Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.mso.anu.edu.au/~geoff/HEA/mundell.pdf Дата изменения: Wed May 31 09:06:47 2000 Дата индексирования: Tue Oct 2 01:00:07 2012 Кодировка: Поисковые слова: galactic plane

Accepted To appear in The Astrophysical Journal.

Parsec-Scale Images of Flat-Spectrum Radio Sources in Seyfert Galaxies
C.G. Mundell Department of Astronomy, University of Maryland, College Park, MD, 20742, USA; A.S. Wilson1 Department of Astronomy, University of Maryland, College Park, MD, 20742, USA;

astro-ph/9908255 23 Aug 1999

J.S. Ulvestad National Radio Astronomy Observatory, P.O. Box O, Socorro, NM, 87801,USA; A.L. Roy2 National Radio Astronomy Observatory, P.O. Box O, Socorro, NM, 87801, USA

ABSTRACT
We present high angular resolution 2 mas radio continuum observations of ve Seyfert galaxies with at-spectrum radio nuclei, using the VLBA at 8.4 GHz. The goal of the pro ject is to test whether these at-spectrum cores represent thermal emission from the accretion disk, as inferred previously by Gallimore et al. for NGC 1068, or non-thermal, synchrotron self-absorbed emission, which is believed to be responsible for more powerful, at-spectrum nuclear sources in radio galaxies and quasars. In four sources T0109,383, NGC 2110, NGC 5252, Mrk 926, the nuclear source is detected but unresolved by the VLBA, indicating brightness temperatures in excess of 108 K and sizes, on average, less than 1 pc. We argue that the radio emission is non-thermal and synchrotron self-absorbed in these galaxies, but Doppler boosting by relativistic out ows is not required. Synchrotron self-absorption brightness temperatures suggest intrinsic source sizes smaller than 0.05,0.2pc, for these four galaxies, the smallest of which corresponds to a light-crossing time of 60 light days or 104 gravitational radii for a 108 M black hole. In one of these galaxies NGC 2110, there is also
1 Adjunct Astronomer, Space Telescope Science Institute 2 Present Address: MPIfR, Auf dem Hugel 69, D-53121 Bonn, Germany


2 extended 0.2 pc radio emission along the same direction as the 400-pc scale jet seen with the VLA, suggesting that the extended emission comes from the base of the jet. In another galaxy NGC 4388, the at-spectrum nuclear source is undetected by the VLBA. We also present MERLIN and VLA observations of this galaxy and argue that the observed, at-spectrum, nuclear radio emission represents optically thin, free-free radiation from dense thermal gas on scales ' 0.4 to a few pc. It is notable that the two Seyfert galaxies with detected thermal nuclear radio emission NGC 1068 and NGC 4388 both have large X-ray absorbing columns, suggesting that columns in excess of ' 1024 cm,2 are needed for such disks to be detectable.
Subject headings: accretion disks|galaxies:active|galaxies:jets| galaxies:nuclei|galaxies:Seyfert

1. Introduction
It has become generally accepted that supermassive black holes SBH lie at the center of most, if not all, galaxies e.g., Richstone et al., 1998; van der Marel, 1999, with some lying dormant and others being triggered into an active phase to produce active galactic nuclei AGN e.g., Haehnelt & Rees, 1993; Silk & Rees, 1998. The power source for this activity is thought to be accretion of material onto the SBH, with the infalling material forming an accretion disk which, depending on detailed conditions, then regulates the fueling rate e.g. Narayan & Yi, 1994; Kato, Fukue & Mineshige, 1998; Blandford & Begelman, 1998. The radius to which these accretion disks extend and hence become more easily observable is not well established, but current AGN uni cation schemes advocate a geometrically thick and clumpy torus e.g. Krolik & Begelman, 1988; Krolik and Lepp, 1989; Pier & Krolik, 1992 or warped thin disk Miyoshi et al., 1995; Greenhill et al., 1995; Herrnstein, Greenhill & Moran, 1996; Pringle, 1996; Maloney, Begelman & Pringle, 1996 which hides the nucleus when viewed edge-on. Our viewing angle with respect to the torus or disk is then responsible for the observed di erences between narrow-line AGNs e.g Seyfert 2's, in which our view of the nuclear broad-line region is obscured edge-on view, and unobscured pole-on view broad-line AGNs e.g. Seyfert 1's. Indirect evidence in support of such tori includes the discovery of broad lines in the polarized hence scattered light of Seyfert 2s Antonucci & Miller, 1985; Tran, 1995, sharp-edged bi-cones of ionized gas e.g., Wilson & Tsvetanov, 1994 photo-ionized by anisotropic nuclear UV radiation perhaps originating from the accretion disk and further collimated by the torus, large gas column densities 1023,25 cm,2 to the nuclei of Seyfert 2's, inferred from photoelectric


3 absorption of soft X-rays Turner et al., 1997 and strong mid-infrared emission in both Seyfert types e.g., Antonucci, 1993; Alonso-Herrero, Ward & Kotilainen, 1996. Recent, high-resolution studies at optical and radio wavelengths have begun to provide more direct evidence for `nuclear' disks on size-scales ranging from the 100-1000-pc diameter dusty disks imaged by HST Ja e et al., 1993; Ford et al., 1994; Carollo et al., 1997 and millimeter interferometry Baker & Scoville, 1998; Downes & Solomon, 1998 to pc-scale disks inferred from HI and free-free absorption studies Mundell et al., 1995; Carilli et al., 1998; Peck & Taylor, 1998; Wilson et al., 1998; Taylor et al., 1999; Ulvestad, Wrobel & Carilli, 1999, down to the 0.25-pc warped, edge-on, Keplerian maser disk in NGC 4258, imaged by the VLBA Miyoshi et al., 1995, Greenhill et al., 1995; Herrnstein, et al., 1996. Theoretical work indicates that UV X-ray radiation from the central engine can heat, ionize and evaporate the gas on the inner edge of the torus Pier & Voit, 1995; Balsara & Krolik, 1993; Krolik & Lepp, 1989. Indeed, simple Stromgren sphere arguments suggest a radius for the ionized region of Rpc = 1:5 N? =1054 s,11=3 ne =105 cm,3 ,2=3 , where N? is the number of nuclear ionizing photons per second and ne is the electron density. Recalling the typical density ne 105,6 cm,3 of the ionized disk in NGC 1068 see below, we expect R 0.3,1.5 pc which is comparable to the tenths of pc to pc-scale resolutions achievable with the VLBA for nearby Seyferts. Recent high angular resolution VLBA radio observations of the archetypal Seyfert 2 galaxy, NGC 1068, by Gallimore et al. 1997, have shown that emission from one of the radio components `S1' may be associated with the inner, ionized edge of the torus. This radio component has a at or rising towards higher frequencies spectrum, suggesting it contains the AGN, and a brightness temperature of up to 4 106 K; it is elongated perpendicular to the inner radio ejecta and extends over 40 mas 3 pc. The radiation mechanism may be free-free thermal emission Gallimore et al., 1997, direct synchrotron emission Roy et al., 1998 or Thomson scattering of a nuclear at-spectrum synchrotron self-absorbed radio core itself not detected by the electrons at the inner edges of the torus Gallimore et al., 1997; Roy et al., 1998. This discovery highlights the possibility of using the VLBA to image the pc-scale disks or tori in other Seyfert galaxies. However, at-spectrum radio sources in AGNs often represent non-thermal synchrotron self-absorbed radio emission with a much higher brightness temperature 108 K than is characteristic of component S1 in NGC 1068. High resolution radio observations are thus required to distinguish between the two emission processes. In the present paper, we report parsec-scale VLBA imaging of ve Seyfert galaxies with at-spectrum radio cores and hundred-pc scale, steep-spectrum, radio jets and lobes. Two of these galaxies also exhibit ionization cones with sharp, straight edges and axes aligned with the radio ejecta. Our goal is to determine whether the at spectrum


4 nuclear radio emission represents thermal emission from the accretion disk obscuring torus or synchrotron self-absorbed emission from a compact radio core source. The paper is organized as follows; Sections 2 and 3 describe the sample selection, observations and reduction techniques whilst in Section 4, the results of the study are presented. Section 5 discusses possible scenarios for the observed radio emission including direct non-thermal radiation from the AGN, emission from supernovae or supernova remnants produced in a starburst, or thermal emission from the accretion disk. The observed brightness temperatures are discussed in the context of the NGC 1068 result and comparison is made with other types of active nuclei such as radio galaxies, radio-loud and radio-quiet quasars. Section 6 summarizes the conclusions. Throughout, we assume H0 =75 km s,1 Mpc,1 and q0 = 0.5.

2. Sample Selection
The radio emission of Seyfert galaxies imaged at resolutions 0001 100 almost always has : the steep spectrum characteristic of optically thin synchrotron radiation. Flat spectrum cores are rare. In order to identify galaxies that may contain radio components similar to `S1' in NGC 1068, we have reviewed both published e.g., Ulvestad & Wilson, 1989, and earlier papers in this series at 6 cm and 20 cm; Kukula et al., 1995 at 3.6 cm and unpublished Wilson, Braatz & Dressel at 3.6 cm VLA `A' con guration surveys and other interferometric studies e.g., Roy et al., 1994. In selecting candidate galaxies for VLBA observations, we used the following criteria: The radio component that is coincident with the optical nucleus the position of which is known to 0002 accuracy e.g., Clements, 1981, 1983, has a at spectrum 0.4, : , between 20 cm and 6 cm or 3.6 cm with the VLA in `A' con guration. This S component must also be unresolved in the VLA `A' con guration at 2cm and or 3.6 cm. The ux density of this component exceeds 5 mJy at 3.6 cm for comparison, the total ux density of component `S1' in NGC 1068 at this wavelength is 14 mJy. There is, in addition, extended, `linear' double, triple or jet-like, steep spectrum radio emission on the hundreds of parsecs kiloparsec scales, or well-de ned, optical ionization cones. The reason for this last criterion is to de ne the axis of ejection of the radio components and thus the expected axis of the accretion disk. We found only six excluding NGC 1068 Seyfert galaxies that satisfy these three criteria in the entire sample of about 130 imaged in the `A' con guration. We omit one of


5 them because of its unfavorable declination 44 , leaving ve for imaging with the VLBA. These galaxies are T0109,383, NGC 2110, NGC 4388, NGC 5252 and Mrk 926.

