Presentation on theme: "Ultraluminous X-ray sources: a mystery for modern X-ray astronomy Tim Roberts."— Presentation transcript:
Ultraluminous X-ray sources: a mystery for modern X-ray astronomy Tim Roberts
Thursday 14th June 2007 Tim Roberts - ULXs2 0.3 – 1 keV ≡ 12 – 40 Å “soft” 1 – 2 keV ≡ 6 – 12 Å 2 – 8 keV ≡ 1.5 – 6 Å “hard” M83 Chandra ACIS-S – true X-ray colours Galaxies in the X-ray regime Point sources – X-ray binaries, supernova remnants Hot phase of the interstellar medium Active nucleus M83 ESO/VLT image M83 - “Southern pinwheel galaxy” ESO VLT
Thursday 14th June 2007 Tim Roberts - ULXs3 X-ray source luminosities Log X-ray luminosity (erg s -1 ) GRBs Stars CVs SNRs NSs & BHs ULXs AGNs QSOs Galactic populations Supermassive black holes
Thursday 14th June 2007 Tim Roberts - ULXs4 The missing link? Only have evidence for stellar-mass ( 10 6 M ) black holes Is there a missing link between them?
Thursday 14th June 2007 Tim Roberts - ULXs5 The Eddington limit Maximum theoretical luminosity for a spherically-accreting object Balance between gravitational pull inwards and radiative push outwards Directly proportional to mass of accretor L Edd ~ 1.3 × (M/M ) erg s -1 ULXs too big for stellar-mass black holes; can’t all be displaced AGN A new, intermediate-mass class of black holes?
Thursday 14th June 2007 Tim Roberts - ULXs6 How do we test this? Need the best instruments Chandra (NASA)XMM-Newton (ESA) Unrivalled X-ray imaging (0.5 arcsecond resolution) High photon collecting area (0.4 m 1 keV) Workhorse instruments for both - CCD imaging spectrometers
Thursday 14th June 2007 Tim Roberts - ULXs7 Black Hole accretion states Photon cm -2 s -1 keV Energy (keV) High (thermal- dominated) ~ 1 – 2 keV disc + PL tail Low/hard Hard PL (Γ ~ 1.5 – 2) dominant, disc absent or truncated, radio jet emission. Least luminous. Very high (steep power-law) Soft PL (Γ > 2.5) plus some hot disc emission. Most luminous. Energy spectra from McClintock & Remillard (2006) Key point - accreting black hole X-ray spectra can be empirically modelled as the combination of an accretion disc spectrum and a power-law (corona)
Thursday 14th June 2007 Tim Roberts - ULXs8 XMM-Newton evidence for IMBHs X-ray spectroscopy – cool accretion discs (Miller et al. 2003) NGC 1313 X-1 T in M kT in ~ 0.15 keV c.f. kT in ~ 1 keV for stellar BHs → ~ 1000 M BHs
Thursday 14th June 2007 Tim Roberts - ULXs9 L X – kT in relationship IMBH candidates occupy separate part of parameter space to stellar- mass BHs Strong evidence for IMBHs as new class underlying luminous ULXs From Miller et al. (2004) L X T 4
Thursday 14th June 2007 Tim Roberts - ULXs10 ULXs in starburst galaxies Multiple ULXs (10+) are found in Starburst galaxies Ongoing star formation ULXs are short-lived Requires an infeasibly large underlying population of IMBHs (King 2004) Alternative: are ULXs in Starbursts “ordinary” high-mass X-ray binaries (HMXBs)? From Gao et al. (2003)
Thursday 14th June 2007 Tim Roberts - ULXs11 In support of stellar-mass BHs How to exceed Eddington limit: Relativistic beaming Radiative anisotropy Truly super-Eddington discs Super-Eddington mass transfer rates in HMXBs can fuel ULXs Blue stellar counterparts – high mass companions? GRS – demonstrates super-Eddington does happen in stellar-mass systems Potential X-ray luminosities for accretion onto a 10 M BH from 2 – 17 M secondaries (Rappaport, Podsiadlowski & Pfahl 2005)
Thursday 14th June 2007 Tim Roberts - ULXs12 Dichotomy Highest X-ray luminosities and cool accretion discs point to IMBHs, but… Other evidence stacking up in favour of smaller black holes Which one is the correct interpretation?
