Multiwavelength Observations of the Gamma-Ray Blazar PKS 0528+134 in Quiescence
aa r X i v : . [ a s t r o - ph . H E ] A p r Accepted for Publication in
The Astrophysical Journal
Multiwavelength Observations of the Gamma-Ray BlazarPKS 0528+134 in Quiescence
N. I. Palma , , M. B¨ottcher , I. de la Calle , I. Agudo , , M. Aller, , H. Aller, , U. Bach ,E. Ben´ıtez , C. S. Buemi , L. Escande , J. L. G´omez , M. A. Gurwell , J. Heidt , D.Hiriart , S. G. Jorstad , , M. Joshi , A. L¨ahteenm¨aki , V. M. Larionov , , P. Leto , Y.Li , J. M. L´opez , B. Lott , G. Madejski , A. P. Marscher , D. A. Morozova , C. M.Raiteri , V. Roberts , M. Tornikoski , C. Trigilio , G. Umana , M. Villata , D.Wylezalek ABSTRACT
We present multiwavelength observations of the ultraluminous blazar-type ra-dio loud quasar PKS 0528+134 in quiescence during the period July to December2009. Four Target-of-Opportunity (ToO) observations with the
XMM-Newton
Satellite in the 0.2 – 10 keV range were supplemented with optical observations Astrophysical Institute, Department of Physics and Astronomy, Clippinger 339, Ohio University, Athens,OH 45701, USA Facultad de Ciencias Espaciales, Universidad Nacional Autonoma de Honduras, Tegucigalpa M.D.C.,Honduras C. A. European Space Astronomy Center, P.O. Box 78, 28691 Villanueva de la Ca˜nada, Madrid, Spain Institute for Astrophysical Research, Boston University, 725 Commonwealth Avenue, Boston, MA 02215,USA.; [email protected]; [email protected]; [email protected] Instituto de Astrof´ısica de Andaluc´ıa, CSIC, Apartado 3004, 18080 Granada, Spain.; [email protected] Department of Astronomy, University of Michigan, Ann Arbor, MI 48109-1042, USA Max-Planck Institut f¨ur Radioastronomie, Auf dem H¨ugel 69, D-53225 Bonn, Germany Instituto de Astronom´ıa, Universidad Nacional Aut´onoma de M´exico, Apdo. Postal 70-264, CP 04510,M´exico INAF - Osservatorio Astrofisico di Catania, Italy Universit´e Bordeaux 1, CNRS/IN2p3, Centre d’Etudes Nucl´eaires de Bordeaux Gradignan, 33175Gradignan, France Harvard-Smithsonian Center for Astrophysics, Cambridge, MA ZAH, Landessternwarte Heidelberg, K¨onigstuhl, D-69117 Heidelberg, Germany Instituto de Astronom´ıa, Universidad Nacional Aut´onoma de M´exico, Apdo. Postal 877, CP 22800,Ensenada, B.C., M´exico Aalto University, Mets¨ahovi Radio Observatory, Mets¨ahovintie 114, FIN-02540, Kylmala, Finland Astronomical Institute, St. Petersburg State University, Universitetsky pr. 28, Petrodvoretz, 198504 St.Petersburg, Russia Isaac Newton Institute of Chile, St. Petersburg Branch, 198504 St. Petersburg, Russia Kavli Institute for Particle Astrophysics and Cosmology, Department of Physics and SLAC NationalAcelerator Laboratory, Stanford University, Stanford, CA 94305, USA INAF, Osservatorio Astronomico di Torino, I-10025 Pino Torinese (TO), Italy University of Cambridge, Department of Physics, Cavendish Laboratory, JJ Thomson Avenue, Cam-bridge, CB3 0HE, UK
Rossi X-rayTiming Explorer (RXTE; 2 – 10 keV) and from
Suzaku (0.5 – 10 keV) as wellas γ -ray data from the Fermi Large Area Telescope (LAT) in the 100 MeV –200 GeV range. In addition, publically available data from the SMARTS blazarmonitoring program and the University of Arizona / Steward Observatory Fermi
Support program were included in our analysis.We found no evidence of significant flux or spectral variability in γ -rays andmost radio bands. However, significant flux variability on a time scale of severalhours was found in the optical regime, accompanied by a weak trend of spectralsoftening with increasing flux. We suggest that this might be the signatureof a contribution of unbeamed emission, possibly from the accretion disk, atthe blue end of the optical spectrum. The optical flux is weakly polarized withrapid variations of the degree and direction of polarization, while the polarizationof the 43 GHz radio core remains steady, perpendicular to the jet direction.Optical spectropolarimetry of the object in the quiescent state suggests a trend ofincreasing degree of polarization with increasing wavelength, providing additionalevidence for an unpolarized emission component, possibly thermal emission fromthe accretion disk, contributing towards the blue end of the optical spectrum.Over an extended period of several months, PKS 0528+134 shows moderate(amplitude .
50 %) flux variability in the X-rays and most radio frequencieson ∼ γ -ray emission.A leptonic single-zone jet model produced acceptable fits to the SEDs withcontributions to the high-energy emission from both synchrotron self-Comptonradiation and Comptonization of direct accretion disk emission. Fit parametersclose to equipartition between the energy densities of the magnetic field andthe relativistic electron population were obtained. The moderate variability onlong time scales, compared to expected radiative cooling time scales, implies theexistence of on-going particle acceleration, while the observed optical polarizationvariability seems to point towards a turbulent acceleration process. Turbulentparticle acceleration at stationary features along the jet therefore appears to bea viable possibility for the quiescent state of PKS 0528+134. Subject headings: galaxies: active — Flat Spectrum Radio Loud Quasars: indi- 4 –vidual (PKS 0528+134) — radiation mechanisms: non-thermal
1. Introduction
Blazars (BL Lac objects and gamma-ray loud flat spectrum radio quasars [FSRQ])are the most extreme type of active galactic nuclei (AGN). They were historically definedthrough extreme flux variability throughout the electromagnetic spectrum, and sometimesstrong and variable linear polarization at radio and optical wavelengths. In the 1990s,observations by
EGRET on board the
Compton Gamma-Ray Observatory revealed large γ -ray fluxes (often dominating the bolometric luminosity of the source) from many blazars.The radio through optical emission from blazars is commonly interpreted as synchrotronemission from ultrarelativistic electrons in a relativistic plasma jet that is closely alignedwith our line of sight ( θ obs < ◦ ). This assertion is supported by the superluminal motionthat most blazars exhibit (e.g., Jorstad et al. 2001; Lister et al. 2009; Piner et al. 2006) aswell as by the observed luminosity and variability timescales observed in these objects. Inextreme cases, variability time scales down to a few minutes have been found in the veryhigh energy (VHE) γ -ray regime (e.g., Albert et al. 2007; Aharonian et al. 2007).Two competing classes of models are currently being considered for the origin of thehigh-energy (X-ray through γ -ray) emission from blazars. In leptonic models, hadrons (pri-marily protons) in the jet (if present in substantial numbers at all), are assumed not to beaccelerated to ultrarelativistic energies. They do not exceed the threshold for photo-pionproduction processes on the low-frequency radiation field in the jet, and proton synchrotronradiation is assumed to be negligible. Therefore, in leptonic models, the high-energy ra-diative output is dominated by Compton scattering of low-frequency photons off relativisticelectrons. In hadronic models, it is assumed that ultrarelativistic protons exist in sufficientnumber. In such a scenario, the protons will dominate the radiative output via proton syn-chrotron radiation and synchrotron and Compton emission from secondary particles. Thoseare produced in photo-pion production and subsequent pion and muon decay and electro-magnetic cascade processes. For a recent review of blazar emission models, see, e.g., B¨ottcher(2010).The mechanism(s) of acceleration of particles to ultrarelativistic energies in blazar jetsare currently very poorly understood. Particle acceleration may be related to relativisticshocks in an unsteady flow (e.g., Marscher & Gear 1985), internal shocks resulting fromthe collision of relativistic plasma blobs ejected at different speeds (e.g., Spada et al. 2001;Mimica et al. 2004; Joshi & B¨ottcher 2010; B¨ottcher & Dermer 2010), re-collimation shocks(e.g., Bromberg & Levinson 2009), or relativistic shear layers in radially stratified jets (e.g., 5 –Stawarz & Ostrowski 2002; Rieger & Duffy 2004, 2006), to name just a few plausible scenar-ios. Signatures that reveal the nature of particle acceleration in blazar jets may be foundboth in spectral and variability features. The nature of the acceleration mechanism is re-flected in the shape of the produced particle spectra. Those, in turn, can be inferred fromthe shape of the non-thermal photon spectra, in particular in the synchrotron part of theSED (e.g., Finke et al. 2008). The dynamics and light travel time effects, in particular inshock acceleration scenarios, will leave distinct imprints in the observed variability features(e.g., B¨ottcher & Dermer 2010).Observational studies of blazars have so far mostly focused on bright, flaring states ofblazars. This is the consequence of observational constraints which make detailed measure-ments of spectral and variability features in X-rays and γ -rays difficult in low flux states.However, blazars are known to spend most of the time in their quiescent state which has sofar received very little attention and is therefore very poorly understood. EGRET detected γ -ray blazars almost exclusively in flaring states, and the simultaneously operating X-raytelescopes (ROSAT, ASCA, RXTE) lacked the sensitivity to measure detailed X-ray spec-tral and variability properties of most blazars in their quiescent states. Therefore, even thequestion whether γ -ray emission persists at all in the quiescent states of blazars remainedan open issue during the EGRET era.The observational situation has dramatically changed with the advent of the new gener-ation of X-ray observatories, in particular Chandra and
XMM-Newton as well as the launchof the
Fermi
Gamma-Ray Space Telescope in June 2008. The
Fermi
Large Area Telescope(LAT, Atwood et al. 2009) is continuously monitoring the entire sky every 3 hours in theenergy range 20 MeV – 300 GeV with about an order of magnitude superior sensitivitycompared to that of EGRET. It routinely detects γ -ray emission from known blazars evenin their quiescent states. A detailed study of the quiescent state of blazars may elucidatewhether the quiescent jet flow is smooth, exhibiting little or no variability, or the quiescentemission consists of the superposition of a rapid succession of “mini-flares”. In particular,a featureless light curve in all bands might indicate the persistence of particle accelerationmechanisms not related to impulsive (shock) events, and might point towards shear-flowacceleration in radially stratified jets, or standing features such as re-collimation shocks inthe quiescent states of blazars.This situation has motivated us to propose Target-of-Opportunity (ToO) observationswith XMM-Newton in AO-8, triggered by an extended quiescent state of a known γ -raybright blazar. We defined a quiescent state of a γ -ray blazar by the object maintaining a >
100 MeV γ -ray flux lower than the lowest flux or upper limit ever determined by EGRET,over at least 2 weeks. The prominent high-redshift γ -ray bright FSRQ PKS 0528+134 6 –fulfilled our pre-specified trigger criterion through its continued γ -ray quiescence for severalmonths prior to September 2009. We therefore triggered our XMM-Newton
ToO observationson PKS 0528+134. The observations consisted of four observations on 2009 September 8,10, 12, and 14. These were coordinated with ground-based radio and optical observations.Simultaneous γ -ray observations were provided by Fermi
LAT.In the following section, we give an overview of the known properties of our target,PKS 0528+134. In §
3, observations and data reduction procedures are described. In §
4, wepresent the results of a flux and spectral variability analysis. The structure of the parsecscale jet of this source is presented in section §
5. Results of our modeling of four simultaneousSEDs obtained during our campaign are discussed in §
6. A discussion on the optical spectralvariability and other relevant issues is presented in section §
7. Finally, we summarize ourresults and draw conclusions in §
8. Throughout this paper, we refer to a spectral index α as the energy index such that F ν ∝ ν − α , corresponding to a photon index Γ ph = α + 1. Weuse a ΛCDM cosmology with Ω m = 0 .
