Kinetic power: the energy channel for low-luminosity active galactic nuclei
Fernandez-Ontiveros, J.A.; Lopez-Lopez, X.; and Prieto, A., 2023
Low Luminosity Active Galactic Nuclei (LLAGN), objects with accreion rate < 10e-3 in Eddington units, may represent the most common and lenthy phase of Black Hole activity. LLAGN are faint UV, optical, IR souces, hence their detection difficulty. This work presents a joint analysis of the multiwavelngth, subarsec resolution, spectral energy distribution and mid-IR nebular lines of nine LLAGN prototype (available here) with the aim at proving the nature and power of their ionizing continuum.
The spectral energy distribution of LLAGN is found to be consistent with pure, compact, jet emission over ten orders of magnitud in frequenzy. THis is a self-absorbed synchrotron emission from radio to the UV and the associated synchrotron self-Compton component from the UV to the X-rays. It shows This two particular characteristics when compared with typical jet spectra in radio galaxies: (i) a very steep spectral slope in the IR-to-UV range (-3.7 < alpha < -1.3; Fnu ~ nu^-alpha); (ii) a high turnover frequency (1.3 mm – 10 um) that separates the optically thick radio from the optically thin IR-to-UV emission. These attributes are explained if the synchrotron continuum is dominated by thermalised particles at the jet base or the corona with high temperatures, with a small fraction of the energy (20%) due to accelerated particles.
The nebular gas excitation is in agreement with photo-ionisation from the inverse Compton radiation (alpha_x ~ - 0.7), which dominates the nuclear continuum shortwards of 3000 Å. No sign of a standard hot accretion disc - the blue-bump - down to few parsecs from the centre is detected in the sample albeit a weak cold disc (<3000K) in the case of Sombrero galaxy nucleus, yet its contribution to the gas excitation is negligible.
The results suggest that the parsec-scale continuum emission of LLAGN is dominated at all wavelengths by an incipient jet, powered by a thermalised particle distribution with high energies on average. It follows that the dominant channel for energy release in these nuclei is kinetic via the jet, rather than radiative. This has implications in the context of galaxy evolution as LLAGN may represent a major but underestimated source of kinetic feedback in galaxies.
The flux distribution of the nine LLAGN in the sample are available at the CDS via anonymous ftp to cdsarc.cds.unistra.fr (130.79.128.5) or via https://cdsarc.cds.unistra.fr/viz-bin/cat/J/A+A/670/A22
Novel BH mass scaling relation inferred from coronal gas excitation
Prieto, A., Rodriguez-Ardila, A., Panda, S. & Marinello, M., 2022
Using bona-fide black hole (BH) mass estimates from reverberation mapping and the line ratio [Si VI] 1.963 micron / Brg-broad as tracer of the AGN ionizing continuum, a novel BH-mass scaling relation of the form log(MBH) = (6.40 +- 0.17) - (1.99 +- 0.37) x log ([SiVI]/Br-broad), dispersion 0.47 dex, over the BH mass interval, 10^6 - 10^8 MBH is found. Following on the geometrically thin accretion disc approximation and after surveying a basic parameter space for coronal lines production, we believe one of main drivers of the relation is the effective temperature of the disc, which is effectively sampled by the [Si VI] 1.963 micron coronal line for the range of BH masses considered. By means of CLOUDY photoionization models, the observed anticorrelation appears formally in line with the thin disc prediction Tdisc ~ MBH^(-1/4).
[Si VII] 1.96 um coronal line emission normalized to the broad component of Br-gamma shown in the X-axis, versus black hole mass derived from reverberation mapping, in the Y-axis. The black line is the best-linear fit to the data, the red-dashed and dotted lines show the one sigma and two sigma deviation respectively. The one sigma dispersion, 0.47 dex, compares with current one from the BH-Mass vs sigma relation in 49 bulges (Gultekin et al. 2006).
