopting a quantitative approach to determine the true nature of this source and the accretion mechanisms triggering its X-ray variability. From the X-ray spectral point of view, the NuSTAR and XMM-Newton data would be qualitatively compatible with what is expected from accreting pulsars with high magnetic field (B ≳ 1012 G; e.g., Kretschmar et al. 2019). In the following, we therefore discuss the similarities and differences between the properties of the X-ray variability shown by J17407 with those from the other binary systems, specifically those hosting strongly magnetised NSs. Nevertheless, we are not excluding a priori that the compact object could be a BH.

With the available photometric data, the spectral type of the optical counterpart is poorly constrained. We showed in Sect. 4 that, among the other possibilities, J17407 could belong to the class of symbiotic stars, with the donor star being a peculiar M-type sub-subgiant star. Although symbiotic stars show X-ray flares, these are longer than in J17407 (e.g., for GX 1+4 ΔLX≈ days, and for 3A 1954+3199Δt ≈ 104 s), and the dynamic range is smaller (ΔLX ≲ 60; Corbet et al. 2008; Bozzo et al. 2022a).

Our study (Sect. 4) and the previous classification of the optical counterpart (Greiss et al. 2011; Kaur et al. 2011) suggest that J17407 might be a LMXB, with a M, F, or G type star, from main sequence to giant. Among LMXBs, some accreting millisecond pulsars (AMXP; see e.g., Papitto et al. 2020) show flares with the same durations as those displayed by J17407, but their dynamic range (ΔLX ≈ 10 − 50) is lower than that observed in J17407 (see e.g., the case of the ‘hiccup’ accretion in IGR J18245−2452, Ferrigno et al. 2014).

Other X-ray binaries showing similar flares are the LMXBs ‘Bursting Pulsar’ GRO J1744−28, the ‘Rapid Burster’ MXB 1730−335, and the HMXB SMC X−1 (see e.g., Bagnoli et al. 2015; Court et al. 2018; Rai et al. 2018, and references therein). These examples exhibit type-II bursts, which typically last for a few tens of seconds (though there are some exceptions; see references above). Generally, these bursts are single peaked (although in some cases they show more complex structures) and there is no evidence of significant spectral variability between the persistent and the flaring states. At odds with J17407, these three sources show type-II bursts only during bright luminosity states where the persistent X-ray luminosity is LX ≈ 1037 − 1038 erg s−1 and the dynamic range of the flares is ΔLX ≈ 10 − 40 (e.g., Bagnoli et al. 2015; Giles et al. 1996; Sazonov et al. 1997).

Different models have been proposed to explain the type-II bursts. Among them, those that can also explain why the other NS LMXBs do not show these bursts belong to the family of the trapped disc models (Spruit & Taam 1993; D’Angelo & Spruit 2010; D’Angelo & Spruit 2012; van den Eijnden et al. 2017). In this model, for some specific values of the magnetic field strength and mass-capture rate (B ≈ 1010 − 1011 G, c roughly ∼10%–45% of the Eddington rate), the interaction between the inner region of the accretion disc and the magnetospheric boundary of the pulsar may interrupt the continuous flow and produce a cycle of accretion events (bursts).

The fluence of the type-II bursts in the Rapid Burster is proportional to the waiting time to the following burst (Lewin et al. 1976). This is not observed in the Bursting Pulsar Kouveliotou et al. (1996) and, to the best of our knowledge, in SMC X−1. Using Fig. 8, we can verify whether or not this is also the case for J17407, but we do not find any significant correlation between the fluence and the waiting time. For these calculations, we considered only flares and waiting times not interrupted by the regular gaps that characterise the NuSTAR observation, where other flares may have occurred.

thumbnail Fig. 8.

NuSTAR burst fluence (3−60 keV) as a function of the time interval between this burst and the next one.

