A host galaxy study of southern narrow-line Seyfert 1 galaxies

We studied seven nearby narrow-line Seyfert 1 (NLS1) galaxies in J - and Ks -bands with redshifts varying from 0.019 to 0.092. This is the first multi-source study targeting the hosts of southern NLS1 galaxies. Our data was obtained with the FourStar instrument of the 6.5 m Magellan Baade telescope at the Las Campanas Observatory (Chile). The aim of our study is to determine the host galaxy morphologies of these sources by using GALFIT. We were able to model six out of the seven sources reliably. Our conclusion is that all of the reliably modelled sources are disk-like galaxies, either spirals or lenticulars. None of these sources present elliptical morphology. Our findings are in agreement with the hypothesis that disk-like galaxies are the main host of jetted NLS1 galaxies. Taking advantage of observations in two bands, we also produced a J − Ks colour map of each source. Five of the six colour maps show significant dust extinction near the core of the galaxy – a feature often seen in gamma-ray-detected jetted NLS1 galaxies, and interpreted to be a consequence of a past minor merger.

NLS1 galaxies are intrinsically different from broad-line Seyfert (BLS1) galaxies, for example, in their black hole masses, Eddington ratios, and their multiwavelength and variability properties.The general consensus is that, unlike BLS1 galaxies, NLS1 galaxies are powered by low-mass central supermassive black holes (M BH ∼ 10 6 M ⊙ -10 8 M ⊙ ; e.g.Peterson 2011;Järvelä et al. 2015;Cracco et al. 2016;Chen et al. 2018).This has been confirmed, for example, by reverberation mapping campaigns (e.g.Wang et al. 2016;Du et al. 2018;Du & Wang 2019).Due to the low black hole mass, NLS1 galaxies are thought to be in an early evolutionary stage with respect to BLS1 galaxies and other classes of AGN (Mathur 2000).They also exhibit high Edding-⋆ irene.varglund@aalto.fi⋆⋆ Dodge Family Prize Fellow in The University of Oklahoma ton ratios, generally between 0.1 and 1, but in some cases also exceeding unity (Boroson & Green 1992;Marziani et al. 2018;Tortosa et al. 2022).
NLS1 galaxies have previously been thought to not have significant radio emission (Komossa et al. 2006;Järvelä et al. 2015).Of the currently identified NLS1 galaxies, roughly one sixth, ∼ 16%, have been detected at radio frequencies: ∼ 6% are weak radio emitters, and ∼ 10% have significant radio emission (Komossa et al. 2006).The rest of the NLS1 galaxies, ∼ 84%, have not been detected at radio frequencies, yet.
Finding disk galaxies, such as BLS1 galaxies, with jets is not uncommon (Keel et al. 2006;Bagchi et al. 2014;Wu et al. 2022;Gao et al. 2023), however, these sources in general have stellar masses similar to those of elliptical galaxies and are thus atypical among disk-like galaxies.NLS1 galaxies, on the other hand, have pseudo-bulges and low Sérsic indices, indicating that their stellar masses are low (Sánchez Almeida 2020).Thus the discovery of relativistic jets in some NLS1 galaxies was unexpected.These jets appear similar, but less powerful, to those in blazars (e.g.Komossa et al. 2006;Zhou et al. 2006;Yuan et al. 2008;Foschini 2011;Foschini et al. 2015;Lähteenmäki et al. 2017Lähteenmäki et al. , 2018)).Typically, relativistic jets manifest strongly in radio and dominate the whole band.However, also examples of jetted sources with extremely weak low-frequency (< 10 GHz) radio emission have been found (Järvelä 2018;Berton et al. 2020;Järvelä et al. 2021;Sbarrato et al. 2021, Järvelä et al. in prep.), questioning the utilisation of the emission in this regime as a proxy for jettedness.The opposite has also been observed, when sources with significant radio emission at low frequencies do not Notes.Columns: (1) Source name, (2) Short name of the source, (3) Redshift, (4) Scale at the redshift of the source, (5) Right ascension, (6) Declination, (7) Total exposure time in J-band, (8) Total exposure time in K s-band, (9) Observation dates, (10) Average seeing during the observations measured from the images in the band used for modelling.
host relativistic jets, even though that would be expected (Järvelä et al. 2022).This anomaly is caused by the presence of significant star formation activity (Caccianiga et al. 2015;Ganci et al. 2019;Berton et al. 2020).
In this paper we examined the hosts of seven southern NLS1 galaxies that are also part of an optical spectral study and a multifrequency radio imaging study.This is the first systematic host galaxy investigation of southern NLS1 galaxies, and will be expanded in the future to allow statistically significant comparison of their radio and optical properties.Our goal was to determine the morphology of the hosts to improve the statistical weight of the recent findings that most NLS1 nuclei are hosted in disklike galaxies.Furthermore, we also studied the colour maps as they could tell us more about the properties of the core, such as its dust content, and shed light on possible merger history.Ultimately, NLS1 galaxies as early-stage AGN, and conveniently residing in the local Universe, can elucidate the evolution of AGN over cosmic time.

