Connections between millimetre continuum variations and VLBI structure in 27 AGN


12 pages

Please download to get full document.

View again

of 12
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.
Connections between millimetre continuum variations and VLBI structure in 27 AGN
  Astronomy & Astrophysics  manuscript no.(will be inserted by hand later) Connections between millimetre continuum variations andVLBI structure in 27 AGN T. Savolainen 1 , K. Wiik  1 , 2 , E. Valtaoja 1 , 3 , S.G. Jorstad 4 , and A.P. Marscher 4 1 Tuorla Observatory, Väisälä Institute for Space Physics and Astronomy, University of Turku, FIN-21500 Piikkiö, Finland 2 Metsähovi Radio Observatory, Helsinki University of Technology, FIN-02540 Kylmälä, Finland 3 Department of Physics, University of Turku, FIN-20100 Turku, Finland 4 Institute for Astrophysical Research, Boston University, 725 Commonwealth Ave., Boston, MA, 02215Received 14 May 2002 / Accepted 27 August 2002 Abstract.  We compare total flux density variations in 27  γ  -ray blazars with structural changes in their parsec-scale jets usingmulti-epoch VLBA observations at 22 and 43 GHz together with data from the Metsähovi quasar monitoring program at 22and 37 GHz. There is a clear connection between total flux density outbursts and VLBI components emerging into the jet.For essentially every new moving VLBI component, there is a coincident total flux density flare, with evolution similar tothat of the component. Furthermore, extrapolated ejection times of the new VLBI components correspond to the beginnings of associated flares. Our results suggest that it is possible to explain all the radio variations as shocks propagating down the jet.A large fraction of the shocks grow and decay within the innermost few tenths of a milliarcsecond and therefore we see themonly as “core flares” in the VLBI images. However, with present data we cannot exclude the possibility that the core itself alsobrightens (and thus contributes to the flare) as a shock passes through it. Key words.  BL Lacertae objects: general – galaxies: active – galaxies: jets – quasars: general – radio continuum: galaxies –techniques: interferometric 1. Introduction Blazars are an interesting and violent subclass of active galac-tic nuclei (AGN), grouping together (although somewhat artifi-cially from a physical point of view) radio-loud quasars and BLLacertae objects. These sources have in common flat cm-waveradio spectrum, high and variable polarization, and pronouncedvariability of the flux density at all frequencies. The superlumi-nal motion observed in these sources together with brightnesstemperatures in excess of the  10 12 K inverse Compton limit(Kellermann & Pauliny-Toth 1968) indicate highly beamedemission from relativistic jets oriented towards the line of sightof the observer. Using modern-day very long baseline interfer-ometry(VLBI)techniqueswecanresolvethejet-likestructuresin blazars on angular scales down to  ∼  0 . 1  milliarcseconds(mas).Relativistic jets also offer an explanation for radio-to-infrared variability of blazars. Marscher & Gear (1985) stud-ied the strong 1983 outburst in 3C 273 and managed to fit theflaring spectra with self-absorbed synchrotron emission. Theyexplained successfully the time-evolution of the flare as be-ing due to a shock wave propagating in the relativistic jet. TheMarscher &Gear model (hereafter MG-model) has threestagesof shock evolution based on the dominant cooling mechanisms Send offprint requests to : T. Savolainen, e-mail: of the electrons: 1) the Compton scattering loss phase, 2) thesynchrotron radiation loss phase and 3) the adiabatic expan-sionlossphase.MG-modelisasimple,analyticalmodel,whichdescribes well the general behaviour of the radio outbursts inAGN (but see the critíque of Björnsson & Aslaken 2000). Themodel was generalized by Marscher et al. (1992) to include theeffects of bending in jets and turbulence on the light curves.Hughes et al. (1985, 1989a, 1989b, 1991) proposed a simi-lar shock model based on a numerical code simulating a piston-driven shock. Their model succesfully explains the lower fre-quency variability, but it does not incorporate radiative energylosses of the electrons, which are important at high frequen-cies and in the earliest stages of the shock evolution. Valtaojaet al. (1992b) presented a generalized shock model describingqualitatively the three stages of the shock evolution (growth,plateau and decay) without