Radio jets in NGC 4151: where eMERLIN meets HST
D. R. A. Williams, I. M. McHardy, R. D. Baldi, R. J. Beswick, M. K. Argo, B. T. Dullo, J. H. Knapen, E. Brinks, D. M. Fenech, C. Mundell, T. W. B. Muxlow, F. Panessa, H. Rampadarath, J. Westcott
MMNRAS , 1–13 (2017) Preprint 8 October 2018 Compiled using MNRAS L A TEX style file v3.0
Radio jets in NGC 4151: where eMERLIN meets HST
D.R.A. Williams, (cid:63) I.M. McHardy, R.D. Baldi, R.J. Beswick, M.K. Argo, , B.T. Dullo, , , J.H. Knapen, , E. Brinks, D.M. Fenech, C. Mundell, T.W.B. Muxlow, F. Panessa, H. Rampadarath, J. Westcott School of Physics and Astronomy, University of Southampton, Southampton, SO17 1BJ, UK Jodrell Bank Centre for Astrophysics, School of Physics and Astronomy, The University of Manchester, Manchester, M13 9PL, UK University of Central Lancashire, Jeremiah Horrocks Institute Preston, UK PR1 2HE Departamento de Astrof´ısica y Ciencias de la Atm´osfera, Universidad Complutense de Madrid, E-28040 Madrid, Spain Instituto de Astrof´ısica de Canarias, V´ıa L´actea S/N, E-38205 La Laguna, Spain Departamento de Astrof´ısica, Universidad de La Laguna, E-38206 La Laguna, Spain Centre for Astrophysics Research, University of Hertfordshire, College Lane, Hatfield, AL10 9AB, UK Department of Physics & Astronomy, University College London, Gower Street, London, WC1E 6BT, UK University of Bath, Claverton Down, Bath, BA2 7AY, UK INAF - IAPS Rome, Via Fosso del Cavaliere 100, I-00133 Roma, Italy
Accepted 2017 August 21. Received 2017 August 21; in original form 2017 July 24
ABSTRACT
We present high-sensitivity eMERLIN radio images of the Seyfert galaxyNGC 4151 at 1.5 GHz. We compare the new eMERLIN images to those from archivalMERLIN observations in 1993 to determine the change in jet morphology in the 22years between observations. We report an increase by almost a factor of 2 in the peakflux density of the central core component, C4, thought to host the black hole, buta probable decrease in some other components, possibly due to adiabatic expansion.The core flux increase indicates an AGN which is currently active and feeding the jet.We detect no significant motion in 22 years between C4 and the component C3, whichis unresolved in the eMERLIN image. We present a spectral index image made withinthe 512 MHz band of the 1.5 GHz observations. The spectrum of the core, C4, is flatterthan that of other components further out in the jet. We use HST emission line images(H α , [O III] and [O II]) to study the connection between the jet and the emission lineregion. Based on the changing emission line ratios away from the core and compari-son with the eMERLIN radio jet, we conclude that photoionisation from the centralAGN is responsible for the observed emission line properties further than 4 (cid:48)(cid:48) (360 pc)from the core, C4. Within this region, several evidences (radio-line co-spatiality, low[O III]/H α and estimated fast shocks) suggest additional ionisation from the jet. Key words: galaxies: active – galaxies: nuclei – galaxies: Seyfert – galaxies: individ-ual: NGC 4151 – galaxies: jets – quasars: emission lines
At the centre of every galaxy is thought to lie a super-massive black hole (SMBH) (Magorrian et al. 1998; Geb-hardt et al. 2000; Ferrarese & Merritt 2000). Broad-bandemission from SMBHs is observed from the X-ray throughto the radio regime. When they accrete matter they turn intoactive galactic nuclei (AGN) (Ho 2008). Most of the AGN inthe local Universe are radio-quiet, defined by Terashima & (cid:63)
E-mail: [email protected]
Wilson (2003) as those where the logarithm of the ratio ofthe radio (5 GHz) to X-ray (2 −
10 keV) luminosity, denotedlog(R X ), is ≥ − © a r X i v : . [ a s t r o - ph . GA ] A ug D.R.A. Williams et al. the jet in the well-known Seyfert 1.5 NGC 4151. It is oneof the brightest AGN in the sky in X-rays (Gursky et al.1971; Boksenberg et al. 1995; Ogle et al. 2000; Wang et al.2010, 2011b) and the radio-brightest of the radio-quiet AGN(Zdziarski et al. 2000). As such it is a great probe of theintermediate regime of radio-loudness, to explore the mech-anisms of jet propagation through the interstellar medium(ISM). NGC 4151 is a nearly face-on ( i ≈ ◦ ) barred spiralgalaxy. It has one of the most precise distance measurementsof an AGN to date, due to dust-parallax measurements, of19 Mpc (H¨onig et al. 2014). This corresponds to an angularscale of ∼
91 pc per arc second.NGC 4151 has been extensively studied in the radio forseveral decades (Wilson & Ulvestad 1982; Johnston et al.1982; Booler et al. 1982; Harrison et al. 1986; Carral et al.1990; Pedlar et al. 1993; Mundell et al. 1995; Ulvestadet al. 1998; Mundell et al. 2003; Ulvestad et al. 2005).The radio structure is characterised by a double-sided jetat PA ∼ ◦ extending from a nucleus with VLBI centre at α J =12 h m m and δ J =+39 d m s (Ul-vestad et al. 2005). Archival VLA/MERLIN observations(Carral et al. 1990; Pedlar et al. 1993; Mundell et al. 1995)detect six radio components along this structure namedC1 to C6 with the naming convention going from westto east. Further resolved components were discovered withVLBA/VLBI images Ulvestad et al. (2005), who renamedthe components A to I (Fig. 1). Throughout this paper weshall use the original C1-C6 nomenclature, unless specifiedotherwise. The core radio component (named C4) is co-incident with the optical nucleus in Mundell et al. (1995)(hereafter referred to as M95) leading to its identificationas the AGN. Here we present deep 1.51 GHz observationswith the upgraded eMERLIN radio interferometer allowing,by comparison with M95, study of changes in the jet mor-phology over a 22 year period.With the exception of the decommissioning of the War-dle (Mk III) antenna, the configuration of eMERLIN is iden-tical to that of MERLIN, providing an angular resolution of150mas at 1.5 GHz. The bandwidth of eMERLIN is widerthat that of MERLIN, leading to improved uv -coverage.The eMERLIN observations of NGC 4151 were made as partof the L egacy e - M ERLIN M ulti-band I maging of N earby G alaxies S urvey - LeMMINGs (Beswick et al. 2014). LeM-MINGS is the second largest of the eMERLIN legacy surveysand consists of observations of all 280 galaxies above δ ≥ ◦ from the Palomar sample of nearby galaxies (Filippenko &Sargent 1985; Ho et al. 1995, 1997a,b,c,d,e, 2003, 2009).A number of observers (Perez et al. 1989; Evans et al.1993; Robinson et al. 1994; Boksenberg et al. 1995; Wingeet al. 1997; Hutchings et al. 1998; Winge et al. 1999; Hutch-ings et al. 1999; Kaiser et al. 2000; Kraemer et al. 2008) haveshown optical emission line images of the nuclear regions ofNGC 4151. Whilst it is generally agreed that photoionisationfrom the AGN is an important contributor to the ionisation,the importance of the radio jet is not so clear. Here, by com-bining the eMERLIN image with HST H α , [O II], and [O III]images, we explore the contribution of the jet in more detail.In section 2 we present the eMERLIN observationsand data reduction. In section 3 we discuss morphologicalchanges between the present image and the previous MER-LIN image. In section 4 we discuss the relationship between Figure 1.
