Rare Finding of a 100 Kpc Large, Double-lobed Radio Galaxy Hosted in the Narrow-line Seyfert 1 Galaxy SDSS J103024.95+551622.7

The source has a radio core (α₂₀₀₀ = 10:30:24.787, δ₂₀₀₀ = 55:16:18.48 with peak flux = 14 mJy beam⁻¹) and extended radio structures. The angular separation between the northeast (NE) (72 mJy beam⁻¹) and southwest (SW) (21 mJy beam⁻¹) peaks is 20.5 arcsec which for the redshift of the host galaxy corresponds to a projected size⁷ of 116 kpc. Along with peaks at the lobes and the radio core coinciding with the host galaxy seen in the SDSS r-band image, there is significant diffuse emission present on either side of the double lobe. This is strong evidence that the NE and SW peaks are not unrelated point sources but constitute a single radio galaxy with FR II-like radio structure with back-flow from hotspot regions causing the diffuse emission. Integrated flux calculated for the whole source from the FIRST image is 155 mJy. So the diffuse emission (subtracting the core and peaks at lobes from the integrated flux ($155-(14+72+21)=48$ mJy)) constitutes the highest 30% of the whole radio source flux density. It may also be noted that the radio source (lobe–core–lobe) does not lie in a straight line and is slightly bent with the NE lobe brighter than the SW counterpart. Also, the presence of diffuse radio emission at the resolution of FIRST can be established from the dimensionless concentration parameter (Ivezić et al. 2002), ${\theta }_{\mathrm{FIRST}}=\sqrt{{S}_{\mathrm{int}}/{S}_{\mathrm{peak}}}$, where S_int and S_peak are the integrated and peak flux densities of the source respectively. θ_FIRST is found to be 1.69, much larger than the value of 1.06 above which a source is defined as "resolved" by Kimball & Ivezić (2008), indicating extended emission. Since FIRST is a high-resolution survey, it is insensitive to the very extended emission originating from the lobes. However, the NRAO VLA Sky Survey (NVSS) at the same frequency provides low-resolution images, thus detecting much more extended emission. Therefore, following Singh et al. (2015), we estimated ${\theta }_{\mathrm{NVSS}-\mathrm{FIRST}}=\sqrt{{S}_{\mathrm{NVSS},\mathrm{int}}/{S}_{\mathrm{FIRST},\mathrm{int}}}$, where ${S}_{\mathrm{NVSS},\mathrm{int}}$ and ${S}_{\mathrm{FIRST},\mathrm{int}}$ are the integrated flux densities obtained from NVSS and FIRST images respectively. The value of ${\theta }_{\mathrm{NVSS}-\mathrm{FIRST}}$ turns out to be 2.18, suggesting the presence of an additional faint low-surface-brightness radio component in the source apart from the extended radio emission detected in FIRST.

To estimate the radio spectral index (α) we collected multi-frequency radio data from the NASA/IPAC Extragalactic Database (NED)⁸ and estimated α using data from 74 MHz to 5 GHz utilizing a linear least-squares fit in the log–log plane. This is shown in the bottom panel of Figure 1. A steep radio spectrum (${S}_{\nu }\propto {\nu }^{\alpha }$) with α = −0.65 ± 0.04 has been found for this object, suggesting extended emission originating from the young radio lobes.

The variability of the source in the optical and mid-IR bands was also investigated using data from the Catalina Real Time Transient Survey (CRTS; Drake et al. 2009) and the Wide-field Infrared Survey Explorer (WISE; Wright et al. 2010) database, respectively. The CRTS optical light curve is shown in the top panel of Figure 2. The amplitude of variability (σ_m) was calculated following Sesar et al. (2007) (see also Ai et al. 2010; Rakshit & Stalin 2017). The source is found to be non-variable in the optical. It is also found to be non-variable in the mid-IR W1 and W2 bands both on short (∼1 day) and long timescales (∼7 yr).

