The Complex Gaseous and Stellar Environments of the Nearby Dual Active Galactic Nucleus Mrk 739

We astrometrically calibrated the VLT/MUSE data cube, using the sources reported by the Gaia Data Release 2 (Lindegren et al. 2018) as a reference. We matched the MUSE positions of Mrk 739E, Mrk 739W, and the star 2MASS J11363074 + 2136018 (Monet et al. 2003) that appears at the top-left corner of the field of view with the Gaia positions. The matching and offset determinations were carried out using the astrometrical calibration tool provided by the Aladin Sky Atlas (Bonnarel et al. 2000). The measured offsets, corresponding to the average displacement of the three sources, are ΔR. A. = −1416 ± 0042 and Δdecl. = 0219 ± 0016. These offsets were applied in the subsequent analysis.

In order to understand the morphologies and kinematics for different atomic species, we analyze here the resulting maps for each emission line. Figures 4 and 5 show the gas distribution for the main emission lines. We identify two clearly separated emission peaks, corresponding to the nuclei of each galaxy, and the arc-shaped clumpy structure observed in Balmer lines, [N II] and [S II], which is mostly located to the west of Mrk 739W. In these emission-line maps, especially in that of Hα λ6563, we can see many features that are not present in the collapsed white-light image shown in Figure 1, such as the fan-like stream of ionized gas in the north or several blobs associated with the aforementioned clumpy distribution to the west of the system. Below, in Section 3.2, we discuss the emission mechanism responsible for the emission of these regions using an optical emission-line diagnostic diagram in order to identify the origin of the ionized gas. The Hβ emission is in general weak, particularly in the south region of the western nucleus, delimited by the cyan contours, where the presence of Hβ absorption lines overlaps with the emission line. The south of the eastern nucleus completely lacks Hβ emission.

Zoom In Zoom Out Reset image size

The spatial distribution of the [O III]λ5007 (hereafter [O III]) emission line starkly contrasts with the Hα map discussed previously. The flux map clearly shows an extended and intense spiraling crest to the north of both nuclei, which does not have a symmetric equivalent to the south or west. Morphologically, neither [O III] nor [O I]λ6300 reveal other obvious or distinguishing features. Other interacting pairs of galaxies show complex and disrupted features at more advanced stages like stellar bridges (Torres-Flores et al. 2020), irregular structures (Sengupta et al. 2017), or unique stellar halos (Foster et al. 2014), suggesting that the lack of those characteristics on Mrk 739 is evidence that we are seeing the merger either before or just after the first passage. The lack of [O III] emission in the western region of Mrk 739E is because we are only reporting the flux of the second component in this region. This choice is due to the relevance of the component either in the morphology, kinematics, or the evolutionary stage of the merger. The noteworthy extended tidal tail on the northeast of Mrk 739 is a ubiquitous feature of the gravitational interaction between galaxies that is formed by gas that is stripped from the outer regions of the interacting systems due to gravitational forces (Toomre & Toomre 1972; Oh et al. 2008, 2015). In addition, Figure 5 presents the distributions for the [N II]λ6583 (left) line and the [S II]λ6717, λ6730 doublet (right).

We compute emission-line flux ratios using the best-fit narrow components for each individual spaxel with sufficient signal, as defined above. In Figure 6, we present maps for the log([O III]λ5007/Hβ), log([N II]λ6583/Hα), log([S II]λ(6717+6731)/Hα), and log([O I]λ6300/Hα) flux ratios, which we use to diagnose the nature of the ionization source (Baldwin et al. 1981; Veilleux & Osterbrock 1987; Kewley et al. 2000, 2001, 2006). The green regions in the first two maps denote areas where Hβ and Hα dominate over [O III] and [N II] lines, respectively, which is indicative of star formation processes. On the other hand, the brown regions correspond to areas where the intensity is roughly equal for both lines. According to Kewley et al. (2006), regions with log([O III]/Hβ) > 1 are most likely dominated by AGN ionization. We identify a clear extended region that has a high [O III]/Hβ ratio, which is likely ionized by nuclear activity to the northeast.