Dust in the central parsec of type 1 AGN: direct evidence for matter inflow to the centre
Prieto, A., Nadonly, K., Fernandez-Ontiveros, J.A. & Mezcua, M., 2021
A parsec-scale dusty torus is thought to be the cause of Active Galactic Nuclei (AGN) dichotomy in the 1/ 2 types, narrow / broad emission lines. In a previous work, on the basis of parsec-scale resolution infrared / optical dust maps it was found that dust filaments, few parsecs wide, several hundred parsecs long, were ubiquitous features crossing the centre of type 2 AGN, their optical thickness being sufficient to fully obscure the optical nucleus. This work presents the complementary view for type 1 and intermediate-type AGN. The same type of narrow, collimated, dust filaments are equally found at the centre of these AGN. The difference now resides in their location with respect to the nucleus, next to it but not crossing it, as it is the case in type 2, and their reduced optical thickness towards the centre, Av <= 1.5 mag, insufficient to obscure at UV nucleus wavelengths. It is concluded that large scale, hundred pc to kpc long, dust filaments and lanes, reminiscent of those seen in the Milky Way, are a common ingredient to the central parsec of galaxies. Their optical thickness changes along their structure, in type 2 reaching optical depths high enough to obscure the nucleus in full. Their location with respect to the nucleus and increasing gradient in optical depth towards the centre could naturally lead to the canonical type 1 / 2 AGN classification, making these filaments to play the role of the torus. Dust filaments and lanes show equivalent morphologies in molecular gas. Available gas kinematic indicates mass inflows at rates < 1Mo / yr
Central kpc of Tyoe 1 NGC 1566, North is up, East to the left.
Left: extinction map, dark color means dusty regions. The nucleus is the white bright pointat the centre. Centre: Colour composition, with the extinction map in red and ALMA CO(3-2) in blue. The nucleus is the pink-color source at the centre. Right: zoom of the central half kpc with the extinction map in red, and ALMA CO(3-2) velocity channels in the range 20 - 80 km / s, extracted from higher angular resolution map, beam 60 mas x 40 mas, shown on top. The low velocity channel in green, the high velocity one, in blue. This selectedvelocity channels isolate one of the nearest-to-the-centre dust filaments (marked with an arrow). Its kinematics unveils how it is progressively losing momentum as aproaching the centre.
From kpcs to the central parsec: feeding star formation and a black hole at the same time
Prieto et al. 2019
Prieto et al. 2005
2005 ESO Press Release
A panchromatic view of the star-forming ring and feeding process in the central kpc of the galaxy NGC 1097 is presented. The assembled IR to UV images at ~ 10 pc resolution allow us to characterize the population of circa 250 clusters in the ring and disentangle the network of filaments of dust and gas that enshroud and feed them.
The ring is a place of intermittent star bursts over the last 100 Myr. Four major episodes covering a proto-cluster phase of 11 mid-IR sources at the molecular clouds core, and two (three) previous bursts with a time separation of 20–30 Myr are identified.
The dust map of the inner few kpc resolves NGC 1097’s two major dust lanes in bundles of narrow, <25 pc width, filaments running along the galaxy’s bar. As they approach the ring, some circularize along it, others curve to the centre to produce a nuclear spiral. We believe these are kpc-scale dust-gas streamers feeding the ring and the black hole.
The total mass in clusters formed in the ring in the last 100 Myr is <10^7 Mo, i.e. less than 1% of the 10^9 Moof molecular gas in the ring; yet, at its current star formation rate, ~ 1.8 Mo/yr, an order of magnitude more in stellar mass should have been produced over that period. This means that the availability of gas in the ring is not the main star formation driver, and other factors, as the rate at which dense gas accumulates in teh ring are more fundamental.
Caption to the figure:
Left panel: RGB (VLT-NACO K-band-red, HST f814W-green and F336W-blue) image - 4 kpc accross - of NGC 1097 circumnuclear ring illustrating the ring of young star clusters, and the network of dust filaments entering the ring. Most dust filaments arrive to the ring via two major dust lanes, one at the North-East and one at the South-West of the ring (identified as the two dark, wide, long features in the image).
Right panel: high contrast image (VLT-NACO J-band after bulge subtraction) of the nuclear spiral of dust filaments traced up 10 pc distance from the BH (at center of the image). The filaments are part of the large scale dust lane that cross the galaxy from kpcs distance. These filaments feed the circumnuclear star forming ring and the black-hole.
In the figure above, 247 stellar clusters in the 1 kpc radius circumnuclear ring are identified with circles in the HST/F336W UV image. Some clusters are only seen from 0.8 um onwards becasue of extinction. North is up, East to the left, scale is in arcsecs. The parsecs-scale SED of some representative clusters in the ring are shown in three separate plots around the HST image. In each plot, the best stellar population fit to the SED is plot over the data points -with errors -in blue. All SEDs are sampled from 0.3 to 2.5 um. In colour, the up to three different bursts of star formation occurring along the ring are identified.
