VLBI IMAGING OF THE DOUBLE PEAKED EMISSION LINE SEYFERT KISSR 1494

The observations were carried out with nine antennas of the VLBA at 1.55 and 4.98 GHz, on 2013 August 10 (Project ID: BK182); the Fort Davis antenna was not involved in the experiment. Fringes were not detected with the Pie Town antenna. Therefore, effectively only eight antennas of the VLBA were used for the observations. The data were acquired with an aggregate bit rate of 1024 Mbits s⁻¹ (two polarization, two baseband channels, bandwidth 128 MHz, and a two-bit sampling rate), in a phase-referencing mode. 1325+436, which is 253 away from the source and has an x, y positional uncertainty of 0.18, 0.28 mas, was used as the phase reference calibrator. 3C 345 and 1315+415 were used as the fringe-finder and phase-check calibrator, respectively. A nodding cycle of 5 minutes (2 minutes on the phase calibrator and 3 minutes on the target) was used for the observations, interspersed by two 5 minute scans on 3C 345 and two 3 minute scans on the phase-check calibrator (only a single scan of the latter at 5 GHz). The experiment lasted a total of three hours, with ≈1.5 hr at each frequency.

The data reduction was carried out using Astronomical Image Processing System (AIPS; version 31DEC12) following standard VLBA-specific tasks and procedures outlined in the AIPS cookbook.⁸ Los Alamos, which was a stable antenna in the middle of the configuration, was used as the reference antenna for the calibration. The amplitude calibration was carried out using the procedure VLBACALA, while the delay, rate, and phase calibration were carried out using the procedures VLBAPANG, VLBAMPCL and VLBAFRGP. The phase calibrator 1325+436 was iteratively imaged and self-calibrated on both phase and phase+amplitude using AIPS tasks IMAGR and CALIB. The images were then used as models to determine the amplitude and phase gains for the antennas. These gains were applied to the target and the final images were made using IMAGR. A round of data-flagging was carried out using the task IBLED on the source SPLIT file, prior to making the images. The radio component detected at 1.6 GHz was offset from the center of the image by 0076, 0120 in right ascension and declination. We ran the task UVFIX on the source SPLIT file to shift the source to the center before producing the final image. We used the elliptical Gaussian-fitting task JMFIT to obtain the component flux densities.

To fit the emission lines in the SDSS spectrum of KISSR 1494 and simultaneously obtain the stellar velocity dispersion from the absorption lines, we first separated the emission line spectrum from the underlying stellar continuum by the following steps. We corrected for reddening using SDSS tabulated E(B − V) values (Schlegel et al. 1998) and then used the pPXF (penalized Pixel-Fitting stellar kinematics extraction) code by Cappellari & Emsellem (2004) to obtain the best fit model for the underlying stellar population. We corrected the spectrum for the redshift of the galaxy as listed in SDSS dr10. The MILES single stellar population models were used as templates of underlying stellar population as they cover a large range of metallicity (M/H ∼−2.32 to +0.22) and age (63 Myr to 17 Gyr). The stellar population contribution to the best fit template is ∼82% and the remaining 18% accounts for the contribution from other components like, the power law component of AGN and Fe ii lines. The stellar velocity dispersion obtained for KISSR 1494 from the absorption lines in the underlying stellar continuum is 206.6 ± 8.2 km s⁻¹.

