Multi-band observations of Swift J0840.7-3516: a new transient ultra-compact X-ray binary candidate
F. Coti Zelati, A. de Ugarte Postigo, T. D. Russell, A. Borghese, N. Rea, P. Esposito, G. L. Israel, S. Campana
AAstronomy & Astrophysics manuscript no. J0840_ms © ESO 2021February 17, 2021
Multi-band observations of Swift J0840.7 − F. Coti Zelati , , A. de Ugarte Postigo , , T. D. Russell , A. Borghese , , N. Rea , , P. Esposito , , G. L. Israel , andS. Campana Institute of Space Sciences (ICE, CSIC), Campus UAB, Carrer de Can Magrans s / n, E-08193, Barcelona, Spaine-mail: [email protected] Institut d’Estudis Espacials de Catalunya (IEEC), Carrer Gran Capità 2–4, E-08034 Barcelona, Spain Instituto de Astrofísica de Andalucía (IAA-CSIC), Glorieta de la Astronomía s / n, E-18008 Granada, Spain DARK, Niels Bohr Institute, University of Copenhagen, Lyngbyvej 2, DK-2100 Copenhagen Ø, Denmark INAF, Istituto di Astrofisica Spaziale e Fisica Cosmica, Via U. La Malfa 153, I-90146 Palermo, Italy Scuola Universitaria Superiore IUSS Pavia, Palazzo del Broletto, piazza della Vittoria 15, I-27100 Pavia, Italy INAF–Istituto di Astrofisica Spaziale e Fisica Cosmica di Milano, via A. Corti 12, I-20133 Milano, Italy INAF–Osservatorio Astronomico di Roma, via Frascati 33, I-00078 Monteporzio Catone, Italy INAF-Osservatorio Astronomico di Brera, Via Bianchi 46, I-23807 Merate (LC), ItalyFebruary 17, 2021
ABSTRACT
We report on multi-band observations of the transient source Swift J0840.7 − Neil Gehrels Swift Observatory . The outburst episode lasted just ∼ ∼ . × − erg cm − s − at peak down to ∼ . × − erg cm − s − in quiescence (0.3–10 keV). Such a marked andrapid decrease in the flux was also registered at UV and optical wavelengths. In outburst, the source showed considerable aperiodicvariability in the X-rays on timescales as short as a few seconds. The spectrum of the source in the energy range 0.3–20 keV was welldescribed by a thermal (blackbody-like) component plus a non-thermal (power law-like) component, and softened considerably asthe source returned to quiescence. The spectrum of the optical counterpart in quiescence showed broad emission features associatedmainly with ionized carbon and oxygen, superposed on a blue continuum. No evidence for bright continuum radio emission wasfound in quiescence. We discuss possible scenarios for the nature of this source, and show that the observed phenomenology pointsto a transient ultra-compact X-ray binary system. Key words. methods: data analysis — methods: observational — techniques: spectroscopic – X-rays: binaries — X-rays: individual(Swift J0840.7 −
1. Introduction
Low-mass X-ray binaries (LMXBs) are systems where a com-pact object (a neutron star or a stellar-mass black hole) accretesmatter from a low-mass ( M (cid:46) M (cid:12) ) donor star (Frank et al.2002). Many LMXBs spend most of their time in quiescencewith little or no mass accretion taking place, but they can un-dergo sporadic accretion outbursts where the X-ray luminositycan increase up to 5–6 orders of magnitude above quiescence, upto L X ∼ − erg s − . The quiescent phase can be as longas decades, whereas the outbursts typically last from weeks tomonths to years (e.g., Tetarenko et al. 2016). Such a transient be-haviour is commonly ascribed to thermal-viscous instabilities inthe accretion disk around the compact object (e.g., Lasota 2001).Ultra-compact X-ray binaries (UCXBs) are a sub-class ofLMXBs that are characterised by orbital periods shorter than80 min and a hydrogen-deficient donor star such as a non-degenerate helium star or a white dwarf (Nelson et al. 1986;Savonije et al. 1986; for a review, see Nelemans & Jonker 2010).Currently, there are 24 known UCXBs with a measured orbitalperiod and a further 15 systems have been classified as candi-date UCXBs. To date, all confirmed UCXBs have been foundto harbour a neutron star accretor, while for some candidates the nature of the accretor has not been established conclusively (e.g.,Bahramian et al. 2017; Sazonov et al. 2020).Due to the peculiar nature of the donor star in UCXBs, thechemical composition of the accreting material in these systemsis expected to be significantly di ff erent from that of LMXBs har-bouring main sequence donor stars. In fact, observations haveshown that the disk material in UCXBs contains elements suchas helium, carbon, and oxygen (and in some cases neon andmagnesium) in large amounts (e.g. Nelemans et al. 2004, 2006;Werner et al. 2006). On the other hand, most LMXBs show sub-stantial emission associated mainly with hydrogen.On 2020 February 5 at 06:35:51 UT, the Burst Alert Tele-scope (BAT) on board of the Neil Gehrels Swift Observatory triggered and located a burst of gamma rays (Evans et al. 2020).Observations with the
Swift
X-ray Telescope (XRT) and the GasSlit Camera (GSC) on board of the
Monitor of All-sky X-rayImage ( MAXI ) revealed an X-ray counterpart to the burst (Ni-wano & MAXI Team 2020), showing considerable flaring activ-ity over the subsequent hours, as detected using the
Swift / XRTand the
Neutron Star Interior Composition Explorer ( NICER )(Iwakiri et al. 2020). These properties and the source proxim-ity to the Galactic plane (Galactic latitude of b ∼ . ◦ ), castdoubts on an interpretation in terms of a “canonical” gamma-ray Article number, page 1 of 11 a r X i v : . [ a s t r o - ph . H E ] F e b & A proofs: manuscript no. J0840_ms burst of extragalactic origin, suggesting a transient source withinthe Galaxy instead. The source was dubbed Swift J0840.7 − ∼
20 hafter the initial burst detection (Melandri et al. 2020; Malesaniet al. 2020).This paper presents the results of follow-up X-ray, UV, op-tical and radio observations of Swift J0840.7 −
2. Observations and data reduction
Table 1 reports a journal of all pointed X-ray observations ofSwift J0840.7 − = h m s .
94, decl. = –35 ◦ (cid:48) (cid:48)(cid:48) . (cid:48)(cid:48) ; Malesani et al. 2020). The spectral analysis was performedusing the xspec fitting package (Arnaud 1996). Hereafter, all un-certainties are quoted at a confidence level (c.l.) of 1 σ , unlessotherwise stated. The alert that led to the discovery of J0840 started with anon-board BAT image trigger (Id. 954304, Evans et al. 2020)and the prompt spacecraft slew. The spectral products and theother pieces of information discussed in Section 3.1 were all ob-tained from a standard analysis of the BAT event data using theinstrument-specific tasks in the FT ools software package v6.28.
