On the search of the elusive Intermediate Mass Black-Holes
M. D. Caballero-Garcia, S. Fabrika, A. J. Castro-Tirado, M. Bursa, M. Dovciak, A. Castellon, V. Karas
aa r X i v : . [ a s t r o - ph . H E ] M a r To appear in “Fourth Workshop on Robotic Autonomous Observatories (2017)”
RevMexAA(SC)
ON THE SEARCH OF THE “ELUSIVE” INTERMEDIATE MASSBLACK-HOLES
M. D. Caballero-Garc´ıa, S. Fabrika,
A. J. Castro-Tirado,
M. Bursa, M. Dovˇciak, A. Castell´on, andV. Karas RESUMENLos agujeros negros de masa intermedia (IMBH) son una hip´otesis a´un no confirmada con total seguridad(como clase de objetos) a partir de evidencias observacionales directas. Su masa ser´ıa mayor que la de losagujeros negros de tipo estelar y menor que la de los monstruosos agujeros negros supermasivos. Unos objetoscandidatos a rellenar este hueco lo constituyen las fuentes de rayos X ultra-luminosas (ULX) en cuyas emisionesse observan propiedades diferentes al caso de los agujeros negros de masa estelar en sistemas binarios. Sinembargo, desarrollos te´oricos y observacionales recientes conducen a la idea de que estas fuentes son, en sulugar, objetos compactos de masa estelar que acretan en un r´egimen inusual de super-Eddington. Por otrolado, las ondas gravitacionales se han visto como una herramienta ´util para encontrar los IMBH. En esteart´ıculo damos una breve descripci´on sobre el descubrimiento de las ondas gravitacionales y su relaci´on conestos agujeros negros de masa intermedia, hasta el momento esquivos.ABSTRACTUltra-Luminous X-ray sources (ULXs) are accreting black holes for which their X-ray properties have beenseen to be different to the case of stellar-mass black hole binaries. For most of the cases their intrinsic energyspectra are well described by a cold accretion disc (thermal) plus a curved high-energy emission components.The mass of the black hole (BH) derived from the thermal disc component is usually in the range of 100-10 solar masses, which have led to the idea that this can represent strong evidence of the Intermediate Mass BlackHoles (IMBH), proposed to exist by theoretical studies but with no firm detection (as a class) so far. Recenttheoretical and observational developments are leading towards the idea that these sources are instead compactobjects accreting at an unusual super-Eddington regime instead. On the other hand, gravitational waves havebeen seen to be a useful tool for finding (some of these) IMBHs. We give a brief overview about the recentadvent of the discovery of gravitational waves and their relationship with these so far elusive IMBHs. Key Words:
Black hole physics — Accretion, accretion-discs — X-rays: general — Galaxies: Formation — Cosmology:Early Universe — gravitational waves
1. INTRODUCTIONUltra-Luminous X-ray sources (ULXs) are point-like, off-nuclear, extra-galactic sources, with ob-served X-ray luminosities (L X ≥ erg s − ) higherthan the Eddington luminosity for a stellar-massblack-hole (L X ≈ erg s − ). The true nature of Astronomical Institute, Academy of Sciences of the CzechRepublic, Boˇcn´ı II 1401, CZ-141 00 Prague, Czech Republic([email protected]). Special Astrophysical Observatory, Nizhnij Arkhyz369167, Russia. Kazan Federal University, Kremlevskaya 18, Kazan420008, Russia. Instituto de Astrof´ısica de Andaluc´ıa (IAA-CSIC), P.O.Box 03004, E-18080, Granada, Spain. Departamento de Ingenier´ıa de Sistemas y Autom´atica,Escuela de Ingenier´ıa Industrial, Universidad de M´alaga(Unidad Asociada al CSIC), Spain. Facultad de Ciencias, Dpto. de ´Algebra, Geometr´ıa yTopolog´ıa, Universidad de M´alaga. Campus de Teatinos, s/nM´alaga, Spain. these objects is still debated (Feng & Soria, 2011;Fender & Belloni, 2012) as there is still no unam-biguous estimate for the mass of the compact objecthosted in these systems.Most ULXs are thought to be powered bysuper-Eddington accretion onto a stellar mass blackhole which can be accomplished (i) by power-ing strong disc winds (Shakura & Sunyaev 1973;Lipunova et al. 1999), (ii) by advecting the radi-ation along with the flow as in radiation pres-sure dominated disc models like Polish doughnuts(Abramowicz et al. 1978; Jaroszynski et al. 1980)and slim discs (Abramowicz et al. 1988), or (iii)both, advection and outflows (Poutanen et al. 2007;Dotan et al. 2011). Luminosities up to 10 erg / scan therefore still be explained by super-Eddingtonmass accretion rates onto stellar mass black holeswhich can have maximum masses up to ∼