The extracted pure emission line spectrum shows signatures of double peaks in prominent emission lines such as [S ii], Hα, Hβ, Hγ, [O i], [O ii], and [O iii]. Our analysis used IDL programs that included the MPFIT function for fitting Gaussians to emission lines. We first modeled each [S ii] doublet line as two Gaussians having equal line widths in velocity space. The [S ii] model was used as a template to fit the narrow [N ii] doublet lines as well as Hα and Hβ narrow lines. This procedure has been described by Reines et al. (2013). The separation between the centroids of the [N ii] narrow components was held fixed and the flux of [N ii] λ6583 to [N ii] λ6548 was fixed at the theoretical value of 2.96. With the inclusion of a broad Hα component in addition to the two narrow components, the reduced χ² improved from 1.49 to 1.09. For the [O iii] doublet lines an additional third broad component improved the fit by ≈20% (χ² improved from 1.67 to 1.35). For the [O ii] doublet lines, however, the broad component could be a consequence of the two blended/unresolved components in one of the doublet lines. The central wavelength of the broad component of Hα is redshifted by 7.23 ± 0.65 Å while [O iii] is blueshifted by 2.64 ± 1.75 Å. The velocity widths of the broad Hα and [O iii] components are 739.4 ± 26.3 km s⁻¹ and 338.6 ± 53.0 km s⁻¹, respectively. The larger width of the Hα broad component compared to the [O iii] broad component could signify a combination of Doppler broadening effects due to the BH and an outflow, both in Hα and the nearby [N ii] doublet lines. The difference in the shifts of the central wavelengths of the broad components of [O iii] and Hα could also be attributable to the same effect. All the prominent lines along with their best fitting components are shown in Figures 2 and 3; their details are noted in Table 1.

A single radio component was detected at 1.6 GHz (Figure 5), but not at 5 GHz. The radio component detected at 1.6 GHz has a peak intensity of ∼360 μJy beam⁻¹, and an integrated flux density of ∼650 μJy in a purely naturally weighted image (made with a ROBUST parameter of +5 in AIPS task IMAGR). The rms noise reached at 1.6 GHz was ∼6.5 × 10⁻⁵ Jy beam⁻¹, making this a 10σ detection. At 5 GHz, where there was no source detection, a final rms noise of ∼3.5 × 10⁻⁵ Jy beam⁻¹ was reached. The deconvolved size of the radio component at 1.6 GHz was obtained from the purely naturally weighted image using the AIPS task JMFIT. This size is ∼7.5 × 5 mas, or ∼8 × 6 pc at the distance of the source.

Zoom In Zoom Out Reset image size

In order to robustly examine the structure of the detected radio component at 1.6 GHz, we created images with different weighting schemes. Figure 5 shows the 1.6 GHz contour image made with uniform weighting using a ROBUST parameter of 0 in IMAGR. The rms noise in this image was ∼7.0 × 10⁻⁵ Jy beam⁻¹. Figure 6 shows a superimposition of the two radio images made with pure natural weighting (with ROBUST = +5, in grayscale), and pure uniform weighting (with ROBUST = −5, in magenta-colored contours). The rms noise in the latter image was ∼1.3 × 10⁻⁴ Jy beam⁻¹. All images have been convolved with a circular beam of size 7 mas, which was the best intermediate value. The radio component is almost completely resolved out in the purely uniformly weighted image. The brightest emission that remains at the source position is at the ∼3σ level. While its integrity needs to be examined with more sensitive VLBI observations in the future, we have shown this image to highlight its lack of centrally concentrated emission.

Zoom In Zoom Out Reset image size

Assuming that three times the rms noise at 5 GHz (=3.5 × 10⁻⁵ Jy beam⁻¹) was an upper limit to the (undetected) source flux density, we derive the 1.6–5.0 GHz spectral index to be α = −1.5 ± 0.5, for a total source flux density of 650 μJy at 1.6 GHz. The rms noise at 1.6 and 5 GHz was used to estimate the spectral index error. This spectral index value is an underestimate because the 1.6 GHz data probes shorter (u, v) spacings than the 5 GHz data and could pick up flux that is missed at 5 GHz. Steep spectrum radio cores are not uncommon in Seyfert galaxies (e.g., Orienti & Prieto 2010; Kharb et al. 2010; Bontempi et al. 2012). The steep radio spectrum rules out thermal or free–free emission as the radiation mechanism and instead supports a non-thermal origin.

