The Spectral Sequence of Blazars - Status and Perspectives
L. Maraschi, G. Ghisellini, F. Tavecchio, L. Foschini, R.M. Sambruna
aa r X i v : . [ a s t r o - ph ] F e b October 31, 2018 19:46 WSPC - Proceedings Trim Size: 11in x 8.5in maraschi THE SPECTRAL SEQUENCE OF BLAZARS – STATUS AND PERSPECTIVES
L. MARASCHI ∗ , G. GHISELLINI and F. TAVECCHIO INAF/Osservatorio Astronomico di Brera, Via Brera 28, 20121, Milano, Italy ∗ E-mail: [email protected]
L. FOSCHINI
INAF/IASF–Bologna, 40129, Bologna, Italy
R. M. SAMBRUNA
NASA/Goddard Space Flight Center, Greenbelt, MD 20771, USA
The present status of the blazar spectral sequence is discussed, including new findings about blazars selected with dif-ferent criteria than the original complete radio-samples. Despite extensive searches of blazars ”breaking” the sequence,the original idea proposed 10 years ago, still seems to hold. On the other hand the forthcoming launch of the
GLAST satellite will provide a new selection band for blazars and blazar related populations as well as fantastic progress onthe spectra and variability behaviour of presently known blazars. The order of magnitude increase in sensitivity of
GLAST will allow to detect γ –rays from jets with lower power and/or lower beaming factor, thus sampling a muchwider population. Keywords : Blazars – BL Lac Objects – Flat-Spectrum Radio Quasars – Relativistic Jets – Gamma-Ray Observations
1. Introduction
The blazar spectral sequence was introduced 10 yearsago [8, 10] in view of understanding the systematicdifferences in broad–band spectral properties fromX–ray selected BL Lacs to radio–selected BL Lacsto Flat Spectrum Radio Quasars (FSRQ). The spec-tral energy distributions (SEDs) of all these sources(unified under the term “blazars”) are dominated bythe beamed non-thermal emission of a relativisticjet. From the results of the
Compton Gamma–RayObservatory , the
Beppo
SAX satellite and Cherenkovtelescopes, it emerged that all SEDs were charac-terized by two peaks, commonly attributed to Syn-chrotron and inverse Compton (IC) radiation respec-tively, emitted by a population of relativistic elec-trons, with seed photons possibly provided by thesynchrotron radiation itself (SSC), the accretion diskor the broad line region (EC). Both peaks appearedto shift to lower frequency from the first class to thenext.Phenomenologically, all the sources of the threecomplete samples available at the time were groupedtogether and divided in radio luminosity bins (eachspanning a decade) irrespective of the original clas-sification and the average SEDs for each bin werecomputed. The results strongly suggested that thesource’s luminosity is a fundamental parameter driv- ing the overall shape of the SEDs, hence the conceptof a “blazar spectral sequence”, implying a link of themain spectral and evolutionary properties of blazarswith the physics of jets and the “central engine”.Modelling the SEDs of individual objects indi-cated that the physical parameters of the radiatingregion in the jets vary systematically with increas-ing luminosity in the sense of an increase of theradiation energy density of the seed photons avail-able for the inverse Compton process and a decrease of the energy of the electrons contributing most tothe emission. A complete description of the spectralmodelling used here is given in [4, 12]. The modelincludes synchrotron and Inverse Compton radiationfrom a population of relativistic electrons whose en-ergy distribution is determined by injection and en-ergy losses in a finite time
R/c where R is the size ofthe emitting region. Jets immersed in strong radia-tion fields will then have electron distributions withspectral breaks at lower energies.The blazar sequence can also be understood interms of a cosmological decrease in the average ac-cretion rate ˙ m onto the central SMBHs. At high–redshift, FSRQs are characterized by high accretionrates and L jet ≃ L disk ; on the other hand, at low–redshift, BL Lac objects do not show signatures of anaccretion disk in the optical band and L jet > L disk , ctober 31, 2018 19:46 WSPC - Proceedings Trim Size: 11in x 8.5in maraschi MS 2141+173z=0.211MS 1340.7+28z=0.905 MS 1234.9+66z=0.852PKS 0402-362z=1.417
Fig. 1. SEDs of the 4 EMSS blazars discussed in this paper, together with the results of our modelling. The dashed line is theassumed contribution of the accretion disk emission. The grey stripe is the sensitivity (5 σ ) of GLAST for one year of exposuretime. implying low radiative efficiency for the accretionflow (i.e. largely sub–Eddington accretion rates) un-less one is prepared to accept that the jet power largely exceeds the accretion power. The transitionfrom near–Eddington to sub–Eddington accretionrates with decreasing redshift is naturally explainedby the cosmological decrease in the availability of gasand increase in the black hole masses at the centersof galaxies [2, 3, 17, 18, 23]. It is important to notethat the same conceptual trend could apply to theFRI/FRII dicothomy [11].It is therefore of paramount importance to testand extend the original blazar sequence suggestion,by removing the bias introduced by the incompleteand non–homogeneous coverage at γ –ray energiesand by producing independent and larger samplesof blazars. With the advent of GLAST both aspectswill receive unprecedented momentum.Here we discuss new objects with peculiar char-acteristics, found as a result of systematic programsor serendipitously, which may already challenge thesequence concept.