80 M ⊙ .These higher mass black holes can be explained by1 CABALLERO-GARC´IA ET AL.direct collapse of metal poor stars (Belczynski et al.2010). The slim disc incorporates the effects of ad-vection. This means that with rising mass accretionrate an increasing fraction of photons gets trappedin the flow, carried inward, and is partly releasedat smaller radii. A typical slim disc has higherflux at soft photon energies below the spectral peakand above it in comparison to a standard thin disc(Straub et al. 2011, 2014).Assuming an isotropic emission, in order to avoidthe violation of the Eddington limit, ULXs mightbe powered by accretion onto Intermediate MassBlack Holes (IMBHs) with masses in the range 10 − M ⊙ (Colbert & Mushotzky 1999; Greene & Ho2007; Farrell et al. 2009). Recently, some studieshave shown evidence of some ULXs being Black HoleBinaries (BHBs, e.g. M 82 X–2; Kong et al. 2007;Caballero-Garc´ıa et al. 2013b). Later it was shownthat M 82 X–2 is a binary but accreting onto a neu-tron star (Bachetti et al. 2014).In this paper we give an overview of the currentmodels used in the analysis of the X-ray data fromULXs and on the masses derived by using them.Later we justify the use of the slim-disc model forthe proper description of these sources. Finally wediscuss on the existence of IMBH and their detec-tions in the electromagnetic spectra and with gravi-tational waves.1.1 . The X-ray properties from Ultra-LuminousX-ray sources The spectra of ULXs show a power-law spectralshape in the 3-8 keV spectral range, together witha high-energy turn-over at 6-7 keV (Stobbart et al.2006; Gladstone et al. 2009; Caballero-Garc´ıa et al.2010), and a soft excess at lower energies (e.g.Kaaret et al. 2006). This soft excess can be modelledby emission coming from the inner accretion discand is characterized by a low inner disc temperatureof ≈ . soft excess imply a muchsmaller mass for the BH in these sources, based onthe idea that the accretion in the disc is not intrinsi-cally standard, in contrast to the majority of BHBs(e.g. see Kajava & Poutanen 2009).1.2 . The standard accretion disc theory The low inner disc temperatures found for someULXs were initially interpreted as an evidence for thepresence of IMBH (Miller et al. 2004). In the stan-dard disc-black body model (i.e. Multi-Color DiscBlackbody or MCD; Makishima et al. 1986, 2000), which is an approximation of the real standard ac-cretion disc theory, the bolometric luminosity fromthe accretion disc is calculated as: L bol = 4 π ( R in /ζ ) σ ( T in / f c ) (1)Then the mass of the BH derived from the innerdisc temperature from the accretion disc is calcu-lated as:M BH M ⊙ = 20 . η ( 0 . ζ ) − ( 1 . in (keV)f c ) − (2)Here f c ≈ . ζ ≈ . T in occurs at a radius somewhat larger than R in (Kubota et al. (1998) give ζ = 0 . η theEddington ratio (L bol = η L Edd ). Setting a typicalinner disc temperature of inner disc temperature of ≈ . BH = 10 − M ⊙ (taking η = 0 . − . The anomalous regime However, it has been seen that in a few BHBs(e.g. XTE J1550-564) when they reach a high lumi-nosity level, the L-T relationship departs from whatit has been shown above. Then it is called that theyhave entered the anomalous regime (Kubota et al.2001, 2004; Gierli´nski & Done 2004).As shown by Kubota et al. (2001), these sourcesgenerally follow the L disc ∝ T relation. Whenreaching a luminosity of several orders of magnitudehigher there is a small departure from this: withincreasing temperature they seem to be slightly un-derluminous. This is particularly pronounced abovekT max = 0 . disc ∝ T . Kubota et al. (2004)refer to this as the apparently standard regime,and suggest that it is associated with a transitionto a slim disc (Abramowicz et al. 1988). However,Gierli´nski & Done (2004) noted that this shouldonly occur above L Edd (Abramowicz et al. 1988;Shimura & Takahara 1995), so it seems more likelyto represent a subtle change in the colour tempera-ture correction factor.A shift of the microquasar GRO J1655-40 tothe right from the L – T relation for a 7 M ⊙ BH(Orosz & Bailyn 1997) is probably related to a highspin of the BH there, which results in a higher disctemperature compared to the Schwarzschild BH.ELUSIVE” INTERMEDIATE MASS BLACK-HOLES 3 apparently standardanomalous (MCD+power-law) (3 component fit) standard log T in log L disk L d i s k ∼ T i n Fig. 1. Schematic picture showing the obtained spectralregimes on the L disc - T in diagram. Thick solid and dashedlines show the source behavior obtained under the MCDplus power-law fit.The luminosity-temperature relationfor super-critically accreting BHs. From Kubota et al.