Zoom In Zoom Out Reset image size

To check on the NLSy1 galaxy classification of SDSS J103024.95+551622.7, we collected the SDSS spectra, removed the effects of Galactic extinction, brought it to the rest frame, and re-analyzed it using the spectral decomposition procedure described by Rakshit et al. (2017) but without subtracting the host galaxy component, which is negligible as the redshift of the object is z = 0.4347. The source was observed twice, hence we fitted both the spectra as shown by the black line in the upper and lower panels of Figure 3 for two different cases: when the broad Hβ component was fitted using a Gaussian function (left) and a Lorentzian function (right). A Lorentzian function provides a better fit to the spectra, having lower reduced-χ² values in both spectra. The use of a Lorentzian function to fit the broad component of the Hβ emission naturally provides narrower FWHMs which are 2170 ± 45 km s⁻¹ and 2140 ± 38 km s⁻¹ for the two epochs compared to 2883 ± 34 km s⁻¹ and 2842 ± 70 km s⁻¹, respectively, found using a Gaussian function. The canonical definition of NLSy1 galaxies pertain to sources with FWHM(Hβ) < 2000 km s⁻¹ (Goodrich 1989). However, the criterion of FWHM(Hβ) < 2200 km s⁻¹ has been used in the initial catalog of NLSy1 galaxies by Zhou et al. (2006) and recently by Rakshit et al. (2017) since the distribution of the broad Hβ line width is smooth with no sharp cutoff between NLSy1 and BLSy1 galaxies. Moreover, a higher cutoff of FWHM(Hβ) < 4000 km s⁻¹ based on the quasar main sequence (Sulentic et al. 2000), a luminosity-dependent cutoff value for the Hβ line width (Laor 2000; Véron-Cetty et al. 2001), or an Eddington ratio cutoff (L/L_EDD ≥ 0.25) (Netzer & Trakhtenbrot 2007) have also been suggested for NLSy1 galaxies. The source studied here is close to the boundary of 2200 km s⁻¹ adopted by Zhou et al. (2006) and Rakshit et al. (2017). However, it has a low [O iii]/Hβ ratio and steep soft X-ray spectral index and an Eddington ratio of 1.01. Putting all the other characteristics together, it is most likely that SDSS J103024.95+551622.7 is an NLSy1 galaxy.

Zoom In Zoom Out Reset image size

We collected information on all NLSy1 galaxies having larger than 20 kpc radio emission from the literature and tabulate them in Table 1. Note that one of the sources, NVSS J095317+283601 in Table 1 has a FHWM of Hβ (2162 ± 201 km s⁻¹) similar to that of SDSS J103024.95+551622.7. A higher FWHM (Hβ) will lead to a higher BH mass. However, all the sources except SDSS J122222.55+041315.7 in Table 1 have BH masses <10⁸ M_⊙. We investigate the relationship between projected size of the radio jet (D_proj) with M_BH and radio power at 1.4 GHz (P_1.4) for NLSy1 galaxies in Figure 4. The red point in each panel represents SDSS J103024.95+551622.7. The magenta polygons represent NLSy1 galaxies that have larger than 20 kpc extended radio structures (see Table 1). No correlation is found between the projected size with M_BH (or jet power). When the sources are put in the WISE W1-band absolute magnitude versus radio power diagram, a strong correlation is found. The points in this plot are the NLSy1 galaxies from Rakshit et al. (2017) which are detected in FIRST and WISE in the W1-band with signal-to-noise ratio > 10. The Spearman correlation test on the full sample yields a correlation coefficient (r) of −0.6 and a very low p-value (p < 1 × 10⁻²⁰), suggesting a significant negative correlation where absolute IR magnitude decreases with radio power. The BH mass of all those objects is plotted against their radio power in the right panel. We found a positive correlation having r = 0.4 which is similar to that found by Liu et al. (2006) between jet power and BH mass, suggesting high-powered jets originate from massive AGNs. From Figure 4 (bottom right panel) it is evident that NLSy1 galaxies with >20 kpc radio structure predominantly occupy the region with large radio power and high BH mass. In terms of the projected radio size, SDSS J103024.95+551622.7 is the second-largest NLSy1 galaxy after SDSS J110006.07+442144.3.⁹ In Figure 4, it follows the $W1-{P}_{1.4}$ anti-correlation and ${M}_{\mathrm{BH}}-{P}_{1.4}$ correlation, like other kpc scale radio-emitting NLSy1 galaxies.

Zoom In Zoom Out Reset image size

From the radio spectrum in Figure 1, we see that the spectrum is steep. Considering the lobes of radio galaxies which are optically thin, the derived spectral index of −0.65 ± 0.04 is very close to the theoretical value of injection spectral index¹⁰ of around −0.62 (Kirk et al. 2000; Konar et al. 2013). If the radio jets have the same spectral index of −0.65 then the inverse Compton effect of the cosmic microwave background (CMB) and radio photons together can give rise to an X-ray power-law spectrum in the 0.1 to 10 kev range whose photon index is supposed to be −(1+0.65) = −1.65 (Konar et al. 2009). However, for the source SDSS J103024.95+551622.7 the photon index estimated from ROSAT data is −2.61, which is much steeper. To resolve this inconsistency, we need further observations in the 0.1–10 keV band. The presence of a strong [O iii] doublet, as seen in Figure 3, indicates strong blackbody emission from the accretion disk, suggesting a standard accretion disk through which the matter is being accreted onto the BH of this source. Therefore, this NLSy1 radio galaxy is a high-excitation radio galaxy in which cold mode accretion is taking place (Allen et al. 2006). Assuming the radio spectral index (−0.65) of this radio galaxy to be the injection spectral index (α_inj), the correlation between α_inj and jet power Q_jet, as published by Konar & Hardcastle (2013) in their Figure 2, we estimated a jet power of this source to be 3 × 10⁴⁴ erg s⁻¹. This suggests that this NLSy1 galaxy has a jet power typical of classical radio galaxies.