The
Swift / XRT (Burrows et al. 2005) monitored extensivelyJ0840 since the burst trigger. Observations were performed witha daily cadence during the first two weeks, about once per weekfrom late February until the end of August 2020 and once ev-ery two weeks since then and until one year after the discoveryof the source. Except for the two prompt post-burst observationsin the windowed timing mode (WT; readout time of 1.8 ms), alldata were acquired in the photon counting mode (PC; 2.51 s).The data sets were processed using standard screening criteria.We extracted source event lists and spectra using a rectangularregion of 40 ×
20 pixels centered on the source for WT-mode dataand a circle with a radius of 20 pixels (1 XRT pixel (cid:39) ×
20 pixels box far from the source for WT-mode data and an annulus centred on source with radii of 40and 80 pixels (free of sources) for PC-mode data. Background-subtracted spectra were extracted separately for the observationsbetween 2020 February 5–9, while a combined spectrum wascreated from all the subsequent pointings, where the source at-tained a steady X-ray intensity level (total exposure time of ∼ NuSTAR (Harrison et al. 2013) observed J0840 on 2020 Febru-ary 8–9 for an on-source exposure time of ∼
42 ks. Data wereprocessed and analysed using nustardas v. 2.0.0 and caldb v. 20200813. For both focal plane modules (FPMs), the sourcesignal-to-noise ratio (S / N) was maximum over the 3–20 keV en-ergy range (S / N ∼
34 for the FPMA and ∼
33 for the FPMB).The subsequent analysis was thus limited to this energy inter-val. Source photons were collected within a circle of radius 80arcsec, while background photons were extracted from a circleof the same size far from the source and on the same detector. Weextracted background-subtracted light curves and spectra using nuproducts . Light curves from each FPM were then combinedto increase the S / N. Spectra of both FPMs were grouped to haveat least 50 counts per energy channel.
The X-ray Timing Instrument (XTI) on board
NICER (Gendreauet al. 2012) monitored J0840 intensively over the first week sincethe burst trigger. We processed and screened the datasets using nicerdas v7a and standard criteria. We additionally extracted thelight curves of all observations in the range 12–15 keV with atime bin of 1 s, and removed instances of background flares byapplying intensity filters. We estimated the background countrate and extracted the background spectra for each observationusing the tool nibackgen
Swift /UVOT
The
Swift
Ultraviolet and Optical Telescope (UVOT; Rominget al. 2005) observed J0840 all along the outburst using di ff erentUV and optical filters. Source photons were extracted adoptinga circle of radius 3 arcsec centered on the position of the target.Background photons were collected from a closeby circle of ra-dius 10 arcsec. Photometry was carried out for all observationsusing the uvotsource task, applying aperture corrections. The Gran Telescopio Canarias (GTC) observed J0840 using theOptical System for Imaging and low-Intermediate Resolution In-tegrated Spectroscopy (OSIRIS; Cepa et al. 2000). An imagingvisit was performed starting on February 18 at 23:46:31 UTC(MJD 58897.991), and consisted of three sets of 3 ×
80 s imagesin the g (cid:48) , r (cid:48) and i (cid:48) filters (see Table 2). The source was observedat high airmass (2.3) and poor seeing (between 1.4 and 2 (cid:48)(cid:48) ).Following this observation, we performed spectroscopystarting on February 20 at 23:55:20 UTC (MJD 58899.997). Theobservations consisted of 3 ×
900 s exposures using a slit widthof 1.0 (cid:48)(cid:48) and the grism R1000B, covering the 3700–7800 Å wave-length range. The spectrophotometric calibration was performedwith respect to the G191-B2B reference star. The reduction andcalibration of the spectra were performed with custom scripts
Article number, page 2 of 11. Coti Zelati et al.: Swift J0840.7 − Table 1.
Log of pointed X-ray observations, spectral parameters and X-ray fluxes of J0840.
Instrument a Obs.ID Start Stop Exposure Count rate b kT BB R BB c F X , obs d F X , unabs d YYYY-MM-DD hh:mm:ss (TT) (ks) (counts s − ) (keV) (km) ( × − erg cm − s − ) Swift / XRT (WT) 00954304000 2020-02-05 06:38:33 2020-02-05 06:40:03 0.1 30.5 ± ± . + . − . ± ± Swift / XRT (WT) 00954304001 2020-02-05 07:52:25 2020-02-05 16:11:53 0.8 22.8 ± ± ± ± ± Swift / XRT (PC) 00954304001 2020-02-05 10:00:19 2020-02-05 16:17:53 2.7 0.52 ± ± . + . − . ± ± NICER / XTI 2201010101 2020-02-05 14:43:34 2020-02-05 21:09:00 1.3 2.15 ± Swift / XRT (PC) 00954304002 2020-02-06 11:13:26 2020-02-06 14:40:52 1.0 0.061 ± ± . + . − . ± ± NICER / XTI 2201010102 2020-02-06 15:27:10 2020-02-06 23:20:50 1.8 3.78 ± Swift / XRT (PC) 00954304005 2020-02-06 22:29:17 2020-02-07 03:02:50 3.3 0.176 ± ± . + . − . ± ± NICER / XTI 2201010103 2020-02-07 00:44:31 2020-02-07 21:03:51 7.2 1.52 ± Swift / XRT (PC) 00954304003 2020-02-07 14:12:57 2020-02-07 16:17:52 2.8 0.084 ± ± . + . − . ± ± NICER / XTI 2201010104 2020-02-08 03:02:14 2020-02-08 23:52:00 2.7 1.04 ± NuSTAR / FPMA 90601304002 2020-02-08 07:01:09 2020-02-09 06:56:09 42.4 0.050 ± e ± . + . − . ± ± NuSTAR / FPMB 90601304002 2020-02-08 07:01:09 2020-02-09 06:56:09 41.5 0.045 ± e ± . + . − . ± ± Swift / XRT (PC) 00954304004 2020-02-08 10:55:58 2020-02-08 17:36:53 3.8 0.060 ± e ± . + . − . ± ± NICER / XTI 2201010105 2020-02-09 01:08:35 2020-02-09 23:06:40 12.7 0.86 ± Swift / XRT (PC) 00954304007 2020-02-09 07:33:15 2020-02-09 17:18:53 4.1 0.029 ± ± . + . − . ± ± Swift / XRT (PC) 00954304008–067 2020-02-10 13:52:00 2021-02-10 17:28:16 207.2 0.0028 ± f ± . + . − . ± ± ( a ) The instrumental setup is indicated in brackets: PC = photon counting, WT = windowed timing. ( b ) Average net count rate. It is in the 0.3–10 keV energy range, except for
NICER (0.5–5 keV) and
NuSTAR (3–20 keV). ( c ) The blackbody radius is calculated assuming a distance of 10 kpc. ( d ) All fluxes are in the 0.3–10 keV energy range. ( e ) Datasets of theseobservations were fitted together. ( f ) Datasets of these observations (59 in total) were merged.
Table 2.
Log of UV / optical photometric observations and source mag-nitudes or upper limits. Instrument(Filter) Start Exposure Magnitude a (TT) (s) Swift / UVOT(
UVW
2) 2020-02-05 06:39:45 246 17.32 ± Swift / UVOT(
UVW
2) 2020-02-05 07:53:45 261 17.02 ± Swift / UVOT(
UVW
2) 2020-02-05 08:15:00 400 17.32 ± Swift / UVOT(
UVW
2) 2020-02-05 10:00:52 197 18.44 ± Swift / UVOT(
UVW
2) 2020-02-05 11:16:12 598 18.51 ± Swift / UVOT(
UVW
2) 2020-02-05 14:16:36 93 > Swift / UVOT(
UVW
2) 2020-02-05 16:13:07 358 > Swift / UVOT(
UVW
2) 2020-02-06 11:13:28 548 > Swift / UVOT(
UVW
2) 2020-02-06 22:29:06 3372 20.27 ± Swift / UVOT(
UVW
2) 2020-02-07 14:13:01 912 > Swift / UVOT(
UVW
2) 2020-02-08 10:56:02 1250 > Swift / UVOT(
UVW
2) 2020-02-09 07:33:19 4041 > GTC / OSIRIS( r ) 2020-02-18 23:47:40 3 × ∼ ( a ) For
Swift / UVOT, magnitudes are reported only for observations performed in outburst.Values are in the Vega system. 5 σ upper limits are quoted in case of non-detection. based on the iraf package. The spectra were then combined andaveraged to increase the S / N. The Australia Telescope Compact Array (
ATCA ) observed thefield of J0840 on 2020 February 11, between 14:50 and 19:10UT (MJD 58890.618–58890.799), for a total on-source timeof ∼ −
638 wasused for bandpass and flux calibration, while PKS B0826 − casa ; McMullinet al. 2007). The field was imaged using a natural weighting (aBriggs robust parameter of 2) to maximise the sensitivity of theobservations. The source was not detected in our radio observa-tions. The reported 3- σ upper-limit on the radio flux density wasdetermined as 3 times the local rms over the source position.
