Ultracompact nature of IGR J17062–6143

Evidence for the ultra-compact nature of IGR J17062–6143

J. V. Hernández Santisteban, V. Cúneo, N. Degenaar, J. van den Eijnden, D. Altamirano, M. N. Gómez, D M. Russell, R. Wijnands, R. Golovakova, M. T. Reynolds and J.M. Miller
Anton Pannekoek Institute for Astronomy, University of Amsterdam, Science Park 904, NL-1098 XH Amsterdam, the Netherlands
Institute of Astronomy, University of Cambridge, Madingley Road, Cambridge CB3 OHA, UK
Observatorio Astronómico de Córdoba, Córdoba, Argentina
CONICET, Consejo Nacional de Investigaciones Científicas y Técnicas, Argentina
Instituto Argentino de Radioastronomía (CCT La Plata, CONICET), C.C.5, (1984) Villa Elisa, Buenos Aires, Argentina
Department of Physics and Astronomy, University of Southampton, Southampton, SO17 1BJ, UK
New York University Abu Dhabi, PO Box 129188, Abu Dhabi, UAE
Department of Astronomy, University of Michigan, 1085 South University Avenue, Ann Arbor, MI 48109, USA
E-mail: j.v.hernandez@uva.nl (JVHS)
Accepted XXX. Received YYY; in original form ZZZ

We present a multi-wavelength study of the persistent low-luminosity neutron star low-mass X-ray binary IGR J17062–6143. The multi-epoch photometric UV to NIR spectral energy distribution (SED) is consistent with an accretion disc . The SED modelling of the accretion disc allowed us to estimate an outer disc radius of cm and a mass-transfer rate M yr, consistent with both theoretical and observational estimates of ultra-compact X-ray binaries (UCXB). In combination with empirical X-ray/NIR relationships, we estimate the orbital period of the system to be hr. In addition, we obtained a low-resolution optical spectrum which revealed a blue continuum and no emission lines. The lack of hydrogen in the spectrum and the size of the accretion disc provide further evidence for an ultra-compact nature of this system.

accretion, accretion discs – stars: neutron – X-rays: binaries – X-rays: individual: IGR J17062–6143
pubyear: 2017pagerange: Evidence for the ultra-compact nature of IGR J17062–6143B

1 Introduction

Low-mass X-ray binaries (LMXBs) are binary star systems where a neutron star (NS) or black hole (BH) accretes from a M companion. The companion typically overflows its Roche lobe, transferring mass onto an accretion disc that surrounds the compact primary. These systems are most easily discovered and studied when the X-ray luminosity liberated in the accretion process is (where is the Eddington limit) due to a high mass-accretion rate. However, LMXBs can also accrete at much lower rate, hence generating a much lower X-ray luminosity.

A number of LMXBs are found to accrete at an X-ray luminosity of for many years. The existence of these very-faint X-ray binaries (VFXBs) is puzzling, as their inferred low mass-accretion rates are below the critical threshold for which H-rich systems are expected to be transient (i.e. exhibiting weeks-months long outbursts of accretion separated by years-decades long quiescent episodes with an X-ray luminosity of ). For H-poor accretion discs, however, this critical mass-accretion threshold is lower (e.g. Tsugawa & Osaki, 1997; Menou et al., 2002; Heinke et al., 2013a; Hameury & Lasota, 2016). It has therefore been proposed that weakly-accreting persistent LMXBs harbour H-poor donor stars (e.g. in’t Zand et al., 2007). Such systems must have very short orbital periods (P hr) for the companion to overflow its Roche lobe, and are therefore referred to as Ultra-Compact X-ray binaries (UCXBs, Nelson:1986aa; Nelemans & Jonker, 2010, for a review).

There is great interest in identifying UCXBs among the population of LMXBs, because these objects are expected to be promising targets for future gravitational wave interferometry experiments (e.g. Nelemans, 2003). Furthermore, UCXBs are interesting laboratories to study the ashes of stellar nuclear burning (e.g. Deloye & Bildsten, 2003). In absence of a direct orbital period measurement (e.g. through the detection of periodic dips/eclipses in the X-ray emission, variations of pulsar arrival times or periodic optical variations), evidence for an UCXB nature can be obtained by searching for the absence of H features in optical spectra (e.g. Nelemans et al., 2004; Nelemans et al., 2006a) or by considering the ratio of the optical over X-ray emission (e.g. Bassa et al., 2006; Bassa et al., 2008; in’t Zand et al., 2009).

IGR J17062–6143 was originally discovered by Integral in 2006 and later identified as an accreting neutron star LMXB when Swift detected a thermonuclear burst in 2012 (Degenaar et al., 2012). Since its discovery, the source seems to have been persistently accreting at a low luminosity and it therefore classifies as a VFXB (Remillard & Levine, 2008; Degenaar et al., 2012; Degenaar et al., 2017; Keek et al., 2017). A detailed study of its X-ray spectrum revealed a broad Fe-K emission line near keV, a hallmark of disc reflection (Fabian & Ross, 2010), which allowed for a measure of the location of the inner accretion disc (Degenaar et al., 2017). This feature, so far not seen in other VFXBs (e.g Armas Padilla et al., 2013; Lotti et al., 2016), suggests that the inner accretion disc in IGR J17062–6143 is truncated at km (). This is in sharp contrast with LMXBs accreting at higher rates, where the inner disc typically resides a factor closer to the compact primary (e.g. Cackett et al., 2010; Ludlam et al., 2017, for sample studies). This suggests that the geometry of the inner accretion flow in this VFXB differs from a standard accretion disc, possibly due to the formation of a radiatively-inefficient accretion flow or because the magnetosphere of the neutron star is pushing gas away (Degenaar et al., 2017). Recently, Strohmayer & Keek (2017) discovered 163.65 Hz pulsations in a single archival 1.2 ks RXTE observation. This implies a magnetic field G, similar to other accreting millisecond X-ray pulsar (AMXP) (e.g. SAX J1808-359 Wijnands & van der Klis, 1998; Mukherjee et al., 2015). The magnetic field value from the truncated inner disc measurement, G, has been found in agreement with the pulsation estimates (van den Eijnden et al., 2017). Also, given the short duration of the RXTE observation, the orbital period could not be constrained imposing a lower limit of minutes.

IGR J17062–6143 suffers relatively little from interstellar absorption compared to other VFXBs and is therefore a particularly promising target to pursue optical observations with the aim to test for an UCXB nature. Apart from its long-term X-ray flux evolution, measurement of the inner disc radius and possible spin period, very little is known about the intrinsic properties of this binary system such as the companion type, orbital period and composition of the accreted material are unknown.

