Powerful jets from accreting black holes: evidence from the optical and infrared
aa r X i v : . [ a s t r o - ph . H E ] J a n In: Black Holes and Galaxy Formation ISBN 0000000000c (cid:13) P OWE RFUL JE T S FROM ACCRE TING B L ACK HOL E S : E VI DENCE FROM T HE OPT I CAL AND I NFRARED . David M. Russell ∗ , University of Amsterdam, Netherlands
Rob P. Fender
University of Southampton, United Kingdom
Keywords: accretion, accretion disks, outflows, jets, radiation mechanisms, X-ray binaries,optical, infrared.
Abstract
A common consequence of accretion onto black holes is the formation of powerful, relativisticjets that escape the system. In the case of supermassive black holes at the centres of galaxies thishas been known for decades, but for stellar-mass black holes residing within galaxies like our own,it has taken recent advances to arrive at this conclusion. Here, a review is given of the evidencethat supports the existence of jets from accreting stellar-mass black holes, from observations madeat optical and infrared wavelengths. In particular it is found that on occasion, jets can dominatethe emission of these systems at these wavelengths. In addition, the interactions between the jetsand the surrounding matter produce optical and infrared emission on large scales via thermal and ∗ E-mail address: [email protected]
David M. Russell, Rob P. Fender non-thermal processes. The evidence, implications and applications in the context of jet physicsare discussed. It is shown that many properties of the jets can be constrained from these studies,including the total kinetic power they contain. The main conclusion is that like the supermassiveblack holes, the jet kinetic power of accreting stellar-mass black holes is sometimes comparable totheir bolometric radiative luminosity. Future studies can test ubiquities in jet properties betweenobjects, and attempt to unify the properties of jets from all observable accreting black holes, i.e. ofall masses.
1. Introduction
There is currently just one way in which we can study black holes phenomenologically inthe lab – investigating the radiation generated by their gravitational influence on matter.Here, the lab is outer space, and the detectors are telescopes sensitive to various differentwavelengths of radiation (radio, infrared, optical, ultraviolet, X-ray, γ -ray). Black holes(BHs) with masses between a few times the mass of the Sun (solar masses; M ⊙ ) and abillion M ⊙ attract local matter, which can convert its gravitational potential energy intoheat and light as it approaches the BH. By definition, BHs are dark and do not directlyproduce radiation we can detect. However it is because they are massive and compact thatthe gravitational forces close to BHs are the most extreme in the universe; as a result theattracted matter is extremely hot and emits high-energy radiation.The gravitational force in newtons (ignoring general relativity; this is just for purposesof demonstration) on a particle of mass m kg due to a massive object of mass M kg is F = GM mr , where G is the gravitational constant G = 6 . × − N m kg − and r is the distance tothe centre of mass of the massive object in metres. The radius of the event horizon (beyondwhich no radiation or matter can escape) of a (non-spinning) BH is R = 2 GMc , where c is the velocity of light, × km s − . From the above equations, the maximumgravitational force from a non-spinning BH (i.e. at the event horizon) is F max ∝ MR ∝ M Close to the event horizon of a M = 10 M ⊙ black hole, the gravitational force is ∼ times that near the surface of the Sun. It is the most compact objects in the universe – theblack holes and neutron stars (a neutron star is a collapsed star with a density of the order ∼ kg m − ) that exert the largest gravitational force possible on a particle (this holdsif general relativistic effects are taken into account). The gravitational potential energyof nearby matter turns into heat and radiation; the extreme gravitational conditions nearBHs make them the brightest objects of high energy emission (X-rays and γ -rays) in theuniverse. We note that in the next decade or so it is likely that the signature of blackvidence of jets produced by black holes from optical and infrared observations. 3holes will also be detected via gravitational waves (via e.g. black hole binary mergers; e.g.Flanagan & Hughes, 1998). This will then become the second method in which we canstudy them observationally.Black hole candidates in our Galaxy are thought to exist as a result of a massive starcollapsing at the end of its life in a supernova explosion. Nuclear fusion in the core ofthe star ceases because its fuel has been depleted, causing the loss of hydrostatic equi-librium and the core collapses under its own gravitational attraction. Most Galactic BHcandidates may be isolated, and accreting matter from the interstellar medium (ISM) atlow, undetectable rates. The BHs that we do detect are in X-ray binaries (XBs) – bi-nary systems in which the BH candidate (of mass M ∼ – M ⊙ or more) and a (rel-atively normal) star are orbiting a common centre of mass. If the surface of the star liesclose to the point between the two objects in which the gravitational force from bothare equal (the L1 point), the outer layers leak as a stream towards the BH. The net an-gular momentum of the gas causes it to form a disc around the BH as the matter spi-rals inwards. The BH is hence being fed, via this accretion disc, by a semi-regular sup-ply of gas from the companion star. As the matter moves inwards its temperature in-creases; the highest energy radiation is emitted from the hottest gas close to the BH. Avisual illustration of an accreting black hole candidate X-ray binary (BHXB) is given here: It is somewhat counterintuitive that it is in these extreme gravitational fields close to BHsthat the highest velocity outflows in the universe are formed. One can use the simple argu-ment however that the matter must lose a large amount of its potential energy and angularmomentum if it is to reach the event horizon, and outflows are one way of removing thisproblem by channelling some of the energy and angular momentum away from the BH;some matter then never reaches the event horizon. Jets of matter and energy travellingclose to the speed of light were first discovered originating from the vicinity of supermas-sive ( M ∼ – M ⊙ ) BHs in active galactic nuclei (AGN) at the centres of galaxies.These are the most massive BHs known in the universe, and their powerful, relativisticjets, and the interactions of those jets with local matter, have been studied for decades (e.g.Baade & Minkowski, 1954; Fanaroff & Riley, 1974). Hotspots indicate where the jets im-pact with the intracluster gas, and lobes of synchrotron-emitting plasma are inflated whichflow back from the hotspot, both of which are often detected at radio frequencies. AGN andtheir jets are now realised to play an important role in the evolution of matter in the universe,affecting to some extent the rate of galaxy growth in the early universe (e.g. Efstathiou,1992; Churazov