Observations and data reduction
Our sample of seven sources was selected from Chen et al. (2018).Our selection criteria were based on the sources' radio properties and the possible presence of jets (Chen et al. 2020).The sources chosen for this study have all been detected at 5.5 GHz, with five of them also having been detected at 1.4 GHz.We have presented the radio flux densities as well as some basic characteristics of our earlier observations in Table 2.The radio spectrum of five of the sources is steep, while the final two sources have a flatter spectrum.Based on the results of Chen et al. (2020), J0952-0136, J1511-2119, and J1615-0936 were expected to harbour relativistic jets.J0952-0136 has been identified as a jetted source also by Doi et al. (2015).To investigate the ra-  (Berton et al. in prep).All of our sources have a low redshift (< 0.1) and were observed with FourStar, mounted on the 6.5 m Magellan Telescope.Five sources were observed in July, 2019, while the rest were observed in February, 2020.One source, J0952-0136, was saturated and could not be modelled.
The field-of-view (FoV) of FourStar is 5.4' × 5.4', with a scale of 0.159 ′′ px −1 .All of our sources were observed in both J-and K s-bands.When observing in J-band, we used 4-point dithering for all sources but J0622-2317 for which we used 6point dithering due to its brightness.In the K s-band the dithering used was 6-point for all sources.
For data reduction, we used IRAF.We first created a flat field image for each source for each respective band, to remove eventual signatures in the data produced by the instrument.This flat was applied to each dithered image.After creating the flat field, we combined the dithered images to produce one image per each dithering sequence.Then we continued by creating a superflat to remove the final effects of the background.As the result, the mean sky value is essentially zero.To finalise the reduction, we aligned the created images and combined them.The data reduction process was done for both filters.After producing the final images, we determined the zero point magnitude for each source in both bands using 2MASS 1 magnitudes.

Data analysis
To perform the photometric decomposition of the host we used the band with better seeing (see Table 1), which in all cases of this study resulted to be the J-band.K s-band images were only used to create the colour maps.Our goal was to determine the host galaxy morphology and the related model parameters that best describe the source.
To perform the photometric decomposition, we followed the same procedure as described in Varglund et al. (2022) with a few exceptions.The PSF model was built primarily with DAOPHOT and secondarily through choosing a star from the image.No separate PSF stars were observed for this project.Furthermore, our cutout image sizes were either ∼ 100 × 100 px or ∼ 200 × 200 px.The size of the image was dependent on the cutout size around the host galaxy we were modelling.The sky background parameter was estimated through two methods: the method described in Varglund et al. (2022) and by using a large annulus that encompassed roughly a fifth of our image.We made sure that no source was inside the annulus, before determining the sky background value.The former method was used to confirm the results of the latter method.Due to the used data reduction process, the sky background value was virtually zero.
The fitting process with GALFIT version 3 (Peng et al. 2010), creating the radial surface plot with the IRAF (Stetson 1987) task ELLIPS E, and the error estimation method was identical to the process in Varglund et al. (2022).The radial surface plot was created to study both the accuracy of the fit as well as to see how the brightness of the components changes as a function of the distance from the core.On average the limiting surface brightness of the images is around 22-23 mag arcsec −2 whereas all of our sources are considerably brighter than the limit, and thus we can model them reliably up to large radii.
In the case of a suspected bar-component based on the bestfit parameters we used an additional diagnostic to determine if the component truly is a bar.To do this, we extracted the position angle (PA) and ellipticity parameters of the isophotes obtained from the ELLIPSE output of the galaxy image.These parameters were plotted against the radius.The plot indicates the presence of a bar if there is a region in which the ellipticity increases while the PA remains mostly constant (∆PA < 20°) (Wozniak et al. 1995;Menéndez-Delmestre et al. 2007).
Finally, since we had images from two bands for each source, we were able to create colour maps of all of our sources.To obtain the colour map, we used the equation below: The data used when calculating the colour image was corrected for the different exposure times of the bands.The images were aligned by using GEOMAP, GEOTRAN, and IMALIGN to guarantee that the images were perfectly aligned.The data processing allowed for the PSFs to be nearly identical, improving the colour maps.In the above equation, zp stands for the zero point magnitude, with the subscript defining the band in question.The contour levels in the maps are defined using the Ksband image sky standard deviation (std) and are at 20, 40, 60, 80, 100, 180, 360, 720, and 1440 × std.