3, Ω Λ = 0 .
7, and H = 70 km s − Mpc − . In thiscosmology, the luminosity distance of PKS 0528+134 is d L = 16 .
2. The Quasar PKS 0528+134
The compact FSRQ PKS 0528+134 is one of the most luminous and most distant γ -rayblazars known, with a redshift of z = 2 .
07 (Hunter et al. 1993). In the high-energy γ -rayband (above 100 MeV), this source was first detected by EGRET during the period 1991April–June (Mattox et al. 1997). Besides EGRET, this source was also detected by the othertwo instruments onboard of CGRO : the Oriented Scintillation Spectrometer Experiment(OSSE) in the 0.05 – 1.0 MeV band (McNaron-Brown et al. 1995), and the Imaging ComptonTelescope (COMPTEL) ( ≈ γ -ray flux was detected in 1993 March, when itreached 10 JyHz (Mukherjee et al. 1996), strongly dominating the bolometric luminosityof the source. PKS 0528+134 is faint in the optical with a mean visual magnitude of m v = 19 . . < A v < l = 191 . ◦ , b = − . ◦ ( α = 5 h m s . , δ = +13 ◦ ′ . ′′
15 J2000), is locatedbehind the translucent molecular cloud B30 (Liszt & Wilson 1993; Hogerheijde et al. 1995).Hence, there have been relatively few optical observations of this source compared to otherbands. 7 –In the radio regime, PKS 0528+134 is regularly monitored by several programs at dif-ferent frequencies. The source shows pronounced radio flux-density variability on timescalesof several months to a few years (Aller et al. 1985; Reich et al. 1993; Zhang et al. 1994;Stevens et al. 1994; Valtaoja & Ter¨asranta 1995; Pohl et al. 1996; Peng et al. 2001; Bach et al.2007). A delay from high frequencies to low frequencies in radio bursts has been identified,and delays of a few months between γ -ray flares and the corresponding radio bursts havebeen found (Mukherjee et al. 1996). Additionally, from VLBI observations performed in the8.4 GHz band over a period of almost 8 years Britzen et al. (1999) found that PKS 0528+134has a bent jet of length ≈ γ -ray bands is associatedwith morphological changes in the radio structure of the jet, and superluminal motion with β ⊥ , app .
30 is found in some of the jet components (Jorstad et al. 2005).Prior to the time period of EGRET, observations at X-ray energies were carried outby the Einstein Observatory in 1980, but no high confidence values for the X-ray flux andspectral index were found due to the low count statistics (Bregman et al. 1985). Startingin March 1991, and later in September 1992, X-ray observations in the energy range 0.07 –2.48 keV, performed with the ROSAT Position Sensitive Proportional Counter (PSPC) wereused to investigate the geometry and physical environment of PKS 0528+134 (Zhang et al.1994; Mukherjee et al. 1996). The continuum emission of this source in the medium – hardX-ray band (0.4-10 keV) was first measured using observations with the Advanced Satellitefor Cosmology and Astrophysics (ASCA) in 1994 and 1995 (Sambruna et al. 1997). FurtherX-ray observations of PKS 0528+134 have been carried out with the Rossi X-ray TimingExplorer (RXTE) during August and September 1996 and May 1999, as well as with
Bep-poSAX in the 0.1 - 10 keV and 15-200 keV bands during 1997 February and March as part of amultiwavelength campaign involving EGRET and ground based telescopes (Ghisellini et al.1999).SEDs of PKS 0528+134 collected during six years of EGRET observations were compiledand modeled with a one-zone leptonic jet model by Mukherjee et al. (1999). It was foundthat during all EGRET detections of the source, the bolometric luminosity was dominatedby its γ -ray output. While, due to often incomplete multiwavelength coverage, the modelingresults of that paper were subject to a large degree of freedom and uncertainty, the observedepoch-to-epoch variability of PKS 0528+134 was found to be consistent with a correlationbetween the γ -ray flux and the bulk Lorentz factor of the emission region along the jet.Given the extreme properties of PKS 0528+134, this blazar has been the target of manyobservations at different wavelengths. However, as for almost all blazars (see § γ -ray quiescence throughout 2009, as revealed by Fermi
LAT minotoring , PKS 0528+134 was therefore found to be an appealing target forour pre-approved XMM-Newton
Cycle 8 ToO observations and multiwavelength campaign.
3. Observations and Data Reduction
The blazar PKS 0528+134 was the target of intensive, simultaneous and quasi-simultaneousobservations at optical (MDM, GASP), radio (GASP), X-ray (
XMM-Newton, RXTE, Suzaku ),and γ -ray ( Fermi
LAT) frequencies during the period September 8 to 18, 2009. In addition,a more extended period of time, throughout July – December 2009, was covered by lessintensive radio, optical, X-ray, and γ -ray monitoring for longer-term (weeks – months) vari-ability studies. In addition to these previously unpublished data, we included publicallyavailable photometric monitoring data from the Small and Moderate Aperture ResearchTelescope System (SMARTS) at the Cerro Tololo Interamerican Observatory, in our datacollection. Specifically, we included B, V, and R-band data from the Yale Fermi/SMARTSproject, covering the entire, extended campaign period. We also included publically availablepolarimetry and spectroscopy data from the University of Arizona – Steward Observatory Fermi support program in our analysis. Most of our optical (BVR) observations were carried out using the 1.3-m McGraw-Hilltelescope of the MDM Observatory on the south-west ridge of Kitt Peak, Arizona. Thetelescope is equipped with a 1024x1024 pixels CCD camera and standard Johnson-CousinsUBVRI filters. Every night during the period September 9 – 19, 2009, sequences of scienceframes on PKS 0528+134 were taken in the B-V-R filters with exposure times of 360, 180,and 180 seconds, respectively, with slight variations depending on atmospheric conditions.All frames were bias-subtracted and flat-field corrected using standard routines in IRAF. In see http://fermi.gsfc.nasa.gov/ssc/data/access/lat/msl lc/ http://james.as.arizona.edu/ ∼ psmith/Fermi/ z = 2 .
07 and the expectation that the time scale (in the cosmological rest frame of theblazar) of brightness variations in FSRQs is & phot routine within the DAOPHOT package of IRAF. The light curves thus constructedare analyzed in section § . Stan-dard bias subtraction and flat field correction were carried out for each frame. For theseimages we also performed differential photometry with an aperture of radius 6” using BVRmeasurements for comparison stars 3, 2, and 1 from Raiteri et al. (1998). We obtainedI-band measurements for the same comparison stars (14 . ± . . ± . . ± .
010 mag, respectively) using comparison stars from the field of PKS 0735+138(Smith et al. 1985) observed just after PKS 0528+134 on 7 nights. The PRISM camerapossesses a polarimeter with a rotating half wave plate that we employed for R-band po-larimetry. Most details of the polarization observations and data reduction can be foundin Jorstad et al. (2010). The polarization data were corrected for the statistical bias associ-ated with the fact that the degree of polarization is restricted to being a positive quantity(Wardle & Kronberg 1974).We also obtained optical data taken using the polarimetric imaging system Polimaattached to the 84 cm telescope at San Pedro M´artir, Baja California, Mexico as part of along-term polarimetric monitoring program of a sample of 35 blazars during 6 photometricnights in October and November 2009. Due to the optical design of Polima, 4 images withdifferent position angles of the polarimeter to measure the Stokes parameters were takenthrough an R-band filter. The data reduction for these frames (correction for bias and pixel-to-pixel variations across the CCD) as well as the photometry has been carried out on eachof the individual images using a dedicated pipeline developed by D. Hiriart. The fluxesof PKS 0528+134 and several stars in the field of view for each PA have been combined.Flux calibration was finally done via the comparison stars 2 and 3 in the field of view from
10 –Raiteri et al. (1998).
In addition to the photometric observations described in the previous sub-section, PKS 0528+134is also regularly monitored with spectroscopic observations through the University of Arizona– Steward Observatory
Fermi
Support program. As a representative example, we show inFigure 1, the optical spectrum of October 19, 2009, which is within the extended period ofour multiwavelength campaign. The source was clearly still in the quiescent state targetedin this work.The spectrum exhibits two distinct emission lines: CIII] λ λ =5860 ˚A, and CIV λ λ = 4740 ˚A. The total measured flux in the linescorresponds to F CIII ] = 1 . × − erg cm − s − and F CIV = 4 . × − erg cm − s − . Inorder to evaluate the de-absorbed fluxes, we evaluate the extinction coefficients A λ with theextinction law of Cardelli et al. (1989) and A V = 2 .
78, yielding A = 3 .
35 and A =2 .
