New Galactic High Mass X-ray Binaries discovered with INTEGRAL
aa r X i v : . [ a s t r o - ph . H E ] O c t New Galactic High Mass X–ray Binaries discovered withINTEGRAL
Lara Sidoli
INAF-IASF, via Bassini 15, I-20133 Milano, Italy
Abstract.
I will review the main observational properties of the new Galactic High Mass X–ray Binaries (HMXBs) discov-ered by the
INT EGRAL satellite in the hard energy range 17–100 keV. About 70% of the newly discovered HMXBs hostOB supergiant companions and show peculiar properties with respect to classical HMXBs detected with previous missions:some of them display huge local absorptions, in excess of 10 cm − (the so-called obscured sources ), while others showfast transient X–ray emission, leading to the definition of a new sub-class of HMXBs, the so-called Supergiant Fast X–rayTransients (SFXTs). Their peculiar behavior is still poorly understood and represents a challenge to theory.
Keywords:
X–ray binaries, X–ray sources, accretion and accretion disks, supergiants
PACS:
NEW GALACTIC HIGH MASS X–RAY BINARIES: THE ROLE OF INTEGRALDISCOVERIES
Our view of the X–ray sky changed since the launch of the
International Gamma-Ray Astrophysics Laboratory inOctober 2002 (
INT EGRAL , [1]): 723 hard X–ray sources have been observed down to a sensitivity limit at mCrablevel [2] in the energy band 17–100 keV, 30% of which are still unidentified.Among all detected sources, 13% are High Mass X–ray Binaries (HMXBs). About 50 objects among the newsources discovered with
INT EGRAL ( IGR sources), have been classified as HMXBs. About 70% of them hostsupergiant donors (see also the IGRs on-line list at http://irfu.cea.fr/Sap/IGR-Sources/ ), almosttripling the number of known Galactic HMXBs with supergiant companions.Among these new HMXBs, two peculiar behaviors have been recognized: • the so-called obscured sources , which display huge amount of low energy absorption (well in excess of10 cm − , thought to be local because of its variability, and produced by the dense wind of the supergiantcompanion; [3, 4]); • a new subclass of transients, the Supergiant Fast X–ray Transients (SFXTs; [5, 6]). Note that a high absorbingcolumn density is not a characterizing property of SFXTs, although a few SFXTs show high and variableabsorption.The discovery of new types of supergiant HMXBs was somehow unexpected: indeed, before
INT EGRAL discov-eries, HMXBs hosting OB supergiants were known to exhibit persistent X–ray emission, driven by the accretion fromthe strong wind of the blue supergiant companions, mainly in narrow orbits.They were thought to be easy to detect because of their luminous persistent emission ( ∼ erg s − ). The numberof Galactic wind-fed HMXBs composed of pulsating neutron stars and OB supergiant companions was limited to afew sources (Vela X–1, 2S0114+650, GX301–2, 1E1145.1–6141, 4U1538–53, X1908+075, XTE J1855–026).Then, after the launch of INT EGRAL , thanks to its sensitivity at hard X–rays and the monitoring strategy of theGalactic plane, several new HMXBs could be discovered and identified, either highly obscured or with fast transientX–ray emission with long duty cycles, very difficult to discover with previous missions.
