The spectral energy distribution of quiescent black hole X-ray binaries: new constraints from Spitzer
E. Gallo, S. Migliari, S. Markoff, J. Tomsick, C. Bailyn, S. Berta, R. Fender, J. Miller-Jones
aa r X i v : . [ a s t r o - ph ] J un The Spectral Energy Distribution of Quiescent Black Hole X-rayBinaries: New Constraints from
Spitzer
Elena Gallo , , Simone Migliari , Sera Markoff , John A. Tomsick , Charles D. Bailyn ,Stefano Berta , , Rob Fender , James C. A. Miller-Jones ABSTRACT
Among the various issues that remain open in the field of accretion onto blackhole X-ray binaries (BHBs) is the way the gas accretes at very low Eddingtonratios, in the so-called quiescent regime. While there is general agreement thatthe X-rays are produced by a population of high-energy electrons near to theBH, the controversy comes about in modeling the contribution from inflowingvs. outflowing particles, and their relative energy budget. Recent
Spitzer obser-vations of three quiescent BHBs have shown evidence for excess emission withrespect to the Rayleigh-Jeans tail of the companion star between 8–24 µ m. Wesuggest that synchrotron emission from a partially self-absorbed outflow might beresponsible for the observed mid-IR excess, in place of, or in addition to, thermalemission from circumbinary material. If so, then the jet synchrotron luminosity,integrated from radio up to near-IR frequencies, exceeds the measured 2-10 keVluminosity by a factor of a few in these systems. In turn, the mechanical powerstored in the jet exceeds the bolometric X-ray luminosity at least by 4 orders ofmagnitude. We then compile the broadband spectral energy distribution (SED)of A0620–00, the lowest Eddington-ratio stellar mass BH with a known radiocounterpart, by means of simultaneous radio, optical and X-ray observations, Physics Department, Broida Hall, University of California Santa Barbara, CA 93106 Chandra Fellow Center for Astrophysics and Space Sciences, 9500 Gilman Dr., University of California San Diego, LaJolla, CA 92093 Astronomical Institute ‘Anton Pannekoek’, University of Amsterdam, Kruislaan 403, 1098 SJ, Amster-dam, NL Space Sciences Laboratory, 7 Gauss Way, University of California Berkeley, CA 94720 Department of Astronomy, Yale University, P.O. Box 208101, New Haven, CT 06520 Dipartimento di Astronomia, Universit`a di Padova, Vicolo dell’ Osservatorio 2, 35122 Padova, IT School of Physics and Astronomy, University of Southampton, Southampton SO17 1BJ, UK
Spitzer data. We are able to fit the SED of A0620–00 witha ‘maximally jet-dominated’ model in which the radio through the soft X-raysare dominated by synchrotron emission, while the hard X-rays are dominated byinverse Compton at the jet base. The fitted parameters land in a range of valuesthat is reminiscent of the Galactic Center super-massive black hole Sgr A*. Mostnotably, the inferred ratio of the jet acceleration rate to local cooling rates is twoorders of magnitude weaker with respect to higher luminosity, hard state sources.
Subject headings:
X-rays: binaries — radiation mechanisms: general — stars:individual (A0620–00, V404 Cyg, XTE J1118+480)
1. Introduction
The
Spitzer Space Telescope offers the opportunity for the first time to identify and char-acterize the properties of highly sub-Eddington Galactic black hole X-ray binaries (BHBs)in the mid-infrared band, a frequency window that is still largely unexplored for these sys-tems, and that can prove to be crucial for our understanding of the overall structure of theaccretion flow in quiescence. The infrared (IR) spectra of BHBs with a low mass donorstar are likely shaped by a number of competing emission mechanisms, among which: repro-cessing of accretion-powered X-ray and ultraviolet photons, either by the donor star surfaceor by the outer accretion disk, direct thermal emission from the outer disk, non-thermalsynchrotron emission from a relativistic outflow and thermal emission from circumbinarydust. We refer the reader to Russell et al. (2006; R06 hereafter), and references therein,for a recent comprehensive work on the optical and near-IR spectral properties of X-raybinaries. Here we wish to stress that, as well as for other wavebands, the relative strengthof each mechanism is known to vary greatly in response to changes in the ‘X-ray state’ ofthe system (see McClintock & Remillard 2006; Homan & Belloni 2005). Throughout thiswork, we shall focus on the IR properties of hard and quiescent low mass BHBs. Such (gen-erally transient) systems are characterized by strong variability, power-law dominated X-rayspectra, and integrated X-ray luminosities that are largely sub-Eddington (roughly betweena few 10 − − − times the Eddington luminosity, L Edd , for the hard state, and below a few10 − L Edd for the quiescent state).In spite of the large degree of uncertainty on the overall geometry of the accretion flow inthis regime, there is general agreement that the X-rays are produced by a population of high-energy electrons near to the BH, and that the accreting gas is highly inefficient at radiating,either as a result of an intrinsically reduced radiative efficiency (Narayan & Yi 1994), orbecause of a substantial mass loss (Blandford & Begelman 1999), or a combination of the two 3 –(e.g. Markoff et al. 2001; Yuan et al. 2005). The hard state is associated with the productionof persistent, partially self-absorbed, synchrotron-emitting outflows with flat/inverted radio-mm spectra (Fender 2001). Such jets appear to survive down to quiescent X-ray luminosities(Gallo et al. 2006), even though sensitivity limitations on current radio telescopes make itextremely difficult to reach the signal-to-noise ratios required to assess their presence for lowluminosity systems farther than 2 kpc or so. There is evidence from large-scale structures thatthe jets’ mechanical power is comparable to the bolometric X-ray luminosity in some hardstate sources (e.g. Cyg X-1, Gallo et al. 2005a; Russell et al. 2007). However, even for thehighest quality spectral energy distribution (SED), disentangling the relative contributionsof inflow vs. outflow to the radiation spectrum and global accretion energy budget can bequite challenging, as illustrated by the emblematic case of XTE J1118+480 in McClintock etal. (2003) and Markoff et al. (2001). Estimates of the total jet power based on its radiationspectrum depend crucially on the assumed frequency at which the flat, partially self-absorbedspectrum turns and becomes optically thin, as the jet ‘radiative efficiency’ depends ultimatelyon the location of the high-energy cutoff induced by the higher synchrotron cooling rate ofthe most energetic particles. Once again, this quantity has proved hard to measure.R06 have collected all the available quasi-simultaneous optical and