59. This yields intrinsic luminosities in the two lines of L CIII ] ≈ . × erg s − and L CIV ≈ . × erg s − .The continuum redward of ∼ F λ ∝ λ . , whichcorresponds to a steep spectrum in frequency space, F ν ∝ ν − . . Optical Polarimetry
R-band photo-polarimetric observations of PKS 0528+134 were also acquired with the 2.2 mtelescope of the Calar Alto Observatory in Amer´ıa, Spain, as part of the MAPCAT program.The data were reduced and calibrated as described in Agudo et al. (2011) and Jorstad et al.(2010).St. Petersburg observations were performed at the 70-cm telescope of Crimean Obser-vatory using ST-7 XME photometer-polarimeter. The standard procedure, including biasand dark subtraction, flat-field correction and calibration relative to comparison stars 1, 2and 3 from Raiteri et al. (1998) was applied.Additional polarimetry and spectropolarimetry data were included from the Universityof Arizona – Steward Observatory Fermi
Support program. Those data included two spec- ∼ iagudo/research/MAPCAT
11 – λ [Angstrom]1e-172e-173e-174e-175e-176e-177e-178e-17 F λ [ e r g / ( c m s A )] C IV λ λ λ . Fig. 1.— Optical spectrum of PKS 0528+134 on December 19, 2009, taken at the StewardObservatory. The two prominent emission lines of CIII] and CIV are labeled and correspondto luminosities of L CIII ] ≈ . × erg s − and L CIV ≈ . × erg s − . The continuumred-ward of ∼ F λ ∝ λ . . 12 – P o l a r i za ti on [ % ] Dec. 1, 2008Jan. 29, 2009
Fig. 2.— Wavelength dependence of the degree of polarization of the optical emission ofPKS 0528+134, from two observations in 2008 December and 2009 January. The plot ishighly suggestive of a systematic trend of increasing polarization with increasing wavelength. 13 –tropolarimetric observations of PKS 0528+134 in 2008 December and 2009 January, whichshowed a rather high ( P &
10 %) degree of polarization, allowing for a meaningful spectrap-olarimetric analysis. At those times, the object was in a similar quiescent state as duringour campaign. The contributions of the CIII] and CIV lines (see previous sub-section) wereassumed to be unpolarized, and their fluxes subtracted from the respective energy bins.Figure 2 shows the degree of polarization as a function of wavelength. The plot is highlysuggestive of a systematic increase of the polarization towards longer frequencies. This maybe interpreted as an increasing contribution of synchrotron emission towards the red end ofthe optical spectrum.
PKS 0528+134 was observed by the
XMM-Newton Observatory (Jansen et al. 2001)between September 8th and 14th 2009 on four consecutive revolutions. Table 1 summarizesthese observations. Here, we focus on the data taken with the EPIC detector, covering theenergy band between 0.2 and 10 keV. EPIC observations were taken in full frame mode andthin optical filter (except the pn exposure in 0600121501 that was taken on small windowmode).
XMM-Newton also has an Optical/UV Monitor Telescope (OM) (Mason et al. 2001),a small 30 cm telescope co-aligned with the main XMM-Newton X-ray telescopes. However,due to the optical faintness of PKS 0528+134 ( & th mag), OM observations did not returnuseful data. Obs. ID Date Exp. (pn) CR (pn) Exp. (MOS1) CR (MOS1) Exp. (MOS2) CR (MOS2)yy/mm/dd (ksec) (cts/sec) (ksec) (cts/sec) (ksec) (cts/sec)0600121401 2009-09-08@05:55UT 9.94(14.28) 0.2675 ± ± ± ± ± ± ± ± ± − − − − − − ± ± Table 1: Summary of PKS 0528+134
XMM-Newton
EPIC observations.
Exp. = Livetimeafter (before) correction due to periods of higher background activity. CR = Count rates,given for the 0.2-10 keV energy range. The quadrant of the pn chip containing the source ofthe pn exposure in observation 0600121701 was lost and hence there is no source information. Data Reduction
The data have been analysed using SASv9.0 (Gabriel et al. 2004) and corresponding calibra-tion files. Event and source lists were obtained for the EPIC detector, following the standard 14 –SAS data reduction procedures. Several filtering criteria have been applied. The event listhas been filtered for time periods of high background activity following the standard proce-dure of removing those time periods with background event rates at energies
E >
10 keVhigher than 1.0 cts/sec and 0.8 cts/sec for EPIC-pn and EPIC-MOS, respectively. The timelosses due to the removal of these periods are up to 30 % depending on the observationand the instrument (three out of the four observations were performed at the end of theirrespective revolutions with the consequent increase in radiation levels towards the end of theobservations). Table 1 shows the livetimes for each observation and instrument before andafter this correction has been performed. The data was also filtered to include only single anddouble (PATTERN ≤
4) pattern events for EPIC-pn and single to quadruple (PATTERN ≤ ≤ R ≤ ∼
40” in radius for the source region.
EPIC Spectral Analysis
Time-averaged spectra have been obtained for each individual observation. The spectra werere-binned in order not to oversample the intrinsic energy resolution of the EPIC camerasby a factor larger than 3, while making sure that each spectral channel contains at least 25background-subtracted counts. Both conditions allow the use of the χ quality-of-fit esti-mator to find the best fit model. Fits were performed in the 0.2 – 10 keV energy range.The spectra from all three EPIC cameras have been used simultaneously during the fittingprocedure. The systematic difference between the EPIC cameras is below ∼ H , ), plus an additional absorbing column density (N H , mol ), primarilydue to heavy elements in the intervening molecular cloud B30, which are not properly tracedby the hydrogen column density. The Galactic hydrogen column density is kept fixed duringthe fitting procedure. The spectral fitting model takes the following form for the differentialphoton flux Φ( E ): 15 –Φ( E ) = e − N H , σ ( E ) · e − N H , mol σ ( E ) · N · E − Γ ph (1)where σ ( E ) is the photoelectric absorption cross-section, with abundances after Anders & Grevesse(1989), and Γ ph the power-law index with normalization N . Errors in the relevant parametersare given at the 90 % confidence level (CL) for any given parameter.Table 2 shows the best fit values for the model considered for all four observations.Figure 3 shows an example of the spectra as determined for one of the observations (composedof EPIC-pn and the two EPIC-MOS spectra) with the best fit model and χ deviation.Similar spectra are derived for the other observations. In addition to our four ToO
XMM-Newton pointings, we monitored the 2.4 – 10 keVX-ray flux of PKS 0528+134 with the
Rossi X-ray Timing Explorer (RXTE) ProportionalCounter Array (PCA), with exposure times of 900 – 3400 s for individual observations. TheX-ray flux measurement entailed the subtraction of an X-ray background model (faint sourcemodel, version 20051128) from the raw spectrum using the standard X-ray data analysissoftware packages FTOOLS and XANADU. We used the program XSPEC to fit the residualphoton spectrum with a power-law model plus photoelectric absorption along the line ofsight.
PKS 0528+134 was also observed with the Suzaku satellite as a part of multi-bandobservations conducted in September 2008, about one year prior to the campaign coveredin this paper. However, both the X-ray flux and γ -ray flux levels were comparable to thosemeasured during the September 2009 campaign, indicating a quiescent state. We thereforeinclude the Suzaku data here for comparison of the spectral and short-term variability prop-erties with those measured by
XMM-Newton one year later, in a similar quiescent state. The
Suzaku satellite features instruments sensitive in the soft X-ray band.Suzaku observations of PKS 0528+134 started on 2008 September 27, 02:38 UT, andlasted until 2008 October 2, 16:12 UT. Since the source was not detected in the HardX-Ray Detector (HXD) data, we considered only the X-Ray Imaging Spectrometer (XIS) 16 – −3 C t s / s / k e V PKS0528+134 EPIC Spectrum (0.2−10 keV)10.5 2 5−4−2024 R e s i du a l s ( S t . D e v . ) Energy (keV)
Fig. 3.— Example of EPIC (pn, MOS1 and MOS2) spectrum for the observation id0600121601 where EPIC data is available for all three cameras. The fit has been donein the 0.2-10 keV energy range with channel grouping > >
25 backgroundsubtracted counts. 17 –instruments. The observation conditions were nominal, although the XIS1 data suffered fromunusually high and variable background, resulting in a total apparent background count rateranging from 1 to 3 counts s − over the entire chip. Nonetheless, since the background-subtracted spectrum determined from the XIS1 data was entirely consistent with that fromXIS0 and XIS3, we included the properly background-subtracted XIS1 data in the spectralfitting (see below).The total exposure time yielding good data accumulated in the pointing was 203 ks.We used the standard ftools data reduction package, provided by the Suzaku ScienceOperations Center, with the calibration files included in the CALDB ver. 4.3.1. The netcount rates were, respectively, 0.10, 0.13, and 0.12 count s − for XIS0, XIS1, and XIS3. Forthe analysis of spectra and light curves, we extracted the counts from a region correspondingto a circle with 260 arc sec radius. We used a region of a comparable size from the same chipto extract the background counts. We plot the resulting total light curve (not background-subtracted) from XIS3, binned in 5400 sec bins, in Figure 4, and note that the backgroundwas steady at the level of ∼ .
015 count s − . The data indicate no significant rapid (a day orless) variability during the Suzaku observation, but show a long-term trend , with a ∼
20 %decrease in flux over the ∼ . XSPEC spectral analysis software (Arnaud1996). For the spectral fitting, we used the standard redistribution files and mirror effec-tive areas generated with
Suzaku -specific tools. We included the counts corresponding tothe energy range of 0.5 – 10.0 keV in our spectral fits. We used all three XIS detectorssimultaneously, but allowed for a small (a few %) variation of the relative normalizations.The source spectrum was modeled as an absorbed power law, with the cross-sections andelemental abundances as given in Morrison & McCammon (1983). Other absorption modelsgive similar results. The best-fit absorbing column was 4 . ± . × cm − , and the photonindex was 1 . ± .
