The nuclear region of NGC 613. I -- Multiwavelength analysis
MMNRAS , 1–18 (2020) Preprint 22 January 2020 Compiled using MNRAS L A TEX style file v3.0
The nuclear region of NGC 613. I - Multiwavelengthanalysis
Patr´ıcia da Silva (cid:63) , R. B. Menezes † , J. E. Steiner ‡ Instituto de Astronomia, Geof´ısica e Ciˆencias Atmosf´ericas, Departamento de Astronomia, Universidade de S˜ao Paulo, 05508-090, S˜ao Paulo, SP, Brazil Instituto Mau´a de Tecnologia, Pra¸ca Mau´a 1, 09580-900, S˜ao Caetano do Sul, SP, Brazil
Accepted 2019 December 28. Received 2019 November 22; in original form 2019 August 16.
ABSTRACT
In this paper, we report a detailed study with a variety of data from optical, near-infrared, X-ray, and radio telescopes of the nuclear region of the galaxy NGC 613 withthe aim of understanding its complexity. We detected an extended stellar emission inthe nucleus that, at first, appears to be, in the optical band, two stellar nuclei separatedby a stream of dust. The active galactic nucleus (AGN) is identified as a variablepoint-like source between these two stellar components. There is a central hard X-rayemission and an extended soft X-ray emission that closely coincides with the ionizationcone, as seen in the [O iii ] λ i ] λ ii regions, eight of them in a star forming ring that is visible in Br γ ,[Fe ii ] λ Key words: galaxies: active – galaxies: individual: NGC 613 – galaxies: nuclei
Active galactic nuclei (AGNs) are phenomena in which asignificant amount of energy is produced by a non-stellarsource. Such energy is believed to be produced by gas accre-tion onto a supermassive black hole (Lynden-Bell 1969; seeNetzer 2013 for more details). The study of the environmentaround AGNs is relevant for the evaluation of their interac-tion with the rest of the host galaxy through the processes offeeding and feedback (Storchi-Bergmann et al. 2009, 2010).The feeding process can be studied by observations of molec-ular emission in the near-infrared (NIR) or millimetre bands.The feedback can be studied, for instance, by observing op-tical emission lines associated with outflows (Harrison et al.2018) and ionization cones or by observing radio jets (Beall2003). The emission of the AGN itself can be observed fromradio to γ -rays. By analysing multiwavelength data it is pos- (cid:63) [email protected] † [email protected] ‡ [email protected] sible to disentangle the complexity of the AGN, its environ-ment, and their interaction.Statistical properties of AGNs are well derived fromstudies of large samples (e.g. Ho 2008; G¨ultekin et al. 2009).However, the interaction of an AGN with its environmentcan be quite complex and individual objects must be stud-ied in detail if we want to understand this complexity.In this work, we present a multiwavelength study of thenuclear region of NGC 613. This is an SB(rs)bc galaxy, lo-cated at 26 ± ii regions that has a radius of about 300 pc(Hummel & Jorsater 1992, Falc´on-Barroso et al. 2014 andAudibert et al. 2019).That the galaxy contains an AGN is confirmed, for ex-ample, by the detection of high-ionization emission lines,such as [Ne v ] and [O iv ], observed with the Spitzer spacetelescope (Goulding & Alexander 2009). Besides that, a ra-dio jet, showing a structure that extends along 5 arcsec with © a r X i v : . [ a s t r o - ph . GA ] J a n Patr´ıcia da Silva et al.
Figure 1.
RGB composition of the images in the filters
F 814 W (red),
F 606 W (green), and
F 450 W (blue) of
HST /WFPC2.The positions and sizes of each field of view (FOV) of GMOS,SINFONI, and SIFS data cubes are represented by the squares,with the orientation N-E. The PA of the GMOS observation is127 ◦ . The total size of the FOV of this image is 13.55 arcsec × position angle (PA) of 6 ◦ (almost perpendicular to the bar)was observed with the Very Large Array (VLA) at 6 and 20cm (Hummel et al. 1987). At 4.86 and 14.94 GHz, three blobswere detected in the central kpc, forming a linear structurewhose PA is 12 ◦ (Hummel & Jorsater 1992). Miyamoto et al.(2017) also detected an elongated structure in 95 GHz withPA= 20 ◦ ± ◦ . All these data show the presence of a nuclearradio jet. Associated with this jet, there is an anisotropicmaser with luminosity of ∼ L (cid:199) (Kondratko et al. 2006).There is also a megamaser (in the same position as the previ-ous one) with isotropic emission and luminosity of ∼ L (cid:199) ,which can also be associated with the nuclear jet (Castangiaet al. 2008).The AGN in NGC 613 is of low luminosity. Using an X-ray luminosity of L − kev = . × erg s − (Castangiaet al. 2013) and a bolometric correction of 20 (Vasudevanet al. 2010), Davies et al. (2017) determined that the AGNbolometric luminosity is 1.6 × erg s − . From the analy-sis of the X-ray Multi-Mirror Mission ( XMM-Newton ) data,Castangia et al. (2013) obtained an intrinsic column densityof 36 + − × cm − and an emission ratio of soft X-rays (2-10keV) to [O iii ] λ ± × cm − ,confirming the highly obscured AGN.The Atacama Large Millimeter/Submillimeter Array(ALMA) observation of molecular lines revealed thatthe nucleus has emission compatible with the ones ofSeyfert/LINER composite and is also being ionized by shockheating. There are very energetic molecular outflows thatare possibly fossils from a phase when this AGN was in ahigher activity and they are also associated with the radiojet, possibly being driven by it (Audibert et al. 2019).Through the decomposition of gas ionization mecha- nisms of the central ∼
35 arcsec ×
25 arcsec, by using aspectral basis that consider ionization by shock, AGN, andstars, an ionization cone aligned with the radio jet was de-tected (Davies et al. 2017). There are also shock waves in theionization cone edges (inside the central 1 kpc ), probablyformed by outflows of gas coming from the AGN that areshocking gas in the interstellar medium.It is believed that circumnuclear star-forming rings arethe result of the strong interaction between the bar (orstrong spiral arms) and the inner gas located in the Lind-blad resonance region (Elmegreen 1994). According to B¨okeret al. (2007, 2008) and Falc´on-Barroso et al. (2008, 2014),the circumnuclear ring of NGC 613 is formed by seven H ii regions. This ring can also be observed in radio wavelengths(Hummel & Jorsater 1992; Miyamoto et al. 2017, 2018; Au-dibert et al. 2019) and, if its structure is circular, the inclina-tion should be 55 ◦ ± ◦ (Hummel & Jorsater 1992). Combeset al. (2019) and Audibert et al. (2019) found that this ring isconnected to the bar in two points, at NW and SE, as B¨okeret al. (2008) previously estimated by using the Hubble SpaceTelescope (HST) images, and that the ring is indeed clumpyas it is shown in the image of Br γ (Falc´on-Barroso et al.2014). There is also a nuclear spiral of molecular gas con-nected with the ring in two different spots (Audibert et al.2019).In this article, we analyse the central region of NGC613 using data cubes obtained with the Gemini Multi-ObjectSpectrograph (GMOS) from the Gemini-South telescope andSoar Integral Field Spectrograph (SIFS) from the SOARtelescope and the archived data from the HST , from ALMA,from the Spectrograph for Integral Field Observations inthe Near Infrared (SINFONI) of the Very Large Telescope(VLT), and from the
Chandra
Space Telescope. The purposeof our analysis is to correlate our results with the informationand data from the literature in order to explain the natureof the observed emission in the centre of this galaxy. Thispaper (Paper I) is part of a comprehensive analysis of NGC613 nucleus and presents the study of the emission of thisgalaxy centre. The study of the stellar and gas kinematicsand stellar archaeology will be presented in Paper II.Section 2 describes the observations and treatmentmethods of the data obtained with the many instrumentsused in this work. Section 3 shows the optical emission-lineanalysis of the observed regions in the centre of NGC 613.The study of X-ray emission is presented in section 4. Wediscuss the data in section 5, presenting the possible scenar-ios, and section 6 summarizes the conclusions of this work.In Appendix A, we present the optical extracted spectra ofthe observed regions and the method that we used to cal-culate the integrated flux of the emission lines, in order tocalculate their ratios.
This work presents data obtained with different instru-ments: Wide-Field Planetary Camera 2 (WFPC2) from
HST , GMOS from Gemini-South telescope, SINFONI fromVLT, SIFS from SOAR telescope, Advanced CCD Imag-ing Spectrometer (ACIS) from
Chandra space telescope andALMA. The following subsections describe the treatmentand conditions of each observation.
MNRAS000
MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 Fig. 1 shows the image of NGC 613 nucleus, obtainedwith
HST , and the fields of view (FOVs) of the GMOS,SINFONI, and SIFS data cubes. The GMOS data cube hasPA = 127 ◦ , so its orientation is different from the otherFOVs, whose PAs = 0 ◦ . By comparing the three data cubes(optical and NIR), the SIFS FOV is the largest, followed bySINFONI (which has the highest spatial resolution) and byGMOS. Hubble Space Telescope images
In this study, we used images obtained from the
HST pub-lic archive. Those images were part of the observation pro-grams 9042 (PI: Smartt, S. J.) and 15133 (PI: Erwin, P.) ,obtained with the WFPC2 instrument on 2001 August 21and WFPC3 on 2018 August 22. For the program 9042, theimages in the filters
F 450 W , F 606 W and
F 814 W weretaken with an integration time of 160 s. Fig. 1 shows theFOV in the central area of the galaxy, whose size is 13.6arcsec × F 814 W filter was 500 s and for the image of
F475 W filter was 700 s.