et al., 2002; McNamara & Nulsen, 2007; Schawinski et al., 2009).There are at least three fundamental questions that remain unanswered regarding AGNjets: • What is the process of particle acceleration? • What is their composition? • How much influence do they have on shaping large-scale structure in the universe? David M. Russell, Rob P. FenderDifficulties arise in answering these questions because various models predict similarobservational properties. For example, the composition of the jets and radio lobes, whichare not charged, could be either baryonic (electrons and protons) or leptonic (electrons andpositrons), both of which result in synchrotron emission which is observed. Similarly, theenergy required to accelerate the jets to relativistic velocities could be tapped either byextraction of the spin energy of the BH, or from the accretion energy in the accretion disc(Blandford & Znajek, 1977; Blandford & Payne, 1982; Meier, 2001); these have not yetbeen distinguished by observations. The power contained in AGN jets is known, at least tosome degree. It is possible to use (successfully in some cases) the jet hotspots and lobesto infer the power input into them, i.e. the energy contained in the jets averaged over time(Burbidge, 1959). It is only in recent years that it has become apparent that accretion onto stellar-mass BH can-didates (and neutron stars) in X-ray binaries commonly produce relativistic, synchrotron-emitting jets in analogy with those of AGN (e.g. Mirabel et al., 1992; Falcke et al., 2004;Fender, 2006; Migliari & Fender, 2006). The name ‘microquasar’ was coined for X-raybinaries with jets due to this analogy. The jets here are found to be associated with specificX-ray regimes (Fender et al., 2004). When the X-ray spectrum is hard, a steady jet ex-ists, the radio luminosity of which correlates with the X-ray luminosity (Gallo et al., 2003,2006). This correlation extrapolates successfully to the radio and X-ray luminosities ofAGN when introducing a mass term which is consistent with that predicted by scalings inblack hole physics (Merloni et al., 2003; Falcke et al., 2004), demonstrating that a single jetformation process alone likely governs the physics of all BHs. When BHXBs make tran-sitions to softer X-ray states, bright discrete ejecta are observed and often resolved at largedistances from the BH (e.g. Tingay et al., 1995; Gallo et al., 2004). The core jet emissionis then quenched, and only returns when the X-ray spectrum again hardens (Fender et al.,2004). XBs vary in luminosity (and hence mass accretion rate) by orders of magnitudeon timescales accessible to us – typically over several weeks or months. By comparison,AGN take millions of years to perform these ‘outburst’ cycles. A recent step forward inour understanding was made when it was proposed that the different X-ray regimes and jetbehaviour of BHXBs can be directly linked to the different classes of AGN (K ¨ording et al.,2006b). Short term variability probes the inner regions of the accretion flow close to theBH, in the vicinity where the jets are formed. These timescales scale with the mass of theBH; for BHXBs it is milliseconds to minutes, whereas for AGN it can be from hours toyears. Minute-to-hour long observations can hence probe these timescales in BHXBs, butdedicated long-term campaigns are required for the same study of AGN.It is uncertain how powerful XB jets are. Few sites of interaction between the jets ofXBs and the ISM have as yet been identified but there are a growing number of examples(see Sections 2.2 and 2.3). These interaction sites can be used to infer the time-averagedpower of XB jets (Kaiser et al., 2004; Gallo et al., 2005). Estimates of the jet power basedon the core jet luminosity have large errors because both the total spectrum of the jet and itsradiative efficiency are poorly constrained (however see K ¨ording et al., 2006a, for a recentadvancement of the latter). Estimates of the jet power are dominated by the highest energyvidence of jets produced by black holes from optical and infrared observations. 5electron distributions in the jet, which generate the highest energy photons, but isolation ofthe jet spectrum can be hampered by other emitting components such as the accretion discat these higher energies (see Section 2.1). Measuring as accurately as possible the jet power( P J ) in different sources and at different luminosities is crucial in understanding the processof jet formation and the overall physics of accretion and the matter and energy BHs inputinto the ISM.It may be that stellar-mass black holes in BHXBs hold the key to understanding AGNand their jets. If the BHXBs are indeed simply scaled-down versions of AGN, with similarphysical structure and behaviour, the three unanswered questions above regarding AGN jetsmay be answered via studies of BHXBs because they vary on accessible timescales, andtheir spectra are not complicated by light from the host galaxies or affected (as much) byorientation and obscuration. In addition, it is possible to model the composition of BHXBjets using information from both the compact jet and jet–ISM impact sites (Heinz, 2006). Many physical components and mechanisms are responsible for optical/infrared (OIR)emission from BHXBs and neutron star X-ray binaries. Fig. 1 is a schematic which il-lustrates the different components of the accretion inflow and outflow (jets), their approxi-mate physical sizes and optical colours, for a BHXB with a low-mass companion star. Thecolours correspond to approximate intrinsic optical ( ∼ V-band) spectral indices. The dis-tances from the centre of the black hole are in logarithmic scale. The dimensions of the com-panion star, accretion stream and accretion disc are based on those quoted in O’Brien et al.(2002) for the BHXB GRO J1655–40. The companion star of GRO J1655–40 has an ap-proximately flat spectral index at V-band ( α = 0 ; Migliari et al., 2007) so it appears yellow.The accretion disc, which produces a multi-temperature black body spectrum is hotter atsmaller radii, and the optical spectral index steepens in this case from α ∼ +0 . at itsouter edge to probably α ∼ +2 . in the inner regions (based on observed values duringoutbursts but these change with disc temperature; e.g. Hynes, 2005). Although hotter, theemitting region is smaller in the inner regions of the disc, and the bulk of the optical emis-sion from the disc originates in the outer regions. The inner edge of the accretion disc is ata radius ∼ − times that of the L1 radius (O’Brien et al., 2002, L1 is the point at whichthe gravitational force from the BH and the star is equal; material from the star is no longergravitationally bound to the star if it crosses this point towards the BH). The hot, opticallythin flow (which may also exist above and below the inner part of the disc) ends at the in-nermost stable circular orbit (ISCO), which is at three times the radius of the event horizon.The radius of the event horizon here is 21 km (we assume a non-spinning black hole ofmass M = 7 M ⊙ ).The outflow illustrated in Fig. 1 comprises a compact, continuously replenished jet pro-ducing synchrotron emission, internal shocks further down the jet, deceleration of jet knotsas they impact the ISM, and a radiative bow shock energized by the jet. All components arebased on observed, resolved phenomena and are expected to produce OIR emission. Weprovide a comprehensive review and discussion of the OIR signatures of jets in Section 2.The compact, conical jet, which is accelerated to relativistic velocities close to the BH,produces synchrotron radiation from electron distributions with energies decreasing with David M. Russell, Rob P. FenderFigure 1. A schematic demonstrating the approximate dimensions and optical colours ofthe components of a low-mass BHXB (see text for details).vidence of jets produced by black holes from optical and infrared observations. 