03, with the acceptable best-fit χ of 3407 for 3512 d.o.f. The Suzaku spec-trum is shown in Figure 5. The observed model 2 – 10 keV flux is 2 . × − erg cm − s − ,with the statistical error of 5 %, which is probably smaller than the systematic error resultingfrom the calibration uncertainty of the Suzaku instruments.The fitted values of the total absorbing column are only marginally consistent betweenthe
Suzaku and
XMM-Newton data sets. This might be in part related to the differencesin calibration of the two instruments. An additional source for the discrepancy might bethe fact that (1) we are using Galactic elemental abundances, which might or might notbe appropriate for the line-of-sight molecular cloud, and (2) the bandpass of
XMM-Newton extends down to 0.3 keV, while for Suzaku it is 0.5 keV. If the elemental abundances in 18 – . . . . R A TE c oun t/ s TIME sPKS 0528+134 Suzaku XIS3 observation
Fig. 4.—
Suzaku
X-ray light curve of PKS 0528+134 from the XIS3 data of 2008 September27 – October 2. t = 0 corresponds to MJD 54736.1771. 19 – − − − . e V ( P ho t on s c m − s − k e V − ) Energy (keV)Suzaku XIS0, XIS1 and XIS3 spectrum of PKS 0528+134, Sep 2008
Fig. 5.— 0.5 – 10 keV X-ray spectrum of PKS 0528+134 from the
Suzaku observations of2008 September 27 – October 2. 20 –the intervening absorber of low-z elements are not the same in the molecular cloud as theGalactic abundances assumed in the spectral fit, the precise value of the best-fit N H mightbe dependent on the bandpass of the instrument and even the details of the effective area atthe energies where the absorption edges due to those elements play an important role.While the X-ray spectral index measured by Suzaku in 2008 September/October isconsistent with the value measured one year later by
XMM-Newton (see § Suzaku
XMM-Newton value. This is inagreement with the ∼ factor 2 – 3 variability measured by RXTE on ∼ weekly time scales(see § γ -Ray Observations Gamma-ray observations of PKS 0528+134 were performed by the
Fermi -LAT. This isa pair-conversion γ -ray telescope sensitive to photon energies greater than 20 MeV. In itsnominal scanning mode, it surveys the whole sky every 3 hours with a field of view of about2.4 sr (Atwood et al. 2009). The data presented in this paper (restricted to the 100 MeV– 200 GeV range) were collected from MJD 54983 (2009 June 1st) to MJD 55193 (2009December 28). For this analysis, the Diffuse event class was used. This is the optimizedclass for point source analysis with minimal residual contamination from charged-particlebackgrounds. To minimize systematics, only photons with energies greater than 100 MeVwere considered in this analysis. In order to limit the contamination from atmospheric γ -raysproduced by interactions of cosmic rays with the upper atmosphere of the Earth, only eventswith zenith angle < ◦ were selected. In addition, time intervals during which the rockingangle was larger than 52 ◦ have been excluded from the analysis, because the bright limb ofthe Earth enters the field of view. The analysis was performed with the standard analysistool gtlike , part of the Fermi-LAT Science Tools software package (version v9r15p5) . TheP6 V3 DIFFUSE set of instrument response functions was applied.Photons were selected in a circular Region Of Interest (ROI) 7 ◦ in radius, centered atthe position of PKS 0528+134. The isotropic background, including the sum of residualinstrumental background and extragalactic diffuse γ -ray background, was modeled by fittingthis component at high galactic latitude (file provided with Science Tools). The Galacticand Isotropic diffuse emission models version “gll iem v02.fit” and “isotropic iem v02.txt” For a documentation of the Science Tools see http://fermi.gsfc.nasa.gov/ssc/data/analysis/documentation/
21 –were used. All point sources in the first Fermi-LAT catalog (Abdo et al. 2010c) lyingwithin the ROI and a surrounding 10 ◦ -wide annulus were considered in the fit and modeledwith single power-law distributions of the form F ( E ) = N ( E/E ) − Γ . Their flux was keptfree whereas their spectral index value was frozen to the value listed in the 1FGL catalog,except for PKS 0528+134 whose index was kept free. Due to limited statistics, the gamma-ray spectrum of PKS 0528+134 was modeled with a power-law distribution in the presentanalysis, although the spectrum measured over 11 months covered by the 1FGL catalogexhibits distinct curvature, characterized by a curvature index of 8.14, corresponding to a4 % probability that the spectrum is adequately represented by a power law (Abdo et al.2010a). The highest energy photon attributed to PKS 0528+134 has an energy of 8.6 GeV.The estimated systematic uncertainty on the flux is 10 % at 100 MeV, 5 % at 500 MeV and20 % at 10 GeV. We obtained radio observations from the GLAST-AGILE Support Program (GASP) ofthe Whole Earth Blazar Telescope (WEBT). In table 3, we list the frequencies in which wecollected data and the observatories participating in the campaign.Centimeter-band observations were obtained with the University of Michigan 26-meterprime focus paraboloid equipped with radiometers operating at central frequencies of 4.8,8.0, and 14.5 GHz. Observations at all three frequencies utilized rotating polarimeter sys-tems permitting both total flux density and linear polarization to be measured. A typicalmeasurement consisted of 8 to 16 individual pointings over a 20 – 40 minute time period.Frequent drift scans were made across stronger sources to verify the telescope pointing cor-rection curves. Observations of program sources were intermixed with observations of a gridof calibrator sources to correct for temporal changes in the antenna aperture efficiency. Theflux scale was based on observations of Cassiopeia A (see, e,g., Baars et al. 1977). Details ofthe calibration and analysis techniques are described in Aller et al. (1985).The 37 GHz observations were made with the 13.7 m diameter Mets¨ahovi radio telescope.The detection limit of the telescope at 37 GHz is of the order of 0.2 Jy under optimalconditions. Data points with a SNR < For details on the background model see http://fermi.gsfc.nasa.gov/ssc/data/access/lat/BackgroundModels.html
22 –
EPIC Power LawPHABS+PHABSObs ID N H , mol Γ ph F . − F − χ /d.o.f. cm − − ergs cm − s − − ergs cm − s − +0 . − . +0 . − . +0 . − . (0.88) 1.35 +0 . − . (1.39) 0.93 (146.2/157)0600121501 0.13 +0 . − . +0 . − . +0 . − . (0.76) 1.10 +0 . − . (1.14) 1.02 (243.3/238)0600121601 0.17 +0 . − . +0 . − . +0 . − . (0.81) 1.14 +0 . − . (1.18) 1.05 (275.2/261)0600121701 0.12 +0 . − . +0 . − . +0 . − . (0.69) 1.18 +0 . − . (1.22) 0.99 (135.6/137) Table 2: Fit performed in the 0.2 – 10 keV energy range. All 3 EPIC instruments havebeen used simultaneously in the fit. The Galactic absorption traced by 21-cm emission fromneutral hydrogen, N H, cm has been fixed to a value of 0 . · cm − according to theLAB Survey of Galactic HI Kalberla et al. (2005). Fluxes given are not absorption corrected(in parenthesis the de-absorbed flux is given). Errors indicate the 90 % CL. NOTE : Forobservation id 0600121701 only MOS1 and MOS2 are available. In this case, the fit has beenperformed in the 0.3 – 10 keV energy range and the fluxes given are between 0.3 – 2.0 and2.0 – 10 keV.Table 3. Frequencies and observatories used to collect radio data
Frequency Observatory05 GHz UMRAO, MEDICINA08 GHz UMRAO, MEDICINA14 GHz UMRAO22 GHz MEDICINA37 GHz Metshovi Radio Observatory (KURP-GIX)43 GHz Noto230 GHz MAUNA KEA (SMA)345 GHz MAUNA KEA (SMA)
23 –from the measurement rms and the uncertainty of the absolute calibration.The 43 GHz Noto observations have been performed using the On The Fly (OTF) scantechnique (scan duration about 20 s) and the telescope gain (K/Jy) was determined as afunction of the elevation using NGC 7027 as primary calibrator. To improve the signal tonoise ratio many scans have been acquired and then averaged for a total integration timeof about 15 minutes. The antenna temperature has been estimated by a gaussian fit of theaverage scan. A more detailed description for the 43 GHz data is given by Leto et al. (2009).Radio observations at the Medicina radio observatory were performed at 5, 8, and 22 GHz,and were analyzed as detailed in Bach et al. (2007).Data at 230 GHz and 345 GHz were collected at the Submillimeter Array (SMA) onMauna Kea and reduced using the MIR data reduction software. SMA flux density measure-ments are produced from a mixture of dedicated flux calibration/monitoring observationsand data from science projects that may utilize a quasar as a calibration standard (typicallyfor gain calibration of the interferometer). Raw visibility data are calibrated to correct foratmospheric absorption and instrumental gain variations, and then referenced to observa-tions of standard flux calibration sources, generally planets and/or moons. More details onthe SMA flux density monitoring program can be found in Gurwell et al. (2007).
4. Variability Analysis
One of the main goals of this multiwavelength campaign was the search for flux andspectral variability of PKS 0528+134 in its quiescent state. We first describe our results onflux variability in § § § The optical light curves of PKS 0528+134 obtained with the 1.3-m McGraw-Hill tele-scope of the MDM Observatory during the core of our multiwavelength campaign (September9 – September 19, 2009), along with the publically available SMARTS data from the sameperiod, are plotted in figure 6. An increase in the brightness of the source of ∼ χ for a fit to a constant flux. Variabil-ity is evident in all three bands with χ ν = 15 .
9, 5 .
2, and 26 .
2, respectively, for the R, V,and B bands. Due to the limited observability period of PKS 0528+134 in any given night 24 –(typically . t optvar . ∼ XMM-Newton observations as well on time scales of a few hours, withinthe individual observations. The
XMM-Newton light curve is plotted in the bottom panel offigure 6. No flux variability is evident. For the 0.2 – 10 keV X-ray flux, a fit to a constantresults in a χ ν of 0.91, thus confirming the absence of significant variability. For the same XMM-Newton observations, we also analyze the intraday variability. For most observationsthe χ ν for a fit to a constant flux is less than 1. In figure 7, we show the light curve fromthe XMM-Newton
MOS1 data of September 10, 2009, as an example. For this particularobservation, χ ν ≈ .