Although having similar spectral range, the data cubes ob-tained with the integral field units (IFUs) of GMOS andSIFS present some significant differences. The SIFS FOV isalmost 6.7 times larger than the GMOS FOV, in area, whichallows us to analyse the circumnuclear region of this galaxyon a larger scale. In order to evaluate the signal-to-noise ra-tio (S/N) of these data cubes, we extracted a spectrum ofa rectangular area with 0.3 arcsec × × S / N GMOS = 29 and S / N SIFS = 27.Considering that and also taking into account the exposuretimes for the GMOS data (930 s) and for the SIFS data(900 s), we conclude that these two instruments are verysimilar, regarding the S/N ratio. As discussed in the follow-ing sections, the difference between the spatial resolution ofthose data cubes is highly significant, since the seeing of theobservation taken with SIFS was considerably larger thanthe seeing of the GMOS observation, resulting in the fullwidth at half-maximum (FWHM) of the point spread func-tion (PSF) in the SIFS data cube being three times largerthan the FWHM of the PSF obtained with the GMOS datacube.
These data were taken using the IFU of the GMOS, in one-slit mode. The data are part of the
Deep IFS View of Nucleiof Galaxies (DIVING3D) survey, which aims to analyse thenuclear regions of all Southern-hemisphere galaxies brighterthan B=12 (Steiner, J. E. et al., in preparation). The observations of NGC 613 were taken on 2015 Jan-uary 25 as part of the observation program GS-2014B-Q-30in the Gemini-South telescope. Three 930 s exposures weretaken with spatial dithering and PA = 127 ◦ . Using the grat-ing R831+G5322, centred in 5850 ˚A, we obtained a spectralresolution of R=4340 and a spectral coverage from 4675 to6828 ˚A.We used data reduction packages, developed by theGemini observatory, in iraf environment. This reductionwas conducted using the following processes: trim determi-nation, bias subtraction, cosmic ray removal ( lacos – vanDokkum 2001), spectra extraction, corrections of gain varia-tions between the spectral pixels (using GCAL-flat images),corrections of gain variations between the spaxels (usingtwilight-flat images), wavelength calibration (using imagesof the CuAr lamp), subtraction of the average spectrumof the FOV corresponding to the sky observation (locatedat 1 arcmin from the object), flux calibration, atmosphericextinction correction, telluric absorption removal, and datacube construction. The resulting three data cubes had spax-els of 0.05 arcsec × × i ] λ The data were taken on 2017 November 23 with SIFS onthe SOAR telescope, during the Science Verification pro-gram. Six 900 s exposures were taken with spatial dithering,in order to obtain a mosaic of the central region of NGC613. Three of these exposures were centred 3 arcsec north ofthe nucleus of the galaxy and the other three were centred 3arcsec south of the nucleus. The dither step for each of thesegroups of three exposures was 0.3 arcsec. The PA of all theobservations was 0 ◦ . The plate scale was 0.3 arcsec/fibre andthe grating used was 700 l/mm, centred in 5650 ˚A, which re-sulted in a spectral coverage from 4250 to 7050 ˚A and spec-tral resolution of R = 4200. During the same night, datafrom the standard star LTT 2415 were taken with exposuretime of 300 s. Since SIFS has an Atmospheric DispersionCorrector (ADC), the differential atmospheric refraction ef-fect was removed from the observed data.The reduction was performed using scripts developed inIteractive Data Language ( idl ) and included the followingsteps: correction of dead fibres, spectra extraction, flat-fieldcorrections (in order to remove gain variations between thespectral pixels and between fibres), wavelength calibration(using images of an HgAr lamp), sky subtraction, and cre-ation of data cubes. After this, processes of flux calibration,atmospheric extinction correction, and telluric absorptionremoval were applied using scripts developed in iraf en-vironment. At the end, we constructed a mosaic with the MNRAS , 1–18 (2020)
Patr´ıcia da Silva et al.
Figure 2.
Image of the GMOS data cube, obtained after the treatment, of the central region of NGC 613 collapsed along the spectralaxis with its flux scale and with its average spectrum corrected for redshift. Note the indication of N-E and the scale of 100 pc.
Figure 3.
Image of the SIFS data cube, obtained after the treatment, of the central region of NGC 613 collapsed along the spectral axiswith its flux scale and with its average spectrum corrected for redshift. Note the indication of N-E and the scale of 200 pc.
Figure 4.
Image of the SINFONI data cube, obtained after the treatment, of the central region of NGC 613 collapsed along the spectralaxis with its flux scale and with its average spectrum corrected for redshift. Note the indication of N-E and the scale of 150 pc. six reduced data cubes and obtained a final data cube withspaxels of 0.3 arcsec × × The data were obtained from the public archive and weretaken on 2005 October 23 and on 2005 November 13 as partof the observation program 076.B-0646(A), PI: B¨oker,T.,with four and five exposures of 300 s, respectively, in theH+K band, which resulted in a spectral resolution of R =1500 and a spectral range from 15000 to 24500 ˚A. The PA of
MNRAS000
MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 the observation was 0 ◦ with fore-optics of 0.25 arcsec. Thisdata set was already published by B¨oker et al. (2008) andFalc´on-Barroso et al. (2014). Here, we use these data super-posed to SIFS data to locate the H ii regions, in order tocalculate the emission-line ratios in the SIFS data cube andto make some comparisons with the other data. In paper II,we use these data to study the stellar and gas kinematics.The reduction of the data was performed using the gasgano software in a process involving the following steps:bad pixel correction, flat-field correction, spatial rectifica-tion in order to remove potential distortions along the FOV,wavelength calibration (using Ne lamp images), sky sub-traction, and data cube construction. The flux calibrationand telluric absorption removal were applied later, usingscripts developed in iraf environment. We obtained ninedata cubes with spaxels of 0.125 arcsec × × λ , was 0.69 arc-sec. Although the data reduction procedure we used is anal-ogous to the one adopted by B¨oker et al. (2008) and Falc´on-Barroso et al. (2014), we also applied a treatment procedureto the reduced data cubes (not used by the aforementionedauthors), in order to obtain certain improvements, as de-scribed in section 2.4. Fig. 4 shows the image of the datacube collapsed along the spectral axis and the data cubeaverage spectrum, showing the main emission lines in thisspectral range. The data cube treatment was applied using scripts writ-ten in idl developed by our group (see Menezes et al. 2014,2015, 2019). We started by performing the differential atmo-spheric refraction correction in the data cubes (except in theSIFS data cube that has ADC). After that, the data cubeswere combined in only one: we combined the three GMOSdata cubes in form of a median. As said previously, we had,after the reduction, nine SINFONI data cubes; they werecombined in order to minimise sky subtraction problems, assome of them had sky emission in excess and some othersexaggerated sky subtraction. The combinations were donefrom 3 to 3 data cubes using medians (alternating betweenthe dates of the observations) until we had only one finaldata cube.SINFONI and SIFS data cubes were spatially re-sampled to have spaxels of 0.0625 arcsec and 0.1 arcsec,respectively. All the data cubes were spatially filtered usingthe Butterworth method (Gonzalez & Woods 2002), in or-der to remove the high spatial frequency noise. Then, theinstrumental fingerprint removal was applied. Lastly, we ap-plied the Richardson–Lucy deconvolution (Richardson 1972and Lucy 1974). For more details about the Richardson–Lucy deconvolution and the other treatment techniques, seeMenezes et al. (2014, 2015, 2019). In GMOS and SIFS datacubes, the PSF variation law was estimated using the stan-dard stars data cubes that were used in the data cubes re-duction processes. For the SINFONI data cube, we used aconstant PSF.As said previously, the SIFS standard star was observedin the same night, so we used these data to estimate the FWHM of the PSF, at 5647 ˚A, that was equal to 2.4 arcsec.The process of deconvolution was performed with six itera-tions. All the treatment was also performed in the standardstar data cube, in order to obtain the FWHM of the PSFafter the deconvolution process (1.6 arcsec).Regarding the GMOS data cube, the best estimate ofthe FWHM of PSF was obtained from the [O i ] λ i ] λ λ , whichpresents a point-like emission in the inner centre of the FOV.The difference of the FWHM of the PSF before and after thedeconvolution process was also lower than we expected. Thedeconvolution was applied with 10 iterations and the result-ing FWHM of the PSF was 0.60 arcsec. It is important tomention that, as discussed in Menezes et al. (2014, 2015,2019), the Richardson–Lucy deconvolution does not com-promise the data in anyway (keeping the flux values andalso the spatial morphology of the structures unchanged).Therefore, although the improvement in the spatial resolu-tion, in this case, is not dramatic, such a technique is worthto be applied. In order to study the gas emission in the data cubes, weapplied a spectral synthesis using the starlight software(Cid Fernandes et al. 2005), with a stellar population spec-tral base created using the Medium-resolution Isaac NewtonTelescope Library of Empirical Spectra (MILES, S´anchez-Bl´azquez et al. 2006) in the optical spectral range (GMOSand SIFS), to trace the stellar continuum and remove it.The spectral synthesis is applied to each spectrum of thedata cube and consists of a linear combination of the basespectra to obtain the observed spectrum. First, we mask allthe emission lines of the treated data cube and then we ap-ply the spectral synthesis. After that, we create a syntheticstellar data cube containing only the synthetic spectra ob-tained in the process. Then, subtract this data cube fromthe original treated one, obtaining a data cube with mainlygas emission that we call gas data cube.In order to obtain the SINFONI gas data cube, weperformed the spectral synthesis using the Penalized PixelFitting (pPXF, Cappellari & Emsellem 2004) method. Thespectral base covers only the K band (Winge et al. 2009).pPXF is a spectral synthesis that uses the spectra from thebase convolved with Gauss–Hermite functions. Then, in thesame way as in the optical, we create a synthetic stellar datacube and subtract it from the original one, obtaining a gasdata cube, in this case, only in the K band. MNRAS , 1–18 (2020)
Patr´ıcia da Silva et al.
