Discovery of annular X-ray emission centered on MAXI J1421$-$613: Dust-scattering X-rays?

X-rays from an X-ray source are scattered by interstellar dust in the line of sight (Overbeck, 1965). A dust-scattering halo is observed surrounding a bright X-ray source which is located behind a large amount of dusts. The dust-scattering halos provide information on the interstellar dust, such as the size distribution and composition of the grains (e.g., Predehl & Schmitt (1995), Draine (2003)). By measuring the temporal variation of a halo, Predehl et al. (2000) constrained a line of sight position of Cygnus X-3.

If the X-ray emission with a short duration time is scattered by dust clouds, ring echoes are detected. Such samples have been detected around several gamma-ray bursts (e.g., Vaughan et al. (2004, 2006); Tiengo & Mereghetti (2006); Vianello et al. (2007); Pintore et al. (2017)) and Galactic sources (e.g., Svirski et al. (2011); Mao et al. (2014)). Only three Galactic samples have clear, well-defined, and large rings with a several arcminute scale (e.g., 1E 1547.0$-$5408, Circinus X-1, and V404 Cygni; Tiengo et al. (2010); Heinz et al. (2015); Vasilopoulos & Petropulou (2016); Heinz et al. (2016)). Most samples are accompanied with multiple rings, and each ring corresponds to a different dust layer or a different outburst. The thickness of the observed rings are different according to the situation; they reflect the duration time of the burst, the thickness of the dust layer, and the spatial resolution of the observation instrument.

An outburst of the new X-ray source MAXI J1421$-$613 was detected by the MAXI Nova Alert System (Negoro et al., 2010) on 2014 January 9 (Morooka et al., 2014). During the outburst, MAXI J1421$-$613 showed at least three type I X-ray bursts, and thus it is categorized into a neutron-star low-mass X-ray binary (Bozzo et al., 2014; Serino et al., 2015). The first burst with the observed flux of $1.7\times 10^{-9}$ erg cm^-2 s^-1 in the 3–10 keV band was detected on January 10 and lasted about 20 s (Bozzo et al., 2014). The second one occurred on January 16 is the brightest among the three. The bolometric flux was calculated to be $7\times 10^{-8}$ erg cm^-2 s^-1 and the duration time was about 36 s (Serino et al., 2015). The third one observed on January 18 has the bolometric flux of $3.2\times 10^{-8}$ erg cm^-2 s^-1 and the duration time of about 40 s (Serino et al., 2015).

After the MAXI alert, the Swift X-Ray Telescope (XRT) began a target-of-opportunity observation on the same day and observed the source approximately every 2 days until February 3. Suzaku performed a follow-up observation since January 31, and the observation continued for three days. But it did not detect the source and thus the 3$\sigma$ upper limit of the source flux of $1.2\times 10^{-13}$ erg cm^-2 s^-1 (0.5–10 keV) was obtained (Serino et al., 2015). Chandra also observed MAXI J1421$-$613 on February 8 with the effective time of 969 s. No source was detected significantly, and the 95%-confidence upper limit on the flux was measured to be $8.1\times 10^{-14}$  erg cm^-2 s^-1 (0.2–10 keV; Chakrabarty et al. (2014)).

Serino et al. (2015) analyzed the spectra during the outburst by utilizing the MAXI Gas Slit Camera (GSC) and the Swift XRT follow-up observations. The authors revealed that the spectra during the outburst can be explained by thermal Comptonization of multi-color disk blackbody emission. The photon index $\Gamma\sim 2$ is a typical for low-mass X-ray binaries and remains almost constant during the outburst. Assuming the empirical maximum luminosity of X-ray bursts, Serino et al. (2015) estimated the maximum distance to be 7 kpc from the observed peak flux.

In this paper, we report on the discovery of an annular X-ray emission around MAXI J1421$-$613 from the Suzaku data. Based on spectral and radial profile analyses of the annular emission, we discuss its possible origin: (1) a supernova remnant or (2) a dust-scattering echo.