Obscured sources: the prototype IGRJ 16318–4848
IGRJ16318–4848 is the first source discovered in 2003 by the IBIS/ISGRI detector on-board
INT EGRAL [7]. The source displayed a variable hard X–ray emission on time scales of 1000 s, with an average flux of × − erg cm − s − (20–50 keV). XMM − Newton follow-up observations allowed to refine the position [8], and revealed an extremely absorbed X–ray spectrum, with an absorbing column density of 2 × cm − , together with prominent iron K a and K b , and nichelK a emission lines ([9], [10], [3], [11]), suggestive of the presence of a dense local envelope ([10], [12], [13]).X–ray observations with Suzaku [14] confirmed the X–ray spectrum rich in iron and nichel emission lines andallowed a truly simultaneous broad band spectroscopy (1–60 keV). These authors confirmed the huge amount ofabsorption, but found a significantly harder X–ray spectrum, well described with a flat power law with a photon indexof ∼ × erg s − and 6.2 × erg s − (2–10 keV). If it is located in theNorma Arm, it lies at a distance of about 4.8 kpc [15]. The optical to MIR spectral energy distribution suggests thepresence of a mid-infrared excess above the stellar spectrum, indicative of dust. This findings, together with the highabsorption, imply dust and dense cool gas enshrouding the whole binary system ([16], [17]).The B[e]-phenomenon is peculiar of very different kinds of stars (from supergiants to pre-main sequence stars)and is characterized by spectra with a near infrared excess (attributed to emission from warm dust) and by forbiddenemission lines (e.g. Fe II) [18]. In particular, B[e] supergiants (sgB[e], [19, 18]) are thought to be in an evolutionarystage intermediate between OB supergiant and Wolf Rayet star [20]. Their properties have been explained in terms ofa two component stellar wind, with the presence of a circumstellar outflowing disk wind [19].After IGRJ16318–4848, INT EGRAL discovered other similar sources displaying high local absorption at softX–rays (larger than 10 cm − ) associated with blue supergiant companions. Thus, IGRJ16318–4848 became theprototype of the so-called obscured HMXBs, being the first and most extreme example of this sub-class of massiveX–ray binaries (e.g. [13, 21]). It is also one of the rare examples of HMXB with a companion showing the B[e]phenomenon [20].
Supergiant Fast X–ray Transients
Supergiant Fast X–ray Transients (SFXTs) are a sub-class of massive X–ray binaries recognized mainly thanks to
INT EGRAL discoveries [5, 6].SFXTs are characterized by transient and fast X–ray emission (composed by short and bright flares, with a durationof a few hours, as observed with
INT EGRAL , [5], [6]) and by the association with OB supergiant companions (e.g.[22], [23], [24], [16], [25]). These two main properties have led to the identification of a new class of 9 members to date(together with several other SFXTs candidates, with fast transient X–ray emission, but still unknown counterparts).Optical/IR spectroscopy allowed the determination of the source distance, implying an X–ray luminosity at the flarepeak of 10 –10 erg s − . INT EGRAL is able to observe the brightest flares, with no persistent emission outside flaring activity. More sensitiveinstruments allow to better investigate the properties of SFXTs when they are not undergoing an outburst.The quiescence, characterized by a very soft (likely thermal) spectrum, has been rarely observed in SFXTs. It isat a luminosity level of ∼ erg s − , as observed in IGR J17544–2619 with Chandra [26]. In literature, the term“quiescence” is often used with different meanings, sometimes simply to indicate the X–ray emission when the sourceis not undergoing a bright outburst (“out-of-outburst” emission), whereas we mean here a specific source state at10 erg s − , where no accretion onto the compact object is present and the spectrum is very soft and thermal [26].This quiescent luminosity implies a very high dynamic range (ratio between the peak luminosity and the luminosityduring quiescence) of 4-5 orders of magnitude.Observations with XMM − Newton caught a few SFXTs in a very low level of emission (L X ∼ erg s − ), showinga hard 0.5–10 keV spectrum and faint very short flares, suggestive of some level of accretion onto the compact object[27, 28].The presence of a very low level of accretion is also suggested by our recent observation of the SFXT IGR J08408–4503 with Suzaku [29] performed in December 2009, which caught the source in an initial low intensity state at4 × erg s − ∼ ∼ cm [29]. IGURE 1.
INTEGRAL results on SFXTs: duty cycles (percentage of time spent in bright flares with respect to the totalobserving time with
INT EGRAL of the source position) calculated from data reported by [31]. The sources are both confirmedand candidate SFXTs observed during the survey of the central region of our Galaxy (see [31] for details).
FIGURE 2.
New Galactic X–ray binaries discovered by
INT EGRAL (red squares and source names) in the Corbet diagram ofspin period versus orbital period. The solid triangles indicate previously known Galactic HMXBs and Be/X–ray transients. Notethat almost all the new Galactic IGRs reported here are HMXBs hosting supergiant companions, except IGR J16393–4643 (likelya Symbiotic system) and IGR J11435–6109 (Be star). The nature of the donor star in IGR J19294+1816 (main sequence Be or OBsupergiant) is still unclear [32].
A monitoring with Swift/XRT of 4 members of the class, spanning about two years of observations (starting inOctober 2006, [33, 34]), demonstrated that the quiescence is a rare state for these transients. Indeed, most of theirlifetime is spent in an intermediate state of emission in the range between 10 and 10 erg s − (0.3–10 keV), witha hard spectrum (well described by a hard power law with photon index of ∼ ∼ ABLE 1.