Figure 5. (a) Image of the GMOS data cube integrated along the spectral axis. The white contours represent the emission of [O i ] λ i ] C ( X C = − . arcsec and Y C = − . arcsec). The position of N2 was defined as the centre of theemission northwest of N1 ( X C = . arcsec and Y C = . arcsec). (b) GB composition of the image in the HST filter
F 450 W (blue) andthe image of the integrated GMOS data cube (green). (c) RG composition of the same GMOS data cube image (green) and the imagein the
HST filter
F 814 W (red). (d) RGB composition in the
HST filters:
F 814 W (red),
F 606 W (green) and
F 450 W (blue). Thecircles indicate the regions N1 and N2. All the images have crosses that indicate the positions [O i ] C and N2, whose sizes represent theuncertainty of 3 σ , taking into account the size of the spaxels of the GMOS data cube. Since the last panel contains only the images from HST , whose spaxels are larger than the GMOS spaxels, the size of the cross differs.
Figure 6. (a) Image of the centre emission of H λ γ invertedimage in flux units (or Br γ in absorption). (b) RG composition of H λ γ (red). The yellow region is where Br γ isin absorption or not being emitted. Both images have indications of the positions of [O i ] C and of the centre of N2, assuming that theposition of [O i ] C is coincident with the peak of the H λ ◦ ,indicated in the image) and distance between the [O i ] C and N2 (centre of the brightest stellar nucleus from the HST images).
Chandra data
The data from
Chandra space telescope were obtained fromthe public archive. The observations were taken on 2014 Au-gust 21 (PI: Garmire, G., program: 15610062), with a 14.1 hexposure, using the ACIS instrument. From the informationof the images headers, referring to the positions and energyof the detected photons, we created a data cube.The data cube was spatially re-sampled. The origi-nal size of the spaxels was 0.492 arcsec and, after the re-sampling, was 0.246 arcsec. After that, a spatial filtering us-ing the Butterworth method was applied and the data cubewas spatially resized, in order to make superpositions (or more direct comparisons) between this data cube and theothers analysed in this work. To do this, we took a referencepoint in this data cube and its values of RA and Dec. andcompared to the other data cubes and also the
HST images,then we centred the images at the same point.We used the data of Mrk 202 to compare PSFs, as shownin section 4. These data were obtained with the same instru-ment, ACIS, and the same process of creation, re-samplingand resizing used in the data cube of NGC 613 was per-formed. The data were taken on 2003 March 16 (PI: Predehl,P., program: 04700038) with a 2.2 h exposure.
MNRAS000
MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 Figure 7.
RG compositions with the CO(3-2) image obtainedfrom ALMA data cube (in green) and the H λ i ] C and of the centre of N2. The size of the crossesrepresents the 3 σ uncertainty taking into account the size of SIN-FONI spaxels. This image was made to match the SINFONI andALMA data using the circumnuclear ring position and also tolocate [O i ] C and the centre of N2 in ALMA data. The data from ALMA were obtained from its public archive.The observations were part of the program 2015.1.00404.S(PI: Combes, F.) and these data were already been publishedin Combes et al. (2019) and Audibert et al. (2019) works.We used here only the data cube of CO(3-2) emission line,whose beam size was 0.43 arcsec × HST images, reinforcing a connection between the bar and thenuclear spiral.
The central region of NGC 613 is a complex environment,since it has emissions of many sources and natures. The emit-ting regions were studied according to the instruments, de-pending on the spatial resolution and FOV. In other words,the inner central region was mainly studied using the GMOSdata cube and
HST images, whereas the circumnuclear re-gion was better studied using SIFS and SINFONI,
Chandra ,and ALMA data cubes. The gas emission, as said previously(see section 2.5), was analysed using the gas data cubes ofGMOS, SIFS and SINFONI.
Table 1.
Right ascension (RA) and declination (Dec.) of [O i ] C ,and the centres of N1 and N2. These coordinates were derivedfrom the HST images, as pointed in Fig. 5d. The relative uncer-tainties of the values are ∼ . arcsec. The absolute uncertaintiescorrespond to the uncertainty of the pointing of the HST .RA Dec.[O i ] C
1h 34m 18.172s –29d 25m 5.885sN1 1h 34m 18.201s –29d 25m 5.74sN2 1h 34m 18.139s –29d 25m 5.99s
The integrated GMOS data cube shows an extended emis-sion with an elongated morphology, consistent with the pres-ence of a double structure (see Fig. 2). This same pattern ispresent both in the emission-line images and in the contin-uum image of this data cube. Since we did not detect anyfeatureless continuum in this object in the GMOS data cube(see Paper II), we can say that the continuum emission fromthis double structure is essentially stellar.The
HST images also present a double structure thatis the same one observed in the GMOS data cube, since ithas the same length ( . ± . arcsec) and orientation (PA= 255 ◦ ± ◦ ), as shown in Fig. 5. It is worth mentioningthat the uncertainties given here and throughout the textare the relative uncertainties, based on the spatial samplingof the images. The absolute uncertainties correspond to theuncertainties of the pointing of the instruments. The emis-sion of [O i ] λ i ] C , whose coordinates are X C = − . ± . arcsec and Y C = − . ± . arcsec in the GMOS FOV. When we super-pose the GMOS and the HST data (Fig. 5a), we see that the[O i ] λ HST image (see Fig. 5d). In the GMOS data cube, thecentre of N2, with coordinates X C = . ± . arcsec and Y C = . ± . arcsec, was defined as the centre of the ex-tended emission, northwest of the [O i ] λ HST images), seeFig. 5(a).The centre of N2, visible both in the
HST and GMOSimages, was used as reference to match the images of thetwo telescopes. The
HST -based coordinates for the centreof N1 and N2 and the position of [O i ] C are given in Table1. We used the position of [O i ] C to superpose the images of HST and SIFS FOV (Fig.1). The SIFS data cube does nothave enough spatial resolution to allow the visualisation ofthe extended emission that we see in the GMOS data cube.Because of that, we were only able to determine the positionof [O i ] C and, therefore, we could not study region N2 byanalysing this data cube.In the case of the SINFONI data cube, we noted that theBr γ inverted image, in flux units, shows a lack or low emis-sion, or even Br γ in absorption, in the inner centre of NCG613. The Br γ in absorption is the most likely feature and itmight be related to the presence of young stellar populations MNRAS , 1–18 (2020)
Patr´ıcia da Silva et al.
Figure 8. (a) GMOS data cube integrated image considering the transmission curve of the
HST filter
F 606 W . (b) Image in the
HST filter
F 606 W convolved with the PSF of the GMOS data cube (0.71 arcsec). Both (a) and (b) had their fluxes normalized at N2. Onecan note that, in the GMOS image, there is an emission excess near [O i ] C and, in the HST image, this excess is located at N2. Thesubtraction of these two images (c) shows a source near [O i ] C (the green cross represents its centre), whose projected distance to [O i ] C is 0.11 arcsec. We named this source v2. The sizes of the crosses represent the uncertainty of 3 σ for the positions of the sources. in regions N1 and N2 (see Paper II), since those stars presentdeeper hydrogen absorption lines than the older ones. Thepattern of the region of Br γ in absorption located at thecentre is similar to the extended emission that we see inthe GMOS data cube (see Fig. 6). When we superpose theimage of the point-like source of the SINFONI data cube(H λ i ] λ i ] C slightly displaced towards the centre and withthe N2 position compatible with the other end of this ex-tended morphology. From that and also from the fact thatthe superposition of SINFONI and SIFS (after assuming thepositions for [O i ] C and for the centre N2) is correct, as itdelineates successfully the circumnuclear ring as observedin the SIFS data cube (see Fig. 10), we concluded that theposition of [O i ] C is, within the uncertainties, the centre ofthe H λ i ] C are different from the ones we obtained in the optical (Ta-ble 1), indicating that the values of RA and Dec. from theSINFONI data cube are not reliable.Fig. 5(a) shows the result of the GMOS data cube col-lapsed along its spectral axis, the positions of [O i ] C and N2(discussed previously) and the white contours that representthe emission of the [O i ] λ i ] λ HST filter
450 W ), whereas such emission is displaced tonorth-east of [O i ] C , but still a significant part of it is insidethe region that delineates N1 (see Fig.5b). At longer wave-lengths ( HST filter
814 W ), we see clearly N1 and N2 (Fig.5c). In this case, N1 seems to have stronger emission in redthan N2 (better seen in panel d). We see in Fig. 5(d) that N1is in a region predominantly red, possibly highly affected bydust, that may cause the displacement of the blue emissionwe see in Fig. 5(b), which might have been coming from theposition [O i ] C (represented by the cross in the image).The Chandra data cube matched the data by the RA and Dec. In section 4, we see that the position of [O i ] C ,N1, and N2 centres are compatible with the centre of thehard X-ray emission, since these data do not have enoughspatial resolution to separate those sources. So we took theposition of [O i ] C as reference to the superposition that wasmade based on the RA and Dec.In order to compare the molecular gas emission obtainedwith the ALMA data cube of CO(3-2), we made a superpo-sition with the molecular emission observed with H λ i ] C and N2 in this context. For discussionof Fig. 7, see section 5.2. When we compare the collapsed GMOS data cube imagesand the
HST
RGB composition we see a difference betweenthe emissions of N1 and N2: in the
HST images, N2 is thebrightest source and, in the GMOS data cube image, N1is brighter than N2. In order to make a more direct com-parison of this feature, the GMOS data cube was integratedalong the spectral axis, taking into account the transmissioncurve of the
HST filter
F 606 W . Besides that, the image ofthe
HST filter
F 606 W was re-sampled to have pixels withthe same sizes of the spaxels of the GMOS data cube (0.05arcsec × HST filter
F 606 W (Fig.8c) shows a source whose centre is compatible with the posi-tion of [O i ] C , within the uncertainties (the distance between MNRAS000
F 606 W (Fig.8c) shows a source whose centre is compatible with the posi-tion of [O i ] C , within the uncertainties (the distance between MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 Table 2.