(2004). . The supercritical regime A recent study of the spectral variability froma sample of ULXs (Kajava & Poutanen 2009), hasshown that the soft excess (i.e. the disc compo-nent fitted from the X-ray spectra of some canonicalULXs) does not follow Eq. 1 but L bol ∝ T − . . Thisin contrast to what is found for many BHBs (Fig. 2)and might indicate that the standard accretion disctheory is not a proper interpretation to the X-rayspectra from ULXs. This implies that the hypothe-sis on which the IMBH idea is relying (i.e. standardaccretion disc theory and the presence of a cold disc)are not valid and it might indicate that these ULXsare not IMBHs as a “class”. In the following wedevelop this idea into more detail.We have seen above that the standard model forsub-critically accreting BHs (Shakura & Sunyaev1973) predicts the relation L ∝ T . Poutanen et al.(2007) develop a model based on super-critical ac-cretion. It is a model taking into account geomet-rical effects and where those related to advectionare not fully taken into account. In spite of itssimplicity it makes the relationship between observ-able amounts (e.g. temperature and luminosity fromthe spectra) and those we want to derive (i.e. themass of the compact object) an easy task. Nev-ertheless, in the following Sec. we will describe afully consistent model based on the slim-disc calcula-tions (Abramowicz et al. 1988; S¸adowski et al. 2011;Bursa et al. 2018, in prep.). Fig. 2. The luminosity-temperature relation for super-critically accreting BHs. From Poutanen et al. (2007).
At super-Eddington accretion rates, three char-acteristic temperatures are identified:the maximal color disc temperature:T c , max = f c T max ≈ . c m − / keV , the color temperature at the spherization radius:T c , sph ≈ . c m − / ˙ m − / keV , the outer photosphere temperature given by:T ph ≈ m − / ˙ m − / keV . (3)where ˙ m = ˙M / ˙M Edd and f c is the correction factorto T max . They are all marked at Fig. 3.A soft, ∼ m o = m − / (1 . f c / T c , sph [keV]) ≈ ⊙ , BH. The observed higher lu-minosities can result from the geometrical beaming.The absolute maximum apparent luminosity usingthis model (taking all the effects of advection andbeaming of the flow) is about 10 ergs − for a 20 M ⊙ BH. 1.5 . The geometry of the disc
A face-on observer would see the emission fromthree separate zones defined by the three character-istic radii (we refer to Poutanen et al. 2007 for moredetails): CABALLERO-GARC´IA ET AL. (cid:1) f observer Tph Rph T (cid:1) R (cid:2) Fig. 3. The model of a supercritical disc with a windfunnel. The figure shows the thin disc (
R > R sp ), thesupercritical disc ( R ≤ R sp ), and the wind funnel con-strained by the radius of the photosphere R ph / sin θ f .From Vinokurov et al. (2013). r < r ph , in , zone A , r ph , in < r < r sph , zone B , r sph < r < r ph , out , zone C . (4)The characteristic disc temperatures can be ob-tained from the Stefan-Boltzmann law (Q rad ( R ) = σ SB T ( R )) .In zone A (i.e. outside from the photosphere),the wind is transparent (i.e. it is momentum-driven)and the radiation escapes unaffected by the outflow.In zone B, the wind is opaque and the energygenerated in the disc is advected by the wind. Theradiation escapes at a radius about twice the energygeneration radius. This does not change the radialdependence of the effective temperature T ∝ r − / (i.e. slim-disc configuration),The outer zone C emits about the Eddington lu-minosity which is produced mostly in the disc atradii r > r sph . The photon diffusion time here issmaller than the dynamical time, thus most of theradiation escapes not far from the radius it is pro-duced. This results in the effective temperature vari-ation close to r − / (i.e. standard accretion regime;Shakura & Sunyaev 1973).1.6 . The need of slim-disc models Typical models to fit the X-ray spectra fromUltra-luminous X-ray sources using XSPEC (e.g.