List of Supergiant Fast X–ray Transients
Source Orbital Period(d) Spin Period(s) References
IGRJ 08408–4503 35 (?) - P orb : [40]IGRJ 11215–5952 164.6 186.78 ± . orb : [35, 46]; P spin : [47, 45]IGRJ 16465–4507 30.243 ± .
035 228 ± orb : [48, 49]; P spin : [50]IGRJ 16479–4514 3.3194 ± . orb : [51]XTE J1739–302 51.47 ± .
02 - P orb : [52]IGRJ 17544–2619 4.926 ± .
001 - P orb : [53]SAXJ 1818.6–1703 30.0 ± . orb : [54, 55]AX J1841.0–0536 - 4.7394 ± . spin : [56]IGR J18483–0311 18.55 ± .
03 21.0526 ± . orb : [57]; P spin : [58] keV) during bright flares is well fitted by an absorbed hard power law (photon index ∼
1) with a high energy cutoff at ∼ Swi f t /XRTmonitoring of other 4 SFXTs during their outbursts [41]. Each ouburst is actually composed by an enhanced averageX–ray emission together with several bright flares lasting a few hours ([42, 44]).Source duty cycles are small, although highly variable from source to source: during the Swift/XRT monitoring, thetime spent in bright outburst is 3%-5% in 3 SFXTs, [34], while with
INT EGRAL it is much smaller (see Fig. 1). Thefigure has been obtained from data reported by [31] and includes both confirmed and candidate SFXTs after sevenyears of
INT EGRAL operations.A few SFXTs are X–ray pulsars (see Table 1 and Fig. 2), thus demonstrating that the compact object is a neutronstar, while in the other sources a black hole cannot be excluded, although the spectral similarity to accreting pulsars issuggestive of a neutron star. The SFXTs spin periods are in the range from 4.7 s to 228 s. The orbital periods are alsovery different, between 3.3 days and 165 days (Table 1). Most of them have been determined from the modulation ofthe X–ray light curve, while in the case of IGRJ 11215–5952 [35] it has been derived from the periodically recurrentoutbursts. The orbital period of ∼
35 days in IGRJ 08408–4503 (Table 1) has been suggested by [40] based on theduration of the flares, which is thought to be linked to the orbital separation, in the framework of the new windaccretion model of [59]. Thus, it needs to be confirmed by timing analysis.Interestingly, a few of the newly discovered HMXBs hosting supergiants lie in the region typical of Be/X–raytransients (IGRJ 11215–5952, IGR J19294+1816, IGRJ 18483–0311) in the Corbet diagram (Fig. 2), probablysuggesting a possible evolutionary link with these transients (see, e.g., [60] and Chaty 2010, these proceedings).
The mystery of SFXTs
SFXTs are massive binaries where the X–ray emission is thought to be produced by the direct accretion of thesupergiant wind onto the compact object, although the possible formation of transient accretion disks have beensuggested [31]. On the other hand, the wind accretion alone cannot explain a so different behavior in two sub-classesof HMXBs (SFXTs vs persistent HMXBs with supergiant donors) where the donor star, the compact object (neutronstar), the orbital parameters seem to be very similar. SFXTs like IGRJ 16479–4514 [51] and IGRJ 17544–2619 [54, 55]are in narrow orbits, even narrower than several persistent HMXBs, thus ruling out the hypothesis of wide eccentricorbits [61] as the main property which makes the difference between these SFXTs and persistent HMXBs containingsupergiants.Different mechanisms have been proposed to explain SFXTs (see [62] for a review): accretion from clumpy winds,where the SFXTs flares are produced by the sudden accretion of a dense clump [26, 63] in a binary system with awide and possibly eccentric orbit [61]; accretion from aspherical winds (with a preferential plane for the outflowingwind), where the flares are triggered by the neutron star passage inside this outflow [45]; centrifugal or magneticbarriers which halt the accretion onto the neutron star for most of the time [64, 65]. Each these mechanisms canhardly explain the behavior of all
SFXTs, whose link with persistent HMXBs remain unclear. It is also possible thatifferent mechanisms are at work together in a single SFXT, or, alternatively, that SFXTs is a non-homogeneous class.Thus, despite the large amount of observational data, there are still several open issues that need to be addressed: theaccretion mechanism, the link between different kinds of HMXBs, the SFXTs evolutionary path and formation history.