Zoom In Zoom Out Reset image size

We combine the optical emission-line ratios, discussed individually above, to obtain a unified classification of the dominating ionization source, as was originally done in the so-called Baldwin, Phillips & Telervich diagram (BPT; Baldwin et al. 1981). Figure 7 shows the BPT diagnostic diagrams for individual regions in the Mrk 739 system. The resolved BPT reveals an extended zone ionized by AGN activity and structures related to the western side of the galaxy merger, where star formation dominates. These maps reveal the widespread influence of the AGN, which is detected out to large distances (5–20 kpc) from the center. The innermost (2–3 kpc) surroundings of the two nuclei are consistent with being ionized by star formation, with composite ionization arising at the edges, where the transition between the central regions and the rest of the galaxy occurs. The yellow dots in the [N II]-BPT diagram correspond to the spaxels belonging to the double-peaked emission line, not shown on the maps for clarity. The region between Kewley et al. (2001) and Kauffmann et al. (2003), denoted by the lime green dots in [N II]-BPT, characterize the edges of the most prominent star-forming regions in the resolved [N II]-BPT. There appears to be a nuclear region around Mrk 739E that is dominated by both composite and star formation mechanisms. This picture is consistent with the scenario proposed by Hopkins et al. (2006) where the gravitational interaction could have driven the gas toward the center, triggering star formation processes. In the particular case of Mrk 739E, we see that in the region of double-peaked emission line, both sets of lines are independently consistent with being ionized by star formation.

Zoom In Zoom Out Reset image size

Finally, the clumpy western region is dominated by star formation, consistent with the results of Koss et al. (2011), who pointed out that Mrk 739W does not reveal evidence for an AGN in the optical/UV due to the high levels of star formation.

The line-of-sight (LOS) velocities and velocity dispersion of the gas were obtained from the Doppler shift and broadening of the three strongest emission lines: Hβ, [O III]λ5007, and Hα. Their spatial distributions are presented in Figure 8, where we report the velocity values in the galaxy rest frame, which correspond to a heliocentric velocity of 8950 ± 5 km s⁻¹ based on ¹²CO(J = 1 → 0) observations (Casoli et al. 1996). As mentioned in Section 2, the width of all emission lines was tied to that of [O III]λ5007. For the [O III], Hβ, and Hα maps, we report the velocity obtained from the narrow Gaussian component that characterizes the systemic excited line emission of the galaxy and the velocity values of the second component, describing the aforementioned isolated structure that is ionized by star formation. Figure 8 highlights a spatially extended region to the north, with velocities close to ∼60 km s⁻¹. Any motion associated with the eastern tidal tail, the western nucleus, and the western side of the galaxy merger appears to be located in the plane of the galaxy, as all have negligible relative velocities. The nuclear region of Mrk 739E reveals a blueshifted velocity of ∼−130 km s⁻¹ in [O III] likely related to outflowing material coming from the eastern nucleus, resulting in a higher [O III] velocity dispersion with values of σ_gas ∼ 150 km s⁻¹, as we see in the bottom-right map of Figure 8. We see an intriguing blueshifted region to the north of the western nucleus with values about ∼−60 to ∼−85 km s⁻¹ and a velocity dispersion between ∼100–120 km s⁻¹ extending north–south toward the Mrk739W nucleus. The velocity maps, in particular Hα, reveal a circular velocity profile surrounding the eastern nucleus with values ranging from ∼−220 km s⁻¹ to ∼+40 km s⁻¹ that might be related with a rotating disk. In the next subsection, we study in detail the origin of this component.

Zoom In Zoom Out Reset image size

The MUSE data reveal a region around the eastern nucleus that appears to be partially decoupled from the rest of the system. The top-right panel of Figure 3 shows a clear double-peaked profile of the emission lines in that region. The presence of double-peaked emission lines in galaxies can be associated with galaxy mergers (Comerford et al. 2018) and dual AGNs (Wang et al. 2009), or even outflows and rotating gas (Greene & Ho 2005; Nevin et al. 2016). Taking advantage of our IFU data, in the bottom -left panel of Figure 8, we morphologically identify a rotating disk-like velocity profile surrounding the eastern nucleus.

