Vol 466 | 8 July 2010 | doi:10.1038/nature09168                                                                           ...
LETTERS                                                                                                                   ...
NATURE | Vol 466 | 8 July 2010                                                                                            ...
LETTERS                                                                                                                   ...
Upcoming SlideShare
Loading in …5
×

A 300 parsec-long jet-inflated bubble around a powerful microquasar in the galaxy ngc 7793

681 views

Published on

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
681
On SlideShare
0
From Embeds
0
Number of Embeds
120
Actions
Shares
0
Downloads
2
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

A 300 parsec-long jet-inflated bubble around a powerful microquasar in the galaxy ngc 7793

  1. 1. Vol 466 | 8 July 2010 | doi:10.1038/nature09168 LETTERSA 300-parsec-long jet-inflated bubble around apowerful microquasar in the galaxy NGC 7793Manfred W. Pakull1, Roberto Soria2 & Christian Motch1Black-hole accretion states near or above the Eddington luminosity of L0.3–10 < 1.8 3 1037 erg s21. Faint, even softer diffuse X-ray emission(the point at which radiation force outwards overcomes gravity) pervades much of the extent of the optical bubble. The morphologyare still poorly known because of the rarity of such sources. strikingly resembles that of an FRII-type powerful active galaxy, dis-Ultraluminous X-ray sources1 are the most luminous class of black playing X-ray and radio hot spots (for example Cygnus A11). The opticalhole (LX < 1040 erg s21) located outside the nuclei of active galaxies. radial velocity of S26 agrees with that of other nearby nebulae in NGCThey are likely to be accreting at super-Eddington rates, if they arepowered by black holes with masses less than 100 solar masses.They are often associated with shock-ionized nebulae2,3, though Nwith no evidence of collimated jets. Microquasars with steady jetsare much less luminous. Here we report that the large nebula S26(ref. 4) in the nearby galaxy NGC 7793 is powered by a black holewith a pair of collimated jets. It is similar to the famous Galacticsource SS433 (ref. 5), but twice as large and a few times more Epowerful. We determine a mechanical power of around a few1040 erg s21. The jets therefore seem 104 times more energetic thanthe X-ray emission from the core. S26 has the structure of aFanaroff–Riley type II (FRII-type) active galaxy: X-ray and opticalcore, X-ray hot spots, radio lobes6 and an optical and X-ray cocoon.It is a microquasar where most of the jet power is dissipated inthermal particles in the lobes rather than relativistic electrons. The very large, shock-ionized nebulae (ultraluminous X-raysource (ULX) bubbles7), with characteristic sizes of ,100–500 pcand ages of around a few 105 yr are much larger and more energeticthan normal supernova remnants. But X-ray luminous sources maybe only a subset of non-nuclear black holes at very high mass accre- 10 arcsec or 190 pction rates. We proposed8 that ionized bubbles might also be foundassociated with black holes that seem X-ray faint, because their radi- Figure 1 | Optical/X-ray image of the 300-pc-diameter jet-inflated bubbleative emission is collimated away from our line of sight, because they S26 in the galaxy NGC 7793. The contours denote the continuum-are transients and currently in a low/off accretion state, or because subtracted Ha emission, which is plotted over a true-colour Chandra/ACIS-they channel most of their accretion power into a jet even at near- S image (ObsId 5 3954; principal investigator, T. G. Pannuti). The projectedEddington mass accretion rates. distance between the X-ray hot spots is ,15 arcsec, which corresponds to Using the Chandra Data Archive, we searched for such systems ,290 pc at the distance of NGC 7793. The Chandra exposure time wasamong unusually