Naturally weighted archival MERLIN image of thecentral, 4 × (cid:48)(cid:48) ( ∼ ×
180 pc) radio structures of NGC 4151, re-reduced with the MERLIN pipeline. The FWHM of the restoringbeam was set to 0.15 (cid:48)(cid:48) × (cid:48)(cid:48) (14 ×
14 pc) and the entire uv -rangewith all 8 antennas in the MERLIN array was used to producethis image in AIPS . For consistency, the contours are the same asFig. 2 in (Mundell et al. 1995): 1.5, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15,20, 25, 30 mJy/beam. The naming conventions are shown abovewith Carral et al. (1990) nomenclature in blue, and the Ulvestadet al. (2005) nomenclature in red.
Figure 2.
Full observed uv -plane of NGC 4151 at 1.5 GHz, us-ing the LeMMINGs (eMERLIN) deep data with all 7 antennasincluded. the radio jet and the optical line emission region and wesummarise our conclusions in section 5. Observations of NGC 4151 were performed at L-band(weighted central frequency of 1.51 GHz) with the eMER-LIN array as part of the LeMMINGs deep sample. Thissub-sample consists of a small number of galaxies for whichparticularly deep pbservations have been taken includ-ing NGC 4151 (this paper), IC 10 (Westcott et al. 2017),NGC 5322 (Dullo et al in prep.) and NGC 6217 (Williamset al in prep). All 7 antennas in the array participated in
MNRAS , 1–13 (2017) adio jets in NGC 4151 the observation on 29th April 2015, including the Lovelltelescope. NGC 4151 was observed on-source for 3.81 hourswith data reduction and imaging following the steps out-lined in the e-MERLIN cookbook and pipeline (Belles et al.2015; Argo 2015) with AIPS (Wells 1985). The full uv -planeof the eMERLIN observations is shown in Fig. 2. The obser-vational set-up used a total bandwidth of 512 MHz, centredon 1.51 GHz. The 512 MHz band was split into 8 intermedi-ate frequencies (IFs) of width 64 MHz and consisting of 128channels in each IF. The calibrator NVSSJ120922+411941(J1209+4119) was used for phase referencing and OQ208and 3C286 were used as the band pass and flux calibratorsrespectively. The target and phase calibrator alternated dur-ing the observing run, with blocks of approximately 2.5 minson the phase calibrator and 7 mins on the target, with theflux and band pass calibrators observed at the end of theobserving run.To calibrate the data, we followed the procedure out-lined in the e-MERLIN cookbook (Belles et al. 2015), a sum-mary of which we include below. Correlation and averagingof the data was performed before the SERPent flagging code,was used to remove the worst instances of radio frequencyinterference (RFI) from the data. The raw data was then in-spected with AIPS tasks
SPFLG and
IBLED to remove any low-level RFI not picked up by the automatic flagger. In additionto the RFI flagging, the first two IFs of the LL polarisationwere flagged on all Lovell baselines due to the inclusion ofa test filter on the antenna. The channels showing no co-herent phase at the ends of each IF were also flagged. It isestimated that approximately 15 per cent of the on-sourcedata were flagged during this process and further calibrationrounds before the final images were made.To begin the calibration procedures, we fitted the off-sets in delays using the
AIPS task
FRING before calibratingfor phase and gain. Band pass solutions were also calculatedwith
BPASS and applied followed by imaging of the phase cal-ibrator, which was self-calibrated until solutions converged.The complex antenna solutions from self-calibration of thephase calibrator were then applied to the target field.
These data were imaged with
IMAGR and phase self-calibration was applied to further refine the data beforethe visibilities were re-weighted using
REWAY to account forvariable sensitivity as a function of antenna, as eMERLINcomprises of an inhomogeneous set of antenna types, fre-quency and time to maximise the resultant sensitivity of thedata. Further self calibration improved the signal-to-noiseand the final image was created with the noise in the nat-urally weighted image of 35 µ Jy as shown in Fig. 3. Due tothe complex nature of this source and its brightness, greatcare was taken to only include real what are deemed to begenuine emission features in the self calibration process. Anamplitude and phase self calibration did not produce stablephases or amplitudes, so only the final phase self calibrateddata are shown. This data set was then used to create allfurther images of NGC 4151. The
SERPent flagging code (Peck & Fenech 2013) is written in
ParselTongue (Kettenis et al. 2006)
Archival data originally published in M95 were re-reduced tocompute accurate positions and flux density measurementsto compare to the new eMERLIN data set. NGC 4151 wasobserved in November 1993 at 1.42 GHz. The MERLIN datawas calibrated with the MERLIN pipeline in AIPS and re-imaged to account for RFI, alter the beam size and re-weightthe data. As the data were originally taken in spectral lineimaging mode to study the neutral hydrogen absorptionin NGC 4151 (see M95), to create a continuum image weflagged all channels that were contaminated by this line toensure the flux measurements on the core component C4were not affected by this absorption feature. The amountof on-source time was 9.8 hours, achieving an rms noise of0.25 mJy/beam. We thus see the large improvement in S/Nachieved by eMERLIN compared to the previous MERLINarray. All stations in the MERLIN array took part in thisobservation including the Lovell telescope. All images pro-duced were made at the same resolution as those publishedby M95, in order to be able to directly compare all data.The re-reduced natural image of the data is shown in Fig. 1together with the contouring scheme from M95. Note thatthe extended low surface brightness structures seen to thesouth of C2 in Fig.2 of M95 is not present in our re-reducedimage, probably due to improved imaging fidelity of thesenew eMERLIN data due to the large bandwidth (512 MHzcompared to 8 MHz) resulting in more complete sampling ofspatial scales.
When comparing the two epochs of data, some caution needsto be taken since archival MERLIN data does not completelymatch the new eMERLIN data in terms of uv -range, num-ber of antennas, bandwidth, central frequency and observingtime.Therefore, to make an unbiased comparison of thesource structure at the two data epochs, we made two addi-tional images with similar conditions for each dataset. The uv -range was limited to between 100 and 1000 k λ where thetwo datasets overlapped. We exclude the now defunct MarkIII (Wardle) antenna from the MERLIN data. All of theimages were made with the same FWHM restoring beamsize of 0.15 × (cid:48)(cid:48) , corresponding to 14 ×
14 pc. The finalimages are shown in Fig. 4 and Fig. 5 based on MERLINand eMERLIN respectively. Subsequently, this enables us tocompare emission on the same spatial scales.The eMERLIN data were then loaded into
CASA to pro-duce an in-band spectral index image with clean and nterms set to 2. The spectral index image is shown in Fig. 6.
Perez et al. (1989) presented [O III] and H α ground-basedimaging of NGC 4151 and show that the ratio of [O III]/H α defines a large ( ∼ (cid:48)(cid:48) , 910 pc) cone-like structure to the000
Perez et al. (1989) presented [O III] and H α ground-basedimaging of NGC 4151 and show that the ratio of [O III]/H α defines a large ( ∼ (cid:48)(cid:48) , 910 pc) cone-like structure to the000 , 1–13 (2017) D.R.A. Williams et al.