Emission-line ratios of the regions detected in the nucleus of NGC 613, calculated from the integrated flux of the emission-linesof the spectra extracted from each region (see Figs. A1, A3 and A4 and appendix A).Regions [O iii ]/H β [N ii ]/H α [O i ]/H α ([S ii ] λ λ α [S ii ] λ ii ] λ ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Figure 9.
Top:
F 814 W filter profile extracted along the PA =255 ◦ ± ◦ (the orientation that passes through [O i ] C and N2) fromthe HST images observed on 2001 (WFPC2) and 2018 (WFPC3).The images were normalized considering the flux in N2. Also theimage from WFPC3 was convolved with an estimate of the PSF ofthe image from WFPC2 for a better comparison. The differencebetween the 2018 data and the 2001 profiles is represented bythe purple curve with the indication of the source of variationV1. We noticed a source between N1 and N2 ( ∼ i ] C ) that has higher luminosity than before. We named thissource V1. Bottom: images of F 814 W HST filter of the innercentre of NGC 613 (a) taken in 2001 and (b) taken in 2018. Bothimages show indication of regions N1 and N2, [O i ] C position andV1. The diameter of the green circles was determined based onthe size of N2 in the 2018 image and the diameter of the cyancircle was taken as the size of V1 in the 2018 image. The crossindicating [O i ] C corresponds to 3 σ of the GMOS spaxel. Table 3.
FWHM and luminosity of the H α line determined foreach region (N1 and N2 from the GMOS data cube spectra andthe other regions from the SIFS data cube spectra). The spectrawere corrected of reddening using H α /H β ratio.Regions FWHM of H α (km s − ) H α luminosity(10 erg s − )N1 592 ± ± ± ± ± ± ± ± ± ± ± ±
25 113 ± ± ± ± ± ± ± ± ± ± ± ± [O i ] C and this source is 0.11 arcsec). Such a source, whichwe call V2, was not detected before. The brightness varia-tion of this source happened in the interval of about 14 yr(between 2001 August and 2015 January, when the obser-vations of HST and GMOS were taken, respectively). Sucha variation could be an evidence of a supernova. However,one can notice a vertical pattern in Fig. 8(c) that is similarto the instrumental fingerprint signature in the GMOS datacube images (see Menezes et al. 2019). We cannot discardthat this image may have some instrumental fingerprint rem-nants, still present in the GMOS data cube after treatment,that generated this pattern. Nevertheless, we know that theinstrumental fingerprint cannot generate emissions that areapproximately point-like as the one that we are seeing.Another way to study variability was possible due tothe availability of recent images from the
HST . By compar-ing images obtained with the same filter (
F 814 W ) in 2001and 2018, we aimed to evaluate in more detail possible fluxvariations of the sources during this period. The 2001 imagewas obtained with WFPC2, while the 2018 image was ob-tained with WFPC3, which has a higher spatial resolution.In order to perform a comparison, first of all, we convolvedthe 2018 image with an estimate of the PSF of the 2001image. After that, we extracted, from each image, a bright-
MNRAS , 1–18 (2020) Patr´ıcia da Silva et al.
Figure 10.
RG composite image, with red representing the im-age of the H α /[N ii ] λ α emission is higher or more rel-evant than the [N ii ] λ γ emission image, from the SINFONI data cube, whichdelineates the star formation ring. We added a dashed yellow linerepresenting approximately the [O iii ] λ γ image, we identified eight emittingregions (named from 1 to 8). In addition to these, we detected re-gions 9 and 10, from the image of the SIFS data cube, as theyare not part of the SINFONI FOV and neither of the ring. Weadded circles to the image representing the positions of N1 andN2, whose sizes are the same of the PSF from the GMOS datacube. All the circles represent the extraction area of the spectrumof each region, in order to calculate their emission-line ratios. ForN1 and N2, we used the GMOS gas data cube to extract theirspectra, because the spatial resolution was better and high enoughto separate those emissions. Since the spectra of regions 1 to 10were extracted from the SIFS gas data cube (because the GMOSFOV does not contain those regions, but only the inner edges ofsome of them), we used the PSF of this data cube for the extrac-tion. Note that the number that indicates each regions follows thesame orientation as Falc´on-Barroso et al. (2014), however Region8 is not the same one the authors named in their article. ness profile from a rectangular region along a PA= 255 ◦ andwith a width of 0.75 arcsec. The two extracted profiles werenormalized in a way that the flux peak (in N2) in both pro-files was equal to 1. The top panel of Fig. 9 shows the twoextracted profiles, together with the curve corresponding tothe difference between them. We notice a flux increase, from2001 to 2018, in a source between N1 and N2, whose positionis not compatible, even at the 3 σ level, with the position of[O i ] C . We named this variable source as V1. The distancebetween V1 and [O i ] C is . ± . arcsec. This variationis better seen by comparing the images of the bottom panelin Fig. 9. If the real position of the AGN is V1, we are see-ing that this AGN is variable. If not, we are seeing anotherindication of supernova or even variations of dust extinctionin the line of sight. Similar to V2, the presence of V1 hasnot been previously reported in the literature. Figure 11.
Diagnostic diagram obtained from the emission-lineratios on table 2. The names of each studied region are indicatednear their respective point. The dark green line represents thetheoretical limit for the starburst emission obtained by Kewleyet al. (2001). The separation between H ii regions and AGNsdetermined by Kauffmann et al. (2003) is indicated by the lightgreen curve. The red line represents the separation between theSeyferts and LINERs emission determined by Schawinski et al.(2007). The other two diagnostic diagrams,log [O iii ]/H β × log ([S ii ] λ λ α and log [O iii ]/H β × log [O i ]/H α representsessentially the same results, and therefore are not displayed in thiswork. Figure 12. [O iii ] λ i ] C and its size theuncertainty of 3 σ . It is known, in the literature, that NGC 613 nucleus has acircumnuclear ring with many star-forming regions (B¨okeret al. 2008; Falc´on-Barroso et al. 2008, 2014). The FOV ofthe GMOS data cube (3 arcsec × MNRAS000
Diagnostic diagram obtained from the emission-lineratios on table 2. The names of each studied region are indicatednear their respective point. The dark green line represents thetheoretical limit for the starburst emission obtained by Kewleyet al. (2001). The separation between H ii regions and AGNsdetermined by Kauffmann et al. (2003) is indicated by the lightgreen curve. The red line represents the separation between theSeyferts and LINERs emission determined by Schawinski et al.(2007). The other two diagnostic diagrams,log [O iii ]/H β × log ([S ii ] λ λ α and log [O iii ]/H β × log [O i ]/H α representsessentially the same results, and therefore are not displayed in thiswork. Figure 12. [O iii ] λ i ] C and its size theuncertainty of 3 σ . It is known, in the literature, that NGC 613 nucleus has acircumnuclear ring with many star-forming regions (B¨okeret al. 2008; Falc´on-Barroso et al. 2008, 2014). The FOV ofthe GMOS data cube (3 arcsec × MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 Figure 13. (a) Total spectrum of the
Chandra data cube of NGC613, (b) spectrum of the circular region whose centre is indicatedby the blue cross in Fig. 14 with the extraction radius of ∼ ∼
20 arcsec × the GMOS data cube. So, to study these regions, we usedthe SIFS and SINFONI data cubes (the latter was the oneused in the literature to characterize the ring in the NIR).The Br γ emission image of the SINFONI data cubeclearly shows eight star-forming regions that compose thering (see Fig. 10 in green). The ring can be observed inthe SIFS data cube, however this data cube does not haveenough spatial resolution to separate each region. On theother hand, as we can see in Fig. 10 (in red), we detected,besides the ring, two regions (named 9 and 10) that seemto be connected to the ring. Falc´on-Barroso et al. (2014)detected seven H ii regions, that are regions 1 to 7 in Fig.10. However, we are considering another region in the ring(Region 8) and their region 8 is actually a region inside thering, but not part of it, that they use for comparison. From the optical gas data cubes, in this case GMOS andSIFS, it is possible to use the emission-line ratios to char-acterize the nature of the emission from the regions we de-tected. Fig. 10 shows the sizes and positions of the areasthat we delineated to extract the spectrum of each region.The spectra of these regions were extracted from this datacube and the diameter of the extraction area was equal tothe FWHM of the PSF of these data (see Fig. 10, magentacircles). For the regions named 1 to 10, we had to makea superposition of the Br γ emission image (which clearlyshows regions 1 to 8 in the ring, in green) and of the H α /[N ii ] λ α emission is more relevant and therefore thepresence of the H ii regions), since we had to know the po-sitions of those regions in this data cube, in order to extractthe spectra to calculate the emission-line ratios. As the spec-tra were extracted from the SIFS data cube, the diameterof the extraction areas of those 10 regions was equal to theFWHM of the PSF of these data (see green circles in Fig.10), with exception of regions 9 and 10 that had their areasand positions estimated directly from the image of H α /[N ii ] λ α and H β lines. In order to calculate the emission-line ratios, we