The innermost region of the water megamaser radio galaxy 3C403
A. Tarchi, A. Brunthaler, C. Henkel, K. M. Menten, J. Braatz, A. Weiss
aa r X i v : . [ a s t r o - ph ] S e p Astronomy & Astrophysics manuscript no. 8317pap c (cid:13)
ESO 2018November 1, 2018
The innermost region of the water megamaser radio galaxy3C 403
A. Tarchi , , A. Brunthaler , C. Henkel , K. M. Menten , J. Braatz , and A. Weiß , INAF-Osservatorio Astronomico di Cagliari, Loc. Poggio dei Pini, Strada 54, 09012 Capoterra (CA), Italye-mail: [email protected] INAF-Istituto di Radioastronomia, Via Gobetti 101, 40129 Bologna, Italy Max-Planck-Institut f¨ur Radioastronomie, Auf dem H¨ugel 69, D-53121 Bonn, Germany National Radio Astronomy Observatory, 520 Edgemont Road, Charlottesville, VA 22903 IRAM, Avenida Divina Pastora 7, 18012 Granada, SpainReceived ; accepted
ABSTRACT
Context.
The standard unified scheme of active galactic nuclei requires the presence of high column densities of gas anddust potentially obscuring the central engine. So far, few direct subarcsecond resolution studies of this material havebeen performed toward radio galaxies.
Aims.
The goal of this paper is to elucidate the nuclear environment of the prototypical X-shaped Fanaroff-Riley typeII radio galaxy 3C 403, the only powerful radio galaxy known to host an H O megamaser.
Methods.
Very Large Array A-array and single-dish Green Bank and Effelsberg 1.3 cm measurements were performedto locate and monitor the water maser emission. Very Long Baseline Interferometry 6 cm continuum observations weretaken to analyze the spatial structure of the nuclear environment at even smaller scales, while the CO J =1–0 and 2–1transitions were observed with the IRAM 30-m telescope to search for thermal emission from a spatially extended,moderately dense gas component. Results.
Positions of the H O maser features and the continuum emission from the core coincide within 5 mas (5.5 pc).Intensities of the two main maser components with (isotropic) luminosities sometimes surpassing 1000 L ⊙ appear to beanti-correlated, with typical timescales for strong variations of one year. If the variations are intrinsic to the cloud(s),the implied angular source size would be < ∼ > ∼ × K. The VLBI continuumobservations support a scenario where a nuclear core, represented by the dominant central radio continuum component,is accompanied by a jet and counterjet, directed toward the western and eastern large scale lobes of the galaxy. COremains undetected, providing a maximum scale size of ∼
50 pc × (500 K/ T b ) / , with T b denoting the brightnesstemperature of the CO J =1–0 line. Possible scenarios that could produce the observed maser emission are outlined.Adopting a mass of several 10 for the nuclear engine, the observed maser features can only be interpreted in terms of anaccretion disk as in NGC 4258, if they solely represent the systemic velocity components. The receding and approachingparts of the putative maser disk are, however, not seen and a secular velocity drift of the observed features is not(yet) apparent. Most likely, the two main maser components mark shocked molecular gas interacting with the nuclearjets. The X-shaped morphology of the radio galaxy may point at a binary nuclear engine. This possibility, greatlycomplicating the nuclear environment of 3C 403, should motivate a number of worthwhile follow-up studies. Key words.
Galaxies: individual: 3C 403 – Galaxies: active – Galaxies: ISM – masers – Radio lines: ISM – Radio lines:galaxies
1. Introduction
The physical conditions in active galactic nuclei (AGN) areunique in the cosmos. Stellar and gas densities are excep-tionally large and enormous amounts of energy and angu-lar momentum are released as the material accretes ontothe supermassive nuclear engine. Performing studies of thestructure, kinematics and excitation of this material is thesole mean available to investigate (super)massive ultracom-pact objects.Because of obscuration and crowding, only a few molec-ular tracers can provide direct information on the in-nermost regions of active galaxies. The H O λ =1.3 cmmaser line, profiting from a long wavelength (no obscura-tion by dust grains), from its enormous luminosity (up to Send offprint requests to : A. Tarchi L ⊙ (isotropically), corresponding to ∼ photons/s;see Koekemoer et al. 1995; Barvainis & Antonucci 2005)and the availability of Very Long Baseline Interferometry(VLBI) networks, is a particularly suitable tracer of this en-vironment. The line samples gas with kinetic temperaturesof several 100 K and densities of ∼ cm − (e.g., Kylafis& Norman 1987, 1991).So-called ”disk-masers” as in NGC 4258 allow us to mapnuclear accretion disks (e.g. Greenhill et al. 1995; Miyoshiet al. 1995; Herrnstein et al. 2005). Analysis of the dynamicsof observed Keplerian disks leads to estimates of the mass ofthe nuclear engine and to a determination of the distance tothe galaxy, independent of any standard candles. So-called”jet-masers” as in Mrk 348 (Peck et al. 2003) are associatedwith the pc-scale jets and provide estimates, through rever-beration mapping, of the speed of the material in the jet. A Tarchi et al.: The innermost region of 3C 403 further class of masers is associated with nuclear outflows(Circinus; Greenhill et al. 2003).Measuring the proper motion of galaxies has been an-other long-standing problem. The solution is to find suf-ficiently luminous ultracompact sources that can be ob-served with VLBI techniques involving phase referenc-ing. Extragalactic H O masers are suitable targets andcould provide, through their measured radial velocities andproper motions, three dimensional velocity vectors in space(Brunthaler et al. 2005, 2007), not only for the masersthemselves but also for the nuclei of their parent galaxies.Based on the assumption that the water masers are lo-cated in circumnuclear tori, (often) aligned with the nucleusand amplifying its continuum, linearly for saturated masersand exponentially for unsaturated ones, one should expecta high detection rate in radio-loud galaxies (in particularthose with the radio axis close to the plane of the sky).However, water megamasers have been found so far only inlow-radio power AGN (P < − W Hz − ), mostlyin galaxies classified as Sy 2s or LINERs (Braatz et al. 1996,1997).A first systematic search for the λ =1.3 cm line in ra-dio galaxies was performed by Henkel et al. (1998). Nomaser was detected in a sample of ∼