J17407 shows some striking similarities with the LMXBs Swift J1858.6−0814, V404 Cygni, and V4641 Sgr. These sources show flares spanning a few orders of magnitude (∼102 − 103) on timescales of ∼10 − 100 s (see e.g., Ludlam et al. 2018; Hare et al. 2020; Wijnands & van der Klis 2000; Rodriguez et al. 2015). V404 Cygni and V4641 Sgr host BHs, while the compact object in Swift J1858.6−0814 was identified to be a NS thanks to the detection of some type I X-ray bursts (Buisson et al. 2020). Despite these similarities, these sources show the flaring activity on top of longer (weeks to months) X-ray outbursts, and the flares show a remarkable spectral variability in the X-ray spectrum, which is at odds with J17407.

The fast and strong variability of J17407 is also reminiscent of that observed by XMM-Newton from the Be/XRBs A0538−66 in 2018. This binary system, located in the Large Magellanic Cloud, hosts a ∼69 ms pulsar orbiting a Be star in Porb ≈ 16.6 days in a highly eccentric orbit (e ≈ 0.72; Ducci et al. 2022; Rajoelimanana et al. 2017). During the XMM-Newton observations of 2018, A0538−66 showed X-ray flares with durations of between 5 and 50 s, and peak X-ray luminosity of up to ∼4 × 1038 erg s−1 (0.2−10 keV). Between the flares, the luminosity was ∼2 × 1035 erg s−1. Therefore, the dynamic range and flare durations of J17407 and A0538−66 are consistent, while the luminosities of J17407 are ∼6 × 102 times lower than in A0538−66, assuming d = 2 kpc. Another similarity between these two sources is that the flares of both show a single-peak or multi-peak structure.

To explain the X-ray variability of A0538−66 observed by XMM-Newton in 2018, different plausible mechanisms were considered. In particular, a gating mechanism in a regime of spherically symmetric inflow was explored in detail. In the proposed scenario, the observed X-ray variability is linked to fast transitions between accretion and supersonic propeller regimes (Ducci et al. 2019b). These accretion regimes and the transitions between them occur under particular conditions, which are mainly driven by the spin period of the pulsar, its magnetic field strength, and the rate of the mass it gravitationally captures (Davies & Pringle 1981; Bozzo et al. 2008).

In the introduction, we reported an optical brightening of about 1 mag in J17407 (Kaur et al. 2011). Notably, A0538−66 also shows fast and bright optical flares (Δt ≲ 1 d, Δmv ≈ 0.5 − 1; rarely up to Δmv ≈ 2.2; Ducci et al. 2016 and references therein). These were explained with reprocessing of the X-ray photons from the accreting pulsar in a cloud surrounding the binary system, and it was also pointed out that it was possible to have a non-negligible contribution to the optical brightening caused by the heating of the surface of the companion star irradiated by the X-ray pulsar (Ducci et al. 2019a).

6. Conclusions

We report the results of quasi-simultaneous NuSTAR and XMM-Newton observations of J17407. These show an astonishing variability, whose properties are so extreme that they are difficult to explain with known accretion mechanisms applied to other accreting binary systems. We show that the variability of J17407 shows some similarities (and also substantial differences) to a few LXMBs, in particular Swift J1858.6−0814, and a remarkable similarity to A0538−66, which, we stress, is an HMXB. The uniqueness of the X-ray flashes displayed by J17407 merits further observations in the X-ray, optical, and infrared bands in order to unveil the nature of the stars of the system and to determine the mechanism that causes its X-ray variability. In particular, future spectroscopic observations in optical near-infrared will be fundamental to constraining the properties of the donor star.


3

XSPEC version 12.12.1c (Arnaud et al. 1996).

4

All spectral parameters frozen, except for the normalisation of the first power law, and the ratio of the normalisation of the second power law to the normalisation of the first power law.

5

All spectral parameters frozen, except for the normalization of the power law.

6

Taken from the VizieR Photometry viewer: http://vizier.cds.unistra.fr/vizier/sed/

8

Obtained using dustmaps, (Green 2018).

9

Whose donor star is an M supergiant (Bozzo et al. 2022a, and references therein).

