Exploring the AGN-Merger Connection in Arp 245 I: Nuclear Star Formation and Gas Outflow in NGC 2992
Muryel Guolo-Pereira, Daniel Ruschel-Dutra, Thaisa Storchi-Bergmann, Allan Schnorr-Müller, Roberto Cid Fernandes, Guilherme Couto, Natacha Dametto, Jose A Hernandez-Jimenez
MMNRAS , 1–20 (2021) Preprint 27 January 2021 Compiled using MNRAS L A TEX style file v3.0
Exploring the AGN-Merger Connection in Arp 245 I: Nuclear StarFormation and Gas Outflow in NGC 2992
Muryel Guolo-Pereira ★ , Daniel Ruschel-Dutra , Thaisa Storchi-Bergmann ,Allan Schnorr-Müller , Roberto Cid Fernandes , Guilherme Couto ,Natacha Dametto , Jose A Hernandez-Jimenez Departamento de Física - CFM - Universidade Federal de Santa Catarina, 476, 88040-900 Florianópolis, SC, Brazil Departamento de Astronomia, Universidade Federal do Rio Grande do Sul. Av. Bento Goncalves 9500, 91501-970 Porto Alegre, RS, Brazil Centro de Astronomía (CITEVA), Universidad de Antofagasta, Avenida Angamos 601, Antofagasta, Chile Departamento de Ciencias Físicas, Universidad Andrés Bello, Fernández Concha 700, Las Condes, Santiago, Chile
Accepted 2021 January 25. Received 2021 January 25; in original form 2019 November 28.
ABSTRACT
Galaxy mergers are central to our understanding of galaxy formation, especially within the context of hierarchical models.Besides having a large impact on the star formation history, mergers are also able to influence gas motions at the centre ofgalaxies and trigger an Active Galactic Nucleus (AGN). In this paper, we present a case study of the Seyfert galaxy NGC 2992,which together with NGC 2993 forms the early-stage merger system Arp 245. Using Gemini Multi-Object Spectrograph (GMOS)integral field unit (IFU) data from the inner 1.1 kpc of the galaxy we were able to spatially resolve the stellar populations, theionisation mechanism and kinematics of ionised gas. From full spectral synthesis, we found that the stellar population is primarilycomposed by old metal-rich stars (t ≥ 𝑍 ≥ . (cid:12) ), with a contribution of at most 30 per cent of the light from a youngand metal-poor population (t ≤
100 Myr, 𝑍 ≤ . (cid:12) ). We detect H 𝛼 and H 𝛽 emission from the Broad Line Region (BLR)with a Full Width at Half Maximum (FWHM) of ∼ − . The Narrow Line Region (NLR) kinematics presents two maincomponents: one from gas orbiting the galaxy disk and a blueshifted (velocity ≈ -200 km s − ) outflow, possibly correlated withthe radio emission, with mass outflow rate of ∼ (cid:12) yr − and a kinematic power of ∼ × erg s − ( (cid:164) 𝐸 𝑜𝑢𝑡 / 𝐿 𝑏𝑜𝑙 ≈ Key words:
Galaxies: individual (Arp 245, NGC 2992) − Galaxies: Seyfert − Galaxies: interactions − Galaxies: stellar content − Galaxies: kinematics and dynamics
The vast majority of galaxies with a spheroid host a supermassiveblack hole (SMBH) in their centres, with a mass range of ∼ − M (cid:12) . In some galaxies, these objects are active and emit intenseradiation due to the accretion of matter onto the SMBH through aninner region called accretion disk (Begelman et al. 1984; Petersonet al. 1998; Event Horizon Telescope Collaboration et al. 2019).Active galactic nuclei (AGN) in Seyfert galaxies are classified astype 1 and 2, depending on the presence of broad emission lines ofthe Balmer series, with intermediate types between these two alsopossible; these broad lines are produced in a region called BroadLine Region (BLR).The importance of AGN in the evolution of the host galaxy isnot completely understood, but there is evidence pointing to a co-evolution scenario between the two. The 𝑀 BH − 𝜎 ∗ relation (Ferrarese& Merritt 2000; Kormendy & Ho 2013) – which relates the mass of ★ E-mail:[email protected] the SMBH with the stellar velocity dispersion in the galactic bulge– and the simultaneous peak between star formation rate (SFR) andSMBH accretion rate in cosmological studies at 𝑧 ≈ © a r X i v : . [ a s t r o - ph . GA ] J a n M. Guolo-Pereira et al. driven winds are a possible mechanism for preventing further growthof the host galaxy. These winds can have outflow velocities as highas 1000 km s − (Rupke & Veilleux 2011; Greene et al. 2012) and aslow as well ≈ −
200 km s − , with mass outflow rates in nearbygalaxies averaging at a few solar masses per year (e.g. Riffel &Storchi-Bergmann 2011; Crenshaw et al. 2015; Revalski et al. 2018).Theoretical studies suggest large scale events like major mergersare the dominant processes leading to SMBH growth at high masses(Menci et al. 2014). At high redshifts (z > 2), major mergers have alsobeen proposed as fuelling mechanisms of the fastest-growing SMBHs(Treister et al. 2012). A major merger can destabilise large quantitiesof gas, driving massive inflows towards the nuclear region of galaxiesand triggering bursts of star formation and nuclear activity (Hopkins& Quataert 2010; Blumenthal & Barnes 2018). Many studies havefound that the most luminous AGN are preferentially hosted by galaxymergers (Schawinski et al. 2010; Glikman et al. 2015; Fan et al. 2016).At lower luminosities, several studies have found a higher incidenceof galaxies with signatures of interactions in AGN hosts as comparedto control samples (Koss et al. 2010) and particularly in close galaxypairs (Ellison et al. 2011; Satyapal et al. 2014) suggesting kinematicpairs are conducive environments for black hole growth.One possible consequence of gas inflow is the circumnuclear starformation (Hopkins et al. 2008; Hopkins & Quataert 2010). Suchmerger-driven inflows can both increase the star formation rate (Elli-son et al. 2013; Pan et al. 2019) and