Spectrally, the double-peaked emission-line region consists of a bluer emission line that becomes gradually redshifted as the position varies from east to west of the eastern nucleus. The originally redder line does not change its velocity with position. Instead, the flux of the redder line decreases until it disappears after reaching the eastern side of the disk, explaining thus the absence of double-peaked emission lines on the eastern side of the disk. We then consider the bluer moving line as the rotating component, while the redder and static line matches the velocity of the extended northern spiraling crest, which is thus related to the large-scale distribution of the galaxy.

There are two possible lines we might consider using to trace the kinematic structure around the eastern nucleus, [O III] and Hα. The [O III] emission line would be cleaner for the kinematic modeling, in order to avoid the broadening problem affecting the Balmer lines. However, the high extinction (see Section 5.1), the presence of blueshifted velocities in the nuclear region, and the spectral resolution of R ∼ 1770, corresponding to a velocity resolution (FWHM) of ∼170 km s⁻¹ for the [O III] line, are not sufficient to study the rotating disk map. The Hα line, thanks to its higher S/N and spectral resolution of FWHM ∼120 km s⁻¹, allows instead for a better kinematic modeling, despite the presence of the BLR contaminating the central portion of the map.

In order to fit the velocity distribution of this rotating component around the eastern nucleus and derive its physical parameters, we perform a Markov Chain Monte Carlo (MCMC) kinematic simulation using the KINematic Molecular Simulation (KinMS ¹⁴ ) package developed by Davis et al. (2013). KinMS generates spectral cubes of synthetic data, used to simulate observations of arbitrary molecular/atomic gas distributions. We used it in order to create a simulated first-moment map, assuming a rotating disk, and then fit the associated parameters using the KinMSpy_MCMC ¹⁵ routine.

We fit a relatively simple model, consisting of a rotating disk with an exponential surface brightness profile given by ${{\rm{\Sigma }}}_{{{\rm{H}}}_{\alpha }}(r)={e}^{\tfrac{r}{{r}_{0}}}$, where r₀ is the disk scale factor. We assume that the disk is rotating with a circular velocity profile that is constant with radius. This assumption is based on the fact that both spatial and spectral resolution do not allow us to distinguish more complex features of the rotating disk. The LOS inclination and position angle of the disk are free parameters of the model, along with the scale factor and the flat velocity of the rotating disk. The angles are measured starting from the north and moving counterclockwise.

The Gibbs sampling MCMC framework with adaptive stepping implements a Bayesian analysis technique to explore the parameter space using a set of walkers from the emcee algorithm (Foreman-Mackey et al. 2013) to find the posterior distribution of the model parameters. A standard log-likelihood function, based on the chi-squared χ² distribution $\mathrm{ln}P=-{\chi }^{2}/2$, was used by each walker in order to maximize the likelihood and determine their next step through parameter space. Assuming Gaussian errors for our free parameters, we can estimate the goodness-of-fit of our model using χ² statistics, ${\chi }^{2}={\sum }_{i}{\left(\tfrac{{\mathrm{data}}_{i}-{\mathrm{model}}_{i}}{{\sigma }_{i}}\right)}^{2}$ for each pixel, i. Since we are studying the first moment of the velocity of the Hα line profile, we need to compute the rms noise, σ, in velocity units. To accomplish this, we made a new but simpler MCMC calculation over a few representative spaxels within the disk region, where the signal of the spectra was modified by random Gaussian noise. The variation was spectrally convolved with a Gaussian kernel matching the spectral resolution of the MUSE instrument. We then fitted the convolved spectra with Gaussian components, as described in Section 2. We repeated this process 1000 times to obtain statistically significant results. For each spectrum, we calculated the LOS velocity for all relevant emission lines. We then took the standard deviation of the 1000 velocities as noise. We found that the noise associated with the measurement of the velocity is mostly constant, at a value of ∼20 km s⁻¹. In order to exclude from the fit the inner region, dominated by the emission from the BLR associated with the AGN, we assign here artificially high rms values.