3. Results
An image trigger is generated when the BAT detects a rateincrease and then finds a point source in a readily-performedimage reconstruction of the sky. In the case of J0840, thepoint source had a S / N = = h m s .
1, decl. = –35 ◦ (cid:48) (cid:48)(cid:48) . duration (the time interval containing 90%of the counts), as computed by the Bayesian blocks algorithm battblocks , was 210 ±
20 s.Several simple models can describe the BAT spectrum (in theT duration range) of J0840 in the 15–150 keV energy range(Evans et al. 2020; Stamatikos & Swift-BAT Team 2020). Forexample, a power law (photon index of Γ = . ± .
1, re-duced chi squared χ r = .
26 for 56 degrees of freedom, d.o.f.),an optically-thin thermal bremsstrahlung ( kT = ± χ r = .
06 for 56 d.o.f.), a blackbody (temperature of kT BB = . ± . χ r = .
12 for 56 d.o.f.), or a cut-o ff power-law( Γ = . ± .
7, cut-o ff energy of E c = + − keV, χ r = .
03 for55 d.o.f.). For the latter model, which gives the best fit, the T fluence was 1 . + . − . × − erg cm − . The X-ray time series reveal considerable variability ontimescales of minutes in all observations performed in outburst,in the form of drops in the count rates as well as erratic flares(Figure 1). We checked the
NICER and
NuSTAR time series forthe presence of periodic dips adopting di ff erent time bins, butfound none. To study the aperiodic time variability, we extractedthe power density spectra for all NICER and
NuSTAR data sets.We sampled the time series of each observation with a time binof 2 − s and calculated power spectra into time intervals of length2 s using the fractional rms-squared normalization. We then re-binned the resulting average spectrum as a geometric series witha step of 0.3. The power denisty spectra are shown in Figure 2.Except for the first NICER observation, the aperiodic variabil-ity of J0840 is characterised in all cases by a noise componentthat increases towards lower frequencies below ≈ P ( ν ) = K + C ν − β , accounting for the white noise (first term)and red noise (second term) components. The best-fitting valuesfor the power-law index and the fractional r.m.s. variability am- Article number, page 3 of 11 & A proofs: manuscript no. J0840_ms C oun t r a t e ( c oun t s s - ) . − k e V S w ift X R T . − k e V N I C ER X T I × × × × × × − − k e V Time (MJD) N uS T AR FP M A + B C O = . E − Fig. 1.
Multi-instrument background-subtracted X-ray light curves of J0840 between 2020 February 5–10, binned at 800 s. The gray shadedregion in the top panel indicates the confidence interval (at 3 σ ) for the quiescent Swift / XRT net count rate computed by stacking all observationsperformed since 2020 February 10 (see Table 1). In most cases, the marker size is larger than the error bars. − − χ Fourier Frequency (Hz) P o w e r d e n s it y [(r m s / m ea n ) H z - ] − χ Fourier Frequency (Hz) − − χ Fourier Frequency (Hz) − − χ Fourier Frequency (Hz) − − χ Fourier Frequency (Hz) − − χ Fourier Frequency (Hz)
NICER (2020-02-05) χ = 0.44 (35) NICER (2020-02-06) β = 1.02±0.06 r.m.s. = 0.69±0.02 χ = 0.59 (35) NICER (2020-02-07) β = 0.91±0.05 r.m.s. = 0.41±0.02 χ = 0.66 (35) NICER (2020-02-08) β = 1.06±0.21 r.m.s. = 0.22±0.05 χ = 0.93 (35) NICER (2020-02-09) β = 0.97±0.09 r.m.s. = 0.32±0.03 χ = 0.52 (35) NuSTAR (2020-02-08/09) β = 0.85±0.07 r.m.s. = 0.92±0.09 χ = 0.88 (35)Fourier Frequency (Hz) P o w e r d e n s it y [(r m s / m ea n ) H z - ] P o w e r d e n s it y [(r m s / m ea n ) H z - ] P o w e r d e n s it y [(r m s / m ea n ) H z - ] Fourier Frequency (Hz) Fourier Frequency (Hz) P o w e r d e n s it y [(r m s / m ea n ) H z - ] P o w e r d e n s it y [(r m s / m ea n ) H z - ] Fourier Frequency (Hz) Fourier Frequency (Hz) Fourier Frequency (Hz)
Fig. 2.
Power density spectra extracted from
NICER (0.5–5 keV) and
NuSTAR (3–20 keV) data. The red solid lines mark the best-fitting model tothe power spectra (see text for details). The best-fitting values for the power-law index, the rms variability amplitudes below 1 Hz and the χ r valueswith the d.o.f. of the fits are also shown. Post-fit residuals are shown in the bottom panels. Uncertainties are at 1 σ c.l. plitudes evaluated at frequencies below 1 Hz for all observationsare shown in Figure 2. We observe large values of the variabilityamplitude, up to 0 . ± .
09 in the
NuSTAR observation. Simi-lar values were derived when averaging power spectra computedover shorter time lengths (down to 2 s), and / or when applying di ff erent geometrical rebinning factors to the averaged spectrum.We found no evidence for variability features over restricted fre-quency ranges (e.g., quasi-periodic oscillations) in any of thepower spectra. Article number, page 4 of 11. Coti Zelati et al.: Swift J0840.7 − P u l s e d f r a c t i o n li m i t( % ) Fig. 3. σ upper limits on the pulsed fraction for any coherent signalwithin the frequency range 0.01–620 Hz, estimated by means of accel-erated searches on the combined NICER datasets over the 0.5–5 keVenergy range (see Section 3.2 for details).
We searched for coherent signals in the
NICER light curvesin the 0.5–5 keV energy range by adopting the recipe outlinedby Israel & Stella (1996) and taking into account also the e ff ectsof signal smearing introduced by the presence of a first periodderivative component ˙ P induced by an orbital motion. We cor-rected the photon arrival times by a factor −
12 ˙ PP t for a grid ofabout 1800 points in the range 5 × − < | ˙ PP (s − ) | < − (seeRodríguez Castillo et al. 2020 for details). No significant peakwas found. Figure 3 shows the 3 σ upper limits to the pulsedfraction (defined as the semi-amplitude of the sinusoid dividedby the average count rate) obtained in two cases: maximizingthe Fourier resolution T − (where T obs is the observation length;black points), and extending the search down to periods of theorder of milliseconds (red points). In the best cases, we obtainedupper limits around 8–12% and 6–7% in the period ranges 1.6–50 ms and 50 ms–1 s, respectively.We also inspected the Swift / XRT data acquired in quiescencefor the presence of a possible modulation of the soft X-ray emis-sion at periods in the range from days to months. We found noevidence for any periodicity on these time scales.