81 is found.A multiwavelength variability analysis for a more extended period of time was madepossible by including RXTE monitoring data. In figure 8, we show the light curves ofPKS 0528+134 in radio (5 and 37 GHz), optical, X-rays (RXTE), and γ -rays (bottom totop). The vertical shaded band highlights the interval of the core campaign where themost intense optical and X-ray observations (including XMM-Newton ) were performed. TheRXTE X-ray data from the period July 14 to December 2, 2009, represent the most extendedcoverage. In this interval, PKS 0528+134 shows significant variability corresponding to areduced χ ν of 3.83 for a fit to a constant flux. The RXTE light curve indicates variabilitywith flux changes of | ∆ F/F | ∼
50 % on a characteristic time scale of t X var ∼ RXTE analysis are systematically higherthan those measured by
XMM-Newton during the same period. We have carefully double-and cross-checked both analyses and confirmed this discrepancy. The ROSAT All Sky SurveyCatalogue does not list any known source in the field of view of
RXTE which may beresponsible for the flux discrepancy. Furthermore, three
RXTE observations were carriedout within a few hours of one of the
XMM-Newton observations. Given this short timeperiod and the systematic offset between the two instruments, rapid variability appearsunlikely to cause the flux discrepancy. The factor of 4 discrepancy seems also too large tobe solely due to calibration uncertainties. A plausible explanation for the systematic offsetcould lie in an uncertain background model for the
RXTE analysis in the region aroundPKS 0528+134, near the Galactic anti-center. While this would cause a constant offsetof the
RXTE flux levels, it will not affect the variability. We are therefore confident thatthe
RXTE flux variability analysis presented here is robust. For the analysis of the spectralenergy distributions in §
6, the flux and spectral information from
XMM-Newton will be used.The optical light curve (second panel from the bottom of Figure 8), including data 25 – V MDMSMARTS R MDMSMARTS B SMARTSMDM
82 83 84 85 86 87 88 89 90 91 92 93 94JD - 2455000 F l u x [ er g c m - s - ] Fig. 6.— MDM and SMARTS optical light curves (RVB) of PKS 0528+134 between Septem-ber 9 and 19, 2009. The bottom panel shows the X-ray light curve corresponding to the four
XMM-Newton observations. 26 – C oun t R a t e [ s - ] Fig. 7.—
XMM-Newton
MOS1 light curve of our ToO observation on September 10, 2009.Each point corresponds to a bin of 500 seconds. No evidence for intra-day variability isfound. 27 –from the MDM, SMARTS, Perkins, San Pedro M´artir, Calar Alto, and Crimean observato-ries, covers the interval from September 9 to November 19, 2009. This is the band wherePKS 0528+134 shows the most significant variations in its flux density, with flaring episodesexhibiting brightness changes of ∆ R . mag on a time scale of . γ -ray flux does not present significantvariability during the extended campaign period (2009 July – December). The best fit foundfor the flux (E >
100 MeV) is (0 . ± . × − ph cm − s − with a test statistic TS = 68.The Test Statistic (Mattox et al. 1996) is defined as twice the difference in log(likelihood)obtained by including the source of interest and omitting it in the source model used inthe gtlike analysis. This flux is slightly lower than the mean flux observed over the firsteleven months of Fermi data (Abdo et al. 2010a), which corresponds to (0 . ± . × − ph cm − s − as represented by the dashed horizontal line in the same panel.Among the frequencies monitored in the radio regime, the 5 GHz and 37 GHz lightcurves had the most extended coverage in time. In the bottom panel of figure 8, we showboth light curves. Only moderate variability with amplitudes of | ∆ F/F | .
20 % is observedat these frequencies. In particular, a decreasing flux tendency is found at the end of the37 GHz light curve. A fit to a constant flux results in χ ν = 3 .
92 and 2.61 for the 5 GHz and37 GHz light curves, respectively.A more detailed analysis of the radio regime, including the light curves in all the mon-itored radio frequencies, is presented in figure 9. The radio light curves show moderatevariability in general. The 8 GHz and 14 GHz bands present significant flux variations with χ ν = 32 .
96 and 10.17, respectively, for a fit to a constant flux. The results of a similaranalysis for all radio frequencies are summarized in Table 4. Variability appears to occur ontime scales of t radiovar ∼ χ ν for fits to a constant flux. Frequency χ ν
28 – F X [ e r g c m - s - ] R M a gn it ud e San PedroMDMPerkinsCalar AltoCrimeanSMARTS F γ [ - ph c m - s - ] Fermi LAT (E > 100 MeV)11 months mean F l ux [ J y ] Fig. 8.— Light curves of PKS 0528+134. From top to bottom: a) the Fermi γ -ray fluxin 15-day integration bins. b) X-ray (RXTE) light curve. c) Optical (R-magnitude) lightcurve including data from the MDM, SMARTS, Perkins, San Pedro M´artir, Calar Alto, andCrimean observatories. d) Radio light curves at 5 and 37 GHz. 29 –
40 50 60 70 80 90 100 110 120 130JD - 245500001234 F l ux [ J y ] Fig. 9.— Radio light curves of PKS 0528+134. The vertical yellow shaded band correspondsto the interval when MDM optical and
XMM-Newton
X-ray observations of this source wereperformed. 30 –frequencies from γ -rays through radio indicates no significant variability in the γ -ray regime,moderate variability in the X-rays ( | ∆ F/F | ∼
50 %) and most radio frequencies ( | ∆ F/F | .
20 %) on time scales of ∼ R . mag in the optical bandson time scales of several hours. Our data are not sampled densely enough for a meaningfulanalysis of time lags among different bands. In order to test whether the optical flux variability discussed in the previous section isassociated with spectral changes, we evaluated color indices B - R and V - R as a measure ofspectral hardness. These were calculated for any pair of B (V) and R magnitudes measuredquasi-simultaneously, i.e., during the same night, by MDM and SMARTS. If conditions weregood enough to extract more than one high-quality (error in magnitude < .
1) V or B and Rband data point per night, magnitude measurements taken within < r = − .
50, for the V - R color vs. R magnitude correlation, and r = − . P ( < r ) = 5 × − and P ( < r ) = 2 × − ,respectively, of obtaining a correlation coefficient more negative than the ones resulting fromour observational data. This seems to provide strong evidence for the presence of a redder-when-brighter trend in PKS 0528+134.Such a redder-when-brighter trend has been observed in other FSRQs, e.g., 3C 454.3(Raiteri et al. 2008), where it has been partially attributed to a contribution of emissionlines from the BLR to the B band. In order to test whether this may also be the cause ofthe color variability described above, we utilize the line fluxes inferred from the spectrumshown in Figure 1. We note that the red-shifted CIV line falls within the B-band range. At 31 – B - R MDMSMARTSMDM (line corrected)SMARTS (line corrected) V - R MDMSMARTSMDM (line corrected)SMARTS (line corrected)
Fig. 10.— Color-magnitude diagrams for B - R (top panel) and V - R (bottom panel) vs.R-band magnitude. Open symbols with dashed error bars indicate the original data; filledsymbols with solid error bars show the data data after correction of all magnitudes for thecontributions from the CIII] and CIV emission lines. 32 – λ = 4740 ˚A, the B-band filter has a transmission coefficient of approximately f λ = 70 % ofits maximum value. Hence, using the bandwidth of the B filter of ∆ ν = 1 . × Hz, wefind that the CIV line will make an effective contribution of F Bν,CIV = F CIV f λ / ∆ ν ≈ . µ Jyto the B-band flux. An analogous calculation of the contribution from the CIII] line to theV and R bands yields F Vν,CIII ] ≈ . µ Jy and F Rν,CIII ] ≈ . µ Jy. We corrected all B, V,and R magnitude values for these line contributions and re-evaluated the color magnitudecorrelations. The resulting points are shown as solid symbols in Figure 10. This leads to anoverall slight shift towards fainter (larger R) magnitudes and redder B - R colors. However,the overall color variability trend and its significance remain unaffected by this correction.We also analyzed the spectral variability in the X-ray regime. For the XMM-Newtondata, we analyzed the variability of the hardness ratio F − keV /F . − keV . Figure 11 illus-trates that the hardness ratio is consistent with being constant over the four XMM-Newton observations performed as part of the core campaign. The RXTE X-ray energy index α wasanalyzed over a period of 150 days as plotted in the top panel of figure 12. A fit to a constantresults in χ ν = 1 .
31 is found, indicating very moderate spectral variability. The comparisonof the
XMM-Newton spectra with the
Suzaku spectrum from 2008 September/October (see § ∼ γ -ray spectral index variability over the extended campaignperiod. As shown in the bottom panel of figure 12, no significant variability is found. Thebest fit found for the spectral index in this period of time is Γ = 2 . ± .
2, which is slightlysofter than the mean spectral index (Γ = 2 . ± .
06) observed over the first eleven monthsof Fermi data (dashed horizontal line Abdo et al. 2010a).
The polarization variability analysis on PKS 0528+134 was performed using the R-band polarimetric observations at the 1.8 m Perkins telescope of Lowell observatory duringa twelve day period from October 15 to 26, 2009, and two measurements from the Universityof Arizona / Steward Observatory
Fermi support program from the same period. As shown infigure 13, the Perkins polarimetry data indicate strong variability of the degree of polarization( χ ν = 20 . ∼ F - k e V [ J y H z ] F . - k e V [ J y H z ]
82 83 84 85 86 87 88 89 90JD - 2455000456 H a r dn e ss R a ti o F /F Fig. 11.— Light curves in soft (0.2 – 2 keV) X-rays (top panel), hard (2 – 10 keV) X-rays(middle), and the hardness ratio. No significant spectral variability was found. 34 – γ -r a y S p ec t r a l I nd e x Mean index over 11 months0.00.51.01.52.02.53.0 X -r a y E n e r gy I nd e x α Fig. 12.— Top panel: X-ray (2 – 10 keV) energy spectral index α vs. time. Bottom: Fermi γ -ray photon spectral index (Γ ph = α + 1) vs. time. The dashed horizontal line correspondsto an index of 2.64, the mean index observed over the first eleven months of Fermi data. 35 – R M a gn it ud e E V P A ( D e g r ee s ) P o l a r i za ti on ( % ) Perkins (R)Steward (500 - 700 nm)
Fig. 13.— Simultaneous photometric and polarimetric data for PKS 0528+134 on sevendays during October 15 – 26, 2009. Top panel: R-band magnitude. Middle panel: Degree ofpolarization. Bottom panel: Electric vector position angle (EVPA). While there is significantvariability in the R-band flux, degree of polarization, and EVPA, we did not find a significantcorrelation between polarization and flux states. 36 –stantial synchrotron contribution to the emission at these wavelengths. We will discussimplications of our polarization results in Section 7.
5. Structure of the Parsec Scale Jet
The quasar PKS 0528+134 is monitored monthly by the Boston University (BU) groupwith the Very Long Baseline Array (VLBA) at 43 GHz within a sample of bright γ -rayblazars . The source was also included in a 2-week campaign of observations of 12 γ -ray blazars organized in 2009 October when 3 additional VLBA epochs at 43 GHz wereobtained (VLBA project S2053). Figure 14 shows the total and polarized intensity imagesof the quasar in 2009 Autumn. The VLBA data were calibrated, imaged, and modeled inthe same manner as discussed in Jorstad et al. (2005). As we did for the optical data, thepolarization parameters were evaluated taking into account the statistical bias as detailedin Wardle & Kronberg (1974). Table 5 gives the parameters (flux, position, size, degree andposition angle of polarization) of the main features seen in the radio jet during this period.Figure 15 shows the light curve of the VLBI core at 43 GHz of the quasar over the lastthree years as monitored by the BU group. According to Figure 15 the parsesc scale jet ofPKS 0528+134 was in a quiescent state in 2009 Autumn. The core was moderately polarized, P ∼ ◦ and 50 ◦ thataligns within 10 ◦ with the jet direction as determined by the position angle of the brightestknot C C
1, has the highest levelof polarization, P ∼
10 %, with the position of polarization perpendicular to the local jetdirection. Although we do not observe superluminal knots in 2009 Autumn, the quasar isknown to have a very high apparent speed in the jet, implying a high bulk Lorentz factor,Γ ∼
30 (Jorstad et al. 2005).