2. New data, new sources
One key–prediction of the sequence concept is the lack of FSRQ with a synchrotron peak in the UV/X–ray energy band ; therefore such objects have been ex- tensively searched (see [20] and references therein).Some authors [1, 15, 21] claimed the discovery ofsuch “anomalous” blazars mainly on the basis oftheir broad band spectral indices, that is ratios ofradio/optical/X–ray fluxes. However a knowledge ofthe spectra within each band is essential to confirmthe claim.For the above reasons we recently started newobservations with
Swift of FSRQs in the only
X–rayselected sample of broad line radio loud AGN [27],derived from the
Einstein Medium Sensitivity Survey (EMSS) [14]. X–ray selection allows to probe RadioLoud Quasars 10–100 times weaker in the radio thanclassical samples and by analogy with the case of X–ray selected BL Lacs, Radio Loud Quasars with ahigh frequency synchrotron peak could be expected.Moreover,
Swift provides not only spectral data inthe X–ray band but also simultaneous fluxes in dif-ferent optical/UV bands, yielding precious informa-tion on the shape of the optical to X–ray SED.
EMSS Blazars
The optical and X–ray data obtained for the first4 FSRQs observed with
Swift are presented in Fig.1 together with archival radio data and theoreticalmodels for the SEDs. The objects are ordered in red-shift and present very similar SEDs: in all cases the ctober 31, 2018 19:46 WSPC - Proceedings Trim Size: 11in x 8.5in maraschi optical emission is likely associated with the “bluebump”, that is the thermal emission from the accre-tion disk. The X–ray emission is rather hard suggest-ing an IC component with “external” seed photons ifascribed to the jet, as chosen in Fig 1. However a con-tribution of X-ray emission from the accretion disk ispossible. The strength of the non thermal (jet) emis-sion relative to the accretion disk, as estimated in Fig1, increases with increasing redshift. This may sug-gest that at low redshifts, where lower X–ray lumi-nosities can be detected, sources with weaker jets rel-ative to the intensity of the accretion disk are found.The weakness may be intrinsic or due to a largerviewing angle causing a lower beaming factor.The γ –ray fluxes predicted from the models areuncertain, as they depend on the shape of the rel-ativistic electron spectrum at relatively modest en-ergies, where the constraints are poor, as well as onthe photon density at the site of the emission re-gion. There is thus some freedom in the parameterchoice and we have followed the criterion to min-imize the observed bolometric luminosity. Despitethis “economic” choice, the predicted γ –ray emissionfalls close to the GLAST sensitivity limit, which isalso shown in Fig. 1. Taking into account variability,it is likely that some of these objects will be detectedby
GLAST , in particular MS 0402 −
362 which hasthe highest X–ray to optical ratio, i.e. the largest jetto accretion power ratio. The
GLAST data wouldprovide important constraints to the model parame-ters.