Figure 3.
New full-resolution eMERLIN image of the central 4 × (cid:48)(cid:48) ( ∼ ×
180 pc) region of NGC 4151 using all 7 eMERLIN antennasand a natural weighting. As in Fig. 1, the entire uv -range was used with a 0.15 (cid:48)(cid:48) × (cid:48)(cid:48) FWHM restoring beam. Contours set are at-0.25, 0.75, 1, 1.5, 2, 3, 4, 5, 9, 16, 25, 36, 49, 64 mJy/beam. south-west of the nucleus of PA ∼ ◦ , misaligned with re-spect to the radio structure. Evans et al. (1993) and Bok-senberg et al. (1995) have presented HST
Planetary Camera(PC) images in [O III] and H α . HST Wide Field PlanetaryCamera 2 (WFPC2) images in [O II], [O III] and H α havebeen presented by (Hutchings et al. 1998, 1999; Kaiser et al.2000; Kraemer et al. 2008).High-resolution HST
Wide-Field Planetary Camera 2(WFPC2) images of NGC 4151 were taken, from the GO-5124 program (PI: H. Ford), in the F336W, F375N ([O II]),F502N ([O III]), F547M, F658N (H α ) and F791W filterswere retrieved from the public Hubble Legacy Archive(HLA ). The data were observed on the 22nd January 1995.Here we re-reduce these data, using the state-of-the-art tech-niques described in Dullo et al. (2016), providing [O II],[O III] and H α emission line images and produce an HST [O III]/H α image which has not been previously presented.These images are considered, together with the eMERLINradio image, in Section 4.In order the create the emission line images we used theWFPC2 PC1 F336W, F375N, F502N, F547M and F658Nimages and followed the procedures outlined in Knapenet al. (2004), S´anchez-Gallego et al. (2012) and Dullo et al.(2016) and compared the narrow-band F375N ([O II]) im-age with the broad-band F336W image to create the [O II]continuum-subtracted emission line image (Fig. 7). Simi-larly, we created the [O III] and H α emission line imagesby comparing the F502N ([O III]) and F658N (H α ) imageswith the F547M image of the galaxy. Figs. 7 and 8 show ournew eMERLIN L-band image overlaid on these three ([O II],[O III] and H α ) line images and on the [O III]/H α emissionline ratio, respectively. http://hla.stsci.edu The V − H dust map of NGC 4151 by Martini et al.(2003, their Fig. 1) shows that the galaxy has spiral dustarms at R > ∼ pc. However, this dust map and the visualinspections of the HST images (see Fig. 7) show that theemission-line regions in the galaxy are only weakly obscuredby dust. Therefore, we did not attempt to correct for thespiral dust arms, although dust extinction may somewhataffect our [O II] line flux measurement since dust absorptionis relatively higher at shorter wavelengths. However, whenwe extract emission line fluxes from the images (Section 4),we take into account the reddening using the extinction lawfrom Calzetti et al. (2000).We note that the coordinates of the radio core (com-ponent C4) in the eMERLIN image (Fig. 3) differ by 0 . (cid:48)(cid:48) HST images.The radio positions are linked to the positions of VLBI phasecalibrator sources. Systematic positional uncertainties are < ∼ HST positions arelinked to the positions of stars in the Guide Star Catalogue(Jenkner et al. 1990) which are typically accurate to 0 . (cid:48)(cid:48) HST positions. We therefore align the radioand optical images so that the maximum surface brightnessat the core of each optical image is coincident with the ra-dio core, C4. We only move the images in Right Ascensionand Declination and do not rotate. To align the radio andoptical data we used astropy to edit the FITS file header forthe central Right Ascension and Declination values. http://docs.astropy.org/ MNRAS , 1–13 (2017) adio jets in NGC 4151 The full-resolution naturally-weighted 1.5 GHz eMERLINimage of NGC 4151 (Fig. 3) shows the six previously knowncomponents. C1-C6, extending over 4 (cid:48)(cid:48) (360 pc), similarto the twin-jet morphology observed in M95. The mor-phology shown in this image is similar to the naturallyweighted MERLIN image (Fig.1) but the core, C4, is def-initely brighter, by a factor of 1.5, in the eMERLIN image.The jets, particularly the western jet, also appear narrower.Below we will consider the morphological changes inthe radio components, from C1 to C6, between the MER-LIN and eMERLIN images, as well as changes in positionsand fluxes. We fitted 2D Gaussian components to each ofthe components C1-6 using the
AIPS task
JMFIT . The peakflux density, integrated flux, RA and Dec, and size of eachcomponent were extracted and are shown in Tables 1 and 2.The positions of the components stated in Table 2 refer tothe self-calibrated data. These positions are consistent withthe values from the dirty image without any self-calibrationapplied, to within 0.001 (cid:48)(cid:48) , much smaller than the FWHM ofthe restoring beam. Note that in those tables we list onlythe error given by
JMFIT from fitting the components. Thetrue error, including the rms noise level on the image, possi-ble contamination by sidelobes from other components anduncertainties in the flux density calibrator, is hard to definebut is likely to be at least 3 times larger. In the followingsections we will focus on each component of the jet in turnto study their morphology, possible movement and flux den-sity change with respect to the two epochs. In terms of theradio morphology, the full resolution images are compared,while flux density variations come from the comparison ofthe uv -range restricted images (Fig. 4 and Fig. 5).We caution, however, that small differences in uv -coverage can have a noticeable effect in imaging, particularlyof low brightness structures. Here, although we are able torestrict the 1D uv -range to be the same for the MERLIN andeMERLIN images (Figs. 4 and 5), we are not able to takethe final step of restricting the 2D uv -coverage of the MER-LIN image to be the same as for eMERLIN as the reductionin sensitivity is then too great to allow useful imaging of lowbrightness structures. Comparing Figs. 4 and 5, the overall peak flux density ofC1 and C2 appears to have decreased. However we cautionthat the different uv -coverage of the MERLIN and eMER-LIN datasets may affect the detection of large scale struc-ture, giving rise to negative bowls underlying the more com-pact emission and hence reducing peak flux densities, in theeMERLIN image. In this respect it is interesting to note thatwith the lower contour levels for the full-resolution eMER-LIN image (Fig. 3), the extended structures of C1 and C2resemble those of the MERLIN image (Fig. 1). C3 is unresolved, even at the highest angular resolutionachievable with eMERLIN. Similar to components C1 andC2, the flux density in C3 decreased by a factor of ∼ uv -coverage but for an unresolved com-ponent the differences in uv -coverage should not be so im-portant. Higher-resolution 1.4 GHz VLBA observations (Ul-vestad et al. 1998) show a faint (2.6mJy total flux density)component at the location of C3. This indicates that theemission associated with component C3 by MERLIN andeMERLIN must be related to an extended and diffuse re-gion, undetectable with the VLBA.The core, where the jet base is located, corresponds tocomponent C4. It is slightly elongated on the western sidetowards C3, where the 4 mJy/beam contour level connectsit to C3 as it did in the previous M95 image. In addition,other weaker extensions are associated with C4 towards thenorth, south and east. We cannot conclude anything abouttheir nature because of their small size and weakness.C4 is the brightest component, with a peak flux densityof ∼
67 mJy/beam, higher than in the MERLIN observationsby nearly a factor of ∼ . × ergs s − . VLBI observation at 18 cm (Ulves-tad et al. 1998; Mundell et al. 2003; Ulvestad et al. 2005)with an angular resolution of a few mas show several com-ponents along the jet axis located within the core C4, whichare not resolved by eMERLIN. The flux density of one ofthese VLBI components (D3b) increased by ∼
30 per cent in4 years between the observations of Ulvestad et al. (1998)and Ulvestad et al. (2005). The larger flux density changesseen between the MERLIN and eMERLIN observations arequite consistent with these VLBI changes and confirm thatC4 contains a currently active AGN core, injecting relativis-tic particles into the inner jet.