de-composed those blended lines in Gaussians functions. Thedecomposition method and the flux and decomposition un-certainties are also discussed in appendix A. Table 2 containsthe emission-line ratios with their respective uncertainties ofthe 12 regions and Fig. 11 shows the diagnostic diagram of[O iii ] λ β × [N ii ] λ α .We can see that, excepting regions N1, N2, 2, 3, and 8,the emission of all other ones is compatible with the emis-sion from H ii regions. N1 and N2 have emission-line ra-tios compatible with the ones of LINERs. Based on the [O iii ] λ β ratio, N2 seems to have a higher ionization de-gree than N1, although the values of this ratio for N1 andN2 are compatible, at the 3 σ level. Regions 2 and 3 haveemission-line ratios compatible with the ones of transitionobjects. The yellow contour of Fig. 10 corresponds to theedges of the emission of [O iii ] λ iii ] λ β ratio) and emission-line ratios compat-ible with the ones of AGNs indicates that the emission fromthose regions is highly contaminated by the ionization cone(being either the result of an overlap of emissions or of anionization of these regions directly by the AGN). Region 8is also very close to the position of the radio jet and outflowand this resulting ionization degree might be due to a con-tamination from the radio jet and outflows coming from theAGN (see section 4).We will discuss in section 4 that the [O iii ] λ α emission line in the spectra of the observed regions. Onecan note that N2 is brighter than N1 (regarding H α emis-sion) and that region 4 is the brightest of the ring. N1 hasthe highest FWHM values, which indicates the presence ofoutflows of gas. K band A map of the D CO coefficient (M´armol-Queralt´o et al. 2009)obtained from the same SINFONI data analysed in this workis shown in Fig. 5 of Falc´on-Barroso et al. (2014). The val-ues of such a coefficient are higher for spectra with a deeperCO absorption band at . µ m. The authors note a slightdecrease in the values of the D CO coefficient at the nucleus,in comparison to the surroundings. If this object presenteda significant emission of a featureless continuum, we would MNRAS , 1–18 (2020) Patr´ıcia da Silva et al.
Figure 14.
Hard X-ray image, 2 to 10 keV, of the
Chandra data cube of NCG 613 and radial profiles of the images of hard X-ray fromNGC 613 and Mrk 202 (a Seyfert 1 galaxy observed with the same instrument, used here as an estimate of the PSF of the data, since ithas only a central emission). The blue cross indicates the position of the peak of the emission and its size the uncertainty of 3 σ . expect shallower absorption CO bands at the nucleus (seeBurtscher et al. 2015), which would result in lower valuesof D CO in that region. Considering the fact that the D CO map presented by Falc´on-Barroso et al. (2014) is quite noisy(due probably to the S/N of SINFONI data cube), it is diffi-cult to confirm whether or not the drop of the values at thenucleus is real. We conclude that the effect of a featurelesscontinuum in these data is, at most, very weak. Therefore,the featureless continuum at the nucleus of NGC 613, in the K band, is either too weak (due to an AGN lower state ofactivity, if it is variable) to be detected or should be highlyobscured in the K band. See section 5.1 for a detailed dis-cussion and a comparison with previous studies. In order to further investigate the nature of the emission inthe centre of NGC 613, we analysed a data cube obtainedwith the
Chandra space telescope. We subtracted the spec-trum of a circular region centred on the nucleus with a radiusof ∼ ix at 0.91 keV, Ne x Ly α at1.03 keV and Ne x Ly cont at 1.35 keV.Fig. 14 shows that the hard X-ray emission is concen-trated, mainly, in the inner centre. In order to see how com-pact the central emission is, we compared its profile withthat of Mrk 202, a Seyfert 1 galaxy (V´eron-Cetty & V´eron2006) observed with the same instrument as NGC 613 (seesection 2.6). The image of Mrk 202 AGN in hard X-raysshows only a central source and no relevant circumnuclearfeatures. The radial profiles of the hard X-ray emission ofthose galaxies are clearly distinct (right-hand panel of Fig. Figure 15.
Image from
Chandra space telescope of the centralregion of NGC 613 in the soft X-rays, with indication of the orien-tation N-E and scale of 200 pc. The PA of the radio jet detectedby Hummel & Jorsater (1992) is represented by the yellow vector(PA = 12 ◦ ), the PA of the outflow observed in the H α channelmap with v= 306 km s − (see Paper II), with PA = 17 ◦ , is rep-resented by the magenta vector and the outflows observed in [O iii ] λ − (PA ∼ –10 ◦ ) and with v = –710 kms − (PA ∼ ◦ ) are represented by the brown vectors. The purplevector with PA = 255 ◦ connects [O i ] C to N2, but [O i ] C and thecentre of N2 are approximately at the same point, because of thelow spatial resolution. [O i ] C position was taken as the centre ofhard X-rays emission, represented by the white cross and its sizeis the 3 σ uncertainty. We also added the red contours that repre-sent the ionization cone observed in the image of the [O iii ] λ000
Chandra space telescope of the centralregion of NGC 613 in the soft X-rays, with indication of the orien-tation N-E and scale of 200 pc. The PA of the radio jet detectedby Hummel & Jorsater (1992) is represented by the yellow vector(PA = 12 ◦ ), the PA of the outflow observed in the H α channelmap with v= 306 km s − (see Paper II), with PA = 17 ◦ , is rep-resented by the magenta vector and the outflows observed in [O iii ] λ − (PA ∼ –10 ◦ ) and with v = –710 kms − (PA ∼ ◦ ) are represented by the brown vectors. The purplevector with PA = 255 ◦ connects [O i ] C to N2, but [O i ] C and thecentre of N2 are approximately at the same point, because of thelow spatial resolution. [O i ] C position was taken as the centre ofhard X-rays emission, represented by the white cross and its sizeis the 3 σ uncertainty. We also added the red contours that repre-sent the ionization cone observed in the image of the [O iii ] λ000 , 1–18 (2020) he nuclear region of NGC 613 ∼
170 pc), while the FWHM of Mrk 202 is 0.99arcsec; the difference is 0.33 arcsec.There is also a clear extended emission with an elon-gated structure that is consistent with being an inclineddisc, as seen in radio, [Fe ii ] λ γ and molecular lines(Fig. 14). The best estimates of the radius and inclination,assuming it is a circular disc, are consistent with the discparameters seen in the NIR and radio (see table 4). Thisdetection of the circumnuclear ring in hard X-rays has notbeen previously reported in the literature. One may wonderwhat the origin of such circumnuclear hard X-ray emissionis. Our interpretation is that it is most likely originated fromsupernova remnants (SNRs) associated with the young stel-lar population. The presence of SNRs in the central region ofNGC 613 has already been proposed by B¨oker et al. (2008)and Falc´on-Barroso et al. (2014). See section 5.2 for a de-tailed discussion.Fig. 15 shows an image of the soft X-ray (0.5–2 keV)emission (in blue) with the [O iii ] λ iii ] λ α and [O iii ] λ NGC 613 has a rich nuclear region. The presence of an AGN(Veron-Cetty & Veron 1986), of a star-forming ring ( Hum-mel & Jorsater 1992, B¨oker et al. 2007, 2008, Falc´on-Barrosoet al. 2008, 2014, Miyamoto et al. 2017, 2018), of a radio jet(Hummel et al. 1987, Hummel & Jorsater 1992, Miyamotoet al. 2017, 2018), of a nuclear spiral (Audibert et al. 2019),of an outflow (detected from the study of the [O iii ] λ The
HST images of NGC 613 show two sources of stellaremission separated by a stream of dust. The projected dis-tance between these two stellar nuclei is 94 ± . ± . arcsec), as seen in HST images. These two stellar sources could be seen in
HST images shown in Falc´on-Barroso et al.(2014), Combes et al. (2019) and Audibert et al. (2019), butwere not discussed by those authors. When we observe theimage of the GMOS data cube, we see that the most intensestellar emission is extended and its orientation is the sameof the one between the double stellar source in the
HST im-ages (PA = 255 ◦ ± ◦ ), as shown in Fig. 5. We named thisemission as regions N1 and N2. We also defined the peakof the [O i ] λ i ] C , which is an emissionline typically associated with regions of partial ionizationnormally seen in AGNs.When we compared the HST images from 2001 and2018, we found that there is a source between N1 and N2that suffered a variation in brightness (Fig. 9). We call thissource, not previously reported in the literature, V1 (see Fig.16). The composition of the
HST images and the CO(3-2) image from the ALMA data cube (Fig.17a) reveals thatthere is a nuclear spiral that passes between N1 and N2and the centre of this nuclear spiral is this variable source(V1) that appears strongly in the
F 814 W filter image from2018. This nuclear spiral was first detected by Combes et al.(2019) and Audibert et al. (2019). V1 suffers obscuration bydust (as we can see in Fig. 1 and that is compatible alsowith previous studies that said that NGC 613 AGN is veryobscured – see Castangia et al. 2013; Asmus et al. 2015) thatcan be a result of the nuclear spiral that seems to bring gasand dust to the centre.Our analysis shows that V1 and [O i ] C are separatedby . ± . arcsec. The positions of these two sources arenot quite compatible, even at the 3 σ level. One may wonderwhere the actual position of the AGN really is, in this case.If the emission of [O i ] λ is extended, this shift can beexplained by differential dust extinction. However, we alsohave to acknowledge that, if we think that the original emis-sion of [O i ] λ i ] C is expected inthe direction of the cone with less extinction. The source V1may be equally reflected but its image was taken with the F 814 W HST filter, suffering, thus, less extinction than[O i ] λ D CO map obtained byFalc´on-Barroso et al. (2014) revealed almost no variationstowards the nucleus, as expected in the case of obfuscationby a featureless continuum from an AGN (Burtscher et al.2015). This is somewhat surprising as the K band is lessaffected by interstellar extinction than the optical. Whenwe look at the image of F 814 W filter of