Our MCMC fitting procedure considers 1.5 × 10⁵ steps, using uniform priors in linear space. The priors, guesses, and resulting best values for the parameters of our model are presented in Table 1. In addition, Figure 9 shows the so-called corner plot, where we can see the posterior distributions of parameters together with the one-dimensional marginalization of the physical parameters that characterize the rotating disk. There is a clear degeneracy between the inclination of the disk and the circular velocity since the relation between the two is given by $\mathrm{vel}\propto \tfrac{{\mathrm{vel}}_{\mathrm{obs}}}{\sin (i)}$. Also, the scale factor is degenerate with the inclination, due to projection effects on the plane of the sky, while the scale factor correlates with the final circular velocity because of the dependency of the velocity on the inclination.

Zoom In Zoom Out Reset image size

For easier visual inspection, in Figure 10 we present the kinematically detached rotating disk of Mrk 739E along with the best-fit model and their residuals. We find a well-defined rotating disk, with a circular velocity of ${237}_{-28}^{+26}\,\mathrm{km}\,{{\rm{s}}}^{-1}$ inclined by ${33}_{-3}^{+5}$ degrees. The model agrees reasonably well with our data, considering the small velocity residuals presented in the right panel of Figure 10. We expect that these low residual values, mainly in regions where the disk velocity is close to 0 km s⁻¹, are produced by a nonoptimal separation in our spectral fitting procedure of the disk component and the component related to the galaxy, due to the spectral resolution. The proper convergence of the parameters within our MCMC treatment is not affected by those residual values. Multiwavelength IFU data at higher spatial resolutions, as those obtained using the MUSE narrow field mode (NFM), as part of ESO program 0104.B-0497(A), would be helpful to understand and better constrain the physical parameters of the rotating disk and disentangle the central AGN-dominated emission from the disk.

Zoom In Zoom Out Reset image size

Our simple rotating disk model allows for the determination of the mass of the material surrounding the eastern nucleus just considering Newtonian physics. Using the flat velocity and scale factor r₀ as a characteristic radius, we can find that the mass within r₀ is given by:

$$\begin{eqnarray}&&M(R)={v}^{2}\displaystyle \frac{{r}_{0}}{G}\end{eqnarray} \tag{ 1 }$$

where G is the gravitational constant, and v and r₀ are the flat velocity and the scale factor, respectively, summarized in Table 1. Thus, inside of a radius of 1.21 kpc centered at the eastern AGN position, the mass is $\mathrm{log}M({M}_{\odot })=10.20\pm 0.06$, almost three orders of magnitude larger than the SMBH mass we estimate for the eastern nucleus in Section 5.2 based on our single-epoch BLR constraints.

Unlike other dual AGNs, e.g., NGC 6240 (Müller-Sánchez et al. 2018; Kollatschny et al. 2020), Mrk 463 (Treister et al. 2018), and Mrk 266 (Mazzarella et al. 2012), Mrk 739 does not show evidence of large-scale structures with high velocity dispersion generated by the ongoing major merger. The system presents narrow lines with a velocity dispersion close to the spectral resolution of MUSE at large scales, while the nuclear region of Mrk 739E displays high velocity dispersion in [O III], likely related to an outflow with a dispersion of σ_gas ∼ 170 km s⁻¹. The dispersion in Mrk 739W is more extended with values close to σ_gas ∼ 120 km s⁻¹. The rotating disk presents a uniform circular velocity and an inclination that indicates that this structure is kinematically decoupled from the rest of the merger.

According to Koss et al. (2011), the Sloan Digital Sky Survey r-band magnitudes obtained from two-dimensional surface brightness fitting are m_r = 13.75 ± 0.15 for Mrk 739W and m_r = 14.03 ± 0.15 for Mrk 739E. Considering that the luminosity at longer wavelengths can be used as a tracer of the stellar mass of the galaxy (Cole et al. 2001; Bell et al. 2003), the ratio between the r-band magnitudes gives a rough estimation of the ratio between the stellar masses of each galaxy in Mrk 739. As a consequence, it can be concluded that the two galaxies have roughly the same mass, therefore qualifying their interaction as a major merger with M/m < 3.