We modelled the broad-band spectrum extracted using quasi-simultaneous data acquired by
NuSTAR and
Swift / XRT (obs.ID:00954304004). We included a renormalization factor in the mod-elling to account for intercalibration uncertainties, as well as forpossible di ff erences in the X-ray flux registered by the two obser-vatories ( NuSTAR continued observing the source for ∼ Swift / XRT observation, over a time intervalof significant X-ray variability; see Table 1 and Figure 1). Forall the models tested, the correction was always smaller than10%. Single-component models such as a blackbody (BB), abremsstrahlung, a multicolour blackbody emission model froman accretion disk ( diskbb in xspec ), or optically thin thermalplasma emission models ( mekal and apec ; Smith et al. 2001)are rejected by the data ( χ r ≥ . χ r = .
16 for 99 d.o.f.). However, based on F -tests, we deem that the broad-band spectrum is better described − − N o r m a li ze d c o un t ss − k e V − c m − SwiftNuSTAR FPMANuSTAR FPMB − . . . R e s i du a l s ( σ ) Fig. 4.
Broad-band X-ray spectrum of J0840 in outburst, extracted fromquasi-simultaneous
Swift / XRT and
NuSTAR data. by a double-component model comprising also a thermal com-ponent at low energies, such as an absorbed BB + PL model.We derived χ r = .
04 for 97 d.o.f. and an F -test probabilityof chance improvement of ∼ × − . The best-fitting parame-ters are N H = . + . − . × cm − , kT BB = . ± .
06 keV, R BB = . + . − . km (assuming a distance of 10 kpc), Γ = . + . − . .The observed flux over the 0.3–20 keV energy band was (3 . ± . × − erg cm − s − A double-blackbody model gives in-stead a much worse description of the data ( χ r = .
56 for 97d.o.f.). More sophisticated models are beyond the scope of thiswork.We did not find evidence for the presence of broad emis-sion / absorption features in the spectrum. We included an addi-tional Gaussian component with centroid allowed to vary in theenergy range 6.4–6.97 keV and set 3 σ upper limits of 40–80 eVon the equivalent width associated with any iron emission linewith widths in the range 0.1–0.4 keV.To assess the overall evolution of the source spectrum andflux along the outburst decay, we then fit the absorbed BB + PLmodel to the
Swift / XRT spectra extracted separately from eachobservation. The limited energy range covered by the XRT to-gether with the relative scarce photon counting statistics at highenergies do not allow us to track in detail the evolution of thethermal and nonthermal spectral components. In the following,we assume that the power-law component is present all alongthe outburst decay down to quiescence and that its slope doesnot vary in time (see Section 4 for a more in-depth discussion onthis assumption). For the joint fits, we thus fixed Γ to the valuederived from the analysis of the broad-band spectrum. The N H was also held fixed to the above value. All other parameters wereallowed to vary. We obtained χ r = .
08 for 674 d.o.f.The results of the spectral analysis are reported in Table 1.We observe a clear overall softening in the spectrum, withthe blackbody temperature decreasing from an initial value of ∼ ∼ ∼ . × − to ∼ × − erg cm − s − over the first 3 hr after the trigger, andthen down to a steady value of ∼ . × − erg cm − s − overthe subsequent few days (here and in the following, all fluxes arequoted in the 0.3–10 keV energy range). J0840 has been linger-ing at such X-ray flux level since then.The field of the source was also covered by XMM-Newton slew observations eight times between November 2004 and May2019, before the outburst onset. J0840 was undetected in allthese pointings. The deepest limit on the count rate at the source
Article number, page 5 of 11 & A proofs: manuscript no. J0840_ms position, < − (3 σ ; 0.2–12 keV), translates into an ob-served flux of < × − erg cm − s − , assuming the Swift / XRTquiescent spectrum. This limit is compatible with the quiescentflux measured in the stacked
Swift / XRT observations performedsince 2020 February 10.
The UV counterpart was detected using the
Swift / UVOT at amagnitude of ∼
17 soon after the burst trigger (here and in thefollowing, magnitudes are reported in the Vega system). It de-cayed to ∼ ∼ σ upper limits of > r (cid:39) . r (cid:39) . / OSIRIS about two weeks later, when the source had al-ready returned to quiescence according to the X-ray observations(see Table 2). The source appears slightly blended with a nearbystar, which may a ff ect the reliability of the photometry. The i -band image is also a ff ected by a blooming spike from a brightred source located about 25 arcsec North. J0840 is detected at anaverage magnitude r (cid:39)
21 in these images.From the Foight et al. (2016) N H – A V relation and the N H = . × cm − derived from the spectral fits of the per-sistent emission, we can estimate A V ∼ . E ( B − V ) ∼ . A UVW ∼ . A r (cid:48) ∼ . UVW ∼ r ∼ / optical flux ratios of F X / F UV ≈
15 and F X / F OPT ≈ The bottom panel of Figure 5 shows the flux-calibrated opti-cal spectrum of J0840 in quiescence, corrected for extinction( E [ B − V ] ∼ .
52; see Section 3.4.1). The spectral shape ofthe continuum emission of J0840 is well fit by a blackbodymodel with a temperature of ≈ iii and O ii ; 6565–6620 Å, which we attribute to a blend ofC ii and O ii lines; and 7210–7270 Å (C ii ) (see Nelemans et al.2004). The profiles of these features are very similar to those de-tected in the UCXB 4U 0614 +
091 (Figure 5). A weaker broadfeature is also seen in the range 7050–7130 Å, which we tenta-tively associate with a blend of lines around He i at 7065 Å. Theequivalent widths of these features are reported in Table 3. Table 3.
Identification and equivalent widths of the most prominentoptical emission lines of J0840.
Line Wavelength range EW(Å) (Å)C iii + O ii ± ii + O ii ± i ± ii ± During the low-luminosity quiescence phase, we did not detecta radio counterpart at the position of J0840 in our ATCA obser-vations, taken on 2020 February 11 (see Figure 6 for the radioimage of the field around J0840). We measured 3- σ upper-limitson the radio flux density of 27 µ Jy / beam and 30 µ Jy / beam at 5.5and 9 GHz, respectively. Combining the two observing bandstogether provided a 3- σ upper limit of 18 µ Jy / beam at a centralfrequency of 7.25 GHz.