HST observed in2018 we see that there is a strong emission in V1 and thepossible explanation, then, is that the AGN, in the epoch
MNRAS , 1–18 (2020) Patr´ıcia da Silva et al. when the SINFONI data were observed (2005), was probablyin a lower state of activity than in 2018.N1 and N2 show emission-line ratios compatible withthose of LINERs (Fig. 11). Davies et al. (2017) obtainedoptical emission-line ratios for the central region of NGC613, but using 3D spectroscopy data with a lower spatialresolution than that of the data used in this work. The valuesof the [O iii ] λ β ratio, which is used in this work asa measure of the ionization degree, in the spectra of N1 andN2 are compatible, at the 3 σ level, mainly due of the highuncertainty of this ratio in the spectrum of N2. Even so,we must consider the possibility that N2 may indeed have ahigher ionization degree than N1.By comparing the HST images with the GMOS datacube images (Fig. 8), we see a variable stellar source, notpreviously reported in the literature, very close to [O i ] C ,inside the area corresponding to N1 (V2 in Fig. 16). Thetime interval of this variability was 14 yr and it might bedue to a supernova, since N1 is a stellar nucleus with thepresence of young stellar populations (see Paper II).As we see in Fig.14, the hard X-ray radial profile showsa width that is larger than the PSF of the instrument, repre-sented here by the radial profile of Mrk 202, a type 1 AGNobserved with the same instrument. Since there is no de-tectable second AGN, it is most likely that this additionalemission might be associated with the radio jet in the innerregion or that these photons are the result of scattering ofthe central AGN emission in the circumnuclear region.The AGN ionization cone is well represented by the [O iii ] λ iii ] λ iii ] λ iii ] λ ix , Ne x Ly α and Ne x Ly cont ), that thereis a strong photoionization in the ionization cone. Consid-ering the lower resolution of the X-ray data, the emissionpeak in soft X-rays is coincident with the position of [O i ] C and of the N1 and N2 centres, but we can also see a strong Figure 16.
Scheme of the scenario proposed by this work inthe central 2 arcsec. The spiral of molecular gas observed withALMA is represented by the blue area. The double stellar nu-cleus is represented by the two grey circles, whose sizes are basedon the emission area of N2 in the
HST images. V1 and V2 arethe two variable sources that were detected by comparing
HST images of filter
F 814 W in different epochs and of filter
F 606W with GMOS data, respectively. Their sizes represent the 3 σ uncertainty of the position considering the GMOS spaxel size inarcsec. [O i ] C is the centre of the emission of [O i ] λ σ uncertainty takinginto account the spaxels of GMOS. The scale of 0.5 arcsec (64 pc)is indicated in the figure. emission towards the jet and the outflows detected in H α and [O iii ] λ As described in section 1, NGC 613 has a well-known cir-cumnuclear ring of star formation. It is clearly seen in theimage of Br γ emission (presented in previous works and herein green in the composition of Fig. 10), indicating, in thiscase, the presence of young stars that ionize the gas of thestar-forming regions. B¨oker et al. (2008) and Falc´on-Barrosoet al. (2014) identified seven H ii regions along this circum-nuclear ring. In this work, we could identify eight distinct H ii regions (one H ii region in addition to the other sevendetected by B¨oker et al. 2008 and Falc´on-Barroso et al.2014). The ring observed in the [Fe ii ] λ γ image. Considering such granularity and also the fact thatthe [Fe ii ] emission is usually associated with shock heatingfrom SNRs, we conclude that there may be SNRs distributedalong the ring. B¨oker et al. (2008) have already suggestedthe presence of SNRs along the cirumnuclear ring, basedon the [Fe ii ] emission from this area and Falc´on-Barrosoet al. (2014) verified that the [Fe ii ]/Br γ ratio from the nu-clear spectrum of NGC 613 also suggests shock heating fromSNRs. We emphasize however that the higher granularity ofthe circumnuclear ring in the [Fe ii ] λ MNRAS000
F 606W with GMOS data, respectively. Their sizes represent the 3 σ uncertainty of the position considering the GMOS spaxel size inarcsec. [O i ] C is the centre of the emission of [O i ] λ σ uncertainty takinginto account the spaxels of GMOS. The scale of 0.5 arcsec (64 pc)is indicated in the figure. emission towards the jet and the outflows detected in H α and [O iii ] λ As described in section 1, NGC 613 has a well-known cir-cumnuclear ring of star formation. It is clearly seen in theimage of Br γ emission (presented in previous works and herein green in the composition of Fig. 10), indicating, in thiscase, the presence of young stars that ionize the gas of thestar-forming regions. B¨oker et al. (2008) and Falc´on-Barrosoet al. (2014) identified seven H ii regions along this circum-nuclear ring. In this work, we could identify eight distinct H ii regions (one H ii region in addition to the other sevendetected by B¨oker et al. 2008 and Falc´on-Barroso et al.2014). The ring observed in the [Fe ii ] λ γ image. Considering such granularity and also the fact thatthe [Fe ii ] emission is usually associated with shock heatingfrom SNRs, we conclude that there may be SNRs distributedalong the ring. B¨oker et al. (2008) have already suggestedthe presence of SNRs along the cirumnuclear ring, basedon the [Fe ii ] emission from this area and Falc´on-Barrosoet al. (2014) verified that the [Fe ii ]/Br γ ratio from the nu-clear spectrum of NGC 613 also suggests shock heating fromSNRs. We emphasize however that the higher granularity ofthe circumnuclear ring in the [Fe ii ] λ MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 Figure 17. (a) RGB composition of the
HST filters
F 814 W in red,
F 475 W in blue, and CO(3-2) from ALMA data cube in green.The red and cyan crosses represent the position of [O i ] C and the centre of N2, respectively, and its size the uncertainty of 3 σ , consideringthe size of the pixel of the HST images (0.03962 arcsec), since those positions were plotted taking the
HST images as reference. (b) RBcomposition of
F 475 W – F 814 W ( B – I in magnitude scale) in red and CO(3-2) from the ALMA data cube in green. The pink andwhite crosses represent the position of [O i ] C and the centre of N2, respectively, and its size the uncertainty of 3 σ , considering the sizeof the pixel of the HST images.
Table 4.
Parameters of the circumnuclear ring calculated fromthe images of Br γ emission, [Fe ii ] λ γ ± ± ii ] λ ± ± λ ± ± ± ± ± ∼ ∼ that the ring was observed both in hard X-ray (see Fig. 14)and radio (Hummel & Jorsater 1992) reinforces the idea ofSNRs along this region. Assuming that the ring is circu-lar, we determined the parameters in Table 4 and they arecompatible with what was estimated by Hummel & Jorsater(1992). Besides that, the radius is also compatible with thedetermination from Audibert et al. (2019).By calculating the emission-line ratios of the eight re-gions identified in the circumnuclear ring, we verified thatall ratios are compatible with the ones of H ii regions, exceptregions 3 and 8, which have ratios compatible with the onesof LINERs (Fig. 11). That can be easily explained when welook at Fig. 10, which shows that those regions are inside orclose to the contour that delineates the ionization cone. Thegas of those regions might be ionized inside the ionizationcone, generating those emission-line ratios. Another possibil-ity is that the emission-line ratios in these two regions resultfrom a contamination by the ionization cone emission. Re-gion 8 that has the highest ionization degree might be alsocontaminated by the radio jet and outflows as we see in Fig.15. In order to study how the molecular emission is related to the whole scenario that we are analysing here, we usedimages from an ALMA data cube of CO(3-2) and the imageof H λ i ] C nor N2 are the centre of the nuclear spiral.And, by matching the circumnuclear ring, certainly, N2 isnot the centre of the emission of H λ HST filtersin Figs. 1 and 5(d), we can see that there is a significant red-dening and obscuration by dust between N1 and N2. Thismay indicate that the spiral is probably bringing dust andgas to the centre of the galaxy. Since the spiral is locatedbetween N1 and N2 and we cannot see a clear double struc-ture in SINFONI data, together with the fact that we didnot detect any difference in the stellar populations of N1 andN2 (Paper II), we may have only one source in the centreof NGC 613. In that case, the dust extinction caused by thenuclear spiral results in an apparent division of the centralsource in the components N1 and N2.Fig. 17(b) shows, in red, areas affected by dust extinc-tion. We clearly see a connection between these areas and themolecular ring. This reinforces the hypothesis of the feedingof the circumnuclear ring by the bar proposed by B¨oker et al.(2008) and Audibert et al. (2019). The morphology also sug-gests little connection between the feeding coming from thebar and the nuclear spiral. We observe the same feature inH α image in Fig. 10 from Gadotti et al. (2019). MNRAS , 1–18 (2020) Patr´ıcia da Silva et al.