DISKBB , DISKPN , KERRBB , BHSPEC , GRAD ; Arnaud1996 ) are based on the thin disc model, which isnot accurate for L > . Edd . Indeed, such modelstend to give incorrect values for BH masses and for See the following web page for a list and description ofthe available models in XSPEC:https://heasarc.gsfc.nasa.gov/xanadu/xspec/manual/ accretion rates (luminosities) above the Eddingtonlimit, as shown in this Section.The standard (thin) disc follows the L ∼ T relation (see Eq. 1). Nevertheless, advectionand obscuration effects cause significant deviationsfrom that relation in the super-Eddington regime(Poutanen et al. 2007). These effects are taken intoaccount in the so-called slim-disc models. This effectis strongly inclination dependent (Poutanen et al.2007) and the luminosity can stay around the Ed-dington limit even if the mass accretion rate is muchhigher ( ˙M ≫ ˙M Edd ). These facts have implicationsfor the spectral modelling, e.g. getting lower innerdisc temperatures given a certain (high) luminosity(Fig. 2 and 4). This is in accordance to what has al-ready been observed in the X-ray spectra from sam-ple of representative ULXs (Poutanen et al. 2007).Instead of using the classical models based on thethin disc theory here we present the results obtainedby using the
SLIMULX model . SLIMULX (Bursa et al.2018, in prep.) is an additive model for thermal con-tinuum emission at high accretion rates to be usedwith the X-ray spectral-fitting tool XSPEC. Themodel provides spectral distribution of black-bodyradiation that is supposed to be emitted from thesurface of a slim accretion disc (Abramowicz et al.1988). A brief description and a first application ofthe model can be found in Caballero-Garc´ıa et al.(2017).The major improvement of the
SLIMULX modelover other commonly used disc models is that itincludes three effects that are dominant at highaccretion rates, and are not present in the standardShakura & Sunyaev (1973); Novikov & Thorne(1973) disc model: • Radial advection of heat, which plays a substan-tial role at higher luminosities, is present andsignificantly modifies the flux of radiation emit-ted at a given radius in the inner disc region. • The inner edge for the disc radiation deviatesfrom ISCO at high luminosity and can be con-siderably closer to the black hole due to the ad-vective transport of heat generated by viscousprocesses. • The location of the effective photosphere dif-fers significantly from the equatorial plane withgrowing luminosity, although the relative verti-cal disc thickness is not large (h/r ≤ http://stronggravity.eu/results/models-and-data/ ELUSIVE” INTERMEDIATE MASS BLACK-HOLES 5from the actual disc photosphere to the observerat infinity (S¸adowski et al. 2009).Bursa et al. (2018, in prep.) fitted the simulated(i.e. faked)
SLIMULX spectra with a thin disc model(
DISKBB ) and the mass was obtained from the fits(Fig. 5). At low ˙M, the fit recovers the original mass(i.e. the one given by the slim-disc model), but athigh mass accretion rate ( ˙M ≥ Edd ) the mass givenby the thin disc model is much higher (note thata mass of 10 M ⊙ was assumed in the SLIMULX sim-ulated spectra). This exercise is representative ofwhat is usually done when fitting the X-ray spectrafrom ULXs in the literature. It appears quite a badapproach to estimate the BH parameters when us-ing thin disc models if the disc is strongly radiationpressure dominated.So we should not use of Eqs. 1 and 2 in the radi-ation pressure dominated (i.e. in the high accretionrate) regime.In the
SLIMULX model the luminosity consists ofthe integration of all the radial flux elements (in theso-called corotating frame) multiplied by a factorthat accounts for the proper transformation of thecorotating frame to the coordinate frame so that thegiven value is indeed the total luminosity of the discignoring geometrical GR effects. In this case the rela-tionship between the mass and any measurable prop-erty of the disc is troublesome. Even although themass is a free parameter of the model, its direct re-lationship with the rest of quantities is complicated.It is generally assumed that the black hole mass isof a certain value (e.g. 10 M ⊙ ) when doing explicitcalculations. Although relevant disc