The role of the viewing angle andGLAST
In order to estimate the effect of different viewingangles on the appearence of the SED of a given ob-ject with respect to the blazar sequence we use theSED model for the well known blazar 3C 454.3 [26].We compute model SEDs with the same physical pa-rameters but different viewing angles ( < ◦ ), as-suming a homogeneous jet. The results are shownin Fig. 2 overlayed on the sequence SEDs. Clearlythe γ –ray emission depends strongly on the viewingangle while the synchrotron emission is less affectedand the blue bump luminosity remains constant. Weconclude that the very existence (up to now) of thesequence is probably due to the fact that the sensi-tivity limits of the initial surveys only allowed themost beamed objects to be detected. This situation will change dramatically with GLAST which will beable to detect blazars at larger viewing angles. Wealso recall that “structured jets” [13] may show sig-nificant γ –ray emission over a larger range of angles. Fig. 2. Example of how the SED of a blazar (3C 454.3)changes by changing the viewing angle, hence the Doppler fac-tor. Note that the EC component varies the most, since thisemission is anisotropic even in the comoving frame. We showfor comparison the SED corresponding of the blazar sequence.
New very luminous blazars
A number of highly luminous (high z ) blazars havebeen discovered recently (e.g. GB 1428+4217, z =4 .
72 [7]; PMN J0525–3343, z = 4 . z = 4 .
276 [28]; Q0906+6930, z = 5 .
47 [22];RBS 315, z = 2 .
69 [25]). All exhibit a very hardX-ray spectrum, which indicates an inverse Comp-ton origin for the high energy radiation in agree-ment with the sequence trends. However few data athard X-rays exist for these objects. Instead, the
BAT instrument onboard
Swift , discovered a new ”hardX-ray selected” blazar, SDSS J074625.87+244901.2,with extremely high luminosity. Its spectrum in the10-100 keV band points to a largely dominant inverseCompton peak in agreement with expectations fromthe sequence [19].Two extremely high luminosity objects discov-ered recently: SDSS J081009.94+384757.0 ( z =3 . z = 3 . synchrotron peak inhard X–rays, violating the sequence dramatically. If ctober 31, 2018 19:46 WSPC - Proceedings Trim Size: 11in x 8.5in maraschi correct, this suggestion would require extremely highvalues of both, magnetic field and particle energies.In order to verify these claims, we reanalised theexisting data from Swift and
INTEGRAL to betterconstrain the SEDs of these two sources. The resultsare displayed in Fig 3. In both cases the SED fromthe optical to the X–ray range appears to be concavepointing to an Inverse Compton origin for the X–rayto γ –ray emission as in the “standard” model for thesequence SEDs [10]. While hard X-ray data for SDSSJ081009.94+384757.0 are lacking, the INTEGRAL data for MG3 J225155+2217, show that its SEDcompares well with SDSS J074625.87+244901.2.Both objects are extreme in luminosity and hard X–ray to optical ratio, suggesting an extension of thesequence to higher luminosities in the sense of aneven higher Compton “dominance” and a γ –ray peakat relatively low energies, in the 1–10 MeV range. GLAST observations will be crucial to understandthe high energy emission of these sources. Note that,at the intensity level recorded recently a
GLAST de-tection of J225155+2217 is expected despite its softspectrum in the 10 MeV–10 GeV range, while thesame is not true for J081009.94+384757.0.
Fig. 3. The SEDs of the two high– z blazars discussed in thetext, together with the proposed models. Intermediate luminosity blazarspeaking at high frequencies
Let us discuss first the well known BL Lac objectPKS 2155–304 (see Fig.4). This object is similar to,but more luminous than Mkn 421 and Mkn 501. In fact its “normal” state as described by the data of [9](green dots) falls well on the sequence with the syn-chrotron peak around 10 Hz [19], while on aver-age Mkn 421 and Mkn 501 peak at somewhat higherfrequencies, in agreement with the sequence. How-ever, during flares, all these objects exhibit a trendof increasing peak frequency with increasing luminos-ity, contrary to the sequence trends . Thus we cautionthat the sequence “concept” applies only to aver-age states. In the exceptional flaring state of July28 2006, PKS 2155–304 changed its peak intensityby almost one order of magnitude, thus moving fromluminosity class 2 to 3 and breaking the sequenceduring its high state.
Fig. 4. The SED of PKS 2155–304 (points) compared to thethe average SEDs corresponding to the blazar sequence. Seetext for details.