Moving to the eastern side of the jet, a 0.4 (cid:48)(cid:48) (36 pc) signifi-cant elongation bridges C4 and C5 (Fig. 3). C5 has a clearextension to the west, similar to that in the 1993 observa-tions. Furthermore, this extension overlaps with a VLBAcomponent G (Mundell et al. 2003). The peak intensity ofcomponent C5 remains mostly unchanged. C6 is the lastcomponent of the eastern jet and is consistent in flux andmorphology with previous MERLIN images. This elementhas a PA 125 ◦ , clearly different from the other components,suggesting a possible bending of the jet spine. A furtherweak component appears east of C6 which is not present inthe previous MERLIN data. This component requires fur-ther investigation with eMERLIN to confirm its presenceand morphology. C4 has definitely increased in flux density (by factor nearly2) in the 22 year period between the MERLIN and eMER-LIN observations. Components C5 and C6 in the easternjet have not changed noticeably. This lack of change leadsconfidence to the finding of a decrease in flux density ofcomponents C3, C2 and C1 in the western jet is real.
MNRAS000
MNRAS000 , 1–13 (2017)
D.R.A. Williams et al.
Table 1.
Flux densities obtained from Figs. 4 and 5, where the resolution of the data is matched, of each component in NGC 4151. Thespectral index was obtained from the spectral index image in Fig. 6, the size of the components as found from the fitting process insection 3 and the minimum energy and magnetic field obtained from this process.Comp. Peak FluxDensityMERLIN(mJy beam − ) IntegratedFlux DensityMERLIN 1993(mJy) Peak FluxDensityeMERLIN(mJy beam − ) IntegratedFlux DensityeMERLIN(mJy) SpectralIndexeMERLIN Component Size ineMERLIN image(mas ) B min (mG) log(U min )(J)C1 5.92 ± ± ± ± × ± ± ± ± × ± ± ± ± × ± ± ± ± × ± ± ± ± × ± ± ± ± × Figure 4.
Naturally weighted archival MERLIN image of the central 3.6 × (cid:48)(cid:48) ( ∼ ×
110 pc) region of NGC 4151, re-reduced with theMERLIN pipeline here. The beam was set to 0.15 (cid:48)(cid:48) × (cid:48)(cid:48) and a restricted uv -range of 100 − λ was used to produce this imagein AIPS to overlap with the same uv -range in the eMERLIN image. All antennas except the Mk III (Wardle) in the MERLIN array wereused so as to include only those antennas in the current eMERLIN array. The contours are -0.75, 1, 2, 3, 4, 5, 6, 7, 8, 9, 16, 25, 36, 49,64 mJy/beam. The naming convention from Carral et al. (1990) are overlaid in black. Figure 5.
Same as Fig. 4 but using the new eMERLIN data and reduced with the eMERLIN cookbook. This included all 7 eMERLINantennas. As in Fig. 4, the uv -range was limited to 100 - 1000 k λ . The intrinsic velocity, v , of a component in a jet over a giventime period is related to the angle that the line of sight sub-tends with the angle of the jet, θ , and its measured apparent transverse velocity v for relativistic jets (Kellermann et al.1989). v = v [ sin ( θ ) + v c cos ( θ )] (1) MNRAS , 1–13 (2017) adio jets in NGC 4151 Table 2.
Astrometric measurements of radio components compared to the position of the core C4 obtained from the self-calibrated data.As the data are self-calibrated, we caution that the absolute values of RA and Dec are not absolute. These positions are compared toobservations first published in M95, but have been re-reduced here (see section 2.1). All positions given are in J2000 co-ordinates, withthe difference calculated from the core C4 in that given image. The difference of these two values is shown in column 8 which is used asa measure of the relative shift of each component. The errors in the position from the fitting process are of the order half the beam size,i.e. ∼ (cid:48)(cid:48) (7 pc).Comp. RA eMERLIN Dec eMERLIN Difference RA MERLIN Dec MERLIN Difference Relative ShiftC1 12 10 32.41726 +39 24 20.6601 1.928 (cid:48)(cid:48)
12 10 32.42493 +39 24 20.6475 1.877 (cid:48)(cid:48) +0.051 (cid:48)(cid:48)
C2 12 10 32.49967 +39 24 20.8362 0.959 (cid:48)(cid:48)
12 10 32.50832 +39 24 20.8694 0.886 (cid:48)(cid:48) +0.073 (cid:48)(cid:48)
C3 12 10 32.54025 +39 24 21.0026 0.466 (cid:48)(cid:48)
12 10 32.54319 +39 24 20.9964 0.464 (cid:48)(cid:48) +0.002 (cid:48)(cid:48)
C4 12 10 32.57989 +39 24 21.0646 - 12 10 32.58279 +39 24 21.0597 - -C5 12 10 32.65199 +39 24 21.3570 0.884 (cid:48)(cid:48)
12 10 32.65756 +39 24 21.3505 0.913 (cid:48)(cid:48) -0.029 (cid:48)(cid:48)
C6 12 10 32.68911 +39 24 21.3844 1.306 (cid:48)(cid:48)
12 10 32.68911 +39 24 21.3885 1.275 (cid:48)(cid:48) +0.031 (cid:48)(cid:48)
Pedlar et al. (1993), Robinson et al. (1994) and Vila-Vilaro et al. (1995) estimate that the pointing angle of theradio jet with respect to the line of sight is ∼ ◦ based ona combination of geometric arguments, velocity images ofthe narrow line region (Winge et al. 1997; Hutchings et al.1998; Winge et al. 1999) and the galactic disc inclination.We take this pointing angle into account when calculatingthe jet speed.To accurately estimate the jet speed, components withreliable positions are needed. Hence only the unresolvedcomponents with no significant extensions can provide a pre-cise astrometric position by fitting them with a 2D-Gaussianprovided by the JMFIT task. This is the case for C3 and C4.As the data is self calibrated, we lose absolute astromet-ric positions of the components between the two epochs ofdata. However we do not lose relative positions and so wecan still measure possible changes in separation between C3and C4. We measure a change in separation of 2 mas butgiven a beam size of 150mas, we do not claim any detectablemovement. Taking the 2mas figure at face value would give avelocity of 0.04c, completely consistent with the upper limitsmeasured from VLBI observations by Ulvestad et al. (2005)but over a period of 4 years.