In addition, the near-IR luminosity, for example, in the K band (2.2 μm) is a good tracer of the stellar mass, since most of the K-band light of the stellar population arises from long-lived giant stars and is less affected by dust obscuration than shorter wavelengths (e.g., Kauffmann & Charlot 1998). The infrared K-band magnitudes of Mrk 739E and Mrk 739W, as reported by Imanishi & Saito (2013), are m_K = 11.37 and m_K = 12.74, respectively. AGN emission weakly obscured by dust, as in Mrk 739E (see Section 5.1), can contribute substantially to the observed K-band flux (Imanishi & Saito 2013). Therefore, we subtracted the contribution of the inner 05 to the total K-band flux (see Table 3 in Imanishi & Saito 2013) in Mrk 739E, assuming that this region is fully dominated by AGN emission. The AGN-corrected K-band magnitude for Mrk 739E is then m_K = 11.83. Following the discussion in Kormendy & Ho (2013), we derive stellar masses of $\mathrm{log}({M}_{* }/{M}_{\odot })=10.86$ and $\mathrm{log}({M}_{* }/{M}_{\odot })=10.50$ for Mrk 739E and Mrk 739W, respectively. The K band yields a more accurate estimation of the stellar mass ratio, where M_east/M_west = 2.2. Mrk 739W is brighter than Mrk 739E in the r band since the young stellar populations (1–2 Gyr), located west of the system, are expected to be hot, have short lifetimes, and contribute more at short wavelengths. The oldest stellar populations in Mrk 739E confirm the fact that this galaxy is more luminous at redder wavelengths, particularly in the K band.

The SFR derived by adopting the Kennicutt (1998) relation, where $\mathrm{SFR}({M}_{\odot }\,{\mathrm{yr}}^{-1})=7.9\,\times \,{10}^{-42}{L}_{{H}_{\alpha }}(\mathrm{erg}\,{{\rm{s}}}^{-1})$, for Mrk 739W is SFR_Hα (M_⊙ yr⁻¹) = 5.3. This relation considers a Salpeter IMF, while the luminosity of the Hα was corrected by stellar absorption and dust extinction based on the Balmer decrement Hα/Hβ flux ratio, as presented in Section 5.1. We further assume that all Hα emission at the west of Mrk 739E is associated with star formation and belongs to Mrk 739W. The SFR of Mrk 739W, based on the UV emission reported by Koss et al. (2011), is SFR_UV(M_⊙ yr⁻¹) = 0.6, while the unresolved SFR for Mrk 739 derived from far-infrared (FIR) is SFR_FIR(M_⊙ yr⁻¹) = 6.9. SFR measurements on Mrk 739E are very unreliable due to AGN contamination. However, according to our optical diagnostic diagram, there are not significant star-forming regions related to the eastern galaxy. Thus, Mrk 739W would lie, within the scatter, on the star-forming main sequence of galaxies (e.g., see Peng et al. 2010; Renzini & Peng 2015), consistent with ongoing star formation and low AGN activity. Mrk 739E, on the other hand, could be undergoing significant star formation quenching, likely related to AGN activity (McPartland et al. 2018), without predominant star-forming regions and old stellar populations.

The Hα to Hβ flux ratio can be used to estimate the extinction by dust in the LOS, as it was proposed by Caplan & Deharveng (1986) based on observations of H ii regions in the Large Magellanic Cloud. In Figure 14, we report the total extinction, A_V, in the V band, based on the Balmer decrement Hα/ Hβ flux ratio. Following the procedure of Domínguez et al. (2013), we considered the reddening curve proposed by Calzetti et al. (2000), an intrinsic Balmer ratio of (Hα/Hβ)_int = 2.86 (for an electron temperature T_e = 10⁴ K; Osterbrock & Ferland 2006) adopting case B recombination, a Milky Way extinction curve and an attenuation law for galactic diffuse interstellar medium R_V = 3.12 (Calzetti et al. 2000).

Zoom In Zoom Out Reset image size

The map shows signatures for relatively high extinction surrounding both nuclei, and in particular, close to Mrk 739E, reaching extinction values A_V ∼ 3 at the location of the rotating disk. We also observe high values of dust extinction associated with the optically diagnosed star-forming regions described in Section 3.2, while most of the galaxy presents more moderate optical extinction between ∼0–2. The extinction map reveals an extended arc-shaped distribution reaching slightly higher values than its surrounding regions. Starting at the north of the eastern nucleus, the dust extinction, with values A_V ∼ 1.5, seems to be defining the perimeter of the leading edge of the potential spiral arms of Mrk 739W, which can just barely be seen in the white-light image in Figure 1. To the west, the distribution follows a semicircular shape that ends where our optical diagnostic diagrams revealed an extended star-forming region.