The central parsecs of M87: jet emission and an elusive accretion disc (Prieto et al. 2016)
We present the first simultaneous spectral energy distribution (SED) of M87 core at a scale of 0.4 arcsec ( ~ 32 pc) across the electromagnetic spectrum. Two separate, quiescent, and active states are sampled that are characterized by a similar featureless SED of power-law form, and that are thus remarkably different from that of a canonical active galactic nuclei or a radiatively inefficient accretion source. We show that the emission from a jet gives an excellent representation of the core of M87 core covering ten orders of magnitude in frequency for both the active and the quiescent phases. The inferred total jet power is, however, one to two orders of magnitude lower than the jet mechanical power reported in the literature. The maximum luminosity of a thin accretion disc allowed by the data yields an accretion rate of < 6 x 10^-5 Mo / yr, assuming 10 per cent efficiency. This power suffices to explain M87 radiative luminosity at the jet frame, it is however two to three order of magnitude below that required to account for the jet's kinetic power. The simplest explanation is variability, which requires the core power of M87 to have been two to three orders of magnitude higher in the last 200 yr. Alternatively, an extra source of power may derive from black hole spin. Based on the strict upper limit on the accretion rate, such spin power extraction requires an efficiency an order of magnitude higher than predicted from magnetohydrodynamic simulations, currently in the few hundred per cent range.
Caption to the figure: M87 SED at 0.4 arcsec resolution (32 pc) -black points. Jet–disc model result -red line- and physical components involved in the model are color -coded. Tabular form -ascii- of M87 SEDs in Prieto et al. 2016)
The nuclear dust lane of Circinus: collimation without a torus
Mezcua et al. 2016
In some AGN, nuclear dust lanes connected to kpc-scale dust structures provide all the extinction required to obscure the nucleus, challenging the role of the dusty torus proposed by the Unified Model. In this letter we show the pc-scale dust and ionized gas maps of Circinus constructed using sub-arcsec-accuracy registration of infrared VLT AO images with optical Hubble Space Telescope images.
We find that the collimation of the ionized gas does not require a torus but is caused by the distribution of dust lanes of the host galaxy on 10 pc scales. This finding questions the presumed torus morphology and its role at parsec scales, as one of its main attributes is to collimate the nuclear radiation, and is in line with interferometric observations which show that most of the pc-scale dust is in the polar direction.
We estimate that the nuclear dust lane in Circinus provides 1/3 of the extinction required to obscure the nucleus. This constitutes a conservative lower limit to the obscuration at the central parsecs, where the dust filaments might get optically thicker if they are the channels that transport material from 100 pc scales to the centre.
In the figure above, left panel: VLT-NACO-Ks / HST-F814W continuum light ratio -efectively a dust map - with contours of [O III], centre: same dust map as in previous with HST-Ha line emission in green and warm VLT-NACO-H2 molecular gas in blue; right: same dust map with VLT-NACO-[Si VII]2.48 um coronal emission in contours. The V-shape defined by the two dust filaments at the centre is evidenced, the ionised gas morphology - its apparent collimation - is determined by the location of the dust filaments.
The central parsec of AGN: Challenges to the torus
Prieto et al. 2014
Type 2 active galactic nuclei (AGN) are by definition nuclei in which the broad-line region and continuum light are hidden at optical/UV wavelengths by dust. Via accurate registration of infrared (IR) Very Large Telescope adaptive optics images with optical Hubble Space Telescope images we unambiguously identify the precise location of the nucleus of a sample of nearby, type 2 AGN. Dust extinction maps of the central few kpc of these galaxies are constructed from optical-IR colour images, which allow tracing the dust morphology at scales of few pc. In almost all cases, the IR nucleus is shifted by several tens of pc from the optical peak and its location is behind a dust filament, prompting to this being a major, if not the only, cause of the nucleus obscuration. These nuclear dust lanes have extinctions Av >~ 3 - 6 mag, sufficient to at least hide the low-luminosity AGN class, and in some cases are observed to connect with kpc-scale dust structures, suggesting that these are the nuclear fueling channels. A precise location of the ionized gas Hα and [Si VII] 2.48 μ coronal emission lines relative to those of the IR nucleus and dust is determined. The Hα peak emission is often shifted from the nucleus location and its sometimes conical morphology appears not to be caused by a nuclear - torus - collimation but to be strictly defined by the morphology of the nuclear dust lanes. Conversely, [Si VII] 2.48 micron emission, less subjected to dust extinction, reflects the truly, rather isotropic, distribution of the ionized gas. All together, the precise location of the dust, ionized gas and nucleus is found compelling enough to cast doubts on the universality of the pc-scale torus and supports its vanishing in low-luminosity AGN. Finally, we provide the most accurate position of the NGC 1068 nucleus, located at the south vertex of cloud B.