6. Spectral Energy Distributions (SEDs) Modeling
In figure 16, we present the SEDs of PKS 0528+134 from radio to γ -rays (blue filledcircles) corresponding to the four XMM-Newton observations (September 8, 10, 11, 14).The optical data have been dereddened assuming A V =2.782, A B = 3 .
62 and A R = 2 . according to the sky dust map given
37 –Table 5.
Parameters of Jet Components
Epoch Knot S [Jy] R [mas] Θ [ ◦ ] a [mas] P [%] χ [ ◦ ](1) (2) (3) (4) (5) (6) (7) (8)16 Sep A ± · · · ± ± C ± ± ± ± − ± C ± ± ± < · · ·
14 Oct A ± · · · ± ± C ± ± ± ± ± C ± ± ± ± ±
516 Oct A ± · · · ± ± C ± ± ± ± − ± C ± ± ± ± ±
820 Oct A ± · · · ± ± C ± ± ± < · · · C ± ± ± ± ±
628 Nov A ± · · · < · · · C ± ± ± < · · · C ± ± ± < · · · Note. — Columns: 1 - epoch of the observation; 2 - component designation; 3 - flux of component; 4- distance of component from the VLBI core; 5 - position angle of component with respect to the core;6 - diameter of component; 7 - degree of polarization of component; 8 - position angle of polarization ofcomponent
38 –Fig. 14.— 43 GHz total ( contours ) and polarized ( color scale ) intensity im-ages of PKS 0528+134 during 2009 Autumn. The highest contour corresponds to S peak =850 mJy/beam, while the yellow color indicates the highest polarized flux of S ppeak =30 mJy/beam, for a beam of 0 . × .
15 mas at PA=-10 ◦ . Total intensity contours corre-spond to 0.25, 0.5, ..., 64 % of the peak. Sticks over the polarized intensity contours indicatethe plane of polarization. The designation of components corresponds to Table 5. 39 – Fig. 15.— Light curve of the VLBI core at 43 GHz. The vertical lines indicate the period ofintensified monitoring in 2009 Autumn. 40 –by Schlegel et al. (1998). We found in sections 4.1 and 4.2 that the Fermi γ -ray flux andspectral index of PKS 0528+134 did not show significant variability in the interval includingthe core campaign. Accordingly, the same γ -ray spectrum was used in the four SEDs. Asin its flare states, in this quiescent state the SED of PKS 0528+134 is characterized bytwo peaks, a low-energy peak between the far infrared and optical spectral bands, and ahigh-energy peak at MeV – GeV energies. As can be seen, in all the SEDs the high-energycomponent dominates the bolometric output by a large amount.We produce model fits to all four SEDs using the equilibrium version of the leptonic one-zone model developed by B¨ottcher & Chiang (2002). This equilibrium model is described inmore detail in Acciari et al. (2009), and we here summarize its main features. The observedelectromagnetic radiation is interpreted as originating from ultrarelativistic electrons (andpositrons) in a spherical emission region of co-moving radius R , which is moving with arelativistic speed β Γ c , corresponding to the bulk Lorentz factor Γ. Depending on the viewingangle θ between the jet direction and the line of sight, the transformations of photon energiesand fluxes is characterized by the Doppler factor D = (Γ[1 − β Γ cos θ ]) − . The size of theemission region is constrained by the shortest observed variability time scale δt var , min through R ≤ cδt var , min D/ (1 + z ) . × ( δt var , min / d) ( D/
10) cm.Ultrarelativistic electrons are assumed to be instantaneously accelerated at a height z above the accretion disk into a power-law distribution in electron energy, E e = γm e c , at arate per unit volume and unit Lorentz factor interval given by Q ( γ ) = Q γ − q with a low- andhigh-energy cutoffs γ and γ , respectively, and injection spectral index q . An equilibriumbetween this particle injection, radiative cooling, and escape of particles from the emissionregion yields a temporary equilibrium state described by a broken power-law. The timescale for particle escape is parameterized through an escape time scale parameter η esc > t esc = η esc R/c . The balance between escape and radiative cooling will lead to a breakin the equilibrium particle distribution at a break Lorentz factor γ b , where t esc = t cool ( γ ).The cooling time scale t cool is evaluated self-consistently taking into account synchrotron,synchrotron-self-Compton (SSC) and external Compton (EC) cooling. The number densityof injected particles is normalized to the resulting power L e in ultrarelativistic electronspropagating along the jet. The magnetic field B in the emission region is pre-specified as afree parameter. It corresponds to a Poynting flux along the jet, L B = πR Γ β Γ c u ′ B where u ′ B = B / (8 π ) is the magnetic field energy density in the co-moving frame. For each modelcalculation, the resulting equipartition parameter, e B = L B /L e is evaluated.Once the quasi-equilibrium particle distribution in the emission region is calculated, ourcode evaluates the radiative output from synchrotron emission, SSC, and EC emission self-consistently with the radiative cooling rates. If the occasional indication of a blue bump in 41 –the optical spectrum can be associated with the accretion disk, we can estimate an accretiondisk luminosity from the corresponding approximate νF ν flux in the UV regime of νF disk ν ∼ × Jy Hz as L D ∼ . × erg s − , which we use for our model fits. The accretion diskemission is modelled as a multi-color blackbody spectrum according to a Shakura & Sunyaev(1973) disk model.In addition to direct accretion disk emission, external radiation may originate as lineemission from the Broad Line Region (BLR). We can estimate the total luminosity of theBLR line emission using the bright quasar template of Francis et al. (1991), normalizedto the observed value of the CIV emission line luminosity of L CIV = 2 . × erg s − .Substantial contributions to the BLR luminosity, in addition to CIII] and CIV observed here,are expected to arise from Fe II, Ly α and Ly β , H β and H γ , Mg II, and He II, among others.The total BLR luminosity is expected to be L BLR ≈ . L CIV ≈ . × erg s − . This is anorder of magnitude lower than the value we adopt for the accretion disk luminosity. However,the relevance to external Compton scattering depends on the photon field energy density inthe rest-frame of the emission region and hence on the geometry and the Doppler boostingof external photons into the co-moving frame of the emission region. The energy density ofthe direct disk emission in the AGN frame is u disk = L D / (4 πz c ) where z is the distanceof the emission region from the central supermassive black hole. If z ≫ R D Γ , where R D is the characteristic radius of the annulus of maximum energy output of the accretiondisk, accretion disk photons will be strongly red-shifted in the emission-region rest frame.However, in the near-field regime, z ≪ R D Γ , even the accretion disk photons enteringthe emission region from behind in the AGN rest frame will still be blue-shifted and theirenergy density enhanced in the emission region rest frame (see, e.g., Dermer & Schlickeiser1993). If the emission region is located within the characteristic radius of the BLR, R BLR ,the BLR radiation field can be treated as approximately isotropic, and it will be blue-shiftedinto the emission-region rest frame, and enhanced by a factor of ≈ Γ compared to its AGNrest-frame value of u BLR = L BLR / (4 πR BLR c ) (Sikora et al. 1994).Lacking knowledge of the size of the BLR and the precise location of the emissionregion in PKS 0528+134, both direct disk and BLR emission contributions remain plausibleas dominant sources of external photons for the external-Compton process. In the following,we have chosen our model parameters (in particular, z ) such that we expect the direct diskemission to dominate. As we will see below, this allows for satisfactory fits to the SED and,in particular, the Fermi -LAT γ -ray emission from PKS 0528+134. Therefore, in order toavoid the introduction of another unconstrained parameter, R BLR , we restrict our modelingefforts to using the direct accretion disk emission as the dominant contributor to the externalradiation field to evaluate the external radiation Compton (ERC) component of the SED. 42 – Frequency Hz10 ν F ν [ J y H z ] Fig. 16.— Spectral energy distribution (SEDs) of PKS 0528+134 corresponding the firstXMM-Newton observation. Blue points correspond to the observational data. The radiodata correspond to the integrated radio flux from single-dish measurements. The orangeradio point at 43 GHz is the flux of the VLBA core component. The continuous linescorrespond to the best fit for each SED. The dotted line represents the synchrotron spectrum,the dot-dot-dashed line corresponds to the Synchrotron self Compton (SSC), the dot-dashedline is the external radiation Compton (ERC), and the short-dashed line is the disk thermalcomponent. 43 –As a representative example, the fit to the SED of PKS 0528+134 during the first
XMM-Newton observation is shown in Figure 16. The parameters used for this particularfit are shown in table 6. The synchrotron (dotted line), synchrotron self Compton (SSC)(dot-dot-dashed line), external radiation Compton (ERC) (dot-dashed line), and the directdisk emission (short-dashed line) components are shown separately, in addition to the totalSED fit curve.We were able to achieve good fits for all four SEDs of PKS 0528+134 in quiescence.Table 7 lists the most relevant parameters used in each fit. Fit parameters close to equipar-tition could be found for all four SEDs. In general, no obvious correlation between thedifferent parameters was found. However, a strong correlation (Pearson’s correlation coeffi-cient r ≈
1) was found between the magnetic field and the optical flux (R-band), reflectingthe synchrotron dominance in the optical band.We point out that our model only includes the emission from the blazar zone, assumedhere to be on sub-pc scales. It is expected that the radio emission originates in the moreextended (pc to kpc scales) jet which is not included in our model. Therefore, our fits under-produce the radio spectra in all SEDs. While most of the radio data shown in Figure 16are obtained by single-dish instruments and therefore represent the integrated flux over allradio components, we have also included the 43 GHz flux from the VLBA core componenton 16 September 2009, just 2 days after the last
XMM-Newton observation. Even this coreradio flux is under-represented by our model, suggesting that the higher-frequency emissionoriginates on even smaller scales than the radio core.