A variety of techniques, such as dynamical, spectral fitting, and scaling methods can be used to estimate the mass of the central SMBH in galactic nuclei (Czerny & Nikołajuk 2010; Shen 2013). For instance, one way to compute this value in AGN is to use scaling relations linking the BLR properties, the optical continuum flux, and specific emission lines through scaling factors (Vestergaard 2002; Marziani & Sulentic 2012; Shen & Liu 2012). Specifically, Kaspi et al. (2000) reported that the BLR size scales with the 5100 Å luminosity as R_BLR ∝ L^0.70±0.3. Following this procedure, Vestergaard & Peterson (2006) presented four empirical relations to estimate the central black hole mass in nearby AGNs based on single-epoch spectra. Specifically, we use Equation (6) from their work:

$$\begin{eqnarray}&&\begin{array}{l}\mathrm{log}{M}_{\mathrm{BH}}/{M}_{\odot }({\rm{H}}\beta )\\ =\ \mathrm{log}\left\{{\left[\displaystyle \frac{\mathrm{FWHM}({\rm{H}}\beta )}{1000\mathrm{km}{{\rm{s}}}^{-1}}\right]}^{2}{\left[\displaystyle \frac{L({\rm{H}}\beta )}{{10}^{42}\mathrm{erg}{{\rm{s}}}^{-1}}\right]}^{0.63}\right\}\\ \quad +\ (6.67\pm 0.03).\end{array}\end{eqnarray} \tag{ 2 }$$

From a spectrum extracted in a region centered on the Mrk 739E nucleus, with a radius of 09, we obtain values of FWHM (Hβ) = 4285 ± 382 km s⁻¹ and a luminosity of log L(Hβ) [erg s⁻¹] = 40.87 ± 0.28, which leads to a derived SMBH mass of $\mathrm{log}{M}_{\mathrm{BH}}/{M}_{\odot }=7.22\pm 0.25$. This is slightly higher than, but consistent with, the value found by Koss et al. (2011) of $\mathrm{log}{M}_{\mathrm{BH}}/{M}_{\odot }=7.04\pm 0.40$, which is based on the FWHM(Hβ) − L_λ (5100 Å) scaling relation from Vestergaard & Peterson (2006).

We can compare the value of the SMBH mass in Mrk 739E with the value obtained from the M_BH–σ_* relation (Ferrarese & Merritt 2000; Gebhardt et al. 2000; Gültekin et al. 2009; Kormendy & Ho 2013). From the stellar population maps presented in Section 4, we estimate the velocity dispersion of the spheroidal component by averaging the bins surrounding the masked region of the eastern nucleus within an aperture of 09, obtaining a value of σ_* = 114 ± 10 km s⁻¹. Then, following Equation (3) in Gültekin et al. (2009), we estimate an SMBH mass of $\mathrm{log}\ {M}_{\mathrm{BH}}/{M}_{\odot }=7.09\pm 0.15$, fully consistent with the value derived from the Hβ broad emission line. This consistency indicates that the eastern nucleus is located, within the scatter, on the M_BH–σ_* relation derived for normal and non-merging galaxies.

The SMBH in Mrk 739E is, on average, almost two orders of magnitude less massive than the typical values of nine nearby (ultra-)luminous infrared late-stage merging galaxies reported by Medling et al. (2015). Furthermore, they found that their sample of black holes does not follow the M_BH–σ_* relation given by McConnell & Ma (2013), and the SMBH sample is overmassive compared to the expected values based on scaling relations, suggesting that the major epoch of black hole growth occurs in the early stages of the merger. Since the SMBH in Mrk 739E lies on the M_BH–σ_* relation, we can speculate that it, in a similar way as the SMBHs reported in Medling et al. (2015), could grow most of its mass during the collision due to this assembly delay that allows mergers to move off the relation, departing from the M_BH–σ_*. This scenario supports the idea that the Mrk 739 system is in an early evolutionary stage, and the SMBHs have not started yet to grow due to the merger. To move off up to 3σ of the relation, the SMBH in Mrk 739E would need to grow by two orders of magnitude, requiring large (>10⁸ M_⊙) amounts of gas to be funneled down to an accretion disk, together with Eddington-limited accretion for a few tens to hundreds of millions of years (Medling et al. 2015). High-resolution observations of the molecular gas can lead us to resolve the amount of available gas that could be accreted onto the SMBHs, and thus understand the extent to which this might be possible, estimating the eventual values that can be reached by the SMBH mass in Mrk 739E.