4. Discussion
In the following, we discuss possible scenarios for the nature ofJ0840, and show that the observed phenomenology is consistentwith a classification as a transient UCXB.The rapid luminosity decay observed in J0840 argues againsta classification as a tidal disruption event (TDE). The luminosityof TDEs, in fact, declines typically over a much more extendedtimescale of months to years, in many cases at a rate that is shal-lower than the canonical power law of the form L ( t ) ∝ t − / pre-dicted by standard theories for TDEs (Auchettl et al. 2017). Theabove properties point instead to a binary system in our Galaxy.J0840 could be then a magnetic cataclysmic variable (mCV), ahigh-mass X-ray binary (HMXB), or an LMXB.mCVs are systems where a magnetic white dwarf ( B WD (cid:38) G) accretes matter from a late-type, low-mass star that over-flows its Roche lobe (for reviews, see Mukai 2017; de Martinoet al. 2020). The X-ray spectrum of mCVs consists typically ofmulti-temperature, optically-thin thermal plasma emission pro-duced in the accretion columns, accounting for the iron complexoften observed in these systems in the energy range 6–7 keV.An ubiquitous feature in the spectra of mCVs is a broad ironfluorescent line at 6.4 keV, with typical equivalent widths in therange ∼ <
100 eV. In this respect, the broad-band spectrum of J0840 inoutburst is clearly di ff erent from what expected from mCVs. Wethus consider unlikely J0840 to be a mCV.HMXBs are systems where a compact object accretes matterfrom a massive ( (cid:38) M (cid:12) ) early-type (O or B) donor star (for areview, see e.g. Reig 2011). The continuum emission of J0840in the optical band can be adequately described by a blackbodymodel with a temperature of about 8000 K. If such emission isindeed dominated by radiation from the donor star (as observedin HMXBs), then the above value for the temperature would bemore consistent with an A-type star rather than with an O or B-type star. The large increase observed in the optical magnitudeof J0840 from quiescence to outburst (see above) would be alsoat odds with what is seen in transient HMXBs, where the bulk ofthe optical / UV emission is provided in the form of steady radia-
Article number, page 6 of 11. Coti Zelati et al.: Swift J0840.7 − Fig. 5.
Top : Optical images acquired by GTC / OSIRIS on 2020 February 18. Saturation blooming due to a nearby star is evident in the i -bandimage. The position of J0840 is marked with a red circle. Bottom : Extinction-corrected optical spectrum of J0840 in quiescence plotted togetherwith a spectrum of the UCXB 4U 0614 +
091 acquired in archival observations with VLT / X-Shooter. The original data are indicated in gray, thesame data convolved with a Gaussian function of width 5 Å are marked in black. The best fitting blackbody model for the continuum of J0840 isoverplotted with an orange line. The most prominent emission features of J0840 are labelled. The carbon and oxygen lines identified by Nelemanset al. (2004) are marked using brown and blue dotted lines. The broad absorption feature at ∼ absorption feature. tion by the donor both in outburst and in quiescence, with only amodest variable contribution from an accretion disk.We then discuss the scenario of a LMXB. The shape of theoptical spectrum of J0840 together with the presence of emis-sion features in quiescence argues against a classification as asymbiotic X-ray binary (SyXB), where a neutron star accretesmass from the stellar wind of an evolved M-type giant donor star (e.g. Masetti et al. 2006, 2007). This then leaves the scenario inwhich J0840 is a LMXB where the compact object accretes fromthe donor via Roche-lobe overflow. Strong emission features have been detected in one SyXB, GX 1 + − erg s − ), that is, in arange where spectral features produced by accretion are expected toemerge over the contribution of the donor (see, e.g., Chakrabarty &Roche 1997). Article number, page 7 of 11 & A proofs: manuscript no. J0840_ms
Swift J0840.7-3516 D ec li n a ti on ( J ) Right Ascension (J2000)
Flux density ( µ Jy/beam)
Fig. 6.
Radio image of the field around J0840. The position of J0840is marked using a green circle with an enlarged radius of 3 arcsec fordisplaying purpose. The synthesized beam is shown in the bottom leftcorner of the panel.
A key diagnostic on the nature of J0840 is provided by itsoptical spectrum in quiescence. This spectrum shows no sign forthe presence of hydrogen emission lines, as instead commonlyobserved in LMXBs with a hydrogen-rich main sequence donorstar. On the other hand, it displays broad emission lines that aregenerally associated with partially ionized carbon (C ii and C iii )and oxygen (O ii ). This phenomenology suggests that a struc-tured accretion disk almost purely made out of carbon and oxy-gen was still present around the accretor soon after the end of theoutburst. In this respect, the optical spectrum of J0840 closelyresembles that of the UCXBs 4U 0614 + −
624 and4U 1626 −
67 (Nelemans et al. 2004, 2006; see also Figure 5).Based on the strong analogy with the above systems, we proposethat J0840 is a new transient UCXB candidate with a carbon-oxygen white dwarf as donor star.The broad-band X-ray spectral shape of J0840 (see Sec-tion 3.3) is typical of LMXBs in outburst, and may be interpretedin terms of repeated inverse Compton up-scattering of soft ther-mal photons onto a population of hot thermal electrons. In thisframework, the main production sites of the thermal photonswould be the inner regions of the optically-thick, geometrically-thin accretion disk (see Shakura & Sunyaev 1973). As discussedin more detail in the next section, we are unable to identify con-clusively the nature of the accretor (black hole or neutron star)with current data. In the case of a neutron star accretor, additionalcontributions to the thermal X-ray emission may be provided bythe accretion-heated star surface (see, e.g., Zampieri et al. 1995;Wijnands et al. 2015) and / or the so-called “boundary layer” thatforms when the disk extends all the way down to the neutronstar such that the in-flowing material spreads out over the starsurface. However, due to the available photon counting statistics,we are not able to discern the possible contribution of multiplecomponents to the observed thermal emission. The location andgeometry of the hot thermal electrons that Compton up-scatterthe thermal photons is instead debated at present: they may bedistributed in an extended cloud above the disk (the so-called “corona”; see e.g., Kara et al. 2019 and references therein) or ina hot inner flow close to the accretor (e.g., Done et al. 2007).The overall spectral softening observed in J0840 along itsoutburst decay may be explained in broad terms by the disk in-stability model for transient UCXBs (and LMXBs in general; seee.g. Hameury & Lasota 2016): the decrease in the temperaturereflects the gradual transition of the disk from a hot ionized stateto a colder neutral state as the mass transfer rate from the donorstar decreases and the system approaches quiescence. The largeincrease of the X-ray flux observed in J0840 might suggest that alarge portion of the disk is brought in the hot state in this system(Hameury & Lasota 2016). This process, in fact, should be par-ticularly e ffi cient in UCXBs, since only a relatively small-sizeddisk can physically fit into their tight binary orbits. In the caseof a neutron star accretor for J0840, an additional contributionto the observed spectral softening might be ascribed to low-levelaccretion onto the star surface. We note, however, that the scarcephoton counting statistics currently available in quiescence fromthe Swift / XRT data ( ∼
540 net counts) precludes a detailed inves-tigation of the evolution of both spectral components throughoutthe outburst and down to quiescence. In this respect, our assump-tion of an unchanged power-law slope is certainly simplistic.The absence of the iron K α fluorescent line in the X-rayspectrum of J0840 in outburst can be tentatively used as a tracerof the chemical composition of the disk and donor star, andmay provide further support to our classification as a UCXB.In many LMXBs, X-ray photons intercepting the inner regionsof the disk are absorbed by highly ionised elements in the diskand re-emitted along the line of sight, producing an additionalemission component known as X-ray reflection. The reprocessedemission shows multiple emission features superimposed to thecontinuum at energies corresponding to transitions in the highlyionised atomic species. Their profiles are broadened and skewedby rotation of the disk material as well as strong Doppler, spe-cial relativistic and general relativistic e ff ects inside the gravi-tational well of the compact accretor (Fabian et al. 1989). Onthe one hand, a broad fluorescent Fe K α line is often detected inLMXBs with hydrogen-rich donor stars (see e.g. Cackett et al.2010). On the other hand, this feature is expected to be signif-icantly attenuated in UCXBs with an anomalous abundance ofcarbon and oxygen, owing to the screening of the presence ofiron and other heavy elements in the carbon-oxygen-dominateddisk material (Koliopanos et al. 2021 and references therein; seealso in’t Zand et al. 2007). Indeed, Koliopanos et al. (2021)found no evidence of emission associated with the iron line inthe X-ray spectra of seven known UCXBs, with upper limits onthe equivalent widths of 8–60 eV (3 σ ). Based on these values,they estimated an oxygen-to-iron ratio which is at least an or-der of magnitude larger than the Solar value, and ascribed thisproperty to the presence of a carbon-oxygen (or even oxygen-neon-magnesium) donor star. In this framework, the upper lim-its inferred on the equivalent width of any iron line in J0840,40–80 eV (see Section 3.3), although not as constraining as thelimits derived by Koliopanos et al. (2021) for other UCXBs, areanyway smaller than the values observed in many hydrogen-richLMXBs, and still support a classification of J0840 as a UCXBwith a carbon-oxygen white dwarf.The enhanced UV / optical emission detected at the outburstpeak and its rapid fading during the earliest outburst phaseshints at a possible correlation with the X-ray emission, and maybe interpreted in terms of enhanced irradiation of the outer re-gions of the disk and the donor star (in analogy with mostLMXBs). The optical counterpart in quiescence is much fainterand blue, suggesting that the optical emission is dominated by Article number, page 8 of 11. Coti Zelati et al.: Swift J0840.7 − the disk in this phase. Assuming a distance in the range 5–10 kpc(see Section 4.1), the dereddened magnitude of J0840 in quies-cence translates into an absolute magnitude for the donor starof M r (cid:48) > . D = M r (cid:48) > . D =
10 kpc),where the values should be considered as upper limits owing tothe unknown contribution of the disk and the donor star to the op-tical emission in quiescence. These values make the donor starof J0840 among the faintest in the population of known LMXBs,and are similar to those estimated for other UCXBs (e.g., Bassaet al. 2006).
