SDSS-IV MaNGA: How Galaxy Interactions Influence Active Galactic Nuclei

The source of fueling for active galactic nuclei (AGN) has been a major topic of debate for many years. The central supermassive black hole (SMBH) requires a constant stream of accreting material to sustain the AGN; however, the angular momentum of the material beyond the galaxy’s very center is too high to be accreted onto the SMBH. Some mechanism must be responsible for reducing this material’s angular momentum before it may become a fuel source for the AGN. It has been long suspected that galaxy interactions may be the primary mechanism of AGN fueling as AGN were often seen within interacting galaxy systems (Adams, 1977). Hydrodynamical simulations of galaxy mergers predict that the gravitational torques between the galaxies create large-scale bars which will funnel gases from the galaxies’ disks towards their centers (Barnes & Hernquist, 1996). These inflowing gases are predicted to trigger new star formation, dilute central metallicities, and potentially fuel the central SMBH.

More recent observational studies have confirmed that galaxy mergers do indeed enhance central star formation (Ellison et al., 2008; Kartaltepe et al., 2012; Patton et al., 2013; Steffen et al., 2021) and dilute central metallicities (Scudder et al., 2012), but the extent to which galaxy interactions influence AGN fueling is still a matter of debate. Some studies have found no correlation between galaxy mergers and AGN (Cisternas et al., 2011; Villforth et al., 2017; Marian et al., 2019; Silva et al., 2021) and other studies have found that AGN are statistically more likely to be in merging galaxies (Ellison et al., 2011, 2013; Lackner et al., 2014; Satyapal et al., 2014; Weston et al., 2017; Goulding et al., 2018).

These disagreements suggest that the methods for selecting galaxy pairs, control galaxies, and AGN need to be carefully considered. It has been previously observed that the AGN fraction in pairs depends strongly on the projected separation between the two galaxies and weakly depends on the stellar mass ratio between the two galaxies (Ellison et al., 2011). Further, late-stage mergers that are coalescing into a single galaxies also show increased AGN fractions (Ellison et al., 2013). This agrees with hydrodynamical simulations which predict that AGN activity generally increases with more advanced merger-stages but especially so as the pair coalesces into a post-merger galaxy (Van Wassenhove et al., 2012; Capelo et al., 2017).

Even while an AGN may be observed in a paired galaxy, the SMBH may be fueled through stochastic processes instead of merger-induced ones. To isolate the two effects, the AGN fraction in paired galaxies needs to be compared to the AGN fraction in isolated galaxies. It is well known that the AGN fraction in galaxies depends on their redshift and stellar mass (Kauffmann et al., 2003). Thus, the AGN excess, the ratio between the fraction of AGN in pairs and isolated galaxies, must be calculated between paired galaxies and control galaxies that have similar stellar masses and redshifts.

Different AGN selection methods also yield different levels of excess AGN in pairs. Ellison et al. (2011) showed that the AGN excess in paired galaxies is higher when optically selecting for composite starburst/AGN galaxies than it is for optically selecting pure AGN galaxies. It was also shown in Satyapal et al. (2014) that the AGN excess nearly doubles by using infrared color cuts to select AGN in comparison to using optical selected AGN. The optical component of the AGN spectrum can be obscured by optically thick dust dredged up by the merger, thus suppressing the excess of optically selected AGN.

Along with the sample selection, the biases of the surveys’ themselves need to be considered. Many of the past works using optically selected AGN have used single fiber spectroscopic surveys, like SDSS (Sloan Digital Sky Survey; Blanton et al. (2017)). Surveys like SDSS have allowed for large samples of paired galaxies; however, such large surveys have some limitations. SDSS has a fiber collision limit of 55″, so that close galaxy pairs will only be found in regions between overlapping tiling plates and in repeated observations. SDSS also only observes targets whose brightness is greater than a r-band apparent magnitude of $r$ $<$ 17.8. These limitations create artificial biases in pair, control, and AGN samples constructed within such surveys. Single fiber surveys also have fixed angular apertures that will cover variable physical radii depending on the galaxy’s redshift. Therefore, the spectra collected from the center of a high redshift galaxy may be diluted by the spectra from its disk.