near-IR data ofa large sample of Galactic X-ray binaries over different X-ray states. The optical/near-IR luminosity of hard/quiescent BHBs correlates with the X-ray luminosity to the power ∼ − L Edd (Gallo etal. 2006; but see Gallo 2007 and Xue & Cui 2007). Combined with the fact that the near-IR emission is largely suppressed in the thermal-dominant state (R06, Figure 4), this leadsto the conclusion that, for the BHBs, the break to the optically thin portion would takeplace in the mid-IR (2-40 µ m). Additional evidence for a synchrotron contribution to theIR band in hard state BHBs comes from variability studies during outbursts (e.g. Hynes etal. 2006; Homan et al. 2005). Indeed, from a theoretical point of view, the break frequency ,here defined as the frequency at which the partially self-absorbed jet becomes optically thin,is inversely proportional to the BH mass: as jet spectral breaks are often observed in theGHz/sub-mm regime in active nuclei, they are expected to occur in the IR-optical band for10 − times lighter objects (see discussion in e.g. Markoff et al. 2001 and references therein).We know however from observations of GX 339–4, the only BHB where the optically thin jetspectrum has been perhaps observed (Corbel & Fender 2002; Homan et al. 2005), that theexact break frequency can vary with the overall luminosity, possibly reflecting changes in themagnetic field energy density, particle density and mass loading at the jet base (Nowak etal. 2005). Determining the location of the jet break as a function of the bolometric luminosityis important to assess the synchrotron contribution to the hard X-ray band, and may evenhighlight substantial differences among different classes of objects. As an example, the fact 4 –that the optically thin jet IR-emission in GX 339–4 connects smoothly with the hard X-raypower law has led to challenge the ‘standard’ Comptonization scenario for the hard X-raystate of BHBs (Markoff et al. 2001). On the contrary, recent Spitzer observations of theultra-compact neutron star
X-ray binary 4U 0614+091 (while in a hard state) revealed thatthe break frequency must take place in the far-IR in this system, effectively ruling out asynchrotron origin for the X-ray power law (Migliari et al. 2006).In addition to the jet,
Spitzer observations of quiescent BHBs should be sensitive topossible emission from circumbinary material. Circumbinary disks may be formed as a re-sult of mass outflow from the accretion disk, and have been invoked as an efficient process forthe removal of orbital angular momentum in addition to gravitational radiation loss and/ormagnetic braking (see Taam & Spruit 2001 in the context of cataclysmic variables). Alter-natively, circumbinary material could be due to the presence of a post-supernova explosionfall-back disk, as argued in the case of the anomalous X-ray pulsar 4U 0142+61 (Wang etal. 2006). Muno & Mauerhan (2006; MM06 hereafter) report on
Spitzer observations of fournearby low mass X-ray binaries: three BHBs plus one neutron star system. Excess mid-IRemission – with respect to the Rayleigh-Jeans tail of the donor blackbody spectrum – isdetected from two (possibly all three) BH systems; MM06 attribute this bump to circumbi-nary dust that is illuminated by the low mass companion star. This would imply that theoptically thick-to-thin jet break occurs in the mm regime, at much lower frequencies than,e.g., inferred by R06.In this paper, we aim to reassess the relative contribution of the various emission compo-nents to the radio/IR/optical spectra of the BHBs A0620–00, V404 Cyg and XTE J1118+480while in the quiescent state. We first report on the re-analysis of
Spitzer observations, focus-ing on the rms estimate in the 24 µ m datasets (Section 2), then proceed by examining theSED of each source (Section 3). The origin of the detected mid-IR excess emission is discussedin Section 4. We finally focus on the broadband SED of A0620–00, a highly sub-Eddington( L X /L Edd ≃ − ) BHB for which we put together previously published radio/X-ray data,the Spitzer data and new optical data, all taken in 2005. We discuss the results of fittingthe whole SED by means of a maximally jet-dominated model in Section 5. A summary isgiven is Section 6. Spitzer observations
The BHBs A0620–00, V404 Cyg and XTE J1118+480 were observed by
Spitzer between2004 October and 2005 May as part of a survey of nearby low-mass X-ray binaries (PI: Muno,Program 3289). Photometry of the three targets was acquired using the Multi-band Imaging 5 –Photometer for
Spitzer (MIPS; Rieke et al. 2004) at 24 µ m and the Infrared Array Camera(IRAC; Fazio et al. 2004) at 8 and 4.5 µ m. The Basic Calibrated Data (BCD) were re-processed and then mosaicked with the m opex software (Makovoz & Marleau 2005), whichuses single, multi-frame, and dual outlier rejection. As discussed by MM06, in the case ofA0620–00, the MIPS image was affected by dark latent features from a previous observation.The artifacts were corrected by dividing each BCD frame by a normalized median frame(based on all BCDs excluding the source). These corrected BCDs were then mosaickedusing m opex. Unique IR counter-parts, consistent with the radio positions, are significantly( > σ ) detected at 4.5 and 8 µ m for all the three sources. The MIPS 24 µ m images of thetargets are shown in Figure 1: V404 Cyg and A0620–00 are detected at the 2-2.5 σ level,while XTE J1118+480 is undetected.For each counterpart, we constructed the observation-specific point-response function(PRF) with p rf estimate, and extracted the source flux using both standard aperture pho-tometry on the background-subtracted image and PRF-fitting (using a pex), taking care tomask foreground stars. Sky subtraction was carried out through the use of multiple 10 arcsecsky apertures placed over an annulus around the source. Table 1 lists the fluxes as measuredusing both aperture photometry and PRF-fitting on the mosaic images (the measured fluxeswere then corrected for interstellar extinction following the standard prescription for thefrequency-variable absorption by Cardelli et al. 1989). The values obtained with the twomethods are consistent with each other within the errors. While they are also consistent,within the errors, with those measured by MM06, we derive systematically larger (typicallyby a factor 3) rms noise levels for the MIPS 24 µ m fluxes. In fact, statistical uncertaintiesrelated to sky subtraction are usually negligible compared to calibration and systematic un-certainties. However, statistical uncertainties can be appreciable – tens of percent – for lowsignal/noise sources (e.g. Dale et al. 2005). At 24 µ m, this is clearly the case for A0620–00and V404 Cyg, which are both affected by high cirrus background, as apparent from Figure 1.
