Spectral Properties of the Soft X-Ray Transient MAXI J0637−430 Using AstroSat

Soft X-ray transients are a subclass of low-mass X-ray binaries (LMXBs) that appear as extremely faint sources (L = 10³⁰–10³³ erg s⁻¹) during most of their lifetime. They are characterized by a nonsteady transfer of mass onto the compact object and they occasionally undergo sporadic outbursts, which occur at intervals of 1–60 yr (Chen et al. 1997; Tetarenko et al. 2016). This causes their X-ray luminosity to increase by a factor of up to 10⁷ (Paradijs & McClintock 1995), and it then decays back to quiescence with an e-folding timescale of ∼30 days (Chen et al. 1997). During the short outbursts, they emit enough X-rays to make them the brightest X-ray sources (L = 10³⁷–10³⁸ erg s⁻¹) in the sky. The occurrence of these outbursts is attributed to instabilities in the accretion disk that are both viscous and thermal in nature (Meyer & Meyer-Hofmeister 1981; Cannizzo et al. 1995; King & Ritter 1998; Lasota 2001). Soft X-ray transients, especially the ones harbouring a black hole, usually go undetected, and are discovered only when they undergo an outburst.

MAXI J0637−430 is one such source, and was first detected by the MAXI/GSC nova alert system during seven scan transits on 2019 November 2–3 in the 2–4 and 4–10 keV bands. The scans revealed the source to be located at R.A. (J2000) = 06^h38^m54^s, decl. (J2000) = −42^h45^m57^s in the soft band (2–4 keV) and at R.A. (J2000) = 06^h37^m43^s, decl. (J2000) = −43^h03^m15^s in the hard band (4–10 keV) at a 90% confidence level (Negoro et al. 2019). Since its first detection, MAXI J0637−430 underwent a considerable increase in its flux from 59 ± 6 mCrab to ∼200 mCrab in the 2–4 keV band and from 32 ± 6 mCrab to ∼50 mCrab in the 4–10 keV band (Negoro et al. 2019). A Target of Opportunity (ToO) observation performed on 2019 November 3 by the Swift mission detected MAXI J0637−430 at R.A. (J2000) = 06^h36^m2359, decl. (J2000) = −42^h52^m041, which was consistent with MAXI's hard band localization. Its X-ray spectrum modeled using an absorbed disk blackbody + power law with inner disk temperature (kT_in) of 0.9 ± 0.1 keV and power-law index (Γ) of 2.3 ± 0.8 indicated that the source underwent an outburst and transitioned from hard to soft spectral state. An optical counterpart with a brightness of u = 14.87 ± 0.02 (Vega) was detected by the UVOT instrument on board Swift at R.A. (J2000) = 06^h36^m2323 decl. (J2000) = −42^h52^m0425. Since there are no known stars at this position, it was deemed that this optical source underwent a significant brightening, as is common for the optical counterpart of black hole low-mass X-ray binaries (BH-LMXBs) during outbursts (Kennea et al. 2019). Follow-up observation by the NuStar mission found the flux of the source to be ∼95 mCrab. Preliminary analysis of the spectrum of the source in the energy range 3–79 keV with a thermal disk blackbody component, power law, and a reflection component yielded kT_in of 0.628 ± 0.004 keV and Γ of 2.40 ± 0.04 (90% confidence errors) (Tomsick et al. 2019). MAXI J0637−430 was also observed in the radio band with the ATCA with flux densities of 66 ± 15 μJy at 5.5 GHz and 60 ± 10 μJy at 9 GHz (Russell et al. 2019). However, as the nature of the radio jet emission could not be deciphered, the source could not be properly classified using the radio/X-ray correlation in X-ray binaries. The source was also observed in the infrared band with the simultaneous imaging camera SIRIUS attached to the 1.4 m telescope Infrared Survey Facility, where the estimated magnitudes in the J, H, and K bands were 17.40 ± 0.01, 17.69 ± 0.02, and 17.96 ± 0.05, respectively (Murata et al. 2019). Spectral analysis of the X-ray data from Swift observations with an absorbed disk blackbody model showed the source to have kT_in of 0.675 ± 0.003 keV (Knigge et al. 2019), which is consistent with the value obtained from NuStar observations. Since its discovery, MAXI J0637−430 was observed by the NICER mission continuously with a cadence of 1–2 days, and was observed to undergo a spectral state transition. It was reported that ~23 days after its discovery the source transitioned into the hard state (Remillard et al. 2020). Subsequent observations by Swift showed that MAXI J0637−430 could possibly be approaching its quiescence level (Tomsick & Lazar 2020). A multiwavelength study of the source was carried out by Tetarenko et al. (2021) using data from Swift-XRT and UVOT, Gemini/GMOS, ATCA, and AAVSO. This study made use of an irradiated accretion disk model (diskir) (Gierliński et al. 2009) to derive the time-series evolution of its spectral parameters over the entire outburst cycle. Analysis of NICER data by Jana et al. (2021) revealed the source to comprise an ultrasoft thermal component (kT_in ≲ 0.6 keV) and a power-law tail. The study also showed that its spectra do not need a thermal component corresponding to the emission from the neutron star surface, thus suggesting that the compact object in MAXI J0637−430 is most likely a black hole. The mass of the black hole inferred from this study is 5–12 M_⊙ for a source distance of d < 10 kpc, and the distance to the source is found to have a lower limit of 6.5 kpc. Baby et al. (2021) found that the 0.5–25 keV spectra of the source could be modeled with a multicolor disk emission (diskbb) convolved with a thermal Comptonisation component (thcomp). Spectral fitting with the kerbb model in conjunction with the soft–hard transition luminosity favors a black hole with a mass of 3–19 M_⊙ and retrograde spin at a distance <15 kpc. A broadband spectral study on NuSTAR data from the source showed that a two-component model, comprising a combination of a multicolor disk blackbody and a thermal Comptonization component, is adequate to fit the spectra only up to 10 keV. When higher energies are considered, scenarios involving a plunging region and reprocessing of returning disk radiation are equally possible (Lazar et al. 2021).