properties fol-low a mass scaling law or do not scale with mass,the results can not be directly rescaled and gener-alized to BH systems of an arbitrary mass, becausethe output of radiative transfer calculation would bedifferent and the predicted spectra also. The ap-plication to stellar mass BH systems is fine over anusually assumed mass range for such systems (few tofew tens of solar masses). Note that higher masses(e.g. IMBHs) can be used (and derived) with thismodel as well (see Straub et al. 2014).1.7 . On the masses derived for the compact objectusing slim-disc models It has been suggested that ULXs appear very lu-minous due to a combination of moderately highmass (IMBHs), mild beaming and mild super-Eddington emission and that ULXs are an inho-mogeneous population composed of more than oneclass (Colbert & Mushotzky 1999; Fabbiano et al.2006). . . L u m i n o s i t y [ L / L E dd ] . . . L u m i n o s i t y [ L / L E dd ] . Fig. 4. The luminosity-temperature relation for super-critically accreting BHs. The two panels correspondto two different inclination angles ( i = 0 ◦ , ◦ , at topand bottom, respectively) of the line of sight to theobserver. Blue/red branches correspond to standardNovikov-Thorne (red) and slim (blue) accretion disc,with the labels indicating the accretion rate (in unitsof Eddington accretion rate). On the horizontal axis,the temperature [keV] denotes the best-fit temperaturefrom both models (this temperature corresponds to theone giving the maximum emission flux in the calculatedspectra). From Bursa et al. (2018, in prep.). Initially, they were supposed to be the IMBHsoriginating from low-metallicity Population III stars(Madau & Rees 2001). Nevertheless, they are notspatially distributed throughout galaxies as it wouldbe expected. On the other hand, IMBHs may beproduced in runaway mergers in the cores of youngclusters (Portegies Zwart et al. 2004). In such cases,they usually should stay within their clusters. Nev-ertheless, it has been found (Poutanen et al. 2013)that all brightest X-ray sources in the Antennaegalaxies are located nearby the very young stellarclusters. NGC 5408 X–1 is also located nearby a CABALLERO-GARC´IA ET AL.
Fig. 5. Simulated
SLIMULX spectra are fittedwith a thin disc model (
DISKBB ; Mitsuda et al. 2004;Makishima et al. 1986) and the mass is obtained fromthe fit (in units of M ⊙ ). The horizontal axis is theaccretion rate (in units of ˙M Edd ). From Bursa et al.(2018, in prep.). young stellar association (Gris´e et al. 2012). Thesestudies concluded that these sources were ejected inthe process of formation of stellar clusters in the dy-namical few-body encounters and that the majorityof ULXs are massive X-ray binaries with the pro-genitor masses larger than 50 M ⊙ . Currently, it isthought that only a handful of ULXs could be con-sidered as potential IMBHs (ESO 243-49 HLX-1 be-tween a few others; Farrell et al. 2009; Sutton et al.2012; Mezcua et al. 2017).Currently the most accepted idea is that the ma-jority of ULXs are powered by accretion onto stellar-mass black holes ( <
100 M ⊙ ) at around or in excessof the Eddington limit (e.g. Colbert & Mushotzky1999; Fabrika & Mescheryakov 2001; King et al.2001; Fabbiano et al. 2006; Poutanen et al. 2007;Liu et al. 2013).Due their brightness, most of them are believedto be BHs. However, recently a new class of ULXswas discovered, through the detection of coherentpulsations: Ultra-luminous X-ray pulsars (ULPs).The presence of pulsations unambiguously identi-fies the compact objects as neutron stars, which aretypically less massive than black holes. In ULPsthe neutron star accretes matter from a compan-ion star at inferred rates much higher than previ-ously expected. Currently three of these systemsare known: M82 X-2 (Bachetti et al. 2014), NGC5907 ULX (Israel et al. 2017a), and NGC 7793 P13(Israel et al. 2017b). Furthermore, a clear path for-ward to obtain a full sample of the ULP populationis missing. In this paper we have described all the modelsused to describe the data from ULXs, from assum-ing low accretion rates (and then high BH mass)to highly super-Eddington accretion rates (then lowcompact object mass). Applying the best mod-els based on slim-disc configuration and includingadvection in the high accretion rate favour eitherlow masses for the compact object (M <