Acknowledgments

We thank the anonymous referee for constructive comments that helped to improve the paper. LD acknowledges Dr. Marilena Caramazza and Dr. Dante Minniti for their helpful suggestions. PR and EB acknowledge financial contribution from the agreement ASI-INAF I/037/12/0. This research has made use of the NuSTAR Data Analysis Software (NuSTARDAS) jointly developed by the ASI Space Science Data Center (SSDC, Italy) and the California Institute of Technology (Caltech, USA). Based on observations obtained with XMM-Newton, an ESA science mission with instruments and contributions directly funded by ESA Member States and NASA. Part of the data analysis was performed using stingray, an open source spectral-timing Python software package for astrophysical data analysis (Bachetti et al. 2022; Huppenkothen et al. 2019a,b). This work made use of Astropy: http://www.astropy.org a community-developed core Python package and an ecosystem of tools and resources for astronomy (Astropy Collaboration 2013, 2018, 2022). It also made use of dustmaps, a unified interface for several 2D and 3D maps of interstellar dust reddening and extinction (Green 2018). This research has made use of the VizieR catalogue access tool, CDS, Strasbourg, France.

References

  1. Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim., 53, 197 [NASA ADS] [CrossRef] [Google Scholar]
  2. Anitra, A., Di Salvo, T., Iaria, R., et al. 2021, A&A, 654, A160 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  3. Arnaud, K. A. 1996, in Astronomical Data Analysis Software and Systems V, eds. G. H. Jacoby, & J. Barnes, ASP Conf. Ser., 101, 17 [NASA ADS] [Google Scholar]
  4. Astropy Collaboration (Robitaille, T. P., et al.) 2013, A&A, 558, A33 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  5. Astropy Collaboration (Price-Whelan, A. M., et al.) 2018, AJ, 156, 123 [Google Scholar]
  6. Astropy Collaboration (Price-Whelan, A. M., et al.) 2022, ApJ, 935, 167 [NASA ADS] [CrossRef] [Google Scholar]
  7. Bachetti, M., Huppenkothen, D., Khan, U., et al. 2022, https://doi.org/10.5281/zenodo.7135161 [Google Scholar]
  8. Bagnoli, T., in’ t Zand, J. J. M., D’Angelo, C. R., & Galloway, D. K. 2015, MNRAS, 449, 268 [NASA ADS] [CrossRef] [Google Scholar]
  9. Bahramian, A., & Degenaar, N. 2023, in Handbook of X-ray and Gamma-ray Astrophysics, eds. C. Bambi & A. Santangelo (Springer Living Reference Work), 120 [Google Scholar]
  10. Bozzo, E., Falanga, M., & Stella, L. 2008, ApJ, 683, 1031 [Google Scholar]
  11. Bozzo, E., Bahramian, A., Ferrigno, C., et al. 2018, A&A, 613, A22 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  12. Bozzo, E., Ferrigno, C., Oskinova, L., & Ducci, L. 2022a, MNRAS, 510, 4645 [NASA ADS] [CrossRef] [Google Scholar]
  13. Bozzo, E., Romano, P., Ferrigno, C., & Oskinova, L. 2022b, MNRAS, 513, 42 [CrossRef] [Google Scholar]
  14. Brazier, K. T. S. 1994, MNRAS, 268, 709 [NASA ADS] [Google Scholar]