Encouraged by the observational campaigns carried out by various satellite- and ground-based telescopes, ToO observations of MAXI J06347−430 were performed using the Soft X-ray Telescope (SXT) and Large Area X-ray Proportional Counter (LAXPC) instruments on board AstroSat in the 0.3–80 keV energy range on 2019 November 8 and 21. Here, we report the results of spectral and temporal studies carried out on MAXI J0637−430 data from the SXT and LAXPC instruments. The details of AstroSat observations and the data reduction procedures are described in Section 2. In Section 3, the light curve and hardness–intensity diagram (HID) are presented. In Section 4, we present the results of spectral analysis. The findings and a summary of the results are discussed in Section 5.

ToO observations of MAXI J06347−430 (Thomas et al. 2019) in the 0.3–80 keV energy range were carried out using SXT and LAXPC on board AstroSat for a total of ∼60 ks on 2019 November 8 (hereafter, Observation 1) and November 15 and 21 (hereafter, Observation 2). We did not include the November 15 data in our study because that contains nine-pointing safety observations for the UVIT instrument on board AstroSat, each with different pointing and offset. Due to this, the spectra from the individual pointings could not be combined because the effective area of the instrument changes with the offset. Moreover, since the LAXPC pointings were also different, flux measurement using SXT+LAXPC data could not be made. AstroSat observations used for our study, marked on the 2–20 keV MAXI light curve in Figure 1, show that Observation 1 was carried out shortly after the outburst peak, whereas Observation 2 was performed in the middle of the outburst decay. The photon counting (PC) mode was employed for observation with SXT, whereas for LAXPC, the observation was carried out in the event analysis (EA) mode . A log of observations used for this study is given in Table 1.