The advent of massive Integral Field Spectroscopy (IFS) surveys resolves a number of the previously mentioned issues. IFS surveys like MaNGA (Mapping Nearby Galaxies at Apache Point Observatory survey; Bundy et al. (2015)) allocate a bundle of fiber optic cables to a single target, giving complete spectroscopic coverage over the whole disk of the galaxy. This allows for the construction of physical apertures based on the target’s redshift instead of being limited to fixed angular apertures. Close galaxy pairs will also fall within the field-of-view of some observations, yielding spectroscopic data for both galaxies.

In the previous paper of this series, Fu et al. (2018) (hereafter; Paper I), we built a sample of 105 galaxy pairs and 14 dual AGN (dAGN) in the fourteenth public data release of the MaNGA survey (MaNGA DR14). To account for MaNGA’s survey biases, we compared the comoving volume densities of AGN in paired galaxies to the expected volume densities. The expected volume density of AGN are calculated from a sample of isolated control galaxies and depended on the stellar mass and redshift of the galaxies. We found that galaxy pairs hosting a single AGN (hereafter; an offset AGN) showed little evidence for an excess volume density of AGN while dAGN showed an excess volume density that increased from 3$\times$ to 6$\times$ at pair separations of 10$-$30 kpc and 1$-$10 kpc respectively.

The results of Fu et al. (2018) suggested that merger events do not enhance the incidence of AGN, except in cases of simultaneous AGN. We refer to the event of simultaneous AGN in a galaxy pair as a correlated AGN. There are a number of possible sources for correlated AGN. The first is from the synchronous fueling of AGN by tidally induced gas inflows, which is predicted to occur in limited windows of projected separation (Van Wassenhove et al., 2012). The second possible source may be widespread radiative shocks from the tidally induced gas inflows. The third cause may be from AGN cross-ionization where the AGN in one of the paired galaxies is able to photoionize the gases in the neighboring galaxy. In the first case, the observed AGN are true AGN, while in the later two cases, one or both of the AGN are false detections.

In a subsequent work, we built a more rigorous methodology for selecting pairs in MaNGA, which we used to study the radial profiles of star formation in galaxy pairs (Steffen et al., 2021) (hereafter; Paper II). Now, the MaNGA survey has finished its final round of observations, increasing the sample from 2618 galaxies in Paper I to 10,130 galaxies. Given the improved pair selection methodology and larger galaxy sample in MaNGA, in this work we revisit the question of AGN volume density in MaNGA galaxy pairs. The larger sample size reveals an excess volume density of offset AGN at close separations which was previously unobserved in Paper I. These new results allow us to separately model the effects of merger-induced AGN and correlated AGN in galaxy pairs.

The paper is organized as follows. In Section 2 we discuss the MaNGA survey which we use in this study. In Section 3 we discuss how we construct our pair sample, in Section 4 we discuss how we select AGN, and in Section 5 we discuss our control sample. In Section 6 we study the volume densities of AGN in our pair sample in comparison with a control sample. In Section 7 we study the [O iii] luminosity between AGN in our pair sample and AGN in our control sample. In Section 8 we discuss the results of this work and in Section 9 we summarize the results of the work. Throughout this work, we adopt the following $\Lambda$CDM cosmology; $\Omega_{m}$ $=$ 0.3, $\Omega_{\Delta}$ $=$ 0.7, and $h$ $=$ 0.7.

MaNGA (Mapping Nearby Galaxies at Apache Point Observatory) is an IFS survey which uses SDSS’s 2.5 m telescope (Gunn et al., 2006) along with two dual-channel BOSS spectrographs (Smee et al., 2013). The survey uses 17 integral field units (IFU) of sizes of 19, 37, 61, 91, and 127 fibers covering 12.5″, 17.5″, 22.5″, 27.5″, and 32.5″on the sky respectively (Drory et al., 2015; Law et al., 2015). MaNGA has a spectral coverage from 3600$-$10300 Å with a resolution of R $\sim$ 2000 and a point spread function of 2.5″ FWHM (Bundy et al., 2015).

With the final public data release (DR17, Abdurro’uf et al. (2021)) MaNGA has surveyed 10,130 galaxies which have been selected from a subset of 41,154 galaxies from the NASA-Sloan Atlas¹¹1NSA v1_0_1; http://www.nsatlas.org (Wake et al., 2017). The selected galaxies are chosen from a redshift range of 0.01 $<$ $z$ $<$ 0.15 and a luminosity range of -17.7 $<$ $\mathcal{M}_{\rm i}$ $<$ -24.0, where $\mathcal{M}_{\rm i}$ is the rest frame i-band absolute magnitudes within the survey’s elliptical Petrosian apertures. The survey is designed to have a flat stellar mass distribution and to have a spatial coverage of 1.5 $R_{\rm eff}$ and 2.5 $R_{\rm eff}$ (where $R_{\rm eff}$ is the 50% half light radius).