3. Radio/Infrared/Optical spectra
We first compile the SEDs of the three systems by putting together the
Spitzer data dis-cussed above, plus optical and radio data available in the literature. For A0620–00, we makeuse of new optical data, presented in Section 3.3.1. Clearly, the non-strict simultaneity of theobservations, combined with the known variability of quiescent BHBs at all wavelengths (e.g.Hynes et al. 2003, 2004), should be kept in mind before drawing any definitive conclusionon the modeling. Figure 2 shows the broadband SEDs of V404 Cyg, XTE J1118+480 andA0620–00, while in the quiescent state, from radio to optical wavelengths. 6 –We first focus on the IR-optical spectra: unlike MM06, we do not compare the dataagainst stellar atmosphere models: the smoothness of our SEDs does not demand a sophis-ticated model which can account for fine spectral features. Most importantly, we aim toquantify the relative goodness of the various models via proper χ fitting, which would bemeaningless if we were to apply stellar atmosphere codes to our sparse data-points. Hence,for each object we first model the IR-optical spectrum with a single temperature blackbody,using the best available estimates for the source distance, inclination and effective tempera-ture. The blackbody approximation is meant to mimic the contribution from the donor star.As shown by MM06, the contribution from the irradiated outer accretion disk is negligiblefor the parameter space relevant to these quiescent systems, at least in the IR band. Thebest-fitting blackbody curves are shown in the left panels of Figure 3, with the fitted pa-rameters and reduced χ given in Table 2. Evidently, the single blackbody model provides apoor fit to the data: excess mid-IR emission, with respect to the Rayleigh-Jeans tail of thedonor/disk, is detected in all three cases.Fitting the data with two blackbodies (Figure 3, middle panels) slightly improves thereduced χ in all cases (Table 3). The temperatures and normalizations of these secondaryblackbodies imply indeed larger physical sizes than the orbital separation, possibly support-ing the circumbinary material interpretation (MM06). However, radio emission has beendetected in two of these sources (V404 Cyg: Hjellming et al. 2000, and A0620–00, Gallo etal. 2006), and interpreted as partially self-absorbed synchrotron emission from a relativisticoutflow. The flat/slightly inverted outflow spectrum must become optically thin at higherfrequencies, possibly in the mid-IR (R06). We thus explore the possibility that the mid-IRexcess might be, at least partly, due to non-thermal emission from a jet. This possibility hasbeen ruled out by MM06 on the basis of far too low fluxes/upper limits at 24 µ m. However,our revised estimates for the 24 µ m rms noise levels leave this possibility open.We choose to fit the radio /IR/optical SEDs with a single blackbody plus a broken powerlaw of the form: F ν = F ν × (cid:26) ( ν/ν ) α , ν < ν b ( ν b /ν ) ( α − α ) ( ν/ν ) α , ν > ν b (1)This is meant to account for a partially self-absorbed synchrotron spectrum with index α = 0 . − . ν b , above which it becomes optically thin withindex α . After running a grid of models with all the six fitting parameters (blackbodytemperature and normalization, plus the four broken power law parameters) free to vary,we choose to fix the index of the optically thin portion to α = − . N ( E ) ∝ E − p with power law index p = +2 . E being theelectron energy; e.g. Fender 2006) and the position of the break to ν b = 10 Hz, in orderto maximize the jet contribution to the mid-IR band. The results of the blackbody plus 7 –broken power law fits are shown in the right panels of Figure 3, with the fitted parametersin Table 4. We discuss below the SED compilation and the results of the modeling on a caseby case basis.
Casares et al. (1993) report on B - V - R - J - H - K band photometry of V404 Cyg takenin 1991 July-August, 2 years after the end of the 1989 outburst that preceded the currentquiescent regime (even though this system, because of its relatively high quiescent X-rayluminosity [ L X /L Edd ≃ − . ], is often considered at the boundary between ‘quiescence’ andthe hard X-ray state). Several later works have established V404 Cyg to be variable by afactor of a few at IR-to-X-ray wavelengths (see e.g. Hynes et al. 2004, Bradley et al. 2007for the X-ray/optical variability; Zurita et al. 2004 for a study of the long term optical/IRvariability, and references therein). The origin of such variability is yet to be well understood,even though there is general agreement that it should take place somewhere in the accretionflow rather than in the hot gas stream/donor star (Shahbaz et al. 2003; Zurita et al. 2003;Hynes et al. 2003; 2004).Over the past few years, V404 Cyg has been known as a relatively stable radio source,with an average flux density of ∼ µ Jy, and a flat/slightly inverted spectrum at GHzfrequencies (Hjellming et al. 2000; Gallo et al. 2005b), interpreted in terms of partially self-absorbed synchrotron radiation from outflowing plasma. The variable nature of this system,combined with the fact that the available data spread an interval of several years (the opticaland
Spitzer data were acquired more than 10 years apart), make it especially difficult to drawdefinite conclusions about the mid-IR emission detected with MIPS (on the other hand, R06showed that the optical-IR luminosity of hard/quiescent state sources scales with the X-rayluminosity to the power 0.6, implying that the X-ray variability should be reduced to someextent in the IR).The top panels of Figure 3 show the IR-optical spectrum of V404 Cyg as fitted with asingle and double blackbody model (left and middle panel, respectively): clearly the lattermodel provides a better fit to the IR-optical data, with χ /d.o.f.=10.3/7 and χ /d.o.f.=1.4/5,respectively. However, these components do not account for the radio emission. Because ofthe flat radio spectrum, it can not be ruled out that the excess emission at 24 µ m might bedue to the high frequency portion of the well-established synchrotron-emitting outflow. Thetop right panel of Figure 3 shows a fit to the radio-IR-optical data with a single blackbodywith T ≃ α = 0 .
02. This two-component model provides as a good fit as the 8 –double blackbody model ( χ /d.o.f.=4.0/9), and it also accounts for the radio emission.This suggests that, in this system, synchrotron emission from a partially self-absorbedoutflow is likely to be responsible for the observed mid-IR excess as much as thermal emissionfrom circumbinary material. As an aside, if such excess were entirely due to circumbinarydisk emission, this would imply that the jet break to the optically thin portion has to occursomewhere in the mm regime, i.e. at lower frequencies than predicted by R06. While thesystem SED could be comfortably reproduced by the sum of two blackbody components plusa broken power law, accommodating the circumbinary material and jet the contribution, thiswould require as many free parameters as data points. Gelino et al. (2006) present B - V - R - J - H - K S band photometry of XTE J1118+480 inquiescence. Due to its high Galactic latitude, XTE J1118+480 is a virtually unabsorbedsource ( A V = 0 .
06) and yet it can be taken as an example of how tricky it can be toinfer the properties and the geometry of the accretion flow based on modeling the SED.For instance, the high quality simultaneous multi-wavelength data acquired while in thehard state (at L X /L Edd ≃ − ) have been successfully modeled in terms of an advection-dominated accretion flow (McClintock et al. 2003; Yuan et al. 2005), as well as using a jetsynchrotron model (Markoff et al. 2001). As shown in Figure 3, middle panels, there isevidence for substantial excess emission at 8 µ m with respect to the donor star tail. Wenotice that the single blackbody model provides a very poor representation of the donor starspectrum: in this case, the actual stellar atmosphere model is certainly more appropriate(see MM06, Figure 2). However, as noticed above, we are interested in constraining thenature of the excess mid-IR emission via proper χ fitting: in this framework, irrespective ofhow well the donor star thermal emission is modeled, our goal is to determine whether fittingthe mid-IR excess with a broken power law model provides a better or worse description ofthe data in a statistical sense. The radio counterpart to XTE J1118+480 is undetected inquiescence, with an upper limit of 0.1 mJy at 8.5 GHz (Mirabel et al. 2001). Because ofthis shallow upper limit, the measured excess at 8 µ m might still be interpreted as due toa partially self-absorbed outflow that extends its power-law spectrum from the radio up tothe IRAC regime. This is illustrated in middle right panel of Figure 3, where a partiallyself-absorbed synchrotron emitting outflow with α = 0 .
