Synchrotron flaring in the jet of 3C 279

Please download to get full document.

View again

of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
We study the synchrotron flaring behaviour of the blazar 3C279 based on an extensive dataset covering 10 years of monitoring at 19 different frequencies in the radio-to-optical range. The properties of a typical outburst are derived from the
    a  r   X   i  v  :  a  s   t  r  o  -  p   h   /   0   6   0   6   6   4   6  v   1   2   7   J  u  n   2   0   0   6 Astronomy & Astrophysics  manuscript no. astro˙ph c  ESO 2014January 16, 2014 Synchrotron flaring in the jet of 3C 279 E.J. Lindfors 1 , 2 , M. T¨urler 3 , 4 , E. Valtaoja 1 , 5 , H. Aller 6 , M. Aller 6 , D. Mazin 7 , C. M. Raiteri 8 , J. A. Stevens 9 , M.Tornikoski 2 , G. Tosti 10 , and M. Villata 8 , 1 Tuorla Observatory, V¨ais¨al¨a Institute of Space Physics and Astronomy, University of Turku, 21500 Piikki¨o, Finland 2 Mets¨ahovi Radio Observatory, Helsinki University of Technology, 02540 Kylm¨al¨a, Finland 3 Geneva Observatory, ch. des Maillettes 51, 1290 Sauverny, Switzerland 4 INTEGRAL Science Data Centre, ch. d’Ecogia 16, CH-1290 Versoix, Switzerland 5 Department of Physics, University of Turku, 20100 Turku, Finland 6 Astronomy Department, University of Michigan, Ann Arbor, MI 48109, United States 7 Max-Planck-Institut f¨ur Physik, M¨unchen, Germany 8 INAF-Osservatorio Astronomico di Torino, Via Osservatorio 20, I-10025 Pino Torinese, Italy 9 Centre for Astrophysics Research, Science and Technology Research Institute, University of Hertfordshire, College Lane, Herts, AL10 9AB 10 Osservatorio Astronomico di Perugia, via Bonfigli, 06123 Perugia, ItalyReceived, accepted ABSTRACT Aims.  We study the synchrotron flaring behaviour of the blazar 3C 279 based on an extensive dataset covering 10 years of monitoring at 19di ff  erent frequencies in the radio-to-optical range. Methods.  The properties of a typical outburst are derived from the observations by decomposing the 19 lightcurves into a series of self-similarevents. This analysis is achieved by fitting all data simultaneously to a succession of outbursts defined according to the shock-in-jet model of Marscher & Gear (1985). Results.  We compare the derived properties of the synchrotron outbursts in 3C 279 to those obtained with a similar method for the quasar3C 273. It is argued that di ff  erences in the flaring behaviour of these two sources are intrinsic to the sources themselves rather than beingdue to orientation e ff  ects. We also compare the start times and flux densities of our modelled outbursts with those measured from radiocomponents identified in Very Long Baseline Interferometry (VLBI) images. We find VLBI counterparts for most of our model outbursts,although some high-frequency peaking events are not seen in the radio maps. Finally, we study the link between the appearance of a newsynchrotron component and the EGRET gamma-ray state of the source at 10 di ff  erent epochs. We find that an early-stage shock componentis always present during high gamma-ray states, while in low gamma-ray states the time since the onset of the last synchrotron outburst issignificantly longer. Thisstatisticallysignificant correlation supports the idea that gamma-ray flaresare associated withthe early stages of shock components propagating in the jet. We note, however, that the shock wave is already beyond the broad line region during the gamma-ray flaring. Key words.  galaxies:active – galaxies:jets – galaxies:quasars:individual:3C279 1. Introduction It is generally accepted that radio outbursts in active galacticnuclei (AGN) are triggered by growing shocks in a relativis-tic jet. There have been many e ff  orts to identify the individ-ual componentsin radioto submillimeter(submm)lightcurves.Litchfield et al. (1995) and Stevens et al. (1995, 1996, 1998)managed to follow the early evolution of individual outburstsby subtracting the quiescent contribution (assumed to be con-stant) from the total spectrum. Valtaoja et al. (1999) identifiedoutbursts by decomposing the variations in millimeter (mm)and centimeter lightcurves into exponential flares. T¨urler et al.