50 Fanaroff-Riley typeI (FR I) galaxies. Subsequently, targets belonging to themore powerful Fanaroff-Riley type II (FR II) class wereobserved by two groups (Tarchi et al., Lara et al., unpub-lished) again with negative results.According to the current paradigm for radio-loud AGN,FR II galaxies host type 2 AGN with narrow optical emis-sion lines and prominent, extended nuclear jets. Since thejets are ejected at a small angle with respect to the planeof the sky, FR II galaxies should contain a nuclear torusbeing viewed approximately edge-on (for a review, see e.g.Urry & Padovani 1995). So far, however, the presence andnature of these hypothesized tori has remained an enigma.The nuclear region of the prototypical FR II galaxy Cyg Ashows evidence for dust and obscuration with gas columndensities up to several 10 cm − . Morphology and physicalparameters of the nuclear region as well as the mechanismtriggering nuclear activity are, however, not yet fully un-derstood (e.g., Conway & Blanco 1995, Carilli & Barthel1996; Fuente et al. 2000; Bellamy & Tadhunter 2004).Very recent modifications to the standard unifiedscheme (hereafter SUS) state that FR I galaxies (or at leastthe majority of them) seem to lack a geometrically andoptically thick molecular torus. This scenario is now sup-ported by an increasing number of studies (e.g., Chiabergeet al. 1999; Perlman et al. 2001; Verdoes Kleijn et al. 2002;Whysong & Antonucci 2004). In order to account for the ab-sence of broad emission lines in narrow–lined FR IIs (withinthe framework of the SUS believed to represent the parentpopulation of radio-loud quasars) a geometrically and opti-cally thick obscuring layer is required instead (Barthel 1989and references therein).Recently, a luminous ( L H O ∼ L ⊙ ) H O megamaserhas been detected in 3C 403 (Tarchi et al. 2003; hereafterTHC), an FR II galaxy at roughly the same distance asCyg A. 3C 403 is the first and so far only powerful radiogalaxy with a detected H O maser.The kpc radio morphology of 3C 403 is typical for X-shaped radio galaxies. The cause of the “X” radio shapeis still debated. Arguments exist that support the originof this peculiar morphology as due to a sudden change in the jet direction. Merritt & Eckers (2002) have argued thatthe sudden change in the jet direction reflects a sudden ( < few yr) change in the black hole spin axis due to a blackhole merger (see also Sect. 5.2.2). Alternative models implyinstead slow precession of the jet axis or lobe backflow (seeDennett-Thorpe et al. 2002, Schoenmakers et al. 2000).The Very Large Array (VLA) λ =3.6 cm maps of Blacket al. (1992; hereafter BBL; their Fig. 13) and Dennett-Thorpe et al. (2002; hereafter DSL; their Fig. 1) show twobright radio lobes towards the east and west with hot-spotsand two weaker, longer and broader wings to the north-west and south-east. The two bright lobes are caused byjets containing a number of prominent knots. Adopting aWMAP cosmology ( H =71, Ω M =0.27, Ω λ =0.73; Spergel etal. 2003), the redshift of z =0.059 leads to an angular sizedistance of D A ∼
230 Mpc, at which 1 mas corresponds to ∼
2. Observations and data reduction
Effelsberg
3C 403 was observed in the 6 − transi-tion of H O (rest frequency: 22.23508 GHz) with the 100-mtelescope of the MPIfR at Effelsberg at four epochs be-tween January 2003 and April 2004. The beam width was40 ′′ . The observations were made with a 18–26 GHz HEMTreceiver that was sensitive to the two orthogonal linearpolarizations. A dual beam switching mode with a beamthrow of 2 ′ and a switching frequency of ∼ ∼ T ∗ A ). Thebeam efficiency was η b ∼ ′′ GBT
Observations were made with the Green BankTelescope (GBT) of the NRAO during six sessions be-tween 2003 October 20, and 2005 February 15. We usedthe 18–22 GHz K-band receiver, which has two beams ata fixed separation of 3 ′ in azimuth. The GBT beamwidthis 35 ′′ at the red shifted maser frequency of 21 GHz, andpointing uncertainties were better than 10 ′′ . The data weretaken in total power mode, and the telescope was noddedbetween two positions on the sky such that the source was The 100-m telescope at Effelsberg is operated by the Max-Planck-Institut f¨ur Radioastronomie (MPIfR) on behalf of theMax-Planck-Gesellschaft (MPG). The National Radio Astronomy Observatory (NRAO) is afacility of the National Science Foundation operated under co-operative agreement by Associated Universities, Inc.archi et al.: The innermost region of 3C 403 3 always in one of the two beams during integration. We useda nod cycle of 2 minutes per position.The spectrometer was configured with two 200 MHzbandpasses, one centered on the systemic velocity of 3C 403and the second redshifted by 180 MHz. No emission was de-tected beyond what is shown in Fig. 1. The zenith systemtemperature was between 35 and 45 K. Atmospheric opacitywas estimated using system temperature and weather data,and ranged from 0.04 to 0.06 at the zenith. We reduced thedata using GBTIDL.
VLA
Water maser emission was observed at the red shiftedmaser frequency of 21 GHz on July 23 and 25, 2003, withthe Very Large Array (VLA) of the NRAO in its A con-figuration for an 8 hour total on-source time. We observedwith two 12.5 MHz IFs centered on the two maser featuresdetected by Tarchi et al. (2003; see the upper spectrum ofFig. 1). Each IF was split into 32 channels providing a chan-nel spacing of ∼ − . The source 1331+305 (2.66 Jy)was used as flux and bandpass calibrator. The point source1950+081, at a distance of 5 . ◦ ∼
360 km s − . The data were Fourier-transformed using nat-ural weighting to create a 256 × ×
64 data cube. The ra-dio continuum was subtracted using the AIPS task UVLSF.This task fits a straight line to the real and imaginary partsof selected channels and subtracts the fitted baseline fromthe spectrum, optionally flagging data having excess resid-uals. In addition, it provides the continuum as a UV dataset, which has been used to create continuum maps of thegalaxy. The restoring beam is 0 . ′′ × . ′′