modify the metallicities gradientsof the galaxies (Barrera-Ballesteros et al. 2015). However, classicalemission-line SFR probes (e.g. Kennicutt 1998) and metallicity cal-ibrations (e.g. Alloin et al. 1979; Pettini & Pagel 2004) cannot beapplied to active galaxies, due to the contamination of the AGN ioni-sation to the emission lines. In this way, the use of stellar populationssynthesis, by performing full spectral fitting, can reveal clues on thestar formation history (SFH) and chemical evolution of AGN hostgalaxies and its content.In this context, we present high spatial and spectral resolution op-tical IFU observations of the inner 1.1 kpc of NGC 2992. NGC 2992is a nearby interacting Seyfert galaxy seen almost edge-on ( 𝑖 ∼ ◦ ,Marquez et al. 1998) at a distance, as measured using the Tully-Fisherrelation, of 38 Mpc (Theureau et al. 2007), which translates into aprojected angular scale of ∼
150 pc per arcsec. Alongside with thestarburst galaxy NGC 2993 and the tidal dwarf galaxy A245N itforms the system Arp 245. Using hydrodynamic simulations, Ducet al. (2000) reported the system is currently at an early stage of theinteraction, about ≈
100 Myr after pericentre passage. Nevertheless,tidal tails have already developed, which can be seen by optical im-ages (Fig. 1), in CO and HI lines (Duc et al. 2000) and infrared (IR)images (García-Bernete et al. 2015).The galaxy nuclear activity has been the subject of several studies,partly due to its variability as seen both in X-rays and in the optical(Trippe et al. 2008), even leading to changes in spectral classifica-tion. Early spectra published by Shuder (1980), Veron et al. (1980),and Ward et al. (1980) all show the presence of a weak broad H 𝛼 component, originated in the BLR, but no detectable correspond-ing broad H 𝛽 component, leading to its original classification as aSeyfert 1.9. Observations from 1994 (Allen et al. 1999), however,display no broad H 𝛼 component, thus having the classification ofSeyfert 2. Gilli et al. (2000) reported a broad H 𝛼 emission again, justto disappear once again when observed in 2006 by L. Trippe et al.(2008). More recently Schnorr-Müller et al. (2016), after subtractingthe narrow component, were able to detect the broad H 𝛽 componentfor the first time, then classifying the galaxy as a Seyfert 1.8.In the radio, 6 cm observations show a double lobe 8-shaped struc-ture of about 8 arcsec ( ∼ − ◦ (Ulvestad &Wilson 1984). From IR observations, Chapman et al. (2000) suggestthe best interpretation is that this structure is related to expandingplasma bubbles, possibly carried by internal jets from the AGN.More recently, using radio polarimetry, Irwin et al. (2017) foundanother double-lobed radio morphology within its spiral disc, whichwas revealed in linearly polarized emission but not in total intensityemission. This second structure by Irwin et al. (2017) is much moreextended than the one found by Ulvestad & Wilson (1984) reachingseveral kpcs from the nucleus, being interpreted by the authors as arelic of an earlier episode of AGN activity.NGC 2992 gas kinematics is complex as found by several longslit spectroscopy studies (e.g. Marquez et al. 1998; Veilleux et al.2001) as well as by several Integral Field Unit observations (García-Lorenzo et al. 2001; Friedrich et al. 2010; Müller-Sánchez et al.2011). These observations show, at most positions, the presence ofa double component line profile. While one component follows thegalaxy rotation curve, the other is interpreted as outflowing gas bythe authors. Using The Multi Unit Spectroscopic Explorer (MUSEBacon et al. 2010) at the Very Larga Telescope (VLT) Mingozzi et al.(2019) clearly show the presence of a kpc-scale, bipolar outflow witha large opening angle.This paper is the first in a series which will explore the possibleconnection between merger and AGN in Arp 245. In this first pa-per, we analyse both the stellar populations and the properties of theionised gas of NGC 2992, in order to better understand the physicalconditions of the nuclear region and the mutual role of nuclear ac-tivity and the interaction may play. In a second paper (Guolo-Pereiraet al., in prep., Paper II) we will observationally explore the effectsof the interaction in NGC 2993. While in the third one (Lösch et al.,in prep, Paper III) we will present a modern version of the hydro-dynamics simulations by Duc et al. (2000) focusing on the possibletriggering of the AGN during the merger.The present paper is organised as follows: in section 2 we de-scribe the observations and data reduction; section 3 we present thestellar populations synthesis and its spatial distribution; the ionisedgas kinematics is described in section 4; ionisation mechanisms arediscussed in section 5; section 6 presents our discussion of the mainresults and their significance; and finally section 7 contains our con-clusions. This work is mostly based on data obtained with the Integral FieldUnit (IFU) of the Gemini Multi Object Spectrograph (GMOS) at theGemini South telescope on the night of February 15, 2018 (Gem-ini project GS-2018A-Q-208, P.I. Ruschel-Dutra). The observationsconsisted of two adjacent IFU fields (covering 5 (cid:48)(cid:48) × . (cid:48)(cid:48) × . (cid:48)(cid:48)
75 sky observationfor atmospheric lines removal.The spatial resolution is limited by seeing, which was at 0 . (cid:48)(cid:48) 𝜎 instr ≈
40 km s − , obtained from the Copper-Argon arclamp lines.The reduction procedures were performed using the Gmos Ifu RE- MNRAS , 1–20 (2021) apping the Inner kpc of NGC 2992 Figure 1.