  15. Buccheri, R., Bennett, K., Bignami, G. F., et al. 1983, A&A, 128, 245 [NASA ADS] [Google Scholar]
  16. Buisson, D. J. K., Altamirano, D., Bult, P., et al. 2020, MNRAS, 499, 793 [CrossRef] [Google Scholar]
  17. Campana, S. 2009, ApJ, 699, 1144 [NASA ADS] [CrossRef] [Google Scholar]
  18. Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 [Google Scholar]
  19. Cash, W. 1979, ApJ, 228, 939 [Google Scholar]
  20. Chambers, K. C., Magnier, E. A., Metcalfe, N., et al. 2016, arXiv e-prints [arXiv:1612.05560] [Google Scholar]
  21. Corbet, R. H. D., Sokoloski, J. L., Mukai, K., Markwardt, C. B., & Tueller, J. 2008, ApJ, 675, 1424 [NASA ADS] [CrossRef] [Google Scholar]
  22. Cornelisse, R., in’t Zand, J. M. M., Kuulkers, E., et al. 2004, Nucl. Phys. B Proc. Suppl., 132, 518 [NASA ADS] [CrossRef] [Google Scholar]
  23. Court, J. M. C., Altamirano, D., Albayati, A. C., et al. 2018, MNRAS, 481, 2273 [NASA ADS] [CrossRef] [Google Scholar]
  24. D’Aì, A., Di Salvo, T., Iaria, R., et al. 2015, MNRAS, 449, 4288 [CrossRef] [Google Scholar]
  25. D’Angelo, C. R., & Spruit, H. C. 2010, MNRAS, 406, 1208 [NASA ADS] [Google Scholar]
  26. D’Angelo, C. R., & Spruit, H. C. 2012, MNRAS, 420, 416 [Google Scholar]
  27. Davies, R. E., & Pringle, J. E. 1981, MNRAS, 196, 209 [CrossRef] [Google Scholar]
  28. de Jager, C., & Nieuwenhuijzen, H. 1987, A&A, 177, 217 [NASA ADS] [Google Scholar]
  29. Del Santo, M., Sidoli, L., Romano, P., et al. 2010, MNRAS, 403, L89 [NASA ADS] [Google Scholar]
  30. Done, C., Gierliński, M., & Kubota, A. 2007, A&ARv, 15, 1 [Google Scholar]
  31. Ducci, L., Sidoli, L., & Paizis, A. 2010, MNRAS, 408, 1540 [Google Scholar]
  32. Ducci, L., Covino, S., Doroshenko, V., et al. 2016, A&A, 595, A103 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  33. Ducci, L., Mereghetti, S., Hryniewicz, K., Santangelo, A., & Romano, P. 2019a, A&A, 624, A9 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  34. Ducci, L., Mereghetti, S., & Santangelo, A. 2019b, ApJ, 881, L17 [NASA ADS] [CrossRef] [Google Scholar]
  35. Ducci, L., Mereghetti, S., Santangelo, A., et al. 2022, A&A, 661, A22 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  36. Ferrigno, C., Bozzo, E., Papitto, A., et al. 2014, A&A, 567, A77 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  37. Foight, D. R., Güver, T., Özel, F., & Slane, P. O. 2016, ApJ, 826, 66 [Google Scholar]
  38. Gaia Collaboration (Vallenari, A., et al.) 2022, A&A in press, https://doi.org/10.1051/0004-6361/202243940 [Google Scholar]
  39. Giles, A. B., Swank, J. H., Jahoda, K., et al. 1996, ApJ, 469, L25 [NASA ADS] [CrossRef] [Google Scholar]
  40. Gonzalez, O. A., Rejkuba, M., Zoccali, M., Valenti, E., & Minniti, D. 2011, A&A, 534, A3 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  41. Götz, D., Mereghetti, S., Mowlavi, N., & Soldan, J. 2004, GRB Coordinates Network, 2793, 1 [Google Scholar]
  42. Grebenev, S. A., & Sunyaev, R. A. 2007, Astron. Lett., 33, 149 [NASA ADS] [CrossRef] [Google Scholar]