Thursday 14th June 2007 Tim Roberts - ULXs13 NGC 1313 X-1 From Miller et al. (2003) kT in ~ 0.15 keV “diskbb” – optically-thick accretion disc power-law – hot, optically-thin corona ULX X-ray spectra revisited Key evidence for IMBHs from “cool disc” in XMM-Newton ULX spectra. 10+ examples But not all ULXs show this spectral form: several have an “inverted” spectrum e.g. NGC 55 ULX (Stobbart et al. 2004), NGC 5204 X-1 (Roberts et al. 2005) Difficult to explain dominant soft power-law physically! kT in ~ 1.16 keV Γ ~ 2.5 M33 X-8 From Foschini et al. (2004)
Thursday 14th June 2007 Tim Roberts - ULXs14 A sample of bright ULXs Look at best archival XMM-Newton data Demonstrate that 2-10 keV spectrum fit by a broken power- law in all of the highest quality data Invalidates IMBH model - hard component is not a simple power-law Stobbart, Roberts & Wilms 2006 Disc Power-law
Thursday 14th June 2007 Tim Roberts - ULXs15 ULX accretion physics Physical accretion disc plus corona model: cool discs, optically-thick coronae ULXs operate differently to common Galactic black hole states, except… “Strong” VHS in XTE J : energetically-coupled corona/ disc (Done & Kubota 2006). Key features are a disc that appears cool as its inner regions are obscured by an energetic, optically- thick corona. From Done & Kubota (2006)
Thursday 14th June 2007 Tim Roberts - ULXs16 Archetypal luminous ULX (L X > erg s -1 ) – top IMBH candidate Deep XMM-Newton obs (110 ks, though > 60% spoiled by bad space weather) Best fit spectrum: cool disc (~ 0.2 keV) plus optically-thick ( ~ 6) corona Holmberg II X-1 Goad, Roberts et al EPIC spectrum of Ho II X-1 Data: pn/MOS1/MOS2 Model: diskpn + comptt
Thursday 14th June 2007 Tim Roberts - ULXs17 X-ray timing measurements – PSDs Courtesy P. Uttley Adapted from Vaughan et al. (2005) Power Spectral Densities for two AGN and Galactic BH Cygnus X-1 Approximate scaling of break frequencies with mass Frequency regime probed by XMM for bright ULXs
Thursday 14th June 2007 Tim Roberts - ULXs18 Ho II X-1: timing Lack of strong variability PSD analysis – compare to classic black hole states Insufficient power for HS May be in L/H or VHS - energy spectrum says latter Similar to “ χ ”-class of GRS in VHS? Band-limited PSD - but don’t see variability, so must be at high-f M BH < 100 M EPIC-pn light-curve of Ho II X-1 (0.3 – 6 keV, 100 s binning)
Thursday 14th June 2007 Tim Roberts - ULXs19 How to progress? X-ray measurements suggest smaller black holes; but no direct mass determination As with Galactic BHs, ultimate test of the compact object mass in ULXs is dynamical studies Mass function f(M) requires measurement of orbital period (P) and velocity amplitude of donor star (K * ), from line features in optical spectrum: f(M) = M X 3 sin 3 i/(M * +M X ) 2 = PK * 3 sin 3 i/2πG But ULX counterparts have m V ~ 22 – 26 (where identified) in crowded fields – not trivial!
Thursday 14th June 2007 Tim Roberts - ULXs20 3 steps to the black hole mass (1) Identify counterpart from HST imaging Time awarded in cycle 14 to complete local ULXs (2) Obtain spectra – suitable features? Gemini spectra awarded for three ULX counterparts (3) Monitoring campaign
Thursday 14th June 2007 Tim Roberts - ULXs21 New HST imaging ACS WFC F606W F330W F435W F606W m V ~ 24m V > 27m V ~ 25
Thursday 14th June 2007 Tim Roberts - ULXs22 New HST imaging (2) m V ~ 22 m V ~ 25 m V ~ 26 Very blue colours - optical emission from accretion discs?
Thursday 14th June 2007 Tim Roberts - ULXs23 Radial velocity studies Gemini-S GMOS spectrum of NGC 1313 X-2 Possible to derive P, K BH from broad He II 4686Å accretion disc line? Done for Galactic BHs (e.g. Soria et al. 1998). Dynamic mass within reach?
Thursday 14th June 2007 Tim Roberts - ULXs24 State of play Break present in 2 – 10 keV spectrum of many ULXs – not expected for sub-Eddington IMBHs Cool disc, optically thick corona - similarities to very high accretion rate Galactic BHs Ho II timing result – mass limit < 100 M New theory - possible to produce BHs of up to 100 M in young (low metallicity) stellar populations Strongly suggests ULXs are larger stellar-mass BHs accreting at around L edd Require dynamical mass measurements to prove this; a work in progress