The brightness temperature (in Kelvin) of the radio component at 1.6 GHz was estimated using the relation,

$$\begin{eqnarray*} {T}_{B}&=&1.8\times 10^{9} (1+ z) \left(\frac{S_{\nu }}{1\,\rm mJy}\right) \left(\frac{\nu }{1\,\rm GHz}\right)^{-2} \left(\frac{\theta _{1}\theta _{2}}{\rm mas^{2}}\right)^{-1}, \end{eqnarray*}$$

where θ₁, θ₂ are the major, minor axes of the best-fitting elliptical Gaussian for a resolved radio component (Ulvestad et al. 2005). T_B of the radio component is ∼1.4 × 10⁷ K. This suggests that the radio emission is related to an AGN rather than a nuclear starburst. Starbursts typically exhibit brightness temperatures of <10⁵ K (Condon et al. 1991).

Assuming equipartition of energy between relativistic particles and the magnetic field (Burbidge 1959), we obtained the minimum pressure, and the particle energy (electrons and protons) at minimum pressure using the relations in O'Dea & Owen (1987),

$$\begin{eqnarray*} L_{{\rm rad}}&=&1.2\times 10^{27} D_L^2 S \nu ^{-\alpha } (1+z)^{-(1+\alpha)} \left(\nu _u^{1+\alpha }-\nu _l^{1+\alpha }\right) (1+\alpha)^{-1},\\ P_{{\rm min}}&=&(2\pi)^{-3/7} (7/12) [c_{12} L_{{\rm rad}} (1+k) (\phi V)^{-1}]^{4/7},\\ B_{{\rm min}}&=&[2 \pi (1+k) c_{12} L_{{\rm rad}} (\phi V)^{-1}]^{2/7},\\ E_{{\rm min}}&=&[\phi V (2\pi)^{-1}]^{3/7} [L_{{\rm rad}} (1+k) c_{12}]^{4/7}, \end{eqnarray*}$$

where L_rad is the radio luminosity in erg s⁻¹, D_L is the luminosity distance in Mpc, z is the source redshift, S is the total flux density in Jy, ν is frequency in Hz, ν_u and ν_l are the upper and lower frequency cutoffs in Hz, respectively, P_min is the minimum pressure in dynes cm⁻², k is the ratio of the relativistic proton to relativistic electron energy, V is the source volume, c₁₂ is a constant depending on the spectral index and frequency cutoffs, ϕ is the volume filling factor, B_min is the magnetic field at minimum pressure in G, and E_min is the particle energy (electrons and protons) at minimum pressure in ergs. The total radio luminosity was estimated assuming that the radio spectrum extends from (ν_l) 10 MHz to (ν_u) 100 GHz with a spectral index of α = −1. Furthermore, it was assumed that k = 1. Table 2 lists equipartition estimates for plasma filling factors of unity and 0.5 (e.g., Blustin & Fabian 2009). The total energy in particles and fields, E_tot, is estimated as E_tot = 1.25 × E_min, while the total energy density, U_tot, in erg cm⁻³ is =E_tot(ϕV)⁻¹.

The lifetime of electrons in the radio component undergoing both synchrotron radiative and inverse-Compton losses on cosmic microwave background (CMB) photons was estimated using the relation (van der Laan & Perola 1969),

$$\begin{equation*} t_{e}\simeq 2.6\times 10^{4} \frac{B^{1/2}}{B^{2}+B_{r}^{2}} \frac{1}{[(1+z)\nu _{b}]^{1/2}}\ {\rm yr,} \end{equation*}$$

where B was assumed to be the equipartition magnetic field B_min, and the break frequency ν_b was assumed to be 1.6 GHz. The magnetic field equivalent to the radiation, which was assumed to be predominantly CMB photons, B_r, was estimated using the relation, B_r ≃ 4 × 10⁻⁶(1 + z)² G. The B_min values of a few mG, log U_tot of ∼−5.0 and electron lifetimes of ∼1000 yr are similar to estimates derived for the parsec-scale radio jets in Seyfert galaxies (e.g., Kharb et al. 2010, 2014a; Orienti & Prieto 2010; Bontempi et al. 2012). The significance of this finding will be discussed in Section 4.4.