1. The rms noiseper channel becomes 0.6 mJy, consistent with the expectedthermal noise.
EVN
3C 403 was observed on May 20, 2004, with theEuropean VLBI Network (EVN) at 6 cm. Eight 8 MHzbands, each at right and left circular polarization, were em-ployed. The total observing time was 10.5 hours. We used4C +02.49 as a phase-reference source and switched every2 minutes between the two sources. The initial calibrationwas performed with the AIPS package. A priori amplitudecalibration was applied using system temperature measure-ments and standard gain curves. Fringes were found in the3C 403 data itself on all baselines, so we used this source asphase-reference and applied the solutions also to 4C +02.49.The data were self-calibrated and mapped using the soft-ware package DIFMAP (Shepherd et al. 1994). We startedwith phase-only self-calibration and later included phase-amplitude self-calibration with solution intervals slowly de-creasing down to one minute. Pico Veleta
3C 403 was observed in a search for CO emis-sion on November 6, 2004, using the IRAM The European VLBI Network is a joint facility of European,Chinese, South African and other radio astronomy institutesfunded by their national research councils. IRAM is supported by INSU/CNRS (France), the MPG(Germany), and the IGN (Spain). receivers tuned to 217.69 GHz. Beamwidths were 23 ′′ and13 ′′ , respectively. Spectra were obtained using the wobblerswitch technique with a beam throw of 60 ′′ and a switch-ing frequency of 0.5 Hz. The data were recorded using the1 MHz (512 channels, 1 MHz channel spacing) and 4 MHz(256 channels, 4 MHz channel spacing) filterbanks for the109 and 218 GHz observations, respectively. The effectivevelocity coverage is 1400 km/s at both frequencies. Typicalsystem temperatures were 160 and 500 K ( T ∗ A ). The fastswitching procedure and the good quality of the resultingbaselines also provide a rough measure of continuum levels(see Sect. 4.1.3). 3C 403 was observed for about 70 minutes(on+off) leading to rms noise levels of 4.2 and 12.5 mJyafter smoothing the 109 and 218 GHz spectral bands tochannel widths of 44 and 50 km s − .
3. The systemic velocity of 3C 403
For the following sections it is of fundamental relevance toassess the accuracy of the reported systemic velocity V sys of 3C 403. Optical emission line measurements indicate V sys = 17688 km s − ( z = 0.059) with no uncertainty value pro-vided (Spinrad et al. 1985 and references therein). Previousstudies indicate that recessional velocities of galaxies de-rived from optical emission lines tend to have a systematicuncertainty, often being smaller than systemic velocities de-rived from H i (e.g. Mirabel & Wilson 1984; Morganti et al.2001). This difference is attributed to a net outflow of gas inthe optically unobscured front side of narrow emission lineregions (e.g. Mirabel & Wilson 1984). Unfortunately, noneutral hydrogen has been so far detected in 3C 403 (seeSect. 5.1.2) to perform a comparison with the optically-derived recessional velocity. Baum et al. (1990), while notproviding a recessional velocity, derived on the basis of [O i ],[N ii ], H α and [S ii ] data a rotation curve of 3C 403 that in-dicates no strong anomalies. From this we conclude thatthe uncertainty in V sys , as derived by Spinrad et al. 1985,should not exceed 100 km s − .
4. Results
Figure 1 shows the water maser spectra from 3C 403 ateleven epochs from January 2003 to February 2005, ob-served with the Effelsberg telescope, the GBT, and the VLAA-array. Past Effelsberg unpublished data from December1997 showed no sign of emission at a 3 σ level of 60 mJy.Fig. 2 displays our most sensitive spectrum, that obtainedwith the VLA indicating the possible presence of three fea-tures that, for the sake of convenience, we indicate as com-ponent I, II, and III.The most obvious change that can be deduced fromFig. 1 is that, in the early spectra, emission is mostlypresent on the red-shifted side (component I), while latermost of the emission (component II) is blue-shifted w.r.t.the systemic velocity. In January and March 2003, com-ponent I had a peak line strength of 20–25 mJy, becom-ing weaker at subsequent epochs and remaining undetectedin the most recent spectra ( < ∼ Tarchi et al.: The innermost region of 3C 403
EFF 2003−03−25VLA 2003−07−23GBT 2005−01−21GBT 2003−10−20EFF 2003−01−11GBT 2004−01−21EFF 2004−01−29GBT 2004−01−31GBT 2004−03−17EFF 2004−04−09GBT 2005−02−15
Fig. 1.
Maser lines in 3C 403 observed with Effelsberg (fourepochs), the GBT (6 epochs), and the VLA (one epoch).The spectrum from the first epoch (upper panel) was al-ready published in THC. The channel spacing is 312 kHz( b = 4.5 km s − ), 195 kHz ( b = 2.8 km s − ), and 391 kHz ( b =5.6 km s − ) for the spectra taken with Effelsberg, the GBT,and the VLA, respectively. Zero velocity corresponds to afrequency of 20996.28 MHz (c z = 17688 km s − w.r.t. theLocal Standard of Rest (LSR)), the galaxy’s recessional ve-locity according to the NASA/IPAC database (NED)to January 2004, component II seems to experience a firstshift from about –45 to –95 km s − and a second one backto about –60 km s − in February 2005. Of course, the pos-sibility that there is actually no shift and we are, instead,witnessing the brightening and fading of different compo-nents is also a possibility. Component III, detected at a lowsignificant level only in the first three spectra, fades belowthe 5 mJy noise level together with component I.While no systematic regular velocity drift involving allthe feature can be confidently observed, it is remarkablethat within about 15 months the maser lines swapped theirintensities. A time scale of one year and peak flux den-sities of order 20 mJy indicate an angular source size of II III I
Velocity (km/s) F l u x d e n s i t y ( m Jy )
05 0−100 100 200
Fig. 2.
The main water maser features in 3C 403 observedwith the VLA A-array on July 23, 2003 (see Sect. 4.1.2 andFig. 1). The dashed vertical line marks the nominal systemicvelocity of the galaxy, V sys = c z = 17688 km s − . Fig. 3.