J0622-2317
According to Chen et al. (2020) this source has a flat in-band radio spectrum, and a diffuse radio morphology.Additional information of this source and all the sources of our sample can be found in Table 2.A visual inspection shows a very clear spiral structure and even hints at the existence of a bar (Fig. 1).
The size of the fitting region was limited due to a nearby, very bright, saturated star.The best-fit results can be seen in Table 3.In addition to the NLS1, also three nearby sources were fitted using the PSF model to obtain a more realistic χ 2 ν for the FoV.We have most likely fitted a disk, or a combination of a disk and a pseudo-bulge component that cannot be separated and modelled individually.The bulge may be unresolved, but its light is still present in the image, thus affecting the model, or the bulge is either not bright enough or does not have different enough properties compared to the disk.The component is small, however, the Sérsic index is high for a disk so it is possible that a bulge component is included.The intensity plots of the source can be seen in Fig. 1.There are clear residuals left due to the spiral arms and some slight residuals near the core of the galaxy.We tried to include a bar into the model, but were unable to obtain good results.
The radial surface plot can be seen in Fig 2 .As expected, the model image curve does not perfectly follow the curve for the galaxy image.This is most likely due to the spiral arms not being completely modelled.The colour map can be seen in Fig. 3. On average, Seyfert galaxies have a J -K s magnitude between 0 and 2.5, and when the magnitude difference is 1.3 or more, a galaxy is considered red (Jarrett 2000).Unlike typical Seyfert galaxies, J0622-2317 is blue.However, the magnitude of the core of the galaxy is above 0, and is thus within a common range found in NLS1 galaxies.As with the other images, the spiral arms are clearly visible in the colour map.

J1511-2119
This is one of the three jetted sources in this study.Its radio map can be found in Chen et al. (2020).We used a DAOPHOT PSF model when fitting.There were ten nearby sources in our FoV, all of which we were able to fit with a single PSF with very little residuals.
Our fit models a combination of a disk and a bulge (Table 4).The Sérsic index of 2.96 does not strongly support any specific morphology, even though indices this high can be found also in spiral galaxies (Elmegreen & Elmegreen 2014).The size of the galaxy is typical for an NLS1 host.The excess residuals (Fig. 4) near the core imply the presence of unmodelled structures, such as leftover fragments of the bulge.The galaxy, model, and Sérsic curves in the radial surface brightness plot, Fig. 5, are all very similar to each other.
The host galaxy of J1511-2119 is disk-like, and based on the Sérsic index, an S0 galaxy.The colour map, shown in Fig. 6, shows colours commonly seen in NLS1 galaxies, with the red most likely signifying dust extinction.The core of the galaxy is extremely red as the J-K s magnitude is ∼2.Beyond the core, the galaxy has very uniform colours.

J1522-0644
We successfully used the PSF model to fit the five nearby sources.The source, model, and residual images can be seen      in Fig. 7.The best-fit parameters are given in Table 5.Along with a PSF, we used two Sérsic functions for fitting this source.
Based on the effective radius of Sérsic 1 it is presumably fitting a disk.The Sérsic index is low for a disk, but not unreasonable.
Based on its parameters, the second Sérsic function most likely models a bar.To further study this possibility, we plotted the PA and ellipticity over the radius, seen in Fig. 10.The PA remains steady throughout the plot.The ellipticity increases toward the higher radii.Based on this plot, it is difficult to confirm whether or not Sérsic 2 truly models a bar.The remaining residuals of the source are mostly remnants of the spiral arms.
The χ 2 ν is relatively high due to the remaining residuals, but the fit is satisfactory and the radial surface plot, Fig. 8, supports the goodness of the fit.The galaxy and the model curves match.The colour magnitude, seen in Fig. 9, of the galaxy is as expected in an NLS1 galaxy.The J-K s magnitude near the core is approximately 1.75, and is thus very red.The redness of the core is most likely due to dust extinction.