(1999)measuredthe properties of synchrotronoutbursts by de- Send o  ff   print requests to : E.J. Lindfors composing the radio-to-submillimeter wavelength lightcurvesinto a series of self-similar flaring events. The method wasfurther developed by T¨urler et al. (2000) who used the shock model of Marscher & Gear (1985) to describe both the aver-age evolution of the outbursts and their individual characteris-tics. This model was found to provide a good description of the lightcurves of 3C 273. In this work, we decompose thelightcurves of a second source, 3C 279. Since our adoptedmethodology is analogous to that used by T¨urler et al. (2000)the results are directly comparable to those found for 3C 273.While in previous studies the outbursts were identified inthe radio-to-submillimeter regime, we here follow the syn-chrotron spectrum up to infrared and optical frequencies.Searches for correlations in the radio-to-optical emission from  2 E. J. Lindfors et al.: Synchrotron flaring in the jet of 3C 279 AGN have been conductedsince the 1970s.One of the most re-cent works was presented in Hanski et al. (2002). In their sam-ple of 20 AGNs they found a clear radio-to-optical correlationin seven cases, a possible correlation in six cases and no corre-lation for the remainder. This agrees well with the findings of previous studies: there appears to be some kind of correlationbetween the radio and optical emission but this correlation isnot seen in all sources at all epochs. The generaltrend seems tobe that all radio outbursts are accompaniedby optical outburstsbut converselynot all optical outbursts have radio counterparts.The source 3C 279 is one of the brightest and most variableblazars at radio wavelengths. As such, it is monitoredregularlyin all radio wavebands from the centimeter to the submillime-ter domain.In the optical,the historicallightcurveshows varia-tionswithatypicalamplitudeofabout2magnitudes,butreach-ing 8 magnitudes during flares (Webb et al. 1990). The weak blue bump of 3C 279 allows us to follow the synchrotronspec-trum up to infrared and optical wavelengths. Indeed, 3C 279 isone of the few objects for which a clear correlationis foundbe-tween the radioandoptical wavebands(Tornikoskiet al. 1994).In this paper we present results of a multifrequencylightcurvedecompositionof 3C 279. The method used is basedon T¨urler et al. (1999; 2000) with some modifications for thespecific case of 3C 279, including an extension of the analy-sis to infrared and optical wavelengths. The average evolutionof outbursts is compared to that of 3C 273. We also compareour results with (1) VLBI maps from Wehrle et al. (2001) andJorstad et al. (2004)to study the connectionbetween the modeloutbursts and VLBI components, and (2) gamma-ray data ob-tained by Hartman et al. (2001) to study the connection be-tween the synchrotron spectra and the gamma-ray state. 2. Data To decompose the radio-to-optical lightcurves of 3C 279 weused data from five radio and seven mm  /  submm wavelengthsas well as four infrared and two optical wavebands. The datasample extends from 1989.0 to 1999.5.The 4.8, 8.0 and 14.5GHz data are from the Universityof Michigan Radio Astronomy Observatory (UMRAO). Someof these data have not previously been published. The 22and 37GHz data and some of the 90GHz data are from theMets¨ahovi Radio Observatory. These data were published byTer¨asranta et al. (1998, 2004). We have also used 90 and230GHz (1.3mm) data from the Swedish-ESO Sub-millimeterTelescope (SEST). SEST data taken prior to 1994.5 were pub-lished by Tornikoski et al. (1996) while more recent data arepublished here for the first time. From the Institut de RadioAstronomie Millim´etrique (IRAM) we have data at 90, 150and 230 GHz (Steppe et al. 1993; Reuter et al. 1997). We alsoused data at 2.0, 1.3, 1.1, 0.85, 0.8, 0.45 and 0.35mm from theJames Clerk Maxwell Telescope (JCMT). The 0.85mm datasetwas published by Robson et al. (2001) while data in the otherwavebands are either from Stevens et al. (1994) or are previ-ously unpublished (post 1993). The dataset of Stevens et al.(1994) also includes the four infrared wavebands used in thiswork.The optical R- and V-band lightcurves are collected fromthe literature, but also contain previouslyunpublisheddata. Wehave used the data from Maraschi et al. (1994), Hartman etal. (1996), Villata et al. (1997), Katajainen et al. (2000) andHartman et al. (2001). The previously unpublished data arefrom the Kungliga Vetenskapsakademien(KVA) telescope andthe Nordic Optical Telescope (NOT) as well as from the obser-vatories of Perugia and Torino. 