Close-up of the region around the nucleus of 3C 403.Shown as a filled triangle, a cross, and a filled square arethe variance-weighted average locations with error bars ofmaser components I, II, and III, respectively. Position off-sets in (RA
J2000 , Dec
J2000 ) are relative to the 22-GHz con-tinuum peak position. < ∼ > ∼ × K,if variations are intrinsic to the cloud(s) and are not trig-gered by fluctuations of the continuum background. Eachindividual component reaches peak isotropic luminositieswell in excess of 1000 L ⊙ . So far, only four other masersources were reported to have a similar power, TXS22226–184 (Koekemoer et al. 1995), Mrk 034 (Henkel et al. 2005),J0804+3607 (Barvainis & Antonucci 2005), and UGC 5101(Zhang et al. 2006). A variation by more than a factor oftwo has not yet been reported in such a luminous masercomponent. Less luminous masers with linewidths in ex-cess of 10 km s − are commonly more stable. O maser
In all VLA velocity channels (see Fig. 2; for a preliminaryreport, see also Tarchi et al. 2005) the emission peaks atthe position RA
J2000 = 19 h m . s
80, Dec
J2000 = 02 ◦ ′ . ′′
2. The nominally estimated error for absolute positionsfor a VLA A-array map is 0 . ′′ archi et al.: The innermost region of 3C 403 5 Fig. 4. − , respec-tively. The zero velocity corresponds to the recessional ve-locity of the galaxy ( V sys = 17688 km s − ).The total integrated intensity (moment-0; inmJy km s − beam − ) VLA plots of components I–IIIshow that they are unresolved at the 0 . ′′ σ rel = p ( θ l / · SN R l ) + ( θ c / · SN R c ) = p (0 . / · + (0 . / · ∼ . ′′ b = 5.5 pc, where θ denotes the restored beam size, and l and c refer to lineand continuum emission, respectively.Fig. 3 presents a close-up of the region around the nu-cleus of 3C 403, showing the relative distribution of masercomponents I, II, and III. The symbols and error barsmark the variance-weighted average locations and uncer-tainties of the three maser components derived fitting,with the AIPS task JMFIT , the maser emission in eachchannel with flux density > − . Our result indicates an upper limit for theoffset between different components of 12 ± ∼ ′ × ′ map, cover-ing the entire radio extent of the galaxy and no additionalmaser spots were detected above the 0.6 mJy level given inSect. 2. Motivated by the detection of megamaser emission, wesearched for the two ground rotational transitions of themost common observable molecule, CO. Unlike 22 GHzH O, the CO (1–0) and (2–1) lines predominantly tracecool, low-density gas. No clear emission is seen down to3 σ noise limits of 13 and 38 mJy for the two respectivetransitions (this refers to the unsmoothed spectra with 1and 4 MHz channel spacing, see Sect. 2). We tentatively de-tect emission at ∼ –300 km s − with respect to the sys-temic velocity (Sect. 3). Evans et al. (2005) also reporteda non-detection of CO toward 3C 403 as part of a surveyof IRAS radio galaxies. Curiously, also in their spectrum For a discussion on the accuracy of the parameters derivedfrom JMFIT, see e.g. Henkel et al. (2004; their Sect. 3) (their Fig. 1) a highly tentative feature is present at about–300 km s − . While the offset is opposite to that expectedfrom optical emission lines (Sect. 3) and while evidence forsuch a feature is too weak to seriously question the reces-sional velocity adopted throughout this paper, a CO spec-trum with higher sensitivity would be desirable.Indicated by the baselines, continuum emission wasalso detected. At 109 GHz, the level is approximately 25 ±
13 mJy (the uncertainty was estimated from the varia-tions of the continuum level in individual scans). No con-tinuum was detected at 218 GHz.
In Fig. 5, we present a naturally weighted VLA A-array22 GHz radio continuum image of the nuclear region of3C 403, produced using maser emission-free channels. Thesource is spatially unresolved with a peak flux density of40 mJy beam − . However, an elongation is apparent in theuniformly weighted image of Fig. 6 that might be a weaksignature of a parsec scale jet visible in the higher resolutionimages presented in Sect. 4.2.2.As for the H O line emission, we have also investigateda ∼ ′ × ′ area in a search for additional radio continuumemission. Apart from the nucleus, among the radio contin-uum features (lobes, radio jets, and bright knots) reportedby BBL and DSL, we have only detected emission at po-sition RA J2000 = 19 h m s , Dec J2000 = 02 ◦ ′ ′′ ,coincident within the errors with the exceptionally brightradio knot F6 (nomenclature according to BBL). The knotin our map (Fig. 7) seems to be resolved into two compo-nents aligned almost perpendicular to the large scale radiojet. Using a 500 k λ taper function, weak emission is alsodetected at the position where two compact structures inthe eastern hotspot (F1 and F2, following BBL) have beenpreviously observed (Fig. 8). The remaining radio emissionfrom 3C 403 detected by DSL is either resolved out or isbelow the detection threshold given in Sect. 2. Milliarcsecond resolution 6-cm VLBI maps of the radio nu-cleus of 3C 403 are shown in Figs. 9 and 10. The visibilitiesof 3C 403 demonstrate that the source is resolved at longerbaselines (Fig. 11). The amplitude drops from 70 mJy at theshortest baselines (10 M λ ) to 40 mJy at 60 M λ .The uniformly weighted VLBI map of 3C 403 is shown inFig. 9. Notable are extensions to the north-east and south-west. We fitted circular Gaussian components to the uv-data, and the best fit with three components gave a reduced χ per degree of freedom of ∼ χ p.d.f to2.2 and 2.9, respectively. On the other hand, adding morecomponents does not significantly improve the fit. Hence,we used a three component model and the results from thisfit are shown in Table 1. We fitted the three componentsmany times with different initial parameters and used thescatter in the results to estimate the uncertainties of thefit parameters. While the flux densities and angular sizes ofthe three components are not strongly constrained by thefit, the position angle is well determined. The orientation of Tarchi et al.: The innermost region of 3C 403 D E C L I NA T I O N ( J2000 ) RIGHT ASCENSION (J2000)19 52 15.86 15.84 15.82 15.80 15.78 15.76 15.74 15.7202 30 25.225.024.824.624.424.224.023.823.623.423.2
Fig. 5.
A naturally weighted 22 GHz VLA radio continuumimage (resolution: 0 . ′′
1, corresponding to ∼
115 pc) of thenuclear region of 3C 403, made using channels free of lineemission. The peak is at 40 mJy beam − . The contours are(–1, 1, 2, 4, ..., 32) × − . D E C L I NA T I O N ( J2000 ) RIGHT ASCENSION (J2000)19 52 15.86 15.84 15.82 15.80 15.78 15.76 15.74 15.7202 30 25.225.024.824.624.424.224.023.823.623.423.2
Fig. 6.
A uniformly weighted 22 GHz VLA radio continuumimage (resolution: 0 . ′′
08, corresponding to ∼
90 pc) of thenuclear region of 3C 403, made using channels free of lineemission. The peak is at 39 mJy beam − . Contours are (–1,1, 2, 4, ..., 32) × − .the pc-scale jet is in very good agreement with the kpc-scalejet seen on VLA images.Since the telescope in Urumqi (China) could not takepart in the observations, all information on the long base-lines is based on only two telescopes, Hartebeesthoek(South Africa) and Shanghai (China). Hence the data onlong baselines, mainly responsible for the uncertainties inthe model fit parameters in Table 1, should be consideredwith some caution. To test the result shown in Fig. 9 we alsoself-calibrated and imaged the data from the short baselinesonly. As before, the amplitude in the visibilities shows aslow decrease towards longer baselines. The resulting mapis shown in Fig. 10. The extension toward the north-eastis visible in this map and there might be also a weak jet D E C L I NA T I O N ( J2000 ) RIGHT ASCENSION (J2000)19 52 17.64 17.62 17.60 17.58 17.56 17.5402 30 34.033.833.633.433.233.032.832.632.432.2
Fig. 7.