7distance from the BH. The optical depth also decreases as the ejected material expands asit flows down the jet. The result is a series of overlapping, self-absorbed synchrotron spec-tra which form an approximately flat ( α ∼ ) radio-to-infrared spectrum, where the syn-chrotron spectrum from each electron energy distribution peaks at a frequency which is ap-proximately inversely proportional to the distance from the jet origin (Blandford & Konigl,1979; Falcke & Biermann, 1996). Stirling et al. (2001) resolved the compact steady jet ofCygnus X–1 at 8.4 GHz. At this radio frequency its extent was 15 milli-arcseconds (mas) ≈ × km adopting a distance to the BHXB of 2.1 kpc (Massey et al., 1995). Assumingthe highest energy electron distributions in the inner regions of the jet produce a synchrotronspectrum that peaks in the near-infrared (NIR; at ν ∼ . Hz, as is implied by some ob-servations; see Section 2.1) the summed spectrum will break to an optically thin one, with α ∼ − . blueward of this ‘jet break’. The corresponding radius of the innermost regionof the synchrotron-emitting compact jet is then ∼ × km. This is also consistent withfast timing observations in the optical, in which variability on timescales corresponding toemitting regions as small as ∼ × km have been shown to most likely originate inthe inner regions of the jet (e.g. Kanbach et al., 2001; Hynes et al., 2003b, see Section 2.1).The opening angle of the radio jet of Cyg X–1 was < ◦ (Stirling et al., 2001); here weassume ◦ . This is consistent with measurements from other BHXBs (Miller-Jones et al.,2006) but much larger than implied by some models (e.g. Kaiser, 2006). The outer regionof the jet that produces the 8.4 GHz radio emission would then be wider than the diameterof the accretion disc (assuming the opening angle is constant along the jet). The compact jetcannot currently be resolved at OIR frequencies. The inner regions closest to the BH ( < km away say) dominate the OIR emission from the compact jet; at a distance of 1 kpc (theclosest known BHXBs are at this distance), this corresponds to an angular size of the orderof micro-arcseconds, which is not resolvable with current detection limits (although sophis-ticated interferometric methods may achieve this in the future; Markoff, 2008). The radioemission from the compact jet originates in a region much larger than this, but this spectralcomponent is orders of magnitude fainter at OIR frequencies and cannot be detected.BHXB jets have however been resolved at radio and X-ray frequencies, and can be seento move at relativistic speeds, at distances larger than ∼ km from the BH (see Section2.2). The emission process is thought to be optically thin synchrotron which peaks at lowfrequencies, and could either be energized by shocks between discrete ejecta travelling atdifferent velocities, or by impacts with the ISM, which decelerate the ejecta. In Fig. 1discrete shocks are shown at distances − km (which have been seen in the radio inSco X–1, actually a neutron star XB; Fomalont et al., 2001) and impacts with the ISM areshown at ∼ km (here observed from the BHXB XTE J1550–564; Corbel et al., 2002).There is just one tentative claim so far of this resolved synchrotron emission detected atOIR frequencies: from GRS 1915+105 (Sams et al., 1996). For the above case of XTEJ1550–564 the authors point out that the synchrotron spectrum from radio to X-ray is justbelow the detection limits of optical data taken with the Very Large Telescope.It has been shown that the impact of the jets of BHXBs with the ISM inflate lobes ofradio plasma, like those seen in FR II radio galaxies powered by the jets of AGN (e.g.Kaiser et al., 2004). Ahead of these lobes there is a radiative shock wave, which producesboth thermal continuum and line emission from the recombination of ionized atoms. Opticalline emission and thermal radio emission from a bow shock powered by a BHXB jet has David M. Russell, Rob P. Fenderrecently been detected – at a distance ∼ × km ( ∼ parsecs) from Cygnus X–1,with a diameter approximately half that distance. This is the furthest and largest structureillustrated in Fig. 1 (see Section 2.3 for discussion of these structures).
2. OIR identifications of jets from black hole X-ray binaries
Traditionally, jets have been identified in BHXBs from the radio regime. This is firstly be-cause they are sometimes resolved, which is a very strong argument for a jet, and secondlybecause no other known emitting components can account for the radio fluxes and spectra(see Fender, 2006, for a review). There is now strong empirical evidence for the spectrumof the compact jet to extend to, and be detected at, OIR frequencies and higher. Althoughthese detections are complicated by other spectral components (which is usually not thecase at radio frequencies), many methods can be adopted to successfully disentangle thecomponents and isolate the jet emission. It is in some cases easier to detect BHXB jets inthe OIR with current facilities. Signatures of OIR jet emission include its spectrum, timingproperties and polarimetry properties, all of which differ from other OIR–emitting compo-nents such as the accretion disc. In Section 2.1 we review these signatures and discuss theobservations of OIR jets in the literature. We stress that although our review focusses on thejet emission, the process that dominates the optical (and in some cases the NIR too) emis-sion of most transient BHXBs is X-ray reprocessing on the surface of the accretion disc.For BHXBs with high-mass stars, and for BHXBs at low accretion rates in quiescence, thecompanion star usually dominates the OIR light.As was discussed in Section 1.3, large-scale, resolved
OIR emission can be produceddirectly and indirectly by the jets of BHXBs. In Sections 2.2 and 2.3 we summarise theknown resolved OIR jets and jet–ISM interaction sites. In Section 3 we briefly discuss howthese observations (OIR compact jets, resolved jets and jet–ISM interactions) can be usedto infer the jet properties, and their implications for jets and accretion onto supermassiveBHs in AGN. A summary is provided in Section 4, and we comment on future work thatwill help to answer the open fundamental questions regarding jets and accretion.