3. Model and Method Inthispaperwestudythemultifrequencylightcurvesof3C279in the context of the shock model of Marscher & Gear (1985).As the shock propagates downstream in the relativistic jet itevolvesthroughthree stages. In the initial (growth)stage, wheninverse Compton losses predominate, the synchrotron self-absorption turnover frequency decreases and the turnover fluxdensity increases. In the second (plateau) stage, synchrotronlosses dominate and the turnover frequency decreases whilethe turnover flux density remains roughly constant. Duringthe third (decay) stage when adiabatic losses dominate, bothturnover frequency and turnover flux density decrease.Bjornsson & Aslaksen (2000) criticize one of the assump-tions of Marscher& Gear (1985)which concernsthe rise of thepeak flux density during the initial Compton stage. Their cor-rection of the expression for the energy density of synchrotronphotonsimpliesa muchshallowerrise ofpeakfluxdensitywithpeak frequency at the beginning of the outburst. On the otherhand,multipleComptonscattering,aprocessnotconsideredbyMarscher & Gear (1985), would actually steepen their derivedrelation. There is therefore no compelling theoretical reason touse a modified version of the srcinal Marscher & Gear (1985)model.In T¨urler et al. (2000), a generalisation of this model wasused to decompose the multifrequency lightcurves of 3C 273.We adopt a similar methodology in this work. The shape of the emission spectrum behind the shock front is assumed tobe that of a simple synchrotron spectrum (with electron en-ergy distribution  N  (  E  )  ∝  KE  − s ) with two spectral breaks,namely the low-frequency break,  ν h , below which  α thick   =  5 / 2and the high frequency break,  ν b , which steepens  α thin  from(1  −  s ) / 2 to  − s / 2. To extend the decomposition into the in-frared and optical we introduce an additional high frequencyexponential cut-o ff  to the synchrotron spectrum correspondingto emission from the most energetic electrons. The sharp low-and high-frequencyspectral breaks of T¨urler et al. (2000) werealso replaced by more realistic smooth transitions. As in themodelling of the micro-quasar GRS 1915 + 105 (T¨urler et al.2004),we introduceda new parameter allowing the ratio  ν h /ν m of the low-frequency spectral break,  ν h , and the synchrotronself-absorption turnover,  ν m , to vary with time.Another significant change is the addition of an underly-ing jet component with constant flux density. This compo-nent is assumed to be an inhomogeneous synchrotron sourcewith a spectrum defined by a high-frequency break fixed at375GHz – as suggested by the quiescent level studies of 3C 279 (Litchfield et al. 1995) – and by two free parameters of the fit defining the synchrotronself-absorptionflux density and  E. J. Lindfors et al.: Synchrotron flaring in the jet of 3C 279 3 Fig.1.  Ten out of nineteen radio-to-optical lightcurves of 3C 279. The points indicate the observed flux density and the solid lineour best fit which is a sum of the underlying jet (long dashed line) and the fifteen outbursts (dotted lines). The first dotted linerepresents the global decay of all outbursts peaking before 1989.0frequency. To be consistent with the lowest infrared measure-ments an exponential cut-o ff   was also added to the underlying jet spectrum at a somewhat arbitrary frequency of 2 · 10 5 GHz.We note, however, that our conclusions are insensitive to thesenumerical values.The specificity of the individual outbursts is modelled byvarying the scale of their evolution in flux density, frequencyand time as done in T¨urler et al. (1999). The more physicalapproach adopted by T¨urler et al. (2000) of varying the onsetvalues of   K   (electron energy density normalization),  B  (mag-netic field strength) and  D  (Doppler factor) was also tried, butwas found to limit the di ff  erences between outbursts too muchto get an equally good fit.The model presented here assumes a constant Doppler fac-tor  D  ∝  R − d  and a magnetic field  B  ∝  R − b perpendicular tothe jet axis as suggested by VLBI polarization measurements(Lister et al. 1998, Marscher et al. 2002). The exponents,  b  and d  , of the evolution of these quantities with radius,  R , of the jetare therefore fixed to  d   =  0 and  b  =  1. We tried varying thesevalues as well, but this did not result in significantly better fits.The resulting model has a total of 77 free parameters tofit the 15 outbursts identified in the lightcurves of 3C 279 from1989to 2000.Twelveparametersare usedto describethe shape  4 E. J. Lindfors et al.: Synchrotron flaring in the jet of 3C 279 and evolution of the synchrotron spectrum for an average out-burst. Three further parameters describe the initial flux densitydecay while two are neededto describe the underlyingjet com-ponent. The remaining 60 parameters are used to describe thestart time andthespecific characteristicsofthe15 di ff  erentout-bursts. The total number of degrees of freedom is 3281. 