We use the definition of radio spectral index, α , where theflux density, S at frequency ν is given by S ν ∼ ν α . A flatspectrum ( α (cid:38) − α (cid:46) − CASA task clean with nterms = 2 to interpolate throughall the intermediate frequencies (Rau & Cornwell 2011). Weconsider only the emission above 2mJy/beam to remove un-real fluctuations due to low S/N. The spectral index imageis presented in Fig. 6. The calculated α values range between − ∼ − α obtained with wider-bandspectra by Pedlar et al. (1993) (5 − −
15 GHz), both with the MERLIN array.The component C3 has a steeper index ∼ −
One possible explanation of a reduction in radio luminosityis simple radiative losses. To estimate the loss time-scale it isnecessary first to estimate the magnetic field strength, whichwe do here.Assuming the minimum energy condition applies to ourtarget, we can estimate the magnetic field B min of each com-ponent along the jet. This assumption folds in informationon the spectral index, size, flux density, the jet compositionand the pitch angle of the magnetic field. We assume thatthe depth of the source is equivalent to the minor axis of thecomponent, unity in the ratio of energy in electrons and ions,no angular dependence on the direction of the magnetic fieldand the line of sight and a frequency range of 0.01 −
100 GHz.The observational constraints come from the integrated fluxdensities, the spectral indices and the deconvolved sizes ofthe components shown in Tab. 1.We compute the magnetic field for all the components.However, as discussed above, the low signal-to-noise ratiocan significantly degrade the accuracy of the fitting parame-ters for each component. Since C3 and C4 are bright resolvedcomponents, the estimates of their magnetic fields are likelymore reliable and are of the order of 1mG. The remainingcomponents show slightly smaller values of B min , decreasingdown the jet. Booler et al. (1982), performing a similar anal-ysis for NGC 4151 using MERLIN data, estimated magneticfields of the same order. MNRAS000
100 GHz.The observational constraints come from the integrated fluxdensities, the spectral indices and the deconvolved sizes ofthe components shown in Tab. 1.We compute the magnetic field for all the components.However, as discussed above, the low signal-to-noise ratiocan significantly degrade the accuracy of the fitting parame-ters for each component. Since C3 and C4 are bright resolvedcomponents, the estimates of their magnetic fields are likelymore reliable and are of the order of 1mG. The remainingcomponents show slightly smaller values of B min , decreasingdown the jet. Booler et al. (1982), performing a similar anal-ysis for NGC 4151 using MERLIN data, estimated magneticfields of the same order. MNRAS000 , 1–13 (2017)
D.R.A. Williams et al.
Figure 6.
CASA eMERLIN spectral index image of the central 3.7 × (cid:48)(cid:48) ( ∼ ×
160 pc) of NGC 4151 at 1.51 GHz. The contours plottedare 2, 4, 9, 16, 25, 36, 49, 64 mJy/beam from the full resolution eMERLIN image. Data were clipped at 2mJy in the full resolution
CASA eMERLIN image so that spurious low signal-to-noise regions were removed. All data were displayed with
CASA and the range ofcolours reduced to − VLBI/VLBA observations resolve further structure incomponents C3 and C4, corresponding to smaller physicalregions in the jet (Ulvestad et al. 1998; Mundell et al. 2003;Ulvestad et al. 2005). Therefore, we calculate B min usingthese higher-resolution data (flux, size and spectral index)for the VLBI central core, D3, located at our component C4and the VLBA component C, located at C3. The B min for D3is 5.38 × − Gauss, while it is 1.35 × − Gauss for VLBAcomponent C. We note that these values are larger by a fac-tor of several tens than the eMERLIN values due mainlyto the smaller size of the component used when calculat-ing B min . Broadly, the range of B min values obtained withthis approach are consistent with the values of the magneticfield of radio jets from the literature for AGN jets/cores(Marscher & Gear 1985; Biretta et al. 1991; Jester et al.2005; Worrall & Birkinshaw 2006). Since the VLBI/VLBA observations do not provide the decon-volved sizes of the components, we used a component size halfthe size of the beam: 1.6 × for VLBI and 2.9 × for VLBA. The synchrotron decay time scale, τ , is related to the mag-netic field strength and observing frequency by τ ∝ B − . ν − . . Although much of the emission from C3 is extendedon scales larger than those probed by VLBI, we can esti-mate the fastest decay time-scale by taking the magneticfield strength from VLBI measurements. This decay time-scale for C3 is ∼
700 years. Thus in 22 years we would onlyexpect a 3 per cent decrease, compared to the observed de-crease of nearly 30 per cent. For the lower surface brightnesscomponents, C1 and C2, the time-scales would be longer.Thus, assuming that the flux density decrease is real, addi-tional energy loss processes are required, e.g. adiabatic ex-pansion losses. Scheuer & Williams (1968) show that, for alinear expansion factor F (and for S ν ∝ ν + α ), the flux densityat a particular frequency will change by F ( α − ) . Thus for α = − ∼ F − . Thus we require only a verysmall expansion (F=1.06) to produce a decrease of 25 percent. The source size is not known but to explain the flux MNRAS , 1–13 (2017) adio jets in NGC 4151 seen by eMERLIN but not detected by VLBA, will probablyexceed a few VLBA beam sizes, i.e. ∼ ∼ − but as the source size and decrease in flux arenot accurately known, this speed is uncertain. However ve-locities of order ∼ − are not unexpected for shocksaround expanding radio sources (Bicknell et al. 1998; Axonet al. 1998; Capetti et al. 1999) and so it is plausible thatadiabatic expansion could explain the decrease in flux of C3.For the lower surface brightness components the char-acteristic radiation decay time-scale is ∼ years, which issimilar to the value estimated by Pedlar et al. (1993). How-ever given our uncertainty in measuring the flux densitychanges we do not speculate further about these compo-nents. Clearly, for C4, injection of new high energy particlesdominates over any radiation or expansion losses. In Fig. 7 we show the HST emission lines images from ournew reduction in H α , [O III], and [O II] and in Fig. 8 weshow an emission line ratio image, [O III]/H α . eMERLINradio contours are superimposed in all cases.In the emission line images (Fig. 7) there is a relativelybright region of emission of size ∼ × (cid:48)(cid:48) (730 pc ×
270 pc)broadly surrounding the radio jet, although the emission lineregion (major PA ∼ ◦ ) and jet (PA ∼ ◦ ) are not exactlyaligned. The [O III]/H α ratio image defines a larger cone-like structure (PA ∼ ◦ ) extending at least 11 (cid:48)(cid:48) (1 kpc) fromthe core. The emission line images show the presence of sub-structures, arches and haloes, indicating a clumpy distribu-tion of the gas.The range of [O III]/H α ratio we show is between 0 and5, based on values obtained for similar ELR for radio galaxies(Baldi et al., in prep). This ratio is a useful diagnostic toolfor investigating the nature of the ionising source, i.e. shocksor photoionisation (Kewley et al. 2006; Allen et al. 2008;Capetti & Baldi 2011). [O III]/H α < α > α values increase from very low values( ∼ . ) near the core and around the radio jet to > . inthe more distant parts of the cone. This change in ratio sug-gests a change in the mechanism of ionisation from shockheating near the core to AGN photoionisation further out.We therefore next consider in more detail whether shocks,such as might be associated with the jet, can explain the ion-isation of the inner, brighter, ionisation region. The brightregion 2 (cid:48)(cid:48) (180 pc) south of the core which is visible in the[O III] and [O III]/Halpha image is a “ghost“ caused by in-ternal reflection (Hutchings et al. 1998). (a)(b)(c) Figure 7.