Both bursts detected from IGR J17062–6143 were of unusually long duration and are indicative of the ignition of a thick layer of He. While such bursts are expected to occur for neutron stars that accrete from H-depleted companions (e.g. in ’t Zand et al., 2005; Cumming et al., 2006), both theoretical calculations and observations suggest that similar bursts can also occur for neutron stars that slowly accrete H-rich material (e.g. Cooper & Narayan, 2007; Degenaar et al., 2010a). Therefore, the burst properties do not give conclusive information about the composition of the accreted material (hence the accretion disc). It is interesting to note that at least one neutron star that displays long X-ray bursts, 4U 0614+091 (Kuulkers et al., 2010) is thought to harbour a CO white dwarf companion based on the lack of H and He features in its optical spectrum (e.g. Nelemans et al., 2004; Nelemans et al., 2006b). This seems to suggest that at least a CO white dwarf companion is not inconsistent with the scenario we envision for IGR J17062–6143. Only marginal evidence for a (semi-)degenerate companion has been found as enhanced oxygen abundance in the high-resolution X-ray spectrum (van den Eijnden et al., 2017).

In this work, we present a multi-wavelength study of IGR J17062–6143: In Section 2, we present observations from X-ray to near-IR. The overall stability of the broad-band flux allows us to construct an average SED and retrieve the accretion disc parameters in Section 3. Combined with lack of H features in our optical spectrum, we argue that IGR J17062–6143 is a new, strong candidate UCXB in Section 4.

2 Observations

2.1 Faulkes Optical photometry

We observed the field of IGR J17062–6143 with the 2-m robotic Faulkes Telescope South (FTS), at Siding Spring, Australia, on 2016 October 4 and 2016 October 6. On both dates, 300-sec exposures were made in three filters; Sloan Digital Sky Survey (SDSS) , and -bands. FTS was equipped with a camera with a pixel scale of 0.304 arcsec pixel and a field of view of arcmin. The images were de-biased and flat-fielded using the automatic Las Cumbres Observatory (LCO) pipeline BANZAI. The seeing as measured from the images was 2.8 and 1.9 on 4 and 6 October, respectively.

We detected a faint source consistent with the position of IGR J17062–6143 in all images. The spatial resolution in the FTS images is too poor to separate out the two NIR sources detected by Magellan (see below), however the position of the optical counterpart is more consistent with that of the southern NIR source and Swift/UVOT coordinates. We find that the optical counterpart is bluer than the surrounding field stars (shown in the bottom panel of Fig. 1), which is consistent with emission from an accretion disc. Photometry was carried out using PHOT in IRAF. Flux calibration was achieved using the known , and magnitudes of four stars in the field of view tabulated in the AAVSO Photometric All-Sky Survey (APASS; Henden et al., 2009). The resulting magnitudes are given in Table 1. The errors include the 1 uncertainty in the comparison star magnitudes, which are small; the error is dominated by the S/N of the X-ray binary.

2.2 Swift UV photometry

IGR J17062–6143 has been observed multiple times by Swift between 2008 and 2016 which provides a rich photometric multi-wavelength follow-up of the system. We downloaded all the available calibrated files from the Swift Data Centre (Evans et al., 2009) and performed aperture photometry using uvotsource (as implemented in HEASOFT v6.18) using a 4 arcsec aperture for the target source and an 11 arcsec aperture for the background region. Although the full dataset includes measurements during the type-I X-ray bursts (Degenaar et al., 2013; Keek et al., 2017), we have excluded them for the analysis in the following sections. We present the average fluxes in every filter Table 1. The errors quoted reflect the r.m.s. The individual measurements are shown in Table 2.

2.3 Magellan NIR photometry

We obtained near-infrared (NIR) photometry with the FourStar camera (Persson et al., 2013) at the 6.5m Baade Magellan Telescope in Cerro las Campanas, Chile. The images were taken in two observing campaigns, 2013 June 16 and 2014 May 8. We observed the source in three filters , and ; the details of the individual observations are presented in Table 1. The telescope was nodded in a AB-AB mode, in order to optimise sky subtraction. The iraf/fsred package (provided by Andy Monson) was used to de-bias, flat-field, align, and co-add the FourStar observations for each object and filter. Aperture photometry was performed using 2MASS sources to determine the zero-point.

The source is clearly detected in all three filters as seen in the NIR finding chart in Fig. 1. The higher quality of the NIR images, with a measured seeing of arcsec, and a smaller pixel scale (0.159" pixel) revealed a fainter source blended in the north-west direction a shown in Fig. 2. We performed PSF photometry using daophot to extract the individual measurements, given in Table 1. The centroid for IGR J17062–6143 is 17:06:16.226(16), -61:42:39.95(23) and for the blended source is 17:06:16.197(22), -61:42:39.58(29). The errors on the positions represent the 90% upper bound statistical uncertainty on the centroid for all stars detected at a similar magnitude. We also note a " rms uncertainty for astrometric solution over the entire field. The latter are explicitly shown in the bottom panel of Fig. 1.

Figure 1: Finding chart for IGR J17062–6143 in the filter. Top: Wide field of view arcmin. Middle: Zoom to the position of IGR J17062–6143 and the blended source to the north-west. The circles on these sources represent the 0.3" uncertainty from the overall astrometric solution. We show the confidence regions as determined by Swift/UVOT (red) and Swift/XRT (black) is also shown for reference (Ricci et al., 2008). Bottom: False colour image made from the optical Faulkes photometry (blue , green and red ). The blue source is compatible with IGR J17062–6143 position obtained from the Magellan images marked as red circles.
Figure 2: Source identification of IGR J17062–6143. Left: -band imaging revealing a fainter blended source to the north-west. Right: A close-up to a star with similar brightness which shows the average PSF of the image. Contours track 10, 20, 30 and 40 counts above the background for both panels.
Facility Filter Date Exposure Magnitude
Å UTC s AB mag
Swift UVW2 1928 20.10 0.05
UVM2 2246 see see 20.40 0.15
UVW1 2600 Appendix Appendix 20.03 0.08
U 3465 § A § A 19.85 0.39
B 4349 20.30 1.69
Faulkes g’ 4770 2016-10-04 300 20.23 0.16
2016-10-06 300 20.29 0.13
r’ 6231 2016-10-04 300 20.30 0.17
2016-10-06 300 20.12 0.10
i’ 7625 2016-10-04 300 20.37 0.19
2016-10-06 300 20.05 0.12
Magellan J 12350 2013-06-16 87 20.54 0.08
H 16620 2013-06-16 306 20.63 0.08
2014-05-17 157 20.44 0.05
Ks 21590 2013-06-16 218 20.98 0.11
2014-05-17 157 20.78 0.10
Table 1: Observation log of the multi-wavelength photometry of IGR J17062–6143. We present only the average magnitudes and r.m.s. (quoted as the error) for the Swift/UVOT data. Otherwise, all errors represent the 1 confidence level. For a complete breakdown of the Swift/UVOT observations, see Table 2 in the Appendix.