The sphere of influence of a black hole is defined as ${r}_{\mathrm{infl}}={{GM}}_{\mathrm{BH}}/{\sigma }_{* }^{2}$ (Merritt 2004), where G is the gravitational constant and ${\sigma }_{* }^{2}$ is the velocity dispersion of the stars of the bulge. Since the eastern SMBH has a mass of $\mathrm{log}{M}_{\mathrm{BH}}/{M}_{\odot }=7.22\pm 0.25$, the sphere of influence of the eastern SMBH is r_infl = 5.57 ± 2.80 pc, which is unresolved with our MUSE observation. These scales are significantly (approximately three orders of magnitude) smaller than the mass and radius of the rotating disk, and hence we conclude that the latter is mostly unaffected by the gravitational field of the SMBH at the center.

The mass of the SMBH in the western nucleus cannot be determined using our MUSE data and the scaling relation from Equation (2). This is because the broad Balmer lines are not visible, likely due to a combination of obscuration and dilution by the host galaxy (bottom-right panel in Figure 3). Based on the X-ray luminosity of Mrk 739W, L_{2–10 keV} = 1.0 × 10⁴² erg s⁻¹, the expected extinction-corrected Hα luminosity, mostly from the BLR, is ${L}_{{{\rm{H}}}_{\alpha }}\sim 1\times {10}^{41}\ \mathrm{erg}\,{{\rm{s}}}^{-1}$, via the Panessa et al. (2006) relation. With an extinction of A_V ∼ 1–2, the expected observed luminosity of the BLR in Mrk 739W would be ${L}_{{{\rm{H}}}_{\alpha }}\sim 1.6\mbox{--}4\times {10}^{40}\ \mathrm{erg}\,{{\rm{s}}}^{-1}$. Since Hα emission in Mrk 739W is extended and slightly more luminous (${L}_{{{\rm{H}}}_{\alpha }}\sim 6.2\,\times {10}^{40}\ \mathrm{erg}\,{{\rm{s}}}^{-1}$), we expect an indistinguishable BLR given the extinction reported in Section 5.1. Higher angular resolution spectroscopy, e.g., from MUSE NFM observations, might resolve the extended narrow line and better recover any possible point-like broad component. Also, it would be useful to obtain near-IR spectroscopy to study the hydrogen Paschen series transitions (e.g., Landt et al. 2013; La Franca et al. 2015) and test whether there is a hidden BLR, though these are rare (e.g., Lamperti et al. 2017; Onori et al. 2017). Given the fact that Mrk739W is optically obscured but shows X-ray emission, the X-ray luminosity along with near-infrared emission lines might also be a promising path to measure the SMBH mass (Ricci et al. 2017b). Assuming that the Mrk 739W SMBH, as Mrk 739E, lies on the M_BH–σ_* relation, we can obtain a rough estimate of its mass. With an average stellar velocity dispersion of σ_* = 91 ± 5 km s⁻¹ within an aperture of 09, we obtain an SMBH mass for the western nucleus of $\mathrm{log}{M}_{\mathrm{BH}}/{M}_{\odot }=6.68\,\pm 0.045$, adopting the M_BH–σ_* relation of Gültekin et al. (2009).