large supernova remnants in nearby galaxies and ,50 ks. The X-ray colours are as follows: red, 0.3–1 keV; green, 1–2 keV;discovered an example in the Sculptor galaxy NGC 7793 (distance of blue, 2–8 keV. All three-colour images in the X-ray band were lightly3.9 Mpc; ref. 9). The optical/radio nebula S264,6 has a size of smoothed with a 1-arcsec Gaussian core. The continuum-subtracted Ha image was taken with the CTIO 1.5-m telescope in 2001 (exposure time of,300 pc 3 150 pc and was originally classified as a supernova rem- 600 s under seeing conditions with a full-width at half-maximum (FWHM)nant candidate; the high flux ratio between the forbidden [S II] lines of 1.2 arcsec) for the Spitzer Infrared Nearby Galaxy Survey25 and ˚ ˚at 6,716 A and 6,732 A and the Balmer Ha line indicates the presence downloaded through the NASA/IPAC Extragalactic Database. We usedof shock-ionized gas. A faint X-ray source10 was known to be asso- matching point-like sources in the full field of NGC 7793 to improve theciated with S26, but it was unresolved in the Rosat observation. The relative astrometry of the Ha and X-ray images. Note the relatively softerX-ray emission is resolved into three point-like sources that are per- X-ray colour of the hot spots, compared with the blue point-like core; evenfectly aligned and match the extent of the major axis of the optical softer, diffuse X-ray emission is detected over the whole nebula. The intrinsicnebula (Fig. 1). We interpret those sources as the core (at the X-ray X-ray luminosity of the southern hot spot is twice as high as the luminosity of the northern hot spot; this is similar to the higher Ha intensity in thatbinary position) and the hot spots (where the jets interact with the region. There is no Ha point-like source or enhanced emission at theambient medium). The core appears to be harder, with an intrinsic position of the X-ray core. The other source of Ha emission a few arc seconds0.3–10-keV power-law luminosity of L0.3–10 < 7 3 1036 erg s21, whereas to the east of S26 is an unrelated H II region. Scaling from the total Hathe hot spots have a softer spectrum and can be fitted by optically thin luminosity of NGC 779325, and assuming a Ha/Hb Balmer decrement ofthermal plasma emission (Fig. 2) with a combined intrinsic luminosity ,3.0, we determined an Hb luminosity of ,1.0 3 1038 erg s21.1 University of Strasbourg, CNRS UMR 7550, Observatoire Astronomique, 11 rue de l’Universite, F67000 Strasbourg, France. 2MSSL, University College London, Holmbury St Mary, ´Surrey RH5 6NT, UK. 209 ©2010 Macmillan Publishers Limited. All rights reserved
  2. 2. LETTERS NATURE | Vol 466 | 8 July 2010 10−2 a N Normalized count (s−1 keV−1) 10−3 10−4 10−5 10−6 2 E Ratio 1 0.5 1 2 5 Energy (keV)Figure 2 | Chandra/ACIS-S spectra of the triple X-ray source in S26. Upper 5 arcsecplot: emission from the two hot spots (added together) and from the centralsource of S26 are shown with red and blue data points, respectively. Theblack lines show the best-fitting models to the thermal lobe emission andpower-law core emission. Lower plot: the ratio between the data and the bmodels. The green line indicates a ratio of 1. The data were extracted from Nthe Chandra archive and analysed with the standard software packagesCIAO and XSPEC. We found that the central source has a hard spectrum(photon index, C 5 1.4 6 0.5) and an intrinsic 0.3–10-keV luminosity ofL0.3–10 < 7 3 1036 erg s21, which is much less than the Eddington limit of aneutron star or a stellar-mass black hole. The emission from the hot spots issignificantly softer and is well fitted by a two-component thermal plasmamodel with kBT1 5 0:26z0:05 keV and kBT2 5 0:96z0:31 keV (contributing {0:08 {0:17roughly the same as the total flux), and no significant absorption above theGalactic line-of-sight column density, NH 5 1.2 3 1020 cm22. We derive Eintrinsic luminosities of L0.3–10 < 1.2 3 1037 erg s21 andL0.3–10 < 0.6 3 1037 erg s21 for the southern and northern hot spots,respectively. The