We utilized the X-ray Imaging Spectrometer (XIS) data (Koyama et al., 2007) aboard Suzaku (Mitsuda et al., 2007). The XIS consists of four X-ray CCD cameras, each placed on the focal plane of the X-Ray Telescope (XRT; Serlemitsos et al. (2007)). A field of view (FOV) of the XIS is $\timeform{17.8^{\prime}}\times\timeform{17.8^{\prime}}$. Three of the sensors (XIS0, 2, and 3) employ front-illuminated (FI) CCDs, while the other (XIS1) has a back-illuminated (BI) CCD. The entire region of XIS2 and one-fourth of XIS0 have been out of function since 2006 November and 2009 June, respectively.

We analyzed data with the analysis software package HEAsoft 6.22.1 and the Suzaku calibration database (CALDB) released in 2016 February. The spectral analysis was performed with XSPEC 12.9.1. The data were screened by the standard event selection criteria for the XIS data processing. The response file (arf) and redistribution file (rmf) were produced by xissimarfgen and xisrmfgen (Ishisaki et al., 2007), respectively. The non-X-ray background (NXB) was estimated by xisnxbgen (Tawa et al., 2008) and was subtracted from spectra and images in the following analysis. For the spectral analysis, we used nearby blank-sky data as background. The observation log is summarized in table 2. Throughout the paper, FI and BI spectra were fitted simultaneously, but only the FI spectra (XIS0$+$3) are displayed in the figures for brevity. Error bars given in figures show the 1$\sigma$ statistical errors.

The X-ray images of the entire FOV of Suzaku XIS in the 0.5–2 keV, 2–5 keV and 5–8 keV bands are given in figure 1, where the position of MAXI J1421$-$613 is shown by the white cross. No significant X-ray emission was found at this position. Another point source was found (black cross in figure 1) at the position of $(\alpha,\delta)_{\rm J2000.0}=(\timeform{14h20m34s.07},\timeform{-61D41^{\prime}00^{\prime\prime}.33})$ with the positional uncertainty of $\timeform{19"}$ (Uchiyama et al., 2008). No catalogued source has been reported so far at this position, hence is newly named as Suzaku J1420.5$-$6141. The spectrum of Suzaku J1420.5$-$6141 was made from a circle with a radius of $\timeform{3^{\prime}}$. The best-fit power-law model gave absorption column density, photon index, and observed flux in the 2–10 keV band to be $N_{\rm H}=4.0^{+0.9}_{-0.8}\times 10^{22}$ cm^-2, $\Gamma=1.6\pm 0.3$, and $F_{\rm{2-10~{}keV}}=(1.4\pm 0.1)\times 10^{-12}$ ergs cm^-2 s^-1, respectively. No time variability within $<10$% was found.

We found clear an annular emission around MAXI J1412$-$613. The width is about $\timeform{6^{\prime}}$ and the flux is reasonably high with uniform distribution except Suzaku J1420.5$-$6141. The spectrum was made from the annular region excluding Suzaku J1420.5$-$6141. The background spectrum was obtained from the nearby position (table 2). These spectra are shown in figure 2. The background subtracted spectrum was fitted with either a phenomenological power-law model or an optically thin thermal plasma model. The best-fit results are given in table 3. The hydrogen column density obtained by the power-law model is $N_{\rm H}=(3.7\pm 0.5)\times 10^{22}$ cm^-2, which is consistent with that of MAXI J1412$-$613 measured by Serino et al. (2015), $N_{\rm H}=4.8^{+1.3}_{-1.1}\times 10^{22}$ cm^-2.

The region of the annular emission is shown on the flux map of CO (Dame et al., 2001) in figure 3. Referring the $N_{\rm H}$ distribution, we divided the region into four quadrants (quadrant 1–4). The spectrum of each quadrant was fitted with a phenomenological power-law model. The photon indices of the four parts are consistent with the best-fit value of the whole annular emission ($\Gamma=4.2$) within the 90% confidence levels. We fixed the photon index to $\Gamma=4.2$ and made other parameters free. The best-fit $N_{\rm H}$ values are $4.5\pm 0.4$, $3.7\pm 0.4$, $3.1\pm 0.3$, and $3.2\pm 0.4$ in the unit of $10^{22}$ cm^-2 for the quadrant 1, 2, 3, and 4, respectively. The largest $N_{\rm H}$ at the quadrant 1 corresponds to the CO peak, while the smallest $N_{\rm H}$ at quadrant 3 is out of the CO peak (Dame et al., 2001).