Determining the nature of the compact object in an unclassifiedLMXB is often challenging. It can be assessed based either on adynamical measurement of the mass of the compact object (viaoptical and near infrared spectroscopy during quiescence; seee.g., Casares et al. 1992), or on the detection of X-ray coherentpulsations and / or thermonuclear X-ray bursts, which would pro-vide a straightforward observational evidence for a neutron staraccretor. To-date, all confirmed UCXBs are known to contain aneutron star accretor, although there are a number of black holecandidates (see e.g., Bahramian et al. 2017). Assuming that ouridentification of J0840 as a UCXB is correct, it is then temptingto conclude that this system hosts a neutron star as well. How-ever, our non-detection of coherent pulsations or thermonuclearX-ray bursts leaves this as an open question.Coherent pulsations are thought to be formed in neutron starsystems where the magnetic field is large enough to truncate theaccretion disk and channel part of the disk material onto a smallregion on the neutron star surface, close to the magnetic poles.The radiation emitted from the heated impact region (hot spot)or from a slab of shocked plasma that forms above it (accretioncolumns) appears modulated at the neutron star spin period, pro-ducing a pulsed emission component. However, most neutronstar LMXBs are non-pulsating systems, and a systematic analy-sis by Patruno et al. (2018) suggested that weak pulsations mightnot form at all in (most) non-pulsating LMXBs. Moreover, threeaccreting neutron star LMXBs have shown coherent X-ray pul-sations only sporadically during their outbursts, with the mostextreme case being represented by Aql X-1, which showed pul-sations over a ∼
150 s interval out of a total observing time of ∼ ffi cient to conclusively rule out a neu-tron star accretor and might be still compatible with either a non-pulsating neutron star LMXB, or an intermittent accreting pulsar.Many LMXBs harboring a neutron star exhibit thermonu-clear (type-I) X-ray bursts due to unstable ignition of hydro-gen and / or helium freshly accreted onto the neutron star surface(see Galloway et al. 2020 for a recent compilation). However,these bursts are not expected to occur in UCXBs with a carbon-oxygen rich donor star due to the lack of hydrogen and heliumin the disk. In fact, very few (or even none) of these eventshave been detected from known UCXBs lacking signatures ofiron emission over a decade or so of monitoring campaigns (Ko-liopanos et al. 2021). Therefore, our non-detection of thermonu-clear bursts from J0840 is, again, not inconsistent with a neutronstar accretor in an UCXB with a carbon-oxygen donor star.The X-ray spectral properties can also be used to help es-tablish the nature of the compact object in a LMXB. Indeed, theX-ray spectra of transient neutron star systems are observed to besignificantly softer than those of black hole systems at luminosi-ties below 10 erg s − (0.5–10 keV; see Wijnands et al. 2015). Clearly, this diagnostic tool strongly relies on the distance to thesystem, which is unknown in the case of J0840. Nevertheless,we can derive some constraints on the distance by consideringthe model for the spiral structure of the Galaxy derived fromthe distribution map of the H ii regions within the Galaxy (Hou& Han 2014). Their distribution and the large absorption columndensity derived from our spectral analysis ( N H (cid:39) . × cm − ;Section 3.3), which is comparable to the total Galactic columnin the direction of J0840 ( N H (cid:39) . × cm − ; Willingaleet al. 2013) suggest that J0840 is located at a distance in therange 5–10 kpc and likely in the Perseus Arm. At the epochof the quasi-simultaneous Swift and
NuSTAR observations ofJ0840, the photon index was Γ ∼ . L X ∼ (1–4) × erg s − (0.5–10 keV; assuming the above-mentioned range for the distance). For such values, the posi-tion of the system on the photon index versus X-ray luminos-ity plane presented by Wijnands et al. (2015) would lie in a re-gion which may be compatible with the track followed by bothblack holes and neutron stars LMXBs, making the nature of theaccretor quite uncertain. Remarkably, the thermal fraction ofthe total X-ray luminosity J0840 in quiescence ( ∼ / or emission mechanisms associated with the magneto-sphere (see e.g. Jonker et al. 2004; Degenaar et al. 2013; Cam-pana et al. 2014; Wijnands et al. 2015). We stress, however, thatthe above value for the thermal fraction of J0840 was estimatedassuming no variation in the power-law slope along the outburst.Deep broad-band observations would be needed to characterizeadequately the quiescent X-ray spectrum and the spectral energydistribution from the optical to the X-rays.In their low-luminosity or quiescent hard states, blackhole and neutron star LMXBs exhibit a non-linear relationshipbetween their radio and X-ray luminosities (e.g., Hannikainenet al. 1998; Corbel et al. 2000; Gallo et al. 2003, 2012; Corbelet al. 2013; Tudor et al. 2017; Gusinskaia et al. 2020). Whilethe behaviour of single systems may di ff er somewhat (e.g.,Tudor et al. 2017; Russell et al. 2018), the analysis of the fullpopulation of black hole and neutron star systems has shownthat black hole LMXBs are typically a factor of ∼
22 moreradio bright than their neutron star counterparts (Gallo et al.2018). As such, a source’s radio and X-ray brightness hasoften been used to help discriminating between a neutron staror black hole primary. Placing our 3- σ radio upper-limit withthe closest in-time X-ray luminosity (1-day separation) on theradio – X-ray luminosity plane (Figure 7) shows that our datais consistent with both a black hole and a neutron star LMXBfor our assumed distances. Hence, the upper-limit on the radioemission during quiescence does not identify the nature of the The optical magnitude of J0840 in quiescence is fainter than the sen-sitivity limit of the
Gaia mission, and indeed the source is not listed inthe
Gaia
Early Data Release 3 (
Gaia
EDR3; Gaia Collaboration et al.2020) and no information on its geometric parallax is available. We note that the prescription by Wijnands et al. (2015) applies tospectra that are modelled with an absorbed PL model. However, thephoton index derived for J0840 using such model,
Γ = . ± .