7. Discussion
As mentioned in the previous section, good fits with a one-zone leptonic SSC + ERC jetmodel were possible with parameters close to equipartition. Our fit parameters, in particular,the magnetic fields of B ∼ γ min ∼ , and jetpowers of L e ∼ erg s − , are in rough agreement with the fit results of Mukherjee et al.(1999). However, we need to point out that the model used in Mukherjee et al. (1999) is notprecisely the same as used in this paper, as it was based on a time-average of an evolvingparticle distribution along the jet. Also, Mukherjee et al. (1999) used a different cosmology(a matter-dominated Universe with q = 0 . Fermi
LAT γ -ray spectra of PKS 0528+134 are systematically softer than the photon indices Γ ph ∼ .
2– 2.6 found during the EGRET era (Mukherjee et al. 1999). Therefore, our fits to the LATSEDs require significantly steeper particle spectral indices. 44 –Table 6. Parameters used in the fit for the SED of PKS 0528+134 corresponding to theXMM-Newton observation of September 8, 2009 (JD 2455082.94).
Parameter Value γ min γ max Injection electron spectral index 3.65Escape time parameter ( t esc = η esc R/c ) η esc = 50Magnetic field [G] 2.05Injection height [pc] z = 0 . erg/s] L = 17Blob radius [cm] 1 . × Black hole mass [ M ⊙ ] 1 . × Observing angle [degrees] θ obs = 3 . D = 19 . L e (jet) [erg/s] 2 . × L B (jet) [erg/s] 2 . × L B /L e Table 7. Relevant fit parameters for the SEDs of PKS 0528+134
SED B [G] F R [Jy Hz] F X [erg cm − s − ] Γ q η L e L B L B /L e z [pc]1 2.05 1.8399e+11 1.6249e-12 20.4 3.65 50 2.30e+45 2.30e+45 1.0 0.132 2.06 1.6570e+11 1.3224e-12 20.1 3.85 50 2.39e+45 2.26e+45 0.95 0.133 3.7 3.8000e+11 1.3702e-12 17.5 3.7 70 1.18e+45 1.73e+45 1.46 0.114 2.21 1.9799e+11 1.4367e-12 21.1 3.65 40.3 2.24e+45 2.40e+45 1.07 0.135
45 –We also need to caution that the model contains a large number of poorly constrainedparameters, and in many cases, different parameter combinations might be able to producesimilarly acceptable fits. Therefore, conclusions about correlations between model parame-ters and observables may not be unique.A similar study of γ -ray bright blazars in quiescence has recently been published byAbdo et al. (2010b). Those authors observed 5 blazars (PKS 0208-512, Q 0827+243, PKS 1127-145, PKS1510-089, and 3C 454.3) in their low-activity state with Suzaku and Swift in X-raysand optical/UV, and analyzed the simultaneous
Fermi -LAT data. All of those blazars showedX-ray continua consistent with a hard (Γ ph ∼ .
5) single power-law spectrum, in agreementwith our
XMM-Newton and
Suzaku results on PKS 0528+134. The broadband SEDs of allfive blazars were dominated by their high-energy ( γ -ray) output, as in PKS 0528+134. Incontrast to our XMM-Newton results on PKS 0528+134, three of the five blazars observed byAbdo et al. (2010b) (PKS 0208-512, PKS 1127-145, and PKS 1510-089) did show significantshort-term variability on time scales of ∼ Fermi -LAT γ -ray flux remaining belowthe lowest EGRET flux or upper limit persistently for at least two weeks, see § γ -ray flux and even γ -ray flares within just a few days of the Suzaku observations.Therefore, the X-ray variability reported in Abdo et al. (2010b) may not be characteristic ofa truly quiescent state of those blazars as investigated in the case of PKS 0528+134 reportedin this paper.In our analysis of optical spectral variability we found a weak anti-correlation betweenthe B-R color and the R-band magnitude. A possible explanation for this softer when brightertrend can be found in the interplay between the two radiation components that contributephotons to the optical flux: the synchrotron component, which is generated in the jet itself,with a steep spectrum, always dominating at low (R-band) frequencies, and slowly varyingemission components associated with the accretion disk and the BLR. The direct accretiondisk emission is expected to peak in the ultraviolet and may therefore contribute at the blueend of the optical spectrum. In Section 4.2, we have shown that the observed color variabilitycan not be attributed to the contribution from the CIII] and CIV emission lines in the opticalspectrum of PKS 0528+134. Therefore, the obvious candidate is direct, thermal accretiondisk emission. The suggestive trend of increasing polarization with increasing wavelengthlends further support to this hypothesis.Given the evidence we found for a substantial accretion disk with a luminosity of L D ∼ . × erg s − , it is worth investigating whether at least part of the observed X-ray 46 –emission may result from Comptonization of soft disk photons in a hot, thermal corona abovethe accretion disk. In the extreme scenario, in which the entire XMM-Newton spectrum isproduced by the corona, Figure 16 illustrates that one would require the energy dissipatedin the corona to be of the same order as that dissipated in the optically thick disk. Inthe case of Comptonization of soft photons with energy far below the X-ray regime, theX-ray spectral index Γ ph ∼ . y = 16 Θ τ T (1 + τ T ), where Θ = kT / ( m e c ) is the dimensionless coronal temperature,through Γ ph = − ± r
94 + 4 y (2)yielding a value of y = 1 .
85. This would imply coronal parameters of Θ τ T (1 + τ T ) = 0 . R . Such a choice might be problematic for two reasons: (1) For R & . × ( D/
20) cm, causality would not allow for variability on time scales of a day or less;(2) as the observed synchrotron emission would require the same magnetic field as chosen inour SED fits, this would result in the energy density in the leptonic particle population beingfar below equipartition with the magnetic field. While both arguments may not strictly ruleout such a scenario, we strongly prefer our fit scenario, close to equipartition, and allowingfor ∼ day scale variability.We found that the shortest observed variability occurs at optical wavelengths. In thestandard interpretation, this is the region near the high-frequency end of the synchrotronspectrum, emitted by the highest-energy electrons with the shortest radiative cooling timescales. The synchrotron cooling time scale for electrons radiating via synchrotron emissionin the V band, is τ obssy ( V ) ≈ . × (1 + z ) / B − / G D − / s (3)Using the redshift of z = 2 .
07, a characteristic magnetic field of B ∼ D = 19 from our fit results, we find an observed synchrotron cooling timeof τ obssy ( V ) ≈ . γ -rayemission, the actual radiative cooling rate is expected to be shorter than the rate expectedfrom synchrotron emission alone, by a factor corresponding to the Compton dominance (theratio of power output in the high-energy vs. the synchrotron component), which is of order10. Therefore, the total radiative cooling time scale of electrons emitting synchrotron in the 47 –V band is likely of the order of ∼ p can produce synchrotronemission with a maximum polarization degree of P max s = ( p + 1) / ( p + 7 /
3) for a perfectlyordered magnetic field (Rybicki & Lightman 1986). Based on our modeling results in Section6, the high-energy (cooled) part of the electron spectrum is expected to have a spectral indexof p = q + 1 ≈ .
6. This would yield a maximum polarization degree of P max s ≈
70 %. Thefact that the actual degree of polarization remains below ∼
10 % indicates the magneticfield is tangled on size scales much smaller than the size of the emission region. The rapid(day scale), apparently random variation of the degree of polarization and the EVPA seemsto indicate a turbulent process resulting in a large number of individual cells with randomlyoriented magnetic fields. Monte Carlo simulations (D’Arcangelo 2010) indicate that for a100 % chaotic magnetic field (i.e., no ordered field component), 150 turbulence cells resultin a degree of polarization of p = (6 ±
3) %, in agreement with the observed values forPKS 0528+134. Given a typical variability time scale in the optical of ∼ R cell ∼ c δt var / ( N / D ) ∼ × cmfor a Doppler factor of D ∼
20 (see §
6) and N = 150 turbulence cells. Comparing theoptical polarization variability to the polarization variability of the radio core (see § &
8. Summary and Conclusions
Over the last two decades PKS 0528+134 has become an important target for multi-wavelength observations because of its high luminosity from radio through γ -rays and the 48 –extreme flux and spectral variability that it shows in its flaring states. In this paper, wehave presented multiwavelength observations of PKS 0528+134 involving the XMM-Newton ,RXTE,
Suzaku , and
Fermi satellites as well as many ground based radio and optical tele-scopes. Our main goal was to characterize this γ -ray loud quasar in a quiescent state toimprove the understanding of SEDs and variability patterns of this source, and of blazars ingeneral. The variability analysis of the collected data, and the construction and modeling offour SEDs of this source, yielded the following results: • No significant short term flux and spectral variability (as determined by data in the corecampaign) was found in γ -rays, X-rays and most radio bands. However, for the sametime interval, significant flux variability with ∆ R . mag on time scales of several hourswas found in the optical, accompanied by a weak spectral softening with increasingflux. The latter trend may be interpreted as a steady contribution of the accretion diskflux at the blue end of the optical spectrum. Optical spectropolarimetry suggests anincreasing degree of polarization towards longer wavelengths, lending further supportto the hypothesis of synchrotron emission contributing an increasing fraction of theemission towards the red end of the spectrum. • Data analysis based on a more extended interval (two months or more) shows nosignificant γ -ray flux and spectral index variations, but moderate flux variability inthe X-rays ( | ∆ F/F | ∼
50 %) and at radio ( | ∆ F/F | .