3. Observations and Reduction 3.1. VLBA Observations
The observations were obtained with the 10-element VLBA Napier et al, 1994 at 8.4 GHz during observing runs in 1997 and 1998, details of which are given in Table 1. Dual circular polarizations Right & Left were recorded for all sources, and only the parallel hands i.e. RR and LL were correlated. T0109,383, NGC 2110 and Mrk 926 were recorded with a 32-MHz bandwidth and two-bit sampling 8 MHz per IF, 4 IFs, 2 polarizations and NGC 4388 and NGC 5252 were recorded with a 16-MHz bandwidth and two-bit sampling. The target sources are too weak to obtain estimates of the phase errors using standard VLBI self-calibration imaging techniques e.g. Walker, 1995; instead the targets were observed in phase referencing mode, in which frequent observations of a nearby bright calibrator are interleaved with target scans and used for fringe tting, which corrects the large phase errors, delays phase variations as a function of frequency and delay rates phase variations as a function of time present in the data Beasley & Conway, 1995. As described below, extending the coherence time in this way improves the signal-to-noise ratio and enables an image of the target source to be made, which can then be used as a starting model for subsequent cycles of self-calibration. Target source plus phase calibrator cycle times are shown in Table 1. This method is similar to that used on smaller, connected-element arrays, such as the VLA known as `phase calibration', but is more problematic for VLBI due to larger and more rapidly varying phase errors. Rapid changes in the troposphere at 8.4 GHz therefore require short switching times to satisfy the condition that the change in atmospheric phase be less than a radian over the switching interval, thus enabling reliable phase connection, without 2-radian ambiguities, for successful imaging of the target source Beasley & Conway, 1995. In addition, less frequent observations were made of a bright calibrator `phase check' source. Data editing and calibration followed standard methods Greisen & Murphy, 1998 and used the NRAO Astronomical Image Processing System aipsvan Moorsel, Kemball & Greisen, 1996. Amplitude scales were determined from standard VLBA antenna gain tables, maintained by NRAO sta , and measurements of Tsys made throughout the run. In addition, all data for source elevations below 5 were removed, and the antennas at Hancock HN, and North Liberty NL were deleted from the NGC 5252 dataset as no


6 fringes were detected to HN, and NL showed poor phase stability due to bad weather. The nal phase corrections, interpolated over time, were used as a guide for additional data editing. Despite short switching times between galaxy and phase calibrator, poor tropospheric conditions and uncertainty in the target source position prevented immediate imaging of the phase-referenced target sources using all the data. Observations of the `phase check' source were therefore used to verify the quality of the phase referencing, before applying the phase corrections to the target sources, and to provide ancillary calibration such as manual pulse calibration and amplitude calibration checks. After imaging the phase calibrator to verify that the corrections derived from fringe tting were valid, phase, delay and rate corrections were applied to the `check source', from phase calibrator scans that were adjacent in time to the check source. Many baselines displayed poor phase coherence at some point in the observing run, preventing a coherent image of the source from being produced initially from the whole dataset. Instead, small time ranges e.g. around 1 hour, within which the ma jority of antennas had less rapidly varying phases, were selected to be used in the initial stages of the imaging process. The `check' source, with calibration applied from the phase calibrator, was imaged for the selected small time range. The resultant image was then used as an input starting model for subsequent cycles of self-calibration. This self-calibration process then enabled the remaining data to be fully calibrated and used to make a nal image of the `check' source. The nal structure, ux and position of each `check' source compared well with previously published images e.g. Browne et al., 1998; Fey & Charlot, 1997 and images produced from our data using self-calibration alone. This method provides an independent consistency check on the phase referencing, increasing our con dence in the images of the target sources. Only one `check' source J0044-3530 was not successfully imaged due to insu cient data i.e. only 3 minutes at very low elevation. The target sources were then imaged using the same method, with natural and uniform data weighting. The uniformly weighted images with robust parameter 0 - Briggs, 1995 are shown in Figure 1. The naturally weighted images, with more sensitivity to extended emission, were used to derive the brightness temperature limits to possible thermal emission from the program galaxies; these limits are 30 lower than those derived from the uniformly weighted images shown in this paper. The uncertainty in the ux scale is taken to be 5 and is included in the total uncertainties in ux densities quoted in Table 2. These errors were derived by adding, in quadrature, the 5 amplitude scale error, the r.m.s. noise in the nal image and the error in the Gaussian tting. The accuracy of the target source positions is dominated by the uncertainty in the


7 position of the phase calibrators 0.4 14 mas; see Table 1. Additional positional errors, due to the transfer of phase corrections from the phase calibrator to the target source, are negligible due to the promixity of each calibrator to its target source.

3.2. MERLIN observations
NGC 4388 was not detected by the present VLBA observations. We therefore obtained and analyzed MERLIN 6-cm 4.993-GHz archival data for NGC 4388, whichwas observed on 7th December, 1992 with six antennas. Phase referencing was performed with regular observations of 1215+113, interleaved throughout the observing run and 3C286 was used for ux and bandpass calibration. A ux of 7.087 Jy for 3C286 was adopted, assuming a total ux density of 7.382 Jy Baars et al., 1977 and correcting for MERLIN resolution e ects. After initial gain-elevation corrections and amplitude calibration using MERLIN software, the data were transferred to aips for all subsequent phase and amplitude calibration, data editing and imaging. Dual polarizations were recorded for a 15-MHz bandwidth, centered at 4.993 GHz, but the right circular polarization data were removed due to instrumental problems, resulting in a nal image of the left circular polarization only Figure 2.

4. Results
Five at-spectrum-core Seyferts, were observed with the VLBA at 8.4 GHz. Four of the ve sources were detected T0109,383, NGC 2110, NGC 5252, Mrk 926 and show compact, unresolved cores with brightness temperatures TB 108 K, total luminosities at 8.4 GHz of 1021 W Hz,1 and sizes, on average, less than 1 pc. In addition to the core emission, NGC 2110 shows extended emission which may represent the inner parts of the radio jets, and NGC 5252 may be marginally extended Figure 1. NGC 4388 is not detected with the VLBA, but is detected at 5 GHz with MERLIN Figure 2. We nd no evidence for emission to a 3- limit of TB 106 K extended perpendicular to the hundred-pc scale radio emission in T0109,383, NGC 2110, NGC 5252 or Mrk 926, as would be expected for emission from an accretion disk, but we discuss the possibility of thermal emission from NGC 4388 Section 5.4. The measured and derived properties of each source are listed in Table 2, while more detailed properties of NGC 2110 and NGC 4388 are given in Tables 3 and 4 respectively. The properties of each source are discussed more fully below. Distances are calculated assuming H0 = 75 km s,1 Mpc,1 and q0 = 0.5, except for NGC 4388 which is assumed to be at the distance of the Virgo cluster, taken to be 16 Mpc.


8

4.1. T0109,383
T0109,383 NGC 424 is a highly inclined 75 early-type RSBr0 a de Vaucouleurs et al. 1991 Seyfert galaxy at a distance of 46.6 Mpc. The nucleus of T0109,383, originally classi ed as a Seyfert type 2 Smith, 1975, exhibits strong line emission from highly ionized species such as Fe vii 5720,6086 and Fe x 6374 Fosbury & Sansom, 1983; Penston et al., 1984. Analysis of the continuum emission from the far IR to the far UV and decomposition of the H Nii blend led Boisson & Durret 1986 to suggest a re-classi cation of T0109,383 to a Seyfert 1. The recent discovery of broad components to the H and H lines, along with emission from Fe ii, con rms the type 1 classi cation Murayama et al., 1998. VLA images of the radio emission at 6 and 20 cm, show the nuclear radio source to consist of an unresolved core with a 20 at spectrum 6 =0.170.07 between 6 and 20 cm, and a weaker, secondary, steep spectrum component '100 4 east of the core Ulvestad & Wilson, 1989. Similar radio : structure is seen in the 8.4-GHz VLA image Braatz, Wilson & Dressel, unpublished, 6 shown in Figure 1, with the core spectrum remaining relatively at 3:5=0.21 between 6 and 3.5 cm Morganti et al., 1999. The results of Gaussian tting to the 8.4-GHz VLBA image Figure 1, given in Table 2, show the sub-pc scale nuclear emission to be unresolved, with a peak brightness of TB 8.1 108 K, adopting a source size smaller than half of the beamsize. The peak and integrated 8.4-GHz VLA uxes for the core, 10.4 mJy beam,1 and 11.2 mJy respectively, are in excellent agreement with those measured from the VLBA image Table 2, indicating that little nuclear emission was missed by the VLBA. A similar peak brightness of 10.4 mJy beam,1 is found in the 3.5-cm ATCA image of Morganti et al. 1999, while their slightly higher integrated ux includes some of the emission ' 100 E and W of the nucleus Ulvestad & Wilson, 1989; Figure 1. The excellent agreement between the nuclear 3.6-cm uxes in observations spanning six years indicates no signi cant variability. In the VLBA image, we detected no extended emission in the N-S direction as might be expected from a parsec-scale, thermal disk if the arcsec-scale, steep spectrum, E-W emission in the VLA image is interpreted as emission from nuclear ejecta brighter than 1.3 106 K 3 in the naturally weighted image and more extended than 0.27 pc half of the beamsize in the naturally weighted image.