The figure above shows the results for NGC1386. The top panels show HST/F814W image with Ks-band continuum contours in white. Blue circles mark the position of the point-like sources used for image alignment, which are also shown in the two inner panels at the top-right corner. The inner 4"x4" region (blue square) is shown in detail in the middle and bottom panels. Bright regions indicate the location of the dust. The middle right panel shows one of the filaments crossing straight into the nucleus (marked with a cross). The bottom left panel shows Halpha in green contours on top of the dust map: it can be seen that the apparent collimated morphology of the ionized gas is defined by the dust filaments. In the bottom right, [SI VII] 2.48 um coronal emission shown in green contours on top of the dust map, shows the true isotropic morphology of the gas. North is up and East is to the left. From (Prieto et al. 2014).
In the figure above, top left: HST/F814W image with Ks-band continuum contours in white. Blue circles mark the position of the point-like sources used for image alignment. The blue square shows the inner 4"x4" in the middle and bottom panels. Top right is the Ks/F814W ratio or dust map. The bright arch at the centre is a dust lane crossing the nucleus. A zoom of this dust lane is shown in all other panels. The position of the nucleus and its error is marked with a cross in the middle panels. Bottom left shows the dust map with with Halpha in contours in green. The morphology of the gas - conical shape- is sharply defined by the large scale dust-lane. North is up and East is to the left. From (Prieto et al. 2014).
High ionization coronal lines: unique tracers of outflows in AGN
Muller-Sanchez et al. 2011
High ionization coronal lines, i.e. lines arising from ionization species with ionization potential, IP > 100 eV, are unique tracers of AGN activity. Very high spatial resolution near-infrared integral field spectroscopy of the coronal line region in nearby Seyfert galaxies reveals for the first time the kinematics of this gas on scales of less than 300 pc from the centre. The near-IR coronal lines [Si VI], [Al IX] and [Ca VIII], all spatially resolved in the integral-field observations, reveal a kinematic pattern of rotation plus radial flow, the latter being the dominant motion. Several observational facts indicate that this radial motion corresponds to outflows: radial acceleration of the gas at different projected distances from the center, gas blueshifts in the direction of the jet pointing to us, redshifts in the direction pointing away from us.
The spatially resolved kinematics of the strongest line, [SiVI] 1.96 um, has been modeled as a combination of an outflow bicone and a rotating disk. The rotating disk appears to be coplanar with those seen on the same scales in lower ionized gas and in the central molecular H2 gas. But the outflow velocities could reach values of hundreds, range in the 120 to 400 km/s, at a few hundred parsec from the center. The measured velocity dispersion in the outflowing gas is also large, ~ 150 km/s, and all together implies relatively large kinetic power in the highly ionised gas outflows, of up to a few percent of the bolometric luminosity of the AGN (Muller-Sanchez et al. 2011).
The figure above shows the spatial resolved structure, up to scales of 150 pc from the center, of high ionization coronal gas in three Seyfert type 2 galaxies from the sample in (Muller-Sanchez et al. 2011). These spatial maps are produced from K-band SINFONI data-cube obtained with adaptive optics at the VLT: the first column is the continuum emission at 2.2 um; second, third and fourth illustrate the morphology of [Si VI] 1.96 um, [Al IX] 2.04 um and [Ca VIII] 2.32 um emission. The central cross marks the 2 um emission peak.
Figure above shows the velocity tomography in [Si VI] 1.96 um in the Seyfert 1.5 galaxy NGC 3783 from Muller-Sanchez's et al. 2011 sample. The channel maps were obtained by integrating the flux within velocity bins of 80 km/s along the emission-line profile. The number in the upper part of each panel corresponds to the central velocity of the bin relative to systemic. The bottom right panel shows the integrated intensity image by summing the velocity slices. The acceleration in this galaxy is mostly observed in the north region.