30 M ⊙ ,e.g. LMC X-3 and NGC5408 X-1 Straub et al.2011; Caballero-Garc´ıa et al. 2017) or high masses(M ∼ > M ⊙ , ESO 243-49 HLX-1 Straub et al. 2014).Therefore BH masses in the range of M = 100 − M ⊙ are hardly to find in the ULX population.Even though initially this population of ULXs wasinitially thought to contain a significant part of theseIMBHs. Indeed, to our knowledge, there is no unambigu-ous detection of the electromagnetic counterpart of aBH with a mass in the range of
M = 80 − M ⊙ .1.8 . X-ray Timing properties of ULXs Black Hole masses scale with the break fre-quency of their Power Density Spectrum (PDS;McHardy et al. 2006; K¨ording et al. 2007). This re-lation holds over six orders of magnitude in mass, i.e.from Black Hole Binaries (BHBs) to Super-MassiveBlack Holes (SMBHs).The PDS and the energy spectra of NGC 5408X–1 (Fig. 6) and M 82 X–1 (for instance) are verysimilar to that of BHBs in the Steep Power-law state(i.e. one of the so-called intermediate states). Never-theless, the characteristic time-scales within the PDSare lower by a factor of a ≈
100 and the X-ray lu-minosity is higher by a factor of a few ≈
10, whencompared to BHBs. This gives BH mass estimatesfor these ULXs of the order of M BH ∼ > − M ⊙ ,from their X-ray timing properties only, as explainedin the following.As described in the previous Sections, determin-ing the mass from the BH in ULXs has been thegoal of several studies. For example, in NGC 5408X–1 there is still no consensus on whether it isan IMBH or a stellar-mass BH. Previous esti-mates from the timing properties (Strohmayer et al.2009; Dheeraj & Strohmayer 2012) indicate a massof M BH ∼ > ⊙ , thus an IMBH, but others(Middleton et al. 2011) indicate a much smaller massof M BH ≤
100 M ⊙ , thus a stellar-mass BH . In It has to be noted here that in low metallicity envi-ronments BHs with masses up to 80 −
130 M ⊙ can still beformed through direct stellar-collapse (Zampieri et al. 2009;Belczynski et al. 2010) and this is why we are referring tothem as stellar-mass BHs.
ELUSIVE” INTERMEDIATE MASS BLACK-HOLES 7
Fig. 6. Average Power Density Spectrum of NGC 5408X–1. From Strohmayer et al. 2009. the first case considered the accretion rate is sub-Eddington, whilst the latter case indicates (near or)super-Eddington accretion.If the Quasi-Periodic Oscillation (QPO) detectedin M 82 X–1 is a fundamental High FrequencyQPO (HFQPO), then it appears at much lower fre-quency ( ≈
50 mHz) than those observed in BHBs(35–450 Hz), by three orders of magnitude. Scalinglinearly with the mass of a stellar-mass BH (10 M ⊙ )if these QPOs have the same origin, the frequencieswe have found lead to a mass of M BH ≈ (10 − ) M ⊙ (taking into account the whole range of possiblevalues of the spin of the BH) for the mass of theBH in M 82 X–1. Other mass estimates, thistime based on spectral-timing scaling relationshipsfrom systems of different mass (Titarchuk & Fiorito2004), have provided a mass estimate of ≈ M ⊙ (Fiorito & Titarchuk 2004). Nevertheless, cautionis required in order to apply relationships basedsolely on the mass of the BH systems. As alreadypointed out by several authors (McHardy et al. 2006for AGNs; Soria 2007 and Casella et al. 2008 forULXs) the accretion rate is an important parame-ter that, together with the mass of the BH, shouldbe considered as the main drivers in these scal-ing relationships. This effect could lead to a muchsmaller mass for the BH (compact object). In ad-dition, it remains to date unclear how timing prop-erties and BH masses are related, and what are theproperties of the accretion states in ULXs comparedto those of Galactic BHBs (Caballero-Garc´ıa et al.2013a; Atapin et al. 2018, in prep.). 