2.4 Swift X-ray photometry

We extracted all available Swift/XRT spectra of IGR J17062–6143 between the NIR and optical observations to obtain a long-term X-ray light curve of the source. We used the Swift/XRT Online Data Products Generator111See http://www.swift.ac.uk/user_objects/index.php (Evans et al., 2009) to extract WT-mode and PC-mode spectra from in total epochs. To each spectrum, we fitted a simple absorbed blackbody plus power law model [tbabs*(bbodyrad+powerlaw)] in xspec v12.9.0 (Arnaud, 1996) and calculated the unabsorbed flux in the keV range. We assumed an absorbing hydrogen column density of (Degenaar et al., 2017).

The resulting long-term light curve is shown in Fig. 3, where the Eddington ratio was calculated assuming an Eddington luminosity of erg/s and a distance of kpc (Keek et al., 2017). The X-ray luminosity varied between , without going into quiescence or outburst in the past decade. Note that the observations at the end of 2015 were taken during a Type-I X-ray burst and its decay (Keek et al., 2017), and that we leave out the single Swift observation during the 2012 Type-I burst for visual clarity.

Figure 3: Long-term light curve of IGR J17062–6143 as observed with /XRT. Note that for visual clarity, we exclude the single observation during the source’s 2012 Type-I X-ray burst. The observations at the end of 2015 also coincide with a Type-I bursts and its decay are shown but not used in the analysis in Section 3.

2.5 Gemini optical spectroscopy

We obtained long-slit spectroscopic observations of IGR J17062–6143 with the Gemini Multi-Object Spectrograph (GMOS) at the 8m Gemini South telescope under a Fast Turnaround program (GS-2016A-FT-24) on 2016 September 27. The GMOS instrument uses the Hamamatsu CCD ( pixels). Six spectra of 900s, three centred at 570 nm and three at 580 nm to avoid the chip gaps, were taken using the the B150 grating (150 l/mm), a slit width of and a binning. The GG455_G0329 blocking filter was also used in order to avoid second order overlap. The chosen set-up resulted in the spectral coverage of the Å wavelength range and a spectral resolution of 3 Å. The seeing during the observation was . A CuAr lamp was also observed for each configuration in order to perform the wavelength calibration.

Spectra were reduced using the iraf-gemini package222http://iraf.noao.edu/. The flux calibration of each spectrum was executed using observations of a standard star, also taken as part of the program and the errors propagated through the gemini pipeline. Finally, in order to increase the signal-to-noise (S/N), the six spectra were combined to obtain a final spectrum with a S/N of at 6500 Å and 7500 Å, and at 8500 Å. The final spectrum was cut to keep the spectral region between 5000 and 9000 Å, as shown in Fig. 4. The bluest part of the spectrum was removed as the filter lowered the response at those wavelengths. On the other side, the reddest part was cut out because the atmospheric absorption becomes too significant and distorts the slope of the continuum. A few features were left in the final spectrum, marked as crosses in Fig. 4. We checked the individual spectra before combining in order to asses the validity of every singular feature. Most of them were tracked to bad background subtraction and cosmic ray removal. We only find the \ionNai doublet 5889, 5895 Å to be present in all six spectra. Given the intrinsic variability of the object and non-simultaneity of the observations, we did not attempted to use the Faulkes photometry to perform a correction on the flux calibration.

3 Results

3.1 A featureless optical spectrum

The optical spectrum of IGR J17062–6143 shows a blue continuum with no emission line features as shown in Fig. 4. We only identify one absorption feature, due to interstellar extinction of \ionNai doublet 5889, 5895 Å. We find an instrumental artefact around 7816 Å, which is coincident with a known \ionHei line. However, after looking at the individual data (only three of the six had information in this region since it lies on the detector gap for the others), we find no evidence for such features.

Figure 4: Flux calibrated optical spectrum of IGR J17062–6143. Telluric transmission spectrum is shown for reference. Regions of telluric absorption have been labelled with and shaded in grey. The crosses show instrumental artefact and bad sky-line subtraction features (see text for details). The locations of H and interstellar \ionNai are marked as well.

The optical spectra of LMXBs are typically dominated by that of the irradiated accretion disc, showing a blue continuum with strong Balmer, Heii and Bowen emission features (see e.g. Charles & Coe, 2006). In occasions, LMXBs can show broad absorption features in high-states and/or outbursts (see e.g. Cornelisse et al., 2009). However, some LMXBs have similarly featureless optical spectra as IGR J17062–6143. In particular, a targeted study of candidate and confirmed UCXBs revealed surprisingly featureless optical spectra (Nelemans et al., 2004; Nelemans et al., 2006a). Whereas the optical spectra of white dwarf analogues of UCXBs (so-called AM CVns) show rich line spectra, detecting emission lines of He, C, N, and O, in UCXBs appears to be only achievable when high signal to noise data is available. Nevertheless, the lack of strong H lines compared to other LMXBs is striking and therefore seems a promising diagnostic to search for UCXBs among LMXBs.

A few LMXBs have indeed been put forward as candidate UCXBs based on the lack of H features in their optical spectra. For instance, the neutron star LMXB A 1246-58 (in’t Zand et al., 2008), which had been previously earmarked as a candidate UCXB based on its low long-term X-ray flux and its small optical over X-ray flux ratio (Bassa et al., 2006). Furthermore, the optical spectrum of the neutron star LMXB 1RXS J180408.9–342058 obtained during its 2015 outburst was nearly featureless apart from the potential detection of a weak Heii line (Baglio et al., 2016). The broad-band X-ray spectrum, X-ray flux amplitude between outburst and quiescence, width of the putative He line, and optical over X-ray flux ratio indeed all point to a relatively compact orbit for that source (Degenaar et al., 2016). In addition, the recently discovered transient X-ray source and LMXB candidate MAXI J1957+032 shows a blue but featureless optical spectrum. Its unusual brief X-ray outbursts (lasting only a few days) may indeed suggest a short orbital period (Mata Sánchez et al., 2017).