While the high brightness temperature supports an AGN origin for the radio emission in KISSR 1494, here we test the idea that the radio emission is star formation related. If we assume that all of the radio luminosity is attributable to star formation, a star formation rate (SFR) can be derived using the relation (Condon 1992),

$$\begin{equation*} \mathrm{SFR\,({M}_\odot \,yr^{-1})=L(W\, Hz^{-1})/5.3\times 10^{21}\, \nu (GHz)^\alpha }. \end{equation*}$$

We find that the upper limit to the SFR (for stellar masses ⩾ 5 M_☉; see Condon 1992) from the 1.4 GHz luminosity (=1.72 × 10²³ W Hz⁻¹) obtained in the kiloparsec-scale FIRST image is ≈43 M_☉ yr⁻¹ for a spectral index of α = −0.8.

The SFR can also be estimated from the narrow Hα line. Adding the flux of the two narrow components of Hα (=2.9 × 10⁻¹⁴ erg cm⁻² s⁻¹) and following the relation in Kennicutt (1998),

$$\begin{equation*} \mathrm{SFR\,({M}_\odot \,yr^{-1})=7.9\times 10^{-42}}\; L_{[\rm H\alpha ]}\, {\rm (erg \,s^{-1})}, \end{equation*}$$

we derive a SFR of ∼1.7 M_☉ yr⁻¹ for L_Hα = 2.16 × 10⁴¹ erg s⁻¹. The SFR in KISSR 1494 derived from its optical spectrum therefore appears to be fairly low, and cannot account for the observed radio emission on kiloparsec-scales, which require the SFR to be >40 M_☉ yr⁻¹. This finding supports the idea that the radio emission on kiloparsec-scales in KISSR 1494 is AGN-outflow related.

Since equal peaked emission lines could arise in a rotating disk around the SMBH (Smith et al. 2012), we use the suggestion of a disk to derive the dynamical mass of the SMBH in KISSR 1494. Using the disk radius r and the mean rotation speed 〈v〉, the mass of the SMBH can be derived as M_BH = (v²r/G). We use the difference between the double peaks in the Hα and Hβ lines (Δλ₁ and Δλ₂) to determine v, where 2v = (1/2)[cΔλ₁/6562.8) + (cΔλ₂/4861.36)], and the denominators are the respective rest wavelengths in air. For Δλ₁ = 6.15 Å and Δλ₂ = 4.74 Å, we obtain v = 142.93 km s⁻¹. As the compact radio source has a dimension of 7.5 × 5 mas, we adopt an arbitrary disk diameter of 7.5 mas (∼8 pc). This yields a BH mass of M_BH = 3.8 × 10⁷ M_☉ for KISSR 1494.

The SMBH mass can also be estimated from the width and the luminosity of the broad component of the Hα line (Figure 2, Table 1). This method assumes that the broad component originates in the BLR and is not due to outflows (Reines et al. 2013), and that the BLR clouds are in virial equilibrium with the SMBH (Kaspi et al. 2005). Using the relation for BH mass from Reines et al. (2013) for a scale factor () of unity:

$$\begin{equation*} \mathrm{log \left(\frac{{M}_{{\rm BH}}}{{M}_\odot }\right)=6.57+0.47 \log} \,(L_{\rm H\alpha })+2.06 \log \,\mathrm{(FWHM_{H\alpha })}, \end{equation*}$$

where Hα line luminosity is in units of 10⁴² erg s⁻¹ and FWHM in units of 10³ km s⁻¹, we obtain M_BH = 3.7 × 10⁶ M_☉, for L_Hα = 9.26 × 10⁴⁰ erg s⁻¹. However, a broad component is also observed in both the [O iii] λλ4959, 5007 lines, which could be attributable to outflows related either to the AGN or nuclear star-formation (e.g., Esquej et al. 2014). If the broad Hα line likewise had a significant outflow component, this BH mass estimate would be unreliable.