ACKNOWLEDGMENTS
This work was supported in Italy by contract ASI/INAF I/088/06/0.
REFERENCES
1. C. Winkler, T. Courvoisier, G. Di Cocco, N. Gehrels, A. Giménez, S. Grebenev, W. Hermsen, J. M. Mas-Hesse, and et al.,
A&A , L1–L6 (2003).2. A. J. Bird, A. Bazzano, L. Bassani, F. Capitanio, M. Fiocchi, A. B. Hill, A. Malizia, V. A. McBride, S. Scaringi, V. Sguera,J. B. Stephen, P. Ubertini, A. J. Dean, F. Lebrun, R. Terrier, M. Renaud, F. Mattana, D. Götz, J. Rodriguez, G. Belanger,R. Walter, and C. Winkler,
ApJS , 1–9 (2010), .3. R. Walter, J. Rodriguez, L. Foschini, J. de Plaa, S. Corbel, T. J.-L. Courvoisier, P. R. den Hartog, F. Lebrun, A. N. Parmar,J. A. Tomsick, and P. Ubertini,
A&A , L427–L432 (2003), arXiv:astro-ph/0309536 .4. R. Walter, J. Zurita Heras, L. Bassani, A. Bazzano, A. Bodaghee, A. Dean, P. Dubath, A. N. Parmar, M. Renaud, andP. Ubertini,
A&A , 133–143 (2006).5. V. Sguera, E. J. Barlow, A. J. Bird, D. J. Clark, A. J. Dean, A. B. Hill, L. Moran, S. E. Shaw, D. R. Willis, A. Bazzano,P. Ubertini, and A. Malizia,
A&A , 221–231 (2005), astro-ph/0509018 .6. I. Negueruela, D. M. Smith, P. Reig, S. Chaty, and J. M. Torrejón, “Supergiant Fast X-ray Transients: a new class of high massX-ray binaries unveiled by INTEGRAL,” in
Proc. of the “The X-ray Universe 2005”, 26-30 September 2005, El Escorial,Madrid, Spain. Ed. by A. Wilson. ESA SP-604, Volume 1, Noordwijk: ESA Pub. Division, ISBN 92-9092-915-4, 2006 , 2006.7. T. Courvoisier, R. Walter, J. Rodriguez, L. Bouchet, and A. A. Lutovinov,
IAU Circ. , 3–+ (2003).8. N. Schartel, M. Ehle, M. Breitfellner, M. Guainazzi, P. Rodriguez-Pascual, M. Santos-Lleo, P. Calderon-Riano, P. Munuera-Gallardo, B. Gonzalez-Garcia, and R. Perez-Martinez,
IAU Circ. , 3–+ (2003).9. J. de Plaa, P. R. den Hartog, J. S. Kaastra, J. J. M. in’t Zand, M. Mendez, and W. Hermsen,
The Astronomer’s Telegram ,1–+ (2003).10. G. Matt, and M. Guainazzi,
MNRAS , L13–L17 (2003), arXiv:astro-ph/0303626 .11. A. Ibarra, G. Matt, M. Guainazzi, E. Kuulkers, E. Jiménez-Bailón, J. Rodriguez, F. Nicastro, and R. Walter,
A&A ,501–507 (2007), arXiv:astro-ph/0611343 .12. M. G. Revnivtsev, S. Y. Sazonov, M. R. Gilfanov, and R. A. Sunyaev,
Astronomy Letters , 587–593 (2003), arXiv:astro-ph/0303274 .13. E. Kuulkers, “An absorbed view of a new class of INTEGRAL sources,” in Interacting Binaries: Accretion,Evolution, and Outcomes , edited by L. Burderi, L. A. Antonelli, F. D’Antona, T. di Salvo, G. L. Israel, L. Piersanti,A. Tornambè, & O. Straniero, 2005, vol. 797 of
American Institute of Physics Conference Series , pp. 402–409, arXiv:astro-ph/0504625 .14. L. Barragán, J. Wilms, K. Pottschmidt, M. A. Nowak, I. Kreykenbohm, R. Walter, and J. A. Tomsick,
A&A , 1275–1278(2009), .15. P. Filliatre, and S. Chaty,