Models of synchrotron emission from compact jets of BHXBs can reproduce the approxi-mately flat ( α ∼ ), optically thick radio spectrum of AGN and BHXBs, which breaks toone which is optically thin (the aforementioned ‘jet break’) at some higher frequency (e.g.Blandford & Konigl, 1979; Markoff et al., 2001, 2005; Kaiser, 2006; Jamil et al., 2008,see Fig. 2 for an example spectrum). For BHXBs the jet break probably lies within theOIR regime (see below). The jet break frequency ( ν b ) depends on the distance of the jetlaunch region from the BH, which is related to the BH mass, and may also have a weakdependence on mass accretion rate and therefore luminosity (Heinz & Sunyaev, 2003;Nowak et al., 2005). In addition to the spectrum, the OIR-emitting jet has been identifiedvia polarimetry and correlations with other wavebands over short timescales ( ∼ secondsand less; fast timing) and long timescales ( ∼ weeks to years; ‘slow timing’, i.e. over manyorders of magnitude of mass accretion rate).vidence of jets produced by black holes from optical and infrared observations. 9Figure 2. An example of a jet model that successfully reproduces the radio, OIR and X-raySED of a BHXB (GX 339–4 in this case); from Markoff et al. (2003). The NIR is modelledas optically thin synchrotron emission from the jet. A synchrotron OIR spectrum:
A flat ( α ∼ ) radio-to-OIR spectrum implied from radio and OIR data alone is consis-tent with (but not direct evidence for) a significant jet contribution to the OIR, since radioemission in BHXBs originates in the jet. Fender (2001) pointed out that this is the casein the spectral energy distributions (SEDs) of the BHXBs V404 Cyg and GRS 1915+105.In fact, Han & Hjellming (1992) showed that the radio and optical light curves of the de-cay of V404 Cyg follow similar shapes (three separate power-law decay slopes) while re-maining proportional in flux ( F ν, radio ∝ F ν, optical ). Similar flat radio–OIR SEDs are seenfor GS 1354–64 (BW Cir) and GRO J0422+32 in the hard state (i.e. when the jet is on;Brocksopp et al., 2001, 2004) and for GRS 1915+105 when its flickering radio jet is on(e.g. Fender & Pooley, 2000; Klein-Wolt et al., 2002). A global correlation between theOIR and X-ray luminosities of 15 BHXBs in the hard state (Russell et al., 2006, see below)implies that all BHXBs approximately follow the correlation L ν, radio ≈ L ν, optical . Never-theless, this correlation alone is not a diagnostic of a jet origin to the OIR luminosity but isconsistent with this interpretation.More convincing is a measure of the intrinsic OIR spectral index. Corbel & Fender(2002) found two components of OIR emission in an SED of GX 339–4 – one red ( α < )and one blue ( α > ). The red component dominates the NIR and is consistent withoptically thin synchrotron emission from the jet, for which we expect α ∼ − . ± . ; thisvalue depends on the electron energy distribution. A ∼ flat radio–NIR spectrum was also0 David M. Russell, Rob P. Fenderevident in the SED. The red NIR component was seen additionally in separate outbursts ofGX 339–4 (Homan et al., 2005), and Motch et al. (1985) measured the power-law optical—NIR spectral index to be α ∼ − . , fairly typical of optically thin synchrotron emission.The SEDs of XTE J1118+480 (Hynes et al., 2000; Fender et al., 2001; Chaty et al., 2003a;Hynes et al., 2003b) and 4U 1543–47 (Buxton & Bailyn, 2004; Kalemci et al., 2005) havesimilar red spectral indices in the hard state, and a NIR-excess was apparent above thespectrum of the accretion disc in XTE J1859+226 (Hynes et al., 2002). SEDs of LMC X–3also reveal an anomalous NIR-excess (Pineault, 1984; Treves et al., 1988), which may bethe first evidence for a jet from a BHXB in the Magellanic Clouds (see also Section 2.3 foran other). OIR SEDs of V404 Cyg, GS 1354–64, GS 2000+25, XTE J1550–564 and GROJ1655–40 are sometimes red ( α < ; Brocksopp et al., 2004; Russell et al., 2006) whichis indicative of either a cold disc (possible for sources at low luminosity) or a synchrotronjet. It is possible to use OIR colour-magnitude diagrams to successfully separate thermaldisc from non-thermal jet emission since the irradiated disc produces a predictable relationbetween OIR colour and magnitude as the temperature of the outer disc (and luminosity)changes, whereas the OIR colour of optically thin synchrotron emission is not expected tochange with disc temperature or luminosity (Maitra & Bailyn, 2008; Russell et al., 2008).Colours redder than expected for a disc (deviations from a colour–magnitude relation for aheated black body) were seen in five BHXBs; the redder colours of some were argued to bedue to the jets.One claim of the jet break itself observed in the NIR SED of GX 339–4(Corbel & Fender, 2002) is based on three data points; follow-up near simultaneous pho-tometry is required to confirm this (the source is known to have large amplitude variabilityon short timescales). However the jet break may have been detected in GRO J0422+32(a self-absorbed synchrotron model fits the double-power-law optical–UV spectrum betterthan an accretion disc model; Hynes & Haswell, 1999; Shrader et al., 1994) and V404 Cyg(Brocksopp et al., 2004; Russell et al., 2006) but these are speculative.A further source of evidence for OIR jets is successful modelling of broadband spectra,incorporating self-consistent physical constituents and emission processes. This approachhas led to further confirmation of OIR jets dominating the SEDs of XTE J1118+480(Markoff et al., 2001) and GX 339–4 (Markoff et al., 2003, an example model from thispaper is shown here in Fig. 2). For GRO J1655–40, and in quiescence for A0620–00, XTEJ1118+480 and V404 Cyg, the jet dominates only at lower frequencies – in the mid-infrared– but makes a low-level contribution to the OIR (Migliari et al., 2007; Gallo et al., 2007). OIR–X-ray correlation in the hard state:
Homan et al. (2005) present a correlation between the quasi-simultaneous NIR andX-ray fluxes of GX 339–4 which span two orders of magnitude in F X during its hard state.The correlation, F NIR ∝ F . is synonymous to the hard state radio–X-ray correlation F radio ∝ F . (Gallo et al., 2006), suggesting again that the radio and NIR emission originsmay be identical. In Russell et al. (2006, 2007b) it was found that the X-ray and OIRhard state data (B-band to K-band) of 15 BHXBs are consistent with a global correlation, L OIR ∝ L . which holds over eight orders of magnitude in L X . Individual BHXBsdo not necessarily display this correlation within their own data sets, but all the data lieclose to the global correlation. However, a correlation with a similar ( L OIR ∝ L . ) slopevidence of jets produced by black holes from optical and infrared observations. 11is expected (van Paradijs & McClintock, 1994) if the OIR is reprocessed emission fromX-rays illuminating the accretion disc. From the observed BHXB correlation alone, thetwo cannot be confidently separated (Russell et al., 2006). An optical/infrared flux drop when the radio jet is quenched:
A strong indication of an OIR jet is a dramatic change of flux during an X-ray statetransition. The compact radio jet is quenched in the soft state, dropping below detectionlimits (Gallo et al., 2003). Homan et al. (2005) saw a drop of a factor 18 (3 magnitudes)in the NIR flux of GX 339–4 when the source entered the soft state. In that time theX-ray flux also decreased but only by a factor ∼ . There are a number of examples ofOIR rises/drops seen during transitions in/out of the hard state; in Fig. 3 we plot the lightcurves and amplitudes of this drop as a function of frequency. It appears that the changein flux is in general stronger in the NIR than in the optical, likely due to the accretion discdominating moreso at higher frequencies because its spectrum is bluer than that of the jet.In at least one case (A0620–00) there is an X-ray rise of similar amplitude to the OIR rise– for this source X-ray reprocessing cannot be ruled out. Moreover, Russell et al. (2006,2007b) found that the NIR (but not the optical) luminosity of all six BHXBs in their study(the first three sources in the above Fig. 3, plus GRO J1655–40, XTE J1720–318 and XTEJ1859+226) is weaker in the soft state than in the hard state at the same X-ray luminosity,by a factor ∼ (this cannot be accounted for by the change in X-ray spectrum). TheNIR soft state data of these six BHXBs lie below the global hard state OIR–X-raycorrelation. In a related result, a NIR-to-UV jet spectrum in GRS 1915+105 is impliedby a drop in the NIR emission line Br- γ when the jet continuum drops, which suggeststhe line is radiatively pumped by high energy photons from the jet (Eikenberry et al.,1998a). However, this seems not to be the case for the H α line in BHXBs since no apparentchange is seen in the equivalent width of this line after state transitions (Fender et al., 2009). Fast timing signatures:
Rapid (on timescales down to milliseconds), high-amplitude optical variability has beenreported from BHXBs since the 1980s (Motch et al., 1982). Fabian et al. (1982) argued thatclouds of plasma with surface brightness temperatures ∼ × K emitting cyclotron radi-ation can account for the rapid optical flares of GX 339–4. Motch et al. (1983) constructeda cross-correlation function (CCF) of simultaneous fast optical and X-ray timing data of GX339–4 during its hard state and discovered a complex behaviour in which (a) a weak positiveoptical response to X-ray flickering peaked after tens of seconds, and (b) an optical–X-rayanti-correlation exists whereby optical variations precede by a few seconds those in X-ray.This ‘precognition dip’ was also later found in the CCF of XTE J1118+480 in the hardstate, in addition to a strong positive optical lag (Kanbach et al., 2001). The optical vari-ability is less smeared than the X-ray, so the optical origin cannot be X-ray reprocessing onthe disc. A further recent study of GX 339–4 in the hard state (Gandhi et al., 2008) showedan optical–X-ray CCF with a weak precognition dip and a large but narrow positive opticalresponse.Two separate lines of reasoning lead to the conclusion that the origin of the fast opticalflickering in XTE J1118+480 (and presumably GX 339–4) is the inner regions of the jet.Hynes et al. (2003b) studied the NIR, optical, UV and X-ray variability properties and by2 David M. Russell, Rob P. Fender
10 11 12 13 14 15 16 17 18 19-80 -60 -40 -20 0 20 40 60 80 100 M a gn it ud e Time after hard-to-soft state change (days)Hard Soft HardXTE J1550-564 (H-band; 2000)4U 1543-47 (J-band; 2002)GX 339-4 (H-band; 2002)GX 339-4 (I-band; 2002)GX 339-4 (V-band; 2002) D r op / r i s e du r i ng s t a t e t r a n s iti on ( m a gn it ud e s ) log [ ν (Hz)] BVRIJHK UXTE J1550-564 hard -> soft 2000XTE J1550-564 soft -> hard 20004U 1543-47 hard -> soft 20024U 1543-47 soft -> hard 2002GX 339-4 hard -> soft 1981GX 339-4 hard -> soft 2002A 0620-00 hard -> soft 1976 Figure 3. Changes in OIR magnitudes of BHXBs over X-ray state transitions. Top panel:OIR light curves of three BHXBs. Bottom panel: Amplitude of the drop/rise during tran-sition, as a function of frequency. The bandpasses (filters) U (ultraviolet) to K (NIR)are indicated. In all cases, the change of flux is a rise when in transition to the hardstate and a drop when the source leaves the hard state (into a softer state). The data arefrom Motch et al. (1985), Kuulkers (1998), Jain et al. (2001), Buxton & Bailyn (2004) andHoman et al. (2005). Errors are typically less than ∼ . mag.vidence of jets produced by black holes from optical and infrared observations. 13isolating the variable component, found the variability to have a single power-law SED fromNIR to X-ray, with a spectral index α = − . , implying an optically thin synchrotron ori-gin. The variability is also more lagged at lower frequencies, which is consistent with syn-chrotron bubbles which become optically thin at lower frequencies as they expand, and notconsistent with either X-ray reprocessing or advection-dominated accretion flow models.Hynes et al. (2006) found a similar spectral index, α = − . for the variable component ofthe NIR emission of XTE J1118+480 during a separate outburst. An alternative approach(Malzac et al., 2004) used a time-dependent model to approximately reproduce the CCF(including the precognition dip) and power spectrum of XTE J1118+480. The model en-visages a jet–disc coupling in which accretion (magnetic) energy can be channelled intoeither the jet power or the X-ray ‘corona’ in a correlated way, the jet producing the opticalemission.Further clues come from optical variability studies of other BHXBs in outburst andquiescence. Brightness temperature arguments were used to rule out X-ray reprocessingin V4641 Sgr and GRO J0422+32 during hard state outbursts, favouring a non-thermaloptical variability origin (Bartolini et al., 1994; Uemura et al., 2002). It was demonstratedthat short ( < sec) flares from V4641 Sgr are likely synchrotron in nature, and longer( > sec) flares are thermal (from their SEDs; Uemura et al., 2004a,b). An optical–X-ray CCF of GRO J1655–40 during a soft state (Hynes et al., 1998) displays a positiveoptical lag on light-travel timescales, with no precognition dip. This favours a reprocessingresponse on the disc surface and further supports the idea that the jet is responsible forthe precognition dip. However, the CCF of SWIFT J1753.5–0127 (a BH or neutron starprimary is debated in this source) displays a deep, wide precognition dip and weak positiveresponse, at a time in which the contribution of the jet to the OIR SED appears to be minimal(Durant et al., 2008, 2009).Rather different variability results come from the BHXB GRS 1915+105 in outburst.This source is unique with respect to its X-ray timing and radio jets – it does not remain inthe canonical hard or soft X-ray states for weeks to months like other BHXB transients, butinstead displays complex states at high accretion rates accompanied by correlated (some-times resolved) radio jet ejections. Discrete, quasi-periodic NIR, mm and radio flares re-peat intermittently on timescales of 30–45 minutes and are correlated with more rapid X-ray variations (e.g. Fender et al., 1997; Ueda et al., 2002; Rothstein et al., 2005). The NIRflares cannot be reprocessed X-rays, and instead are shown to be the precursors of the mmand radio flares – likely single jet ejection events which propagate away from the BHXB,peaking at lower frequencies at larger distances (Mirabel et al., 1998; Fender & Pooley,1998; Eikenberry et al., 1998b; Fender & Pooley, 2000). A similar process may be evidentin SS 433 (this system is most likely a BHXB but may be a neutron star XB). Goranskii et al.(1998) report a variable ‘red’ component in the optical SED – these optical flares lead radioflares from the SS 433 jets by several hours in time. Chakrabarti et al. (2005) also claim aNIR–radio correlation on longer timescales – a broad dip in the NIR and radio light curvesis consistent with the NIR leading the radio by two days.Numerous fast timing studies of BHXBs have been performed during periods ofquiescence, but only one (to our knowledge) makes use of simultaneous X-ray coverage,because most systems are too faint for current X-ray telescopes at these luminosities.Hynes et al. (2004) present a CCF between X-ray flux and H α optical emission line flux,4 David M. Russell, Rob P. Fenderwhich shows a positive correlation consistent with zero lag. This result combined with theobservation of a double peaked H α line profile implies line variability in the disc poweredby X-ray irradiation. Fast optical continuum variations and flaring were reported in quies-cence for at least six BHXBs (A0620–00, GU Mus, MM Vel, V404 Cyg, GRO J0422+32,GS 2000+25; Hynes et al., 2003a; Zurita et al., 2003; Shahbaz et al., 2003, 2004). Therapidity implies (to the authors) localised flares which may occur when magnetic loopsreconnect on the disc, although the power density spectrum, when measured, can resemblethat of an X-ray power density spectrum of a BHXB in the hard state, in which the X-rayvariability comes from the inner regions of the accretion flow. In some cases, the spectralindex of the optical flares themselves have been measured and are quite steep: α = − . for A0620–00 and α = − . for V404 Cyg, which are consistent with an optically thingas at a temperature (8–14) × K. According to the quiescent broadband SEDs of thesetwo BHXBs (Gallo et al., 2007), the jet likely contributes < % (most likely ∼ %) ofthe optical emission at these low luminosities, so the jets are unlikely to be responsible forthe flares. However, the flares are short-lived and could be missed in the broadband SEDs,which are based on mean fluxes over longer timescales. If the flares originate in the jets,the optical–radio jet spectrum would have to be inverted; α ∼ > +0 . for the jets to be asfaint as observed in the radio regime. Polarization signatures:
Fabian et al. (1982) predicted that if the rapid optical flaring of GX 339–4 (Motch et al.,1982) has a cyclotron origin, the flares should be polarized. Optically thin synchrotronemission can be highly (up to 70%) polarized if the local magnetic field is ordered (e.g.Bj¨ornsson & Blumenthal, 1982). This polarization signature is commonly detected (atlevels up to tens of per cent) in the radio regime from discrete, optically thin jet ejections(e.g. Hannikainen et al., 2000) and finally confirmed observationally in 2008 at OIRfrequencies. One BHXB (GRO J1655–40) and two neutron star XBs (Sco X–1 andCyg X–2) have linearly polarized NIR emission which is stronger at lower frequencies(Shahbaz et al., 2008; Russell & Fender, 2008). The polarization levels, ∼ –7% suggest atwisted magnetic field geometry, but the polarization is probably masked by non-polarizedlight from other spectral components. The polarization of Sco X–1 is variable ontimescales of minutes, which could be explained by a variable flux level of the jet orby a changing magnetic field geometry. The polarization angles imply a magnetic fieldorientation which differs from source to source and may change in time for the same source(Shahbaz et al., 2008; Russell & Fender, 2008). Two further BHXBs, XTE J1118+480and XTE J1550–564, have low levels of polarization ( ∼ . % in the optical and 1–2%in the NIR, respectively) which require follow-up observations to confirm their origin(Schultz et al., 2004; Dubus & Chaty, 2006). Baryonic jets:
The remarkable precessing jets of SS 433 (of which the compact object may be a BHor a neutron star) have signatures at radio, optical and X-ray frequencies. As well as beingresolved in the radio and X-ray regimes (e.g. Schillemat et al., 2004; Migliari et al., 2005),broad optical emission lines with moving velocity shifts up to 40,000 km s − are shown tobe from discrete jets moving ballistically at the same velocity as the radio jets: 26% of thevidence of jets produced by black holes from optical and infrared observations. 15speed of light, at distances < km from their launch region (e.g. Murdin et al., 1980;Margon et al., 1984; Vermeulen et al., 1993). This is the only evidence for baryonic rela-tivistic jets from an X-ray binary. Ionized atoms in the jets are recombining as they cool, andX-ray spectra reveal numerous lines from highly ionized atoms in the jets (Marshall et al.,2002) – their chemical composition therefore includes heavy elements and not pure pairplasma.SS 433 is surrounded by a large spherical supernova remnant W50. The compact objectin SS 433 is thought to be the remnant of the supernova explosion which resulted in W50.The supernova remnant is 10,000 years old. It could be that the recent supernova enrichedthe photosphere of the companion star with heavy elements, which is now being accretedtowards the compact object and swept up into the jets. If this is so, we may not expect to seeheavy elements in the jets of other XBs because they are much older, and the photospheresof the companion stars are no longer enriched with heavy elements from the supernovaexplosion.A further BHXB that has high-velocity, baryonic outflows is V4641 Sgr. Here, the H α optical and Br γ NIR emission lines are very broad with blue wings and variable, indicatingan outflow with velocity ∼ % of the speed of light (Chaty et al., 2003b). This outflow ishowever not the same as the highly relativistic (much faster) jet in this source. The compact jet, as described in previous Sections, is seen (in the case of Cyg X–1;Stirling et al., 2001) to extend ∼ − km away from the BH. This was measured at oneparticular radio frequency – at lower frequencies the compact jet may appear longer still,but will also perhaps be fainter. The compact jet makes only a low contribution to the OIRregime at these large distances from the BH. However bright, discrete, moving ejecta areseen at radio and sometimes X-ray frequencies at distances between ∼ and ∼ kmfrom the BH, which also emit OIR radiation (Fig. 1). This emission is not part of the over-lapping synchrotron spectra that build up the compact jet, but comes from isolated ‘knots’downstream. Two processes could be responsible: a fast jet plasma cloud catching up withand colliding with a slower ejection, or a plasma cloud that plows into a denser region ofthe ISM. Both processes result in synchrotron emission. Although such large-scale movingjets have never been detected at OIR frequencies, we discuss here that OIR jet emissiondoes exist on these scales, and could be detected with future observations.The large-scale, moving jets of XTE J1550–564 were detected at radio and X-rayfrequencies and were shown to be decelerating, likely due to interaction with the ISM(Corbel et al., 2002). The radio spectrum and X-ray flux of the same ejection are consistentwith a single power-law of spectral index α = − . , typical of optically thin synchrotronemission. Deep optical observations were performed with the 8-m Very Large Telescopebut no optical counterpart of the ejections were found. The derived optical upper limits wereonly just greater than the interpolation of the synchrotron power-law, suggesting that witha slightly higher signal-to-noise ratio (a deeper image or a larger telescope) the large-scaleejection would be detected. Deep NIR observations may have been more successful; thissource suffers a moderate level of foreground extinction, which affects the optical regimemoreso than the NIR.6 David M. Russell, Rob P. FenderTable 1. The predicted NIR K-band magnitudes of known extended radio jets of BHXBs,assuming α = − . from radio to NIR. θ is the angular distance in arc-seconds of theradio knot from the core. The 2.2 µ m values of F ν are predicted intrinsic flux densities andK mags are the corresponding apparent K-band magnitudes after interstellar extinction istaken into account. For foreground extinction we adopt A K = 0 . A V ; estimated valuesof A V are from Jonker & Nelemans (2004), Chapuis & Corbel (2004) and Eikenberry et al.(2001). BHXB ——————– Radio jet knot ——————– – 2.2 µ m flux – A V F ν Freq. θ Reference F ν K mag(mJy) (GHz) ( µ Jy)
Moving jets :XTE J1550–564 3.6 4.8 29” Corbel et al. (2002) 2.7 21.6 5.0GRO J1655–40 120 2.3 0.73” Tingay et al. (1995) 55 18.1 3.7GX 339–4 0.47 4.8 6.9” Gallo et al. (2004) 0.36 23.6 3.9V4641 Sgr 10 4.9 0.6” Hjellming et al. (2000) 7.7 19.9 1.0GRS 1915+105 2.1 5.0 0.23” Miller-Jones et al. (2005) 1.6 23.8 20
Large-scale lobes :1E 1740.7–2942 2 1.5 45” Mirabel et al. (1992) 0.68 27.1 40GRS 1758–258 0.52 5.0 23” Rodr´ıguez et al. (1992) 0.41 23.9 7.9