A naturally weighted 22 GHz VLA radio contin-uum image (resolution: 0 . ′′
1, corresponding to ∼
115 pc)of the bright radio knot (F6, following DSL) in 3C 403,obtained from channels free of line emission. The peakis at 1.5 mJy beam − . Contours are (–1, 1, 1.5, 2) × − . D E C L I NA T I O N ( J2000 ) RIGHT ASCENSION (J2000)19 52 19.40 19.35 19.30 19.25 19.20 19.15 19.10 19.0502 30 38373635343332
Fig. 8.
A 500 k λ tapered 22 GHz VLA radio continuumimage (resolution: 0 . ′′
33, corresponding to ∼
375 pc) of thecompact hotspot structures (F1 and F2, following BBL) in3C 403, made using channels free of line emission. The peakis at 2.5 mJy beam − . The contours are (–1, 1, 1.5, 2, 2.5) × − .component towards the south-west. From this we concludethat the weak extensions on mas scales are real and are notcaused by amplitude calibration errors.Assuming the presence of a core and one or two nuclearjet components, there are three interpretations: archi et al.: The innermost region of 3C 403 7 Fig. 9.
Full resolution EVN image of the nucleus of 3C 403at λ =6 cm. The contours start at 0.2 mJy and increasewith a factor of 1.5. The beam size of the observations is1.87 mas × . ◦ Fig. 10.
Low resolution EVN image of the nucleus of 3C 403at λ =6 cm using only the continental EVN antennas. Thecontours start at 0.35 mJy and increase with a factor of 1.5.The beam size is 6.05 mas × . ◦ – Component A is the nucleus. Then we are dealing witha one sided jet towards the south-west. – Component B is the nucleus and components A and Care jet and counterjet.
Fig. 11.
The visibilities of 3C 403 clearly show that thesource is resolved on the longer baselines.
Table 1.
Model fit of the three components A, B, and C tothe uv-data. S is the flux density, d denotes the separationfrom the phase center. Comp. S [mJy] d [mas] Size [mas] P.A.A 9.4 ± ± ± ±
3B 45.2 ±
10 0.00 0.15 ± ± ± < ± – Component C is the nucleus and a one sided jet goestowards the north-east.
5. Discussion
To select the most likely of the three scenarios outlinedin Sect. 4.2.2 and to identify the location of the radio nu-cleus, we compare our continuum VLBI maps with the largescale structure of 3C 403. The orientation of component A(P.A.=73 ◦ ; Table 1) is in very good agreement with theeastern large scale jet and the position of a bright knot(73 ◦ ) in its lobe (e.g., BBL, DSL). The south-western jet(component C), as seen in Fig. 9, seems to bend slightly tothe west, thus pointing towards the western large scale jet.In view of the limited uv-coverage w.r.t. long baselines (seeSect. 4.2.2), we consider this latter coincidence as tentative.The presence of the western and eastern large scale jetswith similar flux densities suggests that the jets must lieclose to the plane of the sky. The inner 10 pc of 3C 403(Figs. 9 and 10) show a similar morphology. Thus a scenariowith B representing the core and A and C marking jet andcounterjet is by far the most likely one. If we see the jet andthe counterjet with similar strengths, the orientation of thejets must be close to the plane of the sky. This view is sup-ported by Kraft et al. (2005; hereafter K05). They presentresults from Chandra
X-ray observations and describe theX-ray spectrum of the nucleus as a superposition of twopower-law continuum components with a 6.4 keV Fe line.
Tarchi et al.: The innermost region of 3C 403
One of these power-law components is heavily absorbing( N H ∼ × cm − ).If the X-ray photons are predominantly scattered radi-ation not passing through the main body of the obscur-ing torus, the high column density derived by K05 wouldonly be a lower limit. Since no evidence was found for asignificant amount of dust in observations with the HubbleSpace Telescope (Martel et al. 1999), K05 conclude that theabsorption is due to material close to the nuclear engine,possibly forming a molecular torus. Spectroscopic stud-ies at optical wavelengths revealed narrow emission lines(Tadhunter et al. 1993), indicating that the broad line re-gion is hidden behind an edge-on dusty torus. i and CO Morganti et al. (2001) searched for H i absorption in 3C 403and got an upper limit of N H < × cm − ( T spin =100 K). Large differences in H i and X-ray column densitiesare quite common in Seyfert galaxies (e.g. Gallimore et al.1999). The striking discrepancy between X-ray (Sect. 5.1.1)and H i column densities indicates the predominance of ei-ther ionized or molecular gas along the line of sight towardthe nuclear source of 3C 403. While both kinds of environ-ment may coexist in a small volume surrounding the centralengine, the maser emission provides substantial evidence forthe latter.Why are we then not seeing CO? For a quasi-thermallyexcited line, seen in emission, beam dilution may becomeimportant in an object as distant as 3C 403. With the ob-served brightness temperature as a function of the intrinsicone as well as source and beam solid angle ( T obsb = [ T intb × Ω torus ]/[Ω torus + Ω beam ]) we can determine the small-est toroid size, Ω torus , we would have detected. With a COline detection threshold of T obsmb = 1.8 mK and a 109 GHzbeam size of 23 ′′ (Sect. 2), we obtain Ω torus ∼ . ′′ × (500 K/ T intb ) / that corresponds to a linear scale of ∼
50 pc × (500 K/ T intb ) / . Elliptical galaxies are characterized byhigh stellar velocity dispersions efficiently heating any coolgas and creating an X-ray emitting plasma that is greatlyconfining the volume occupied by molecular clouds (e.g.,Wiklind & Henkel 2001). The lack of dust emission (Martelet al. 1999) supports a small size of the molecular complexin the nuclear region and is therefore consistent with ournon-detection of CO.For the absence of deep CO J =1–0 absorption againstthe nuclear radio continuum source (the absorption mustreach a significant fraction of the continuum level, otherwiseit would not be detected; see Sect. 4.1.3 and Fig. 4), thereare several possible explanations. (1) There is no molecu-lar gas in front of the nuclear source. In view of the col-umn density (K05) this is not likely. (2) The torus is partlyor entirely molecular but has a very small fractional COcolumn density in the J = 0 and 1 states. This may be aconsequence of a high temperature of the gas, possibly cou-pled with an intense radiation field (resulting in excitationtemperatures of several 100 to ∼ O maser hotspots(Sect. 4.1.1) does not provide any guideline, because 22 GHzH O emission may originate from much higher density gas then the ground rotational lines of CO. (4) Alternatively,it is also possible that any existing absorption would be‘swamped’ by emission at the same velocity interval. O maser emission
The main result of our VLA observations is that the two22 GHz H O velocity components of 3C 403 do not arisefrom the extended lobes of the system, but are aligned withthe central radio continuum source within a ∼ O emission beyondthe Magellanic Clouds, there are a total of 51 known mega-maser sources ( L H O >