20 %) frequencies on time scalesof ∼ XMM-Newton spectra of 2009 September andthe
Suzaku spectrum of 2008 September/October suggests that the X-ray spectral indexremains stable in the quiescent state of PKS 0528+134 even throughout substantial(factor ∼ • We constructed four SEDs of PKS 0528+134 in quiescence. Our results show thateven in the quiescent state, the bolometric luminosity of PKS 0528+134 is stronglydominated by its γ -ray emission, although the γ -ray spectra are significantly steeperthan found during the EGRET era. • We fitted the four SEDs with a leptonic combined SSC+ERC jet model. In thismodel, the low energy component is produced by the sum of the synchrotron processin the jet and the disk luminosity, and the high energy emission is due to the sum ofthe synchrotron self-Compton and the external radiation Compton contributions. Fitparameters close to equipartition were found for all SEDs. • The moderate variability in most wavelength bands, compared to the expected shortradiative cooling time scale, implies the persistence of particle acceleration on long 49 –time scales. This may favor acceleration scenarios based on standing features, such asre-collimation shocks.This work was supported by NASA through XMM-Newton Guest Observer Programawards NNX08AD67G and NNX09AV45G, Chandra Guest Observer Program award GO8-9100X, and Fermi Guest Investigator Program award NNX09AT82G.Norman I. Palma Cruz thanks the Fulbright Program and the National AutonomousUniversity of Honduras for making his stay in the U. S. possible during the interval that thisresearch took place.The
Fermi
LAT Collaboration acknowledges generous ongoing support from a numberof agencies and institutes that have supported both the development and the operation of theLAT as well as scientific data analysis. These include the National Aeronautics and SpaceAdministration and the Department of Energy in the United States, the Commissariat `al’Energie Atomique and the Centre National de la Recherche Scientifique / Institut Nationalde Physique Nucl´eaire et de Physique des Particules in France, the Agenzia Spaziale Italianaand the Istituto Nazionale di Fisica Nucleare in Italy, the Ministry of Education, Culture,Sports, Science and Technology (MEXT), High Energy Accelerator Research Organization(KEK) and Japan Aerospace Exploration Agency (JAXA) in Japan, and the K. A. Wallen-berg Foundation, the Swedish Research Council and the Swedish National Space Board inSweden.The VLBA is an instrument of the National Radio Astronomy Observatory, a facilityof the NSF, operated under cooperative agreement by Associated Universities, Inc.Additional support for science analysis during the operations phase is gratefully acknowl-edged from the Istituto Nazionale di Astrofisica in Italy and the Centre National d’ ´EtudesSpatiales in France.This paper is partly based on observations carried out at the German-Spanish Calar AltoObservatory, which is jointly operated by the MPIA and the IAA-CSIC, and on observationswith the Medicina and Noto telescopes operated by INAF — Istituto di Radioastronomia.Calar Alto data were acquired as part of the MAPCAT (Monitoring AGN with Po-larimetry at the Calar Alto Telescopes) project.Acquisition of the MAPCAT data is supported in part by MICIIN (Spain) grantsAYA2007-67267-C03-03 and AYA2010-14844, and by CEIC (Andaluc´ıa) grant P09-FQM-4784. 50 –”The Submillimeter Array is a joint project between the Smithsonian Astrophysical Ob-servatory and the Academia Sinica Institute of Astronomy and Astrophysics, and is fundedby the Smithsonian Institution and the Academia Sinica.”The Mets¨ahovi team acknowledges the support from the Academy of Finland to ourobserving projects (numbers 212656, 210338, and others)The research at Boston University (BU) was funded in part by NASA Fermi GuestInvestigator grants NNX09AT99G and NNX08AV65G, and by the NSF through grant AST-0907893. The PRISM camera at Lowell Observatory was developed by K. Janes et al. atBU and Lowell Observatory, with funding from the NSF, BU, and Lowell Observatory.D.M. acknowledges support from Russian RFBR foundation via grant 09-02-00092.The research at UMRAO was funded in part by NSF grant AST-0607523 and by NASAgrants NNX09AU16G and NNX10AP16G. The operation of UMRAO is made possible byfunds from the University of Michigan.G. Madejski acknowledges support from NASA through
Suzaku
Guest Observer grantNNX08AZ89G.The Steward Observatory
Fermi
Support Program is supported by NASA through
Fermi
Guest Investigator Program grants NNX08AW56G and NNX09AU10G.
REFERENCES
Abdo, A. A., et al., 2010a, ApJ, 715, 429Abdo, A. A., et al., 2010b, ApJ, 716, 835Abdo, A. A., et al., 2010c, ApJS, 188, 405Acciari, V. A., et al., 2009, ApJ, 707, 612Agudo, I., et al., 2011, ApJ, 726, L13Aharonian, F., Akhperjanian, A. G., Bazer-Bachi, A. R., et al., 2007, ApJ, 664, L71Albert, J., et al., 2007, ApJ, 669, 862Aller, H. D., Aller, M. F., Latimer, G. E., & Hodge, P. E. 1985, ApJS, 59, 513Anders, E. and Grevesse N. 1989; Geochimica et Cosmochimica Acta 53, 197. 51 –Arnaud, K. A. 1996; in ASP Conf. Ser. 101: Astronomical Data Analysis Software andSystems (ADASS) V, 17.Atwood, W. B., et al., 2009, ApJ, 697, 1071Baars, J. W. M., et al. 1977, A&A, 61, 99Bach, U., et al., 2007, A&A, 464, 175B¨ottcher, M. & Chiang, J., 2002, ApJ, 403, 663B¨ottcher, M., 2010, in proc. “Fermi Meets Jansky”, Eds. T. Savolainen, Ed. Ros, R. W.Porcas, J. A. Zensus, p. 41B¨ottcher, M., & Dermer, C. D., 2010, ApJ, 711, 445Bregman, J. N., Glassgold, A. E., Huggins, P. J., & Kinney, A. L. 1985, ApJ, 291, 505Britzen, S., Witzel, A., Krichbaum, T. P., Quian, S. J., & Campbell, R. M. A&A, 341, 418Bromberg, O., & Levinson, A., 2009, ApJ, 699, 1274Cardelli, J. A., Clayton, G. C., & Mathis, J. S., 1989, ApJ, 345, 245D’Arcangelo, F. D., 2010, Ph.D. Thesis, Boston UniversityDermer, C. D., & Schlickeiser, R., 1993, ApJ, 416, 458Dickey, J. M., & Lockman, F. J., 1990, ARAA, 28, 215Dingus, B. L., Bertsch D. L., Diegel S. W., et al., 1996, ApJ, 467, 589Finke, J. D., Dermer, C. D., & B¨ottcher, M., 2008, ApJ, 686, 181Francis, P. J., Hewett, P. C., Foltz, C. B., Chaffee, F. H., Weyman, R. J., & Morris, S. L.,1991, ApJ, 373, 465Gabriel, C., Denby, M., Fyfe, D. J. et al. 2004; in ASP Conf. Ser. 314: Astronomical DataAnalysis Software and Systems (ADASS) XIII, 759.Ghisellini, G., et al. 1999, A&A, 348, 63Gurwell, M.A., Peck, A.B., Hostler, S.R., Darrah, M.R., Katz, C.A., & the SMA Team 2007.”Monitoring Phase Calibrators at Submillimeter Wavelengths”, From ZMachines toALMA:(Sub)millimeter Spectroscopy of Galaxies, Charlottesville, VA. ASP Confer-ence Series, Vol. 375, A.J. Baker, J. Glenn, A.I. Harris, J.G. Mangum, and M.S. Yun,eds., p 274. 52 –Hogerheijde, M. R., Geus, E. J., Spaans, M., Langevelde, H. J., & Dishoeck, E. F., 1995,441, L93Hunter, S. D., Bertsch, D. L., Dingus, B. L., et al., 1993, ApJ, 409, 134Jansen, F. et al. 2001; A&A, 365, L1.Jorstad, S.G., Marscher, A.P., Mattox, J.R., Wehrle, A.E., Bloom, S.D., & Yurchenko, A.V.,2001, ApJS, 134, 181Jorstad, S. G., et al. 2005, AJ, 130, 1418Jorstad, S. G., et al. 2010, ApJ, 715, 362Joshi, M., & B¨ottcher, M., 2010, ApJ, in pressKalberla, P. M. W., Burton, W. B., Hartmann, Dap, Arnal, E.M ., Bajaja, E., Morras, R.,and P¨oppel, W. G. L. 2005; A&A, 440, 775.Leto, P., Umana, G., Trigilio, C., et al., 2009, A&A, 507, 1467LLister, M. L., et al., AJ, 138, 1874Liszt, H. S., & Wilson, R. W., 1993, ApJ, 403, 663Marscher, A. P., & Gear, W. K., 1985, ApJ, 298, 114Mason, K. O. et al. 2001; A&A, 365, L36.Mattox, J. R., et al., 1996, ApJ, 461, 396Mattox, J. R., Schachter, J., Molnar, L., et al., 1997, ApJ, 481, 95McNaron-Brown, K., Johnson, W. N., Jung, G. V., et al., 1995, ApJ, 451, 575Mimica, P., et al., 2004, A&A, 418, 947Morrison, R., & McCammon, D., 1983, ApJ, 270, 119Mukherjee, R., Dingus, B. L., Gear, W. K., et al., et al., 1996, ApJ, 470, 831Mukherjee, R., B¨ottcher, M., Hartman, R. C., Sreekumar, P., Thompson, D. J., Mahoney,W. A., Pursimo, T., Sillanp¨a¨a, A., & Takalo, L. O., 1999, ApJ, 527, 132Peng B., et al. 2001, ApJ, 551, 172 53 –Piner, B. G., Bhattarai, D., Edwards, P. G., & Jones, D.L., 2006, ApJ, 640, 196Pohl, M., et al. 1996, A&AS, 120, 529Raiteri, C. M., et al. 1998, A&AS, 130, 495Raiteri, C. M., et al., 2007, A&A, 473, 819Raiteri, C. M., et al., 2008, A&A, 491, 755Reich, W., et al. 1993, A&A, 273, 65Rieger, F. M., & Duffy, P., 2004, ApJ, 617, 155Rieger, F. M., & Duffy, P., 2006, ApJ, 652, 1044Rybicki, G. B., & Lightman, A. P., 1986, “Radiative Processes in Astrophysics”, Wiley-VCH.Sambruna, R. M., et al. 1997, ApJ, 474, 639Schelegel, D. J., Finkbeiner, D. P., Davis, M., 1998, ApJ, 500, 525Shakura, N. I., & Sunyaev, R. A., 1973, A&A, 24, 337Sikora, M., Begelman, M. C., & Rees, M. J., 1994, ApJ, 421, 153Smith et al. 1985, AJ, 90, 1184Spada, M., et al., 2001, MNRAS, 325, 1559Stawarz, L., & Ostrowski, M., 2002, ApJ, 578, 763Stevens, J. A., et al. 1994, ApJ, 437, 91Str¨uder, L., Briel, U., Danner, K. et al 2001; A&A, 365, L18.Ter¨asranta, H., Tornikoski, M., Mujunen, A. et al. 1998, A&AS, 132, 305Valtaoja, E., Ter¨asranta, H. 1995, A&A, 297, L13Wall, J. V., & Peacock, J. A., 1985, MNRAS, 216, 173Wardle, J. F. C., & Kronberg, P. P., 1974, ApJ, 194, 294Yurchenko, A.V., 2001, ApJS, 134, 181Zhang Y.F., Marscher A.P., Aller H.D., et al., 1994, ApJ, 432, 91 54 –