One claim of a resolved NIR jet from a BHXB does exist in the literature – that of GRS1915+105 (Sams et al., 1996). The jet here is not resolved itself, but appears as a residual tothe south-west of the point-spread-function of GRS 1915+105 in the K-band. The implieddistance of the jet from the BH is ∼ × km. This could be the NIR counterpart ofa discrete ejection with a broadband radio-to-X-ray SED like the one of XTE J1550–564above.Extended radio jets (either moving knots or apparently stationary ‘lobes’) are now de-tected from ∼ BHXBs. Here, we investigate whether these structures may be observableat OIR frequencies, adopting the assumption that the radio emission is optically thin (ob-servations indicate this is the case), with α = − . extrapolated to OIR. We calculate theapparent K-band (2.2 µ m) magnitude of the knots given the known foreground extinctionto each source. The results are given in Table 1. We see that some of the predicted jetmagnitudes are bright enough to be detected with current large telescopes. The brightest, atK ∼ mag is however only 0.7 arc-seconds from the core, so high-resolution imaging isrequired, but possible.In addition, the angular size of the resolved X-ray jets of SS 433 are a few arcseconds(Migliari et al., 2005), so it may be possible to detect the OIR counterparts to these jets withdeep observations, possibly in line emission such as shifted H α .SS 433 actually also provides the best evidence yet from the optical regime of acircumbinary disc surrounding an XB – a separate, non-outflowing large-scale structurearound the binary (Blundell et al., 2008). This result is implied from two stationary H α emission lines; one redshifted and one blueshifted, that suggest an orbiting velocity of ∼ km s − .vidence of jets produced by black holes from optical and infrared observations. 17 At the ‘front’ of the jet there should, in analogy with AGN jets, be a hot spot where the jetslams into dense matter and is decelerated, and a bow shock shell of compressed, shockedgas that surrounds synchrotron-emitting radio plasma (e.g. Kaiser et al., 2004, an alterna-tive is a gradual deceleration of the jet over large distances by a low-density ISM). Theshell emits bremsstrahlung radiation and recombination lines. The first confirmed bowshock shell associated with the jet of a BHXB was recently discovered as a bremsstrahlung-emitting radio ring and a recombination line-emitting optical ring, ∼
10 parsecs away fromCyg X–1 in the direction of its known radio jet (Gallo et al., 2005; Russell et al., 2007a).This (see Fig. 4) is the largest known optically-emitting structure powered by the jet of aBHXB.Cooke et al. (2007) showed that optical emission line ratios of the nebula close to theBHXB LMC X–1, which was previously thought to be photoionized by the X-rays of theBHXB (Pakull & Angebault, 1986) revealed the nebula is in some regions shock-excited,and argued this shock excitation arises from the jets of LMC X–1. This is then the secondBHXB jet-powered bow shock nebula.Filaments of optical line emission are also seen where the two jets of SS 433 havebroken through the W50 supernova remnant that surround the XB (e.g. Boumis et al., 2007).These filaments are shock-excited, compressed gas, energized by the jets of SS 433. Ratherthan a bow shock morphology, the filaments are instead localised regions of shock waveswhich may contain reflected or secondary shocks. Similar filaments are observed inside thelarge bow shock nebula of Cyg X–1. It is interesting to note that the emission line ratios ofthe Cyg X–1 nebula, LMC X–1 nebula and SS 433 filaments all imply shock waves withvelocities around ∼ km s − .It was argued (Heinz, 2002) that a high density region of the ISM is needed for a jet-powered bow shock to form and appear bright. XBs with high-mass companions typicallyoccur in denser regions of the ISM than those with low-mass companions. The fact thatthe three above jet–ISM interaction sites are associated with the jets of XBs with high-masscompanions supports this hypothesis.In addition, it was recently demonstrated that most XBs with low-mass companionsare probably travelling through the ISM at velocities greater than a required limit for jet-powered bow shock nebulae to form and remain powered by the jet (Heinz et al., 2008). Inthese cases, synchrotron plasma generated by the jets may form trails behind the XB, if theXB velocity exceeds that of the expanding plasma bubble. These expanding trails will havea shock front where they interact with the ISM. This shocked gas may have been recentlydetected for a low-mass neutron star XB SAX J1712.6–3739 (Wiersema et al., 2009). Here,two approximately straight stripes of H α emission appear to originate from the location ofthe XB, analogous to the trails envisaged by Heinz et al. (2008). Follow-up imaging andspectroscopy will constrain the properties and nature of this nebulosity.Nebulae discovered near the XBs GRO J1655–40, LS 5039 and GRS 1009–45 are can-didate jet-powered bow shocks and require follow-up observations to confirm these associ-ations (Russell, 2008; Miller-Jones et al., 2008).8 David M. Russell, Rob P. FenderFigure 4. Image of the Cyg X–1 ring nebula in optical line emission (Russell et al., 2007a).The bow shock is powered by the jet of Cyg X–1. H α emission is shown in red, [O III] ingreen and continuum emission in blue. The nebula appears yellow since the [O III]/H α fluxratio is ∼ , indicating a bow shock velocity > km s − . The edge of an H II region canbe seen red in the lower left of the image. The compact radio jet of Cyg X–1 (lower right)is ∼ times smaller in apparent length (Stirling et al., 2001, see Fig. 1).vidence of jets produced by black holes from optical and infrared observations. 19
3. The jet properties as revealed from OIR analyses
In Section 2 it is established that the jets of BHXBs can be detected directly and indirectlyvia numerous methods from optical or infrared observations. We now discuss how thesestudies can help to constrain the properties of the jets. Due to simple scaling laws that existtheoretically and empirically between the small BHs in XBs and the massive BHs in AGN(Section 1.2; see Fender, 2009, for a review), it is possible to infer the properties of AGNjets by studying BHXB jets.From OIR observations of the compact jet, we can essentially uncover its OIR spec-trum (which is usually optically thin; α ∼ − . ), variability properties and polarizationlevel/orientation. The OIR flux and spectral index of the compact jet, when combined withthe spectrum at lower and higher frequencies, reveals the broadband spectrum of the jetwhich can then be compared to the most advanced jet models to infer its properties; forexample its dimensions, Lorentz factor, confinement, and the total jet power. So far, it hasbeen estimated in some cases that the jet power exceeds the radiative power (which can bederived from the bolometric luminosity). A further dimension of information can constrainthe properties more tightly: time-dependent jet modelling. In particular, changes in the jetproperties as a function of mass accretion rate and X-ray state will probe the jet physics.Scaling relations with accretion rate and black hole mass can be extrapolated to BHs inAGN, providing predictions that can be tested with observations. If consistent, it could beshown that all BH jets are ubiquitous.The OIR jet emission of BHXBs originates near the jet base where it is launched andaccelerated, so fast timing and polarization studies could provide a wealth of informationabout the physical conditions here (for example the time-dependent magnetic field config-uration or the accretion energy extraction mechanism), providing constraints for jet for-mation theories. A separate piece of information about the compact jets comes from themoving emission lines from precessing jets when they are seen. These lines are signaturesof ‘heavy’ (atomic rather than leptonic) jets and they can also be used to tightly constrainthe velocity of the jets.Extended, moving jet knots detected at OIR frequencies would help to constrain the jetenergy losses at large distances downstream. Measurements of deceleration coupled withinformation about the ISM may tell us how receptive jets are to producing radiation whencolliding with dense matter.Finally, OIR emission lines from ring bow shocks and trailed bow shocks can be usedas independent methods of calculating the time-averaged jet power and their composition(Gallo et al., 2005; Heinz, 2006). For BHXBs in which this is known and the broadbandspectrum of the compact jet has also been measured, the radiative efficiency of the core jetcan be derived. Again, these properties can be scaled up to AGN jets. Essentially, howBHXB jets interact with the ISM has direct implications for how AGN jets energize theintracluster medium, shaping galaxy formation.