10 L ⊙ ). Most of these have not yetbeen studied in spatial detail. Omitting 3C 403, the samplecontains two pure jet-masers (apparent interaction betweena nuclear jet and a dense molecular cloud), while 16 sourcesappear to be disk-masers arising from an accretion disk.The latter is either based on detailed observations as in thecase of NGC 4258 (e.g., Miyoshi et al. 1995) or, a little morespeculative, on the measured lineshape (see, e.g., NGC 6323in Braatz et al. 2004). H O megamaser emission associatedwith a nuclear outflow is only known for one source of thesample, the Circinus galaxy. Even here, however, only apart of the emission is associated with the outflow, com-plementing a nuclear accretion disk viewed approximatelyedge-on (Greenhill et al. 2003).In view of the known sample of megamaser galaxies, anassociation of the maser in 3C 403 with a nuclear outflowappears to be unlikely. So far all detected ‘jet-maser’ com-ponents are observed at one side w.r.t. the systemic velocity(Henkel et al. 2005). We should keep in mind,however, that3C 403 is an (elliptical) FR II and not a (spiral) Seyfert 2or LINER galaxy and that ”disk-masers” may be identi-fied more easily than ”jet-masers”. Nevertheless, observedlineshapes and the relatively large number of disk maserssuperficially hint toward H O emission from an accretiondisk as the most plausible interpretation of the H O linesfrom 3C 403.If the emission were part of an accretion disk, at whichgalactocentric radius would it arise? Assuming that thetwo spectral components originate from the tangentiallyseen parts of a nuclear accretion disk seen approximatelyedge-on ( i ∼ ◦ ), we obtain (Fig. 1) a rotation velocityof V rot = V obs sin − i ∼
100 km s − . From Bettoni et al.(2003; their Table 3) we obtain M BH = 10 . M ⊙ as themass of the nuclear engine. Combining these two parame-ters and assuming the presence of Keplerian rotation (as inNGC 4258, see Miyoshi et al. 1995; Herrnstein et al. 1999,2005), this yields with R = 0 . (cid:20) M BH M ⊙ (cid:21) (cid:20) V obs sin − i km s − (cid:21) − (cid:20) D Mpc (cid:21) − masan angular distance of 245 mas, corresponding to a galacto-centric radius of R GC ∼
275 pc. This size conflicts with the archi et al.: The innermost region of 3C 403 9 result of our VLA observations (Sect. 4.1.2) that, indicat-ing that the continuum and the H O emission are arisingfrom the same 5.5 pc (5 mas) sized region, imply a max-imal galactocentric radius for the H O masers of 2.75 pc(2.5 mas).Is it possible to reconcile this value with the possibil-ity that the maser arises in an accretion disk like that inNGC 4258? A 2.75-pc radius disk rotating at 100 km s − would imply a black hole mass M BH = 10 . M ⊙ , a fac-tor of 100 lower than that derived by Bettoni et al. 2003.Alternatively, we might consider a disk inclined by ∼ ◦ orless. However, this would describe an almost face-on disk,a scenario not only different from that of NGC 4258 butalso seemingly contradicting the presence of jets orientedapproximately along the plane of the sky.To summarize, the masers in 3C 403 are peculiar inthree different ways. (1) They are detected at the core ofa radio galaxy. (2) The two main ultraluminous velocitycomponents are highly time variable and (3) the Keplerianmodel seems to fail . If the main velocity components donot represent the two tangentially viewed parts of the mas-ing disk but instead only one part and the systemic fea-tures, this would lead to V rot ∼
200 km s − and with M BH = 10 . M ⊙ (Bettoni et al. 2003) to a galactocentric dis-tance of 60 mas (70 pc). In this case the disagreement withobservations is slightly less severe but the linear scale isstill inconsistent with the results from our VLA A-arraymeasurements (Sect. 4.1.2). Thus, the accretion disk sce-nario requires relatively strict constraints to be applied toour case unless we are only viewing the systemic velocityfeatures of the putative maser disk (see below). Cloud stability considerations : A main difference be-tween Seyfert and FR II galaxies (equivalent to radio loudtype 2 quasars according to the standard unified scheme)is the mass of the nuclear engine. While in Seyfert galaxiesmasses range from ∼ to a few 10 M ⊙ (e.g., Greenhill etal. 1996, 1997; Herrnstein et al. 1999, Henkel et al. 2002),masses in radio galaxies and quasars reach 10 M ⊙ , with3C 403 hosting a nucleus that is not far below this lattervalue (Bettoni et al. 2003). Could clouds in an NGC 4258-like disk that is characterized by a particularly large rota-tional velocity and small galactocentric radius be stable inan environment dominated by such a large central mass?The masers in NGC 4258 might arise from well confineddiscrete clumps that are encompassing only a small fractionof the entire volume of the warped circumnuclear disk. For aroughly spherical clump at distance R GC from the centralengine and neglecting magnetic fields (see Modjaz et al.2005), a density of n (H ) > ∼ π G (cid:18) V rot R GC (cid:19) is required to reach stability against tidal disruption (Starket al. 1989). G is the gravitational constant. For NGC 4258,i.e. R GC ∼ V rot ∼ − (Herrnstein etal. 1999, 2005), the resulting density becomes n (H ) ∼ × cm − . This is close to thermalization (Kylafis et al.1991). For commonly accepted collisional excitation thiseither hints at excitation by a two temperature gas (e.g.,Kylafis et al. 1987) or at emission from turbulent, warm, dense molecular debris that does not form self-gravitatingclouds.Most circumnuclear maser disks are characterized bysmaller rotation velocities and higher galactocentric dis-tances, significantly reducing the required density for sta-bility against tidal disruption so that here the assump-tion of long lived masing clumps giving rise to collision-ally excited masers is less of a problem (e.g., n (H ) > ∼ cm − in NGC 1068; see Greenhill et al. 1996; Gallimoreet al. 2001). Assuming for a putative accretion disk in3C 403, where only the systemic spectral features are seen(Sect. 5.2.1), R GC = 0.2 pc (as in NGC 4258), the corre-sponding minimum density for cloud stability reaches, how-ever, 10 cm − , which is prohibitively large for collisionallyexcitated maser emission by a one temperature gas. Couldthen a lack of stable, well ordered clouds explain the rapidvariability of the maser features in 3C 403? At the maximalgalactocentric distance permitted by our VLA data, R GC ∼ > ∼ cm − . The jet-maser scenario : While in Seyfert and LINERgalaxies statistics favor accretion disk over jet-masers (Sect.5.1.1), the two main H O velocity components are (1) quitebroad and (2) offset by several 10 km s − from the nominal(NED) systemic velocity of the galaxy. These are the typicalspectral properties of jet masers (Peck et al. 2003; Henkel etal. 2005). The flares of velocity components I and II (Fig. 2)may arise from shocked regions at the interface between theenergetic jet material and the molecular gas of the cloud thejet is boring through. The time delay between componentsI and II (Fig. 1) and the apparent red- (component I) andblue-shift (component II) with respect to the systemic ve-locity may be caused by two clouds at opposite sides of thenucleus that have a slightly different galactocentric radius,possibly being part of a circumnuclear molecular ring. A binary supermassive black hole in 3C403?