4.2. NGC 2110
NGC 2110 was initially classi ed as a Narrow Line X-ray Galaxy, NLXG, Bradt et al., 1978, and lies in an S0 E host galaxy Wilson, Baldwin & Ulvestad, 1985 at a distance


9 of 30.4 Mpc. Such NLXG's have a su cient column of dust to the nucleus to obscure the broad line region, thus leading to a Seyfert 2 classi cation of the optical spectrum, but an insu cient gas column to attenuate the 2 10 keV emission, so the hard X-ray luminosity is comparable to those of Seyfert 1's Weaver et al., 1995a; Malaguti et al., 1999. Early radio observations found NGC 2110 to be a strong radio source Bradt et al., 1978 and subsequent VLA imaging Ulvestad & Wilson, 1983; 1984b showed symmetrical, jet-like radio emission, extending 400 in the N-S direction and straddling a central compact core. A more recent VLA A-con guration image at 3.6 cm, obtained by Nagar et al. 1999 and shown in Figure 1, contains a wealth of complex structure. Ulvestad & Wilson 1983 found 20 the spectrum of the core to be relatively at spectral index 6 0.360.05 between 20 6 cm and 6 cm, but becoming steeper 2 0.960.09 between 6 cm and 2 cm assuming no time variability. Using the 3.6-cm core ux measurement of Nagar et al. 1999 and 6 3 ignoring variability or resolution e ects gives spectral indices of 3:6 = 0.61 and 2:6 = 1.31, also suggesting a steepening of the spectrum at higher frequencies. The radio continuum emission of NGC 2110, imaged with the VLBA at 3.6 cm and shown in Figure 1, consists of a compact core, presumably the nucleus, and slightly extended emission which is most pronounced to the north. The results of tting a single-component Gaussian are given in Table 2; the fact that the integrated ux is signi cantly higher than the peak ux also suggests the source is resolved. Resolved structure is also evident in the time-averaged u; v data not shown, consistent with an unresolved point source with a ux densityof 8 mJy superimposed on an extended halo" with approximate dimensions of 2.5 N-S 0.5 E-W mas. Preliminary two-component Gaussian ts to the image are also consistent with an unresolved point source and an extended component. We therefore subtracted an 8-mJy point source in the u; v plane using the aips task uvsub positioned at the peak of the 3.6 cm VLBA image, and studied the residual emission. This emission is extended both north and south of the core by 0.7 mas, consistent with emission from the inner regions of the northern and southern jets. Using the brightness of 8 mJy beam,1 for the unresolved component and assuming an upper limit to the source size of 0.94 0.36 mas half of the beamsize, we nd TB 6.0 108 K. In addition to the core and extended emission, the Gaussian ts suggest the presence of a third component, centered 1.95 mas north of the core; its size and direction of elongation are not well constrained. This component may be a knot in the northern jet. A summary of the tted properties of each component is given in Table 3. The total VLBA-detected ux density of the source zero baseline ux measured in the uv plane is 30 mJy. This ux density is lower than the previously measured VLA core ux of 77.6 mJy at this frequency Nagar et al., 1999, presumably due to the high spatial


10 resolution of the present observations and missing short spacings of the VLBA compared to the VLA, thereby reducing our sensitivity to extended structure. This may also explain why we detect no VLBA counterpart to the small eastern extension present in the 3.6-cm VLA image, which contains about 3.6 mJy of ux and extends approximately 0005 east of : the core Nagar et al., 1999. Alternatively, the extension in the VLA image may be a result of instrumental e ects caused by the source position being close to the celestial equator and the short duration of the snapshot observation, an e ect termed `equator disease' Antonucci & Ulvestad, 1985. In the VLBA image, we detect no extended emission in the E-W direction such as might be expected from a parsec-scale thermal disk brighter than 3.1 106 K 3 in the naturally weighted image, and more extended than 0.07 pc one half of the E-W beamsize in the naturally weighted image.

4.3. NGC 4388
NGC 4388 is a nearby, edge-on spiral galaxy SBsb pec - Phillips & Malin, 1982 which is thought to lie close to the centre of the Virgo cluster Phillips & Malin, 1982 and may be tidally disturbed by nearby cluster core galaxies M84 or IC3303 Corbin, Baldwin & Wilson, 1988. Ionization cones extend approximately perpendicular to the disk Pogge, 1988; Corbin et al., 1988; Falcke, Wilson & Simpson, 1998 and the kinematics of the ionized gas in the narrow line region NLR shows a complex combination of rotation and out ow Corbin et al. 1988; Veilleux, 1991; Veilleux et al., 1999. The nucleus is variously classi ed as Seyfert type 1 or 2, with the high galactic inclination and obscuring dust lanes making unambiguous classi cation di cult Falcke et al., 1998. Shields & Filippenko 1988 report broad, o -nuclear H emission, but subsequent IR searches for broad lines such as Pa Blanco, Ward & Wright, 1990; Ruiz, Rieke & Schmidt, 1994 and Br and Br Veilleux, Goodrich & Hill, 1997 have failed to detect a broad nuclear component. Previous radio continuum images of NGC 4388 Stone et al., 1988; Carral, Turner & Ho, 1990; Hummel & Saikia, 1991; Falcke et al., 1998 show complex, extended structure, both along the galactic plane and perpendicular to it. A recent 3.5 cm VLA image of the extended radio emission Falcke et al., 1998 shows, in more detail, the `hour-glass'-shaped : radio out ow to the north of the galactic plane, and the compact 100 9 separation central double, which were suggested by earlier images. In Section 4.3.1 we concentrate on the radio emission from the northern component of the compact radio double, which shows a at spectrum up to 2 cm Carral, Turner & Ho, 1990 and is thoughttobethe nucleus, and in Section 4.3.2, we discuss the extended emission to the SW.


11
4.3.1. The nucleus

As stated earlier, NGC 4388 is not detected in the 8.4-GHz VLBA observations, with a 3- brightness temperature limit of TB 2.2 106 K =63.2 Jy beam with a beam size of 2.52 1.46 mas in the naturally weighted map, with a factor 1.7 applied to correct for decorrelation due to residual imperfections in the phase referencing corrections, estimated using the check source. We do, however, detect emission from NGC 4388 at 6cm with MERLIN. The uniformly weighted MERLIN image Figure 2 shows emission from two components, labelled M1 and M2, the stronger of which we identify with the nucleus and discuss in more detail here, while M2 is discussed in Section 4.3.2. The nuclear component M1, has a peak brightness of 1.2 mJy beam,1 which corresponds to a brightness temperature TB 2.4 104 K at 5 GHz beamsize 91 39.5 mas, see Table 4. The nucleus is unresolved in the MERLIN data, indicating that the source size is intermediate between the MERLIN and VLBA beam sizes. However, a combination of the MERLIN and VLBA results with published spectral index information can further constrain the source size. Earlier radio observations of NGC 4388 have found the nuclear spectrum to be at from 1.49 GHz to 15 GHz. The spectral index was measured to be = 0.26 between 1.49 GHz and 4.86 GHz with a relatively large beamsize of 100 2 Hummel & Saikia, 1991 and : Carral et al. 1990 report a at spectrum up to 15 GHz with an upper limit to the nuclear size of 70 mas. Including the VLA 8.4 GHz core ux of Kukula et al 1995 suggests that the spectrum of the nucleus may be very slightly inverted between 8.4 GHz and 15 GHz = 0.05 but within the errors it can be taken as at. We therefore used the measured MERLIN 5-GHz peak ux to derive predicted VLBA 8.4 GHz uxes of the nucleus, for spectral indices of both = 0.0 and 0.26, and converted these predicted uxes to brightness temperatures, assuming the source is unresolved by the representative VLBA beamsize of 2.52 1.46 mas. These predicted temperatures are listed in Table 4 and are above the detection threshold of the VLBA observations for a source size equal to or smaller than the VLBA beam. The larger predicted brightness temperature, for a source size equal to the VLBA beam, of TB ' 8.3 106 K is, however, only 3.8 times greater than our 3- , VLBA detection limit and so the solid angle of the source need only be 3.8 times larger than the VLBA beamsize to be undetected. We therefore constrain the size of the nucleus to be 3.7 mas =0.0 or 0.3 pc. Sensitive, high angular resolution VLBA observations at lower frequencies such as 2.3 GHz and 1.4 GHz are required to determine the actual size and structure of the nucleus in NGC 4388.


12
4.3.2. Col limated radio emission

In addition to the core emission at 5 GHz, the MERLIN image of NGC 4388 shows a second weak component SW of the nucleus labelled M2 in Figure 2. This component is only 0.2500 away from the nucleus, lying along PA 211 with respect to the core, and should not be confused with the stronger, more distant radio component seen in previous VLA images e.g. Falcke et al, 1998, which lies 100 9 from the nucleus in PA 201 . To : establish the reality of this weak MERLIN component, the VLA 8.4-GHz data published by Falcke et al., 1998 were re-examined. The uniformly weighted image, shown in Figure 3a, has a rather elongated synthesized beam of size 0007 00022 PA 74 , but a bridge of : : emission is clearly visible, connecting the two main radio components labelled V1 & V2. A similar bridge of emission was seen in the 4.86-GHz VLA image of Hummel & Saikia 1991 and the 8.4-GHz VLA snapshot image of Kukula et al. 1995. The crosses indicate the positions of the nucleus and the weak component M2 visible in the MERLIN 5 GHz image. Figure 3b shows a super-resolved image produced from the same VLA data, with a circular synthesized beam of 000 22. Again the crosses mark the MERLIN components, and : an extension of the 8.4 GHz VLA emission is seen at the location of the weak MERLIN component M2, further suggesting that the MERLIN emission is real. The super-resolved image also suggests that a well-collimated jet of emission is emanating from the core V1 along PA 210 for 100 5, before changing direction at V2. The MERLIN component M2 : would then be an inner part of this collimated radio ejection. Falcke et al 1998 nd a good association between V2 and a `spike' of optical line emission, suggesting interaction of the radio jet with a cloud in the NLR. The `hooked' shape of the southern radio jet, suggested by Fig. 3b, is reminiscent of similarly well-collimated radio jets seen in an increasing number of Seyferts e.g., Mrk 34 - Falcke et al., 1998; Mrk 3 - Kukula et al. 1993; NGC 1068 - Wilson & Ulvestad, 1983; NGC 2110 - Ulvestad & Wilson, 1983; Nagar et al., 1999, which are de ected or terminate at radio hot spots. However, sensitive high angular resolution radio observations are required to image the detailed structure of the proposed radio jet in NGC 4388.