Zoom In Zoom Out Reset image size

SXT is a focusing telescope equipped with a charge-coupled device (CCD) camera that performs X-ray imaging in the 0.3–8.0 keV energy range with a spectral resolution of ∼150 eV at 6 keV (Singh et al. 2016). SXT data from MAXI J06347−430 were processed using the standard SXT pipeline—AS1SXTLevel2-1.4b. ³ This yielded Level 2 event files for individual orbits of the observation, which were then merged into one master event file using the SXT Event Merger Tool. ⁴ The merged event file was then used to extract source images with the help of XSELECT V2.4k. The source was selected as the region between 8' and 5' (inner radius) for Observations 1 and 2, respectively, and 15' (outer radius) to reduce the effect of pile-up of the CCD. The response matrix ⁵ and background ⁶ files provided by the SXT Payload Operations Centre (POC) were used for the analysis. An off-axis auxiliary response file (ARF) was created with the sxt_ARFModule (see footnote 4). SXT data in the range 0.5–5.0 keV for Regions 1 and 3, and 0.5–4.8 keV for Region 2 were used (Figure 4) because the data quality above and below these energy ranges was poor.

LAXPC is a cluster of three co-aligned proportional counters (LAXPC-10, LAXPC-20, LAXPC-30) that operates in the 3–80 keV energy range with an absolute temporal resolution of 10 μs (Yadav et al. 2016; Agrawal 2017; Antia et al. 2017). Data from LAXPC were processed with the LAXPCSOFT (Format A) ⁷ to obtain event files, good time interval (GTI) files, light curves, source and background energy spectra, response matrix files, and power density spectra. LAXPC-20 data alone were used for our study because it was reported by the POC that LAXPC-10 underwent an abnormal change in its gain on 2018 March 28 and LAXPC-30 was not operational during this time. Moreover, as the energy spectrum above 20 keV was background-dominated, spectral studies using LAXPC were restricted to the energy range 4–20 keV.