particle density, nX, of the X-ray-emitting, shocked thermalplasma (assumed to have solar abundance) can be estimated asnX < 2h23/2 cm23, where h is the diameter of the emitting hot spots in unitsof arc seconds. A characteristic hot spot size of ,1 arcsec is suggested by themarginally resolved Chandra image. This gives a mass for the X-ray-emittinggas of ,500h23/2 solar masses, more than can be supplied by a donor starorbiting the black hole. It probably represents a mix between the very hot butdilute jet gas that has passed through the reverse (Mach) shock and the much 5 arcsecdenser material behind the foreward (bow) shock, which is advancing intothe interstellar medium (ISM). Alternatively, a very steep power-lawspectrum with C 5 5.7 6 1.4 and NH 5 (4 6 2) 3 1021 cm22 cannot Figure 3 | The stellar and high-excitation nebular content of S26.presently be excluded, but seems to contradict both the low observed optical a, Optical-continuum greyscale image of S26, taken with the FORS1reddening4 and the expected slope for a synchrotron component extending instrument on the ESO Very Large Telescope (VLT) in 2002 andfrom the radio to the X-ray bands. Error bars, 1s. downloaded from the public archive. The narrowband filter used for this ˚ ˚ image was centred at 5,105 A, with a FWHM of 61 A (exposure time of 1,600 s under seeing conditions with a FWHM of 1.0–1.5 arcsec). The size7793, effectively ruling out a chance superposition of a background and orientation of the image are as in Fig. 1. The positions of the X-ray coreactive galactic nucleus. and hot spots have been overplotted as red circles of radius 0.7 arcsec. Note ˚ Optical narrow-band He II 4,686 A observations and nearby con- the relative brightness of the optical counterpart to the X-ray core; wetinuum imaging observations (Fig. 3) show that NGC 7793 emits estimate B < 23 mag, corresponding to MB < 25 mag. b, Continuum- ˚significant amounts of He21 R He1 4,686 A recombination photons ˚ subtracted greyscale FORS1 image in the He II 4,686 A emissionwith approximately the same spatial distribution as Ha emission. This ˚ ˚ (narrowband filter centred at 4,684 A, with a FWHM of 65 A; exposure timeimplies that either the central star is both extremely luminous and and seeing conditions were the same as for the image in a). The image wassufficiently hot to photoionize the large He III region (that is, smoothed with a 0.6-arcsec Gaussian core, to highlight the extended nebular ˚ line emission. The observed flux ratio between He II 4,686 A emission andTeff $ 80,000 K), or that the gas has been ionized by a very fast shock the H II Balmer lines suggests shock ionization with vshock < 275 km s21.wave. We will show below that the second possibility is realized in S26. ˚ Note also the point-like He II 4,686 A emission from the core, for which we The images show (Fig. 3a) that the optical counterpart of the X-ray ˚ estimate an equivalent width of ,30 A. The green lines show the positioncore is a blue stellar object with an absolute magnitude of and width of the slit used to acquire the spectrum shown in Fig. 4. ˚MB < 25 mag. Moreover, we detect an excess of stellar 4,686 A emis- ˚sion (Fig. 3b) with an equivalent width of ,30 A. This is similar to We have now obtained key optical spectroscopic observations thatwhat we expect from an early-type Wolf–Rayet star (type WNE). As allow us to measure the expansion (shock) velocity of S26 directlyan aside, the mass donor in IC 10 X-112, the most massive stellar-mass and thus estimate the energetics of the system. From the half-width atblack hole presently known13 and which is also embedded in a large zero intensity of the main emission lines, we estimate that the gas issynchrotron bubble, is an evolved star of this type. This bubble and supersonically expanding with a maximum velocity envelope ofS26 have sometimes been attributed to a hypothetical hypernova vexp < 250 km s21 (Fig. 4). An independent estimate of the shock ve-event6,14,15. locity can be derived from the relative strength of the He II 4,686 A˚210 ©2010 Macmillan Publishers Limited. All rights reserved