4. Results The best-fit decomposition of the radio-to-optical lightcurvesinto a series of 15 self-similar outbursts is shown in Fig. 1. Ithas a reduced  χ 2 value of 8.72. This relatively high value canbe considered as acceptable here keeping in mind that the aimof the modelling is to derive the average properties of a typicaloutburst and that the specificity of each outburst is only mod-elled very crudely in order to minimise the number of free pa-rameters. Furthermore, the analytical shock-in-jet model itself is ofnecessity asimplistic idealizationof thecomplexemissionof actual jets (G´omez 2005) and thus a perfect match of themodel to the data cannot be expected. At the lowest frequen-cies, the fit describes the smooth shape of the lightcurves verywell. At90GHz, thegeneralfluxdensitylevelofthefit is good,but during the biggest outbursts (in 1991, 1993.5, 1994 and1996.5)the modelflux densityremains belowthe observations.The same tendencyalso applies to all mm-bandlightcurvesandto the optical R- and V-band lightcurves. Part of this fit (until1994.0) has been published in Lindfors et al. (2005). 4.1. Average evolution of the synchrotron spectrum  The average evolution of an outburst in 3C 279, as defined by12 fit parameters, is shown in Fig. 2. In Fig. 2c we see theaverage outburst spectra. It peaks at  ∼ 150GHz, which is thefrequency at which the transition from the Compton stage tothe synchrotron stage takes place. The peak flux density is 6.6Jy. The second stage transition in the average spectra is at  ∼ 20GHz and its flux density is 2.2 Jy.Theoverallshapeoftheevolutionis quitedi ff  erentfromtheone derived for 3C 273 (T¨urler et al. 2000). The most obviousdi ff  erenceis thedecreasingfluxdensityduringthesecondstageoftheevolutionwhichis quitedi ff  erentfromthetypicalplateauseen in 3C 273. In the model, this di ff  erence can be ascribedto the high value of the parameter  k   defining the decrease of the normalization  K   ∝  R − k  of the electron energy distribution  N  (  E  )  ∝  K E  − s with jet opening radius,  R . The derived value of  k   is 4 . 0 which is far above the value of   k  ad  =  2( s + 2) / 3  =  2 . 81(with  s  =  2 . 25) predicted for an adiabatic jet flow. The srci-nal Marscher & Gear model adopted only adiabatic compres-sion to heat the electrons in the shock. While this seems tobe a good assumption for 3C 273 (k  = 3.03), our high  k  -valuefound for 3C 279 suggests that the emitting electrons are sub- ject toimportantnon-adiabaticcoolingprocesseswithinthejet.This might be due, at least in part, to synchrotron and inverseCompton radiative losses in the undisturbed underlying jet.As mentioned in T¨urler et al. (2000), the three parameters r  ,  k   and  d   are di ffi cult to constrain uniquely by the fit. In par-ticular, the e ff  ect of   k   >  k  ad   and of   d   >  0 are similar, both pro-ducing a decreasing flux density during the synchrotron stageof the evolution. The high value of   k   might therefore also indi-cate that our assumption of   d   =  0 is not valid and thus suggestthat the Doppler factor tends to decrease during the evolutionof the shock. This could be due to a decelerating jet flow or toa geometry in which the jet bends away from the line of sight(see Jorstad et al. 2004).Another di ff  erence is that the flattening of the spectral in-dex by  ∆ α thin  = + 0 . 5 (predicted to occur at the synchrotron-to-adiabatic stage transition) is found to be much less abruptfor 3C 279 than for 3C 273, and for 3C 279 it starts muchearlier during the initial Compton stage. This can be seenby the position of the start of the flattening ( t  f  ) in Fig. 3,which is much earlier than found for 3C 273 (see Fig. 4. of T¨urler et al. (2000)). A similar behaviour was also found forGRS 1915 + 105 (T¨urler et al. 2004).Apart from  k  , the values of two other physical parametersof the jet are found to be similar to those derived for 3C 273(T¨urler et al. 2000). The index,  s , of the electron energy dis-tribution is found here to be  s  =  2 . 25 ( s  =  2 . 05 for 3C 273)and the opening radius  R  ∝  L r  of the jet with distance,  L , alongthe flow is found to be slightly non-linear ( r   =  0 . 78) indicatingthat, as for 3C 273 ( r   =  0 . 