Top Left: LRGB image composition of Arp 245 from Block (2011). Top Right: GMOS r’ band acquisition image of NGC 2992. Middle Left: Zoomedacquisition image, and the Ulvestad & Wilson (1984) radio emission in purple (contours : 3, 6, 9, 15, 30, 50, 70, 90 per cent of the peak of 7.2 mJy). The dashedrectangles show the two IFU FoV’s. Middle Right: IFU continuum mean flux in the FoV. The black cross represents the continuum nuclear position in all figuresof the paper. All maps of the paper were smoothed using a Gaussian kernel with FHWM with half of the spatial resolution, see section 2. Bottom, from top tobottom: Spectra corresponding to the regions marked as A, B and C, with spectral flux density in units of 10 − erg s − cm − Å − . North is up and East to theleft in this and all figures of the paper. Duction Suite (GIREDS) . The process comprised the usual steps:bias subtraction, flatfielding, trimming, wavelength calibration, tel-luric emission subtraction, relative flux calibration, the building ofthe data cubes, and finally the registering and combination of the 12individual exposures. The final cube was built with a spatial samplingof 0 . (cid:48)(cid:48) × . (cid:48)(cid:48)
1, resulting in a FoV with ≈ GIREDS is a set of python scripts to automate the reduction of GMOS IFUdata, available at GitHub. the borders and superimposing the FoVs, it results in a total usefulcombined FoV, with total angular coverage of 29 arcsec around thenucleus. The combined FoV is centred at RA = 9h45m42s and Dec= -14 𝑜 (cid:48) (cid:48)(cid:48) , while the continuum peak is shifted at Δ RA = 0.1 (cid:48)(cid:48) and Δ Dec = -0.1 (cid:48)(cid:48) .Many Gemini data cubes show certain “instrumental fingerprints”,in the form of vertical stripes in the reconstructed data cube. Suchfeatures also have specific spectral signatures. In order to identifyand remove these instrumental features, we use a technique based
MNRAS , 1–20 (2021)
M. Guolo-Pereira et al. on Principal Component Analysis (PCA, Menezes et al. 2014). Thefingerprint removal is fundamental for the stellar population synthesisbecause its effect is more apparent for lower instrumental counts.The data cube spectra were corrected for Galactic extinction usingthe Cardelli et al. (1989) extinction curve (CCM) for an 𝐴 𝑉 = . − (see section 4).We show in the upper panel of the Fig. 1 an LRGB image ofthe Arp 245 system (Left) and the acquisition image taken with ar’ filter (Right). In the middle panel we show a zoomed version ofthe acquisition image superimposed with the the Ulvestad & Wilson(1984) figure of 8-shaped radio emission of the galaxy (Left) and themean continuum flux in the IFU’s FoV (Right), while in the bottompanels we show three sample spectra from representative regionsmarked, respectively from top to bottom, A, B and C. The prominentdust lane seen in the acquisition image, and in the stellar continuumimage from the IFU, has a reported 𝐴 𝑉 greater than 3 (Colina et al.1987; Schnorr-Müller et al. 2016). Spectra from this region of highextinction have no measurable stellar component above the noiselevel, as exemplified by the spectrum extracted from C. The nuclearspectrum shows a broad H 𝛼 component, and a faint broad H 𝛽 isalso present, in contrast with its original Seyfert 1.9 classificationbut in agreement with its the current state, Seyfert 1.8, proposed bySchnorr-Müller et al. (2016). In order to obtain the spatially resolved SFH, and to fit the gasemission features, free from the stellar population contamination, weperformed stellar population synthesis on the data cube by employingthe starlight full spectra fitting code (Cid Fernandes et al. 2005).Briefly, starlight fits an observed stellar spectrum 𝑂 𝜆 with amodel 𝑀 𝜆 built from a linear combination of 𝑁 spectral components.Dust attenuation is treated as if due to a foreground screen andparametrised by the 𝑉 -band extinction 𝐴 𝑉 . Stellar kinematics ismodelled with a Gaussian line-of-sight velocity distribution centredat velocity 𝑣 ∗ , and with a dispersion 𝜎 ∗ . Besides the model spectrum 𝑀 𝜆 , the code outputs the corresponding values of 𝐴 𝑉 , 𝑣 ★ , 𝜎 ★ , anda population vector (cid:174) 𝑥 containing the percentage contribution of eachbase population to the flux at a chosen normalisation wavelength 𝜆 ,which in our fits was chosen as 𝜆 = 5500 Å. The code minimises the 𝜒 of the fit, though a more convenient and intuitive figure of meritto assess the quality of the fit is through adev parameter, defined asthe mean value of | 𝑂 𝜆 − 𝑀 𝜆 | / 𝑂 𝜆 over the fitted wavelengths.We note that, while there are many different spectral fitting codesand techniques available, cross-comparison of these various tech-niques have shown to yield rather consistent results (e.g. Kolevaet al. 2008; Ferré-Mateu et al. 2012; Maksym et al. 2014; Mentz et al.2016; Cid Fernandes 2018). Therefore, rather than trying differentfitting codes, we chose to maintain starlight and investigate thedifferences introduced by the choice of models and stellar libraries,which are known to produce variations in the recovered properties(e.g. Maraston & Strömbäck 2011; Chen et al. 2010; Wilkinson et al.2017; Baldwin et al. 2018; Dametto et al. 2019). This section discusses the stellar population models and stellar li-braries used in this work, and how we can account for the contributionof a featureless continuum due to the AGN in NGC 2992. The key ingredients in the stellar population synthesis methodare the spectral base elements, i.e. those spectra available to thefitting code to combine in order to recover the galaxy properties.The base elements are, usually, Simple Stellar Population (SSP),i.e. a collection of stars with same age and metallicity, taken froman evolutionary model which are constructed using a stellar spectralibrary and an Initial Mass Function (IMF).In this work we have used SSPs from two models: Bruzual &Charlot (2003, BC03), which uses the STELIB (Le Borgne et al.2003) spectral library, and Maraston & Strömbäck (2011, M11),which employs the MILES (Sánchez-Blázquez et al. 2006) spectrallibrary. For both models/libraries we use the same Chabrier (2003)IMF and instantaneous star formation bursts. We found varying theIMF produces only minor variations in the recovered SFH whencompared to the changes introduced by using distinct models. Theformer is outside the scope of this work, and we refer the readerinterested in this particular effect to Chen et al. (2010), Ge et al.(2019) and references therein. From the SSPs provided by BC03we choose a set of 45 representative elements divided into threemetallicities (0 . , . , . (cid:12) ) and an age range covering from t(Gyr) = .
001 to t(Gyr) =
13. From those available by M11 with the MILESlibrary we chose a set of 40 SSPs with age and metallicity coverage assimilar as possible to the BC03 ones: three metallicities (0 . , . , . (cid:12) ) and ages ranging from t(Gyr) = .
007 to t(Gyr) =
13. Thecomplete list of SSPs used from each model is shown in Table 1.It is important to mention that the MILES Library does not providemetal-poor stars ( 𝑍 = . (cid:12) ) younger than 55 Myr (0.055 Gyr),and neither metal-rich stars ( 𝑍 = . (cid:12) ) younger than 100 Myr (0.1Gyr), while BC03/STELIB provide the same age range for the threemetallicities (see Table 1).We define a reduced population vector with three age ranges: 𝑡 <
100 Myr; 100 Myr ≤ 𝑡 ≤ 𝑡 > . 𝑥 𝑌 , 𝑥 𝐼 and 𝑥 𝑂 , where‘ 𝑌 (cid:48) , ‘ 𝐼 (cid:48) and ‘ 𝑂 (cid:48) stands for young, intermediate and old respectively.The division in a reduced population vector has been used in severalworks in stellar population synthesis (e.g. Cid Fernandes et al. 2005;Riffel et al. 2010a; Cid Fernandes et al. 2013; González Delgadoet al. 2015; Mallmann et al. 2018), while the range values may beslightly distinct from one another, the ones adopted here are the mostcommon. Also, for the young population the cutting value of 100 Myris the estimated age of the pericentre passage between NGC2992 andits companion (Duc et al. 2000).When dealing with stellar populations in Seyfert galaxies, a powerlaw ( 𝐹 𝜆 ∝ 𝜆 − 𝛼 ) featureless continuum (FC) should be added to thespectral base (Koski 1978) to account for the AGN emission. Valuesfor 𝛼 are between 0 . − . 𝛼 = . 𝛼 index for the optical spectrum of NGC 2992. We extractedtwo integrated spectra: an inner one, hereafter 𝑆 ( 𝜆 ) , defined as thesum of spectra within a circular region with radius of 0 . (cid:48)(cid:48)
4, and anexternal one, 𝐸 ( 𝜆 ) , in an annulus with inner radius of 1 (cid:48)(cid:48) and outerradius of 1 . (cid:48)(cid:48)
4, both centred at the peak of continuum emission. Wemodel the 𝑆 ( 𝜆 ) spectrum as a combination of 𝐸 ( 𝜆 ) and a power lawFC ( 𝐹 𝜆 ∝ 𝜆 − 𝛼 ), both reddened by the same factor and normalised inthe same wavelength, 𝜆 = MNRAS , 1–20 (2021) apping the Inner kpc of NGC 2992 Table 1.