HST emission line images of NGC 4151 for (a) H α , (b)[O III] and (c) [O II]. The full resolution eMERLIN radio contoursare plotted on top and are at 2, 4, 9, 16, 25, 36, 49, 64 mJy/beam.In all images, north is up and east is to the left. All three imagescorrespond to the central ∼ × (cid:48)(cid:48) ( ∼ ×
550 pc) of the nucleusof NGC 4151.MNRAS000
550 pc) of the nucleusof NGC 4151.MNRAS000 , 1–13 (2017) D.R.A. Williams et al.
Figure 8.
Image showing the ratio of the optical [O III] to H α line emission from HST imaging of the central ∼ × (cid:48)(cid:48) ( ∼ × < to remove spurious artefacts. The larger-scale structure of the ionisation region is seen south west of thecore. This lines up more closely with the PA of the overall line emission ∼ ◦ but it is clear that the inner region bounded by the radiojet is at a slightly different angle at ∼ ◦ . In this image, north is up and east is to the left. The ratio of the radio luminosity to the emission line lu-minosities provide powerful diagnostics of the source of theionisation, (e.g., Pakull et al. 2010). Bicknell et al. (1998)present the [O III] and radio luminosities of various samplesof AGN. They state that the [O III] luminosities are con-sistent with the predictions of a model for ionisation basedon the expansion of radio lobes (Bicknell et al. 1997). Herewe therefore calculate the emission line luminosities of theELR.To determine the H α , [O III] and [O II] luminositiesof the brighter extended ELR surrounding the radio jet weselect a circle of radius 2 (cid:48)(cid:48) (180 pc), centred on C4. We per-formed a standard analysis, using the task RADPROF in IRAF(Tody 1986, 1993) for the aperture photometry of the
HST images F658N, F502N, F375N. To estimate the emissionline fluxes from the ELR we must first measure and removethe large unresolved component from the AGN nucleus. Wetherefore measure the flux within a radius of 0.125 (cid:48)(cid:48) (11 pc) centred on the core. We subtract this flux from the fluxwithin the 2 (cid:48)(cid:48) radius area. We convert the extracted fluxesinto physical units using the parameter
PHOTFLAM , the flux-density normalisation value, and the image bandwidths. Wecorrect the fluxes for reddening using the Calzetti dust ex-tinction law (Calzetti et al. 2000). Furthermore, we esti-mate the contamination from other emission lines fallingwithin these bands. For F658N, to estimate the H α we takeinto account the [N II] doublet, using the observed ratio of[N II] λ α from the optical spectrum (Ho et al.1997a). Only the [O III] λ λλ ,3726,3729 doubletfalls in the F375N band. Hence, the total ELR fluxes withinthe 2 (cid:48)(cid:48) radius but not including the unresolved contribu-tion from the nucleus, are 1.54 × − , 1.10 × − and1.43 × − erg s − cm − , respectively for H α , [O III], and[O II]. Assuming a distance of 19 Mpc for our target, thecalculated luminosities are 6.65 × , 4.74 × and 6.16 × erg s − . The ratio of the [O III] to radio luminosityhere is consistent with the observations presented by Bick-nell et al. (1998). Changes in the positioning and size of the MNRAS , 1–13 (2017) adio jets in NGC 4151 subtraction region of the central core is the major source oferror in measurement of line fluxes from the ELR and maybe up to a factor of 3.Shock models can be used to predict line luminosities.Nelson et al. (2000) used the MAPPINGS II code to estimatethe expected H β luminosity from shocks in NGC 4151. Herewe repeat that analysis using results from the MAPPINGS III code (Dopita & Sutherland 1996; Allen et al. 2008) simi-lar to the analysis of eMERLIN and
HST observations ofM 51b by Rampadarath et al. (submitted). First, takingH β /H α =0.29 from long-slit spectra (Ho et al. 1997a), thenfrom our measured value of H α luminosity (above) we findthat L H β = 1.93 × erg s − . The ratio of H β luminosityto luminosity in ionising radiation produced from shocks is ∝ v − . s , where v s is the velocity of the shock. Here, in agree-ment with Nelson et al. (2000), we find that a low velocityshock overpredicts L H β and we require v s > − ,similar to the calculated jet expansion speed, to reduce thepredicted L H β to the observed level. We note that the MAP-PINGS III code assumes a relatively high conversion efficiencyof jet kinetic power into ionising radiation ( ∼ H β consistent with lower shock velocities. We alsonote that Weaver et al. (1977) refers to low velocity stellarwinds but faster shocks with velocities > − areobserved in narrow line regions of other nearby radio-brightSeyferts, similar to NGC 4151 (Axon et al. 1998; Capettiet al. 1999).The emission line ratios [O III]/H β and [O II]/H β canprovide another constraint on the shock velocities (Allenet al. 2008). Our ratio values from the 2 (cid:48)(cid:48) apertures are ∼ β and ∼
30 for [O II]/H β . These ratios imply v s (cid:29) − , broadly consistent with the velocities derivedfrom L H β .We qualitatively compare the energetics from the AGNradiation with those of the radio jets. We estimate the ki-netic jet power from the radio core luminosity according tothe empirical relation found by Merloni & Heinz (2007) byusing 5 GHz core components from VLA data. For our tar-get, we use the 5 GHz data obtained in Pedlar et al. (1993)and hence calculate a jet power of . × erg s − . We canestimate the AGN radiative power from the X-ray nuclearemission. Wang et al. (2011a) measure the nuclear X-rayemission using Chandra as 1.13 × − erg s − cm − , corre-sponding to a luminosity of . × erg s − , although theydo note that this is variable. Therefore the values derivedfor the energetics of the AGN radiation and of the radiojets are consistent assuming uncertainties in the Merloni &Heinz (2007) relation.Mundell et al. (2003) concluded that the AGN radiationfield is the main source of ionising power in the ELR, butthat the radio jet interaction with line-emitting clouds couldcontribute to the observed ionisiation. By fitting ChandraX-ray spectra, Wang et al. (2011b) also argued that shockscould make up to 12 per cent of the ionised extended emis-sion. Generally, our results are consistent with the conclu-sions of Mundell et al. (2003) and Wang et al. (2011b) aboutthe jet contribution to the ionisation of the ELR. Combinedwith changing ionisation parameter in our [O III]/H α image,we therefore conclude that in the ELR at greater distancesfrom the AGN core than the radio jet e.g. further away fromC4 than 360 pc, that photoionisation from the AGN is the only significant source of ionising power for the ELR. How-ever in regions close to the radio structure ( ≤
360 pc), the jetcannot be ruled out as at least one component of the ionis-ing power. As we cannot fully disentangle the contributionof the two main ionising sources, we cannot quantitativelyestimate the fractional ionisation due to the radio jet relativeto the AGN illumination in the ELR from our results.