2. GRAVITATIONAL WAVES AS A NEWWINDOW TO THE UNIVERSEAs reported by Abbott et al. (2016) the firstgravitational-wave (GW) transient was identified indata recorded by the Advanced Laser InterferometerGravitational-wave Observatory (LIGO) detectorson 2015 September 14. The event, initially desig-nated G184098 and later given the name GW150914,is described in detail elsewhere. By prior arrange-ment, preliminary estimates of the time, significance,and sky location of the event were shared with 63teams of observers covering radio, optical, near-infrared, X-ray, and gamma-ray wavelengths withground- and space-based facilities. As this eventturned out to be a binary black hole merger, thereis little expectation of a detectable electromagnetic(EM) signature . GW150914 is consistent with theinspiral and merger of two BHs of masses 36 +5 − and29 ± ⊙ , respectively, resulting in the formationof a final BH of mass 62 ± ⊙ at a distance of410 +160 − Mpc.As reported by Abbott et al. (2017) on 2017 Au-gust 17 a binary neutron star coalescence candi-date (later designated GW170817) with merger time12:41:04 UTC was observed through gravitationalwaves by the Advanced LIGO and Advanced Virgodetectors. The Fermi Gamma-ray Burst Monitorindependently detected a gamma-ray burst (GRB170817A) with a time delay of ≈ . at a distance of 40 +8 − Mpc and with com-ponent masses consistent with neutron stars. Thecomponent masses were later measured to be in therange 0.86 to 2.26 M ⊙ . An extensive observing cam-paign was launched across the electromagnetic spec-trum leading to the discovery of a bright optical tran-sient (SSS17a, now with the IAU identification ofAT 2017gfo) in NGC 4993 less than 11 hours af-ter the merger. The follow-up observations supportthe hypothesis that GW170817 was produced by themerger of two neutron stars in NGC 4993 followedby a short gamma-ray burst (GRB 170817A) and akilonova/macronova powered by the radioactive de-cay of r-process nuclei synthesized in the ejecta.In Fig. 7 there is an interactive graphic featuringall the BH detected by LIGO (including GW170608),and the recently announced Neutron Stars, plus allthe other compact objects known from electromag-netic measurements.2.1 . Detection of the missing IMBHs Black holes, the ultra-compact remnants of verymassive stars, are prime candidates for emitting de- CABALLERO-GARC´IA ET AL.
Fig. 7. Interactive graphic featuring all of our BlackHoles (including GW170608), our recently announced NeutronStars,and all the other compact objects known from electromagnetic measurements. From LIGO collaboration. tectable gravitational waves in this way. While cur-rent detection work focuses on commonly-expectedbinary systems of black holes with masses a fewtimes that of the Sun, it is possible that thereare detectable populations of black holes whichhave masses larger than this, i.e. hundreds orthousands times the mass of the Sun. These aredubbed IMBHs, as their masses lie in the currently-unobserved range between solar mass black holes andthe supermassive black holes in the center of galax-ies. The gravitational-wave interferometer networkof LIGO and Virgo are able to measure ripples fromthe final moments of such intermediate mass binaryblack holes merging together.Candidate IMBHs, nevertheless are still rela-tively rare (see our discussion above). Gravitational-wave detection of intermediate mass black hole bi-nary mergers confirm their existence and provide in-formation on the abundance of such systems in theUniverse, as well as precisely measuring the compo-nent masses. The presence or absence of these signals will shed light on star formation in the early universe,provide important information about the dynamicsand structures of globular clusters, and give cluesabout the formation of supermassive black holes.Black holes are just that – black. They are thecorpses of dead stars, so massive and compact thatnot even light can escape them. They literally can-not be seen, and before LIGO came online, their ex-istence could only be inferred by their gravitationaleffect on their neighbors or because of light (often X-rays) emitted by hot gas falling into the black hole.But an isolated black hole is invisible. It interactsvia gravity and, even then, it only emits gravitationalradiation when it is moving. So detectors like LIGOor Virgo are the only way to see them. They areessentially black hole telescopes.With even just a few observations of gravita-tional waves, the LIGO measurements are perplex-ing. Prior to 2016, it was thought that there weretwo classes of black holes: stellar-mass black holes,with masses no more than about 10 times that of ourELUSIVE” INTERMEDIATE MASS BLACK-HOLES 9Sun, and massive, monstrous black holes at the cen-ter of galaxies with masses in the range of hundredsof thousands to billions of solar masses.Black holes with masses in the range of ∼ > − M ⊙ were unexpected. And yet, that’s just whatLIGO (and now LIGO plus Virgo) have observed.It has been shown that BHs in the mass range of30 −