SSPs used in the stellar population synthesis.Model BC03 M11Spectral Library STELIB MILESMetallicity ( 𝑍 (cid:12) ) 0.2 1.0 2.5 0.5 1.0 2.0Ages (Gyr) 0.001, 0.003, 0.005,0.01, 0.025, 0.04,0.1, 0.3, 0.6, 0.9,1.5, 3.0, 5.0,11.0, 13.0 0.055, 0.065,0.075, 0.085,0.1, 0.15,0.3, 0.6,0.9 0.007, 0008,0.009, 0.01,0.015, 0.025,0.04, 0.055,0.065, 0.075,0.085, 0.1,0.15, 0.3, 0.9,1.5, 3.0, 5.0,11.0, 13.0 0.1, 0.15,0.3, 0.6,0.9, 1.5,3.0, 5.0,11.0, 13.0 (cid:18) 𝑆 𝑚 ( 𝜆 ) 𝑆 𝑚 ( 𝜆 ) (cid:19) = (cid:20) ( − 𝒑 ) (cid:18) 𝐸 ( 𝜆 ) 𝐸 ( 𝜆 ) (cid:19) + 𝒑 (cid:18) 𝜆𝜆 (cid:19) − 𝜶 (cid:21) × − . 𝜹𝑨 𝑽 𝑞 𝜆 (1)where 𝑆 𝑚 ( 𝜆 ) is the model spectrum, 𝑝 is the fraction contributionof the FC, 𝛿𝐴 𝑉 is the difference in extinction between the 𝑆 ( 𝜆 ) and 𝐸 ( 𝜆 ) regions. The values which minimise the residuals between 𝑆 ( 𝜆 ) and 𝑆 𝑚 ( 𝜆 ) are 𝑝 = 0.23, 𝛿𝐴 𝑉 = 2.2 and 𝛼 =1.7.It should be noted we are over-simplifying by assuming that 𝑆 ( 𝜆 ) is the combination of 𝐸 ( 𝜆 ) and an FC, i.e. the stellar spectrum ofboth are the same. However, we argue this is just a method to get adata-based value for 𝛼 , instead of just assuming a fixed value whichis even less physically justified. Also, no further interpretation willbe given in this regard besides the adoption of this 𝛼 value to producethe FC base element . In order to perform the spatially resolved stellar synthesis withenough signal to noise ratio ( 𝑆 / 𝑁 ) we employed the Voronoi tes-sellation technique (Cappellari & Copin 2003) in our data cube, witha 𝑆 / 𝑁 target set to 20. Results from Cid Fernandes et al. (2005)show at this level starlight is able to reliably recover input pa-rameters. Some spectra, however, failed to reach the desired 𝑆 / 𝑁 ,even when combining several tens of spaxels, specifically those ob-scured by the galaxy’s dust lane. Henceforward we will just analyseand show the results for those voronoi zones which the combined-spectra has 𝑆 / 𝑁 ≥
10, relying again on Cid Fernandes et al. (2005)tests that show it to be the minimum value to the reliability of thecode. starlight measures the 𝑆 / 𝑁 as the ratio between the meanand the Root-Mean-Squared (RMS) flux in a user-selected region ofthe spectrum (which should be as featureless as possible), the rangeadopted here is a Δ 𝜆 = 20 Å wide window around the normalisation Further validation of this approximation will be presented throughout thepaper, once the almost constant radial profiles shown in Fig. 3 is taken intoaccount, and the inferred stellar 𝐴 𝑉 estimates, roughly consistent with thosederived from the Balmer decrement (see also e.g. Fig. A.1 in Mingozzi et al.(2019) for consistent results) will be presented in appendixes A and B. wavelength as measured from the combined-spectra of the Voronoizone.In order to be able to explore both the effects of the FC componentand the change of model/library, we run the starlight code in thespaxels of the binned cube using three distinct spectral bases: • BC03 : Only stellar SSP components from the BC03/STELIBmodels, see Table 1 for ages and metallicities. • BC03+FC : Same as BC03 stellar SSPs, with the addition of anFC component ( 𝐹 𝜆 ∝ 𝜆 − . ) to the spectral base. • M11+FC : Stellar SSP components from M11/MILES, see Ta-ble 1 for ages and metallicities, in addition to the FC component.Since the effects of the presence/absence of FC component canbe tested comparing
BC03 with
BC03+FC , and the change ofmodel/library is explored comparing
BC03+FC with
M11+FC , wedid not find necessary to test the synthesis using
M11 without FC,therefore we kept our analyse within these three configurations.The FC component is available for starlight only in the spaxels ofthe inner 1.1 arcsec , which equals 3 𝜎 of the Gaussian point spreadfunction. The reason for not allowing the FC to be used throughoutthe field is that there is a high level of degeneracy between the FC anda young reddened population (further discussion in Cid Fernandes& González Delgado (2010) and the next section), and also becausethere is no physical reason to have an FC emission outside the nucleussince it is emitted by the spatially unresolved accretion disk. We willcall the percentage contribution of the FC to total light of a spaxelas 𝑥 𝑎𝑔𝑛 . During the synthesis all the emission lines were masked(see Fig. 2), the mask is the same in all spaxels with exception to thecentral ones (r < 1.1 (cid:48)(cid:48) ) where H 𝛽 and [N ii] +H 𝛼 masks are broader.The sum of 𝑥 𝑌 , 𝑥 𝐼 , 𝑥 𝑂 , 𝑥 𝑎𝑔𝑛 values is always 100 per cent for allspaxels.The maps of the recovered properties from the three distinct basesare shown in Fig. A1, Fig. A2 and Fig. A3 at the Appendix A. Thesefigures show the maps for 𝑥 𝑌 , 𝑥 𝐼 , 𝑥 𝑂 , 𝑥 𝑎𝑔𝑛 , 𝐴 𝑉 , the light-weightedstellar mean age (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) defined