There are at least two examples, however, of black hole LMXBs with orbital periods of hr that lacked H features during some phases of their outbursts: Swift J1753.5–0127 (Cadolle Bel et al., 2007; Zurita et al., 2008) and Swift J1357.2–0933 (e.g. Torres et al., 2011; Corral-Santana et al., 2013). Although lack of H in the optical spectrum may thus not strictly imply an ultra-compact nature, we show in the next sections that in case of IGR J17062–6143 both its spectral energy distribution (SED) and ratio of its optical/NIR and X-ray fluxes also support an UCXB nature.

3.2 X-ray/Optical Correlations

In order to justify the use of the multi-wavelength data across three years to construct the persistent SED of IGR J17062–6143, we analyse the intrinsic variability and correlations using simultaneous X-ray/UV/optical data from Swift. We excluded epochs concerning the two type-1 X-ray bursts (Degenaar et al., 2013; Keek et al., 2017). We have searched for possible correlation between the 2-10 keV X-ray band and the different UV/optical observations taken with Swift/UVOT data. We performed a Pearson-rank test for each filter and found only the -band to reject the null-hypothesis (no correlation) at , with a p-value of . We note that IGR J17062–6143 presents UV/optical variability even at similar X-ray levels as shown in Fig. 5, hence it makes difficult to separate and discern any correlation. Therefore, we conclude that the observed scatter in the UV/optical is roughly independent of the X-ray flux and consistent (within a certain scatter) throughout all epochs. We will take into account this variability in our SED modelling presented in the following sections.

Figure 5: Swift/UVOT UV/optical bands as a function of X-ray flux in the 2-10 keV band, during the persistent state of IGR J17062–6143.

3.3 Spectral energy distribution

The broadband SED of IGR J17062–6143 is shown in Fig. 6. The UV to NIR data follows an extincted clear power law behaviour, . We have also calculated the reddening as a free parameter. We used as our goodness-of-fit parameter and the 1 confidence intervals were obtained by scaling the errors so . We estimated a power law index , consistent with the plateau formed by a steady-state accretion disc (Lynden-Bell, 1969; Frank et al., 2002) and E(B-V). This latter value of extinction is lower than estimates of simultaneous NuStar and Chandra X-ray spectral fits, which render column densities of cm (Degenaar et al., 2017) i.e. E(B-V) (using transformations of Predehl & Schmitt, 1995).

Given the lack of any donor features and the spectral index of the photometry, we have assumed that the UV-NIR wavelength range is dominated entirely by the accretion disc. In order to retrieve physical parameters, which are more informative than a simple power law, we have used the model of an irradiated accretion disc as in Chakrabarty (1998, Eq. 10 to 15). This model consists of a thick geometrically thin accretion disc (Shakura & Sunyaev, 1973; Frank et al., 2002) which emission is modified by the central X-ray source. Therefore, the temperature profile is defined by the combination of internal viscous heating and shallow X-ray heating. We have assumed a disc albedo and a canonical mass for the NS of M M. In order to compare the disc model to the available photometry ( and , their associated 1 uncertainties), we applied the reddening of the object to the model spectrum and performed synthetic photometry for every filter in our sample. We have excluded the spectrum from our analysis due to an unreliable absolute flux calibration (see Sec. 2.5). To derive the best fit parameters, we explored the parameter space by using a MCMC procedure as implemented in emcee (Foreman-Mackey et al., 2013)333http://dan.iel.fm/emcee/current/.. The likelihood function that we employed to retrieve the best fit parameters, where the uncertainties are Gaussian and independent, is given by


where the total variance is defined as . We have added a fractional scatter common to the dataset, in order to reflect the intrinsic variability and non-simultaneity of the observations. Also, this added scatter will reflect any contamination arising from the dim companion detected at NIR wavelengths (see §2.3). The choice of priors for our MCMC procedure is described below.

During the energetic 2015 Type-I burst, signatures of photospheric radius expansion were observed, which is an indication that the Eddington limit was indeed reached (Keek et al., 2017), hence a distance to the system of kpc can be inferred. This measurement is perhaps more accurate than the kpc estimate inferred from the 2012 burst (Degenaar et al., 2013), since the 2015 Type-I burst had softer photon coverage near its peak (using MAXI; 2–20 keV) than the 2012 one (using Swift/BAT; 15–50 keV) and should thus provide more reliable constraints on the soft  keV black body emission. We used Keek et al. (2017) measurement and its associated error as a Gaussian prior for the distance. The galactic extinction was obtained from the hydrogen density column measured in the X-ray spectral fitting (Degenaar et al., 2017; Keek et al., 2017). We therefore employed values of E(B-V) (with typical for the Milky Way, Predehl & Schmitt, 1995), where the uncertainty reflects the spread of values obtained, as a Gaussian prior and restricted to positive values E. For the rest of the disc parameters, inner radius , outer radius and mass-transfer rate , we assumed log-uniform priors. We also imposed the condition that .

The SED with the best fit is shown in Fig. 6 (joint and marginal posterior distributions are shown in Appendix B). We found the outer radius of the disc to be well constrained cm and an upper limit of the inner radius at 2 of cm. This is not surprising given the lack of information in the far-UV ( Å) region of the spectrum which produces a wide range of solutions, explicitly shown in the random realisations of Fig. 6. The inclination of the system is also unconstrained, however we can impose a 2 upper limit of , consistent with X-ray reflection modelling (Degenaar et al., 2017; van den Eijnden et al., 2017). We obtain a  M yr which is consistent with observational (Cartwright et al., 2013; Heinke et al., 2013b) and theoretical (Deloye & Bildsten, 2003; Sengar et al., 2017) estimates of persistent UCXBs. This evident discrepancy between the and that inferred from the low-luminosity is discussed in the following section.

We have included the best X-ray fitting components (bbody+gaussian+powerlaw) to the simultaneous XMM-Newton and NuStar data presented in van den Eijnden et al. (2017) in Fig. 6. Our SED fitting suggests that the thermal component in the X-ray spectrum (bbody) is very unlikely from the disc, and more in line with thermal emission from the NS or boundary layer. This was previously proposed for this and other VFXBs based on the inferred temperature and emitting radius (e.g. Armas Padilla et al., 2011; Wijnands et al., 2015; Degenaar et al., 2017), but now for the first time seen very clearly by studying the multi-wavelength SED of a VFXB.

Figure 6: Broadband SED fitting of IGR J17062–6143 modelled as an irradiated accretion disc shown as the best fit model (red line). The UV to NIR data has been dereddened by the best value found from the SED fit. The X-ray spectrum was simulated from the best fit to simultaneous XMM-Newton and NuStar data (powerlaw+bbody+gaussian) taken from van den Eijnden et al. (2017). The black lines show random realisations from our MCMC analysis. The unconstrained inner radius and inclination in the SED fit results in a large spread at shorter wavelengths Å.