J1615-0936
We tried to model the PSF with DAOPHOT and by directly using a nearby star.Based on our results, the DAOPHOT PSF provided a better model.The unsatisfactory fit of the nearby source increases the χ 2 ν of the fit.We modelled the source with a PSF, a Sérsic function, and an exponential disk function.The best-fit parameters can be found in Table 6.Despite several attempts and different approaches, the χ 2 ν remained quite high.Based on our best-fit results, the Sérsic represents a pseudo-bulge and the exponential disk function naturally fits a disk.There are some clear residuals left, as seen in Fig. 11.The residuals of the NLS1 are most likely caused by the bright AGN that was hard to model due to the lack of a bright enough PSF star.
The χ 2 ν of the fit is not ideal, however, it does not always imply that the fit is physically unreasonable (Dewsnap et al. 2022).Indeed, based on the radial surface brightness plot, seen in Fig. 12, our model is accurate, as there is virtually no deviation between the model and galaxy curves, and the obtained components are physically reasonable.Based on our results, the host galaxy of J1615-0936 is disk-like, resembling a lenticular galaxy.This is supported by the axial ratio of the disk component being so small, 0.29 +0.00 −0.00 .The colour map of J1615-0936 is presented in Fig. 13.Based on it, the core has a J-K s colour magnitude of roughly 1.5, which means that the core of the galaxy is very red.The rest of the galaxy presents uniform colours.Overall, the colours of the galaxy are as would be expected in an NLS1 galaxy.

J1628-0304
We used a Sérsic function, an exponential disk function, and a PSF to fit the galaxy.The best-fit parameters are shown in Table 7. DAOPHOT provided the best PSF model, but it is not perfect, as evidenced by the remaining residuals, seen in Fig. 14, near the core of the galaxy.The Sérsic function models a bar component.The galaxy is small, also proven by the small effective radius of the exponential disk component.We plotted the ellipticity and PA against the radius of the isophotes, seen in Fig. 17, to study the possible bar.In the plot the ellipticity can be seen increasing while the PA is more or less constant.The plot strongly suggests that the Sérsic function indeed models a bar component.
The χ 2 ν for this source is quite high.However, there are virtually no high surface brightness residuals left.The remaining residuals are likely caused by the imperfect PSF model, which we could not improve despite several attempts with different methods.A major issue with the PSF modelling was the lack of bright enough stars in the FoV -none of them are brighter than the AGN, and probably cannot model the wings of the PSF accurately.However, as discussed earlier, the χ 2 ν value does not alone determine if the fit is physically justifiable and we want to emphasise the importance of taking into account also the obtained components, the residuals, and the comparison between the galaxy and and the model, in Fig. 15, when evaluating the fit.Based on all these aspects, the obtained results are adequate.The host galaxy of J1628-0304 is disk-like.
The colour map can be seen in Fig. 16.The colours of this galaxy are interesting as the reddest part is not at the core, but rather surrounds it.The structure could be due to a star forming ring or a minor merger (Iodice et al. 2014).However, the resolution of our images does not allow a detailed view of the possible minor merger and due to this, the feature is most likely a star forming ring.Circumnuclear star forming rings have been frequently observed in NLS1 galaxies (Winkel et al. 2022).Nonetheless, the core has colour magnitude of ∼1.5 and is thus very red.

J1646-1124
This source has a high χ 2 ν value (Table 8).Despite this, when studying the residuals, seen in Fig. 18, and the best-fit parameters, the fit is acceptable.A PSF was fitted on the five nearby sources as well as on the galaxy.In addition, the galaxy required one Sérsic function, with the function most likely modelling a bulge.The Sérsic index is high for a spiral galaxy, though still within the range of accepted values for them.We attempted fitting additional components, but could not achieve a physically justifiable fit.Our issues mostly stem from the AGN being so bright.A fit with three Sérsic functions yielded a better χ 2 ν ( 2), but one of the components consistently remained unresolved, likely trying to fit parts of the point-like AGN.The resolved components modelled a bulge and a disk component.The alternativefit results can be seen in Table 9.
Some residuals still remain near the galaxy core, and are most likely due to the issues with the PSF model.There is indication of a spiral-like structure surrounding the galaxy, starting from its west side, and bending anti-clockwise toward east.However, without further proof, the assessment of the residuals remains a speculation.Due to the χ 2 ν value, the residuals, and the best-fit parameters, we do not deem it possible to irrefutably determine the host galaxy of J1646-1124.However, there is evidence, for example, in the alternative fit, that J1646-1124 is disklike, and based on the parameters, an S0.
The radial surface plot, seen in Fig. 19, shows that despite the difficulties in the modelling, the model fits the galaxy very well up to the radius of 6 kpc.The colour map of the source can be seen in Fig. 20.The colours are as expected for an NLS1 galaxy, with the core being very red (∼2.2).The redness is most likely due to dust extinction.Some slight saturation, resulting in the polygonal shape, due to the K s-image, can be seen in the colour map.