In the lower left panel of Figure 4, we see that there is a strong correlation between W1 and P_1.4. Most of the luminosity in the W1-band of WISE comes from the old stars in the bulge. This means the greater the W1 luminosity, the heavier the bulge, and the heavier the BH mass (Kormendy & Richstone 1995). With a heavier BH mass, we might get higher jet power provided the Eddington-scaled mass accretion rate and the spin of the BHs are in a narrow range. We indeed get a correlation between BH mass and the radio power at 1.4 GHz (which is a proxy of jet power). Therefore, we can expect that all NLSy1 galaxies are accreting with Eddington-scaled mass accretion rate within a very narrow range. If spin of the BH has a role in determining the jet power, then all NLSy1 galaxies must have spin within a very narrow range (or jet power has no dependence on the spin). Therefore, the correlations in the left and right bottom panels are as expected.

Similarly, the absence of correlation in the top right and top left panels are as expected. Let us consider the top left panel, where source size versus M_BH is plotted. If the source size depended only on the jet power, then we would perhaps obtain a correlation between source size and BH mass. However, the source size depends on jet power (Q_jet), density (ρ_amb) of the ambient medium, and the angle (θ) of the jet with the line of sight. Even if the jet power is correlated with BH mass, the scatter of the values of ρ_amb and θ from source to source will spoil any correlation between source size and BH mass. Since BH mass and P_1.4 are correlated, we do not expect to get any correlation between source size and P_1.4, as there is no correlation between source size and BH mass. A larger sample will help to address this.

NLSy1 galaxies are more often associated with spiral or disk galaxies with rich gaseous central regions (Deo et al. 2006; Ohta et al. 2007). Not only in NLSy1 galaxies but also in general Seyfert galaxies, radio-loud galaxies with over 100 kpc large radio lobes are rare. While almost all large (over 100 kpc) radio galaxies are hosted in elliptical galaxies, only six of them are known to be hosted in disk/spiral galaxies with large radio lobes. In Figure 5, we plot the optical g − r color against r magnitude of NLSy1 galaxies from Rakshit et al. (2017) in dots (magenta) along with SDSS quasars in the background (black). Our target is represented by a star symbol having a g − r color of 0.10 mag. According to Banerji et al. (2010), a galaxy can be defined as a spiral if the observer-frame g − r < 0.6 mag (see also Tempel et al. 2011). Thus, SDSS J103024.95+551622.7 can be considered as a spiral galaxy following the above criteria. According to the plot, the majority of NLSy1 galaxies are expected to be spiral. However, deep imaging observations of NLSy1 galaxies are needed to confirm if their hosts are indeed spirals. Nevertheless, a few examples of spiral host radio galaxies are known, such as J0313-192 (Ledlow et al. 1998), Speca (Hota et al. 2011), J0314-1906, J2345-0449, J0836+0532, and MCG+07-47-10 (Hota et al. 2016). It is likely that such spiral host radio galaxies could be more frequent in the early universe, prior to the quasar/radio galaxy era (z ∼ 2–3) and when spiral galaxies were more prevalent than the merger-remnant ellipticals (Hota et al. 2016). Thus, this new object reported here, SDSS J103024.95+551622.7, hosted in a spiral galaxy, may represent an opportunity to investigate non-standard modes of relativistic jet production in passively evolving galaxies with smaller BH masses which may have been more common in the early universe, prior to the rise of merger-remnant ellipticals and supermassive BHs.

Zoom In Zoom Out Reset image size

The observed core dominance parameter (r_obs; Orr & Browne 1982) can be calculated from the ratio of core (S_core) to lobe (S_lobe) flux densities. At 1.4 GHz from FIRST, the source has an integrated core flux density of 32.7 mJy with the SW and NE lobes having integrated flux densities of 21.9 mJy and 95.3 mJy respectively. The observed core dominance parameter is thus r_obs = 0.28. This value is higher than that known for radio galaxies with FR I and FR II radio morphologies having median values r_obs = 0.022 and 0.003 respectively (Morganti et al. 1997). We note that blazars show extreme core dominance, which is caused by Doppler boosting (Chen et al. 2015). We calculate the Doppler factor (δ) following Doi et al. (2012)