8), the jet opening angle is slowlydecreasing with distance suggesting that some jet collimationprocess is at work.We alsonote,aswas foundforGRS 1915 + 105(T¨urleretal.2004), that there is a tendency for the synchrotron emission of the outbursts to start out inhomogeneous but to become homo-geneous as the shock evolves with time. This is manifested as adecreasing  ν h /ν m  ratio which a ff  ects the shape of the spectrumat frequencies below the turnover as can be seen in Fig. 2c.Another di ff  erence is that in 3C 273 the outbursts are veryshort-lived at higher frequencies, while in 3C 279 at all fre-quencies above  ∼ 150GHz it takes  ∼  0.45 years to reach thepeak flux density. Since the rise time at all such high frequen-cies is the same these outbursts peak simultaneously. This isshown in Fig. 3. This result, together with the finding that theplateau stage is a bit shorter than in 3C 273, might suggest thatin 3C 279 the magnetic field energy density is relatively lowerthan the electron energy density.3C 279 is an archetypical blazar while 3C 273 is not atrue member of the blazar class. Most conspicuously, it is notrapidly variable in the optical band and has low optical polar-ization. However, it exhibits blazar-like optical behaviour atlow levels, diluted by thermal radiation (Impey et al. 1989,Valtaoja et al. 1990, Valtaoja et al. 1991) and has thereforebeen dubbed a mini-blazar by Impey et al. (1989).As expectedwithin the context of the unification scheme, 3C 273 has arather large viewing angle, estimated to be around 10 degrees(L¨ahteenm¨aki& Valtaoja 1999,Savolainen et al. 2006).This issmaller than for ordinary quasars, but larger than for blazars;for 3C 279 the estimated viewing angle is around 2 degrees(Lindfors et al. 2005 and references therein).In the simplest unification models, orientation is the onlyparameter. One might therefore expect that the borderlineblazar 3C 273 would appear similar to 3C 279 if the view-ing angle were decreased. However, some of the di ff  erenceswe find between 3C 273 and 3C 279 (non-adiabatic jet flowin 3C 279, adiabatic in 3C 273; magnetic energy density lower  E. J. Lindfors et al.: Synchrotron flaring in the jet of 3C 279 5 Fig.2.  Logarithmic evolution of thesynchrotron emission of an averageoutburst in 3C 279. The three dimen-sional (log  ν , log  S  , log  t  ) representa-tionis shownin( a ),whilethe Cartesianprojections of this surface are shownin ( b ), ( c ) and ( d ).  b ) Frequency ver-sus time representation with contoursstarting at a flux density of 0.02 Jy andspaced by 0.3 dex.  c ) The synchrotronspectra at di ff  erent times spaced by0.2 dex.  d ) The lightcurves at di ff  er-ent frequencies spaced by 0.2 dex. Thethick solid line with arrows traces thetime evolution of the spectral turnover,whereas the dashed line connects thepeak of the lightcurves at di ff  erent fre-quencies. The filled circles refer to thespecific characteristics of the outbursts(see Section 4.2) and the vertical barsin ( b ) and ( c ) show the frequency cov-erage of the 19 lightcurves. Fig.3.  Model lightcurves of the average outburst in 3C 279at di ff  erent frequencies spaced by 0.1 dex (grey lines). Thesix highlighted lightcurves are at frequencies,  ν , defined bylog( ν  /  GHz) =  3.5, 2.5,...,1.0, in order of increasing timescales.t r  and t p  refer to transitions from Compton to synchrotronstage (t r  = 0.45 years) and from synchrotron to adiabatic stage(t p  = 3.34 years) respectively.t f   refers to the time when the flat-tening of the spectral index starts (t f   = 0.046 years).than electronenergydensityin 3C 279whereas in 3C 273thereis little di ff  erence) seem intrinsic to the sources and cannotbe explained by di ff  erences in viewing angles alone. However,we note that variations in the viewing angle as a function of time  /  radial distance could play a role in the observed di ff  er-ences between these two sources. 4.2. Characteristics of Different Outbursts  The evolution of each of the 15 individual outbursts is allowedto deviate in scale from the average outburst. For this we needthree parameters corresponding to the logarithmic shifts alongthe three axes of Fig. 2, namely time, ∆ log t  , frequency, ∆ log ν and flux density,  ∆ log S  . The values obtained for these shiftsare given in Table 1 and their distribution is shown graphicallywith the filled circles in Fig. 2.We attempted to understand the physical srcin of the dif-ferences seen between the outbursts by looking for correla-tions between  ∆ log S  ,  ∆ log ν  and  ∆ log t  . Contrary to resultsfor 3C 273 (T¨urler et al. 1999) we find no correlation betweenany of the shifts, and Fig. 2 shows clearly that the shifts do notalign with any axis. It is also clear that the outbursts do not dif-fer mainly in amplitude as found for GRS 1915 + 105(T¨urler etal. 2004) but also in peaking frequency and duration.To establish that our decomposition is not purely mathe-matical but corresponds to some physical reality, we comparethe outbursts suggested by our fit to observed VLBI compo-nents. 3C 279 has been monitored with VLBI since the 1980s,and for the period considered in this paper the VLBI data cov-erage is good. In Table 2, the start times ( T  0 ) of the outbursts,as derived from our modelling, are compared with the extrapo-
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!