  43. Green, G. 2018, J. Open Source Softw., 3, 695 [NASA ADS] [CrossRef] [Google Scholar]
  44. Green, G. M., Schlafly, E., Zucker, C., Speagle, J. S., & Finkbeiner, D. 2019, ApJ, 887, 93 [NASA ADS] [CrossRef] [Google Scholar]
  45. Greiss, S., Steeghs, D., Maccarone, T., et al. 2011, ATel, 3688, 1 [NASA ADS] [Google Scholar]
  46. Hare, J., Tomsick, J. A., Buisson, D. J. K., et al. 2020, ApJ, 890, 57 [Google Scholar]
  47. Harrison, F. A., Craig, W. W., Christensen, F. E., et al. 2013, ApJ, 770, 103 [Google Scholar]
  48. Heinke, C. O., Tomsick, J. A., Yusef-Zadeh, F., & Grindlay, J. E. 2009, ApJ, 701, 1627 [NASA ADS] [CrossRef] [Google Scholar]
  49. Herpich, F. R., Ferreira Lopes, C. E., Saito, R. K., et al. 2021, A&A, 647, A169 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  50. Huppenkothen, D., Bachetti, M., Stevens, A. L., et al. 2019a, ApJ, 881, 39 [Google Scholar]
  51. Huppenkothen, D., Bachetti, M., Stevens, A. L., et al. 2019b, J. Open Source Softw., 4, 1393 [NASA ADS] [CrossRef] [Google Scholar]
  52. in’t Zand, J. J. M. 2005, A&A, 441, L1 [CrossRef] [EDP Sciences] [Google Scholar]
  53. Jansen, F., Lumb, D., Altieri, B., et al. 2001, A&A, 365, L1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  54. Jonker, P. G., van der Klis, M., & Groot, P. J. 2003, MNRAS, 339, 663 [NASA ADS] [CrossRef] [Google Scholar]
  55. Kaur, R., Heinke, C., Kotulla, R., et al. 2011, ATel, 3695, 1 [NASA ADS] [Google Scholar]
  56. King, A. R., & Wijnands, R. 2006, MNRAS, 366, L31 [NASA ADS] [Google Scholar]
  57. Kouveliotou, C., van Paradijs, J., Fishman, G. J., et al. 1996, Nature, 379, 799 [NASA ADS] [CrossRef] [Google Scholar]
  58. Kretschmar, P., Mereghetti, S., Hermsen, W., et al. 2004, ATel, 345, 1 [NASA ADS] [Google Scholar]
  59. Kretschmar, P., Fürst, F., Sidoli, L., et al. 2019, New Astron. Rev., 86, 101546P [CrossRef] [Google Scholar]
  60. Lewin, W. H. G., Doty, J., Clark, G. W., et al. 1976, ApJ, 207, L95 [NASA ADS] [CrossRef] [Google Scholar]
  61. Ludlam, R. M., Miller, J. M., Arzoumanian, Z., et al. 2018, Atel, 12158, 1 [Google Scholar]
  62. Madsen, K. K., Forster, K., Grefenstette, B., Harrison, F. A., & Miyasaka, H. 2022, J. Astron. Telesc. Instrum. Syst., 8, 034003 [NASA ADS] [CrossRef] [Google Scholar]
  63. Magnier, E. A., Schlafly, E. F., Finkbeiner, D. P., et al. 2020, ApJS, 251, 6 [NASA ADS] [CrossRef] [Google Scholar]
  64. Mereminskiy, I., Lutovinov, A., Sunyaev, R., et al. 2020, ATel, 13630, 1 [NASA ADS] [Google Scholar]
  65. Minniti, D., Lucas, D., & VVV Team 2017, VizieR Online Data Catalog: II/348 [Google Scholar]
  66. O’Donnell, J. E. 1994, ApJ, 422, 158 [Google Scholar]
  67. Papitto, A., Falanga, M., Hermsen, W., et al. 2020, New Astron. Rev., 91, 101544 [Google Scholar]
  68. Pecaut, M. J., & Mamajek, E. E. 2013, ApJS, 208, 9 [Google Scholar]
  69. Rai, B., Pradhan, P., & Paul, B. C. 2018, RAA, 18, 148 [Google Scholar]
  70. Rajoelimanana, A. F., Charles, P. A., Meintjes, P. J., et al. 2017, MNRAS, 464, 4133 [NASA ADS] [CrossRef] [Google Scholar]