$$\begin{eqnarray}&&{\delta }_{\mathrm{core}}^{(3-\alpha )}=\displaystyle \frac{{r}_{\mathrm{obs}}}{{r}_{\mathrm{int}}},\end{eqnarray} \tag{ 1 }$$

where r_int is the intrinsic core dominance parameter and α is the core spectral index. We assumed the core spectral index to be flat with α = 0. The calculation of r_int needs an estimation of the intrinsic core power. We derived the intrinsic core power using the empirical relation known for radio galaxies and given by $\mathrm{log}{P}_{c}=(0.62\pm 0.04)\times \mathrm{log}{P}_{t}+(7.6\pm 1.1)$ (Giovannini et al. 2001), where P_c is the core radio power at 5 GHz and P_t is the total radio power at 408 MHz. Using the core flux at 5 GHz and the total flux at 365 MHz from NED, we found r_int = 0.008. This is much lower than r_obs, suggesting strong Doppler boosting in the core. Using Equation (1) we estimated δ = 3.3.

To constrain the jet speed (β_core = v_core/c) and inclination angle (θ_core) we used the relation ${\delta }_{\mathrm{core}}=\sqrt{(1-{{\beta }^{2}}_{\mathrm{core}})}/(1-{\beta }_{\mathrm{core}}\,\cos {\theta }_{\mathrm{core}})$. This relation is highly degenerate; however, we obtained values of β_core > 0.87 and θ_core < 12°. These values are similar to those obtained by Doi et al. (2012) for a sample of NLSy1 galaxies, suggesting that radio-loud NLSy1 galaxies are viewed pole-on. Similar low-inclination values in the range of 10°–15° were also obtained by Richards & Lister (2015) for other kpc scale jetted NLSy1 galaxies. Considering a viewing angle of 12°, we found the deprojected size of the source to be ${D}_{\mathrm{de} \mbox{-} \mathrm{proj}}\,\gt 116/\sin 12^\circ \gt 557\,\mathrm{kpc}$.

The source exhibits a two-sided structure at kpc scales as seen in Figure 1. Therefore, the flux ratio of the approaching and receding lobes (R_F) were used to calculate the speed of the kpc scale lobes following Richards & Lister (2015)

$$\begin{eqnarray}&&{R}_{F}={\left[\displaystyle \frac{(1+{\beta }_{\mathrm{kpc}}\cos \theta )}{(1-{\beta }_{\mathrm{kpc}}\cos \theta )}\right]}^{3-\alpha },\end{eqnarray} \tag{ 2 }$$

where the spectral index α = −0.65. Using, R_F = 4.35, we obtained ${\beta }_{\mathrm{kpc}}\cos \theta =0.2$. Assuming the inclination angle to remain constant between the core and kpc scale lobes, i.e., θ = 12°, we obtained β_kpc ∼ 0.2. Considering the expansion speed of β_kpc and a two-sided expansion, we estimated the kinetic age of the source using ${t}_{\mathrm{kinetic}}={D}_{\mathrm{de} \mbox{-} \mathrm{proj}}/2{\beta }_{\mathrm{kpc}}c$, where ${D}_{\mathrm{de} \mbox{-} \mathrm{proj}}$ is the deprojected size of the source. We obtained t_kinetic > 4 × 10⁶ yr.

We also estimated the limiting radiative age of this radio source as ${\tau }_{\mathrm{rad}}=50.3\tfrac{{B}^{1/2}}{{B}^{2}+{B}_{\mathrm{IC}}^{2}}\tfrac{1}{\sqrt{{\nu }_{\mathrm{br}}(1+z)}}$, where B is the lobe magnetic field expressed in nT, B_IC = 0.318(1 + z)² nT is the CMB equivalent magnetic field (Konar et al. 2006) and ν_br is the break frequency (in GHz) observed in the radio spectrum. From the above expression, we get the highest radiative age, when $B=\tfrac{{B}_{\mathrm{IC}}}{\sqrt{3}}$. Using $B=\tfrac{{B}_{\mathrm{IC}}}{\sqrt{3}}=0.378$ nT, we obtained a radiative age of 20.2 Myr for the break frequency of 5 GHz. Since there is no break observed in the radio spectrum, the highest observed radio frequency is the lower limit of the break frequency. Thus, ν_br > 5 GHz, corresponds to τ_rad < 20.2 Myr. Both the kinetic and radiative age measurements agree well and provide a strong constraint on the age of SDSS J103024.95+551622.7 to be >10⁶ yr. This also lies within the age range of 10⁵–10⁷ yr found for other radio-loud NLSy1 galaxies by Doi et al. (2012) and Richards & Lister (2015). It is thus likely that NLSy1 galaxies are young AGNs (Mathur et al. 2001).