Net light curves of the source were obtained in the 0.7–7.0 keV range from the SXT instrument; and 4.0–5.0 keV and 5.0–30.0 keV ranges from the LAXPC instrument. These light curves were binned to ∼50 s. It is seen that during the beginning of the outburst decay, i.e., in Observation 1, the source intensity remains fairly constant at ∼30 counts s⁻¹ and ∼28 counts s⁻¹ in the 4.0–5.0 keV (panel 2 in Figure 2) and 5.0–30 keV (panel 3 in Figure 2) ranges respectively. This then changes as the count rates in both energy ranges increase by a small factor toward the end of the observation. This jump in the count rate is reflected in the hardness–time diagram too (panel 4 in Figure 2). In comparison, the LAXPC net flux along with the hardness ratio is seen to decrease monotonically through the latter part of the outburst decay, i.e., in Observation 2 (panels 2, 3, and 4 in Figure 3). Using LAXPC-20 data of both the observations, a combined, 50 s binned HID was generated with hardness defined as the ratio of counts in the 5.0–30.0 keV range to counts in the 4.0–5.0 keV range and intensity defined as the sum of counts in the 4.0–30 keV range. From the pattern traced by the HID, it is not possible to determine whether the source showed characteristics of the q-diagram exhibited by BH-LMXBs (Remillard & McClintock 2006) or the Z or Atoll pattern exhibited by neutron star LMXBs (Hasinger & van der Klis 1989). However, it is seen that the HID showed variability in hardness from ∼0.9 to ∼1.9 (Figure 4). Further, in order to investigate the hardness–intensity relation in the energy range <4 keV, data from the SXT instrument, corresponding to the three regions in the LAXPC-HID were used to generate an HID (Figure 5). The hardness of this SXT-HID was defined as the ratio of counts in the 1.0–7.0 keV range to counts in the 0.3–1.0 keV range and the intensity was defined as the sum of counts in the 0.3–7.0 keV range, for time corresponding to all three regions from the LAXPC-HID. Regions 1 and 3 in this SXT-HID show variations in average flux and hardness, whereas Region 2 has too few data points to draw a definitive conclusion.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The simultaneous broadband X-ray spectral coverage of AstroSat with the SXT and LAXPC instruments was used to perform flux-resolved spectroscopy. This was done by dividing the LAXPC-HID into three regions: Regions 1, 2, and 3. The divisions were made such that each region denotes an isolated cluster of points in the LAXPC-HID (Figure 4). Simultaneous GTIs for both the SXT and LAXPC instruments were generated, using which simultaneous spectra for both instruments were generated for all three regions of the LAXPC-HID. The spectra were then fit with the spectral modeling tool XSPECversion: 12.10.1o (Arnaud 1996) in the energy ranges 0.5–5.0 keV from the SXT instrument for Regions 1 and 3, and 0.5–4.8 keV for Region 2; whereas the energy range 4.0–20 keV from the LAXPC instrument was chosen for all three regions. These energy ranges were chosen because the spectra below 0.5 keV and above 4.8 keV (from the SXT instrument) showed very high residuals and the spectra above 20.0 keV (from the LAXPC) was dominated by the background. A multicolor blackbody model, diskbb (Mitsuda et al. 1984), was used along with a convolution Comptonization model, simpl (Steiner et al. 2009), in order to explain the emission from the accretion disk and thermally Comptonized corona, respectively. The energy ranges for the spectral fits were extended using energies 0.01 100 500 log to supply an energy-binning array. To account for absorption in the interstellar medium, we used the Tuebingen–Boulder interstellar medium absorption model tbabs, with the solar abundance table given by Wilms et al. (2000). Further, a multiplicative constant factor was included to address uncertainties caused by cross-calibration of the SXT and LAXPC instruments. As prescribed by the POC, a systematic error of 3% was added to all spectral fits (Bhattacharya 2017). In addition to this, a gain fit was performed for the SXT data to account for the nonlinear change in the detector gain. The slope of the gain was frozen to 1, leaving the offset to vary. There were positive residuals around 6.4 keV indicating the possible presence of a disk reflection feature. A Gaussian component with its line energy frozen at 6.4 keV was later added to account for this. The addition of the Gaussian component yielded a small change in Δχ² from 1.03 and 0.92 to 1.0 and 0.76 in Regions 1 and 2 respectively, whereas for Region 3 it remained constant. The model combination constant * tbabs (simpl * diskbb + Gaussian) yielded good fits for all the three regions, with Δχ² of ∼0.9 (Figure 6). The best-fit spectral parameters of this fit are given in Table 2 and the corresponding spectra are given in Figure 6. The norm of diskbb remains fairly constant in all three regions. Hence, in order to understand the observed variation in the HID, the spectral fit was repeated with the norm of diskbb fixed at 2075. This value was obtained by fitting a constant through the norms of all three regions. Further, the unabsorbed total and disk flux were calculated in the 0.5–20 keV range using the cflux model. The best-fit parameters of this spectral fit are given in Table 3.