27, plus a ∼ χ is improved with respect to thedouble blackbody model ( χ /d.o.f. = 12.0/3 vs. 6 . /
3, respectively for the double blackbodyand blackbody plus broken power law model). Within the blackbody plus power-law model, 9 –the fitted values for the blackbody temperature and normalization are consistent, within theerrors, with the inferred values for the donor star (namely ∼ ≃ R ⊙ ; Gelino etal. 2006). The fitted radio spectral index is consistent with hard state sources (Fender 2001),and predicts a GHz flux density lower than 5 µ Jy, practically undetectable with current radiofacilities over reasonable integration times (the rms noise level for a 24 hr integration withthe VLA is about 5 µ Jy at 8.5 GHz; however, planned upgrades, such as the eMERLIN andEVLA will be able to probe such flux density levels in hrs-long exposures).
We construct the SED of A0620–00 by means of radio, IR, optical and X-ray observa-tions, all taken in 2005; the optical/near-IR data were acquired by the Small and ModerateAperture Research Telescope System (SMARTS ) consortium, using the Cerro Tololo Inter-American Observatory (CTIO) 1.3 m together with ANDICAM , a dual-channel imagercapable of obtaining optical and IR data simultaneously. A0620–00 was observed through I - V - H filters on 2005 August 18, one day before the beginning of the (strictly simultane-ous) Chandra/VLA observations (taken on 2005 August 19-20; Gallo et al. 2006), while the Spitzer data discussed above were acquired on 2005 March 06 (MIPS) and March 25 (IRAC).SMARTS data were calibrated using data from previous nights and were processed andreduced using standard IRAF aperture photometry routines. The measured magnitudes wereconverted into fluxes using the SMARTS photometric zero-points; we used a color excess of E ( B − V )=0 . ± .
02 (Wu et al. 1976), and corrected for extinction following again thestandard prescription for the frequency-variable absorption by Cardelli et al. (1989). Theresults are summarized in Table 5. Interestingly, all of the measured magnitudes are brighterthan the maximum magnitude from the previously published quiescent light-curves (see Table1 in Gelino et al. 2001, reporting on optical and IR observations of A0620–00 between 1976and 2001), and from 0.5-0.7 mag brighter than the mean magnitudes. However, given theobserved trend over the past few years of increasing brightness in this source, it seems veryunlikely that these results require a sudden flare. This however has to be kept in mindwhen inspecting the whole SED of A0620–00, in particular when comparing the 2005 March
Spitzer observations with the optical, near-IR values given by Gelino et al. (2001).
10 –
Significant excess emission with respect to the Rayleigh-Jeans portion of the donor’sblackbody spectrum is detected at 8 and 24 µ m. As shown in the middle-bottom panel ofFigure 3, the sum of two blackbodies ( ∼ χ /d.o.f.=2.0/2). The detection of a radio counterpart to A0620–00 strongly suggeststhat this quiescent system is powering a synchrotron-emitting outflow (Gallo et al. 2006).Arbitrarily assuming a flat spectrum for the partially self-absorbed portion of the jet, thiswould have to become optically thin at frequencies lower than 10 Hz for it not to contributeto the mid-IR excess. Alternatively, the whole radio-IR-optical spectrum can be well fit bythe sum of ∼ α = 0 . χ /d.o.f.=7.8/3).
4. Origin of the mid-IR excess: implications for the jet power
The
Spitzer observations of three quiescent BHBs discussed above show evidence forexcess emission in the mid-IR band; while it may possible to reproduce the emission between2 − × Hz with a blackbody whose temperature is consistent with the shown temperaturesof the secondary stars, it would be difficult to explain the excess at 10 Hz with any modelfor which the temperature is high enough so that 10 Hz is in the Rayleigh-Jeans portionof the blackbody spectrum. Thus, two main possibilities arise to account for the measuredexcess: thermal emission from cool (hundreds of K) circumbinary material, or synchrotronemission from outflowing plasma. The latter hypothesis was dismissed by MM06 on the basisof far too low 24 µ m fluxes/upper limits. Our estimates for the statistical uncertainties onthe 24 µ m observations, however, reinstate this possibility.Under the assumption that non-thermal synchrotron emission is at the origin of themeasured IR-excess, we can estimate the amount of power stored in the outflows. Integratingthe partially self-absorbed jet spectra up to 10 Hz, and assuming a (conservatively low)jet radiative efficiency of 5%, and no Doppler boosting (see Fender 2001), we obtain jetpowers in the range ∼ × erg s − , for A0620–00 and XTE J1118+480, the lowerEddington ratio sources, up to ∼ × erg s − , for V404 Cyg (see Table 6). Under theseassumptions, the total jet power exceed the measured X-ray luminosities (between 2-10 keV)in quiescence by a factor 50 at least. Assuming that the steep X-ray power laws observedin quiescent BHBs (with average photon index Γ ≃
2; e.g. Corbel et al. 2006) extendup to ∼
100 keV, where a spectral cutoff is observable in higher Eddington-ratio systems,the bolometric (0.1-100 keV) X-ray luminosities are likely to exceed the measured 2-10 keVluminosities by a factor of a few. Therefore, this regime of L j , tot > ∼ L X fits the definition of 11 –‘jet-dominated’ state put forward by Fender et al. (2003). The above estimates of L j , tot arebased on a conservative radiative efficiency for the synchrotron process of 5%; as such, theyrepresent strict lower limits. Alternatively, we can estimate the total jet power followingthe formalism by Heinz & Grimm (2005), where the monochromatic radio core emission ( L r ,in units of 10 erg s − ) of three well studied radio galaxies was directly compared to theradio lobe emission, and combined with a self-similar jet model (Heinz & Sunyaev 2001)in order to calibrate the ratio of mechanical vs. radiative power of partially self-absorbedjets. They proposed that the jet kinetic power of both super-massive and stellar size BHscan be estimated from the core radio luminosity as: L j , tot = 6 . × L / (1 . − α r / W . erg s − , where α r is the radio spectral index over the partially self-absorbed regime, andthe parameter W . carries the (quite large) uncertainty on the radio galaxy calibration.The normalization value by Heinz & Grimm is roughly in agreement to that estimated byK¨ording et al. (2006): here, for flat spectrum radio sources, the jet power (at the hard tosoft state transition) is expressed as: L j , tot < ∼ . × ( f / . η/ . L (12 / erg s − , f being the fraction of outer mass accretion rate that is not expelled via winds/outflows, and η the standard accretion efficiency. Either way, the inferred total jet power would exceedthe bolometric X-ray luminosity by at least 4 orders of magnitude for the three quiescentBHBs under consideration. It is worth mentioning that, independently of normalization andefficiency factors, in all three cases the jet synchrotron luminosity, integrated up to 10 Hz (that is neglecting the optically thin portion), already exceed the measured 2-10 keVluminosities by a factor of a few (Table 6, right column).In contrast, if thermal emission from circumbinary disk material is entirely responsiblefor the measured mid-IR excess, this would imply that the jet spectrum breaks at muchlower frequencies, perhaps in far-IR/mm regime, lowering the above estimates by a factorof ten at least. A final test to assess the origin of the measured excess could be variabilitystudy in the mid-IR, possibly coordinated with the radio.