4 Discussion

There are two VFXBs that have optical properties apparently ruling out an UCXB nature: 1RXH J173523.7–354013 shows very strong H emission in its optical spectrum (Degenaar et al., 2010b), whereas the companion of M15 X-3 appears to be too optically bright for an UCXB (Arnason et al., 2015). An alternative explanation for their sustained low X-ray luminosity may be that the magnetic field of the NS is interacting with the accretion flow and prevents a higher mass-accretion rate (Wijnands, 2008; Degenaar et al., 2014; Heinke et al., 2015). Such an explanation was tentatively also proposed for IGR J17062–6143, since X-ray spectroscopy suggests that its inner accretion disc is severely truncated away from the NS (Degenaar et al., 2017). However, our present study shows that this source is also a strong candidate UCXB.

Future photometric and/or spectroscopic studies may reveal the orbital period of the system and unequivocally determine the nature of the system. Nonetheless, we can use indirect methods to determine robust range of acceptable to aid further observational campaigns. Such methods concern the relative contributions of the optical/infrared to X-ray fluxes (van Paradijs & McClintock, 1994; Revnivtsev et al., 2012). As the central X-ray source illuminates the accretion disc, a fraction of the incident energy is reprocessed modifying the emitted SED. In addition, at longer wavelengths (such as the NIR) the donor may contribute a significant fraction of the total luminosity.

We have plotted the X-ray/optical luminosities against a large sample of BHs and NSs in Fig. 7. As clearly shown, both classes of LMXBs have a different correlation, where BHs are a factor brighter in optical than NS LMXBs. Taking into account both estimates of the distance, IGR J17062–6143 lies comfortably within the NS track. However, despite its position at the lower end of the NS track, we caution against any inference on the orbital period from the low optical luminosity alone (e.g. van Paradijs & McClintock, 1994). In particular, IGR J17062–6143’s luminosity is similar to other NS LMXBs with widely different orbital periods such as SAX J1808.4-3658 (2.013 hr, Chakrabarty & Morgan, 1998), 4U 1608-52 (12.89 hr, Wachter et al., 2002) and Aql X-1 (18.95 hr, Chevalier & Ilovaisky, 1998).

Figure 7: L - L correlation for BH and NS. Data was taken from Russell et al. (2006) and Russell et al. (2007). IGR J17062–6143 (square) follows the NS track for both estimates of the distance to the system. We show the UCBX 4U06414+09 (circle) as a reference for a typical system.

On the other hand, the empirical relationship developed by Revnivtsev et al. (2012) provides a new framework to characterise persistent LMXBs, such as IGR J17062–6143. At NIR wavelengths, the system is assumed to be dominated by the reprocessing of the central X-ray flux in the accretion disc and the secondary star. The use of exclusive persistent NS-sources provides a more homogeneous dataset (small variability and mass-dependence) to perform the calibration of this relationship. This method describes the absolute magnitude in the -band (in Vega system) as a function of the orbital period and X-ray luminosity,




In order to calculate the luminosity, we used a distance of  kpc (Keek et al., 2017), an extinction value of , a correction on the extinction for the band  (Rieke & Lebofsky, 1985) and a fixed . We find, for both NIR -band measurements, an orbital period of hr and hr444If we use a value for kpc (Degenaar et al., 2013), we obtain even shorter orbital periods hr.. The 1 uncertainties on the orbital periods were calculated via a Monte Carlo simulation.

Alternatively, we can constrain the orbital period by using the size of the accretion disc, measured in our SED modelling. The outer parts of the disc will be heavily affected by the tidal interaction of the companion thus regulating its maximum size (Paczynski, 1977). This tidal radius, , can be approximated by


where is the orbital separation and the mass ratio of the system. On the other hand, the minimum outer disc radius, is determined by the angular momentum of the particles after they exit the inner Lagrangian point. This can be approximated by (Verbunt & Rappaport, 1988)


valid for . We can find a family of solutions of and as a function of donor mass given a fixed orbital period, as shown in Fig. 8. Any allowed orbital period of the system should reside, for a given donor mass (assuming a canonical mass for the neutron star M M), between both curves. This allows us to exclude systems with P hr since the estimates of the disc are smaller than the minimum radius . On the other hand, the lower limit provided by the X-ray pulsations (Strohmayer & Keek, 2017) provides a constrain on the minimum mass of the donor M, thus allowing for a wide range of possible donors. However, at this shorter orbital periods, the donor of IGR J17062–6143 is expected to be a (semi-)degenerate object (such as WD or a helium star Sengar et al., 2017) to physically fit inside the Roche-lobe of the secondary (e.g. 4U 1626-67, Levine et al., 1988). Given the the mass transfer rate estimate and the lower limit on the orbital period, we cannot discard the scenario where the system remained in the LMXB phase close to its period minimum and never underwent a second phase of mass transfer (Sengar et al., 2017).

Figure 8: The tidal radius of an accretion disc as a function of donor mass. We compare the estimate from the SED modelling (red line) to the predicted tidal (black lines) and circularisation (blue lines) radius for a given orbital period. The and confidence levels of R are shown as the red bands. The smallest orbital period represents the lower limit of 17 minutes obtained from the X-ray pulsations (Strohmayer & Keek, 2017).

The low X-ray luminosity of IGR J17062–6143 contrasts with the higher mass transfer rate from the donor implied by our SED modelling as well as theoretical predictions for UCXBs (Sengar et al., 2017). If the X-ray luminosity from the blackbody component is due to accretion onto the NS surface (van den Eijnden et al., 2017), erg s, we can obtain an estimate on the mass accretion rate. Assuming a totally efficient conversion of the accretion power into , we find M yr. This would require that the system, in order for mass conservation to hold, ejects % of the in-falling material. One possible mechanism to reconcile this is that inner accretion disc is truncated by the NS magnetosphere, which drives a propeller outflow (Illarionov & Sunyaev, 1975). Hints of such a (propeller-)outflow ocurring in IGR J17062–6143 are present in the high-resolution X-ray spectra as oxygen-rich high-velocity absorption and emission features (van den Eijnden et al., 2017).