4.4. NGC 5252
NGC 5252, an S0 galaxy de Vaucouleurs et al., 1991 with a Seyfert type 1.9 nucleus Acosta-Pulido et al., 1996, exhibits a dramatic bi-cone of ionized gas Tadhunter & Tsvetanov, 1989; Wilson & Tsvetanov, 1994; Acosta-Pulido et al., 1996 extending 35 kpc from the nucleus along PA 165 and argued to be ionized by anisotropic nuclear UV radiation. The intrinsic anisotropy of the ionizing radiation was nicely con rmed by HI


13
21-cm images, which show neutral hydrogen lling the regions outside the bi-cone Prieto

&Freudling, 1993, 1996. Observations with the VLA Wilson & Tsvetanov, 1994 show a radio structure consisting of a central, compact core, with a relatively at spectrum between 20 cm and 20 6 6 cm 6 = 0.22, steepening at shorter wavelengths 3:6 = 0.78, and weaker emission extending 200 north PA 345 and south PA 175 of the core. A second compact radio component, seen 2200 north PA 8: 9 of the nucleus, may be associated with NGC 5252, lying close to the PA of the extended core emission and coinciding with a region of high excitation ionized gas Wilson & Tsvetanov, 1994, or may be a background source Morse et al., 1998. More recent radio images by Nagar et al. 1999 at 20 cm and 3.6 20 cm con rm the overall spectrum of the core 3:6 = 0.32 and the radio continuum features seen in the earlier images. The 8.4-GHz VLBA image of the core of NGC 5252 is shown in Figure 1, along with the larger scale 20-cm VLA image from Nagar et al., 1999, and the results of Gaussian tting to the emission are detailed in Table 2. The deconvolved size of the core is less than half of the beamsize and so we classify the emission as unresolved. The peak brightness of 7.9 mJy beam,1 corresponding to TB 4.2 108 K is in agreement with that measured by Nagar et al. 1999 7.9 mJy beam,1 and slightly higher than that found by Wilson & Tsvetanov 1994 6.7 mJy beam,1. The integrated VLBA ux 9.1 mJy is also very similar to that of Nagar et al. 1999 9.3 mJy. The similarity of the VLBA- and VLA-measured nuclear uxes indicates that little emission was missed by the VLBA. The fact that the measured integrated VLBA ux is 15 higher than the peak and the possible visible extension to the SW Fig. 1 are evidence for weak extended emission. The quality of the data, however, make this detection tentative and higher sensitivity observations are required to con rm the extension and determine its structure.

4.5. Mrk 926
Mrk 926 MCG-2-58-22, the most distant ob ject in our sample at z=0.0473, is a type 1 Seyfert and was rst identi ed as a Seyfert through its luminous Lx = 5 1044 erg s,1 X-ray emission Ward et al. 1978. Later X-ray observations found a relatively at X-ray spectrum although detailed model ts are controversial - George et al., 1998 and Weaver et al. 1995b suggested that their ASCA observations of Mrk 926 were consistent with a bare" Seyfert 1 nucleus. X ray variability, with a 14 year timescale, provides an upper limit of 4.3 pc for the size of the emitting region Weaver et al., 1995b.


14 The 8.4 GHz VLA image Braatz, Wilson & Dressel, unpublished, shown in Figure 1, shows a compact core with weak E-W extensions con rming the earlier 6-cm image Ulvestad & Wilson, 1984a. The nuclear radio spectrum is relatively at between 6 cm 6 20 and 3.6 3:6 = 0.24, but steepens at longer wavelengths 6 = 1.1. This steepening of the spectrum may, in part, be due to inclusion of the extended E-W emission in the low resolution 20-cm observation Wilson & Meurs, 1982, but is also consistent with a lack of free-free absorption toward the nucleus, supporting the X-ray classi cation as a bare" Seyfert 1 nucleus. The 8.4-GHz VLBA image Figure 1 shows an unresolved source. The results of Gaussian tting to the VLBA detection of Mrk 926 and corresponding derived quantities are given in Table 2. A peak brightness of 7.5 mJy beam ,1 is measured in the VLA image while the peak brightness in the VLBA image is 4.6 mJy beam,1 . No additional extended ux is present in the VLBA data as can be seen from the image to a 3- limit of TB 1.3 106 K and the agreement between the peak and integrated ux densities Table 2. The upper limit to the size of the unresolved core is 1.3 0.5 pc half of the beamsize. The fact that less ux is detected with the VLBA than the VLA, however, suggests additional emission, extended on scales between the VLA 00027 00021 : : and VLBA beams 2.8 1.0 mas, or time variability.

5. Discussion 5.1. Compact Cores and Flat Spectra
In contrast to the commonly observed high brightness temperature, at spectrum cores in quasars, blazars and radio galaxies e.g. Pearson et al., 1998; Kellerman et al., 1998, Hough et al., 1999, at spectrum cores in Seyfert galaxies appear to be rare e.g. de Bruyn & Wilson 1978; Sadler et al., 1995; Morganti et al., 1999, with only 10 detected in the sample of de Bruyn and Wilson 1978. High brightness temperature, at spectrum cores are thought to represent the synchrotron self-absorbed base of a relativistic jet produced by the central engine e.g., Peterson, 1997, while the steep spectrum components are considered to be associated with shocks along the jet. Despite the small linear extents of Seyfert radio jets compared with those in quasars and radio galaxies, the relative proximity of Seyferts permits higher linear resolution to be achieved at comparable angular resolutions to studies of quasars and radio galaxies. However, various factors, such as free-free absorption by broad line region BLR or other dense gas and dominance by steep spectrum emission, may make it di cult to detect at spectrum radio cores in Seyfert galaxies. In addition, evidence is mounting that Seyfert radio jets are not relativistic Ulvestad et al., 1999, but instead may be thermally dominated Bicknell et al., 1998; Wilson &Raymond,


15 1999, thereby removing Doppler boosting of the putative at spectrum core emission as a means for increasing its detectability. As shown by recent studies of NGC 1068 Gallimore et al., 1997; Roy et al., 1998 an alternative explanation for at spectrum cores in Seyferts is emission from the obscuring torus itself, either as thermal bremsstrahlung Gallimore et al., 1997, direct synchrotron radiation Roy et al., 1998 or scattered synchrotron radiation from a self-absorbed core Gallimore et al., 1997; Roy et al., 1998. Here, we discuss our results in this context, examining whether the emission is likely to be thermal or non-thermal, central engine or starburst related and whether, if associated with a central engine, Doppler boosting could be important.

5.2. Nuclear emission - thermal or non-thermal?
High angular resolution 8.4-GHz VLBA observations of the at-spectrum nuclear component, `S1', in NGC 1068 have revealed an edge-on, disk-like structure, aligned roughly perpendicular to the radio jet axis, with an extent of 0.5 pc and a brightness temperature ranging from 5 105 to 3.7 106 K Gallimore et al., 1997. In contrast, we nd high TB 108 K, unresolved 1 pc cores in T0109,383, NGC 2110, NGC 5252 and Mrk 926, consistent with non-thermal emission. The brightness temperature of a synchrotron self-absorbed source is TB ' me c2 3k ' 2.0 109 K, where is the Lorentz factor, me the electron mass, c the speed of light and k Boltzmann's constant. The maximum TB for such a source is 1012 K, limited by the inverse Compton catastrophe", in which cooling by inverse Compton scattering quickly reduces the brightness temperature to 1011,12 K e.g., Kellermann & Pauliny-Toth, 1969; Kellermann & Pauliny-Toth, 1981. Readhead et al. 1994 argue that a more physical limit is the equipartition brightness temperature, which limits the intrinsic brightness temperature in the emission rest frame to be 1011 K consistent with average observed values -Lahteenmaki, Valtao ja & Wiik, 1999. Although observed TB 's as high as 1016 K have been inferred from intra-day variability seen in some blazars e.g., Crusius-Waetzel & Lesch, 1998, and in one extreme case, intra-hour variability in the quasar PKS 0405,385, suggesting TB 1021 K if the variability is intrinsic to the source or TB 5 1014 K if explained by interstellar scintillation of a source smaller than 5as Kedziora-Chudczer et al., 1997, the brightness of these ob jects is thought to be extremely Doppler boosted by relativistic out ows. Our measured lower limits to the TB of T0109,383, NGC 2110, NGC 5252 and Mrk 926 are in the range 2,8 108 K and are therefore consistent with synchrotron self-absorption, but do not require Doppler boosting. For =1, the intrinsic brightness temperature of a self-absorbed


16 source is 2 109 K, suggesting source sizes smaller than 0.2,0.3 mas, or 0.05,0.2pc, for these four galaxies, the smallest of which corresponds to a light-crossing time of 60 light days or 104 gravitational radii for a 108 M black hole. Such small sources may be resolved at 8 GHz with future space VLBI missions such as ARISE Ulvestad, Gurvits & Lin eld, 1997; Ulvestad & Lin eld, 1998. The spectra of the nuclei of NGC 5252 and NGC 2110 appear to steepen towards higher frequencies as discussed in Section 4, consistent with the cores becoming optically thin to synchrotron self-absorption at 6 cm as suggested for NGC 5252 by Wilson & Tsvetanov, 1994. Alternatively, free-free absorption by ionized gas, with optical depths 1.4 GHz 0.6, could account for the observed spectral indices between 20 cm and 6 cm for T0109, NGC2110 and NGC5252, assuming that the intrinsic emission is optically 20 thin synchrotron with 6 = 0.7. No evidence of thermal disk-like emission, extended perpendicular to the collimation axis, is found in these sources to a 3- limit of 106 K see Figure 1. Any emission similar to the brighter, inner 0.5-pc disk in NGC 1068, could just be spatially resolved in T0109,383, NGC 2110 and NGC 5252, but not in more distant ob jects like Mrk 926. Nevertheless, these four Seyferts are dominated by the compact, high TB core emission which is completely di erent from NGC 1068, which shows no compact, unresolved core. The absence of a bright core in NGC 1068 may re ect absorption by a BLR cloud or by the probably thermal disk seen in radio continuum. As we discuss below, a high column density of ionized gas in such a disk is needed for detectable thermal radio emission and the disk could hide the compact, high brightness core through free-free absorption. The column density towards the nucleus of NGC 1068 is thought to be so high 1026 cm,2 Matt et al., 1997 that it is optically thick to Compton scattering and the nucleus is totally hidden from view even in hard X-rays. The X-ray inferred column densities for three of the core-dominated Seyferts are signi cantly lower at 2.4 1022 cm,2 for NGC 2110 Weaver et al., 1995a, 3.4 1022 cm,2 for NGC 5252 Turner et al., 1997 and 5.7 1020 cm,2 for Mrk 926 Weaver et al., 1995b. Given the detection of the sub-pc scale, non-thermal radio sources, the disk cannot be optically thick to free-free absorption if the synchrotron self-absorbed radio core is seen through it. We may then calculate an upper limit to the thermal radio emission from the disk by assuming 8.4 GHz 0.5 and that the entire column density inferred from X-ray photoelectric absorption is fully ionized. As the disks in NGC 2110 and NGC 5252 are expected to lie along the beam minor axis, we assume a conservative upper limit to the disk diameter of twice the minor axis beamsize and use the height to diameter ratio of 2:1 as found for NGC 1068, giving maximum disk dimensions of 0.22 0.11 pc and 0.86 0.43 pc