Asidefrom the mass of the nuclear engine, there may be anotherfundamental difference between 3C 403 and NGC 4258 thatis worth mentioning. On a large spatial scale, 3C 403 showsnot the standard two-sided radio jet, but a pair of suchjets (see Sect. 1). The apparently younger one is orientedroughly NE-SW, the more extended, weaker one in the SE-NW direction. It is thus one of the rare examples of anX-shaped radio galaxy (e.g., Figs. 1 and 2 of Kraft et al.2005). This morphology is believed to be caused by thecoalescence of two supermassive black holes (e.g., Gopal-Krishna et al. 2003; Liu 2004) whose interaction modifiesthe inclination of the pre-existing nuclear disk of the dom-inant galaxy. Such a scenario may lead to a complexitythat is much higher than that encountered in the nuclearenvironment of NGC 4258, yielding a puzzling number ofpossible disk- and jet-maser configurations that have to beconstrained by future measurements. In view of the numberof free parameters in such an environment, it remains to beseen whether the anticorrelation of the main two maser fea-tures was a rare accidental event or whether it will providean important clue for a better understanding of the so farpoorly explored circumnuclear regions of X-shaped radiogalaxies.
6. Conclusions and outlook
We have investigated the luminous megamaser in the X-shaped FR II galaxy 3C 403 using single-dish monitoring ofthe line features and VLA A-array observations. The twomain maser features show extreme intensity variations dur-ing a two-year period. The position of the maser as derivedfrom the VLA maps is consistent with a location within afew pc of the radio nucleus of the galaxy.In addition, we have used the EVN to study the galaxy’snuclear radio continuum on parsec scales. A dominant cen-tral component and two extensions toward the south-westand north-east are identified. Because these parsec-scaleextensions show position angles that are compatible withthose of the large scale jets and because these large scalejets show similar flux densities, an interpretation in termsof a dominant nuclear core and a two-sided jet is plausible.No clear sign of CO emission or absorption has beendetected. If emission and absorption are not accidentallycanceling each other, this limits the extent of a low densitymolecular gas component to a size of order < ∼
200 pc.If the mass of the nuclear engine is as large as previouslyestimated and if the nuclear accretion disk is oriented (asexpected) almost edge-on, an NGC 4258-like accretion diskscenario is only viable, if all the detected H O features aresystemic, arising from the front or back side of the disk.There is, however, no observational evidence for the tan-gentially viewed approaching and receding parts of the disk.A secular velocity drift of the observed components is alsonot seen. In view of linewidths and deviations from the sys-temic velocity, an interpretation in terms of ”jet-masers”,arising from molecular clouds shocked by the nuclear jet(s),is the most plausible one.Since 3C 403 is an X-shaped radio galaxy, it may host abinary nuclear engine, greatly enlarging the number of po-tential scenarios that should be observationally constrainedby future measurements. These imply single-dish masermonitoring over long time spans (1) to study the lifetime ofthe various components, (2) to check whether the anticorre-lation of the two main features is a typical or an accidentalevent and (3) to find, possibly, some evidence for a system-atic velocity drift. The latter would be a clear sign for thepresence of a nuclear disk even if the red- and blue-shiftedparts of this disk remain undetectable (see Wilson et al.1995 and Braatz et al. 2003 for such a previously identi-fied case). Another very important measurement would bethe simultaneous observation of the maser features and theradio continuum at (sub-)milliarcsecond resolution. Thiscould directly show whether the maser features are asso-ciated with the nuclear jets or whether we should rejectthis scenario. In any case it would shed light onto the mor-phology of the nuclear environment of a prototypical X-shaped galaxy. Finally, systematic monitoring of the radiocontinuum, H O and X-ray spectra may provide through re-verberation mapping important correlations that would beessential to assess the physical state and three-dimensionalmorphology of the circumnuclear medium surrounding a10 . M ⊙ AGN.
Acknowledgements.
AT would like to thank Karl Heinz Wissmannand his collaborators for pleasant discussions during the makingof this work. AB is supported by the Priority Programme 1177of the Deutsche Forschungsgemeinschaft. We wish to thank J.Moran for critical comments. This research has made use of theNASA/IPAC Extragalactic Database (NED) which is operated by theJet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.We have also made use of NASA’s Astrophysics Data System.