5. A maximally jet-dominated model for the quiescent state
Ultimately, as discussed by McClintock et al. (2003), while there is general agreementthat the X-ray emission in quiescent BHBs comes from high-energy electrons near the BH,the disagreement comes about in: i) attributing the emission to outflowing vs. inflowingelectrons; ii) modeling the electron distribution as thermal vs. non-thermal (or hybrid). TheSEDs of quiescent BHBs, as well as low-luminosity AGN are often examined in the contextof the advection-dominated accretion flow model (ADAF; Narayan & Yi 1994), whereby thelow X-ray luminosities would be due to a highly reduced radiative efficiency, and most of the 12 –liberated accretion power disappears into the horizon. Alternatively, building on the work byFalcke & Biermann (1995) on AGN jets, a jet model has been proposed for hard state BHBs.The model is based upon four assumptions: 1) the total power in the jets scales with thetotal accretion power at the innermost part of the accretion disk, ˙ mc , 2) the jets are freelyexpanding and only weakly accelerated via their own internal pressure gradients only, 3) thejets contain cold protons which carry most of the kinetic energy while leptons dominate theradiation and 4) some fraction of the initially quasi-thermal particles are accelerated intopower-law tails. Markoff et al. (2001) argued that jet synchrotron emission could accountfor the broad continuum features of the simultaneous radio through X-ray observations ofXTE J1118+480 while in the hard state. This same model could also explain the broadspectral features of 13 quasi-simultaneous radio/X-ray observations of GX 339–4, and wasable to reproduce the observed non-linear radio/X-ray correlation in this system (Corbelet al. 2003) by varying the amount of power that is channeled in the jet (Markoff et al.2003). Based on the required reflection signatures a new model was developed (Markoffet al. 2005; MNW05 hereafter) which could reproduce the simultaneous radio/X-ray dataof hard state systems (GX 339–4 and Cygnus X-1) via radiation from a compact, mildlyrelativistic jet, combined with a truncated thermal disk. In particular, the X-ray emissioncan be interpreted as a combination of optically thin synchrotron emission predominantlyfrom an acceleration region ∼ −
100 gravitational radii along the jets, plus external(thermal disk photons) and synchrotron self-Compton emission from the base of the jets, ina region associated with a magnetic compact corona. The radio through the soft X-rays aredominated by synchrotron emission, while the hard X-rays are mostly Comptonization, withweak reflection. This ‘maximally jet-dominated model’ was intended to explore the possibilitythat the ‘hot electron corona’ and ‘jet base’ may be intimately related, or, in the extremecase, synonymous (we refer the reader to MNW05 for a fuller description). This model hasbeen tested extensively on simultaneous radio and X-ray data, and for a number of hard stateBHBs. The mid-IR portion of the spectrum is clearly crucial in order to put constraints onthe optically thick-to-thin jet breaks, as demonstrated by the
Spitzer observations of theneutron star X-ray binary 4U 0614+091 (Migliari et al. 2006) and the BHB GRO 1655–40(Migliari et al. , submitted to ApJ).In the following we attempt to fit the radio through X-rays SED of A0620–00 in qui-escence via the maximally jet-dominated model, where full details can be found in theAppendix of MNW05. The choice of A0620–00 (over e.g. V404 Cyg, for which the ra-dio spectrum is well constrained) is motivated by the fact that, with the exception of the
Spitzer data, the observations were acquired nearly-simultaneously (the VLA/
Chandra ob-servations were strictly simultaneous, while the SMARTS observations were taken only oneday apart). As a comparison, the broadband SED of V404 Cyg is built on datasets that were 13 –taken over 10 years apart. In addition, A0620–00 has been in quiescence for over 30 years,and is considered as a stable and moderately variable system, while V404 Cyg is known tovary in flux by a factor of a few within hours (e.g. Hynes et al. 2003).
The fitting was performed with the
Interactive Spectral Interpretation Sys-tem ( ISIS ; Houck & De Nicola 2000). As outlined in MNW05, the fitting is initiated outsideISIS in order to avoid local minima, using unfolded data that yield a set of starting param-eters for which the reduced χ is lower than 2. We have decided to fix several parameterswhich previously have been allowed to vary, in some cases because the results of fitting themodel to several hard state sources suggest that there may be canonical values, and secondlybecause of the low count rates. In spite of the large luminosity difference between A0620–00( L X /L Edd ≃ − ) and other sources whose hard state spectra were successfully fitted by thejet model, such as XTE J1118+480 (Markoff et al. 2001), GX339–4 and Cygnus X–1 (Markoffet al. 2005), simultaneous VLA/ Chandra observations of A0620–00 in quiescence have shownthat the non-linear radio/X-ray correlation for hard state BHBs appears unbroken all theway down to 10 − L Edd , arguing for no substantial difference between hard and quiescentstate (Gallo et al. 2006; but see Xue & Cui 2007 and Gallo 2007). On the other hand, recenthigh statistics X-ray observations of hard state BHBs seem to show that a geometrically thindisk is present and extends close to the innermost stable orbit already at 10 − L Edd (Miller etal. 2006a, 2006b; Rykoff et al. 2007). As such solution would be very difficult to maintain at10 − L Edd , these authors conclude that a major transition has to take place at intermediateluminosities. Consequently, in light of the large degree of uncertainty over the nature andgeometry of the accretion flow in quiescence, this must be considered as an exploratory study.The model is most sensitive to the fitted parameter N j , which acts as a normalization,though it is not strictly equivalent to the total power in the jets (see discussion in MNW05).It dictates the power initially divided between the particles and magnetic field at the base ofthe jet, and is expressed in terms of a fraction of L Edd . Once N j is specified and conservationis assumed, the macroscopic physical parameters along the jet are determined assuming thatthe jet power is roughly shared between the internal and external pressures. The radiatingparticles enter the base of the jet where the bulk velocities are lowest, with a quasi-thermaldistribution. Starting at location z acc in the jets, a free parameter, a fraction 85% of theparticles are accelerated into a powerlaw with index p , also a fitted parameter. The maximumenergy of the accelerated leptons is calculated by setting the acceleration rate to the local 14 –cooling rates from synchrotron and inverse Compton radiation at z acc . If the accelerationprocess is diffusive Fermi acceleration, the acceleration rate depends on the factor f = ( u acc /c ) f sc , where u acc is the shock speed relative to the bulk plasma flow, and f sc is the ratioof the scattering mean free path to the gyro-radius. Because neither plasma parameteris known, we fit for their combined contribution via f , which thus reflects the efficiency ofacceleration . The particles in the jet radiatively cool via adiabatic expansion, the synchrotronprocess, and inverse Compton up-scattering; however, adiabatic expansion is assumed todominate the observed effects of cooling. A weak thermal accretion disk is assumed tobe present, with an inner disk temperature (somewhat arbitrarily) fixed at T = 10 K,or ∼
90 eV (inner disk temperatures between 50–200 keV are typically obtained for higherEddington ratio sources). This component is also included in the Figure 4 and its photons areconsidered for local inverse Compton up-scattering. However they are negligible comparedto the photons produced by synchrotron radiation. The other main model parameters arethe electron temperature T e , and the equipartition parameter between the magnetic field andthe radiating (lepton) particle energy densities, k . A blackbody with temperature 4900 K,consistent with the companion star (Casares et al. 1993), is added to the model to accountfor the optical emission. An additional blackbody component has been also added to thefit, with normalization free to vary, in order to account for possible contribution from theouter disk. These photons are also included in the Comptonization. The ratio of the ‘nozzle’(i.e. the pre-acceleration region) length to its radius has been fixed to 1 .