5 Conclusions

We have performed a multi-wavelength analysis of the ultra-compact binary candidate IGR J17062–6143. The low-resolution optical spectrum shows a blue-continuum consistent with an accretion disc however no emission lines are observed, suggesting a H-poor companion. We used multi-epoch UV, optical and NIR photometry to model the accretion disc and retrieve physical parameters such as the outer disc radius and mass transfer. Both estimates are consistent with properties of systems with orbital periods  hr. In addition, we employed empirical relations to estimate an orbital period between 0.3-0.5 hr. Further simultaneous multi-wavelength observations will allow to construct a more accurate SED of the system and reduce the uncertainties of the accretion disc modelling. The low extinction of the system makes IGR J17062–6143 an ideal candidate to explore UCXBs, and explore the the connection between the accretion inflow and the outflow produced by the interaction with the neutron star.


We thank the referee for comments and suggestions that significantly improved this work. JVHS, ND, and JvE are supported by a Vidi grant awarded to ND by the Netherlands Organization for Scientific Research (NWO). JVHS acknowledges partial support from “NewCompStar", COST Action MP1304 and thanks the IoA, Cambridge for their hospitality during the exchange visit. ND is also supported by a Marie Curie grant from the European Commission (contract no. FP-PEOPLE-2013-IEF-627148). DA acknowledges support from the Royal Society. RW is supported by an NWO Top grant, module 1. VC is supported by a grant awarded by the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Argentina. We thank Andy Monson for developing and supporting FSRED and for help installing that package. We are grateful to Neil Gehrels and the Swift duty scientists for making ToO observations of IGR J17062–6143 possible. We also thank German Gimeno and the Gemini duty scientists for the observation of IGR J17062–6143 under a FT observing program. The Faulkes Telescopes are maintained and operated by the Las Cumbres Observatory (LCO). We acknowledge the use of public data from the Swift data archive. This research made use of astropy, a community-developed core Python package for Astronomy (Astropy Collaboration et al., 2013), matplotlib (Hunter, J. D., 2007) and aplpy (Robitaille & Bressert, 2012).