The Seyfert galaxy NGC 4151, with twin radio jets, is theradio-brightest of the so-called ’radio-quiet’ AGN and is thusone of the few such AGN for which it is possible to observetemporal changes in the structure of its jets. Here we presenthigh-resolution 1.5 GHz images obtained as part of the LeM-MINGs legacy survey with the eMERLIN array. We comparethese images with those made 22 years previously with thestructurally very similar MERLIN array (M95). These im-ages clearly show that the central AGN (component C4)has brightened by almost a factor 2. The components inthe eastern jet (C5, C6) do not appear to have changedbut components in the western jet, particularly C3, seem tohave decreased in intensity. These observations show thatthe AGN core, C4, is still very much active in NGC 4151and still injecting particles into the inner jet but energy lossmechanisms are dominating further out in the jet.We detect no significant change in the separation be-tween the core, C4, and the only other unresolved compo-nent in the eMERLIN image, C3. The resultant upper limitto the jet velocity ≤ ∼ few mG, consistent with similar MERLIN observations inthe literature. Magnetic fields derived using VLBI observa-tions are factors of 10 higher. The characteristic synchrotrondecay time-scale, inferred from the magnetic field derivedfrom VLBI observations for the unresolved bright compo-nent C3, is of the order of ∼
700 years, longer than that re-quired to account for the flux decrease in 22 years. However,we find that given adiabatic losses, a very small linear ex-pansion factor of 6 per cent can produce the measured fluxdensity decrease of 25 per cent.We present newly-reduced high-resolution optical emis-sion line images (H α , [O III], and [O II]) from HST of thenucleus of NGC 4151. The misalignment between the radioand optical line emission regions has been discussed by pre-vious authors, e.g. Evans et al. (1993) and Pedlar et al.(1993) and are not discussed further here. However we douse these data to investigate the origin of the ionisation of MNRAS000
700 years, longer than that re-quired to account for the flux decrease in 22 years. However,we find that given adiabatic losses, a very small linear ex-pansion factor of 6 per cent can produce the measured fluxdensity decrease of 25 per cent.We present newly-reduced high-resolution optical emis-sion line images (H α , [O III], and [O II]) from HST of thenucleus of NGC 4151. The misalignment between the radioand optical line emission regions has been discussed by pre-vious authors, e.g. Evans et al. (1993) and Pedlar et al.(1993) and are not discussed further here. However we douse these data to investigate the origin of the ionisation of MNRAS000 , 1–13 (2017) D.R.A. Williams et al. the ELR as a function of distance from the AGN. We notethat the [O III]/H α ratio close to the AGN, near the radiojet, is best produced by shocks whereas, far from the AGN( ≥
360 pc), the ratio is more consistent with photoionisation.We also note that the ratio of radio to emission line lumi-nosity shown here is similar to that found by Bicknell et al.(1998) for various samples of radio galaxies and is consis-tent with models for ELR ionisation based on expandingradio lobes. Although our estimate is very uncertain, thederived expansion velocity of C3 is in agreement with thehigher shock velocities needed to avoid overproduction ofH β emission. We also note that the brightest parts of theELR are those directly surrounding the radio jet. We there-fore conclude that although the parts of the ELR furtherfrom the core than the eMERLIN radio jet ( ≥
360 pc) areionised largely by photons from the AGN, in the region ofthe ELR close to the radio jet, the jet is probably also asignificant contributor to the ELR ionisation.Although it is clear that there is a two-sided radio jetinteracting with the ISM in NGC 4151, it remains to be seenwhether it is characteristic of all low-luminosity AGN or isin fact a curious, but intriguing exception to the rule (Ul-rich 2000). To explore the ISM-jet connection on a broaderscale, we require systematic studies of large samples of low-luminosity AGN at sub-arcsec-resolution. Radio observa-tions of nearby sources such as those currently being madewith the LeMMINGs survey are able to resolve sub-kpc scaleradio emission to investigate jet formation. However, only inconjunction with multi-wavelength data of comparable an-gular resolution (such as that obtained from
HST and
Chan-dra ) can we learn about the physics of AGN feedback intothe ISM.
ACKNOWLEDGEMENTS
We thank the anonymous reviewer for their commentsand revisions. We acknowledge funding from the MayflowerScholarship from the University of Southampton afforded toDW to complete this work. This publication has also re-ceived funding from the European Union’s Horizon 2020 re-search and innovation programme under grant agreement No730562 [RadioNet]. IMcH thanks the Royal Society for theaward of a Royal Society Leverhulme Trust Senior ResearchFellowship. RDB and IMcH also acknowledge the support ofSTFC under grant [ST/M001326/1]. JHK acknowledges fi-nancial support from the European Union’s Horizon 2020 re-search and innovation programme under Marie Sk(cid:32)lodowska-Curie grant agreement No 721463 to the SUNDIAL ITNnetwork, and from the Spanish Ministry of Economy andCompetitiveness (MINECO) under grant number AYA2016-76219-P. DMF wishes to acknowledge funding from an STFCQ10 consolidated grant [ST/M001334/1]. EB and JW ac-knowledge support from the UK’s Science and TechnologyFacilities Council [grant number ST/M503514/1] and [grantnumber ST/M001008/1], respectively. FP has received fund-ing from the European Union’s Horizon 2020 Programmeunder the AHEAD project (grant agreement No 654215).We also acknowledge Jodrell Bank Centre for Astrophysics,which is funded by the STFC. eMERLIN and formerly,MERLIN, is a National Facility operated by the Univer-sity of Manchester at Jodrell Bank Observatory on behalf of STFC. Some of the observations in this paper were madewith the NASA/ESA Hubble Space Telescope, and obtainedfrom the Hubble Legacy Archive, which is a collaboration be-tween the Space Telescope Science Institute (STScI/NASA),the European Space Agency (ST-ECF/ESAC/ESA) andthe Canadian Astronomy Data Centre (CADC/NRC/CSA).DW would also like to thank Alessandro Capetti, Sam Con-nolly, Sadie Jones and Anthony Rushton for useful discus-sions.