The bulge stellar velocity dispersion as derived by our line-fitting is σ_⋆ = 206.6 ± 8.2 km s⁻¹. Following the M_BH–σ_⋆ relation for late-type galaxies (McConnell & Ma 2013):

$$\begin{equation*} \mathrm{log\left(\frac{{M}_{{\rm BH}}}{{M}_\odot }\right)=8.07+5.06 \log \left(\frac{\sigma _\star }{200\; km \,s^{-1}}\right)}, \end{equation*}$$

we derive a BH mass, M_BH = 1.4 ± 1.0 × 10⁸ M_☉, which is ≈4 times larger than the dynamical mass limit derived above. It is evident that the M_BH–σ_⋆ relation has a lot of scatter when it comes to late-type galaxies.

All three estimates of BH mass are very crude. For the lack of a more reliable measurement, we have adopted the BH mass derived from the M_BH–σ_⋆ relation in the rest of the paper. In Table 3 we list all the BH mass estimates along with their caveats.

We use the combined flux of the narrow components of the [O iii] λ5008 line (=1.21 × 10⁴¹ erg s⁻¹) to estimate the bolometric luminosity (L_bol) using the relation from Heckman et al. (2004): L_bol/L_[O iii] ≈ 3500, where L_[O iii] is the monochromatic continuum luminosity at a 5000 Å  rest frame. We derive L_bol = 4.2 × 10⁴⁴ erg s⁻¹. For a BH mass of 1.4 × 10⁸ M_☉, the Eddington luminosity ($\mathrm{\equiv 1.25\times 10^{38} {M}_{{\rm BH}}/{M}_\odot }$) is ≈1.7 × 10⁴⁶ erg s⁻¹ and the Eddington rate (≡ L_bol/L_Edd)  is ∼ 0.02. The accretion rate in KISSR 1494 seems to be typical of Seyfert galaxies in the literature (e.g., Ho 2008). In short, the double peaked emission line Seyfert KISSR 1494 does not distinguish itself from other Seyfert galaxies in terms of its BH mass and accretion properties.

The presence of a single radio component per se does not rule out the binary BH scenario. One possibility could be that the two BHs have separations less than the size of the detected radio component, i.e., if they are separated by <8 pc. However in this case, the NLR clouds giving rise to the double peaks will have to be restricted to those spatial scales as well, making this an unattractive proposition. The NLR clouds typically exist on spatial scales of tens to hundreds of parsecs from the central engine (e.g., Crenshaw et al. 2000). In addition, it is unclear what media (equivalent to dusty tori around BLR clouds) could shield the spatially extended NLR clouds of the two BHs from mixing and losing their individual identities.

An alternate scenario could be that a second BH is indeed present, but has associated radio emission that is below the rms noise in the final image (∼6.5 × 10⁻⁵ Jy beam⁻¹ at 1.6 GHz). There is a well established correlation between the core radio luminosity at 5 GHz (L₅, W Hz⁻¹) and the BH mass for nearby early-type galaxies (Franceschini et al. 1998): $L_5\propto M_{{\rm BH}}^{2.73}$. What is more relevant to KISSR 1494 is that for a handful of late-type galaxies considered in this study, the BH mass exponent is still close to 2.7. Assuming the 1.6–5.0 GHz core spectral index to be −1, we estimate that the upper limit to the radio luminosity at 5 GHz from the second BH would be ∼1.5 × 10²⁰ W Hz⁻¹. From the L₅–M_BH relation, this implies that the mass of the second BH would be ≲ 2.5 × 10⁷ M_☉. At the upper limit, this BH mass lies close to the M_BH–σ_⋆ mass of the primary BH (that emits at radio frequencies) within error limits. It would be surprising though if a major merger took place in KISSR 1494 in its past and left no tell-tale signatures of a disturbance, like tidal tails, shells or loops. In addition, this merger apparently left the spiral features of the host galaxy intact. Major mergers of spiral galaxies are expected to lead to the formation of elliptical galaxies (e.g., Schweizer 1982). This disfavors the suggestion of a second BH with a mass nearly equal to the primary BH (with the associated radio emission), but having only faint (undetectable) radio emission. It is also instructive to remember that Burke-Spolaor (2011) found only two binary AGNs in a sample of 3114 radio-bright AGNs observed with VLBI, attesting to the rarity of sources with binary back holes.