17 for NGC 2110 and NGC 5252 respectively3 . The assumed upper limits to the disk size, column density and free-free optical depth yield lower limits to both the electron densities ne 7.3 104 cm,3 and 2.6 104 cm,3, and electron temperatures Te 3 104 K and 1.8 104 K for NGC 2110 and NGC 5252 respectively, and therefore a maximum predicted 8.4-GHz ux of 1.4 Jy for both, undetectable with present observations. These calculations indicate that thermal radio emission from an accretion disk is most likely to be detected from nuclei with high 1024 cm,2 X-ray inferred column densities, as maybethe case for NGC 4388 see Section 5.4. Thus, further insight into the nature of thermal radio cores in Seyferts might be gained with milliarcsecond-resolution observations of galaxies with high 1024 cm,2 X-ray inferred column densities.

5.3. Seyfert nuclei - starburst or accretion-powered central engine?
Akey question is whether the nuclear power source in radio-quiet quasars, Seyferts and ultra-luminous infrared galaxies is a compact starburst or accretion onto a supermassive black hole. Whilst emission from hot stars in the torus might account for the featureless continua in Seyfert 2's Fernandes & Terlevich 1995; Gonz alez-Delgado et al. 1998, it seems that starbursts cannot provide the necessary collimation to produce radio jets. The existence of radio jets is, therefore, often used as an indication of the presence of a black hole plus accretion disk. Although some Seyferts are now known to possess strikingly collimated jets e.g., Nagar et al., 1999; Kukula et al., 1999, the resolution of the radio images is often insu cient to demonstrate the high degree of collimation seen in radio galaxies and radio-loud quasars. Although compact starbursts may co-exist with AGNs in some Seyferts e.g., Heckman et al., 1997; Gonz alez-Delgado et al., 1998; Carilli et al., 1998, we have argued that the sub-pc scale, high brightness radio emission from T0109,383, NGC 2110, NGC 5252 and Mrk 926 is dominated by the central engine. Radio emission from a starburst region consists of synchrotron radiation from supernova remnants SNR's plus thermal free-free emission from HII regions. The brightness temperature of such a region cannot exceed 105 K at 1 GHz Condon, 1992 and so the high brightnesses of our Seyfert nuclei, along with their sub-pc sizes rule out a starburst origin for the radio emission. In fact, these Seyfert nuclei show similar brightness temperatures to some radio-quiet quasars Blundell & Beasley,
An edge-on disk geometry is unlikely for Mrk 926, a Seyfert 1, and the inferred column density for this galaxy is muchlower than that of NGC 2110 or NGC 5252. Thus the predicted radio ux from Mrk 926 will be lower than from NGC 2110 or NGC 5252.
3


18 1998 and LINERs Falcke et al., 1999, which are also argued to be non-thermal emission from black-hole powered central engines rather than compact starbursts. In NGC 4388, the well-collimated radio jet also suggests the presence of an AGN, despite the lack of an unresolved, high brightness nucleus. We can also rule out individual or a collection of extremely bright radio supernovae RSN as an explanation for the Seyfert core emission. Although these rare, bright RSN, of the kind observed in NGC 891 van Gorkom et al. 1986, designated SN1986J and classi ed as an unusual type II radio supernova Weiler, Panagia & Sramek, 1990, can display high brightness temperatures at the peak of their light curves e.g. 109 Kat 6 cm at the peak of the SN1986J light curve, their uxes increase and decrease over a timescale of a few years. The 6-cm ux of SN1986J doubled in approximately three years and halved again over the following three years. Measured VLA uxes for our Seyfert cores have typically varied by less than 30 over 10 years and are therefore inconsistent with an RSN interpretation. In addition, the maximum inferred 8.4-GHz luminosity for SN1986J of 8 1020 W Hz,1 6 calculated using the 6-cm ux at the peak of the light curve, 3:6 = 0.7, and assuming isotropic emission at a distance of 8.96 Mpc is somewhat lower than the luminosities of our Seyfert nuclei Table 2. We therefore conclude that the high brightness temperatures 108 K, small sizes 1 pc and absence of strong ux variations over 10-year timescales in T0109,383, NGC 2110, NGC 5252 and Mrk 926 are probably indicative of synchrotron self-absorption close to a jet-producing central engine. The lower limits to the brightness temperatures do not require relativistic motions, which is consistent with the non-relativistic proper motions observed in two Seyfert galaxies by Ulvestad et al., 1999.

5.4. Thermal Bremsstrahlung Emission in NGC 4388
Unlike the other four galaxies discussed in this paper, no high brightness temperature compact radio core is detected in NGC 4388 at 8.4 GHz. The observationally inferred brightness temperature of 2.4 104 K TB 2.2 106 K the lower limit at 5 GHz, the upper limit at 8.4 GHz, see Section 4.3.1 is too low for synchrotron self absorption to be important. Instead we consider a model in which the emission is optically thin, thermal bremsstrahlung, compatible with the observed at radio spectrum, from a gas with an electron temperature of Te K. Our bremsstrahlung model consists of a thermal plasma of uniform temperature, Te , and density, ne, that lls the emitting volume, V , with a lling factor f . The maximum


19 source volume is set by the MERLIN beamsize and assumes an ellipsoidal geometry with semi-axes of 3.6 1.5 1.5 pc, while the minimum source size, as discussed in Section 4.3.1, is 0.3 pc, for which an edge-on disk geometry is assumed disk diameter 0.4 pc, height 0.2 pc. Figures 4a,b,c show the allowed range of densities ne, opacities at 4.993 GHz and ionized gas column densities Ne as a function of Te , within the limits set by the possible source sizes. As shown by the shaded area in Figure 4b, the gas becomes optically thin at 4.993 GHz for Te 104:5 K, suggesting ne 1.6 104 f ,0:5 cm,3 for the MERLIN size limit, or Te 106:8 K and ne 1.8 106 f ,0:5 cm,3 for the smaller source size set by the VLBA limit. The larger electron density, for f =1, is intermediate between that expected in Seyfert narrow line regions e.g., 103 cm,3 - Koski, 1978 and broad line regions 109 cm,3 , consistent with the small size of the region, and similar to ionized gas densities of 105 107 cm,3 inferred from free-free absorption in the inner parsec of Mrk 231 and Mrk 348 Ulvestad et al., 1999. Similarly high electron temperatures and densities have been derived for thermal emission from the torus gas in NGC 1068 Gallimore et al., 1997, thus implying that we may be seeing the same phenomenon in NGC 4388. The lower limit to the electron column density of Ne 7 1022 cm,2 Figure 4c is compatible with the total column of 4.2 1023 cm,2 inferred from the photoelectric absorption seen in the ASCA observed X-ray spectrum of NGC 4388 Iwasawa et al., 1997. Demanding equality of ionized and total columns would imply Te 106 K. If the radio emission is indeed thermal bremsstrahlung, wemay calculate the predicted, intrinsic H ux from this gas using e.g., Ulvestad, Wilson & Sramek, 1981: e FH erg cm,2 s,1 = 0:62 T0:5 K g,1; Z; T H S mJy; e e where S is the radio ux density, H = 3:01 10,14T=104,0:85 and g is the Gaunt factor. Taking S = 2.1 mJy at 5 GHz, we predict FH 4.0 10,13 erg cm,2 s,1 for Te 104:5 K Figure 4d. However the observed Oiii ux of 1.86 10,13 erg cm,2 s,1 , through 1005300 Corbin et al., 1988 or 200300 5 apertures Colina, 1992, and the Oiii : : to H ratio of 11.2 Phillips & Malin, 1982, result in a mean observed H ux of only 1.7 10,14 erg cm,2 s,1 which we treat as an upper limit to the H ux from the MERLIN source due to the large optical apertures compared to the radio sizes, a factor of 24 lower than our predicted value if Te = 104:5 K. If our assumption of a thermal origin for the radio emission is correct, the di erence between observed and predicted FH suggests AV 3.0 magnitudes of extinction towards the nucleus for Te =104:5 K. Using the H ux measured by Dahari & De Robertis 1988 of 5.2 10,14 erg cm,2 s,1 results in a lower extinction of AV 1.9 mags. In addition, no broad Pa , Br or Br lines are detected towards the nucleus Blanco et al., 1990; Ruiz et al., 1994; Veilleux et al, 1997,


20 while a broad, o -nuclear H line Shields & Filippenko, 1988 is detected 400 from the nucleus, consistent with scattered emission from the broad line region. A temperature in excess of 108 K would be required to explain the low observed H ux in the absence of extinction and therefore, given the implication from optical studies that NGC 4388 harbors an obscured Seyfert type 1 nucleus, the moderately large value of extinction suggested by the radio data does not seem unreasonable. Alternatively, not all of the radio emission might be thermal, and the base of a radio jet could provide a non-thermal contribution. However, any signi cant non-thermal contribution would be di cult to reconcile with the at radio spectrum, given the low observed brightness temperature. We therefore favor the thermal emission model for the nuclear radio emission in NGC 4388.