REFERENCES
Allen M. G., Groves B. A., Dopita M. A., Sutherland R. S., Kew-ley L. J., 2008, ApJS, 178, 20Argo M., 2015, preprint, ( arXiv:1502.04936 )Axon D. J., Marconi A., Capetti A., Macchetto F. D., SchreierE., Robinson A., 1998, ApJ, 496, L75Belles P.-E., Beswick R., Argo M., Jackson N., Muxlow T.,Richards A., 2015, e-MERLIN CookBook - version 3.0, 2015February. JBCABeswick R., Argo M. K., Evans R., McHardy I., Williams D. R. A.,Westcott J., 2014, in Proceedings of the 12th European VLBINetwork Symposium and Users Meeting (EVN 2014). 7-10October 2014. Cagliari, Italy.. p. 10Bicknell G. V., Dopita M. A., O’Dea C. P. O., 1997, ApJ, 485,112Bicknell G. V., Dopita M. A., Tsvetanov Z. I., Sutherland R. S.,1998, ApJ, 495, 680Biretta J. A., Stern C. P., Harris D. E., 1991, AJ, 101, 1632Boksenberg A., et al., 1995, ApJ, 440, 151Booler R. V., Pedlar A., Davies R. D., 1982, MNRAS, 199, 229Calzetti D., Armus L., Bohlin R. C., Kinney A. L., Koornneef J.,Storchi-Bergmann T., 2000, ApJ, 533, 682Capetti A., Baldi R. D., 2011, A&A, 529, A126Capetti A., Axon D. J., Macchetto F. D., Marconi A., Winge C.,1999, ApJ, 516, 187Carral P., Turner J. L., Ho P. T. P., 1990, ApJ, 362, 434Condon J. J., 1987, ApJS, 65, 485Condon J. J., Ransom S. M., 2016, Essential Radio AstronomyDopita M. A., Sutherland R. S., 1996, ApJS, 102, 161Dullo B. T., Mart´ınez-Lombilla C., Knapen J. H., 2016, MNRAS,462, 3800Evans I. N., Tsvetanov Z., Kriss G. A., Ford H. C., Caganoff S.,Koratkar A. P., 1993, ApJ, 417, 82Ferrarese L., Merritt D., 2000, ApJ, 539, L9Filippenko A. V., Sargent W. L. W., 1985, ApJS, 57, 503Gebhardt K., et al., 2000, ApJ, 539, L13Ghisellini G., Haardt F., Matt G., 2004, A&A, 413, 535Gursky H., Kellogg E. M., Leong C., Tananbaum H., GiacconiR., 1971, ApJ, 165, L43Harrison B., Pedlar A., Unger S. W., Burgess P., Graham D. A.,Preuss E., 1986, MNRAS, 218, 775Ho L. C., 2008, ARA&A, 46, 475Ho L. C., Filippenko A. V., Sargent W. L., 1995, ApJS, 98, 477Ho L. C., Filippenko A. V., Sargent W. L. W., 1997a, ApJS, 112,315Ho L. C., Filippenko A. V., Sargent W. L. W., Peng C. Y., 1997b,ApJS, 112, 391Ho L. C., Filippenko A. V., Sargent W. L. W., Peng C. Y., 1997c,ApJS, 112, 391Ho L. C., Filippenko A. V., Sargent W. L. W., 1997d, ApJ, 487,568Ho L. C., Filippenko A. V., Sargent W. L. W., 1997e, ApJ, 487,591Ho L. C., Filippenko A. V., Sargent W. L. W., 2003, ApJ, 583,159 MNRAS , 1–13 (2017) adio jets in NGC 4151 Ho L. C., Greene J. E., Filippenko A. V., Sargent W. L. W., 2009,ApJS, 183, 1H¨onig S. F., Watson D., Kishimoto M., Hjorth J., 2014, Nature,515, 528Hutchings J. B., et al., 1998, ApJ, 492, L115Hutchings J. B., et al., 1999, AJ, 118, 2101Jenkner H., Lasker B. M., Sturch C. R., McLean B. J., SharaM. M., Russel J. L., 1990, AJ, 99, 2082Jester S., R¨oser H.-J., Meisenheimer K., Perley R., 2005, A&A,431, 477Johnston K. J., Elvis M., Kjer D., Shen B. S. P., 1982, ApJ, 262,61Jones S., McHardy I., Moss D., Seymour N., Breedt E., UttleyP., K¨ording E., Tudose V., 2011, MNRAS, 412, 2641Jones S., McHardy I., Maccarone T. J., 2017, MNRAS, 465, 1336Kaiser M. E., et al., 2000, ApJ, 528, 260Kellermann K. I., Sramek R., Schmidt M., Shaffer D. B., GreenR., 1989, AJ, 98, 1195Kettenis M., van Langevelde H. J., Reynolds C., Cotton B., 2006,in Gabriel C., Arviset C., Ponz D., Enrique S., eds, Astronom-ical Society of the Pacific Conference Series Vol. 351, Astro-nomical Data Analysis Software and Systems XV. p. 497Kewley L. J., Groves B., Kauffmann G., Heckman T., 2006, MN-RAS, 372, 961Knapen J. H., Stedman S., Bramich D. M., Folkes S. L., BradleyT. R., 2004, A&A, 426, 1135Kraemer S. B., Schmitt H. R., Crenshaw D. M., 2008, ApJ, 679,1128Magorrian J., et al., 1998, AJ, 115, 2285Marscher A. P., Gear W. K., 1985, ApJ, 298, 114Martini P., Regan M. W., Mulchaey J. S., Pogge R. W., 2003,ApJS, 146, 353Merloni A., Heinz S., 2007, MNRAS, 381, 589Mundell C. G., Pedlar A., Baum S. A., O’Dea C. P., GallimoreJ. F., Brinks E., 1995, MNRAS, 272, 355Mundell C. G., Wrobel J. M., Pedlar A., Gallimore J. F., 2003,ApJ, 583, 192Mundell C. G., Ferruit P., Nagar N., Wilson A. S., 2009, ApJ,703, 802Nelson C. H., Weistrop D., Hutchings J. B., Crenshaw D. M.,Gull T. R., Kaiser M. E., Kraemer S. B., Lindler D., 2000,ApJ, 531, 257Ogle P. M., Marshall H. L., Lee J. C., Canizares C. R., 2000, ApJ,545, L81Pakull M. W., Soria R., Motch C., 2010, Nature, 466, 209Peck L. W., Fenech D. M., 2013, Astronomy and Computing, 2,54Pedlar A., Kukula M. J., Longley D. P. T., Muxlow T. W. B.,Axon D. J., Baum S., O’Dea C., Unger S. W., 1993, MNRAS,263, 471Perez E., Gonzalez-Delgado R., Tadhunter C., Tsvetanov Z.,1989, MNRAS, 241, 31PRau U., Cornwell T. J., 2011, A&A, 532, A71Robinson A., et al., 1994, A&A, 291, 351S´anchez-Gallego J. R., Knapen J. H., Wilson C. D., Barmby P.,Azimlu M., Courteau S., 2012, MNRAS, 422, 3208Scheuer P. A. G., Williams P. J. S., 1968, ARA&A, 6, 321Terashima Y., Wilson A. S., 2003, ApJ, 583, 145Tody D., 1986, in Crawford D. L., ed., Proc. SPIEVol. 627, In-strumentation in astronomy VI. p. 733Tody D., 1993, in Hanisch R. J., Brissenden R. J. V., BarnesJ., eds, Astronomical Society of the Pacific Conference SeriesVol. 52, Astronomical Data Analysis Software and SystemsII. p. 173Ulrich M.-H., 2000, A&ARv, 10, 135Ulvestad J. S., 2003, in Zensus J. A., Cohen M. H., Ros E., eds,Astronomical Society of the Pacific Conference Series Vol. 300,Radio Astronomy at the Fringe. p. 97 Ulvestad J. S., Roy A. L., Colbert E. J. M., Wilson A. S., 1998,ApJ, 496, 196Ulvestad J. S., Wong D. S., Taylor G. B., Gallimore J. F., MundellC. G., 2005, AJ, 130, 936Vila-Vilaro B., et al., 1995, A&A, 302, 58Wang J., Fabbiano G., Risaliti G., Elvis M., Mundell C. G., Du-mas G., Schinnerer E., Zezas A., 2010, ApJ, 719, L208Wang J., Fabbiano G., Elvis M., Risaliti G., Mundell C. G.,Karovska M., Zezas A., 2011a, ApJ, 736, 62Wang J., et al., 2011b, ApJ, 742, 23Weaver R., McCray R., Castor J., Shapiro P., Moore R., 1977,ApJ, 218, 377Wells D. C., 1985, Data Analysis in Astronomy. Springer US,Boston, MAWestcott J., et al., 2017, MNRAS,Wilson A. S., Ulvestad J. S., 1982, ApJ, 263, 576Winge C., Axon D. J., Macchetto F. D., Capetti A., 1997, ApJ,487, L121Winge C., Axon D. J., Macchetto F. D., Capetti A., Marconi A.,1999, ApJ, 519, 134Worrall D. M., Birkinshaw M., 2006, in Alloin D., ed., LectureNotes in Physics, Berlin Springer Verlag Vol. 693, Physics ofActive Galactic Nuclei at all Scales. p. 39Wrobel J. M., 2000, ApJ, 531, 716Zdziarski A. A., Poutanen J., Johnson W. N., 2000, ApJ, 542, 703This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000