Zoom In Zoom Out Reset image size

Table 2. Best-fit Model Spectral Parameters

Model	Parameter	Region 1	Region 2	Region 3
tbabs	NH (1020 cm−2)	3.300 ± 0.006	2.120 ${}_{-0.021}^{+0.045}$	1.250 ± 0.005
simpl	Γ	2.00${}_{-0.16}^{+0.14}$	1.95${}_{-0.20}^{+0.17}$	2.46 ± 0.03
	FracSctr	0.016${}_{-0.003}^{+0.004}$	0.031${}_{-0.009}^{+0.010}$	0.100${}_{-0.008}^{+0.009}$
diskbb	kTin (keV)	0.610 ±0.005	0.650 ± 0.030	0.520 ± 0.007
	Norm	2103${}_{-152}^{+165}$	1455${}_{-465}^{+714}$	2083${}_{-98}^{+104}$
Gaussian	Line (keV)	6.4 (fixed)	6.4 (fixed)	6.4 (fixed)
	Width (keV)	1.06${}_{-0.40}^{+0.35}$	1.49${}_{-0.47}^{+0.42}$	0.20${}_{-0.37}^{+0.56}$
	Norm (10−3)	1.39${}_{-0.50}^{+0.60}$	3.24${}_{-1.40}^{+1.60}$	0.34 ± 0.30
Reduced χ2 /dof		1.03/363	0.76/109	1.33/460

Download table as:  ASCIITypeset image

Table 3. Best-fit Model Spectral Parameters with Fixed diskbb Norm

Model	Parameter	Region 1	Region 2	Region 3
tbabs	NH (1020 cm−2)	3.31 ± 0.006	${3.87}_{-0.03}^{+0.04}$	1.12 ± 0.005
simpl	Γ	${2.00}_{-0.16}^{+0.14}$	${2.00}_{-0.17}^{+0.15}$	2.49 ± 0.03
	FracSctr	${0.016}_{-0.003}^{+0.004}$	0.028 ± 0.007	0.100 ± 0.008
diskbb	kTin (keV)	0.610 ± 0.003	0.620 ± 0.005	0.520 ± 0.005
	Norm	2075 (fixed)	2075 (fixed)	2075 (fixed)
Gaussian	Line (keV)	6.4 (fixed)	6.4 (fixed)	6.4 (fixed)
	Width (keV)	${1.06}_{-0.40}^{+0.35}$	${1.32}_{-0.42}^{+0.34}$	${0.37}_{-0.36}^{+0.56}$
	Norm (10−3)	${1.39}_{-0.50}^{+0.60}$	${3.14}_{-1.30}^{+1.50}$	${0.40}_{-0.20}^{+0.30}$
Unabsorbed disk flux	(10−9 erg cm−2 s−1)	${5.31}_{-0.95}^{+1.05}$	${4.77}_{-0.84}^{+1.20}$	${2.58}_{-0.95}^{+1.05}$
Unabsorbed total flux	(10−9 erg cm−2 s−1)	${5.53}_{-0.95}^{+1.05}$	${5.15}_{-0.85}^{+1.18}$	${3.07}_{-0.96}^{+1.04}$
Unabsorbed disk flux/total flux		0.96	0.92	0.84
L/LEdd		0.025	0.023	0.141
Reduced χ2/dof		1.02/364	0.77/110	1.34/461

Download table as:  ASCIITypeset image

In addition to this, we also carried out temporal analysis with data from the LAXPC instrument in the 4–30 keV range. However, it did not yield substantial results.