6. Conclusions
We have used the VLBA at 8.4 GHz to study ve Seyfert nuclei that contain at spectrum radio sources, in order to determine whether the at-spectrum nuclear radio emission, detected in VLA studies, represents thermal emission from the accretion disk obscuring torus or synchrotron self-absorbed emission from a compact radio core source. Four of the ve sources were detected T0109,383, NGC 2110, NGC 5252, Mrk 926 and show compact, unresolved cores with brightness temperatures, TB 108 K, total luminosities at 8.4 GHz of 1021 WHz,1 and sizes, on average, less than 1 pc. We conclude that the sub-pc scale radio emission in these sources is non-thermal and self absorbed and, hence, dominated by the central engine. In addition to the core emission, NGC 2110 shows extended emission which may represent the inner parts of the radio jets. However, we nd no evidence of thermal disk-like emission, extended perpendicular to the collimation axis, in any of these sources to a 3- limit of 106 K. The putative nucleus of NGC 4388 is not detected with the VLBA but is detected with MERLIN at 5 GHz. The observationally inferred brightness temperature of 2.4 104 K TB 2.2 106 K the lower limit at 5 GHz, the upper limit at 8.4 GHz is too low for synchrotron self absorption to be important. Instead we have proposed a model in which the emission is optically thin, free-free thermal bremsstrahlung emission from a gas with an electron temperature of Te 104:5 K and density ne 1.6 104 f ,0:5 cm,3 where f is the volume lling factor. The larger inferred values of Te 106:8 K and ne 1.8 106 f ,0:5 cm,3 for the smaller source size set by the VLBA limit, are similar


21 to the values of 106:8 K and 106:8 cm,3 found for thermal emission from the torus gas in NGC 1068 Gallimore et al., 1997, thus implying that we may be seeing the same phenomenon in NGC 4388. Sensitive VLBA observations of NGC 4388 at 1.4 GHz or 2.3 GHz are required to spatially resolve the emitting region and determine its exact physical properties. We thank Pierre Ferruit, Neil Nagar and Dave Shone for useful discussions, and Peter Thomasson for help with the MERLIN data. The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. MERLIN is a U.K. national facility operated by the University of Manchester on behalf of PPARC. This research has made use of: the United States Naval Observatory USNO Radio Reference Frame Image Database RRFID, NASA's Astrophysics Data System Abstract Service ADS and the NASA IPAC Extragalactic Database NED, which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. This research was supported by the National Science Foundation under grant AST 9527289 and NASA under grant NAG 81027.


22

REFERENCES
Acosta-Pulido, J.A., Vila-Vilar B., P o, erez-Fournon, I., Wilson, A.S. & Tsvetanov, Z.I., 1996, ApJ, 464, 177 Alonso-Herrero, A., Ward, M.J. & Kotilainen, J., 1996, Vistas in Astronomy, 40, 221 Antonucci, R.R.J., 1993, ARA&A, 31, 473 Antonucci, R.R.J. & Miller, J.S., 1985, ApJ, 297, 621 Antonucci, R..R.J. & Ulvestad, J.S., 1985, ApJ, 294, 158 Baars, J.W.M., Genzel, R., Pauliny-Toth, I.I.K. & Witzel, A., 1977, A&A, 61, 99 Balsara, D.S. & Krolik, J.H., 1993, ApJ, 401, 109 Baker, A.J. & Scoville, N.Z., 1998, BAAS, 192, 3605 Bassani, L., Dadina, M., Maiolino, R., Salvati, M., Risaliti, G., Della Ceca, R., Matt, G. & Zamorani, G., 1999, ApJS, 121, 473 Beasley, A.J. & Conway, J.E., 1995, in ASP Conf. Series 82, Very Long Baseline Interferometry and the VLBA, eds. J.A. Zensus, P.J. Diamond & P.J. Napier San Francisco: ASP, 327 Bicknell, G.V., Dopita, M.A., Tsvetanov, Z.I. & Sutherland, R.S., 1998, ApJ, 495, 680 Blanco, P. R., Ward, M. J., &Wright, G. S. 1990, MNRAS, 242, 4P Blandford, R.D. & Begelman, M.C, 1998, MNRAS, 303, L1 Blundell, K.M. & Beasley, A.J., 1998, MNRAS, 299, 165 Boisson, C. & Durret, F., 1986, A&A, 168, 32 Bradt, H.V., Burke, B.F., Canizares, C.R., Green eld, P.E., Kelley, R.L., McClintock, J.E., Koski, A.T. & Van Paradijs, J., 1978, ApJ, 226, L111 Briggs, D.S., 1995, Ph.D. Thesis, N.M. Institute of Mining &Technology Browne, I.W.A., Wilkinson, P.N., Patnaik, A.R. &Wrobel, J.M., 1998, MNRAS, 293, 257 Carral, P., Turner, J. L., & Ho, P. T. P. 1990, ApJ, 362, 434 Carilli, C.L., Wrobel, J.M. & Ulvestad, J.S., 1998, AJ, 115, 928


23 Carollo, C.M., Franx, M., Illingworth, G.D. & Forbes, D.A., 1997, ApJ, 481, 710 Clements, E.D., 1981, MNRAS, 197, 829 Clements, E.D., 1983, MNRAS 204, 811 Colina, L., 1992, ApJ, 386, 59 Condon, J.J., Huang, Z.-P., Yin, Q.F. & Thuan, T.X., 1991, ApJ, 378, 65 Condon, J.J., 1992, in Testing the AGN Paradigm", AIP Conference Proceedings 254, eds. Holt, S.S., Ne , S.G., Urry, C.M., 629 Corbin, M.R., Baldwin, J.A. & Wilson, A.S., 1988, ApJ, 334, 584 Crusius-Waetzel, A.R. & Lesch, H., 1998, A&A, 338, 399 Dahari, O. & De Robertis M.M., 1988, ApJS, 67, 249 de Bruyn, A.G. & Wilson, A.S., 1978, A&A, 64, 433 de Vaucouleurs, G., de Vaucouleurs, A., Corwin, H.G., Jr., Buta, R.J., Paturel, G. & Fouqu P., 1991, Third Reference Catalogue of Bright Galaxies Springer, Berlin e, Downes, D. & Solomon, P.M., 1998, ApJ, 507, 615 Falcke, H., Ho, L.C., Ulvestad, J.S., Wilson, A.S. & Nagar, N.M., 1999, in Proceedings of the International Symposium on Astrophysics Research And Science Education at The Vatican Observatory", ed. C. Impey, in press Falcke, H., Wilson, A.S. & Simpson, C., 1998, ApJ, 502, 199 Fey, A. L. & Charlot, P. 1997, ApJS, 111, 95 Fernandes, R.C., Jr. & Terlevich, R., 1995, MNRAS, 272, 423 Ford, H.C., et al., 1994, ApJ, 435, L27 Fosbury, R.A.E. & Sansom, A.E., 1983, MNRAS, 204, 1231 Gallimore, J.F., Baum, S.A. & O'Dea, C.P., 1997, Nature, 388, 852 George, I.M., Turner, T.J., Netzer, H, Nandra, K., Mushotzky, R.F. & Yaqoob, T., 1998, ApJS, 114, 73


24 Gonz alez-Delgado, R.M., Heckman, T., Leitherer, C., Meurer, G., Krolik, J., Wilson, A.S., Kinney, A. & Koratkar, A., 1998, ApJ, 505, 174 Greisen E.W. & Murphy, P.P., 1998, The AIPS Cookbook, chapter 9, online at http: www.cv.nrao.edu aips cook.html Greenhill, L.J., Jiang, D.R., Moran, J.M., Reid, M.J., Lo, K.Y. & Claussen, M.J., 1995, ApJ, 440, 619 Haehnelt, M.G. & Rees, M.J., 1993, MNRAS, 263, 168 Heckman, T.M., Gonz alez-Delgado, R., Leitherer, C., Meurer, G.R., Krolik, J., Wilson, A.S., Koratkar, A. & Kinney, A., 1997, ApJ, 482, 114 Herrnstein, J.R., Greenhill, L.J. & Moran, J.M., 1996, ApJ, 468, L17 Hough, D.H. et al., 1999, ApJ, 511, 84 Hummel, E. & Saikia, D.J., 1991, A&A, 249, 43 Iwasawa, K., Fabian, A.C., Ueno, S., Awaki, H., Fukazawa, Y., Matsushita, K. & Makishima, K., 1997, MNRAS, 285, 683 Ja e, W., Ford, H.C., Ferrarese, L., van den Bosch, F & O'Connell, R.W., 1993, Nature, 364, 213 Kato, S., Fukue, J. & Mineshige, S., 1998, Black-hole accretion disks", eds. S. Kato, J. Fukue, and S. Mineshige; Publisher: Kyoto, Japan: Kyoto University Press, 1998 Kedziora-Chudczer, L., Jauncey, D.L., Wieringa, M.H., Walker, M.A., Nicolson, G.D., Reynolds, J.E., Tzioumis, A.K., 1997, ApJ 490, L9 Kellermann, K.I. & Pauliny-Toth, I.I.K., 1969, ApJ, 155, L71 Kellermann, K.I. & Pauliny-Toth, I.I.K., 1981, ARA&A, 19, 373 Kellermann, K.I., Vermeulen, R.C., Zensus, J.A. & Cohen, M.H, 1998, AJ, 115, 1295 Koski, A.T., 1978, ApJ, 223, 56 Krolik, J.H. & Begelman, M.C., 1988, ApJ, 329, 702 Krolik, J.H. & Lepp, S., 1989, 347, 179