5, based on resultsin MNW05. The inclination angle between the jet axis and line of sight i has been fixedto 43 ◦ , the mass fixed to 9.7 M ⊙ and the distance to 1.2 kpc, according to the recentresults by Froning et al. (2007). We wish to stress that adopting the system parametersinferred by Gelino et al. (2001) –i.e. 11 M ⊙ for the BH mass and i =40.75 ◦ – does not resultin a substantial change of the fitted parameters. Starting with parameter values similarto those found in other hard state BHBs, we have obtained a reasonable fit to the data,with χ /d.o.f.=14.3/11. The best fit model is shown in Figure 4, with parameters and 90%confidence error bars given in Table 7. Most of the free parameters have landed in ranges which we are starting to recognize as‘typical’ based on higher luminosity sources such as Cyg X-1 and GX 339-4 (MNW05),GRO J1655-40 (Migliari et al., submitted to ApJ) and the low luminosity AGN M81*(Markoff et al., in prep.). Interestingly, the two main differences appear to be related to theacceleration and equipartition. In higher luminosity sources we have found ratios of mag-netic energy density to the energy densities in radiating particles on the order of ∼ −
5, 15 –while here our best fit value actually favors a slight domination of the particle energy overthe magnetic field (0 . < k < . f compared to thelocal cooling rates. We find f to be around two orders of magnitude lower for A0620–00than in higher luminosity sources. Interestingly, the only other black hole we can studycurrently with similarly weak accretion is Sgr A*, the Galactic Center super-massive BH. Infact, the jet model was first developed in simplified form by Falcke & Markoff (2000), withthe aim to determine whether the same kind of model that could explain the inverted radiospectrum of Sgr A* could also account for the newly discovered X-ray emission (Baganoffet al. 2000) . They concluded that the SED of Sgr A* does not require a power law ofoptically thin synchrotron emission after the break from its flat/inverted radio spectrum.Therefore, if the radiating particles have a power-law distribution, it must be so steep asto be indistinguishable from a Maxwellian in the optically thin regime, i.e. they must beonly weakly accelerated . Here we have shown that something similar, albeit less extreme,is occurring in the quiescent BHB A0620–00; either scenario implies that acceleration in thejets is absent or very inefficient at 10 − − − L Edd .
6. Summary
We compile the radio/IR/optical spectra of three quiescent BHBs: V404 Cyg, XTEJ1118+480 and A0620–00, for which we also present new optical SMARTS observations.Re-analysis of the archival
Spitzer
MIPS data for these systems yields systematically highervalues for the statistical uncertainties related to sky subtraction with respect to the standard ∼
10% value that is typically quoted for bright point-like sources. While our revised valuesfor the 24 µ m fluxes are still consistent with those given by MM06 at the 3 σ level, theyallow for a different interpretation of the measured mid-IR excess with respect to the tail ofthe donor star thermal component. We suggest that non-thermal emission from a jet couldbe responsible for a significant fraction (or all) of the measured excess mid-IR emission. In this framework, radically different particle distributions, such as power laws and Maxwellians, mayresult in similar fits as long as the characteristic particle energy (minimum and peak energy, respectively forthe power law and the Maxwellian) is similar. See MNW05, Appendix.
16 –While this possibly may not rule out the presence of circumbinary material, we argue thatthe radio/IR/optical spectra of the three BHBs under consideration do not require – in astatistical sense – the presence of an additional thermal component. A variability studycould definitively address the question on the origin of the mid-IR excess, as, contrary tonon-thermal jet emission, circumbinary disk emission is expected to be steady.If non-thermal emission from a partially self-absorbed outflow is indeed responsible forthe measured mid-IR excess, then the synchrotron luminosity of the jet, even excludingoptically thin radiation from the base, exceeds the measured 2-10 keV luminosity by a factorof a few in all three systems. In turn, the jet mechanical power in quiescence is greater thanthe bolometric (0.1-100 keV) X-ray luminosity by several ( > ∼
4) orders of magnitude.We proceed by focusing on A0620–00, the lowest Eddington-ratio BHB with a knownradio counterpart, and construct its quiescent SED by adding VLA,
Spitzer , SMARTSand
Chandra data. In spite of the non-simultaneity of the
Spitzer observations with theradio/optical/X-ray observations (which were taken over a two day period), we fit its broad-band SED of A0620-00 with a maximally jet-dominated model (MNW05). This is the firsttime that such a complex model is applied in the context of quiescent BHBs, and with thestrong constraints on the jet break frequency cut-off provided by the
Spitzer data in the mid-IR regime. In terms of best-fitting parameters, the major difference with respect to higherluminosity sources for which this model has been successfully tested is in the value of theacceleration parameter f compared to the local cooling rates, which turns out to be two or-ders of magnitude lower for A0620–00. This weak acceleration scenario is reminiscent of theGalactic Center super-massive BH Sgr A*. Within the jet model working hypothesis, bothSEDs are in fact consistent with the hard X-ray emission stemming primarily from inverseCompton processes in a corona/jet base which is dominated by quasi-thermal particles.E.G. is funded by NASA through Chandra
Postdoctoral Fellowship grant number PF5-60037, issued by the
Chandra
X-Ray Center, which is operated by the Smithsonian Astro-physical Observatory for NASA under contract NAS8-03060. J.A.T. acknowledges partialsupport from
Spitzer contract number 1278068. C.D.B. is funded by NSF grant AST-0407063. S.B. acknowledges support by the Ing. Aldo Gini Foundation. We are gratefulto Mike Nowak for providing us with the analysis scripts for ISIS.