Discussion
We were able to model six out of the seven sources in our sample.The only source we could not model was J0952-0136.A summary with the final verdicts of each source can be found in Table 10.Two of our modelled sources, J1511-2119 and J1615-0936, have been speculated by Chen et al. (2020) to be jetted based on their radio spectra.All six of our sources have disklike hosts.The only uncertain host morphology is for J1646-1124, however, our results still suggest the possibility of disklike morphology.Three of our sources show evidence for being in lenticular galaxies.However, it is unclear if the morphology truly is lenticular or if the results are caused by the AGN being too bright to have a reliable PSF model from the FoV.We were able to model a disk component in five of our sources, which naturally indicates disk-like morphology of the galaxy.A bar component can be found in two sources, J1522-0644 and J1628-0304.Based on the PA and ellipticity versus the major axis of the isophote plots, J1628-0304 is very likely to host a bar.We were able to model a bulge component in four sources.The modelled bulges have a slightly high Sérsic index to be considered pseudo-bulges in disk-like galaxies.The commonly used
limit for distinguishing between pseudo-and classical bulges is n = 2, where indices below 2 are associated with pseudo-bulges and indices above with classical bulges (Orban de Xivry et al. 2011).However, the sources that have bulges with Sérsic indices above 2 are the sources we suspect of being S0s and it has been suggested that S0/a sources can have pseudo-bulges with Sérsic index ≈ 2.4 with a standard deviation of 0.66 (Kormendy & Kennicutt 2004).With this in mind, the modelled bulges are most-likely pseudo-bulges, which is in agreement with what is seen in disk-like galaxies.
Only a few studies on NLS1 galaxy stellar masses exist, and we were able to find a stellar mass estimate for three of our sources: J0952-0136 (M star ∼ 1.5 × 10 9 M ⊙ ) (Pan et al. 2021), J0622-2317 (M star ∼ 1.3 × 10 10 M ⊙ ) (Saulder et al. 2016), and J1511-2119 (M star ∼ 1.3 × 10 11 M ⊙ ) (Koss et al. 2021).Two of these sources have a low stellar mass, while J1511-2119 has a high stellar mass.The high stellar mass is not surprising as the galaxy is most likely an S0.The stellar mass of J0952-0136 being low is interesting as it is a jetted source, however, as we do not have information of the stellar mass of other jetted NLS1 galaxies, no further conclusions can be drawn.Based on the relation of the stellar mass, and the effective radius and Sérsic index of the NLS1 galaxies of this paper, the stellar mass of these sources is most likely low (Sánchez Almeida 2020).
When studying the evolution of AGN, the nearby environment, both group-and galaxy-scale, is a crucial factor to take into account.This is due to the nearby environment being capable of reshaping the galaxy as well as the galaxy dynamics.It has been hypothesised that jetted NLS1 galaxies and interaction/mergers might be linked (e.g.Järvelä et al. 2017;Olguín-Iglesias et al. 2020;Berton et al. 2021).However, there is no clear consensus on the topic yet.None of the sources in this paper present clear mergers or interaction.This could be for a variety of reasons, one of the possible reasons being the lack of resolution for seeing minor mergers.It has been shown before that minor mergers are capable of triggering nuclear activity, an example of this is IRAS 17020+4544.Although this source is a jetted NLS1 galaxy, there is no clear evidence of mergers in optical/NIR imaging, but there is clear evidence of a small companion galaxy based on molecular gas observations (Salomé et al. 2021;Järvelä et al. 2022).
It has also been speculated that γ-ray-emitting NLS1 galaxies have experienced more minor mergers than non-γ-NLS1 galaxies with significant radio emission (Olguín-Iglesias et al. 2020).The same study also stated that the γ-NLS1 galaxies have bulge colours that are significantly different from their hosts with red NIR colour, J − K ∼ 2, while the non-γ-ray NLS1 galaxies have bulge colours that are similar to the rest of the galaxy.The red nuclei is, according to the authors, most likely due to either the bulge showing enhanced star formation or the AGN being dust-embedded as a consequence of a past minor merger.Contradictory to the results in Olguín-Iglesias et al. ( 2020), all but one of the sources in this paper showcase the nucleus being significantly more red than the rest of the galaxy.The only exception to this is J0622-2317, however, it is possible that the colouration difference in this source is due to the spiral arms beginning directly from the core, thus "hiding" the core under the spiral arms.
In case the speculation by Olguín-Iglesias et al. ( 2020) is indeed correct, it is entirely possible that the colour of the core could be a hint of previous minor mergers also in our sources.However, unlike in Olguín-Iglesias et al. ( 2020), the nuclei of our sources do not present any clear difference in colours between jetted and non-jetted NLS1 galaxies.Based on the current understanding, it is impossible to say whether or not mergers are needed for triggering powerful relativistic jets in NLS1 galaxies.