We drew a scheme to define the scenario in the central 2 arc-sec, taking into account all the phenomena that we describedin this paper (see Fig. 16). This scheme is what we definedhere as being the scenario that explains the coexistence ofthe multiple structures studied in this work. The nuclearspiral (in blue) passes between N1 and N2 and may bringgas and dust from the circumnuclear region. Regions N1and N2 are represented by the two grey circles, correspond-ing to the two components of the double stellar nucleus asappears in the
HST images. Since they are separated by astream of dust and the nuclear spiral, we are not entirelysure if they are separated regions. There are, therefore, twohypotheses to explain them: they are probably part of onecentral extended structure (as seen in Fig. 6) and the sep-aration might be an obscuration effect caused by the dust.The second hypothesis assumes that they are two separatedstructures orbiting the central AGN in the plane of the ob-servation, since we did not detect any difference of velocitiesin both regions (see Paper II). In Paper II we will resumethis discussion.V1 and V2, in Fig. 16, are the two variable sources.We believe that V1 is the AGN, since it is in the centre ofthe nuclear spiral and it is a very strong point-like sourcein the
HST images. If this is correct, the AGN has variableactivity that was also proposed by Audibert et al. (2019)when looking for possible fossil molecular outflows. N1 andN2 might be ionized by the central AGN. [O i ] C is not com-patible with the position of V1 and this is probably due todifferential extinction together with emission and reflectionof the [O i ] λ Multiwavelength analysis has shown that the galaxy NGC613 has a rich nuclear environment. The study of data cubesfrom telescopes in the NIR, optical, X-ray, and radio bands,besides
HST images, led us to the following findings: • In the optical band, the central region of NGC 613is characterized by an apparent double stellar nucleus. Wecalled the two stellar nuclei as N1 and N2. The least bright-est nucleus, as seen in the
HST images, coincides, within theerrors, with the emission of [O i ] λ i ] C in this work. The brightest nucleus as seen inthe HST image was defined as N2. The separation betweenthe two stellar components, as seen in the
HST images, is ∼
94 pc ( ∼ • The spectrum extracted from N1 has emission-linesratios compatible with the ones of LINERs, presenting fairlybroad forbidden lines (FWHM ∼
665 km s − ). On the otherhand, the spectrum from N2 has also emission-line ratios ofLINERs, but with slightly higher ionization degree than N1. • By comparing the images in the filter
F 814 W fromthe
HST observed in 2001 and in 2018, we detected evidenceof variability in a point between N1 and N2, whose distancefrom [O i ] C is 0.24 arcsec. In 2001 this source was not de-tected and in 2018 it was clearly visible. The position of [O i ] C and of this variable source are not compatible, but theyare very close to each other. Since this variable source ispoint-like in the HST images and is the centre of all nuclearstructure, it might be the central variable AGN. Therefore,the shift between the centre of the [O i ] λ i ] C , and this variable central source (the AGN in thiscase) might be due to scattering of the AGN emission andemission of the ionization cone, since the shift is towards theionization cone. • When we compare the optical data from
HST andGMOS, observed in 2001 and 2015, respectively, we verifythat N1 also suffered a variability in brightness. In 2001this source was fainter than in 2015. Considering the timeinterval and also that N1 has young stellar populations, thismight be an evidence of a supernova. • We found extended soft X-ray emission, closely asso-ciated with the ionization cone, seen in [O iii ] λ • The hard X-ray emission, besides having a strong cen-tral component, also presents a circumnuclear structure inthe form of a ring that has geometric parameters similar tothose seen in Br γ , [Fe ii ] λ λ ii ] λ γ image, rein-forces this hypothesis. • The profile of the central hard X-ray emission isslightly broader than the PSF by 0.33 arcsec. We interpretthis as possibly due to circumnuclear scattering. • From the optical (SIFS) and NIR (SINFONI) datawe identified 10 H ii regions. Eight of them are part of thecircumnuclear ring (already observed in other works), whilethe other two are further away. • There are at least three H ii regions in the circumnu-clear ring whose emission-line ratios are affected by emissionfrom the ionization cone or partially ionized by the centralsource. As a consequence, they present spectra that resemblethe ones of LINERs. • We confirm, by analysing the
HST data together withthe CO(3-2) image from ALMA, that the molecular gas ringis being fed by the bar along two arms. The nuclear spiral,which passes between N1 and N2, might be bringing dustand gas to the centre, causing the obscuration. Also it mightbe dividing an extended central stellar structure in two, thatis the double stellar nucleus that we see in the optical band,due to dust obscuration.
ACKNOWLEDGEMENTS
This work is based on observations obtained at the Gem-ini Observatory (processed using the Gemini iraf pack-age), which is operated by the Association of Universitiesfor Research in Astronomy, Inc., under a cooperative agree-ment with the NSF on behalf of the Gemini partnership: theNational Science Foundation (United States), the NationalResearch Council (Canada), CONICYT (Chile), the Aus-tralian Research Council (Australia), Minist´erio da Ciˆencia,Tecnologia e Inova¸c˜ao (Brazil), and Ministerio de Ciencia,
MNRAS000
MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 Tecnolog´ıa e Innovaci´on Productiva (Argentina). This workis also based on observations made with the NASA/ESA
Hubble Space Telescope , obtained from the Data Archive atthe Space Telescope Science Institute, which is operated bythe Association of Universities for Research in Astronomy,Inc., under NASA contract NAS 5-26555. This research hasalso made use of the NASA/IPAC Extragalactic Database(NED), which is operated by the Jet Propulsion Laboratory,California Institute of Technology, under contract with theNational Aeronautics and Space Administration. This re-search has also made use of data obtained from the
Chandra
Data Archive and the
Chandra
Source Catalog, and softwareprovided by the
Chandra
X-ray Center (CXC) in the appli-cation packages ciao , chips , and sherpa and used data fromobservations collected at the European Southern Observa-tory under ESO programme 076.B-0646(A). This paper alsomakes use of the following ALMA data: ADS/JAO.ALMA2015.1.00404.S. ALMA is a partnership of ESO (representingits member states), NSF (USA) and NINS (Japan), togetherwith NRC (Canada), MOST and ASIAA (Taiwan), andKASI (Republic of Korea), in cooperation with the Republicof Chile. The Joint ALMA Observatory is operated by ESO,AUI/NRAO and NAOJ. We thank CNPq (Conselho Na-cional de Desenvolvimento Cient´ıfico e Tecnol´ogico), undergrant 141766/2016-6, and FAPESP (Funda¸c˜ao de Amparo `aPesquisa do Estado de S˜ao Paulo), under grant 2011/51680-6, for supporting this work. We also thank professor Giusep-pina Fabbiano for offering suggestions on Chandra data cubeanalysis and Dr. Roderik Overzier for reading and revisingthis article.