For a bolometric luminosity of 1 × 10⁴⁵ erg s⁻¹ and an SMBH mass of $\mathrm{log}{M}_{\mathrm{BH}}/{M}_{\odot }=7.04\pm 0.4$, Koss et al. (2011) found that the Eddington ratio, defined as λ_Edd = L_bol/L_Edd, ¹⁶ is λ_Edd = 0.71 for Mrk 739E. For the SMBH mass value of $\mathrm{log}{M}_{\mathrm{BH}}/{M}_{\odot }=7.22\pm 0.25$ that we derived, we estimate a λ_Edd = 0.48. Assuming that the SMBH mass of Mrk 739W obtained from the M_BH–σ_* relation is correct, the Eddington ratio for Mrk 739W is λ_Edd = 0.033 for a bolometric luminosity of 2 × 10⁴³ erg s⁻¹ derived in Koss et al. (2011). Hence, the SMBHs seems to be in different accretion scenarios. While our measurement of the Eddington ratio of Mrk 739E is slightly lower than the value reported by Koss et al. (2011), it is still one of the highest ratios among the Swift BAT AGNs (Vasudevan et al. 2010). On the other hand, the Eddington ratio of Mrk 739W is consistent with the median λ_Edd value of the Swift-BAT AGN Spectroscopic Survey (Koss et al. 2017). High-resolution infrared imaging in the K (2.2 μm) and L bands (3.8 μm) and red K–L colors, sensitive to buried AGNs (Imanishi & Saito 2013), could not find evidence for an AGN in Mrk 739W, suggesting nonsynchronous mass accretion onto SMBHs (Van Wassenhove et al. 2012), consistent with the Eddington ratio estimations for Mrk 739E and Mrk 739W.

The VLT/MUSE observations presented here can provide valuable insights into the current merger stage of the dual AGN, Mrk 739. Morphologically, we present a schematic illustration of the dual AGN in Figure 15. Mrk 739W resembles a spiral-barred galaxy, with spiral arms that spatially overlap with Mrk 739E. This overlap likely explains the double narrow-line component region (top-right panel in Figure 3), where the redshifted emission-line component is likely associated with Mrk 739W's bar, and the blueshifted one would come from Mrk 739E's disk region. The eastern side of the bar is the origin of two spiral arms, one of them extending up to the western star-forming region, while the other is morphologically related to the spatially extended region to the north side of Mrk 739, the so-called spiraling crest. The southern arm extends up to the eastern side of Mrk 739E.

Zoom In Zoom Out Reset image size

Dynamically, the common velocity between the eastern tidal tail (light-blue in Figure 15) and Mrk 739W in the ionized gas suggests that the tail can be associated with the western galaxy rather than with the eastern one. The spiraling crest with velocities of ∼60 km s⁻¹ (Figure 8) is associated with the leading arm of the northern spiral arms of Mrk 739W. The stellar kinematics suggest that Mrk 739W is mostly face-on, with a stellar velocity gradient of only ∼60 km s⁻¹ from north to south (bottom-left panel in Figure 12).

Since we only distinguish Mrk 739E by its rotating disk of ionized gas and its old, blueshifted, and metal-poor population on the southeastern edge of the galaxy, we claim that the young stellar populations of Mrk 739W are hiding and extinguishing the eastern galaxy from a foreground position, making Mrk 739E nearly absent on luminosity-weighted maps close to the center of the system, and highly extincted with values of A_V ∼ 3. The high stellar velocity dispersion region on the eastern side of the eastern nucleus thus represents an interface region where both galaxies appear in projection.

The morphology and dynamics of the dual AGN Mrk 739 are hence consistent with an early stage of the collision, where the foreground galaxy is a young star-forming galaxy that is interacting with its background elliptical companion. Since Mrk 739W's AGN does not show evidence of being actively accreting, as we discussed in Section 5.2, we propose that the nuclear activity of Mrk 739E has been the main ionizing mechanism of the northwestern spiral arms, similar to the phenomenon known as "Hanny's Voorwerps," as seen for example in IC 2497 (see, e.g., Józsa et al. 2009; Lintott et al. 2009; Keel et al. 2012; Sartori et al. 2016). This scenario suggests that the eastern AGN ionized the northern structures of Mrk 739W from its background position.

We emphasize the need for high-resolution multiwavelength and spatially resolved spectroscopic data, covering optical, near-IR, and submillimeter wavelengths, to dynamically model the system in order to understand the exact evolutionary stage of this complex merging galaxy. In particular, this data set would allow us to characterize the behavior of the gas that is actively feeding both AGNs, the distribution of the dense and cold gas close to the nuclei that work as a reservoir for future accretion onto both SMBHs, the stellar structures that could be produced by the interaction, and stellar formation related to the western galaxy.