A newly discovered double-double candidate microquasar in NGC 300
We present the discovery of a powerful candidate microquasar in NGC 300, associated with the S 10 optical nebula (previously classified as a supernova remnant). Chandra images show four discrete X-ray knots aligned in the plane of the sky over a length of 150 pc. The X-ray emission from the knots is well fitted with a thermal plasma model at a temperature of 0.6 keV and a combined 0.3–8 keV luminosity of 10 erg s. The X-ray core, if present at all, does not stand out above the thermal emission of the knots: this suggests that the accreting compact object is either currently in a dim state or occulted from our view. We interpret the emission from the knots as the result of shocks from the interaction of a jet with the interstellar medium (possibly over different epochs of enhanced activity). Cooler shock-heated gas is likely the origin also of the optical bubble and lobes near the X-ray structure, detected in images from the Hubble Space Telescope and the Very Large Telescope. In the radio bands, we observed the region with the Australia Telescope Compact Array, and discovered an elongated radio nebula (about 170 55 pc in size) with its major axis aligned with the chain of Chandra sources. The radio nebula has an integrated GHz radio luminosity of 10 for a distance of 1.88 Mpc. The morphology, size and luminosity of the extended X-ray, optical and radio structure suggest that NGC 300-S 10 belongs to the same class of powerful ( erg s) microquasars as SS 433, Ho II X-1 and NGC 7793-S 26.
keywords:accretion, accretion disks – stars: black holes – X-rays: binaries
Collimated jets can carry away as kinetic energy a significant fraction of the total accretion power of a compact object (e.g., 1999PhR...311..225L; 2004MNRAS.355.1105F; 2006MNRAS.372...21A; 2013ApJ...762..103K; 2014Natur.515..376G). This energy is imparted into the surrounding gas, and often creates large-scale structures such as bubbles, lobes and hotspots, detected both around supermassive black holes (e.g., 1984ApJ...285L..35P; 1984RvMP...56..255B; 1998MNRAS.296..445H; 2005ApJ...622..797K) and around off-nuclear (stellar-mass) compact objects (e.g., 1992Natur.358..215M; 1998AJ....116.1842D; 1999ARA&A..37..409M; 2002Sci...298..196C; 2005Natur.436..819G; 2010Natur.466..209P). In the case of supermassive systems, the injection of mechanical power has substantial feedback effects on the evolution of the host galaxy and the growth of the black hole itself (e.g., 2012ARA&A..50..455F; 2012NJPh...14e5023M; 2013Sci...341.1082M; 2015ARA&A..53..115K). In stellar-mass systems (often referred to as “microquasars”), jets may not be powerful enough to affect the global structure of their host galaxy; however, they still impart significant kinetic energy into the interstellar medium (ISM) and their cumulative effect can influence the evolution of galactic magnetic fields (2005MNRAS.360.1085F; 2008ApJ...686.1145H). Thus, the most powerful microquasars in the Milky Way and nearby galaxies are excellent laboratories for studies of how jets impact their surroundings. An advantage of microquasars is that, among them, we find the nearest examples of steady, highly super-Eddington accretion in the local universe (2004ASPRv..12....1F; 2007MNRAS.377.1187P; 2014Sci...343.1330S). Instead, most nuclear supermassive black holes grow at or below the Eddington limit (e.g., 2010MNRAS.402.2637S; 2010MNRAS.408L..95K; 2012MNRAS.425..623L) except perhaps during the heavily obscured early phase of quasar growth (e.g., 2003ApJ...596L..27K; 2014ApJ...784L..38M; 2015ApJ...804..148V; 2017ApJ...836L...1T), or during transient episodes of tidal disruption events (e.g., 2009MNRAS.400.2070S; 2014ApJ...781...82C; 2016ApJ...819L..25A).
There is both observational and theoretical evidence for jets from candidate super-Eddington compact accretors in ultraluminous X-ray sources (ULXs: 2011NewAR..55..166F; 2017ARA&A..55..303K). Magneto-hydrodynamic (MHD) simulations (e.g., 2005ApJ...628..368O; 2011ApJ...736....2O; 2014ApJ...796..106J; 2015MNRAS.453.3213S; 2017PASJ...69...92K; 2017MNRAS.469.2997N) indicate that powerful outflows are launched from the inner part of the super-critical disk, where radiation pressure forces dominate over gravitational forces; the outflows generate a lower-density polar funnel, inside which a relativistic jet is collimated and accelerated. However, it is unclear whether or not this polar funnel and associated jets are a necessary component of every system in the super-critical regime (regardless of the mass, spin, magnetic field and nature of the compact object) or instead there are alternative solutions with and without a relativistic jet. For example, if the super-critical accretor is a strongly magnetized neutron star, the inner disk may be truncated at the magnetospheric radius and the thick outflow funnel may not form (2015MNRAS.454.2539M). It is also still actively debated (2012MNRAS.419L..69N; 2013ApJ...762..104S; 2013MNRAS.431..405R; 2014MNRAS.439.1740M) whether jets are primarily powered by the spin-down of the compact object via a Blandford-Znajek process (1977MNRAS.179..433B; 2011MNRAS.418L..79T; 2015MNRAS.454L...6M) or instead by the accretion power released in the disk (1982MNRAS.199..883B).