The MaNGA survey has three main subsamples; the Primary sample ($\sim$47% of the survey), the Secondary sample ($\sim$37% of the survey), and the Color-Enhanced sample ($\sim$16% of the survey). The Primary sample consists of the galaxies which are covered out to 1.5 $R_{\rm eff}$ while the Secondary sample consists of the galaxies which are covered out to 2.5 $R_{\rm eff}$. The Color-Enhanced sample is designed to target rare galaxies such as high-luminosity blue galaxies and low-luminosity red small galaxies. The Color-Enhanced sample is covered out to 1.5 $R_{\rm eff}$ and can be combined with the Primary sample to make the Primary+ sample. In this work we will be using both the Primary+ and Secondary samples which we will refer to as the Main sample. MaNGA also contains a number of ancillary observations that were specifically designed to target galaxy mergers and AGN. We decide not to use these ancillary targets as they will bias our sample.

We construct the luminosity function for the selected galaxies to check if it follows the expected luminosity distribution of galaxies. We find that galaxies above an r-band absolute magnitude of $\mathcal{M}_{\rm r}$ $\geq$ -19 are under-sampled in our selection. We decide to remove these faint objects from our sample to avoid incompleteness. The preceding restrictions give us 8,585 MaNGA IFUs to work with for this study.

Previous pair studies with MaNGA have found pairs by cross-matching MaNGA targets with galaxies in its parent catalog, the NSA catalog (Pan et al., 2019; Steffen et al., 2021). To utilize the volume density method (Described in Section 6), we will compile a pair sample that is entirely contained within the MaNGA survey’s 2392 arcmin² footprint. We first identify and spectroscopically classify discrete objects in the MaNGA survey’s footprint. We name this catalog the mangaObj and describe its construction fully in Steffen et al. 2022b (in preparation).

Briefly, we first overlay the SDSS photometric catalog, photoObj, over MaNGA’s fields-of-view. The cross-matched objects from the photoObj catalog are then visually inspected. SDSS’s object deblending algorithm frequently place multiple targets on clumpy extended objects such that a single galaxy will be assigned multiple targets. We remove these over-deblended objects and then visually classified the remaining objects with MaNGA’s spectroscopy. The catalog contains 11,072 galaxies that have redshifts which are similar ($z$ $<$ 0.01) to the MaNGA target galaxies or are MaNGA target galaxies themselves. 9,470 of these galaxies are within the 8,585 MaNGA IFUs that we have selected for this study.

We use the spectral fitting code, spfit²²2https://github.com/fuhaiastro/spfit, to model the spectra in MaNGA’s data-cubes (Paper I). spfit is designed to simultaneously fit stellar continuum and emission lines following the Penalized Pixel-Fitting method (pPXF; Cappellari & Emsellem (2004); Cappellari (2017)) using the simple stellar population (SSP) models from MIUSCAT (Vazdekis et al., 2012). We use the redshift of the MaNGA target, from the NASA-Sloan Atlas (NSA), as the best guess of the redshift for all objects in the MaNGA field. Objects with substantially different redshifts, $\Delta v\geq$ 2000 km s^-1, will fail to be properly modeled. spfit provides emission line fluxes, equivalent widths, and gas kinematics calculated from the fitted emission lines and stellar kinematics and stellar masses from the stellar continuum.

We use the galaxies with redshifts that are similar to the MaNGA target galaxy ($z$ $<$ 0.01) as candidates for our pair sample. We extract and model spectra for each object in the mangaObj catalog with spfit through a 1 kpc radius circular aperture centered on the galaxy’s photometric center. We use the kinematics and stellar masses from spfit in the subsequent pair selection.

The MaNGA survey imposes a natural projected separation limit of $\sim$85 kpc across the diameter of an IFU assuming the largest IFU size, 32.5″ and the survey’s upper redshift limit, $z$ $=$ 0.15. In most cases, the MaNGA target galaxy is in the center of the IFU, so the effective projected separation limit is $\sim$42 kpc. We require that the MaNGA target galaxy is within 10% of the radius of the IFU and that a paired galaxy be within the inscribed circle that is 1″ from the hexagonal IFU boundary. This eliminates any geometric biases that may arise from off-center MaNGA targets or objects falling in one of the IFU’s corners. With these geometric requirements, the projected separation range of our pair sample varies from 1 kpc $-$ 36 kpc.