going into details, thus providinga framework for comparison between the theory and observa-tions. Total flux density (TFD) monitoring campaigns, whichprovide nearly fully sampled flux curves at radio wavelengths,and VLBI images, which allow us to map the parsec-scalestructure of the blazar jets, are the two main observational toolsfor constraining theoretical models.In VLBI observations of blazars, bright knots of emissionreferred to as “components” are seen. These components lineup to form jet-like features appearing in various forms fromvery straight to heavily bent structures. The so-called “core”  2 T. Savolainen et al.: Connections between millimetre continuum variations and VLBI structure in 27 AGN is the point where the jet becomes visible. The core is pre-sumed stationary (see Bartel et al. 1986), but most of the otherVLBI components move outward in the jet at apparent super-luminal speeds. However, in some sources there are also sta-tionary components other than the core. These may be due, forexample, to interactions between the jet and the surroundinginterstellar medium.According to the shocked jet models, moving componentsin the VLBI maps are interpreted as shocks propagating downthejet.However, therehasbeen adearthofconclusive evidencelinking VLBI components with radio flux variations; only arelatively small number of individual sources have been in-vestigated thus far. One of the first studies linking TFD vari-ations with moving knots in the VLBI maps was carried outby Mutel et al. (1990). They found that each of four majorTFD outbursts of BL Lac between 1980 and 1988 can be as-sociated with the emergence of a new superluminal compo-nent. Abraham et al. (1996) estimated the ejection times of seven VLBI components in 3C 273 and noticed that all ejec-tions were related to increases in the single-dish flux density atfrequencies higher than 22 GHz. Türler et al. (1999) also stud-ied 3C 273 by decomposing multi-frequency light curves intoa series of self-similar flares. They found good correspondencebetween the ejection times of the VLBI components and thebeginning times of the flares. Krichbaum et al. (1998) have re-ported a correlation between mm-VLBI component ejectionsand local minima in the 90 GHz total flux density curve of PKS 0528+134. For 3C 345, which is one of the best observedsources with VLBI at 22 GHz, Valtaoja et al. (1999) wereable to associate VLBI components with individual millimetreflares. Similar correlations were also found for PKS 0420 − 014(Britzen et al. 2000) and for 3C 279 (Wehrle et al. 2001).In our study, we compare for the first time two large datasets: multi-epoch VLBA images of 42 blazars (Jorstad et al.2001a) detected at 0.1–3 GeV by EGRET and TFD data fromthe mm-wave Metsähovi Radio Observatory quasar monitor-ing program. A description of our data is given in section 2.Our aim is to establish connections between TFD variationsand structural changes in the jets. The results from the analy-sis, as we will show in section 3, strongly support the shocked jet model.The VLBI core is the dominant component in almost all thecases studied. The core region was usually also highly variable,being responsible for most of the observed TFD variability inthese sources. Variations in the VLBI core flux are reportedin the literature quite often (see, e.g., the results of the recentVLBA monitoring of 3C 279 by Wehrle et al. 2001). Since thecore is usually assumed to be the apex of the jet, the implicitassumption is that a core flare results from a change in the jetflow parameters. However, according to our study, these varia-tions are rather related to moving VLBI components that blendwith the radio core. This will be discussed in section 4. 2. The data Our data set consists of observations from two separate cam-paigns, namely the VLBA monitoring of EGRET-detectedblazars by Jorstad et al. (2001a) and the Metsähovi RadioObservatory quasar monitoring program. Jorstad et al. (2001a)monitored a sample of 42  γ  -bright blazars at 22 and 43 GHzwith VLBA between 1993 and 1997. For 27 of these sources(see Table 1), there were enough TFD variation data availablefrom Metsähovi monitoring (Teräsranta et al. 1998) to reliablyidentify large outbursts in the flux curves.Selection criteria for sources observed by Jorstad et al.