References
Barthel, P. D. 1989, ApJ, 336, 606Barvainis, R., & Antonucci, R., 1994, AJ, 107, 1291Barvainis, R., & Antonucci, R., 2005, ApJ, 628L, 89Baum, S. A., Heckman, T., & van Breugel, W. 1990, ApJS, 74, 389Bellamy, M. J., & Tadhunter, C. N., 2004, MNRAS, 353, 105Bettoni, D., Falomo, R., Fasano, G., & Govoni, F., 2003, A&A, 399,869Black, A. R. S., Baum, S. A., Leahy, J. P., et al. 1992, MNRAS, 256,186 (BBL)Braatz, J. A., Wilson, A. S., & Henkel, C. 1996, ApJS, 106, 51Braatz, J. A., Wilson, A. S. & Henkel, C. 1997, ApJS, 110, 321Braatz, J. A., Wilson, A. S., Henkel, C., Gough, R., & Sinclair, M.2003, ApJS, 146, 249Braatz, J. A., Henkel C., Greenhill L. J., Moran J. M., & Wilson A.S., 2004, ApJ, 617, 29Brunthaler, A., Reid, M., Falcke, H., Greenhill, L. J., & Henkel, C.2005, Science, 307, 1440Brunthaler, A., Reid, M., Falcke, H., Henkel, C. & Menten, K. M.2007, A&A, 462, 101Carilli, C. L., & Barthel P. D. 1996, A&A Rev., 7, 1Chiaberge, M., Capetti, A., & Celotti, A. 1999, A&A, 349, 77Conway, J. E., & Blanco, P. R. 1995, ApJ, 449, L131Dennett-Thorpe, J., Scheuer, P. A. G., Laing, R. A., et al. 2002,MNRAS, 330, 609 (DSL)Evans, A. S., Mazzarella, J. M., Surace, J. A., et al. 2005, ApJS, 159,197Fanti, C., Pozzi, F., Dallacasa, D., et al. 2001, A&A, 369, 380Fuente, A., Black, J. H., Mart´ın-Pintado, J., et al. 2000, ApJ, 545,L113Gallimore, J. F., Baum, S. A., O’Dea, C. P., Pedlar, A., & Brinks, E.,1999, ApJ, 524, 684Gallimore, J. F., Henkel, C., Baum, S. A., et al. 2001, ApJ, 556, 694Gopal-Krishna, Biermann, P. L., Wiita, P. J. 2003, ApJ, 594, 103Greenhill, L. J., Jiang, D. R., Moran, J. M., et al., 1995, ApJ, 440,619Greenhill, L. J., Gwinn, C. R., Antonucci, R., & Barvainis, R. 1996,ApJ, 472, L21Greenhill, L. J., Moran, J. M., & Herrnstein, J. R. 1997, ApJ, 481,L23Greenhill, L. J., Booth, R. S., Ellingsen, S. P., et al. 2003, ApJ, 590,162Henkel, C., Wang, Y. P, Falcke, H., Wilson, A. S., & Braatz, J. A.1998, A&A, 335, 463Henkel, C., Braatz, J. A., Greenhill, L. J., & Wilson, A. S. 2002, A&A,394, L23Henkel, C., Tarchi, A., Menten, K. M., & Peck, A. B. 2004, A&A,414, 117Henkel, C., Peck, A. B., Tarchi, A., et al. 2005, A&A, 436, 75Herrnstein, J. R., Moran, J. M., Greenhill, L. J., et al. 1999, Nature,400, 539Herrnstein, J. R., Moran, J. M., Greenhill, L. J., & Trotter, A. S.2005, ApJ, 629, 719Koekemoer, A. M., Henkel, C., Greenhill, L.J., et al. 1995, Nature,378, 697Kondratko, P. T., Greenhill, L. J., Moran, J. M., et al. 2006, ApJ,638, 100Kraft R. P., Hardcastle M. J., Worrall D. M., & Murray S. S. 2005,ApJ, 622, 149 (K05)Kylafis, N. D., & Norman, C. A. 1987, ApJ, 323, 346Kylafis, N. D., & Norman, C. A. 1991, ApJ, 373, 525Liu, F. K. 2004, MNRAS, 347, 1357Martel, A. R., Baum, S. A., Sparks, W. B., et al., 1999, ApJS, 122,81Mauersberger, R., Wilson, T. L., & Henkel, C. 1988, A&A, 201, 123Merritt, D., & Eckers, R. D., 2002, Science, 297, 1310Mirabel, I. F., & Wilson, A. S. 1984, ApJ, 277, 92Miyoshi, M., Moran, J., Herrnstein, J., et al. 1995, Nature, 373, 127Modjaz, M., Moran, J. M., Kondratko, P. T., & Greenhill, L. J. 2005,ApJ, 626, 104Morganti, R., Oosterloo, T. A., Tadhunter, C. N., et al. 2001, MNRAS,323, 331O’Dea, C. P., & Baum, S. A. 1997, AJ, 113, 148 archi et al.: The innermost region of 3C 403 11
Peck, A. B., Henkel, C., Ulvestad, J. S., et al. 2003, ApJ, 590, 149Perlman, E. S., Sparks, W. B., Radomsky, J., et al. 2001, ApJ, 561,L51Schoenmakers, A. P., de Bruyn, A. G., R¨ottgering, H. J. A., van derLaan, H., & Kaiser, C. R. 2000, MNRAS, 315, 371Shepherd M. C., Pearson T. J., & Taylor G., 1994, BAAS, 26, 987Spergel, D. N., Verde, L., Peiris, H. V., Komatsu, E., Nolta, M. R.,et al. 2003, ApJS, 148, 175Spinrad, H., Djorgovski, S., Marr, J., & Aguilar, L. 1985, PASP, 97,932Stark, A. A., Bally, J., Wilson, R. W., & Pound, M. W. 1989,IAU Symp., 136, The Center of the Galaxy, Kluwer AcademicPublishers, Dordrecht, p129Tadhunter, C. N., Morganti, R., di Serego-Alighieri, S., Fosbury, R.A. E., & Danziger, I. J. 1993, MNRAS, 263, 999Tarchi, A., Henkel C., Chiaberge, M., & Menten, K. M., 2003, A&A,407, L33 (THC)Tarchi, A., Henkel, C., Chiaberge, M., et al. 2005, ApSS, 295, 117Urry, C. M., & Padovani, P. 1995, PASP, 107, 803Verdoes Kleijn, G. A., Baum, S. A., de Zeeuw, P. T., & O’Dea, C. P.2002, AJ, 123, 1334Whysong, D., & Antonucci, R. 2004, ApJ, 602, 116Wiklind, T., & Henkel, C. 2001, A&A, 375, 797Wilson, A. S., Braatz, J. A., & Henkel, C. 1995, ApJS, 455, L127Zhang, J.S., Henkel, C., Kadler, M., et al. 2006, A&A, 450, 933
APPENDIX: 4C +02.49
In order to find suitable phase calibrators for VLBI ob-servations, we observed nine bright and compact sourcesfrom the NRAO VLA Sky Survey (NVSS) with an angu-lar separation of < ◦ with respect to 3C 403. The 22 GHz( λ ∼ ∼ ± ± ± >
10 kpc for a turnover fre-quency <
100 MHz. This cannot be tested, since, so far, noredshift has been derived for this source.The two compact components of 4C+02.49 were clearlydetected with the EVN at λ ∼ ± ± ∼
220 mJy is re-solved out. This is likely the main cause for the seeminglyinverted spectra when comparing the EVN 6 cm with theVLA 1.3 cm flux densities.