We use the relative line-of-sight velocities of the galaxies to remove projected pairs from the sample. From spfit, we have stellar and gas kinematics for the fitted spectra. We chose to use the stellar kinematics for the relative velocity cuts since galaxies hosting AGN may have large offsets due to AGN outflows. We then set a relative line-of-sight velocity of $\Delta v\leq$ 500 km s^-1 to select kinematic pairs. A cut at 500 km s^-1 is subjective; however, in Paper I we observed that a more or less stringent requirement on the relative line-of-sight velocity has little impact on the resulting pair sample.

The MaNGA survey’s stellar mass measurements are derived from the NSA catalog. The algorithm in the NSA catalog frequently under-deblends close objects such that the stellar mass given for a MaNGA target in a pair will actually be the combined stellar mass of both galaxies. Separating this stellar mass between the two paired galaxies is not a simple task as these galaxies often overlap each other and have irregular morphologies.

As a simple attempt to resolve this problem, we use the stellar masses calculated from the inner 1 kpc radius circle from spfit to define our mass ratios. These mass ratios are then used to partition the stellar masses from the NSA catalog between the galaxies in pairs. The mass ratio, $\mu$, is defined as;

Where $M_{\rm t}^{SPFIT}$ and $M_{\rm c}^{SPFIT}$ are the stellar masses from spfit modeled spectra extracted from a 1 kpc circular aperture for the MaNGA target and its companion respectively. This mass ratio is then used to partition the stellar mass from the NSA catalog such that the MaNGA target’s stellar mass, $M_{\rm t}$, is calculated as,

and the stellar mass of the companion galaxy, $M_{\rm c}$, is then calculated as,

For the remainder of the paper, we will be referring to the mass ratio, $\mu$, in logarithmic space such that log$(\mu)$ $=$ log$(M_{\rm t}^{SPFIT}/M_{\rm c}^{SPFIT})$. For the pair selection, we impose a logarithmic mass ratio cut of $|$log($\mu$)$|$ $\leq$ 1.0 to remove insignificant companions from the sample.

The pair selection process can be briefly summarized as the following. We require that:

1. 1.

   the pair be in either MaNGA’s Primary+ or Secondary samples.

2. 2.

   the MaNGA target has a r-band absolute magnitude that is less than $\mathcal{M}_{\rm r}$ $\leq$ -19.

3. 3.

   each paired observation is entirely contained within a single IFU.

4. 4.

   the MaNGA target galaxy is within 10% of the radius of the MaNGA IFU from the IFU’s center.

5. 5.

   the galaxy pair be within the inscribed circle that is 1″ from the IFU boundary.

6. 6.

   the relative line-of-sight velocity between the galaxies be within $\Delta v\leq$ 500 km s^-1.

7. 7.

   the mass ratio between the galaxies be within $|$log($\mu$)$|$ $\leq$ 1.0.

Using the above criteria, we find 391 pairs, 35 triplets, 1 quadruplet, and 1 quintuplet within the 8,585 MaNGA IFUs. If we define a major merger as paired galaxies whose mass ratio is below $|$log($\mu$)$|$ $\leq$ 0.5 and a minor merger as paired galaxies whose mass ratio is between 0.5 $<$ $|$log($\mu$)$|$ $\leq$ 1.0, we find 178 major mergers and 213 minor mergers among the 391 paired galaxies. Going forward, we will focus on the galaxy pairs in order to simplify the resulting statistical analysis.

We classify AGN in our sample following the BPT (Baldwin et al., 1981) and WHAN (Cid Fernandes et al., 2011) methods. For the BPT method, we use the emission line flux ratios, log([O iii]/H$\beta$) and log([N ii]/H$\alpha$), extracted from a 1 kpc aperture over the centers of the galaxies. We classify AGN from star formation on the BPT diagram using the empirical separation between star-forming galaxies and AGN from Kauffmann et al. (2003) (K03),

the theoretical maximum starburst line from Kewley et al. (2001) (K01),

and the empirical separation between Seyferts and LINERs from (Schawinski et al., 2007) (S07),

We determine a galaxy to be on the AGN branch if it is above the K03 line. We then split the AGN branch into further categories. We classify the galaxy as composite starburst/AGN if it is between the K01 and K03 lines. We classify the galaxy as LINER if it is above the K01 line but below the S07 line. We classify the galaxy as a Seyfert if it is above both the K01 and S07 lines (Figure 1).