One example of a super-Eddington source with collimated jets is Ho II X-1. This jetted ULX () has a distinct triple radio structure: two outer knots from a previous outburst of expanding ejecta and a third inner knot resulting from a more recent () ejection event (2014MNRAS.439L...1C; 2015MNRAS.452...24C). The inner central knot also appears variable, fading by a factor of at least 7.3 over (2015MNRAS.452...24C). There must be significant energy being imparted into the ISM to inflate the Ho II X-1 radio nebula (2012ApJ...749...17C). 2015MNRAS.452...24C argue that the Ho II X-1 radio nebula is inflated by flaring events from a transient jet, rather than a continuous jet. In several other cases, ULXs are surrounded by large (100–300 pc in diameter) bubbles of shock-ionized gas (2002astro.ph..2488P; 2008AIPC.1010..303P), but there is no direct evidence to attribute the mechanical power to a collimated jet rather than a broad outflow.
At slightly lower radiative luminosities, we have the candidate super-Eddington microquasar M 83-MQ1 (2014Sci...343.1330S). Unlike Ho II X-1, the X-ray emission from the central source () is not sufficient enough to exceed the ULX threshold ( and above). However, the mechanical power output (), inferred from optical and infrared diagnostic emission lines, places the source in the super-Eddington regime.
Another example of a powerful (candidate super-Eddington) off-nuclear microquasar with even lower core radiative luminosity, is S 26 in the nearby galaxy NGC 7793 (2010Natur.466..209P; 2010MNRAS.409..541S). S 26 consists of a hard non-thermal X-ray core, two radio and X-ray hotspots, radio lobes, and a radio/optical cocoon with a projected size of 300 150 pc. The hotspots are thought to be termination shocks as the jet impacts the ISM, with the X-ray emission (thermal plasma) and radio emission (synchrotron) coming from different physical processes (2010MNRAS.409..541S). While the core X-ray luminosity () is relatively low, 2010Natur.466..209P find a time-averaged mechanical power output of a few , suggesting that super-Eddington accretion is indeed taking place, for a stellar-mass accretor (). Clearly, for sources like S 26 and M 83-MQ1, the impact on their surroundings is longer-lived than their radiative powers.
Both S 26 and M83-MQ1 are likely more powerful analogues of the Galactic microquasar SS 433 (), whose powerful jet (; 1980MNRAS.192..731Z; 2004ASPRv..12....1F; 2007A&A...463..611B; 2011MNRAS.414.2838G; 2017MNRAS.467.4777F) is observed interacting with the surrounding bubble, W50, inflating the radio ‘ears’ (1998AJ....116.1842D) and creating X-ray hotspots (2007A&A...463..611B). In all three of these sources the mechanical power dominates over the observed radiative power, although it is possible that most of the directly emitted X-ray photons are occulted from our view by optically thick material in the disk plane, if the sources are seen at high inclination. These microquasars, along with the ULX Ho II X-1, are all distinct jet sources and it is currently unclear what properties of the central source and ISM leads to such a diverse range of observables. A larger sample of super-Eddington microquasars is required to understand how these objects impart their mechanical energy into the surrounding medium, what fraction of such sources have collimated jets, how the jet depends on compact object type/mass, and how the jet properties relate to those of jets in sub-Eddington microquasars.