In this work, we analyzed the SXT and LAXPC data from AstroSat observations (2019 November 8 and 21) of MAXI J0637−430 in the 0.5–20 keV energy range. The analysis revealed three distinct clusters in the LAXPC-HID of the source (Figure 4). However, these clusters do not form a clear pattern to give insights regarding the exact nature of the source. Monitoring the source through its entire outburst cycle with NICER and MAXI has shown its HID to exhibit signatures of the various states of a BH-LMXB in outburst (Baby et al. 2021; Jana et al. 2021). Flux-resolved spectral analysis showed that the spectra can be characterized by a multicolor blackbody component arising from the accretion disk along with a thermal Comptonization component. Our choice of model, diskbb, to characterize the soft component is in agreement with erstwhile studies carried out on the source. However, different models have been used to characterize its hard component: powerlaw (Tetarenko et al. 2021), nthcomp (Jana et al. 2021; Lazar et al. 2021), and thcomp (Baby et al. 2021). The accretion disk temperatures, 0.61, 0.65, and 0.52 keV in Regions 1, 2, and 3 respectively, point to a cool disk. This is in agreement with the studies carried out by Tetarenko et al. (2021), Jana et al. (2021), and Baby et al. (2021), where the disk temperature is seen to decay from ∼0.6 to ∼0.1 keV during the course of the outburst. The slightly lower value of disk temperature in Region 3 correlates with the region being harder in the HID (Figure 4). The diskbb norms of 2103, 1455, and 2083 in Regions 1, 2, and 3 indicate constant accretion disk radius throughout both the observations, which is consistent with that exhibited by many BH-LMXBs in the soft state (Done et al. 2007). Similar results have been found by Baby et al. (2021), who used the convolution model thcomp along with diskbb to characterize the AstroSat spectra of the source. Assuming a source distance of 10 kpc, inclination angle of 70°, and a color hardening factor of 1.7 (Shimura & Takahara 1995), we calculated the inner disk radius to be ∼98, ∼81, and ∼97 km in Regions 1, 2, and 3 respectively. For a black hole of 20 M_⊙, keeping the disk normalization constant at 2075, we estimated the inner disk radius to be ∼6 R_g. This is the distance at which the innermost stable circular orbit is located for a nonrotating black hole. The increase in scatter fraction (FracSctr) (obtained using simpl model) from 0.016${}_{-0.003}^{+0.004}$ in Region 1 to 0.031${}_{-0.009}^{+0.01}$ in Region 2 reflects the increased count rate in the LAXPC-HID. In addition, the presence of a Gaussian line at 6.4 keV points to a reflection feature from the accretion disk. The width of this line (∼1 keV in Regions 1 and 2) shows that it is broadened due to relativistic effects in the vicinity of the central compact object. It is to be noted that the width and the norm of the Gaussian component are significantly smaller in Region 3. The unabsorbed total flux and disk flux of the source in the 0.5–20 keV range imply Eddington fractions of 0.025, 0.023, and 0.0141 for Regions 1, 2, and 3 respectively (Table 3). The ratio of unabsorbed disk flux to the total flux (∼0.9) in all three regions (Table 3) suggests that the total flux is dominated by emission from the accretion disk. This consistent disk-dominated flux combined with the photon power-law index of ⪆2, across all three regions of the HID, shows the source to be in the soft state.

We carried out flux-resolved spectral studies on two observations (∼60 ks) of the soft X-ray transient source MAXI J0637−430 in the 0.5–20 keV energy range using the SXT and LAXPC instruments on board AstroSat. Spectral analysis shows the source to have a cool accretion disk having a temperature of ∼0.55 keV with a reflection feature at 6.4 keV. The value of photon index of ⪆2 and the ratio of unabsorbed disk flux to the total disk of ∼0.9 indicate that MAXI J0637−430 was in the soft spectral state when observed by AstroSat. Also, it is seen that the value of the disk normalization is consistent with being constant (Table 2) and points to an accretion disk with an inner disk radius of 11.1 R_g. This is observed in several BH-LMXBs in the soft state. We conclude from our study that MAXI J0637−430 is a strong candidate for a black hole X-ray binary. Further observations and in-depth studies of the source during its future outbursts are essential to confirm its nature and unravel other physical parameters.

We thank the SXT POC at TIFR, Mumbai and the LAXPC POC team for their support, timely release of data and providing the necessary software tools. This work has made use of software provided by HEASARC. The authors acknowledge the financial support of ISRO under AstroSat Archival Data Utilization Program (No. DS-2B-13013(2)/9/2019-Sec.2 dated 2019 April 29). This publication uses data from the AstroSat mission of the Indian Space Research Organisation (ISRO), archived at the Indian Space Science Data Centre (ISSDC). One of the authors (S.B.G.) thanks the Inter-University Centre for Astronomy and Astrophysics (IUCAA), Pune for the Visiting Associateship. The authors thank the anonymous referee for the valuable suggestions/comments, which greatly helped in improving this manuscript.