  71. Rodriguez, J., Cadolle Bel, M., Alfonso-Garzón, J., et al. 2015, A&A, 581, L9 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  72. Romano, P. 2015, J. High Energy Astrophys., 7, 126 [Google Scholar]
  73. Romano, P., Barthelmy, S. D., Krimm, H. A., et al. 2011, ATel, 3685, 1 [NASA ADS] [Google Scholar]
  74. Romano, P., Ducci, L., Mangano, V., et al. 2014, A&A, 568, A55 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  75. Romano, P., Bozzo, E., Mangano, V., et al. 2015, A&A, 576, L4 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  76. Romano, P., Bozzo, E., Esposito, P., et al. 2016, A&A, 593, A96 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  77. Romano, P., Evans, P. A., Bozzo, E., et al. 2023, A&A, 670, A127 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  78. Sazonov, S. Y., Sunyaev, R. A., & Lund, N. 1997, Astron. Lett., 23, 286 [NASA ADS] [Google Scholar]
  79. Scargle, J. D., Norris, J. P., Jackson, B., & Chiang, J. 2013, ApJ, 764, 167 [Google Scholar]
  80. Sguera, V., Bazzano, A., Bird, A. J., et al. 2006, ApJ, 646, 452 [NASA ADS] [CrossRef] [Google Scholar]
  81. Shakura, N., Postnov, K., Sidoli, L., & Paizis, A. 2014, MNRAS, 442, 2325 [Google Scholar]
  82. Shaw, A. W., Heinke, C. O., Degenaar, N., et al. 2017, MNRAS, 471, 2508 [NASA ADS] [CrossRef] [Google Scholar]
  83. Sidoli, L. 2017, in XII Multifrequency Behaviour of High Energy Cosmic Sources Workshop (MULTIF2017), 52 [Google Scholar]
  84. Sidoli, L., Belloni, T., & Mereghetti, S. 2001, A&A, 368, 835 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  85. Smith, L. C., Lucas, P. W., Kurtev, R., et al. 2018, MNRAS, 474, 1826 [Google Scholar]
  86. Spruit, H. C., & Taam, R. E. 1993, ApJ, 402, 593 [CrossRef] [Google Scholar]
  87. Staubert, R., Trümper, J., Kendziorra, E., et al. 2019, A&A, 622, A61 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  88. Strüder, L., Briel, U., Dennerl, K., et al. 2001, A&A, 365, L18 [Google Scholar]
  89. Tomsick, J. A., Chaty, S., Rodriguez, J., Walter, R., & Kaaret, P. 2008, ApJ, 685, 1143 [NASA ADS] [CrossRef] [Google Scholar]
  90. Turner, M. J. L., Abbey, A., Arnaud, M., et al. 2001, A&A, 365, L27 [CrossRef] [EDP Sciences] [Google Scholar]
  91. van den Eijnden, J., Bagnoli, T., Degenaar, N., et al. 2017, MNRAS, 466, L98 [NASA ADS] [CrossRef] [Google Scholar]
  92. van der Klis, M. 2006, Compact Stellar X-ray Sources, 39 [CrossRef] [Google Scholar]
  93. van Paradijs, J. 1998, in The Many Faces of Neutron Stars, eds. R. Buccheri, J. van Paradijs, & A. Alpar, NATO Advanced Study Institute (ASI) Series C, 515, 279 [NASA ADS] [CrossRef] [Google Scholar]
  94. Wachter, K., Leach, R., & Kellogg, E. 1979, ApJ, 230, 274 [NASA ADS] [CrossRef] [Google Scholar]
  95. Wijnands, R., & van der Klis, M. 2000, ApJ, 528, L93 [NASA ADS] [CrossRef] [Google Scholar]
  96. Wilms, J., Allen, A., & McCray, R. 2000, ApJ, 542, 914 [Google Scholar]