An important caveat to the discussion above must be mentioned. It is possible that the detected radio component at 1.6 GHz is not coincident with the position of the central SMBH, but is instead a bright portion of a radio outflow, like a compact jet knot or a "hot spot." However, it is not possible to verify this with the present day optical telescopes. Only a more sensitive VLBI study at multiple wavelengths may detect either more radio components from the radio outflow or reveal the spectral index information with greater confidence, and thereby elucidate the true nature of the detected radio feature. For the purposes of this paper, we have assumed that the detected radio component is indeed coincident with the SMBH in KISSR 1494.

Smith et al. (2012) have proposed that equal-peaked AGNs, of which KISSR 1494 is an example, could have their line emission originating in a rotating ring with relatively little gas at small radii. The fact that the radio component detected at 1.6 GHz in KISSR 1494 is not compact and gets completely resolved out in some image weighting schemes (see Figure 6) can be reconciled with the fact that the radio emission is from a disk or a ring/torus. Could the same rotating disk or ring/torus give rise to the double peaked emission lines? This is unlikely because of the small size of the radio feature (∼8 × 6 pc)—the NLR clouds are expected to be distributed on much larger 10–100 pc scales, extending beyond the parsec-scale torus. However, it is possible that the disk is much more extended than the region probed by VLBI and the NLR clouds form part of the outer regions of this disk. The missing diffuse radio emission going from kiloparsec to parsec-scales could also arise in an extended disk emitting at radio frequencies. Mulchaey et al. (1996) have demonstrated that if the NLR is in a disk then the V-shaped ionization cones will always be observed in Seyfert galaxies, irrespective of our line of sight being inside or outside of the cones. The prevalence of ionization cones therefore favors NLR disks.

Gallimore et al. (2004) have suggested that the flat/inverted spectrum radio core in the Seyfert galaxy NGC 1068 is thermal bremsstrahlung (free–free) emission from the inner hot ionized edge of the torus. Roy et al. (1998) have discussed the possibility that the radio core in NGC 1068 is direct synchrotron emission from the torus instead. The steep spectrum along with the moderate brightness temperature T_B ∼ 10⁷ K in the radio core of KISSR 1494 does not favor synchrotron self-absorption from the torus as an explanation, but rather optically thin synchrotron emission. For synchrotron self-absorption to occur, T_B > 10⁹ K is typically required. Steep-spectrum parsec-scale radio emission from another Seyfert galaxy, NGC 4388, has been suggested to originate in a disk (Giroletti & Panessa 2009). The size of the radio emitting disk in NGC 1068 is ∼0.4 pc with a T_B = 2.6 × 10⁶ K, while the size of the radio disk in NGC 4388 is ∼0.5 pc with a T_B = 1.3 × 10⁶ K. The NLR in NGC 1068 and NGC 4388 are probably not disky because their SDSS spectra do not exhibit double-peaked emission lines, unless their NLR disks are nearly face-on. If the radio component in KISSR 1494 is indeed emission from a disk, its size is at least 10 times larger than disks in NGC 1068 and NGC 4388. Its brightness temperature is five to ten times higher as well. The significance of this difference is unclear at present.

If a radio outflow impacts the gas clouds in the NLR, then the approaching and receding emission-line gas clouds could give rise to the double peaked emission line spectra. The lack of a clear parsec-scale core-jet structure is not consistent with a powerful radio outflow in KISSR 1494. However, a broad wind-like parsec-scale radio outflow cannot be ruled out.