It has been shown that the ionization spectra from hot low-mass evolved stars (HOLMES) can create line ratios which mimic other classifications on the BPT diagram, especially LINERs (Stasińska et al., 2008). Cid Fernandes et al. (2011) showed that the EW(H$\alpha$) is a good tracer for these HOLMES. An EW(H$\alpha$) cut of 3Å is typically used to remove retired galaxies; however, we decide to use a stricter EW(H$\alpha$) $<$ 6Å to remove wAGN (weak AGN). We find that using the stricter EW(H$\alpha$) $=$ 6Å cut improves the observed volume density of AGN in the pair sample.

We also visually inspect the spectra of each of the identified galaxies for broadened emission lines. These broadened emission lines are evidence of high velocity gases near the SMBH and are a guaranteed signature of AGN activity. Such galaxies are known as broad-line AGN (BLAGN) or Type I Seyferts. If a galaxy in our sample possesses broadened emission lines, we include it in our AGN sample, regardless of its BPT or WHAN classification.

In total, we have 1045 AGN out of the 9,470 identified galaxies, such that about 10% of our galaxy sample hosts an AGN. Of these, 713 are composite starburst/AGN galaxies, 97 are LINERs, 158 are Type II Seyferts, and 77 are Type I Seyferts. In the pair sample there are 105 AGN, of which 62 are composite starburst/AGN galaxies, 18 are LINERs, 17 are Type II Seyferts, and 8 are Type I Seyferts. Among the galaxy pairs with AGN, we find 25 dual AGN (dAGN) systems, which we list in Table 1, and we find 55 pair systems with a single AGN (hereafter offset AGN).

Since we want to isolate merger induced effects in galaxy pairs, we need a sample of control AGN in isolated galaxies to set a baseline. Our control sample consists of 7811 isolated galaxies that have no companions within the MaNGA IFU with a line-of-sight velocity separation of $\Delta$v $\leq$ 2000 km s^-1. Within these control galaxies there are 872 AGN of which; 613 are composite starburst/AGN galaxies, 70 are LINERs, 126 are narrow-line Seyferts, and 63 are BLAGN.

In order to study AGN rates between controls and pairs, we need to control for other factors which may influence stochastic AGN rates. It has been shown in previous works that the rate of AGN is related to the stellar mass and redshift of the host galaxy. The MaNGA survey has a limited redshift range, therefore, we do not expect that the AGN fraction in MaNGA will have a significant dependence on redshift. We represent the model AGN fraction, $f^{\rm mod}_{\rm AGN}$, as a function of stellar mass, $M$. We show the AGN fraction as a function of the stellar mass for the control and pair samples in Figure 2. We find that the AGN fraction as a function of the stellar mass can be fit well with a log-normal function.

Even though the AGN fraction will have no significant evolution in our redshift range, we include the redshift term, $z$, in our model. We fix the power-law index to 4 so that it is consistent with the AGN luminosity function for galaxies with $z\leq 1$ (Ueda et al., 2003). We then fit the parameters of the stellar mass’s log-normal function and find that $f_{0}=0.12$, $b=10.56$, and $\sigma=0.44$.

We previously calculated the AGN fraction in Paper I and found the following fit parameters, $f_{0}=0.22$, $b=10.57$, and $\sigma=0.54$. The amplitudes of the AGN fraction are substantially different between the works which is largely due to the stricter EW(H$\alpha$) cuts that we implement in this work. If we recalculate the AGN fraction while lowering our EW(H$\alpha$) cut to 3Å, we find that $f_{0}=0.19$. This value is consistent with our previously calculated value from Paper I. We can now use this model of the AGN fraction (assuming an EW(H$\alpha$) cut of 6Å) to estimate the expected volume densities of AGN in our pair sample.