4. Conclusion
To summarise, due to the rapid observational and theoretical advances in jets from BHXBsin the last decade or so, including the finding of direct links with AGN jets, this may be the0 David M. Russell, Rob P. Fender‘golden era’ of our understanding of the fastest objects in the universe – relativistic jets –and how they are produced via the process of accretion onto compact objects. Perhaps acrucial and satisfactory future goal would be to unify the properties of jets from all accretingobjects, including the slowest, weakest jets (for example from young stellar objects) and thefastest, most powerful jets (most likely associated with γ -ray bursts).From an OIR observer’s perspective, there are a number of short term studies in thefield of BHXB jets that will help lead us forward in our understanding towards the longterm goals: • OIR polarization and fast timing analyses of the isolated optically thin jet in manyBHXBs, and in different X-ray states when the jet is not bright, to constrain the ubiq-uity of the properties of the inner regions of the jet; • Track the position of the jet break in the spectrum, with changes in mass accretionrate, which constrains the jet power and jet production theories; • Resolve the OIR counterparts of discrete, moving radio ejecta; • Discover more jet-powered bow shock nebulae and infer the time-averaged jet powerfor a sample of BHXBs, again testing ubiquity; • Compare the properties of BHXB jet-powered structures to AGN lobes and bowshocks, and also to OIR-emitting shock-excited bubbles associated with ultraluminousX-ray sources (which could be intermediate-mass BHs), of which there are many (e.g.Pakull & Gris´e, 2008). • Eventually, it may be possible to resolve the compact jet at OIR frequencies indirectlyusing interferometry (Markoff, 2008), which will open up new avenues of studyingthe inner regions of the jets and accretion flow.
Acknowledgements . DMR would like to thank Piergiorgio Casella for useful discussionsregarding the interpretations of multi-wavelength fast-timing analyses. DMR acknowledgessupport from a Netherlands Organization for Scientific Research (NWO) Veni Fellowship.
References
Baade, W.; Minkowski, R.
ApJ
ApJS
ApJ
MNRAS
ApJ
MNRAS
ApJ
MNRAS
MNRAS
NewA
ApJ
ApJ
MNRAS
A&A
MNRAS
MNRAS
MNRAS
ApJL
ApJ
Science
A&A
ApJ
MNRAS
MNRAS
ApJ
ApJ
ApJ
A&A
A&A
A&A
MNRAS
MNRAS
Compact Stellar X-Ray Sources ; Cambridge University Press: Cambridge,U.K., 2006; pp 381 (eds. Lewin, W. H. G.; van der Klis, M.)Fender, R. P.
Disc-jet coupling in black hole X-ray binaries and active galactic nuclei ; inThe Jet Paradigm - From Microquasars to Quasars, Lect. Notes Phys. 794 (ed. Belloni,T.) (arXiv:0909.2572)Fender, R. P.; Pooley, G. G.; Brocksopp, C.; Newell, S. J.
MNRAS
MNRAS
MNRAS
MNRAS
MNRAS
MNRAS
PhRvD
ApJ
MNRAS
MNRAS
Nature
MNRAS
ApJ
MNRAS
ARep
ApJ
ApJ
A&A
ApJ
MNRAS
ApJ
ApJ
ApJ
ApJ
MNRAS
MNRAS
ApJ
MNRAS
MNRAS
MNRAS
ApJ
ApJ
ApJ
Proceedings of Science , 2008,
Proceedings of the VIIMicroquasar Workshop: Microquasars and Beyond, 1-5 September 2008, Foca, Izmir,Turkey (arXiv:0811.3320)Jonker, P. G.; Nelemans, G.
MNRAS
MNRAS
ApJ
ApJ
Nature
MNRAS
MNRAS
MNRAS
NewAR
ApJ
MNRAS
ApJ
Proceedings of Science , 2008,
Proceedings of the VII Microquasar Workshop:Microquasars and Beyond, 1-5 September 2008, Foca, Izmir, Turkey (arXiv:0811.3601)Markoff, S.; Falcke, H.; Fender, R.
A&A
A&A
ApJ
ApJ
ApJ
ARA&A
ApJ
MNRAS
MNRAS
MNRAS
ApJ
MNRAS
MNRAS
AIPC
Nature
A&A
A&A
A&A
SSRv
MNRAS
ApJ
MNRAS
Nature
AIPC
A&A
ApJ
ApJ
Ph.D. thesis
MNRAS
MNRAS
MNRAS
MNRAS
Proceedings of Science , 2008,
Proceed-ings of the VII Microquasar Workshop: Microquasars and Beyond, 1-5 September 2008,Foca, Izmir, Turkey (arXiv:0811.2919)Sams, B. J.; Eckart, A.; Sunyaev, R.
Nature
ApJ
AAS
BaltA
MNRAS
MNRAS
ApJ
ApJ
MNRAS
Nature
ApJ
ApJ
PASJ
PASJ
PASJ
A&A
A&A
MNRAS