Many stellar nurseries in the centre of an old galaxy: the elliptical NGC 1052 given birth again (Fernandez-Ontiveros et al. 2010)
High spatial resolution near-infrared images of the central 2 kpc of the near elliptical galaxy NGC 1052 reveal a total of 25 compact sources randomly distributed in the region. Fifteen of them exhibit Hα luminosities an order of magnitude above the estimate for an evolved population of extreme horizontal branch stars. Their Hα equivalent widths and optical-to-NIR spectral energy distributions are consistent with them being young stellar clusters aged <7 Myr. We consider this to be the first direct observation of spatially resolved star-forming regions in the central kiloparsec of an elliptical galaxy. The sizes of these regions are <~11 pc and their median reddening is E(B-V)~ 1 mag. According to previous works, NGC 1052 may have experienced a merger event about 1 Gyr ago. On the assumption that these clusters are spread with a similar density over the whole galaxy, the fraction of galaxy mass (5 x 10e-5) and rate of star formation (0.01 Mo yr-1) involved suggest the merger event as the possible cause for the star formation we see today (Fernandez-Ontiveros et al. 2010). NGC 1052 is at 18 Mpc distance.
In the figure above, left: SDSS colour composite (g- ,r- ,i- filters) image of NGC 1052; right: composite VLT/NACO Ks-band (red), HST/WFPC2 Halpha-continuum free (blue) of the central 2 kpc region. Star forming regions with Halpha-EW>~ 50A -15 in total- are marked with a circle. 10 more compact regions with Halpha-EW <50A are also in the same field. These ones are better seen in unsharp-masked VLT/NaCo Ks-band image (shown in Fernandez-Ontiveros et al. 2010)
Parsecs-scale spectral energy distributions of AGN (Prieto et al. 2010)
We are compiling spectral energy distributions (SEDs) with very high spatial resolution for some of the nearest, most well studied, active galactic nuclei (Prieto et al. 2010). These genuine AGN-core SEDs, mostly from Seyfert galaxies are characterized by two main features: an IR bump, with the maximum in the 2 -10 um range, and an increasing X-ray spectrum with frequency, in the 1 to ~200 keV region. These dominant features are common to Seyfert type 1 and 2 objects alike. In detail, type 1 AGNs are clearly distinguished from type 2s by their high spatial resolution SEDs: type 2 AGN exhibit a sharp drop shortward of 2 um, with the optical to UV region being fully absorbed; type 1s show instead a gentle 2 um drop ensued by a secondary, partially-absorbed optical to UV emission bump.
The figure above shows the average SED template of type 1 Seyfert nuclei in blue - it includes NGC 3783, NGC 1566 and NGC 7469 - and in red of type 2 Seyfert nuclei - Circinus, NGC 1068, NGC 5506 and NGC 7582. Prior averaging, each SED was set to its rest frame system and then normalized to the mean value of its nuFnu distribution. Tabular form of these templates can be retrieved in this link.
The location of warm dust in the nearest AGN (Reunanen et al. 2010)
In order to set an upper limit to the size of the putative torus, we have been imaging the nearest AGN with the VLT at 11 and 18 um, achieving at these wavelengths spatial resolutions of a few tens of parsec at the nucleus of these objects. For some of the objects, the obtained subarcsec images at 18 um are unique. Two major results were found:
1. The bulk of the mid-IR emission is concentrated on an unresolved central source within a size of less than 5 to 130 pc, depending on the object distance. Further resolved emission is detected in 70% of the sample in the form of circumnuclear star-forming rings or diffuse nuclear extended emission.
2. In all cases, the nuclear fluxes at 11 and 18 um represent minor contribution of that measured by IR satellites at the nearest energy bands of 12 and 25 um. This contribution ranges from 30% to less than 1%. Either a low surface brightness component extending over galactic scales or strong extra-nuclear IR sources – e.g. HII regions in spiral arms – have to be the main source of the IR satellites emission. In either case, the contribution of these components dwarfs that of the AGN at mid-IR wavelengths.