As noted in Section 1, the total radio flux density of KISSR 1494 with the VLA FIRST survey at 1.4 GHz is ∼23 mJy. This flux density drops to 0.65 mJy at 1.6 GHz with VLBI (present study). Hence most of the radio flux density comes from a diffuse radio component on scales between a few arcseconds and 10 mas. AGN-driven outflows have indeed been observed on scales of a few kiloparsec in Seyfert galaxies, especially with sensitive radio observations (e.g., Kharb et al. 2006, 2014b; Gallimore et al. 2006). If we assume that the kiloparsec-scale emission is lobe emission from a broad AGN outflow in KISSR 1494, then we can estimate the time-averaged kinetic power (Q) of this outflow following the relations in Punsly & Zhang (2011) for radio-powerful sources (see also Willott et al. 1999):

$$\begin{eqnarray*} {Q}&\approx& 1.1\times 10^{45} [(1+z)^{1+\alpha } Z^2 F_{151}]^{6/7} {\rm {erg\, s^{-1}}},\\ {Z}&\equiv& 3.31 - 3.65 [(1+z)^4 - 0.203 (1+z)^3\\ &&+ 0.749 (1+z)^2 + 0.444 (1+z) + 0.205]^{-0.125}, \end{eqnarray*}$$

where F₁₅₁ is the 151 MHz flux density and z is the source redshift. We derive F₁₅₁ using the 1.4 GHz flux density from FIRST and a lobe spectral index of α = −0.8, and obtain a kinetic power of Q ≈ 1.5 × 10⁴² erg s⁻¹. This Q value is typical of outflows in low-luminosity AGNs (e.g., Mezcua & Prieto 2014).

Blundell & Kuncic (2007) have suggested that the radio core emission in radio-quiet quasars is thermal free–free emission from a slow dense disk wind. The radio data of KISSR 1494 are consistent with a disk wind that emits synchrotron emission instead. It could indeed be wind emission from the tenuous corona region above the accretion disk, which has a sufficient supply of relativistic electrons and strong magnetic fields (e.g., Merloni & Fabian 2002; Laor & Behar 2008). In this scenario, the radio emission does come from a less collimated outflow compared to a relativistic jet. As mentioned by Marscher (2006), the corona itself could be the low altitude portion of a wind flowing from the accretion disk. The double-peaked emission lines in the source could also be explained if the NLR clouds are a part of this broader outflow. Even the equal peaks of the Hα and Hβ emission lines can be explained if they arise in a highly symmetric outflow close to the central engine, whose intensities are not significantly affected by dust obscuration.

We conclude that the radio continuum emission and the double-peaked emission lines in the spectrum of KISSR 1494 are consistent primarily with two explanations: (A) the mas-scale radio emission is the parsec-scale base of a synchrotron wind from the magnetized corona above the accretion disk. The NLR clouds are also a part of this outflow and extend to larger scales, probably of the order of tens to hundreds of parsecs (e.g., Crenshaw et al. 2000; Ruiz et al. 2005), and (B) the milliarcsec-scale radio emission comes from the (static, as in not outflowing) inner ionized edge of the accretion disk or torus; this disk extends to tens or hundreds of parsecs and also incorporates the rotating NLR clouds. In both these cases the radio emission is from the synchrotron mechanism. While the extensive literature on the presence of parsec-scale radio outflows in Seyferts (e.g., Giroletti & Panessa 2009) supports explanation A, the fact that KISSR 1494 is an "equal-peaked" AGN supports the disk explanation B. However, as mentioned in Section 4.3, a highly symmetric AGN-driven NLR outflow where the line intensities from the receding flow are not significantly obscured by dust, could also support explanation A in "equal-peaked" AGNs. Therefore, although our present data cannot clearly distinguish between explanations A and B, there appears to be more supporting evidence for explanation A, i.e., a non-thermal coronal disk wind. Finally, the radio and NLR emission may be decoupled. The radio emission may come from an outflow while the NLR may be in a disk, and vice versa.

If the radio core in KISSR 1494 is indeed an outflowing coronal wind, it is interesting to note that its properties (magnetic field strength, total energy density, electron lifetime, kinetic power) are no different from the radio "jet" properties derived in other Seyfert galaxies. This supports the idea that radio "jets" and outflowing coronal winds are indistinguishable in Seyferts. The implication of this finding needs to be explored further.