(2001a) with the VLBA were: (1) detection by EGRET (E >100 MeV; Hartman et al. 1999); (2) flux density at 37 GHz  1 Jy; and (3) declination (J2000) ≥ − 30 ◦ . These criteria givea sample containing roughly 60% of the known  γ  -ray blazars(Hartman et al. 1999). The VLBA observations were made athigh radio frequencies giving  ≈  0.1 – 0.3 mas resolution and ≈  10  mas map size. High resolution allows us to study the in-ner parts of the jet and possibly see how the shock formation isconnected to the flaring behaviour of these sources. However,as we discuss below, VLBA maps at 43 GHz in many cases stillhave insufficient resolution to separate the new shock in the jetfrom the core before the millimetre flare is over.Since its beginning in 1980, the Metsähovi quasar monitor-ing program has been the most comprehensive such programat high radio frequencies. The Metsähovi sample contains 157individual sources including about 100 of the brightest radio-loud AGN in the Northern hemisphere (declination   − 10 ◦ ),which are observed at 22, 37 and 87 GHz (see Teräsranta et al.1998 for details). The sample also includes the Northern 2 Jycatalogue of flat spectrum sources (Valtaoja et al. 1992a) ful-filling the following criteria:  δ   ≥  0 ◦ ,  α Kuhr (2 . 7 − 5  GHz )  ≥− 0 . 5 ( S   ∝  ν  α ) , with  α  taken from the catalogue of Kühr etal. (1981), and  S  max (22  GHz )  ≥  2  Jy. Of the 27 sources inour study, 13 belong to this 2 Jy catalogue. Of the 14 sourcesthat do not belong, 5 have declination below  0 ◦ and the restwere fainter than 2 Jy at 22 GHz prior to 1992. In our study,we used 22 and 37 GHz Metsähovi data from 1990 to 1998 to-gether with 22 and 43 GHz VLBA maps. Comparing 37 GHzTFD-data with 43 GHz VLBA maps is justified by the typicallyflat spectra of our sources in the millimetre region.Our sample of 27 sources with good VLBI and TFD dataconsists of 12 high optical polarization quasars (HPQs), 7 lowoptical polarization quasars (LPQs), 7 BL Lacertae objects(BLOs) and one object classified as a radio galaxy (GAL). Thepercentage of each class of radio-loud AGN is given in Table 2for our sample, for the 2 Jy sample (Valtaoja et al. 1992a), andfor a sample containing the EGRET blazar identifications thathave a high probability of being correct (Mattox et al. 2001).As one can see, our sample is very similar to the  γ  -ray blazarsas well as to the 2 Jy sample representing the brightest radio-loud AGN. [The one radio galaxy in our sample, 0446+112, isa less certain EGRET identification, which is the reason why itis not included in the list by Mattox et al. (2001).] This supportsthe notion that the results presented in this paper are applicableto all  γ  -ray blazars and, to some extent, to radio-loud AGN ingeneral.  T. Savolainen et al.: Connections between millimetre continuum variations and VLBI structure in 27 AGN 3 Table 1.  List of sources in our sample. Here  z   is the redshift and  N   is the total number of VLBA observations.Source Other desig. Class  z   Epochs 22 GHz 43 GHz  N  0202+149 HPQ 0.833 1995-97 – + 40219+428 3C 66A BLO 0.444 1995-97 + + 70234+285 CTD 20 HPQ 1.207 1995-97 + – 40235+164 AO 0235+164 BLO 0.94 1995-96 – + 60420 − 014 OA 129 HPQ 0.915 1995-97 + + 80446+112 GAL 1.207 1995-97 – + 40458 − 020 HPQ 2.286 1995-97 – + 50528+134 LPQ 2.07 1994-97 + + 110716+714 BLO >0.2 1995-97 + – 90804+499 OJ 508 HPQ 1.43 1996-97 + – 30827+243 LPQ 2.046 1995-97 + + 60836+710 HPQ 2.17 1995-97 + + 70851+202 OJ 287 BLO 0.306 1995-96 – + 70954+658 BLO 0.367 1995-96 + – 31101+384 Mkn 421 BLO 0.031 1995-97 + – 81156+295 4C 29.45 HPQ 0.729 1995-97 + – 51219+285 ON 231 BLO 0.102 1995-97 + – 31222+216 4C 21.35 LPQ 0.435 1996-97 + – 21226+023 3C 273 LPQ 0.158 1993-95 + + 51253 − 055 3C 279 HPQ 0.538 1993-97 + + 101510 − 089 HPQ 0.361 1995-97 – + 51606+106 LPQ 1.226 1994-97 + + 51611+343 DA 406 LPQ 1.401 1994-97 + + 111633+382 4C 38.41 LPQ 1.814 1994-96 + – 61741 − 038 HPQ 1.054 1995-97 + – 22230+114 CTA 102 HPQ 1.037 1995-97 – + 72251+158 3C 454.3 HPQ 0.859 1995-96 – + 12 3. Connections between millimetre flux curvesand VLBI components In order to compare TFD and VLBI events in our data, wedecompose the total flux density variations in the Metsähovi22/37 GHz flux curves into exponential flares of the form ∆ S  ( t ) =  ∆ S  max e ( t − t max ) /τ  , t < t max ∆ S  max e ( t max − t ) / 1 . 3 τ  , t > t max .  (1)Here  ∆ S  max  is the maximum amplitude of the flare,  t max  is theepoch of the flare maximum and  τ   is the flare rise timescale. Ithas been shown earlier that all TFD variations can be modelledto surprising accuracy with a small number of flares consist-ing of an exponential rise, sharp peak and exponential decaysuperposed on a constant baseline flux (Valtaoja et al. 1999).This decomposition (described in more detail by Valtaoja et al.1999) helps us to identify and isolate individual events as wellas to estimate amplitudes and timescales of the outbursts.Next we plotted the TFD decompositions and the flux vari-ations of the VLBI components for each source. Two exam-ples illustrate the results of this comparison: 1633+382 (4C38.41) and PKS 2230+114 (CTA 102) are shown in Fig. 1 andinFig.2.Thecomponent identifications canbefoundinJorstadet al. (2001a). Even at first glance, it is evident that there is aclear connection between the millimetre continuum variationsand the VLBI component fluxes. Whenever there are enoughVLBA observations, the summed flux curve of the VLBI com-ponents is similar to the continuum flux curve; only the ampli- Table 2.  The percentage of each class of the radio-loud AGN in threedifferent samples.[HPQ= high optical polarization quasar, LPQ= lowoptical polarization quasar, BLO = BL Lacertae object, GAL = radiogalaxy and N/A = No exact classification available]Class Our sample 2 Jy sample EGRET blazarsHPQ 44 % 28 % 37 %LPQ 26 % 32 % 30 %BLO 26 % 28 % 26 %GAL 4 % 8 % 0 %N/A 0 % 4 % 7 % tude of the former is ∼ 90 % that of the latter. This is expectedwith the missing 10% of the flux in the VLBI maps probably just due to the insensitivity of high-frequency VLBI to diffuseemission. There is a slight time shift between the 37 GHz TFDcurves and the 43 GHz VLBI component flux curves. This isunderstandable according to the shock models, since the max-imum amplitude of the flare moves from high frequencies tolower frequencies as the shock evolves.A much more interesting result is that for every superlumi-nal ejection seen in the VLBA data, the TFD decompositionshows a coinciding flare. We examine ejections having zeroepochs after the year 1990. For most of our sources, MetsähoviTFD monitoring is rather sparse before this and therefore notsuitable for our comparison. We exclude two ejections becauseof large gaps in the Metsähovi flux curve at their zero epochs  4 T. Savolainen et al.: Connections between millimetre continuum variations and VLBI structure in 27 AGN 1994 1994.5 1995 1995.5 1996 1996.5 1997 1997.5 199801234    F   l  u  x   [   J  y   ] 1633+382 (4C 38.41) − 22 GHz 22 GHz TFD dataModel fit curve19941994.519951995.519961996.519971997.5199800.511.522.53    F   l  u  x   [   J  y   ] Modelled peaks 1995.21995.4 Model flare1994 1994.5 1995 1995.5 1996 1996.5 1997 1997.5 19980123 Year    F   l  u  x   [   J  y   ] VLBI − 22 GHz Core B3 B1 B2 Indiv. knots Fig.1.  An example (quasar 1633+382) of a graph containing TFD measurements, exponential flare model fits, and individual VLBI componentflux density vs. time. The top panel presents total flux density observations at 22 GHz from the Metsähovi quasar monitoring program (dots) andour fit (curve), which is a sum of individual exponential model flares superposed on a constant baseline flux. The modelled flares are shown inthe middle panel with the epochs of the maximum flux density indicated. The bottom panel displays the flux evolution of the VLBI componentsat 22 GHz from the VLBA observations by Jorstad et al. (2001a), whose component designations we adopt; the errors in flux density of theVLBI components are approximately 0.05 Jy. (the observation gap in 1994). We require that there be at leastthree observations of the ejected component and that the ob-served flux density of the component be greater than 0.1 Jy (theapproximate noise level of Metsähovi observations) at sometime. In our data, there are 29 ejections of VLBI componentsfulfilling the above criteria (see Table 3). The TFD flares corre-sponding to these 29 ejections are identified by comparing thecomponent ejection times with the beginning times of the TFDflares, as well as by comparing the light curves of the VLBIcomponents with those of the decomposed TFD flares.We define the beginning of an exponential TFD flare as t 0 , TFD  =  t max  −  τ  , where  τ   is the variability timescale(e-folding time). This definition gives the point where S  ( t 0 , TFD ) =  S  max e  . While there is no mathematical sense indefining the beginning of an exponential function, in realitythere must be a starting point to a flare. We could instead es-timate the beginning of the flare as the previous local mini-mum of the flux curve ( t lm ). If we compare  t lm  to  t 0 , TFD  foran outburst that starts just after a local minimum, we see thatthe average time difference between the two is 0.0 years witha standard deviation of 0.4 years (see Fig. 3). Therefore, theaverage values of   t 0 , TFD  and  t lm  are the same. When two ormore closely spaced outbursts blend together, the local mini-mum is no longer a good indicator of the start of the flare. Insuch a case the local minimum is near the peak rather than thebeginning of the later flare. Hence,  t 0 , TFD  is a more reliableand practical starting point to a flare.We compare the extrapolated ejection epochs of the super-luminal knots (from Jorstad et al. (2001a)) with the beginningtimes of the TFD flares. In 28 of the 29 cases we find a TFDflare that occurred within 0.5 yr of the ejection epoch. The onlyexception is component E2+B1 of 3C 279, for which  t 0 , VLBI is not very well determined.The frequency of large TFD flares ( ∆ S >  0 . 3 · S  quiescent )estimated from the Metsähovi data is 1 per 1.6 years. On theother hand, the frequency of   observed   superluminal ejections  T. Savolainen et al.: Connections between millimetre continuum variations and VLBI structure in 27 AGN 5 1993 1994 1995 1996 1997 1998 19990246810    F   l  u  x   [   J  y   ] PKS 2230+114 (CTA 102) − 37 GHz 37 GHz TFD dataModel fit curve19931994199519961997199819990246    F   l  u  x   [   J  y   ] Modelled peaks 1997.71998.21994.71993.8Model flare1993 1994 1995 1996 1997 1998 19990246Year    F   l  u  x   [   J  y   ] VLBI − 43 GHz Core B3 B2 B1 C Indiv. knots1996.3 Fig.2.  Another sample TFD curve, exponential flare model fits, and individual VLBI component TFD values as a function of time. The 1998.48data is from VLBA observations by Fredrik Rantakyrö (Rantakyrö et al. 2002). See the caption of Fig. 1 for details. is approximately 1 per 2.3 years. Using these values we cal-culate the probability that a superluminal ejection could occurby random chance within a time interval  dt  before or after thebeginning of the TFD flare. The results are given in Table 4.For every applied  dt  range the expected number of random oc-currences is clearly much lower than the observed number of coincidences. The probability that 28 out of 29 ejections wouldbe observed to occur randomly within 0.5 yr of the beginningsof TFD flares is  <  10 − 7 . Hence, the correspondence between t 0 , VLBI  and  t 0 , TFD  is real at a very high level of significance.We therefore find that, at high radio frequencies, the startof a TFD flare precedes the arrival of a new superluminal knotat the position of the brightness centroid of the core of the jet.However, we do not have enough VLBI data to say if the con-verse is true, i.e., whether there is a new VLBI component forevery TFD flare. When we see the new VLBI component forthe first time, the flux of the TFD flare is usually already de-creasing. This behaviour is analysed in section 4, in which wediscuss the so-called core flares.The mean time difference between the zero epoch of theVLBI components and the beginning of the TFD flares  ∆ t  = t 0 , VLBI  − t 0 , TFD  is  +(0 . 19 ± 0 . 04)  yr (ignoring componentE2+B1 in 3C 279). The extrapolated ejection time of a VLBIcomponent is therefore  ∼  0 . 2  yr  after   the beginning of theassociated TFD flare, on average. This may indicate that theproper motion of a typical VLBI knot accelerates during theearly stages in the component’s evolution. On the other hand,we note that  t 0 , VLBI  is the moment when the component is co-incident with the  brightness centroid   of the core. In this case,the TFD flare might begin when the disturbance that creates theshock first hits the inner edge of the core, which would occurbefore  t 0 , VLBI .The fluxes of the VLBI components and the decomposedTFD flares are correlated. In Fig. 4 we plot the VLBI compo-nent fluxes vs. the decomposed TFD flare fluxes at the VLBIepochs (from our exponential-flare model fits). The Spearmancorrelationcoefficientofthisgraphis r S  = 0 . 817 ;theprobabil-ity that  r S  would be this high from uncorrelated data ∼ 10 − 14 .The linear Pearson correlation coefficient  r P  = 0 . 76 , whichcorresponds to a probability of  ∼  10 − 12 that the correlation isby chance. Furthermore, in 59% of the cases the fluxes differby less than a factor of two. There is therefore a clear connec-
Related Documents
View more...
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!