as: (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) = (cid:205) 𝑁 ∗ 𝑗 = 𝑥 𝑗 𝑙𝑜𝑔 ( 𝑡 𝑗 ) (cid:205) 𝑁 ∗ 𝑗 = 𝑥 𝑗 (2)where 𝑥 𝑗 are the elements of the population vector (cid:174) 𝑥 , i.e. the fraction The radius(r) = . , 1–20 (2021) M. Guolo-Pereira et al. contribution of each component to the total light, 𝑡 𝑗 is the ages of 𝑗 -thcomponent, and 𝑁 ∗ is the number of stellar spectra in the base (SSPs).The figures in Appendix A also shows the light-weighted stellar meanmetallicity (cid:104) Z (cid:105) for each reduced population vector component 𝑥 𝑌 , 𝑥 𝐼 and 𝑥 𝑂 , respectively (cid:104) Z (cid:105) 𝑌 , (cid:104) Z (cid:105) 𝐼 and (cid:104) Z (cid:105) 𝑂 , defined as: (cid:104) Z (cid:105) = (cid:205) 𝑡 𝑓 𝑗 = 𝑡 𝑥 𝑗 𝑧 𝑗 (cid:205) 𝑡 𝑓 𝑗 = 𝑡 𝑥 𝑗 (3)where 𝑡 and 𝑡 𝑓 are respectively the lower and upper age limit ofthe reduced population vector components as defined in the previoussection.The upper panel of Fig. 2 shows an example fit for a spaxel at 𝑟 = . (cid:48)(cid:48)
6, the residuals between the observed and the fitted spectra andthe emission line mask. At the bottom panels the age/metallicity/FCdecomposition for the three syntheses are shown as an example ofthe small differences resulting from the change of spectral base (seesubsection 3.3 and subsection 3.4).In Fig. 3 we show the radial profiles of 𝑥 𝑌 , 𝑥 𝐼 , 𝑥 𝑂 , 𝑥 𝑎𝑔𝑛 and (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) , in bins of 0 . (cid:48)(cid:48)
45 ( ≈
70 pc), with points representing themedian, and the coloured region represents the 25 and 75 percentilesof the azimuthal distribution for the given radial bin. In Fig. 4 weshow the values of (cid:104) Z (cid:105) for each one of the bases integrated over theentire FoV.In general terms the presence of a large contribution of old metal-rich SSPs and a smaller, but significant, contribution of young metal-poor stars seems to be independent of the presence of an FC com-ponent, or the change of model/library. The detailed analyses of thedifferences caused by the distinct spectral bases available to the codewill be better explored in the sections below. As shown by Fig. 3, the differences in the
BC03 and
BC03+FC synthesis are seen only at the inner spaxels and are linked to thepresence of the FC component in the spectral base. For the externalspaxels ( 𝑟 > . (cid:48)(cid:48) 𝑥 𝑂 values range between 75per cent and 85 per cent, with all spaxels having super-solar meanmetallicity, (cid:104) Z (cid:105) 𝑂 > . (cid:12) , and the majority of them being closeto the upper limit of the models, (cid:104) Z (cid:105) 𝑂 ≈ . (cid:12) . The 𝑥 𝑌 valuesare between 15 per cent and 25 per cent, with most spaxels havinga sub-solar mean metallicity, (cid:104) Z (cid:105) 𝑌 ≤ . (cid:12) , with a considerableportion of them at the lower limit of the model, i.e. (cid:104) Z (cid:105) 𝑌 = . (cid:12) .Some properties are identical in all runs even at the inner spaxels, 𝐴 𝑉 is very high in the innermost spaxels ( 𝐴 𝑉 > 𝐴 𝑉 ≈ . 𝐴 𝑉 map shown at Appendix A). Similarly, the contribution fromintermediate SSPs, 𝑥 𝐼 , is null at most spaxels and reaches a maximumvalue between 10 and 20 per cent in very few regions.Comparing the inner spaxels of BC03 and
BC03+FC , we see thatthe presence of an FC component in the base causes a decrease in 𝑥 𝑂 by just a few per cent. Larger differences, however, are seen in thecontribution from young stars, especially where an FC componentis not available ( BC03 ). In this case, there is a sharp increase in 𝑥 𝑌 from a steady 20 per cent at 𝑟 > (cid:48)(cid:48) to more than 40 per cent at thecentre. On the other hand, the ( BC03+FC ) base produces a roughlyconstant 𝑥 𝑌 at 𝑟 < . (cid:48)(cid:48) , while 𝑥 𝑎𝑔𝑛 increases from ∼ (cid:48)(cid:48) to ∼
20 per cent in the innermost spaxels. This effect is caused by the above-mentioned degeneracy between an FC spectrum and ayoung-reddened SSP (Cid Fernandes et al. 2004; Cid Fernandes &González Delgado 2010). As shown by Cardoso et al. (2017), fitting aspectrum that contains an AGN power law-like continuum with onlystellar components causes the solution to have an excess of youngpopulations, which leads to a drop in the (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) (see the bottompanel of Fig. 3). Although such degeneracy can cast some doubts inthe exact values of the 𝑥 𝑌 and 𝑥 𝑎𝑔𝑛 decomposition in BC03+FC , thefact that the solution at the centre has 𝑥 𝑌 values compatible with theones of the external regions of the FoV, and that the FC componentcontribution follows a radial profile compatible with an unresolvedsource (i.e increasing from 𝑥 𝑎𝑔𝑛 ≈ 𝑥 𝑎𝑔𝑛 ≈
20 per centat the continuum peak) is strong evidence in favour of the existenceof an FC emission in NGC 2992.