This log-normal distribution for the stellar masses of AGN in MaNGA was previously seen in Paper I and Sánchez et al. (2022). The AGN fraction peaks around log($M/M_{\odot}$) $=$ 10.5$-$11.0 because of the selection biases imposed by the BPT+WHAN selection criteria (Section 4). AGN become rarer in lower mass galaxies because they host less massive black holes. If these black holes accrete material with the same Eddington ratio as the more massive galaxies, they will be less luminous than the black holes in the higher mass galaxies. These less luminous AGN will have a smaller ionized region which may be diluted in our 1 kpc radial aperture so that they may be missed by our BPT analysis. The AGN fraction may decline again at high stellar masses because high mass galaxies tend to be quiescent and tend to have lower gas fractions. A gas-poor galaxy hosting an AGN may be missed by BPT analysis since they will not produce strong emission lines. These selection biases further emphasize the importance of the control sample in our analysis.

The source of the powerful emission we observe from AGN is a result of material accreting onto the black hole. Unfortunately, direct measurements of the bolometric luminosity for AGN are difficult as they emit across several wavelength regimes and many of those wavelength regimes are obscured. In the optical regime, the [O iii] luminosity is often used as a proxy for the bolometric luminosity (Heckman & Best, 2014). The [O iii] line is useful as it is strong enough to be visible in most galaxies and it is linearly proportional to the bolometric luminosity.

Since the [O iii] luminosity can be used as a tracer of the black hole accretion rate, we want to use this line to see if the black holes in paired galaxies accrete materials at higher rates than those found in isolated galaxies. We extract our luminosities from the central 1 kpc of the galaxies and correct our luminosities for reddening using the reddening curve from Calzetti et al. (2000), assuming $R_{\rm V}$ $=$ 3.1 and case-B recombination, H$\alpha$/H$\beta$ $=$ 2.86. As we had shown in Figure 2, the fraction of AGN in our control sample is primarily a function of the galaxy’s stellar mass. We plot the [O iii] luminosity as a function of the stellar mass and redshift in Figure 7. We see that the [O iii] luminosity increases with higher stellar masses and the [O iii] luminosity is almost constant with redshift up to $z$ $<$ 0.12.

For a better comparison, we match each paired galaxy with control galaxies whose stellar masses are within 0.1 dex and whose redshift are within 0.01. We require that each paired galaxy have at least 20 control galaxies to match with and if they do not, they are excluded from the analysis. We then take the difference between the luminosity of the paired galaxy and the median luminosity of its mass and redshift-matched controls. For this analysis, we only use the MaNGA target galaxy in each pair and we use total stellar masses from the NSA catalog for both the control galaxies and the paired galaxies. We decide to use the total stellar masses from the NSA catalog for the paired galaxies because the MaNGA survey has a tight stellar mass to redshift distribution. The paired galaxies whose stellar masses are split by our mass ratios often fall off of the survey’s stellar mass$-$redshift distribution such that they find no similar control galaxies.

In Figure 8, we present the excess [O iii] luminosity of the AGN in paired galaxies with respect to the control sample. We take the mean excess luminosity in projected separation bins, which are 5 kpc wide, and calculate the vertical error bars with the bootstrap resampling method mentioned in Section 6. Below 15 kpc, we see that the AGN in pairs have a [O iii] luminosity that is 0.2 dex (1.6$\times$) higher than those in the control galaxies. This falls to zero beyond 15 kpc. Overall, we find that the [O iii] luminosity of the paired galaxies is consistent with their control galaxies. About half of the AGN in pairs show enhanced [O iii] luminosities, while the other half have lower luminosities than their controls. We do see that some paired galaxies have [O iii] luminosities that are 10$-$20 times brighter than their control galaxies, though we find no correlation between these pairs and other merger parameters (like stellar mass, redshift, mass ratio, or the presence of tidal features).

The relative [O iii] luminosities have also been studied in previous works. Ellison et al. (2013) found that the AGN in post-merger galaxies have [O iii] luminosities that are 0.9 dex (8$\times$) higher than the AGN in controls and that the AGN in paired galaxies have [O iii] luminosities that are 0.5 dex (3$\times$) higher than controls at separations of 5 kpc. This enhanced [O iii] luminosity falls to 0.3 dex (2$\times$) between 20$-$25 kpc. The [O iii] luminosity differences that we find are lower than those found in Ellison et al. (2013), by $\sim$0.3 dex. Jin et al. (2021) also compares the surface brightness of [O iii] between the AGN in pairs and the AGN in control galaxies using the MaNGA survey. They find that the [O iii] surface brightness is 0.13 dex higher in the AGN in pairs than in the AGN in isolated galaxies which is consistent with what we find, although, due to the wide errors in our work, we conclude that the [O iii] luminosity of the AGN in our paired galaxies is consistent with their control galaxies. Based on this, we infer that the black hole accretions rates between isolated and interacting galaxies are similar. This could mean that either merger-induced fueling occurs at a similar rate to stochastically induced fueling or that stochastic fueling is the dominant mode of AGN fueling even in our pair sample.