Conclusions
We have studied six NLS1 galaxies that have been observed in J− and K s-bands.The observations were carried out with the FourStar NIR instrument.The goal of our study was to investigate what morphology the host galaxies of our sources have.We also created colour maps for all of our sources.These maps were created to help study the properties of the core of the source.Furthermore, we also wanted to study possible interaction in our sources.The main findings of our study are: 1.All of our sources are most likely located in disk-like galaxies, including the possibly jetted NLS1 galaxies.2. The colour maps show that the core of NLS1 galaxies strongly tends toward the red-side, with no significant difference between the cores of jetted and non-jetted NLS1 galaxies.3. The redness of the core is most likely due to either dust extinction or due to previous minor mergers.
With over 100 studied hosts of NLS1 galaxies, jetted and non-jetted, the host galaxies of all types of NLS1 nuclei can be stated to be predominantly disk-like.What is still mostly missing is the information regarding the host galaxies of NLS1 sources at high redshifts, and more observations will be needed to determine if their properties are similar to their low-redshift counterparts.Suitable instruments for these studies would be, for example, the Very Large Telescope as well as the JWST.Multiwavelength observations are crucial for understanding the role of mergers and interaction in jetted NLS1 galaxies.For this, observations with the Atacama Large Millimeter Array at sub-mm would be a key factor.NLS1 galaxies offer a unique opportunity to study the interplay of different phenomena, such as jets, outflows, and star formation, at an early stage of AGN evolution, and estimate the impact an active nucleus has on the galaxy evolution.

Fig. 2 :
Fig. 2: Radial surface brightness profile plot of J0622-2317.The blue line represents the galaxy component, the dashed red line shows the model, the dotted green line marks the PSF, and finally the dashed pink line shows the Sérsic component.The error of each component is shown by the shaded area surrounding each curve.

Fig. 8 :
Fig. 8: Radial surface brightness profile plot of J1522-0644.The blue line represents the galaxy component, the dashed red line shows the model, the dotted green line marks the PSF, and finally the dashed pink line shows the Sérsic component.The error of each component is shown by the shaded area surrounding each curve.

Fig. 10 :
Fig. 10: Position angle and ellipticity of the Sérsic 1 function plotted against the major axis of the isophote of J1522-0644.

Fig. 12 :
Fig. 12: Radial surface brightness profile plot of J1615-0936.The blue line represents the galaxy component, the dashed red line shows the model, the dotted green line marks the PSF, the dashed pink line shows the Sérsic component, the final component, the exponential disk, is shown with dotted light blue.The error of each component is shown by the shaded area surrounding each curve.

Fig. 15 :
Fig. 15: Radial surface brightness profile plot of J1628-0304.The blue line represents the galaxy component, the dashed red line shows the model, the dotted green line marks the PSF, the dashed pink line shows the Sérsic component, the final component, the exponential disk component is shown with dotted light blue.The error of each component is shown by the shaded area surrounding each curve.

Fig. 17 :
Fig. 17: Position angle and ellipticity of the Sérsic function plotted against the major axis of the isophote of J1628-0304.

Fig. 19 :
Fig. 19: Radial surface brightness profile plot of J1646-1124.The blue line represents the galaxy component, the dashed red line shows the model, the dotted green line marks the PSF, the dashed pink line shows the Sérsic 1 component, the final component, Sérsic 2, is shown with dotted light blue.The error of each component is shown by the shaded area surrounding each curve.

Table 1 :
Basic properties of the sample and the observations.

Table 2 :
Radio properties of the sample (F = flat, S = steep), and the black hole mass estimates.