  • Armas Padilla et al. (2011) Armas Padilla M., Degenaar N., Patruno A., Russell D. M., Linares M., Maccarone T. J., Homan J., Wijnands R., 2011, MNRAS, 417, 659
  • Armas Padilla et al. (2013) Armas Padilla M., Degenaar N., Wijnands R., 2013, MNRAS, 434, 1586
  • Arnason et al. (2015) Arnason R. M., Sivakoff G. R., Heinke C. O., Cohn H. N., Lugger P. M., 2015, ApJ, 807, 52
  • Arnaud (1996) Arnaud K. A., 1996, in Jacoby G. H., Barnes J., eds, Astronomical Society of the Pacific Conference Series Vol. 101, Astronomical Data Analysis Software and Systems V. p. 17
  • Astropy Collaboration et al. (2013) Astropy Collaboration et al., 2013, A&A, 558, A33
  • Baglio et al. (2016) Baglio M. C., D’Avanzo P., Campana S., Goldoni P., Masetti N., Muñoz-Darias T., Patiño-Álvarez V., Chavushyan V., 2016, A&A, 587, A102
  • Bassa et al. (2006) Bassa C. G., Jonker P. G., in’t Zand J. J. M., Verbunt F., 2006, A&A, 446, L17
  • Bassa et al. (2008) Bassa C., et al., 2008, The Astronomer’s Telegram, 1575
  • Cackett et al. (2010) Cackett E. M., et al., 2010, ApJ, 720, 205
  • Cadolle Bel et al. (2007) Cadolle Bel M., et al., 2007, ApJ, 659, 549
  • Cartwright et al. (2013) Cartwright T. F., Engel M. C., Heinke C. O., Sivakoff G. R., Berger J. J., Gladstone J. C., Ivanova N., 2013, ApJ, 768, 183
  • Chakrabarty (1998) Chakrabarty D., 1998, ApJ, 492, 342
  • Chakrabarty & Morgan (1998) Chakrabarty D., Morgan E. H., 1998, Nature, 394, 346
  • Charles & Coe (2006) Charles P. A., Coe M. J., 2006, Optical, ultraviolet and infrared observations of X-ray binaries. pp 215–265
  • Chevalier & Ilovaisky (1998) Chevalier C., Ilovaisky S. A., 1998, IAU Circ., 6806
  • Cooper & Narayan (2007) Cooper R., Narayan R., 2007, ApJ, 661, 468
  • Cornelisse et al. (2009) Cornelisse R., et al., 2009, A&A, 495, L1
  • Corral-Santana et al. (2013) Corral-Santana J. M., Casares J., Muñoz-Darias T., Rodríguez-Gil P., Shahbaz T., Torres M. A. P., Zurita C., Tyndall A. A., 2013, Science, 339, 1048
  • Cumming et al. (2006) Cumming A., Macbeth J., in ’t Zand J., Page D., 2006, ApJ, 646, 429
  • Degenaar et al. (2010a) Degenaar N., et al., 2010a, MNRAS, 404, 1591
  • Degenaar et al. (2010b) Degenaar N., et al., 2010b, MNRAS, 404, 1591
  • Degenaar et al. (2012) Degenaar N., Altamirano D., Wijnands R., 2012, The Astronomer’s Telegram, 4219
  • Degenaar et al. (2013) Degenaar N., Miller J. M., Wijnands R., Altamirano D., Fabian A. C., 2013, ApJ, 767, L37
  • Degenaar et al. (2014) Degenaar N., et al., 2014, ApJ, 792, 109
  • Degenaar et al. (2016) Degenaar N., et al., 2016, MNRAS, 461, 4049
  • Degenaar et al. (2017) Degenaar N., Pinto C., Miller J. M., Wijnands R., Altamirano D., Paerels F., Fabian A. C., Chakrabarty D., 2017, MNRAS, 464, 398
  • Deloye & Bildsten (2003) Deloye C. J., Bildsten L., 2003, ApJ, 598, 1217
  • Evans et al. (2009) Evans P. A., et al., 2009, MNRAS, 397, 1177
  • Fabian & Ross (2010) Fabian A. C., Ross R. R., 2010, Space Sci. Rev., 157, 167
  • Foreman-Mackey (2016) Foreman-Mackey D., 2016, The Journal of Open Source Software, 24
  • Foreman-Mackey et al. (2013) Foreman-Mackey D., Hogg D. W., Lang D., Goodman J., 2013, PASP, 125, 306
  • Frank et al. (2002) Frank J., King A., Raine D. J., 2002, Accretion Power in Astrophysics: Third Edition
  • Hameury & Lasota (2016) Hameury J.-M., Lasota J.-P., 2016, A&A, 594, A87
  • Heinke et al. (2013a) Heinke C. O., Ivanova N., Engel M. C., Pavlovskii K., Sivakoff G. R., Cartwright T. F., Gladstone J. C., 2013a, ApJ, 768, 184
  • Heinke et al. (2013b) Heinke C. O., Ivanova N., Engel M. C., Pavlovskii K., Sivakoff G. R., Cartwright T. F., Gladstone J. C., 2013b, ApJ, 768, 184
  • Heinke et al. (2015) Heinke C. O., Bahramian A., Degenaar N., Wijnands R., 2015, MNRAS, 447, 3034
  • Henden et al. (2009) Henden A. A., Welch D. L., Terrell D., Levine S. E., 2009, in American Astronomical Society Meeting Abstracts #214. p. 669
  • Hunter, J. D. (2007) Hunter, J. D. 2007, Computing In Science & Engineering, 9, 90
  • Illarionov & Sunyaev (1975) Illarionov A. F., Sunyaev R. A., 1975, A&A, 39, 185
  • Keek et al. (2017) Keek L., Iwakiri W., Serino M., Ballantyne D. R., in’t Zand J. J. M., Strohmayer T. E., 2017, ApJ, 836, 111
  • Kuulkers et al. (2010) Kuulkers E., et al., 2010, A&A, 514, A65
  • Levine et al. (1988) Levine A., Ma C. P., McClintock J., Rappaport S., van der Klis M., Verbunt F., 1988, ApJ, 327, 732
  • Lotti et al. (2016) Lotti S., et al., 2016, ApJ, 822, 57
  • Ludlam et al. (2017) Ludlam R. M., et al., 2017, ApJ, 836, 140
  • Lynden-Bell (1969) Lynden-Bell D., 1969, Nature, 223, 690
  • Mata Sánchez et al. (2017) Mata Sánchez D., Charles P. A., Armas Padilla M., Buckley D. A. H., Israel G. L., Linares M., Muñoz-Darias T., 2017, MNRAS, 468, 564
  • Menou et al. (2002) Menou K., Perna R., Hernquist L., 2002, ApJ, 564, L81
  • Mukherjee et al. (2015) Mukherjee D., Bult P., van der Klis M., Bhattacharya D., 2015, MNRAS, 452, 3994
  • Nelemans (2003) Nelemans G., 2003, in Centrella J. M., ed., American Institute of Physics Conference Series Vol. 686, The Astrophysics of Gravitational Wave Sources. pp 263–272 (arXiv:astro-ph/0310800), doi:10.1063/1.1629441
  • Nelemans & Jonker (2010) Nelemans G., Jonker P. G., 2010, New Astron. Rev., 54, 87
  • Nelemans et al. (2004) Nelemans G., Jonker P. G., Marsh T. R., van der Klis M., 2004, MNRAS, 348, L7
  • Nelemans et al. (2006a) Nelemans G., Jonker P. G., Steeghs D., 2006a, MNRAS, 370, 255
  • Nelemans et al. (2006b) Nelemans G., Jonker P. G., Steeghs D., 2006b, MNRAS, 370, 255
  • Paczynski (1977) Paczynski B., 1977, ApJ, 216, 822
  • Persson et al. (2013) Persson S. E., et al., 2013, PASP, 125, 654
  • Predehl & Schmitt (1995) Predehl P., Schmitt J. H. M. M., 1995, A&A, 293, 889
  • Remillard & Levine (2008) Remillard R. A., Levine A. M., 2008, The Astronomer’s Telegram, 1853
  • Revnivtsev et al. (2012) Revnivtsev M. G., Zolotukhin I. Y., Meshcheryakov A. V., 2012, MNRAS, 421, 2846
  • Ricci et al. (2008) Ricci C., Beckmann V., Carmona A., Weidenspointner G., 2008, The Astronomer’s Telegram, 1840
  • Rieke & Lebofsky (1985) Rieke G. H., Lebofsky M. J., 1985, ApJ, 288, 618
  • Robitaille & Bressert (2012) Robitaille T., Bressert E., 2012, APLpy: Astronomical Plotting Library in Python, Astrophysics Source Code Library
  • Russell et al. (2006) Russell D. M., Fender R. P., Hynes R. I., Brocksopp C., Homan J., Jonker P. G., Buxton M. M., 2006, MNRAS, 371, 1334
  • Russell et al. (2007) Russell D. M., Fender R. P., Jonker P. G., 2007, MNRAS, 379, 1108
  • Sengar et al. (2017) Sengar R., Tauris T. M., Langer N., Istrate A. G., 2017, preprint
  • Shakura & Sunyaev (1973) Shakura N. I., Sunyaev R. A., 1973, A&A, 24, 337
  • Strohmayer & Keek (2017) Strohmayer T., Keek L., 2017, ApJ, 836, L23
  • Torres et al. (2011) Torres M. A. P., Steeghs D., Jonker P. G., Rauch M., 2011, The Astronomer’s Telegram, 3143
  • Tsugawa & Osaki (1997) Tsugawa M., Osaki Y., 1997, PASJ, 49, 75
  • Verbunt & Rappaport (1988) Verbunt F., Rappaport S., 1988, ApJ, 332, 193
  • Wachter et al. (2002) Wachter S., Hoard D. W., Bailyn C. D., Corbel S., Kaaret P., 2002, ApJ, 568, 901
  • Wijnands (2008) Wijnands R., 2008, in Bandyopadhyay R. M., Wachter S., Gelino D., Gelino C. R., eds, American Institute of Physics Conference Series Vol. 1010, A Population Explosion: The Nature & Evolution of X-ray Binaries in Diverse Environments. pp 382–386 (arXiv:0801.0953), doi:10.1063/1.2945081
  • Wijnands & van der Klis (1998) Wijnands R., van der Klis M., 1998, Nature, 394, 344
  • Wijnands et al. (2015) Wijnands R., Degenaar N., Armas Padilla M., Altamirano D., Cavecchi Y., Linares M., Bahramian A., Heinke C. O., 2015, MNRAS, 454, 1371
  • Zurita et al. (2008) Zurita C., Durant M., Torres M. A. P., Shahbaz T., Casares J., Steeghs D., 2008, ApJ, 681, 1458
  • in ’t Zand et al. (2005) in ’t Zand J., Cumming A., van der Sluys M., Verbunt F., Pols O., 2005, A&A, 441, 675
  • in’t Zand et al. (2007) in’t Zand J. J. M., Jonker P. G., Markwardt C. B., 2007, A&A, 465, 953
  • in’t Zand et al. (2008) in’t Zand J. J. M., Bassa C. G., Jonker P. G., Keek L., Verbunt F., Méndez M., Markwardt C. B., 2008, A&A, 485, 183
  • in’t Zand et al. (2009) in’t Zand J. J. M., Jonker P. G., Bassa C. G., Markwardt C. B., Levine A. M., 2009, A&A, 506, 857
  • van Paradijs & McClintock (1994) van Paradijs J., McClintock J. E., 1994, A&A, 290, 133
  • van den Eijnden et al. (2017) van den Eijnden J., et al., 2017, MNRAS, pp XXX–XXX

Appendix A Swift photometry

We present in Table 2 all the individual measurements obtained with Swift/UVOT used in the SED modelling in AB mag with the X-ray flux associated for each pointing. Details on the data reduction can be found in Section 2.2 for UVOT and Section 2.4 for XRT.