In order to give a constraint on the cloud position, adopting the Galactic rotation curve by McClure-Griffiths & Dickey (2007) of $R_{\rm o}=8.5$ kpc and $V_{\rm o}=220$ km s^-1, we plotted the correlation map of the LSR velocity and the distance from the sun (see figure 3). The LSR of $\sim 43$ km s^-1 gives two solutions for the distance of the cloud of $\sim 2.6$ kpc and $\sim 9$ kpc.

We made four annular spectra of \timeform3.’5–\timeform5’, \timeform5’–\timeform6’, \timeform6’–\timeform7’, and \timeform7’–\timeform8’ and investigated spectral variation with a radius. We fitted the spectra with the phenomenological power-law model with fixing the $N_{\rm H}$ value of 3.7$\times$10²² cm^-2. No clear differences in the spectral shape were found.

We also extracted spectra of the annular emission from the three periods and examine the spectral change with time. The periods are named as period 1, 2 and 3 (see table 3). The spectrum was fitted with the phenomenological power-law model with fixing the $N_{\rm H}$ value of 3.7$\times$10²² cm^-2. The best-fit spectral parameters are listed in table 3.

In order to examine the expansion of the annular emission, we made a radial profile from three different periods of the observation and fit the radial profiles with a gaussian. The gaussian centroid of period 1, 2, and 3 was measured to be $\timeform{5.^{\prime}4\pm 0.^{\prime}2}$, $\timeform{5.^{\prime}9}\pm{0.^{\prime}2}$, and $\timeform{5.^{\prime}9\pm 0.^{\prime}3}$, respectively.

We also checked whether the annular emission is detected by other satellites. Swift performed two observations of MAXI J1421$-$613 with the Photon Counting (PC) mode in almost the same period as the Suzaku observations; one observation started on January 30 (Obs ID $=$ 00033098009) and the other one started on February 1 (Obs ID $=$ 00033098010). The total exposure time of the two observations is 2.4 ks. We extracted a radial profile centered on the position of MAXI J1421$-$613 in the 1–5 keV band using the Swift data. The number of counts in each radial bin are divided by the corresponding area. We also divided the profile by the effective exposure time derived from the exposure map in order to take in to account telescope vignetting, the CCD bad pixels and columns, and the attitude variations. We found that the profile has a hint of a peak at $\sim\timeform{6^{\prime}}$. The count rate of the annular emission is measured to be $0.03$–$0.05$ counts s^-1, which approximately corresponds to the flux of $(2$–$3)\times 10^{-12}$ erg cm^-2 s^-1. These values are consistent with the radius and the flux of the annular emission observed by Suzaku. However, the swift data could not constrain the spectral parameters due to insufficient statistics. We also checked the Chandra/ACIS-S data, but we cannot distinguish the annular emission due to a short exposure time (1 ks) and the low net counts of 68 within the whole FOV (the 2–5 keV band).

We note that the preceding analysis with Suzaku by Serino et al. (2015) failed to find the annular emission, and hence the present paper reports the new discovery of the annular emission. Our detection of the annular emission with no bright point source (figure 1) catches very rare chance of the dust scattering echo; MAXI J1421$-$613 showed an outburst phase (10^-9–10^-10 erg cm^-2 s^-1) within a short duration of $<$ 10 days whereas the scattered X-rays are observed after MAXI J1421$-$613 re-entered to a long quiescent phase ($<10^{-13}$ erg cm^-2 s^-1). The steeper $\Gamma$ of the annular emission of $\sim 4.2$ than that of MAXI J1421$-$613 of $\sim 2.1$, and the time history of the radius and the flux is all consistent with a dust scattering echo. If MAXI J1421$-$613 shows an outburst again, one would be able to observe a dust scattering echo when they observe the object more than ten days after the outburst.

We thank all the members of the Suzaku and MAXI teams. We are grateful to Dr. Katsuji Koyama and Dr. Yasuharu Sugawara for their valuable comments. K.K.N. is supported by Research Fellowships of JSPS for Young Scientists. This work was supported by JSPS and MEXT KAKENHI Grant Numbers JP16J00548 (KKN) and JP17K14289 (MN).