05, isonly slightly larger than the value derived from the (statistically moreacceptable) BB + PL model. Therefore, our considerations are not af-fected significantly by the adopted model. Article number, page 9 of 11 & A proofs: manuscript no. J0840_ms Unabsorbed 1-10 keV X-ray luminosity (erg s − ) - G H z R a d i o l u m i n o s i t y ( e r g s − ) Black hole LMXBsNeutron star LMXBsAMXPsSwift J0840.7-3516 5kpcSwift J0840.7-3516 10kpc
Fig. 7.
The radio and X-ray luminosities of accreting black holes andneutron stars. The quasi-simultaneous radio and X-ray luminosities ofJ0840 during quiescence are shown using the red square and orangediamond for the assumed distances of 5 and 10 kpc, respectively. Weshow the population of black hole LMXBs (black circles), neutronstar LMXBs (blue squares), and accreting millisecond X-ray pulsars(AMXPs; green triangles). The gray dashed line indicates the best-fitcorrelation for the BH systems from Gallo et al. (2018). Our radio andX-ray observations of J0840 do not di ff erentiate between a black holeor neutron star accretor. Data taken from Bahramian et al. (2018). accretor. During quiescence, more sensitive radio observationsthat detect the source, or monitoring during a new outburstmay be able to discriminate between the two classes using thismethod.As a final remark, we note that for similar binary orbital pe-riods, neutron star systems are typically brighter in the X-raysthan black hole systems in quiescence. In this respect, the rela-tively large value for the quiescent X-ray luminosity of J0840, L X ∼ (1–4) × erg s − , is far more consistent with a neutronstar accretor than with a black hole accretor (assuming an orbitalperiod in the range of those measured for UCXBs; see e.g. Fig. 3by Armas Padilla et al. 2014, and references therein).
5. Conclusions
J0840 was discovered as it entered a powerful X-ray / UV / opticaloutburst on 2020 February 5. This transient episode was charac-terized by an increase in the X-ray flux of a factor of ≈ abovequiescence and lasted overall only ∼ ∼ ∼ .
3, respectively. The UV emission rapidly decayed byabout 3 mag in ∼ Swift , while the optical emission decreased by about 1 mag in ∼ . ≈ ffi cient to draw a firm conclusion on thenature of the accretor, we favour the scenario of a neutron staraccretor based on the multi-band properties observed both in out-burst and in quiescence.By definition, confirmation of J0840 as an UCXB can onlybe obtained through the measurement of the orbital period. Amethod that proved to be successful in relatively low-inclinationsystems has been the detection of a periodic optical modulationin photometric observations (e.g., Zhong & Wang 2011; Wanget al. 2015). This arises from X-ray heating of the donor star bythe irradiating X-ray source and to the variation of the visiblearea of the heated face as a function of the orbital phase. Alter-natively, an assessment of the orbital period can be achieved viatime-resolved optical spectroscopy through radial velocity mea-surements of the broad emission features that trace the orbitalmotion of the accreting material close to the accretor. Given theoptical faintness of J0840 in quiescence, dedicated observationswith 8-m class optical telescopes seem warranted to nail downthe nature of this system. In turn, the determination of the orbitalperiod would allow to restrict the parameter space for a moresensitive search for pulsed emission in existing X-ray data, pos-sibly giving tighter constraints on the nature of the accretor. Acknowledgements.
We thank F. Harrison and J. Stevens for scheduling Targetof Opportunity observations with
NuSTAR and ATCA in the Director’s Discre-tionary Time. This research is based on observations made with the Gran Tele-scopio Canarias (
GTC ), installed at the Spanish Observatorio del Roque de losMuchachos of the Instituto de Astrofísica de Canarias in the island of La Palma,under Director’s Discretionary Time (code GTC2020-142). We are indebted toAntonio Cabrera and the GTC sta ff for their e ff orts in performing the GTC ob-servations. We also thank Peter Jonker for comments on the manuscript. TheNuSTAR mission is a project led by the California Institute of Technology, man-aged by the Jet Propulsion Laboratory and funded by NASA. The Australia Tele-scope Compact Array is part of the Australia Telescope National Facility whichis funded by the Australian Government for operation as a National Facility man-aged by CSIRO. We acknowledge the Gomeroi people as the traditional ownersof the ATCA observatory site. This research made use of: the NuSTAR DataAnalysis Software (NuSTARDAS) jointly developed by the ASI Science DataCenter (ASDC) and the California Institute of Technology; software provided bythe High Energy Astrophysics Science Archive Research Center (HEASARC),which is a service of the Astrophysics Science Division at NASA / GSFC andthe High Energy Astrophysics Division of the Smithsonian Astrophysical Ob-servatory; APLpy (Robitaille & Bressert 2012), an open-source plotting packagefor Python hosted at http://aplpy.github.com ; and Astropy (Astropy Col-laboration et al. 2013, 2018), a community-developed core Python package forAstronomy. IRAF is distributed by the National Optical Astronomy Observatory,which is operated by the Association of Universities for Research in Astronomy(AURA) under a cooperative agreement with the National Science Foundation.We also made use of the following software: CASA (McMullin et al. 2007),HEASOFT v. 6.28, HENDRICS v. 5.1 (Bachetti 2018), XSPEC v. 12.11.1 (Ar-naud 1996). FCZ and AB are supported by Juan de la Cierva fellowships. FCZ,AB and NR are supported by the ERC Consolidator Grant “MAGNESIA” (nr.817661) and acknowledge funding from grants SGR2017-1383 and PGC2018-095512-BI00. TDR acknowledges financial contribution from the agreementASI-INAF n.2017-14-H.0. We thank support from the COST Action “PHAROS”(CA 16124).