In this section we discuss differences in the recovered propertiesbetween
BC03+FC and
M11+FC bases and their derivations. Wedo not intend to make a detailed analysis of the models and itsderivations, nor to suggest one is a better choice than the other foruse with this type of data. Such discussions can be found in otherstudies, e.g. Baldwin et al. (2018); Ge et al. (2019). Instead, wemerely want to test the stability of the inferred stellar properties withchanges in models.Besides the distinct stellar libraries used in these model libraries,they also employ different evolutionary prescriptions (see Conroy2013, for a review on stellar evolutionary models). BC03 is con-structed using the isochrone synthesis approach, while M11 usesthe fuel consumption theory. The treatment of the thermally puls-ing asymptotic giant branch (TP-AGB) phase is also a major topicof disagreement between different models. For example, Damettoet al. (2019) found a relative excess of young populations and lack ofintermediate-age components when Near-Infrared (NIR) stellar pop-ulations synthesis are performed with BC03 SSPs, favouring resultsachieved with the M11 models. Nevertheless, these effects are muchmore important when modelling stellar populations in the NIR, beingsignificantly less relevant in the optical.Both 𝑥 𝑎𝑔𝑛 and 𝐴 𝑉 are very similar to those obtained with BC03+FC , but the population vectors have notable differences. Forinstance, 𝑥 𝑂 values for M11+FC are systematically lower by ∼ BC03+FC , while at the sametime both 𝑥 𝑌 and 𝑥 𝐼 are conversely higher for M11+FC . Given theincrease of both 𝑥 𝑌 and 𝑥 𝐼 and the decrease in 𝑥 𝑂 , the light-weightedmean stellar age (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) for M11+FC is systematically lower thanfor
BC03+FC by ∼ . (cid:12) < (cid:104) Z (cid:105) 𝑂 ≤ . (cid:12) )and the young populations have a solar to sub-solar mean metallicity( (cid:104) Z (cid:105) 𝑌 ≤ . (cid:12) ), as shown in the middle panels of Fig. 4. However,while the values of (cid:104) Z (cid:105) 𝑌 in BC03+FC are mostly sub-solar ( (cid:104) Z (cid:105) 𝑌 < . (cid:12) ) the values in M11+FC are all solar ( (cid:104) Z (cid:105) 𝑌 = . (cid:12) ). Thisdiscrepancy is probably caused by the fact that the MILES librarydoes not contain SSPs with a sub-solar metallicity younger than 55Myr (see Table 1). For instance, in the example of Fig. 2, while BC03 and
BC03+FC recover its young populations as a combination ofthree SSPs with ages of 1 Myr, 5 Myr and 10 Myr, the two firstwith sub-solar metallicity ( 𝑍 = . (cid:12) ) and the later with solar MNRAS000
BC03+FC recover its young populations as a combination ofthree SSPs with ages of 1 Myr, 5 Myr and 10 Myr, the two firstwith sub-solar metallicity ( 𝑍 = . (cid:12) ) and the later with solar MNRAS000 , 1–20 (2021) apping the Inner kpc of NGC 2992 F l u x () Observed (r = 0.6")Fit (BC03+FC)4500 5000 5500 6000 6500 [ Å ] R e s i d u a l Mask10 x j ( % ) BC03 : Av=2.66, adev=3.99, < log t * > = 9.12 Z Z Z F10 x j ( % ) BC03 + FC : Av=2.60, adev=3.93, < log t * > = 9.38 Z Z Z F10 Age [yr] x j ( % ) M11 + FC : : Av=2.77, adev=3.95, < log t * > = 9.27 Z Z FC Figure 2.
Example of stellar population synthesis fit. Top: Observed (black) and synthetic (red) spectra both in units of 10 − erg s − cm − Å − in a spaxel atradius equal to 0 . (cid:48)(cid:48)
6. Middle: Residual between the observed and synthetic spectra, masked emission lines are shown in purple. Bottom: Age/metallicity and FCdecomposition for the three synthesis,
BC03 , BC03+FC and
M11+FC . The dashed lines show the division between 𝑥 𝑌 , 𝑥 𝐼 and 𝑥 𝑂 SSPs. metallicity ( 𝑍 = . (cid:12) ), the M11+FC recovers it as both 9 Myrand 10 yr solar metallicity SSPs, and the fitting code would not beable to fit such a young component with lower metallicity preciselybecause MILES library does not provide an SSPs with 𝑍 = . (cid:12) younger than 55 Myr. Therefore, we can see these (cid:104) Z (cid:105) 𝑌 = . (cid:12) values in the synthesis using M11+FC base as an upper limit for theyoung population metallicity.As mentioned, above, we do not claim any of the three synthesesto be the one that better describes the data. In fact, there are nosignificant differences in the adev parameters among the bases: adev values range from 3 per cent at the inner spaxels up to 8 per cent atlarger radius. Recovered properties produced with the distinct modelspresent the same general trend (see subsection 3.2). However, weargue the bases which include an FC component are more physicallyjustified. Both the large decrease in (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) in the inner spaxels, whenthis component is ignored, and the fact that the FC components fluxeshave a spatial profile characteristic of a point-like source (FWHM ≈ . (cid:48)(cid:48)
8, see Fig. 7) are strong evidence of the existence of an FC likeemission in NGC 2992. It is worth mentioning even though we arenot discussing results for the stellar kinematics, we have made sureall the three decompositions were kinematically consistent with eachother, and no non-physical value was fitted to 𝑣 ∗ , 𝜎 ∗ parameters,which would cause an incorrect subtraction of the continuum usedin emission line fitting procedures of the next sections. Emission line structure of the central regions of AGN is very com-plex, often possessing multiple velocity components along the lineof sight (e.g. Villar-Martín et al. 2011; Veilleux et al. 2013; McElroyet al. 2015). For the type-1 and intermediate-type AGNs, the presence of both narrow and broad emission lines are seen; these emission linesare believed to come from two distinct parts of the AGN. The first isnamed the "Broad Line Region" (BLR). Its typical size, as deducedfrom the broad line variability, is 10-100 light-days, therefore it isspatially unresolved by current observations, meaning that its spatialprofile is the same as the point spread function (PSF) of the obser-vation. BLR emission is characterised by very broad (1000 km s − M. Guolo-Pereira et al. x O [ % ] (BC03) (BC03+FC) (M11+FC) x I [ % ] x Y [ % ] x a g n [ % ] Radius [arcsec] l o g t * [ y r ] Figure 3. Radial profiles of the recovered properties for BC03 (gold), BC03+FC (red) and M11+FC (blue) in bins of 0 . (cid:48)(cid:48) 45 ( ∼ 70 pc). From top tobottom: Per cent contribution to the total light from young ( 𝑥 𝑂 ), intermediate( 𝑥 𝐼 ) and old ( 𝑥 𝑌 ) SSPs, from the FC ( 𝑥 𝑎𝑔𝑛 ) and the light-weighed meanstellar age ( (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) ). 