The source RX J1456.0+5048 ( z = 0 . Swift dataanalysed by us are shown in Fig. 5 superimposedon the sequence average SEDs. The object exhibitsa peak at high frequency (10 Hz) with a ratherlarge peak X–ray luminosity (10 erg s − ), fallingin the central region of the sequence where the aver-age SEDs nearly cross each other. The optical spec-trum of this source exhibits very weak emission lines,approaching a BL Lac classification which reinforcesthe similarity to PKS 2155–304. ctober 31, 2018 19:46 WSPC - Proceedings Trim Size: 11in x 8.5in maraschi The intriguing case of RGBJ1629+401
This blazar is particularly interesting, since it ex-hibits very unusual properties. It is known sincesome years and was suggested to be a FSRQ ( z =0 . Fig. 5. The SED of the two “blue” quasars discussed in thetext, together with the proposed model. For RGB J1629+401we show two possible models, to show (blue) the effect of apossible X–ray corona associated to the disk accretion lumi-nosity. In both cases we have superposed (thin solid lines) theaverage SEDs corresponding to the blazar sequence.
Its optical spectrum shows strong but narrowemission lines a . Due to its high optical luminosity( M = −
23) it was classified as a radio–loud NarrowLine Type 1 Quasar by Komossa et al. [16, 29] in asystematic search for radio–loud NL objects of whichNL Seyfert 1 galaxies are the most common. Themass of the black hole is estimated to be 2 × so-lar masses and its X–ray luminosity is highly super–Eddington. However if the X–ray luminosity is at-tributed to a beamed jet, as suggested by the SEDand by the strongly “inverted” spectral index in theradio, the latter estimate may be reduced. We notethat the radiation energy density U ext derived fromthe observed luminosity of the narrow emission lines,assuming a distance from the black hole consistentwith the line width ( v < U ext issmall, our model associates to this blazar a relativelylarge γ peak , and then this source falls in the middleof the γ peak ∝ U − branch of Fig. 6, though tech-nically it can be defined as a quasar from the pointof view of the brightness of its optical nucleus. Wetherefore believe that it is reasonable to hypothesizethat this object hosts a strong relativistic jet. We re-call that NL Type 1 quasars are rare among quasarsand ”radio loudness” is rarer in NL quasars than inBL quasars [16]. Thus it is not surprising that suchobjects were not found in the radio–loud samples ex-plored previously. Clearly RGB J1629+401 poses anumber of questions that will need further studies.A detection at γ -ray energies could confirm the jetmodel but is not guaranteed at this average intensitylevel. Fig. 6. The relation between γ peak and the total (magneticplus radiative) energy density as seen in the comoving frame.We have labelled the sources discussed here and compare themto the sample of blazars discussed in [4].
3. Discussion and Conclusions
Despite the efforts to search for objects which mayviolate the sequence trends no “strong outliers” havebeen found. Some claims from other authors seemunjustified, as we have shown.The existence of the sequence can be traced backto a strong physical link between the energy of theelectrons emitting most of the power and the total(radiative plus magnetic) energy densities. As dis- a http://cas.sdss.org/astrodr6/en/tools/explore/obj.asp?ra=247.2554958&dec=40.1332306 ctober 31, 2018 19:46 WSPC - Proceedings Trim Size: 11in x 8.5in maraschi cussed by [10] and [12] this trend can be the resultof the balance between the cooling rate (measuredby the amount of total energy density) and the (al-most universal) acceleration rate of the electrons.The most powerful sources have a large amount ofmagnetic and radiation energy density, determininga severe cooling and thus a small value for the equi-librium Lorentz factor of the electrons. On the con-trary, BL Lacs are characterized by a low level ofcooling, explaining the large electron Lorentz factorsin these sources.In Fig. 6 the parameters derived for the sourcesdiscussed in this paper are compared with those de-termined for the group of blazars recently consid-ered by [4]. Although RGB J1629+401 does not de-viate significantly from other blazars in this parame-ter plane it should be stressed that its other charac-teristics, notably the estimated black hole mass andoptical narrow line spectrum are at variance withpreviously known blazars. Searching for more objectslike RGB J1629+401, could extend the “blazar phe-nomenon” to low masses and provide useful hintsto understand the link between relativistic jets andblack hole masses.High– z objects show extreme ratios of inverse-Compton to synchrotron emission, which – with ref-erence to the adopted SED model – corresponds to ahigh energy density of external radiation. It is easy toanticipate that in the study of this external Comptoncomponent GLAST will play a crucial role.