  3. 3. NATURE | Vol 466 | 8 July 2010 LETTERS high power for a non-nuclear accreting source. Using the formalism [N II] Hα [N II] outlined in refs 7, 19, we find the hydrogen particle density of the ISM into which S26 expands to be n < 0.7 cm23. The mechanical energy of the swept-up shell is 1/2Mvexp2 5 (15/77)Qjett (ref. 17), where M 5 (4p/3)1.38mpnR3 (mp being the mass of the proton). This yields the same value for Qjet as derived above. The long-term-averaged jet power observed in S26 is an order of magnitude higher than what is estimated for the already very power- ful, mildly relativistic jets from the Galactic microquasar SS4335,20. Perhaps more importantly, we find that Qjet exceeds the apparent X-ray luminosity of the core by a factor of about 104. A similar 5 arcsec dominance of mechanical power over (observed) radiative power in SS433 has traditionally been ascribed to absorption of X-rays along 500 km s–1 our line of sight (source seen along the plane of the accretion disk21). The discovery of a second, even more extreme X-ray-faint jet source,Figure 4 | Spectrum of Doppler-broadened emission lines of S26 indicating with a mechanical power well above the Eddington limit for a stellar-an expansion velocity of 250 km s21. Part of a medium-resolution (FWHM, ˚,0.7 A) long-slit ESO VLT FORS2 spectrum (taken in October 2009) mass black hole, suggests that there may be accretion modes that ˚covering the wavelength range between 3,600 and 7,200 A, with the slit channel most of the available power into jets rather than photons,running across the eastern body of S26 as shown in Fig. 3. The exposure even at extremely high mass accretion rates22.times for the red and blue settings were 2,200 s under good seeing conditions With its rich, interconnected structure of optical, X-ray and radio(FWHM, 0.7–0.8 arcsec). The reduced spectra were analysed using ESO- emission, S26 provides new observational tests of our physical under-MIDAS routines. The dispersion is along the horizontal axis with increasing standing of accretion-powered astrophysical sources and of theirwavelengths towards the right. In the vertical (spatial) direction, one pixel energy transfer into the surrounding gas. S26 is the missing linkcorresponds to ,0.2 arcsec; the total extent of S26 is ,10 arcsec along the between the similarly large and energetic ULX bubbles (in whichslit. Narrow, constant-intensity, vertical emission lines are due tospectroscopically unresolved atmospheric O and OH night glow; the most there is yet no direct evidence of collimated jets) and the Galactic ˚prominent emission lines from S26 are, from left to right, [N II] 6,548 A, Ha jet source SS433, with its comparatively smaller, fainter nebula, W50. ˚and [N II] 6,584 A. The ‘bulged’ appearance, mostly visible in the intense Ha It is also a true non-nuclear analogue of powerful FRII-type activeemission, reflects the remarkable kinematics of the emitting material in S26. galaxies, with their persistent radio lobes.The small extent in wavelength at the lower and upper nebular boundaries The jet power we have derived for S26 is orders of magnitude largerreflects the small velocities of the emitting gas along our line of sight (that is, than the value that would be derived from the radio luminosity of themostly perpendicular motion). The central part of the line covers the whole nebula using the relations for microquasar lobe dynamics23. Therange of radial velocities across the nebula, which we assign to expansive reason for this discrepancy is the underlying assumption that mostmotion from zero up to 250 km s21 around a central velocity that is close tothe apparent radial velocity in that part of NGC 7793. The density of the ISM of the power is converted into non-thermal relativistic particles(n < 0.7 cm23; see text) is high enough that the cooling time behind the (including the radiating