All Tables

Table 1.

Summary of the X-ray observations.

Table 2.

Best-fit spectral parameters to describe the first part (black points in Fig. 3), low luminosity state (blue points in Fig. 3), and high luminosity state (i.e., flares; red points in Fig. 3).

Table 3.

Energy fluxes of the optical counterpart of J17407.

All Figures

thumbnail Fig. 1.

NuSTAR images (3–79 keV) of the J17407 field. Top panels: images for the modules A (left) and B (right) obtained using the entire observation. Bottom panels: images obtained during the first part of the observation (t < 59836.372534 MJD), where J17407 is faint and its flux constant (see Fig. 2). The red circle shows the extraction regions centred on the target. Magenta ellipses and circles show the extraction regions used for the background. Colorbars in the bottom show the count rate for each pixel.

In the text
thumbnail Fig. 2.

NuSTAR light curve in the energy range 3−60 keV obtained through the Bayesian block segmentation method. Top panel: light curve over the entire observation period. Bottom panels: three zoomed-in sections of the light curve (corresponding to the three grey-shaded areas in the top panel) to better show the typical time duration and structures of the flares. The ‘holes’ among bins, especially noticeable in the low-luminosity state of the top panel, arise from the passage of NuSTAR through the South Atlantic Anomaly regions.

In the text
thumbnail Fig. 3.

XMM-Newton and NuSTAR light curves and HRs. Panel a: XMM-Newton/pn light curve (0.2−12 keV). Panel b: NuSTAR light curve (3−60 keV). Panel c: NuSTAR light curve in the energy band 3−9 keV. Panel d: NuSTAR light curve in the energy band 9 − 60 keV. Panel e: NuSTAR HRs as a function of time. Panel f: NuSTAR HRs as a function of rate. The first part of the light curve, where the flares are absent, is displayed in all panels with black points (diamonds for XMM-Newton, squares for NuSTAR). Flares: Red triangles. Inter-flares: Blue circles.

In the text
thumbnail Fig. 4.

XMM-Newton (black: pn; red: MOS1; green: MOS2) and NuSTAR (blue: module A) spectra of J17407 during the first part of the light curve (black points in Fig. 3), fitted with an absorbed power law (see Table 2). The lower panel shows the residuals of the fit.

In the text
thumbnail Fig. 5.

NuSTAR (blue circles: module A; red triangles: module B) spectra of J17407 during the low luminosity state between flares (blue circles in Fig. 3). Top panel: spectra are fitted with an absorbed power law with a high-energy cutoff. Middle panel: residuals of the same fit of the top panel. Bottom panel: residuals of the fit of the spectra with an absorbed power law. See Table 2 for the best-fit parameters.

In the text
thumbnail Fig. 6.

NuSTAR (blue circles: module A; red triangles: module B) spectra of J17407 during the high luminosity state (i.e., flares; red triangles in Fig. 3). Top panel: spectra are fitted with two power law, both absorbed, one of them with a high-energy cutoff. Middle panel: residuals of the same fit of the top panel. Bottom panel: residuals of the fit of the spectra with an absorbed power law. The purpose of this panel is to highlight the presence of a feature in the residuals at ∼8.6 keV, which requires the use of a more complex model, as described in Sect. 3.2. See Table 2 for the best-fit parameters, also with other spectral models.

In the text
thumbnail Fig. 7.

Best-fit values of effective temperature, radius, and extinction obtained from the fit of the photometric SED of J17407 (see Table 3) with an absorbed blackbody model, with two different methods. Method one: best-fit values of T and R from the fit of the photometric SED, assuming distances in the range 1–30 kpc and for three values of Av (see main text for more details). These solutions show up as three horizontal lines whose colours reflect the values assumed for Av (see vertical colour bar on the right). A grid of distances (in units of kpc) in black is overplotted. Method two: “Γ-shape” coloured area shows the best-fit parameters ( χ red 2 $ \chi^2_{\rm red} $ < 2) T and R for different values of d, where the dependency of Av on the distance is based on Green et al. (2019). For this method, the distances (in units of kpc) from 1.6 to 9.4 kpc are overplotted in red. The most relevant spectral classes for our analysis are shown in solid (main sequence), dashed (sub-giant), and dot-dashed (giant) red lines. The inset figure shows an example of a best-fitting SED (dashed red line). Blue points are the photometric measurements (see Table 3). The fit residuals (observed-model)/error are shown in the lower panel.

In the text
thumbnail Fig. 8.

NuSTAR burst fluence (3−60 keV) as a function of the time interval between this burst and the next one.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.