In this work, we identified a sample of 391 galaxy pairs within the fields-of-view of 8,585 MaNGA IFUs. We identified 105 AGN in the pair sample using optical BPT analysis. Among these, we found 25 dAGN systems. We found that galaxy pairs have a greater fraction of AGN than control galaxies in the MaNGA survey. We summarize the findings of this work below.

1. 1.

   We found that the AGN in galaxy pairs have a volume density that is 2.5$\times$ higher than what would be expected by random pairing at close separations. This excess of AGN disappears around 20 kpc. In dAGN, we found a volume density of AGN that is 40$\times$ higher than what would have been expected by random pairing at close separations. This excess of AGN falls to 1.6$\times$ at separations of 20 kpc.

2. 2.

   We model the observed AGN volume density in galaxy pairs to account to stochastic AGN fueling, merger-enhanced fueling, and correlated AGN activity. We demonstrate that excess AGN are induced in mergers not only from synchronous AGN activity, but also that the merger event can increase the natural rate of AGN fueling.

3. 3.

   Our model predicts that 11.5% of offset AGN will be converted to dAGN at a projected separation of 10 kpc. Our model further predicts that merger-induced fueling will enhance the rate of stochastic fueling by a factor of 1.27$\times$ at 10 kpc.

4. 4.

   We found that the [O iii] luminosity of the AGN in paired galaxies is consistent with the AGN in mass and redshift matched control galaxies. This demonstrates that galaxy interactions may increase the likelihood of activating an AGN, but they do not necessarily induce more luminous AGN than stochastically induced AGN.

In this work, we confirm our results obtained in (Paper I) in that we have been able to demonstrate that galaxy mergers are able to induce AGN activity in the centers of galaxies. We also show that the [O iii] luminosities of the AGN induced by galaxy interactions are consistent with the AGN in isolated galaxies. This indicates that stochastic fueling may still be the dominant fueling mechanism in close pairs. This also demonstrates how it is difficult to separate effects of tidally induced and stochastically gas-inflows in galaxy pairs.

We thank the anonymous referee for useful comments that helped improve the manuscript. J.S. and H.F. acknowledge support from the National Science Foundation (NSF) grants; AST-1614326 and AST-2103251. YSD acknowledges the support from the National Key R&D Program of China via grant No.2022YFA1605000, and the National Natural Science Foundation of China grants 12273051 and 11933003. ICG would like to acknowledge support from DGAPA-UNAM grant IN113320. Funding for the Sloan Digital Sky Survey IV has been provided by the Alfred P. Sloan Foundation, the U.S. Department of Energy Office of Science, and the Participating Institutions. SDSS acknowledges support and resources from the Center for High-Performance Computing at the University of Utah. The SDSS web site is www.sdss.org.

SDSS is managed by the Astrophysical Research Consortium for the Participating Institutions of the SDSS Collaboration including the Brazilian Participation Group, the Carnegie Institution for Science, Carnegie Mellon University, Center for Astrophysics — Harvard & Smithsonian (CfA), the Chilean Participation Group, the French Participation Group, Instituto de Astrofísica de Canarias, The Johns Hopkins University, Kavli Institute for the Physics and Mathematics of the Universe (IPMU) / University of Tokyo, the Korean Participation Group, Lawrence Berkeley National Laboratory, Leibniz Institut für Astrophysik Potsdam (AIP), Max-Planck-Institut für Astronomie (MPIA Heidelberg), Max-Planck-Institut für Astrophysik (MPA Garching), Max-Planck-Institut für Extraterrestrische Physik (MPE), National Astronomical Observatories of China, New Mexico State University, New York University, University of Notre Dame, Observatório Nacional / MCTI, The Ohio State University, Pennsylvania State University, Shanghai Astronomical Observatory, United Kingdom Participation Group, Universidad Nacional Autónoma de México, University of Arizona, University of Colorado Boulder, University of Oxford, University of Portsmouth, University of Utah, University of Virginia, University of Washington, University of Wisconsin, Vanderbilt University, and Yale University.