130 M ⊙ can be formed through BH-BH mergersin low-metallicity environments (see recent simula-tions from Marchant et al. 2016, 2017). They couldjustify the discovery of the recent IMBHs discoveredby LIGO so far ( ≤
130 M ⊙ ; Fig. 7). But observation-ally, IMBHs with masses in the range 100 − M ⊙ are extremely rare to find. They could form (partof) the unseen dark matter proposed to exist inthe galactic haloes by cosmological studies (see e.g.Mediavilla et al. 2017, who tested for this possibil-ity). 3. DISCUSSION AND CONCLUSIONSUltra-Luminous X-ray sources (ULXs) are ac-creting black holes for which their X-ray propertieshave been seen to be different to the case of stellar-mass black hole binaries. For most of the cases theirintrinsic energy spectra are well described by a coldaccretion disc (thermal) plus a curved high-energyemission components. The mass of the black hole(BH) derived from the thermal disc component isusually in the range of 100 − M ⊙ , which have ledto the idea that this might represent strong evidenceof the Intermediate Mass Black Holes (IMBH), pro-posed to exist by theoretical studies but with no firmdetection (as a class) so far. But the mass estimationdepends on the accretion rate.We have discussed the masses obtained by us-ing different models considering different accretionrate regimes. Masses of the IMBH in the range ofM = 80 − M ⊙ are hardly to obtain. Also, recenttheoretical and observational developments are lead-ing towards the idea that ULXs are instead stellar-mass compact objects accreting at an unusual super-Eddington regime instead, therefore favouring lowermass estimates for the compact objects.On the other hand, gravitational waves have beenseen to be a useful tool for finding (some of) theseIMBHs. We have given a brief overview aboutthe recent advent of the discovery of gravitationalwaves and their relationship with these so far elusiveIMBHs.MCG, MB and MD acknowledge support pro-vided by the European Seventh Frame-work Pro-gramme (FP7/2007-2013) under grant agreement n ◦ Abbott, B. P., Abbott, R., Abbott, T. D., et al., 2016,ApJ, 826, 13Abbott, B. P., Abbott, R., Abbott, T. D., et al., 2017,ApJ, 848, 12Abramowicz, M., Jaroszynski, M., & Sikora, M. 1978,A&A, 63, 221Abramowicz, M. A., Czerny, B., Lasota, J. P., &Szuszkiewicz, E. 1988, ApJ, 332, 646Arnaud, K. A., 1996, ASPC, 101, 17Atapin, K., Fabrika, S. & Caballero-Garcia, M. D., 2018,in prep.Bachetti, M. et al., 2014, Natur, 514, 202Belczynski, K., Bulik, T., Fryer, C. L., et al. 2010, ApJ,714, 1217Bursa, M. et al., 2018, in prep.Caballero-Garc´ıa, M. D. & Fabian, A. C., 2010, MNRAS,402, 2559Caballero-Garc´ıa, M. D., Belloni, T. M. & Wolter, A.,2013a, MNRAS, 435, 2665Caballero-Garc´ıa, M. D., Belloni, T. & Zampieri, L.,2013b, MNRAS, 436, 3262Caballero-Garcia, M. D., Bursa, M., Dovˇciak, M. et al.,2017, CoSka, 47, 84Casella, P., Ponti, G., Patruno, A. et al., 2008, MNRAS,387, 1707Colbert, E. J. M. & Mushotzky, R. F., 1999, ApJ, 519,89Dheeraj, P. R. & Strohmayer, T. E., 2012, ApJ, 753, 139Dotan, C. & Shaviv, N. J. 2011, MNRAS, 413, 1623Ebisawa, K., Mitsuda, K. & Hanawa, T., 1991, ApJ, 367,213Fabbiano, G., 2006, ARA&A, 44, 323Fabrika, S. & Mescheryakov, A., 2001, IAUS, 205, 268Farrell, S. A., Webb, N. A., Barret, D. et al., 2009, Natur,460, 73Fiorito, R. & Titarchuk, L., 2004, ApJ, 614, 113Gierli´nski, M. & Done, C., 2004, MNRAS, 347, 885Gladstone, J. C., Roberts, T. P. & Done, C., 2009, MN-RAS, 397, 1836Greene, J. E. & Ho, L. C., 2007, ApJ, 670, 92Gris´e, F., Kaaret, P., Corbel, S. et al., 2012, ApJ, 745,123Hutchings, J. B., Crampton, D., & Cowley, A. P., 1983,ApJ, 275, 43Israel, G. L., Belfiore, A., Stella, L., et al., 2017, Sci, 355,817Israel, G. L., Papitto, A., Esposito, P., et al., 2017, MN-RAS, 466, 48Jaroszynski, M., Abramowicz, M. A., & Paczynski, B.1980, Acta Astron., 30, 10 CABALLERO-GARC´IA ET AL.