Acknowledgments
This work has made use of the
NASA/IPAC Ex-tragalactic Database (NED) , which is operated bythe
Jet Propulsion Laboratory , California Institute ofTechnology , under contract with the
National Aero-nautics and Space Administration , and of data ob-tained from the
High Energy Astrophysics ScienceArchive Research Center (HEASARC) , provided by
NASA’s Goddard Space Flight Center . This workhas been partially supported by contracts ASI-INAFI/023/05/0 and I/088/06/0.
References [1] Bassani L., Landi R., Malizia A., et al.,
ApJ , L1,(2007) [2] B¨ottcher M. & Dermer C.D.,
ApJ , 86 (2002)[3] Cavaliere A. & D’Elia V.,
ApJ , 226 (2002)[4] Celotti A. & Ghisellini G.,
MNRAS in press, 2008(astro–ph/0711.4112)[5] Donato D., Ghisellini G., Tagliaferri G., Fossati G.,
A&A , 739 (2001)[6] Fabian A.C., Celotti A., Iwasawa K., McMahon R.G.,Carilli C.L., Brandt W.N., Ghisellini G. & Hook,I.M.,
MNRAS , , 373 (2001)[7] Fabian A.C., Iwasawa K., McMahon R.G., Celotti A.,Brandt W.N. & Hook M. MNRAS , , L25 (1998)[8] Fossati G., Maraschi L., Celotti A., Comastri A.,Ghisellini G., MNRAS , 433 (1998)[9] Foschini L., Ghisellini G., Tavecchio F., et al.,
ApJ , L81 (2007)[10] Ghisellini G., Celotti A., Fossati G., Maraschi L.,Comastri A.,
MNRAS , 451 (1998)[11] Ghisellini G. & Celotti A.,
A&A , L1 (2001)[12] Ghisellini G., Celotti A., Costamante L.,
A&A ,833 (2002)[13] Ghisellini G., Tavecchio F. & Chiaberge M., 2005,
A&A , , 401[14] Gioia, I.M., Maccacaro T., Schild R.E., et al., ApJS , 567 (1990)[15] Giommi P., Massaro E., Padovani P., et al., A&A , 97 (2007)[16] Komossa S., Voges W., Xu D., et al., AJ , 531(2006)[17] Maraschi L., in 20 th Texas Symposium on Relativis-tic Astrophysics , (AIP Conference Proceedings 586,Melville, NY, 2001)[18] Maraschi L. & Tavecchio F.,
ApJ , 667 (2003)[19] Maraschi L., Ghisellini G. & Tavecchio F., in
Thefirst GLAST symposium , AIP Conference Proceed-ings, 921, pp. 160-162 (2007)[20] Padovani P.,
Astrophys. Space Sci , 63 (2007)[21] Padovani P., Costamante L., Ghisellini G., GiommiP., Perlman E.,
ApJ , 895 (2002)[22] Romani R.W., Sowards–Emmerd D., Greenhill L. &Michelson P.,
ApJ , , L9 (2004)[23] Sambruna R.M., Gliozzi M., Tavecchio F., MaraschiL., Foschini L., ApJ , 146 (2006)[24] Sambruna R.M., Markwardt C.B., Mushotzky R.F.,et al.,
ApJ , 23 (2006)[25] Tavecchio F., Maraschi L., Ghisellini G., et al.,
ApJ , 980 (2007)[26] Tavecchio F., Maraschi L., Wolter G., et al.,
ApJ , 900 (2007)[27] Wolter A. & Celotti A.,
A&A , 527 (2001)[28] Yuan W., Matsuoka M., Wang T., Ueno S., KuboH. & Mihara T.,
ApJ , , 625 (2000)[29] Zhou H.Y. & Wang T.G., Chin. J. Astron. Astro-phys.2