REFERENCES
Baganoff F. et al. 2000, Bulletin of the American Astronomical Society, 32, 1184 17 –Blandford R. & Begelman C., 1999, MNRAS, 303, L1Bradley C., Hynes R., Kong A., Haswell C., Casares J., Gallo E., 2007, ApJ, in press(arXiv:astro-ph/07062652v1)Cardelli J., Clayton, G., Mathis, J., 1989, ApJ, 345, 245Casares J., Charles P. A., Naylor T., Pavlekno E. P., 1993, MNRAS, 265, 834Corbel S., Tomsick J., Kaaret P., 2006, ApJ, 636,Corbel S., Nowak M., Fender R. P., Tzioumis A. K., Markoff S., 2003, A&A, 400, 1007Corbel S. & Fender R., 2002, ApJ, 573, L35Dale D. et al. 2005, ApJ, 633, 857Falcke H. & Markoff S., 2000, A&A, 362, 113Falcke H. & Biermann P. L., 1996, A&A, 308, 321Fazio G. et al. 2004, ApJS, 154, 10Fender R. P., 2006, in Lewin W. H. G., van der Klis M., eds, Compact Stellar X-Ray Sources.Cambridge Univ. Press, CambridgeFender R. P., Belloni T., Gallo E., 2004, MNRAS, 355, 1105Fender R. P., Gallo E., Jonker P. G., 2003, MNRAS, 343, L99Fender R. P., 2001, MNRAS, 322, 31Froning C., Robinson E., Bitner M., 2007, ApJ, in press (arXiv:astro-ph/07040267v1)Gallo E., 2007, in Proc. of ‘The Multicolored Landscape of Compact Objects and theirExplosive Origins’, AIP Conf. Proc. (arXiv:astro-ph/0702126v1)Gallo E. et al. 2006, MNRAS, 370, 1351Gallo E., Fender R. P., Hynes R. I., 2005b, MNRAS, 356, 1017Gallo E., Fender R., Kaiser, C., Russell, D., Morganti, R., Oosterloo, T., Heinz, S., 2005a,Nature, 436, 819Gallo E., Fender R. P., Pooley G. G., 2003, MNRAS, 344, 60 18 –Gelino D., Balman S., Kiziloglu ¨U., Yilmaz A., Kalemci E., Tomsick J., 2006, ApJ, 642, 438Gelino D. M., Harrison T. E., Orosz J. E. 2001, AJ, 122, 2668Heinz S. & Sunyaev R., 2003, MNRAS, 343, L59Heinz S. & Grimm H.-J. 2005, ApJ, 633, 384Hjellming P. M., Rupen M. P., Mioduszewski A. J., Narayan R., 2000, ATel 54Homan J., Buxton, M., Markoff, S., Bailyn, C., Nespoli, E., Belloni, T., 2005, ApJ, 624, 295Homan J. & Belloni T., 2005, A&SS, 300, 107Houck J. & De Nicola L., 2000, in ASP Conf. Ser., Vol. 216, Astronomical Data AnalysisSoftware and Systems IX, eds. N. Manset, C. Veillet, D. Crabtree (San Francisco:ASP), 591Hynes R. et al. 2006, ApJ, 651, 401Hynes R. et al. 2004, ApJ, 611, L125Hynes R. I. et al. 2003, MNRAS, 345, 292K¨ording E., Fender R., Migliari S., 2006, MNRAS, 369, 1451Makovoz D. & Marleau F., 2005, PASP, 117, 1113Markoff S., Nowak M. A., Wilms J., 2005, ApJ, 635, 1203 (MNW05)Markoff S. & Nowak M. A., 2004, ApJ, 609, 972Markoff S., Nowak, M., Corbel, S., Fender, R., Falcke, H., 2003, A&A, 397, 645Markoff S., Falcke H., Fender R., 2001, A&A, 372, L25Miller J., Homan, J., Steeghs, D., Rupen, M., Hunstead, R. W., Wijnands, R., Charles, P.A., Fabian, A. C., 2006a, ApJ, 653, 525Miller J., Homan, J., Miniutti G., 2006b, ApJ, 652, L113McClintock J. E., Remillard R. A., 2006, in Lewin W. H. G., van der Klis M., eds, CompactStellar X-Ray Sources. Cambridge Univ. Press, CambridgeMcClintock J. E. et al. 2003, ApJ, 593, 435 19 –Migliari S., Tomsick, J., Maccarone, T., Gallo, E., Fender, R., Nelemans, G., Russell, D.,2006, ApJ, 643, L41Mirabel I. F., Dhawan V., Mignani R., Rodrigues I., Giglielmetti F., 2001, Nature, 413, 139Muno M. & Mauerhan J., 2006, ApJ, 648, L135 (MM06)Nowak, M.; Wilms, J., Heinz, S., Pooley, G., Pottschmidt, K., Corbel, S., 2005, ApJ, 629,1006Narayan R., Yi I. 1994, ApJ, 428, L13Rieke G. et al. 2004, ApJS, 154, 25Russell D., Fender R., Gallo E., Kaiser C., 2007, MNRAS, 376, 1341Russell, D., Fender, R., Hynes, R., Brocksopp, C., Homan, J., Jonker, P., Buxton, M., 2006,MNRAS, 371, 1334 (R06)Rykoff E., Miller J., Steeghs D., Torres M., 2007, submitted to ApJ(arXiv:astro-ph/0703497v1)Shahbaz T., Dhillon V. S., Marsh T. R., Zurita C., Haswell C. A., Charles P. A., Hynes R.I., Casares J., 2003, MNRAS, 346, 1116Taam R. & Spruit H., 2001, ApJ, 561, 329Wang Z., Chakrabarty D., Kaplan D., 2006, Nature, 440, 772Wu C.-C., Aalders J., van Duinen R., Kester D., Wesselius P., 1976, A&A, 50, 445Xue Y. & Cui W., 2007, A&A, 466, 1053Yuan F., Cui W., Narayan R., 2005, ApJ, 620, 905Zurita C., Casares J., Hynes R., Shahbaz T., Charles P., Pavlenko E., 2004, MNRAS, 352,877Zurita C., Casares J., Shahbaz T., 2003, ApJ, 582, 369
This preprint was prepared with the AAS L A TEX macros v5.2.