ApJ , 469–484 (2004), arXiv:astro-ph/0408407 .16. F. Rahoui, S. Chaty, P.-O. Lagage, and E. Pantin,
A&A , 801–813 (2008), .17. S. Chaty, F. Rahoui, C. Foellmi, J. A. Tomsick, J. Rodriguez, and R. Walter,
ArXiv e-prints 0802.1774 (2008), .18. H. J. G. L. M. Lamers, F. Zickgraf, D. de Winter, L. Houziaux, and J. Zorec,
A&A , 117–128 (1998).19. F. Zickgraf, B. Wolf, O. Stahl, C. Leitherer, and G. Klare,
A&A , 421–430 (1985).20. J. S. Clark, I. A. Steele, R. P. Fender, and M. J. Coe,
A&A , 888–896 (1999).21. S. Chaty, and P. Filliatre,
Chinese Journal of Astronomy and Astrophysics Supplement , 104–109 (2005).22. L. J. Pellizza, S. Chaty, and I. Negueruela, A&A , 653–658 (2006), arXiv:astro-ph/0605559 .23. N. Masetti, M. L. Pretorius, E. Palazzi, L. Bassani, A. Bazzano, A. J. Bird, P. A. Charles, A. J. Dean, A. Malizia,P. Nkundabakura, J. B. Stephen, and P. Ubertini,
A&A , 1139–1149 (2006), astro-ph/0512399 .24. I. Negueruela, D. M. Smith, T. E. Harrison, and J. M. Torrejón,
ApJ , 982–986 (2006), arXiv:astro-ph/0510675 .25. E. Nespoli, J. Fabregat, and R. E. Mennickent,
A&A , 911–917 (2008), .26. J. J. M. in’t Zand,
A&A , L1–L4 (2005), arXiv:astro-ph/0508240 .27. R. González-Riestra, T. Oosterbroek, E. Kuulkers, A. Orr, and A. N. Parmar,
A&A , 589–594 (2004), astro-ph/0402293 .28. E. Bozzo, L. Stella, C. Ferrigno, A. Giunta, M. Falanga, S. Campana, G. Israel, and J. C. Leyder,
ArXiv e-prints (2010), .29. L. Sidoli, P. Esposito, and L. Ducci,
ArXiv e-prints (2010), .30. D. J. Mullan,
ApJ , 303–312 (1984).1. L. Ducci, L. Sidoli, and A. Paizis,
MNRAS pp. 1262–+ (2010).32. J. Rodriguez, J. A. Tomsick, A. Bodaghee, J. Zurita Heras, S. Chaty, A. Paizis, and S. Corbel,
A&A , 889–894 (2009), .33. L. Sidoli, P. Romano, V. Mangano, A. Pellizzoni, J. A. Kennea, G. Cusumano, S. Vercellone, A. Paizis, D. N. Burrows, andN. Gehrels,
ApJ , 1230–1235 (2008),
ArXiv:0805.1808 .34. P. Romano, V. La Parola, S. Vercellone, G. Cusumano, L. Sidoli, H. A. Krimm, C. Pagani, P. Esposito, E. A. Hoversten, J. A.Kennea, K. L. Page, D. N. Burrows, and N. Gehrels,
ArXiv e-prints (2010), .35. L. Sidoli, A. Paizis, and S. Mereghetti,
A&A , L9–L12 (2006), astro-ph/0603081 .36. D. Götz, M. Falanga, F. Senziani, A. De Luca, S. Schanne, and A. von Kienlin,
ApJL , L101–L104 (2007), arXiv:astro-ph/0612437 .37. P. Romano, L. Sidoli, V. Mangano, S. Vercellone, J. A. Kennea, G. Cusumano, H. A. Krimm, D. N. Burrows, and N. Gehrels,
ApJL , L137–L140 (2008), arXiv:0805.2089 .38. L. Sidoli, P. Romano, V. Mangano, G. Cusumano, S. Vercellone, J. A. Kennea, A. Paizis, H. A. Krimm, D. N. Burrows, andN. Gehrels,
ApJ , 120–127 (2009), .39. L. Sidoli, P. Romano, P. Esposito, V. L. Parola, J. A. Kennea, H. A. Krimm, M. M. Chester, A. Bazzano, D. N. Burrows, andN. Gehrels,