𝑥 𝑎𝑔𝑛 is null for BC03 by definition. The 𝑥 𝐼 values for BC03 are not shown, however they are exactly the same as BC03+FC . Figure 4. Values of (cid:104) Z (cid:105) 𝑌 , (cid:104) Z (cid:105) 𝐼 and (cid:104) Z (cid:105) 𝑂 integrated over the entireFoV for each base, from left to right, respectively, BC03 , BC03+FC and M11+FC . Points are the median values, and the coloureds bars are the 25 and75 percentiles. The dashed black lines show the metallicity limit for each ofthe models. In order to fit the emission lines in NGC 2992 we use the IFSCubepackage (Ruschel-Dutra 2020). IFSCube is a python based soft-ware package designed to perform analysis tasks in data cubes ofintegral field spectroscopy, mainly focused on the fitting of spectralfeatures, such as atomic lines. It allows for multiple combinations ofGaussian and Gauss-Hermite functions, with or without constraintsto its parameters. Stellar continuum subtraction, as well as pseudocontinuum fitting if necessary, are performed internally, and in thiscase, we used the residual spectra from the stellar population synthe-sis performed with starlight. Initial guesses for the parameters areset for the first spaxel, usually, the one with the highest 𝑆 / 𝑁 , and allsubsequent spectra use the output of neighbouring successful fits asthe initial guess.When dealing with complex emission, the standard approach isto model the line profile with a combination of Gaussian functions.Visual inspection of our IFU shows distinct velocity components,evidenced by large asymmetries in the line profiles, are present inover one-third of the spectra in the FoV. As a result, we decided tomodel these components separately and to investigate whether theypossessed any meaningful physical information, following previousworks (e.g. Villar-Martín et al. 2011; McElroy et al. 2015; Fischeret al. 2018).For every fitted emission line, we determined if the multi-component model was statistically justified, and not a better fit purelyby virtue of the extra model parameters, by performing a series off-tests. The f-test is a standard statistical test to gauge whether ahigher order model is preferable to a simpler model when fitting aparticular data set. We set the false rejection probability for the lowerorder model to 10 − . The higher this threshold is set, the harder itis to justify a more complex model and the more likely it is to beover-fitting noise. The reader interested in more detailed analyses anda step-by-step deduction of the application of f-test in astrophysicsemission line fitting is referred to Freund (1992) and statistical ref-erences therein.As an Intermediate-type Seyfert nucleus, NGC 2992 presentsboth BLR and NLR emission. The BLR emission lines (H 𝛼 andH 𝛽 ) are blended with some NLR emission lines (H 𝛼 , [ N ii ] 𝜆 [ N ii ] 𝜆 𝛽 ), leading to degeneracies if both NLR and BLRemission lines are fitted, spaxel by spaxel, at the same time withall parameters free. Fortunately, the fact that the BLR is not re-solved, i.e. its kinematics and its Balmer decrement (H 𝛼 /H 𝛽 ) can not http://github.com/danielrd6/ifscubeMNRAS , 1–20 (2021) apping the Inner kpc of NGC 2992 vary spatially, can be used to decrease this degeneracy. Our chosenmethodology consists on first fitting an integrated spectrum to fix theBLR emission line profile (kinematics and Balmer decrement), andlater using this BLR spectrum as a fixed component, except for theflux, on the spatially resolved NLR fits.We extracted a spectrum centred at the continuum peak, using avirtual aperture of 3 𝜎 = . (cid:48)(cid:48) 𝜎 refers to the Gaussian point spreadfunction). In this spectrum we fitted [ N ii ] 𝜆 [ N ii ] 𝜆 𝛼 (both BLR and NLR) and H 𝛽 (both BLR and NLR), adding Gaus-sian components to the emission lines until the addition of a newcomponent does not increase the fitting quality, as measured by thef-test described above. The final fit has shown that two Gaussiansare necessary to fit each one of the BLR Balmer emission lines. Thisresulting profile, meaning the superposition of two Gaussian func-tions, hereafter called as BLR Component, can be seen in panel (1)of Fig. 5, and is characterised by a FWHM BLR = − anda Balmer decrement of (H 𝛼 /H 𝛽 ) BLR = . 2. These values are inconsiderable agreement with the ones fitted by Schnorr-Müller et al.(2016), FWHM BLR = − and (H 𝛼 /H 𝛽 ) BLR = . 2, con-sidering these authors used three Gaussian functions to model theprofile and the BLR of NGC 2992 has been shown to vary in a fewyears period (L. Trippe et al. 2008).In order to resolve the kinematics of the NLR and obtain mapsof emission line ratios, we fitted the following eight strongest emis-sion lines present in the data cube: H 𝛽 , [ O iii ] 𝜆 , 𝜆 𝛼 , [ N ii ] 𝜆 , 𝜆 [ S ii ] 𝜆 , 𝜆 ± 400 km s − andthe velocity dispersion up to 400 km s − (FWHM ≈ − ).We added one more set of Gaussians to the fit, meaning that eachemission line is modelled by up to two Gaussian components. Toreduce the number of free parameters, each set of Gaussians wasrequired to have the same velocity and velocity dispersion (meaningtheir relative wavelengths were fixed), while their fluxes were allowedto vary. The reason for this is each component represents a kinemat-ically distinct part of the gas, meaning that they may have differentionisation states determined by their relative fluxes. This means that,for example, the single-component set of Gaussians fit to the eightemission lines instead of having twenty four (eight lines × three pa-rameters) free parameters, only had nine (a single velocity, 𝑣 , andvelocity dispersion, 𝜎 , amplitudes of H 𝛼 , H 𝛽 , the two [S ii] and thestronger [N ii] and [O iii], while the amplitudes of the weaker [N ii]and [O iii] were set fixed as 1/3 of the stronger ones, as given by theirQuantum Mechanics’ probabilities (hereafter we call [ O iii ] 𝜆 [ N ii ] 𝜆 [ S ii ] 𝜆 , 𝜆 , therefore the size is maximum for [O iii] (Fig. 6) and H 𝛼 (Fig. 11), a little bit smaller for [N ii] (Fig. 5) and the [S ii] doublets(not shown) and much smaller for the low S/N line, H 𝛽 (not shown).However, the general pattern is the same for all emission lines: thedata are well fitted by a single-Gaussian in most of the FoV, but ina region, from the nucleus to the end of the FoV in the northwest(NW) direction two Gaussians are needed.As can be seen in both figures, for those spaxels in which theemission lines are well behaved and symmetric, a single-Gaussianfunction, which we will call Component 1, is enough to fit the data.However, in the spaxels where two components