Filter MJD Exposure Time Magnitude [2-10 keV]
s UVOT System mJy erg s cm
UVW2 54587.45660 277.0 19.97 0.12 0.037 0.004 1.478 0.148
UVW2 54588.19469 1061.9 20.11 0.08 0.033 0.002 1.268 0.127
UVW2 54588.05937 283.2 20.18 0.13 0.031 0.004 1.268 0.127
UVW2 54592.43834 790.6 20.15 0.08 0.032 0.002 2.325 0.232
UVW2 54592.33448 224.1 20.07 0.14 0.034 0.004 2.325 0.232
UVW2 54593.74933 1184.3 20.13 0.07 0.032 0.002 1.991 0.199
UVW2 54593.61518 460.4 20.15 0.11 0.032 0.003 1.991 0.199
UVW2 54594.61436 2391.6 20.08 0.06 0.034 0.002 1.991 0.199
UVW2 54594.27713 66.7 20.08 0.24 0.034 0.007 1.991 0.199
UVW2 54595.28439 3163.7 20.08 0.05 0.034 0.002 1.934 0.193
UVW2 54595.01637 558.8 20.15 0.10 0.032 0.003 1.157 0.116
UVW2 54597.42468 4067.9 20.11 0.05 0.033 0.002 1.576 0.158
UVW2 54597.02051 627.1 20.18 0.09 0.031 0.003 1.576 0.158
UVW2 54661.53944 6809.6 20.03 0.05 0.035 0.001 1.598 0.160
UVW2 54661.33972 1021.9 20.08 0.07 0.034 0.002 1.598 0.160
UVM2 57624.60002 805.1 20.72 0.16 0.019 0.003 0.546 0.055
UVM2 57630.18016 992.4 20.90 0.16 0.016 0.002 0.306 0.031
UVM2 57639.42057 1088.9 20.40 0.11 0.025 0.003 0.358 0.036
UVM2 57646.85150 551.5 20.97 0.23 0.015 0.003 0.323 0.032
UVM2 54587.78830 82.7 20.46 0.37 0.024 0.008 1.268 0.127
UVM2 54588.06322 213.8 20.12 0.19 0.033 0.006 1.268 0.127
UVM2 54592.33732 130.4 19.97 0.22 0.037 0.008 2.325 0.232
UVM2 54593.62108 294.3 19.80 0.14 0.044 0.006 1.991 0.199
UVM2 54594.27809 47.5 20.91 0.64 0.016 0.009 1.991 0.199
UVM2 54595.02385 414.6 20.16 0.14 0.031 0.004 1.157 0.116
UVM2 54598.03805 833.0 19.99 0.09 0.037 0.003 1.576 0.158
UVW1 54587.45231 141.5 20.04 0.19 0.034 0.006 1.478 0.148
UVW1 54588.05504 141.5 20.11 0.20 0.032 0.006 1.268 0.127
UVW1 54590.10843 437.2 19.93 0.12 0.038 0.004 1.032 0.103
UVW1 54590.00866 121.8 20.44 0.31 0.024 0.007 1.032 0.103
UVW1 54592.14843 130.5 20.08 0.22 0.033 0.007 2.325 0.232
UVW1 54593.60824 230.1 19.86 0.14 0.041 0.005 1.991 0.199
UVW1 54594.27599 33.2 19.93 0.36 0.038 0.013 1.991 0.199
UVW1 54595.00798 279.3 19.83 0.12 0.042 0.005 1.157 0.116
U 54587.45361 70.6 19.87 0.26 0.042 0.010 1.478 0.148
U 54588.05633 70.6 19.73 0.24 0.047 0.010 1.268 0.127
U 54590.00952 16.4 20.55 1.57 0.022 0.032 1.032 0.103
U 54592.33206 55.9 19.45 0.22 0.061 0.012 2.325 0.232
U 54593.61033 114.9 19.56 0.17 0.055 0.009 1.991 0.199
U 54594.27633 16.5 19.10 0.31 0.084 0.024 1.991 0.199
U 54595.01050 139.5 19.87 0.19 0.042 0.007 1.157 0.116
U 54600.41665 3073.0 19.68 0.05 0.050 0.002 1.548 0.155
U 54600.18262 339.6 19.74 0.15 0.047 0.007 1.675 0.167
U 54660.40377 993.9 19.70 0.07 0.049 0.003 1.598 0.160
U 57644.49897 1932.3 20.38 0.10 0.026 0.002 0.435 0.044
U 57644.46425 982.6 20.57 0.17 0.022 0.003 0.435 0.044
B 54587.45449 70.6 20.04 0.56 0.035 0.018 1.478 0.148
B 54588.05722 70.6 20.73 1.02 0.018 0.017 1.268 0.127
B 54592.33277 55.9 19.36 0.36 0.065 0.021 2.325 0.232
B 54593.61173 114.9 20.15 0.49 0.031 0.014 1.991 0.199
B 54594.27658 16.5 21.72 5.05 0.007 0.034 1.991 0.199
B 54595.01220 139.5 19.80 0.32 0.043 0.013 1.157 0.116
Table 2: UV, optical and X-ray photometry from Swift. Since the Swift X-ray flux measurements were performed for each individual snapshot, there are UVOT entries with repeated values.

Appendix B SED Posterior distributions

We present the joint and marginal posterior distributions of the accretion disc parameters obtained in the SED fit in Fig. 9 (see Sec. 3.3). We used corner.py (Foreman-Mackey, 2016) to visualise the MCMC chains.

Figure 9: Posterior probability distributions for the accretion disc parameters. Colour scale contours show the joint probability for every combination of parameters. Contours represent the 0.5, 1 , 2 and 3 levels. Marginal posterior distributions are shown as histograms with the median and 1 marked as dashed lines.
Comments 0
Request Comment
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
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description