References
Armas Padilla, M., Wijnands, R., Degenaar, N., et al. 2014, MNRAS, 444, 902Arnaud, K. A. 1996, in Astronomical Society of the Pacific Conference Series,Vol. 101, Astronomical Data Analysis Software and Systems V, ed. G. H.Jacoby & J. Barnes, 17Astropy Collaboration, Price-Whelan, A. M., Sip˝ocz, B. M., et al. 2018, aj, 156,123Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558,A33Auchettl, K., Guillochon, J., & Ramirez-Ruiz, E. 2017, ApJ, 838, 149
Article number, page 10 of 11. Coti Zelati et al.: Swift J0840.7 − Bachetti, M. 2018, HENDRICS: High ENergy Data Reduction Interface fromthe Command ShellBahramian, A., Heinke, C. O., Tudor, V., et al. 2017, MNRAS, 467, 2199Bahramian, A., Miller-Jones, J., Strader, J., et al. 2018, Radio / X-ray correlationdatabase for X-ray binariesBassa, C. G., Jonker, P. G., in’t Zand, J. J. M., & Verbunt, F. 2006, A&A, 446,L17Burrows, D. N., Hill, J. E., Nousek, J. A., et al. 2005, Space Sci. Rev., 120, 165Cackett, E. M., Miller, J. M., Ballantyne, D. R., et al. 2010, ApJ, 720, 205Campana, S., Brivio, F., Degenaar, N., et al. 2014, MNRAS, 441, 1984Casares, J., Charles, P. A., & Naylor, T. 1992, Nature, 355, 614Casella, P., Altamirano, D., Patruno, A., Wijnands, R., & van der Klis, M. 2008,ApJ, 674, L41Cepa, J., Aguiar, M., Escalera, V. G., et al. 2000, in Society of Photo-OpticalInstrumentation Engineers (SPIE) Conference Series, Vol. 4008, Proc. SPIE,ed. M. Iye & A. F. Moorwood, 623–631Chakrabarty, D. & Roche, P. 1997, ApJ, 489, 254Corbel, S., Coriat, M., Brocksopp, C., et al. 2013, MNRAS, 428, 2500Corbel, S., Fender, R. P., Tzioumis, A. K., et al. 2000, A&A, 359, 251de Martino, D., Bernardini, F., Mukai, K., Falanga, M., & Masetti, N. 2020,Advances in Space Research, 66, 1209Degenaar, N., Wijnands, R., & Miller, J. M. 2013, ApJ, 767, L31Di Salvo, T. & Sanna, A. 2020, arXiv e-prints, arXiv:2010.09005Done, C., Gierli´nski, M., & Kubota, A. 2007, A&A Rev., 15, 1Evans, P. A., Tohuvavohu, A., & Neil Gehrels Swift Observatory Team. 2020,GRB Coordinates Network, 26982, 1Fabian, A. C., Rees, M. J., Stella, L., & White, N. E. 1989, MNRAS, 238, 729Fitzpatrick, E. L., Massa, D., Gordon, K. D., Bohlin, R., & Clayton, G. C. 2019,ApJ, 886, 108Foight, D. R., Güver, T., Özel, F., & Slane, P. O. 2016, ApJ, 826, 66Frank, J., King, A., & Raine, D. J. 2002, Accretion Power in Astrophysics: ThirdEditionGaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2020, arXiv e-prints,arXiv:2012.01533Gallo, E., Degenaar, N., & van den Eijnden, J. 2018, MNRAS, 478, L132Gallo, E., Fender, R. P., & Pooley, G. G. 2003, MNRAS, 344, 60Gallo, E., Miller, B. P., & Fender, R. 2012, MNRAS, 423, 590Galloway, D. K., in’t Zand, J., Chenevez, J., et al. 2020, ApJS, 249, 32Gendreau, K. C., Arzoumanian, Z., & Okajima, T. 2012, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 8443,Space Telescopes and Instrumentation 2012: Ultraviolet to Gamma Ray, ed.T. Takahashi, S. S. Murray, & J.-W. A. den Herder, 844313Gusinskaia, N. V., Hessels, J. W. T., Degenaar, N., et al. 2020, MNRAS, 492,2858Hameury, J. M. & Lasota, J. P. 2016, A&A, 594, A87Hannikainen, D., Hunstead, R., & Campbell-Wilson, D. 1998, New A Rev., 42,601Harrison, F. A., Craig, W. W., Christensen, F. E., et al. 2013, ApJ, 770, 103Hou, L. G. & Han, J. L. 2014, A&A, 569, A125in’t Zand, J. J. M., Jonker, P. G., & Markwardt, C. B. 2007, A&A, 465, 953Israel, G. L. & Stella, L. 1996, ApJ, 468, 369Iwakiri, W., Gendreau, K., Enoto, T., et al. 2020, The Astronomer’s Telegram,13456, 1Jonker, P. G., Galloway, D. K., McClintock, J. E., et al. 2004, MNRAS, 354, 666Kara, E., Steiner, J. F., Fabian, A. C., et al. 2019, Nature, 565, 198Koliopanos, F., Péault, M., Vasilopoulos, G., & Webb, N. 2021, MNRAS, 501,548Lasota, J.-P. 2001, New A Rev., 45, 449Malesani, D. B., Izzo, L., Palmerio, J., et al. 2020, GRB Coordinates Network,26989, 1Masetti, N., Landi, R., Pretorius, M. L., et al. 2007, A&A, 470, 331Masetti, N., Orlandini, M., Palazzi, E., Amati, L., & Frontera, F. 2006, A&A,453, 295McMullin, J. P., Waters, B., Schiebel, D., Young, W., & Golap, K. 2007, in As-tronomical Society of the Pacific Conference Series, Vol. 376, AstronomicalData Analysis Software and Systems XVI, ed. R. A. Shaw, F. Hill, & D. J.Bell, 127Melandri, A., Covino, S., Fugazza, D., D’Avanzo, P. I., & REM Team. 2020,GRB Coordinates Network, 26986, 1Mukai, K. 2017, PASP, 129, 062001Nelemans, G. & Jonker, P. G. 2010, New A Rev., 54, 87Nelemans, G., Jonker, P. G., Marsh, T. R., & van der Klis, M. 2004, MNRAS,348, L7Nelemans, G., Jonker, P. G., & Steeghs, D. 2006, MNRAS, 370, 255Nelson, L. A., Rappaport, S. A., & Joss, P. C. 1986, ApJ, 304, 231Niwano, M. & MAXI Team. 2020, GRB Coordinates Network, 26987, 1Patruno, A. & Watts, A. L. 2021, Accreting Millisecond X-ray Pulsars, ed. T. M.Belloni, M. Méndez, & C. Zhang, Vol. 461, 143–208Patruno, A., Wette, K., & Messenger, C. 2018, ApJ, 859, 112Reig, P. 2011, Ap&SS, 332, 1 Robitaille, T. & Bressert, E. 2012, APLpy: Astronomical Plotting Library inPythonRodríguez Castillo, G. A., Israel, G. L., Belfiore, A., et al. 2020, ApJ, 895, 60Roming, P. W. A., Kennedy, T. E., Mason, K. O., et al. 2005, Space Sci. Rev.,120, 95Russell, T. D., Degenaar, N., Wijnands, R., et al. 2018, ApJ, 869, L16Savonije, G. J., de Kool, M., & van den Heuvel, E. P. J. 1986, A&A, 155, 51Sazonov, S., Paizis, A., Bazzano, A., et al. 2020, New A Rev., 88, 101536Shakura, N. I. & Sunyaev, R. A. 1973, A&A, 500, 33Smith, R. K., Brickhouse, N. S., Liedahl, D. A., & Raymond, J. C. 2001, ApJ,556, L91Stamatikos, M. & Swift-BAT Team. 2020, GRB Coordinates Network, 27010, 1Tetarenko, B. E., Sivako ff , G. R., Heinke, C. O., & Gladstone, J. C. 2016, ApJS,222, 15Tudor, V., Miller-Jones, J. C. A., Patruno, A., et al. 2017, MNRAS, 470, 324Wang, Z., Tziamtzis, A., Kaplan, D. L., & Chakrabarty, D. 2015, PASA, 32, e035Werner, K., Nagel, T., Rauch, T., Hammer, N. J., & Dreizler, S. 2006, A&A, 450,725Wijnands, R., Degenaar, N., Armas Padilla, M., et al. 2015, MNRAS, 454, 1371Willingale, R., Starling, R. L. C., Beardmore, A. P., Tanvir, N. R., & O’Brien,P. T. 2013, MNRAS, 431, 394Zampieri, L., Turolla, R., Zane, S., & Treves, A. 1995, ApJ, 439, 849Zhong, J. & Wang, Z. 2011, ApJ, 729, 8 List of Objects ‘GX 1 +
4’ on page 7‘4U 0614 + − −
67’ on page 867’ on page 8