REFERENCES
Asmus D., Gandhi P., H¨onig S. F., Smette A., Duschl W. J., 2015,MNRAS, 454, 766Audibert A., et al., 2019, A&A, 632, A33Balmaverde B., et al., 2012, A&A, 545, A143Beall J. H., 2003, Chinese Journal of Astronomy and Astro-physics, 3, 373Bianchi S., Guainazzi M., Chiaberge M., 2006, A&A, 448, 499B¨oker T., Falcon-Barroso J., Knapen J. H., Schinnerer E., Al-lard E., Ryder S., 2007, in Vazdekis A., Peletier R., eds, IAUSymposium Vol. 241, Stellar Populations as Building Blocksof Galaxies. pp 497–498, doi:10.1017/S1743921307008873B¨oker T., Falc´on-Barroso J., Schinnerer E., Knapen J. H., RyderS., 2008, AJ, 135, 479Burtscher L., et al., 2015, A&A, 578, A47Cappellari M., Emsellem E., 2004, PASP, 116, 138Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345, 245Castangia P., Tarchi A., Henkel C., Menten K. M., 2008, A&A,479, 111Castangia P., Panessa F., Henkel C., Kadler M., Tarchi A., 2013,MNRAS, 436, 3388Cid Fernandes R., Mateus A., Sodr´e L., Stasi´nska G., GomesJ. M., 2005, MNRAS, 358, 363Combes F., et al., 2019, A&A, 623, A79Davies R. L., et al., 2017, MNRAS, 470, 4974Elmegreen B. G., 1994, ApJ, 425, L73Evans D. A., Lee J. C., Kamenetska M., Gallagher S. C., KraftR. P., Hardcastle M. J., Weaver K. A., 2006, ApJ, 653, 1121Falc´on-Barroso J., B¨oker T., Schinnerer E., Knapen J. H.,Ryder S., 2008, in Bureau M., Athanassoula E., BarbuyB., eds, IAU Symposium Vol. 245, Formation and Evo- lution of Galaxy Bulges. pp 177–180 ( arXiv:0709.0353 ),doi:10.1017/S1743921308017584Falc´on-Barroso J., Ramos Almeida C., B¨oker T., Schinnerer E.,Knapen J. H., Lan¸con A., Ryder S., 2014, MNRAS, 438, 329Gadotti D. A., et al., 2019, MNRAS, 482, 506Gonzalez R. C., Woods R. E., 2002, Digital image processing (2nded., Upper Saddle River, NJ: Prentice Hall)Goulding A. D., Alexander D. M., 2009, MNRAS, 398, 1165Greene J. E., Pooley D., Zakamska N. L., Comerford J. M., SunA.-L., 2014, ApJ, 788, 54G¨ultekin K., et al., 2009, ApJ, 698, 198Hardcastle M. J., Croston J. H., 2005, MNRAS, 363, 649Harrison C. M., Costa T., Tadhunter C. N., Fl¨utsch A., KakkadD., Perna M., Vietri G., 2018, Nature Astronomy, 2, 198Ho L. C., 2008, ARA&A, 46, 475Hummel E., Jorsater S., 1992, A&A, 261, 85Hummel E., Jorsater S., Lindblad P. O., Sandqvist A., 1987,A&A, 172, 51Kauffmann G., et al., 2003, MNRAS, 346, 1055Kewley L. J., Dopita M. A., Sutherland R. S., Heisler C. A.,Trevena J., 2001, ApJ, 556, 121Kinkhabwala A., et al., 2002, ApJ, 575, 732Kondratko P. T., et al., 2006, ApJ, 638, 100Lucy L. B., 1974, AJ, 79, 745Lynden-Bell D., 1969, Nature, 223, 690Marinucci A., Risaliti G., Wang J., Nardini E., Elvis M., FabbianoG., Bianchi S., Matt G., 2012, MNRAS, 423, L6M´armol-Queralt´o E., et al., 2009, ApJ, 705, L199May D., Steiner J. E., 2017, MNRAS, 469, 994Menezes R. B., Steiner J. E., Ricci T. V., 2014, MNRAS, 438,2597Menezes R. B., da Silva P., Ricci T. V., Steiner J. E., May D.,Borges B. W., 2015, MNRAS, 450, 369Menezes R. B., Ricci T. V., Steiner J. E., da Silva P., Ferrari F.,Borges B. W., 2019, MNRAS, 483, 3700Miyamoto Y., Nakai N., Seta M., Salak D., Nagai M., Kaneko H.,2017, Publications of the Astronomical Society of Japan, 69,83Miyamoto Y., Seta M., Nakai N., Watanabe Y., Salak D., IshiiS., 2018, Publications of the Astronomical Society of Japan,70, L1Nasonova O. G., de Freitas Pacheco J. A., Karachentsev I. D.,2011, A&A, 532, A104Netzer H., 2013, The physics and evolution of active galactic nu-clei. Cambridge University Press, CambridgeOsterbrock D., 2006, Astrophysics of gaseous nebulae and activegalactic nuclei. University Science Books, Sausalito, CaliforniaRicci T. V., Steiner J. E., Menezes R. B., 2011, ApJ, 734, L10Richardson W. H., 1972, Journal of the Optical Society of Amer-ica (1917-1983), 62, 55S´anchez-Bl´azquez P., et al., 2006, MNRAS, 371, 703Schawinski K., Thomas D., Sarzi M., Maraston C., Kaviraj S.,Joo S.-J., Yi S. K., Silk J., 2007, MNRAS, 382, 1415Storchi-Bergmann T., McGregor P. J., Riffel R. A., Sim˜oes LopesR., Beck T., Dopita M., 2009, MNRAS, 394, 1148Storchi-Bergmann T., Lopes R. D. S., McGregor P. J., RiffelR. A., Beck T., Martini P., 2010, MNRAS, 402, 819Vasudevan R. V., Fabian A. C., Gandhi P., Winter L. M.,Mushotzky R. F., 2010, MNRAS, 402, 1081Veron-Cetty M.-P., Veron P., 1986, A&AS, 66, 335V´eron-Cetty M. P., V´eron P., 2006, A&A, 455, 773Wang J., et al., 2011, ApJ, 742, 23Winge C., Riffel R. A., Storchi-Bergmann T., 2009, ApJS, 185,186Young A. J., Wilson A. S., Shopbell P. L., 2001, ApJ, 556, 6van Dokkum P. G., 2001, PASP, 113, 1420MNRAS , 1–18 (2020) Patr´ıcia da Silva et al.
APPENDIX A: SPECTRA OF THE OBSERVEDREGIONS AND EMISSION-LINEDECOMPOSITIONS
As explained in section 3.4, we extracted a spectrum ofeach identified region in Fig. 10. The extraction was per-formed using circular areas, whose diameter was taken asthe FWHM of each instrument: for regions 1 to 10, we usedthe SIFS data cube and, for N1 and N2 regions, we used theGMOS data cube. Figs. A1, A3 and A4 show the blue andred parts of the extracted spectra of each region.One can notice that the spectra from regions N1, N2, 2,3, 7, and 8 have blended emission lines. In order to calculatethe emission-line ratios of each region, it was necessary todecompose the blended lines in two Gaussian sets. First, wedetermined an uncertainty to the flux associated with eachwavelength of the spectrum. For this, we calculated the stan-dard deviation of the values in a specific wavelength range,without any emission line, of the extracted spectra. Thesevalues of standard deviation were taken as the uncertaintyof the fluxes in the average wavelengths of the ranges used inthis estimate. Then, we made an interpolation of the values,in order to obtain a value of uncertainty to each wavelengthof the spectrum.The first Gaussian fit involved the lines [S ii ] λλ ii ] was fitted by a sum of twoGaussians functions (green+blue), each one with a specificwidth and velocity.It is important to mention that the maximum and min-imum values admitted for the ratio of the integrated fluxesof the [S ii ] λ ii ] λ ii ] λλ α as a sum of two Gaussians for each line, which resulted in aset of three green Gaussians and a set of three blue Gaus-sians in Figs. A2 and A5. We assumed that, in each set, theGaussians have the same width and redshifts as the corre-sponding Gaussians of the [S ii ] λλ ii ] λλ α /H β ratio (Balmer decrement). By using this ratioand also the extinction law determined by Cardelli et al.(1989), we applied the interstellar extinction correction toeach spectra. Then, we calculated the integrated fluxes ofthe H β , [O iii ] λ i ] λ ii ] λλ α e [S ii ] λλ This paper has been typeset from a TEX/L A TEX file prepared bythe author. MNRAS000
As explained in section 3.4, we extracted a spectrum ofeach identified region in Fig. 10. The extraction was per-formed using circular areas, whose diameter was taken asthe FWHM of each instrument: for regions 1 to 10, we usedthe SIFS data cube and, for N1 and N2 regions, we used theGMOS data cube. Figs. A1, A3 and A4 show the blue andred parts of the extracted spectra of each region.One can notice that the spectra from regions N1, N2, 2,3, 7, and 8 have blended emission lines. In order to calculatethe emission-line ratios of each region, it was necessary todecompose the blended lines in two Gaussian sets. First, wedetermined an uncertainty to the flux associated with eachwavelength of the spectrum. For this, we calculated the stan-dard deviation of the values in a specific wavelength range,without any emission line, of the extracted spectra. Thesevalues of standard deviation were taken as the uncertaintyof the fluxes in the average wavelengths of the ranges used inthis estimate. Then, we made an interpolation of the values,in order to obtain a value of uncertainty to each wavelengthof the spectrum.The first Gaussian fit involved the lines [S ii ] λλ ii ] was fitted by a sum of twoGaussians functions (green+blue), each one with a specificwidth and velocity.It is important to mention that the maximum and min-imum values admitted for the ratio of the integrated fluxesof the [S ii ] λ ii ] λ ii ] λλ α as a sum of two Gaussians for each line, which resulted in aset of three green Gaussians and a set of three blue Gaus-sians in Figs. A2 and A5. We assumed that, in each set, theGaussians have the same width and redshifts as the corre-sponding Gaussians of the [S ii ] λλ ii ] λλ α /H β ratio (Balmer decrement). By using this ratioand also the extinction law determined by Cardelli et al.(1989), we applied the interstellar extinction correction toeach spectra. Then, we calculated the integrated fluxes ofthe H β , [O iii ] λ i ] λ ii ] λλ α e [S ii ] λλ This paper has been typeset from a TEX/L A TEX file prepared bythe author. MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 Figure A1.
Blue and red intervals of the spectra of regions N1 and N2 extracted from the GMOS data cube.
Figure A2.
Decomposition of the blended H α + [N ii ] λλ ii ] λλ , 1–18 (2020) Patr´ıcia da Silva et al.
Figure A3.
Blue and red intervals of the spectra of regions 1, 2, 3, 4, and 5 extracted from the SIFS data cube.MNRAS000
Blue and red intervals of the spectra of regions 1, 2, 3, 4, and 5 extracted from the SIFS data cube.MNRAS000 , 1–18 (2020) he nuclear region of NGC 613 Figure A4.
Blue and red intervals of the spectra of regions 6, 7, 8, 9, and 10 extracted from the SIFS data cube.MNRAS , 1–18 (2020) Patr´ıcia da Silva et al.
Figure A5.
Decomposition of the blended H α + [N ii ] λλ ii ] λλ000