MNRAS , 258–262 (2009), .40. P. Romano, L. Sidoli, G. Cusumano, P. A. Evans, L. Ducci, H. A. Krimm, S. Vercellone, K. L. Page, and et al.,
MNRAS ,45–51 (2009), .41. L. Sidoli, P. Romano, L. Ducci, A. Paizis, G. Cusumano, V. Mangano, H. A. Krimm, S. Vercellone, D. N. Burrows, J. A.Kennea, and N. Gehrels,
MNRAS , 1528–1538 (2009), .42. P. Romano, L. Sidoli, V. Mangano, S. Mereghetti, and G. Cusumano,
A&A , L5–L8 (2007), arXiv:0704.0543 .43. L. Sidoli, P. Romano, V. Mangano, G. Cusumano, S. Vercellone, J. A. Kennea, A. Paizis, H. A. Krimm, D. N. Burrows, andN. Gehrels,
ApJ , 120–127 (2009), .44. R. A. Rampy, D. M. Smith, and I. Negueruela,
ApJ , 243–249 (2009), .45. L. Sidoli, P. Romano, S. Mereghetti, A. Paizis, S. Vercellone, V. Mangano, and D. Götz,
A&A , 1307–1315 (2007), arXiv:0710.1175 .46. P. Romano, L. Sidoli, G. Cusumano, S. Vercellone, V. Mangano, and H. A. Krimm,
ApJ , 2068–2074 (2009), .47. J. Swank, D. Smith, and C. Markwardt,
Astron. Tel. , 1–+ (2007).48. V. La Parola, G. Cusumano, P. Romano, A. Segreto, S. Vercellone, and G. Chincarini,
MNRAS , L66–L70 (2010), .49. D. J. Clark, A. B. Hill, A. J. Bird, V. A. McBride, S. Scaringi, and A. J. Dean,
MNRAS , L113–L117 (2009), .50. A. Lutovinov, M. Revnivtsev, M. Gilfanov, P. Shtykovskiy, S. Molkov, and R. Sunyaev,
A&A , 821–829 (2005), arXiv:astro-ph/0411550 .51. C. Jain, B. Paul, and A. Dutta,
MNRAS , L11–L15 (2009), .52. S. P. Drave, D. J. Clark, A. J. Bird, V. A. McBride, A. B. Hill, V. Sguera, S. Scaringi, and A. Bazzano,
MNRAS pp. 1257–+(2010), .53. D. J. Clark, V. Sguera, A. J. Bird, V. A. McBride, A. B. Hill, S. Scaringi, S. Drave, A. Bazzano, and A. J. Dean,
MNRAS ,L75–L79 (2010), .54. J. A. Zurita Heras, and S. Chaty,
A&A , L1–L4 (2009), .55. A. J. Bird, A. Bazzano, A. B. Hill, V. A. McBride, V. Sguera, S. E. Shaw, and H. J. Watkins,
MNRAS , L11–L15 (2009), .56. A. Bamba, J. Yokogawa, M. Ueno, K. Koyama, and S. Yamauchi,
PASJ , 1179–1183 (2001), arXiv:astro-ph/0110423 .57. A. M. Levine, and R. Corbet, The Astronomer’s Telegram , 1–+ (2006).58. V. Sguera, A. B. Hill, A. J. Bird, A. J. Dean, A. Bazzano, P. Ubertini, N. Masetti, R. Landi, A. Malizia, D. J. Clark, andM. Molina,
A&A , 249–257 (2007), arXiv:astro-ph/0702477 .59. L. Ducci, L. Sidoli, S. Mereghetti, A. Paizis, and P. Romano,
MNRAS , 2152–2165 (2009), .60. Q. Liu, H. Li, and J. Yan,
Science in China G: Physics and Astronomy , 130–134 (2010).61. I. Negueruela, J. M. Torrejón, P. Reig, M. Ribó, and D. M. Smith , 252–256 (2008).62. L. Sidoli, Advances in Space Research , 1464–1470 (2009).63. R. Walter, and J. Zurita Heras, A&A , 335–340 (2007), arXiv:0710.2542 .64. S. A. Grebenev, and R. A. Sunyaev,
Astronomy Letters , 149–158 (2007).65. E. Bozzo, M. Falanga, and L. Stella, ApJ683