leptons) and into magnetic fields. These inshock, ,200v1004.4/Zn yr (ref. 26), where v100 5 vshock/(100 km s21) and Z is turn are thought to provide the pressure that drives the expansion ofthe metallicity relative to the solar value, is smaller than the age of the the microquasar lobes. But much more thermal energy may be storedbubble. Therefore, the cooling/recombination zone behind the shock is in non-relativistic protons and nuclei. Much higher kinetic jet powerlargely complete, that is, the shock is largely radiative27, in agreement with than previously thought, possibly even surpassing the Eddingtonthe presence of relatively strong low-ionisation or neutral species. One such luminosity, has recently been advocated for several powerful FRII- ˚diagnostic species is [O I], which emits the 6,300 A line4 with Il6,300/ type active galaxies24; this suggests that most of the energy in theirIHb < 0.63. cocoons is carried by thermal particles, as seems to be the case for S26.line with respect to the H I Balmer emission. We measure a flux ratio of Received 22 January; accepted 5 May 2010.Il4,686/IHb < 0.09–0.12. From a standard library of radiative shocks16,we find that this diagnostic line ratio implies vshock < 275 6 25 km s21 1. Ward, M. Luminous X-ray sources in spiral and star-forming galaxies. Phil. Trans. R. Soc. Lond. A 360, 1991–2003 (2002).if precursor emission is taken into account, in remarkable agreement 2. Pakull, M. W. & Mirioni, L. Optical counterparts of ultraluminous X-ray sources.with the kinematic velocity estimate. Furthermore, the diagnostic [O Preprint at Æhttp://arxiv.org/abs/astro-ph/0202488æ (2002).III] forbidden line ratio Il5,007/Il4,363 corresponds to an electron tem- 3. Pakull, M. W. & Mirioni, L. Bubble nebulae around ultraluminous X-ray sources.perature of Te < 27,000 K, which is typical of shock-ionized gas and Rev. Mex. Astron. Astrofis. (Ser. Conf.) 15, 197–199 (2003).inconsistent with photoionization, which would yield electron tem- 4. Blair, W. P. & Long, K. S. Identification of supernova remnants in the Sculptor galaxies NGC 300 and NGC 7793. Astrophys. J. Suppl. Ser. 108, 261–277 (1997).peratures closer to ,10,000 K. 5. Fabrika, S. The jets and the supercritical accretion disk in SS433. Astrophys. Space We are now in a position to estimate the energetics of this jet- Phys. Rev. 12, 1–153 (2004).inflated bubble reliably. To this end, we use the well-known self- 6. Pannuti, T. G. et al. An X-ray, optical and radio search for supernova remnants insimilar expansion law17,18 as a function of time, t, for wind-driven the nearby Sculptor group Sd galaxy NGC 7793. Astrophys. J. 565, 966–981 (2002).stellar bubbles and jet-driven radio lobes: R < 0.76(Qjet/r0)1/5t3/5. ´ 7. Pakull, M. W., Grise, F. & Motch, C. in Populations of High Energy Sources in GalaxiesHere Qjet is the long-term-averaged jet power, R is the (mean) radius (eds Meurs, E. J. A. & Fabbiano, G.) 293–297 (IAU Symp. 230, Cambridge Univ.of the bubble expanding with velocity vexp 5 dR/dt into the ISM, the Press, 2006).density of which, r0, is assumed to be constant. The first conclusion we ´ 8. Pakull, M. W. & Grise, F. in A Population Explosion: The Nature & Evolution of X-raycan draw is that the characteristic age of the bubble is t 5 (3/5)(R/ Binaries in Diverse Environments (eds Bandyopadhyay, R. M., Wachter, S., Gelino, D. & Gelino, C. R.) 303–307 (AIP Conf. Proc. 1010, American Institute of Physics,vexp) < 2 3 105 yr. The jet power, Qjet, can be derived from standard 2008).bubble theory17, which implies that a fraction of 27/77 of Qjet is emitted 9. Karachentsev, I. D. et al. Distances to nearby galaxies in Sculptor. Astron.by a fully radiative shock expanding into the ISM. Furthermore, radi- Astrophys. 404, 93–111 (2003).ative shock models19 predict that for an observed vshock < 275 km s21, 10. Read, A. M. & Pietsch, W. ROSAT observations of the Sculptor galaxy NGC 7793.0.77% of the radiated energy appears as Hb photons if we include the Astron. Astrophys. 341, 8–22 (1999). 11. Wilson, A. S., Smith, D. A. & Young, A. J. The cavity of Cyg A. Astrophys. J. 644,effect of a radiative precursor. This implies that the Hb luminosity L9–L12 (2006).(,1.0 3 1038 erg s21; see Figure 1) can be used to estimate the jet power 12. Clark, J. S. & Crowther, P. A. On the Wolf-Rayet Counterpart to IC 10 X-1. Astron.directly. We find that Qjet < 5 3 1040 erg s21, which is a remarkably Astrophys. 414, L45–L48 (2004). 211 ©2010 Macmillan Publishers Limited. All rights reserved