The figure above shows narrow band continuum images at 18 um of two of the galaxies analyzed in our nearest-AGN study. Left is the type 2 Seyfert galaxy NGC 7582, right is the type 1 NGC 7469, both with circumnuclear star forming regions. The very central warm dust in these nuclei is distributed in a region less than 30 pc- in NGC 7582, less than 100 pc- in NGC 7469, FWHM (Reunanen, Prieto & Siebenmorgen 2010).
The centre of the prototype starburst galaxy NGC 253 at parsecs scale (Fernandez-Ontiveros et al. 2009)
VLT adaptive optics images of one of the nearest starburst galaxy NGC 253 - 4 Mpc - allow us to resolve the central 150 pc of this galaxy in more than 37 IR-bright star forming regions, a factor of three larger that is seen in the optical with the HST. The achieved angular resolution in the IR, down to 200 mas and thus comparable to that of available 2 cm maps, permitted us a very accurate IR--radio registration, with eight of these sources being found to have a radio counterpart.
These nuclear regions are characterized by Halpha equivalent widths of about 80 A, relatively large extinction of Av ~7 mag, luminosities, Lbol of about 10e7 Lo, and average sizes of ~3 pc. The compiled high spatial resolution spectral energy distributions (SEDs) of these regions look very similar, all are characterised by a maximum at ~20 um and a moderate emission bump in the 1 - 2 um range. We find that these features can be reproduced by considering a dominant contribution of very young stellar objects (YSO) to the integrated light from these regions: emission from the hot dust surrounding these protostars naturally accounts for the observed IR features and in turn for the overall SED shape. The current best estimate for the age and mas of these these nuclear regions are 0.1 - 3 Myr and 10e5 Mo respectively.
Due to their similar SEDs, and the fact that most of these regions are spatially resolved, an average SED of the 37 circumnuclear clusters has been produced. This template, spanning the radio - IR - optical range, is so far the most genuine representation of an extragalactic circumnuclear star-forming region (available here in tabular form)
Following the accurate IR-radio registration of the central 150 pc region of NGC 253, no X-ray, optical or IR counterpart to the radio core position are found. For the last twenty years this radio core has been assumed to harbor an AGN but the lack of a counterpart seen at IR wavelengths call into question its supposed AGN nature. The analysis of its SED suggests that this source may instead represents a scaled up version of Sgr A* at the Galactic Centre (Fernandez-Ontiveros et al. 2009).
We determined the position of the stellar kinematic center, from VLT / SINFONI 2D spectroscopy using the K-band CO band heads as tracers, and found it compatible within the errors, with either the historically presumed radio core position, 0.7" off-, or with an hard X-ray source, 0.4"off- the stellar kinematic center. As both could be a manifestation of nuclear activity, we consider the two as possible galactic nucleus candidates (Muller-Sanchez et al. 2010).
VLA / A-array 2 cm- and VLT adaptive optics J-, Ks- and L- bands composite- image of the central 400 pc region of NGC 253. A total of 37 regions above 3 sigma of the local background are identified within the central 150 pc -square- in the IR, eight of them have a radio counterpart at 2 cm. Most of the regions are spatially resolved in the VLT image, with median size FWHM ~ 3.3 pc (0.17 arcsecs). Best spatial resolution was achieved in L-band, FWHM ~ 0.13 arcsecs. North up, East left.
Unveiling the active nucleus in obscured AGN and the size of the torus (Prieto et al. 2010)
VLT adaptive optics images in the infrared of the nearest AGN are revealing the real spatial location of the active nucleus. Dust obscuration has being leading to wrong identifications in the optical in many cases. The subarcsec resolutions achieved at near-IR wavelengths is critical for identifying with high accuracy the true nucleus location. We find that in many, well known Seyfert 2 galaxies, the nucleus remains hidden up to about 1 um. Yet, from 2 um onward, it becomes the most outstanding source in the galaxy.
The figure above shows examples of obscured nuclei shortward of 1 um: e.g. NGC 1068 and NGC 5506. These are color images, J – K, of the central kpc region of these galaxies, the nucleus location is the bright source at the centre. In both cases, this bright source is unresolved down to the achieved resolution at the VLT with adaptive optics: FWHM < 0.06 arcsec in K-band. This value sets already an upper limit to the size of any putative torus at 2 um, and in turn of the location of the warmer dust, to be less than 4 pc in NGC 1068, 18 pc in NGC 5506, in diameter (Prieto & Meisenheimer 2004; Prieto et al. 2010).