25 Kukula, M.J., Ghosh, T., Pedlar, A., Schilizzi, R.T., Miley, G.K., De Bruyn, A.G. & Saikia, D.J., 1993, MNRAS, 264, 893 Kukula, M.J., Pedlar, A., Baum, S.A. & O'Dea, C.P., 1995, MNRAS, 276, 1262 Kukula, M.J., Ghosh, T., Pedlar, A. &Schilizzi, R.T, 1999, ApJ, in press Lahteenmaki, A., Valtao ja, E. & Wiik, K., 1999, ApJ, 511, 112 Ma, C. et al., 1998, AJ, 116, 516 Malaguti, G. et al., 1999, A&A, 342, L41 Maloney, P.R., Begelman, M.C. & Pringle, J.E., 1996, ApJ, 472, 582 Matt, G. et al., 1997, A&A, 325, L13 Miyoshi, M., Moran., J., Herrnstein, J., Greenhill, L., Nakai, N., Diamond, P. & Inoue, M., 1995, Nature, 373,127 Morganti, R., Tsvetanov, Z.I., Gallimore, J. & Allen, M.G., 1999, A&AS, in press Morse, J.A., Cecil, G., Wilson, A.S. & Tsvetanov, Z.I., 1998, ApJ, 505, 159 Mundell, C.G., Pedlar, A., Baum, S.A., O'Dea, C.P., Gallimore, J.F. & Brinks, E., 1995, MNRAS, 272, 355 Murayama, T., Taniguchi, Y., Iwasawa, K., 1998, AJ, 115, 460 Nagar, N.M., Wilson, A.S., Mulchaey, J.S. & Gallimore, J.F., 1999, ApJS, 120, 209 Napier, P.J., Bagri, D.S., Clark, B.G., Rogers, A.E.E., Romney, J.D., Thompson, A.R. & Walker, R.C., 1994, Proc IEEE, 82, 658 Narayan, R. & Yi, I., 1994, ApJ, 428, L13 Pearson, T.J. et al., 1998, in ASP Conference Series 144, Radio Emission from Galactic and Extragalactic Compact Sources, eds. J.A. Zensus, G.B. Taylor, & J.M. Wrobel San Francisco: ASP, 17 Peck, A.B. & Beasley, A.J., 1998, in ASP Conf. Series 144, Radio Emission from Galactic and Extragalactic Compact Sources, eds. J.A. Zensus, G.B. Taylor, & J.M. Wrobel San Francisco: ASP, 155 Peck, A.B. &Taylor, G.B., 1998, ApJ, 502, L23


26 Penston, M.V., Fosbury, R.A.E., Boksenberg, A., Ward, M.J. & Wilson, A.S., 1984, MNRAS, 208, 347 Peterson, B.M., 1997, An Introduction to Active Galactic Nuclei, Cambridge: Cambridge University Press Phillips, M.M. & Malin, D.F., 1982, MNRAS, 199, 905 Pier, E.A. & Krolik, J.H., 1992, ApJ, 399, L23 Pier, E.A. &Voit, G.M., 1995 ApJ, 450, 628 Pogge, R.W., 1988, ApJ, 332, 702 Prieto, M.A. & Freudling, W., 1993, ApJ, 418, 668 Prieto, M.A. & Freudling, W., 1996, MNRAS, 279, 63 Pringle, J.E., 1996, MNRAS, 281, 357 Readhead, A.C.S., 1994, ApJ, 426, 51 Richstone, D. et al., 1998, Nature, 395, 14 Roy, A.L., Colbert, E.J.M., Wilson, A. S., Ulvestad, J.S., 1998, ApJ, 504, 147 Roy, A.L., Norris, R.P., Kesteven, M.J., Troup, E.R. & Reynolds, J.E., 1994, ApJ, 432, 496 Ruiz, M., Rieke, G. H., & Schmidt, G. D. 1994, ApJ, 423, 608 Sadler, E.M., Slee, O.B., Reynolds, J.E. & Roy, A.L., 1995, MNRAS, 276, 1373 Shields, J.C. & Filippenko, A.V., 1988, ApJ, 332, L55 Silk, J. & Rees, M.J., 1998, A&A, 331, L1 Smith, M.G., 1975, ApJ, 202, 591 Stone, J.L., Jr., Wilson, A.S. & Ward, M.J., 1988, ApJ, 330, 105 Tadhunter, C. & Tsvetanov, Z.I., 1989, Nature, 341, 422 Taylor, G.B., O'Dea, C.P., Peck, A.B. & Koekemoer, A.M., 1999, ApJ, 512, L27 Tran, H.D., 1995, ApJ, 440, 578 Turner, T.J., George, I.M., Nandra, K. & Mushotzky, R.F., 1997, ApJS, 113, 23


27 Ulvestad, J. S., Gurvits, L. I., & Lin eld, R. P. 1997, in High Sensitivity Radio Astronomy, ed N. Jackson & R. Davis Cambridge: Cambridge University Press, 252 Ulvestad, J. S., & Lin eld, R. P. 1998, in ASP Conf. Series 144, Radio Emission from Galactic and Extragalactic Compact Sources, ed. J. A. Zensus, G. B. Taylor, & J. M. Wrobel San Francisco: ASP, 397 Ulvestad, J.S., Roy, A.L., Colbert, E.J.M. & Wilson, A.S., 1998, ApJ, 496, 196 Ulvestad, J.S., Wilson, A.S. & Sramek, 1981, ApJ, 247, 419 Ulvestad, J.S. & Wilson, A.S., 1983, ApJ 264, L7 Ulvestad, J.S. & Wilson, A.S., 1984a, ApJ, 278, 544 Ulvestad, J.S. & Wilson, A.S., 1984b, ApJ, 285, 439 Ulvestad, J.S. & Wilson, A.S., 1989, ApJ, 343, 659 Ulvestad, J.S., Wrobel, J.M., Roy, A.L., Wilson, A.S., Falcke, H. & Krichbaum, T.P., 1999, ApJ, 517, L81 Ulvestad, J.S., Wrobel, J.M. & Carilli, C.L., 1999, ApJ, 516, 134 van der Marel, R.P., 1999, AJ, 117, 744 van Gorkom, J., Rupen, M., Knapp, G., Gunn, J, Neugebauer, G. & Matthews, K., 1986, IAUC, 4248, 1 van Moorsel, G., Kemball, A., & Greisen, E., 1996, in ASP Conf. Series 101, Astronomical Data Analysis Software and Systems V, eds., G.H. Jacoby & J. Barnes San Francisco: ASP, 37 Veilleux, S., 1991, ApJ, 369, 331 Veilleux, S., Bland-Hawthorn, J., Cecil, G., Tully, R.B. & Miller, S.T, 1999, ApJ, 520, Veilleux, S., Goodrich, R.W. & Hill, G.J., 1997, ApJ, 477, 631 Walker, R.C., 1995, in ASP Conf. Series 82, Very Long Baseline Interferometry and the VLBA, eds. J.A. Zensus, P.J. Diamond & P.J. Napier San Francisco: ASP, 247 Ward, M.J., Wilson, A.S., Penston, M.V., Elvis, M., Maccacaro, T. & Tritton, K.P., 1978, ApJ, 223, 788


28 Weaver, K.A., Mushotzky, R.F., Serlemitsos, P.J., Wilson, A.S., Elvis, M. & Briel, U., 1995a, ApJ, 442, 597 Weaver, K.A., Nousek, J., Yaqoob, T., Hayashida, K. & Murakami, S., 1995b, ApJ, 451, 147 Weiler, K.W., Panagia, N. & Sramek, R.A., 1990, ApJ, 364, 611 Wilson, A.S., Baldwin, J.A. & Ulvestad, J.S., 1985, ApJ, 291, 627 Wilson, A.S. & Meurs, E.J.A., 1982, A&AS, 50 217 Wilson, A.S. &Raymond, J.C., 1999, ApJ, 513, 115 Wilson, A.S. et al., 1998, ApJ, 505, 587 Wilson, A.S. & Tsvetanov, Z.I., 1994, AJ, 107, 1227 Wilson, A.S. & Ulvestad, J.S., 1983, ApJ, 275, 8 Wrobel, J.M., Patnaik, A.R., Browne, I.W.A & Wilkinson, P.N., 1998, BAAS, 30, 1308

A This preprintwas prepared with the AAS L TEX macros v4.0.


29 Fig. 1.| Radio continuum images of the sources detected at 8.4 GHz with the VLBA T0109,383, NGC 2110, NGC 5252 and Mrk 926. The 8.4-GHz VLBA images are shown in the right hand panels, with the restoring FWHM beamsize shown as an ellipse in the bottom left corner of each image and linear scale marked in the bottom right corner. A VLA image of the larger scale structure of each source is shown in the corresponding left hand panel from Nagar et al. 1999 for NGC 2110 and NGC 5252, and from unpublished observations by J.A. Braatz, L.L. Dressel & A.S. Wilson for T0109,383 and Mrk 926. Contour levels for the VLBA images and unpublished VLA images are given in Table 5. Fig. 2.| MERLIN 5-GHz radio continuum image of NGC 4388. The nucleus is marked M1 and the secondary weak feature possibly a component in the radio jet is marked M2. The beamsize is indicated by an ellipse in the lower left corner. The contour levels, in multiples of 3r.m.s., are 2, 1, 1, 2, 3, 4 0.3 mJy beam,1 or TB = 2, 1, 1, 2, 3, 4 5.9103 K beamsize 91.0 39.5 mas. Fig. 3.| a VLA 8.4-GHz radio continuum image full resolution image produced from data presented by Falcke et al., 1998 of the 1009 central double source in NGC 4388 each : component is marked V1 and V2, with V1 taken to be the core; b Super-resolved VLA image, from the same u; v data as used for a, revealing collimated ejection from the core and bending of the jet at V2. The crosses indicate the two MERLIN components M1 and M2, where M1 is coincident with the core and M2 coincides with a protrusion in the 8.4-GHz VLA contours in the direction of the jet. The contour levels for both images are 1, 1, 2, 4, 8, 16, 32 90 Jy beam,1 and the beamsizes shown in the lower left corner of each image are a 000 70 00022 and b 00022 00022. : : : : Fig. 4.| Range of possible values of a electron density, ne, b free-free opacity at 4.993 GHz, , c column density of ionized gas, Ne, and d predicted H ux, assuming the radio emission from NGC 4388, at 4.993 GHz, is free-free thermal bremsstrahlung emission from an ionized gas with an electron temperature, Te . Upper and lower limits to the plotted quantities are derived from the lower and upper limits to the source size Section 4.3.1; we assumed the maximum source volume MERLIN limit corresponds to an ellipsoid with semi-axes 3.6 1.5 1.5 pc, and the minimum volume VLBA limit corresponds to an edge-on disk with diameter, 0.4 pc, and height, 0.2 pc. The permitted values for an optically thin plasma as indicated by the at radio spectrum are indicated by the shading.