20 –Table 1. Spitzer observations of quiescent black hole binaries.
Target Flux ( µ Jy)Method 4 . µ m 8 . µ m 24 . µ mV404 Cyg Ap. Photometry 3336 1820 414 ± ± < < ± ± A V =2.8 (Shahbaz et al. 2003); XTE J1118+480: A V =0.06(Gelino et al. 2006); A0620–00: A V =1.2 (Wu et al. 1983).
21 –Table 2. Single blackbody fits to the IR-optical spectra
Target
R/D T fit T star χ / d.o.f.(1) (2) (3) (4) (5)V404 Cyg 5.4 ± ±
80 5500 10.3/7XTE J1118+480 0.7 ± ±
113 4250 69.6/5A0620–00 2.0 ± ±
104 4900 21.8/4Note. — Columns are: (1) Source name; (2) Fitted star ra-dius over distance,
R/D , times 10 − ; (3) Fitted star tempera-ture, T fit , in K ; (4) Star temperatures as found in the literature, T star , in K (references for the star temperature and system in-clination and distance are the same as listed in the caption ofTable 1 for the extinction values); (5) Fitted χ over degrees offreedom (d.o.f.).
22 –Table 3. Double blackbody fits to the IR-optical spectra
Target (
R/D ) T fit , ( R/D ) T fit , χ / d.o.f.(1) (2) (3) (4) (5) (6)V404 Cyg 5.09 ± ±
94 30 ±
18 489 ±
169 1.4/5XTE J1118+480 0.55 ± ±
150 4 ± ±
140 12.0/3A0620–00 1.7 ± ±
149 23 ±
10 393 ±
83 2.0/2Note. — Columns are: (1) Source name; (2)&(4) Fitted blackbody radius overdistance, times 10 − ; (3)&(5) Fitted blackbody temperature, in K ; (6) Reduced χ . Subscripts 1 and 2 indicate the first and second blackbody components.
23 –Table 4. Blackbody + broken power law fits to the radio-IR-optical spectra.
Target (
R/D ) T fit F ν α χ /d.o.f.(1) (2) (3) (4) (5) (6)V404 Cyg 5.0 ± ±
94 448 ±
189 0.02 ± ± ±
211 62 ±
23 0.27 ± ± ± ± ± ν = 10 Hz, in µ Jy (the broken power-law expression is given in equation 1;we fixed ν b = 10 Hz and α = − . ν b ; (6)Reduced χ .
24 –Table 5. A0620–00: SMARTS observations
Band UT start mag a Flux b ( µ Jy) V ± ± I ± ± H ± ± a Un-dereddened values. b De-reddened values (adopting A V =1.2), allowing for anextra 0.05 mag uncertainty due to systematic calibrationerrors.
25 –Table 6. Jet power
Target α D L j , tot L j , rad /L X (1) (2) (3) (4) (5)V404 Cyg 0.022 4 > × > × > × ν b = 10 Hz; (3) Distance, in kpc; (4)
Total (kinetic + radiative) jet power, in erg s − ; (5) Ratio betweenthe radiative jet power, integrated up to ν b , and the quiescentX-ray luminosity L X , between 2–10 keV. L j , tot is calculatedassuming no Doppler boosting, and a (conservative) 5% ra-diative efficiency; as such, it represents a strict lower limit tothe total jet power. Accordingly, L j , rad = 0 . × L j , tot onlyaccounts for the partially self-absorbed synchrotron emissionfrom the jet. Quiescent X-ray luminosities are taken from:V404 Cyg: Garcia et al. (2001), Kong et al. (2002), Hyneset al. (2004). XTE J1118+480: McClintock et al. (2004).A0620–00: Kong et al. (2002), Gallo et al. (2006).
26 –Table 7. Jet Model for A0620–00 N H N j r z acc T e p f k BB norm (1) (2) (3) (4) (5) (6) (7) (8) (9)3 . +0 . − . . +0 . − . . +2 . − . +272 − . +0 . − . . +0 . − . . +1 . − . . +0 . − . . +0 . − . Note. — Columns are: fitted (1) Hydrogen equivalent column density, in 10 cm − ; (2) Modelinternal normalization, expressed in units of 10 − L Edd : it dictates the power dived by the particlesand the magnetic field at the base; (3) Jet base (or ‘nozzle’) radius, in units of gravitational radii r g = GM BH /c ; (4) Acceleration region, z acc , in r g ; it sets the location along the jet at which (afraction of) the particles start being accelerated; (5) Temperature of the relativistic quasi-Maxwelliandistribution with which the leptons enter the jet, in 10 K; (6) Power law index of the acceleratedelectron distribution, p , where N ( E ) ∝ E − p ; (7) Acceleration parameter, f , in units of 10 − : sets thebalance between particle acceleration and radiative plus adiabatic cooling, such that the quasi-thermalparticles be energized into a power-law tail; (8) Equipartition parameter, k = ( u B /u rad ): the ratiobetween the energy density in radiating leptons and the magnetic field energy density; (9) Internaldisk blackbody normalization, in 10 erg s − . We fixed the BH mass, distance and inclination ofA0620–00 to: 9.7 M ⊙ , 1.2 kpc and 43 ◦ (Froning et al. 2007), yielding χ = 1 .
3. Similar parameters,within the errors, are obtained adopting a mass of 11 M ⊙ and inclination of 40.75 ◦ (Gelino et al. 2001).Error bars are given at the 90% confidence level.
27 –Fig. 1.—
Spitzer
MIPS 24 µ m images of V404 Cyg, XTE J1118+480, and A0620–00. Whitecircles (with 2 arcsec radius) mark the position of the radio counterparts, from MERLINand VLA observations for V404 Cyg (R. Spencer and M. Rupen, private communications);VLA for A0620–00 (Gallo et al. 2006); MERLIN for XTE J1118+480 (Fender et al. 2001).The fields of view of V404 Cyg and A0620–00 are evidently affected by high backgroundcontaminations, resulting in high statistical uncertainties related to sky subtraction. Forreference, 1 MIPS pixel corresponds to 1.2 arcsec in size. North is at the top, and East is tothe left of these images. 28 –Fig. 2.— Composite radio/IR/optical spectra of quiescent black hole binaries. V404 Cyg :radio data from Gallo et al. (2005b), taken in 2002; IR data from this work, taken in 2004-2005; optical photometry from Casares et al. 1993, taken between 1990-1992.
A0620–00 :radio data from Gallo et al. (2006), acquired in August 2005; IR and optical data from thiswork. The data span a period of 5 months, with nearly simultaneous radio/optical coverage.
XTE J1118+480 : radio upper limit from Mirabel et al. (2001); IR data from this work;optical photometry from Gelino et al. (2006). 29 –