Based on our early optical observational efforts to understand the nature of ultraluminous X-ray sources in nearby galaxies (2002astro.ph..2488P; 2003RMxAC..15..197P) we noted that several ULX nebulae had indeed previously been catalogued (e.g., 1997ApJS..108..261B; 1997ApJS..112...49M; 1997ApJS..113..333M) as unusually extended optically selected candidates for supernova remnants. Some outstanding examples previously reported or presently studied by the Strasbourg members of the present collaboration include Holmberg IX X-1 (1995ApJ...446L..75M; 2011ApJ...734...23G), M81 X-6 (SNR #22/23, 1997ApJS..112...49M; 2003RMxAC..15..197P), NGC 5585 X-1 (SNR #1, 1997ApJS..112...49M; Soria et al in prep), NGC 2403 X-1 (SNR #14/15, 1997ApJS..113..333M; Pakull et al. in prep) and microquasar S26 in NGC 7793 (1997ApJS..108..261B; 2010Natur.466..209P; 2010MNRAS.409..541S). As mentioned earlier, the latter source is not ultraluminous in X-rays, but emits ultraluminous mechanical power. Moreover, the microquasar S 26 displays a linear triple X-ray source morphology reminiscent of a much larger radio galaxy like Cyg A with its central black hole and (facing) X-ray/radio hot spots. We recall here that also the Galactic jet-source SS 433 and its radio/X-ray nebula W 50 would display such a triple point-like source morphology (e.g., 2011MNRAS.414.2838G) if observed at a distance of a few Mpc.
Searching for additional S 26/SS 433-type microquasars among supernova remnant candidates in nearby galaxies we noticed an intriguing source previously identified as a bright optical SNR by 1980A&AS...40...67D (listed as source number 2 in their Table 3; see also the finding chart in their Fig. 3). It is located in the nearby, late-type spiral NGC 300, at a distance of Mpc (2005ApJ...628..695G). The same optical source was observed and studied in more detail by 1997ApJS..108..261B, who list it as NGC 300 S 10 = DDB2. In both studies, the SNR identification is based on the high ratio between [S ii]6716,6732 and H line emission; line ratios [S ii]:H are indicative of shock-ionized gas (1973ApJ...180..725M; 1978A&A....63...63D). What is striking about S 10 is that we found four X-ray sources spatially resolved by Chandra, aligned in the plane of the sky and associated with the shock-ionized H emission (Figure 1). This is very unusual for an SNR; instead, we interpret the X-ray appearance as an unambiguous signature of a jet. Henceforth, we refer to those four X-ray sources as knots 1 through 4 (Figure 1). Spurred by this discovery, we then observed the field in the radio band with the Australia Telescope Compact Array (ATCA). Previous studies reported associated radio emission (2000ApJ...544..780P; 2004A&A...425..443P). With our new ATCA data, we found a bright, elongated radio nebula, overlapping with the X-ray jet (Figure 2). This is further evidence for the presence of collimated, relativistic ejections.
In this paper, we report on those discoveries, and analyse the multiband properties of S 10, using a combination of archival and new data to probe the connection between the X-ray jet, radio nebula and accretion phases of the central engine. In Section 2 we outline our data reduction techniques; in Section LABEL:sc:results we present our X-ray, radio and optical results; and in Section LABEL:sc:discussion we discuss the energetics of S 10 and compare it to super- and sub-Eddington jet sources.
2 Data Analysis
2.1 X-ray observations
NGC 300 has been observed by Chandra a total of five times. However, in one of those observations (ID 9883) the candidate microquasar target of our study does not fall on any of the chips. A second short observation (ID 7072), taken with HRC-I, does not have the sensitivity required to detect the source. Thus, we only used three of the five observations for our X-ray data analysis (Table LABEL:obs_tab): ACIS-I observations 12238, 16028 and 16029. We downloaded the data from the public archive, and reprocessed them using standard tasks from the Chandra Interactive Analysis of Observations (ciao) Version 4.9 software package (2006SPIE.6270E..1VF). We filtered out high particle background intervals. For our imaging analysis, we used HEASARC’s DS9 visualisation package. After we identified a number of discrete X-ray sources associated with the target of our study (as discussed in Section LABEL:xray_results_sec), we used the ciao task specextract to extract the background-subtracted spectrum for each source, in each observation. For the two southernmost sources (knots 1 and 2), we extracted the source counts from circular regions of radius ; for the other two sources (knots 3 and 4), we used elliptical regions of axes (position angle ), to avoid contamination from the brighter, neighbouring sources. A local background region was selected, approximately 3 times larger than the source regions. Spectral fitting was performed using XSPEC version 12.9.1 (1996ASPC..101...17A). Because of the low number of counts for each source, to test the goodness of our fits we used XSPEC’s implementation of W-statistics, which is Cash statistics (1979ApJ...228..939C) modified for a background-subtracted spectrum.
|Telescope||Obs ID/Filter/||Obs Date||Exposure|
|ATCA||5.5/9 GHz||2015-10-21||4.53 hr|
|5.5/9.0 GHz||2015-10-22||10.75 hr|
|5.5/9.0 GHz||2015-10-23||10.32 hr|
|5.5/9.0 GHz||2016-08-25||9.95 hr|
|5.5/9.0 GHz||2016-08-26||9.86 hr|