30 Table 1: Observing parameters for the 8.4-GHz VLBA observations of Seyfert galaxies and calibrators. Measured positions, derived from Gaussian tting, for the check sources are also given. The surveys from which the positions of the phase calibrators were selected are listed below as footnotes, along with the corresponding positional accuracy.
Target Seyfert Date of observations Phase calibrator Position used J2000 Source + phase cali-brator cycle time `Check' calibrator Position used J2000 Position measured Times on `check' cal. Online fringe- nder
a;b

T0109,383 07 Aug 1997 J0106,4034 01h06m45.1080s 40 340 1900 960 :
a

NGC 2110 9 Aug 1997 J0541,0541 05h41m38.0834s 05 410 4900 428 :
b

NGC 4388 12 Jun 1998 J1207+1211 12h07m12.625s 12 110 4500 89 :
c

NGC 5252 14 Jun 1998 J1320+0140 13h20m26.7938s 01 400 3600 786 :
d

Mkn 926 21 Jul 1997 J2255,0844 22h55m04.2398s 08 440 0400 022 :
a

2+ 1min
J0044,3530 00h44m41.229s 35 300 4100 63 : | 3 1 min 3C454.3
c

4+1 min
J0607,0834 06h07m59.699s 08 340 4900 98 : 06h07m59.6975s 08 340 4900 994 : 4 1 min DA193

3 + 1 min

3+ 1min

4+ 1min
J2246,1206 22h46m18.232s 12 060 5100 28 : 22h46m18.2313s 12 060 5100 262 : 4 1 min 3C454.3

J1214+0829 J1359+0159 12h14m59.914s 13h59m27.1478s 08 290 2200 53 01 590 5400 543 : : 12h14m59.913s 13h59m27.1512s 08 290 2200 56 : 01 590 5400 531 : 12 1 min 13 1 min 3C273
d

3C273

Total time on source 2.5 hrs 3.2 hrs 6.0 hrs 5.8 hrs 3.3 hrs Ma et al., 1998 2.5, 0.4 mas Browne et al., 1998;Wrobel et al., 1998 14 mas Peck & Beasley, 1998 5 mas

Table 2: Results of VLBA observations. Galaxy names, positions, peak and integrated uxes from single component Gaussian tting, and beamsizes  are listed columns 1 5. Lower limits to the brightness temperatures TB K, column 6 are derived from the given peak ux, assuming an upper limit to the source size of half of the beamsize; for NGC 2110, the peak ux of the unresolved component given in Table 3 is used instead. The luminosity at 8.4 GHz L, column 7 is calculated using the total ux density and assuming isotropic emission at the distance of the galaxy D, column 8; obtained assuming H0 = 75 km s,1 Mpc,1 and q0 = 0.5.
Seyfert T0109,383 NGC 2110 NGC 5252 Mkn 926
a a

See Table 3

Fitted position 8.4-GHz 8.4-GHz  J2000.0R.A. Peak Flux Total Flux mas J2000.0Dec. mJy beam,1 mJy pc 01 11 27.6413 9.9 0.6 11.0 0.7 1.96 0.63 38 05 00.477 0.45 0.14 05 52 11.3762 16.6 0.8 29.51.5 1.88 0.72 07 27 22.513 0.28 0.11 13 38 15.8698 7.9 0.4 9.1 0.5 2.01 0.95 +04 32 33.513 0.91 0.43 23 04 43.4776 4.6 0.2 5.0 0.3 2.82 1.02 08 41 08.629 2.64 0.95 for properties of multiple components present in NGC 2110.

T K 108 K
B

8.1 6.0 4.2 1.7

L D 1021 Mpc W Hz,1 2.7 46.6 3.1 8.8 2.1 30.4 92.4 191.4


31

Table 3: Results of Gaussian tting to the emission from NGC 2110; three components are detected, which are identi ed column 1 as an unresolved `core', north-south `jet' emission and a northern `knot'. The positions of each component column 5 relative to the core position column 2 are given with positive o sets north of the core. Peak and integrated uxes are given in columns 3 & 4, along with deconvolved sizes and position angles column 6, 7 for the marginally resolved components.
Fitted position J2000.0R.A. J2000.0Dec. Unresolved `core' 05 52 11.3762 07 27 22.513 N-S `jet' Component Northern `knot' 8.4 GHz Peak Flux mJy beam,1 8.0 0.5 8.70.5 5.0 0.3 8.4-GHz O set of tted Deconvolved size Total Flux component centroid, mas mJy mas relative to `core' 7.5 0.4 0 0 13.1 0.7 6.0 0.3 +0.21 +1.95 1.41 0.36 0.3 0.3 0.75 0.12 0.4 0.12 PA deg 0 8.5 0.8 137 90

Table 4: Upper table a shows the results of Gaussian tting to the 5-GHz core emission in NGC4388, imaged with MERLIN. Fitted position, peak brightness, integrated ux density and beam size are given. The lower table b shows the corresponding derived brightness temperature TB K calculated using the measured MERLIN peak brightness and beamsize and the predicted VLBA brightness temperatures at 8.4 GHz, assuming the spectral indices given, the VLBA beamsize given and that the source is unresolved with the VLBA. These predicted VLBA brightness temperatures exceed the observed value, allowing a lower limit to the source size to be determined Section 4.3.1.
Seyfert NGC4388

a Results of Gaussian tting to 5 GHz MERLIN image of NGC4388.
Fitted position B1950.0R.A. B1950.0Dec. 12 23 14.646 12 56 20.20
B

5-GHz Peak Flux 5-GHz Integrated Flux mJy beam mJy 1.20.1 VLBA Beamsize 2.521.46 mas 0.230.14 pc 2.10.3 Assumed from =5 to 8.4 GHz 0 0.26

MERLIN Beamsize mas 91.0 39.5

b Derived MERLIN 5-GHz and predicted VLBA 8.4-GHz brightness temperatures.
Derived T 104 K at 5 GHz 2.4
Predicted

T 106 K at 8.4 GHz 8.3 7.3
B


32

Table 5: Upper table lists contour levels column 2, plotted as multiples of 3 r.m.s in the image column 3, and beamsizes column 4 for the VLBA images in Figure 1. Brightnesses mJy beam,1 are converted to TB 106 K using the corresponding VLBA beamsizes. Lower table lists contour levels column 2 for the unpublished VLA images of T0109,383 and Mrk 926 shown in Figure 1, plotted as multiples of 3 r.m.s in the image column 3. The VLA beamsizes are given in column 4.
Seyfert T0109,383 NGC 2110 NGC 5252 Mkn 926 T0109,383 Mkn 926 8.4-GHz VLBA Contour Levels 3 r.m.s. Beamsize multiples of 3 r.m.s. mJy beam,1 106 K mas 1, 1, 2, 4, 8, 16 0.45 9.2 1.96 0.63 1, 1, 2, 4, 8, 16, 32 0.28 5.3 1.88 0.72 1, 1, 2, 4, 8, 16 0.30 4.0 2.01 0.95 1, 1, 2, 4, 8, 16 0.14 1.2 2.82 1.02 8.4-GHz VLA Contour Levels 3 r.m.s. Beamsize multiples of 3 r.m.s. mJy beam,1 00 1, 1, 2, 4, 8, 16, 32, 64 0.11 0.41 0.17 1, 1, 2, 4, 8, 16, 32 0.12 0.27 0.21


Figure 1

T0109-383
0 5 10

NGC 2110
0 2 4 6 8
0 20 40 60

0

5

10

15

-38 20 54.0 54.5

-38 05 00.468

-07 27 22.504

VLA - 8.4 GHz
00.470

VLBA - 8.4 GHz
-07 28 00

VLA - 8.4 GHz
22.506

VLBA - 8.4 GHz
DECLINATION (J2000)
22.508 22.510 22.512 22.514 22.516 22.518

DECLINATION (J2000)

DECLINATION (B1950)

55.0 55.5 56.0 56.5 57.0 57.5 58.0 58.5 01 09 09.95 09.90 09.85 09.80 09.75 09.70 09.65 09.60 09.55 RIGHT ASCENSION (B1950)

DECLINATION (B1950)

00.472 00.474 00.476 00.478 00.480 00.482 00.484 00.486 01 11 27.642

01

02

03

33

4 mas 0.91 pc
27.641 RIGHT ASCENSION (J2000)

04

22.520 22.522

4 mas 0.59 pc
05 52 11.376 RIGHT ASCENSION (J2000)

05 49 46.55 46.50 46.45 46.40 46.35 46.30 46.25 46.20 RIGHT ASCENSION (B1950)

NGC 5252
0 5 10

Mkn 926
0 2 4 6 8
-08 57 19.0 0 2 4 6

0

1

2

3

4

VLA - 1.4 GHz
04 48 10
DECLINATION (B1950)

04 32 33.522

-08 41 08.620

VLBA - 8.4 GHz
33.520
DECLINATION (J2000)
DECLINATION (B1950)

VLA - 8.4 GHz
08.622
DECLINATION (J2000)
19.5

VLBA - 8.4 GHz

33.518 33.516 33.514 33.512 33.510 33.508 33.506

08.624 08.626 08.628 08.630 08.632 08.634 08.636

05

20.0

00

47 55

20.5

50

21.0

45

33.504
13 35 45.0 44.5 44.0 43.5 RIGHT ASCENSION (B1950)

4 mas 1.8 pc
13 38 15.870 RIGHT ASCENSION (J2000)

21.5 23 02 07.25 07.20 07.15 07.10 RIGHT ASCENSION (B1950)

08.638 23 04 43.478

4 mas 3.8 pc
43.477 RIGHT ASCENSION (J2000)


34

Figure 2

12 56 20.4

20.3

DECLINATION (B1950)

20.2
M1

20.1

20.0
M2

19.9

12 23 14.66

14.65 14.64 RIGHT ASCENSION (B1950)

14.63


Figure 3

0

1

2

3

0

1

2

3

12 39 44.5

(a)
V1

12 39 44.5

(b)
M1 M2

44.0

44.0

43.5
DECLINATION (J2000)
DECLINATION (J2000)

43.5

35

43.0

43.0

42.5

42.5

42.0

V2

42.0

41.5

41.5

41.0

41.0

40.5 12 25 46.85 46.80 46.75 46.70 RIGHT ASCENSION (J2000) 46.65

40.5 12 25 46.85 46.80 46.75 46.70 RIGHT ASCENSION (J2000) 46.65


36

Figure 4