  4. 4. LETTERS NATURE | Vol 466 | 8 July 201013. Silverman, J. M. & Filippenko, A. V. On IC 10 X-1, the most massive known stellar- 25. Kennicutt, R. C. Jr et al. An Ha imaging survey of galaxies in the local 11 Mpc mass black hole. Astrophys. J. 678, L17–L20 (2009). volume. Astrophys. J. Suppl. Ser. 178, 247–279 (2008).14. Lozinskaya, T. A. & Moiseev, A. V. A synchrotron superbubble in the IC 10 galaxy: 26. Dopita, M. A. & Sutherland, R. S. Astrophysics of the Diffuse Universe 200-204 a hypernova remnant? Mon. Not. R. Astron. Soc. 381, L26–L29 (2007). (Springer, 2003).15. Asvarov, A. I. Radio emission from shell-type supernova remnants. Astron. 27. Raymond, J. C. et al. Spatial and spectral interpretation of a bright filament in the Astrophys. 459, 519–533 (2006). Cygnus Loop. Astrophys. J. 324, 869–892 (1988).16. Allen, M. G., Groves, B. A., Dopita, M. A., Sutherland, R. S. & Kewley, L. J. The MAPPINGS III library of fast radiative shock models. Astrophys. J. Suppl. Ser. 178, Acknowledgements This work was based on observations made with ESO 20–55 (2008). telescopes at the Paranal Observatory under program 084.D-0881 (M.W.P.), and17. Weaver, R., McCray, R., Castor, J., Shapiro, P. & Moore, R. Interstellar bubbles. II. made use of the ESO/ST-ECF Science Archive facility, which is a joint collaboration Structure and evolution. Astrophys. J. 218, 377–395 (1977). of the ESO and the Space Telescope European Coordinating Facility. Data were18. Kaiser, C. R. & Alexander, P. A self-similar model for extragalactic radio sources. obtained from the Chandra Data Archive and software was provided by the Chandra Mon. Not. R. Astron. Soc. 286, 215–222 (1997). X-ray Center. R.S. acknowledges support from an Early-Career Leverhulme19. Dopita, M. A. & Sutherland, R. S. Spectral signatures of fast shocks. I. Low-density Fellowship and hospitality at the University of Sydney during part of this work. model grid. Astrophys. J. Suppl. Ser. 102, 161–188 (1996). Author Contributions M.W.P. designed the study and analysed the optical20. Begelman, M. C., Hatchett, S. P., McKee, C. F., Sarazin, C. L. & Aarons, J. Beam spectroscopic observations. R.S. contributed the X-ray spectral analysis and the models for SS433. Astrophys. J. 238, 722–730 (1980). relative astrometric calibrations of the multiband datasets. C.M. carried out the21. Begelman, M. C., King, A. R. & Pringle, J. E. The nature of SS433 and the reductions and analysis of the optical imaging data. All authors discussed the ultraluminous X-ray sources. Mon. Not. R. Astron. Soc. 370, 399–404 (2006). results and made substantial contributions to the manuscript.22. Bogovalov, S. V. & Kel’ner, S. R. Dissipationless disk accretion. Astron. Rep. 49, 57–70 (2005). Author Information Reprints and permissions information is available at23. Heinz, S. Radio lobe dynamics and the environment of microquasars. Astron. www.nature.com/reprints. The authors declare no competing financial interests. Astrophys. 388, L40–L43 (2002). Readers are welcome to comment on the online version of this article at24. Ito, H., Kino, M., Nozomu, K., Isobe, N. & Shoishi, Y. The estimate of kinetic power www.nature.com/nature. Correspondence and requests for materials should be of jets in FR II radio galaxies: existence of invisible components? Astrophys. J. 685, addressed to M.W.P. (manfred.pakull@astro.unistra.fr) or R.S. 828–838 (2008). (rsoria@physics.usyd.edu.au).212 ©2010 Macmillan Publishers Limited. All rights reserved

×