were statistically sig-nificant, the profiles show an asymmetry in the form of a blueshiftedwing (see panels (1) and (2) in Fig. 5 and Fig. 6) requiring the additionof a second Gaussian, which we will call Component 2.In Fig. 7 we show the spatial and radial profile of the BLR Com-ponent. As expected for an unresolved structure, this component isdetected just in the innermost spaxels, with a spatial distribution ofa point-like source, equal to that of the stars in the acquisition image(FWHM ≈ . (cid:48)(cid:48) 8, see section 2); at radii larger than 3 𝜎 (1.1 arcsec) itis almost undetectable. Let us now discuss the spatially resolved properties of the NLR Com-ponents 1 and 2 and their possible physical interpretation. Component1 is present in the entire FoV. In the top panels of Fig. 8 we show, fromleft to right, [O iii] flux, velocity dispersion ( 𝜎 ), already correctedby instrumental dispersion. In the middle left panel, we present theradial velocity ( 𝑣 ). We show the [O iii] flux as an example of themorphology of this component, which is similar to the continuummorphology, with the peak of the emission shifted by only ∼ . (cid:48)(cid:48) 𝑑𝑒𝑥 ) towards thedust lane region. The 𝜎 values range from 50-100 km s − and showa very disturbed pattern. The 𝑣 map shows values from -130 to 120km s − and although it presents some disturbance, a rotation patternis clearly discernible, with the SW portion receding and the north-eastern (NE) approaching. We have considered the west side of thegalaxy is the near one, following the literature (e.g. Marquez et al.1998; Veilleux et al. 2001); this explains why the emission is seenmainly to the east and less to the west where it is attenuated by thedust lane in the line-of-sight.In order to quantify how much the velocity field deviates from apure rotation and to derive both the major axis PA and the inclinationof disk we fit the Bertola et al. (1991) model. The model assumes aspherical potential with pure circular orbits, in which the observedradial velocity at a position ( 𝑅 , 𝜓 ) in the plane of the sky given bythe relation: 𝑣 ( 𝑅, 𝜓 ) = 𝑣 sys + 𝐴𝑅 cos ( 𝜓 − 𝜓 ) sin 𝜃 cos 𝑝 𝑖 (cid:110) 𝑅 (cid:2) sin ( 𝜓 − 𝜓 ) + cos ( 𝜓 − 𝜓 ) (cid:3) + 𝑐 𝑐𝑜𝑠 𝑖 (cid:111) 𝑝 / (4)where 𝑣 𝑠𝑦𝑠 is the systemic velocity, 𝑖 the inclination of the disk (with 𝑖 = 0 for a face-on disk), 𝐴 is the amplitude of the curve, c is theconcentration parameter and 𝜓 the line of nodes PA. The parameter Because the f-test detection is sensible to how much the second component,the more complex model, stands out from the noise. Therefore in low S/Nlines, it is less likely the f-test will require a complex model.MNRAS , 1–20 (2021) M. Guolo-Pereira et al. -3 -2 -1 0 1 2 3 X (arcsec) Y ( a r c s e c ) E N 12 3 -3 -2 -1 0 1 2 3 X (arcsec) -3-2-10123 Y ( a r c s e c ) 12 3 F l u x () (1) Component 1Component 2BLR ComponentTotal FitObserved6500 6550 6600 6650 6700 6750wavelength [ ]0.50.00.50.00.40.81.21.6 F l u x () (2) F l u x () (3) [10 erg cm s Å ] Figure 5. Top left: Map of the continuum emission at 𝜆 = 5500 Å and map of the preferred number of Gaussian components to the [N ii] emission-lines. Top rightand bottom: Examples of the H 𝛼 , [ N ii ] 𝜆 , 𝜆 [ S ii ] 𝜆 , 𝜆 − erg s − cm − Å − and the dashed vertical lines show rest-framewavelength of the lines. -3 -2 -1 0 1 2 3 X (arcsec) Y ( a r c s e c ) E N -3 -2 -1 0 1 2 3 X (arcsec) -3-2-10123 Y ( a r c s e c ) F l u x () (1) Component 1Component 2Total FitObserved4940 4960 4980 5000 5020wavelength [ ]0.50.00.50.00.51.0 F l u x () (2) F l u x () (3) [10 erg cm s Å ] Figure 6. Same as Fig. 5 but for [O iii] emission lines.MNRAS000 Same as Fig. 5 but for [O iii] emission lines.MNRAS000 , 1–20 (2021) apping the Inner kpc of NGC 2992 -3 -2 -1 0 1 2 3 X (arcsec) -3-2-10123 Y ( a r c s e c ) H (BLR Comp.) Radius [arcsec] F l u x H ( B L R C o m p . ) [ e r g c m s ] H (BLR Comp.)FWHM/2 0.4" 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Radius [arcsec] F l u x F C ( = Å ) [ e r g c m s ] BC03+FCM11+FC0.01.02.03.04.0 [ e r g c m s ] Figure 7. Top: Spatial distribution of the BLR Component. Middle: Ra-dial projection of the BLR Component profile, vertical axis units of10 − erg s − cm − . Both show the point-like source emission of the BLR,with a FWHM ≈ . (cid:48)(cid:48) 𝑝 defines the form of the rotation curve, varying in the range between1 (logarithm potential) and 1.5 (Keplerian potential). To avoid thedegeneracy between the parameters, we restricted the disk inclinationto have values of 𝑖 = 70 ± 10 (based on Duc et al. 2000, photometricanalyses) and we assumed the kinematical centre to be co-spatialwith the peak of the continuum emission. A Levenberg-Marquardtleast-squares minimisation was performed to determine the best fit-ting parameters and the uncertainties were obtained by making onehundred Monte Carlo iterations. We show the best parameters andtheir uncertainties in Table 2.The PA of the line of nodes was measured as 32 ± ◦ , in agreementwith previous values from the literature: PA = 32 ◦ (Veilleux et al.2001, hereafter V01), PA= 30 ◦ (Marquez et al. 1998, hereafter M98).V 𝑠𝑦𝑠 and i are also in reasonable agreement with other studies ofthe galaxy: V01 found, respectively, 2335 ± 20 km s − and 68 ◦ , whileM98 found 2330 km s − and 70 ◦ . The velocity field amplitude (A =132 ± − ), though, is much lower than the ones found by bothV01 and M98 (225 ± 20 km s − and 250 km s − ). This discrepancy isprobably due to the fact that both have used long-slit data in theirstudies, therefore being able to map the velocity field up to severalkpc. Thus having a better constraint in the behaviour of the field at a Table 2. Best Fitting Parameters Bertola et al. (1991) Model.Parameters Values 𝑣 𝑠𝑦𝑠 ± − ] 𝐴 ± − ] 𝑝 𝑎 𝑐 ± 𝜓 ◦ ± ◦ 𝑖 ◦ ± ◦ 𝑎 In the model the parameters 𝑝 is allowed to vary between 1 and 1.5, thefitted value was equals to 1 in all the Monte Carlo iterations. larger radius, while our data can only recover the internal portion ofthe velocity field.The modelled velocity field is shown in the middle right panel ofFig. 8, and the residuals between the observed field and model areshown in the bottom left panel of the same figure. The uncertaintymap for the Comp. 1 𝑣 field, as measured by the Monte Carlo simu-lations described in the subsection 4.1, is shown in the bottom rightof Fig. 8. While most of the field is well represented by the disk ro-tation model ( | 𝑣 − 𝑣 𝑀𝑜𝑑𝑒𝑙 | < 20 km s − ), there is one region at theeastern border with residuals up to 40 km s − . Since the maximumuncertainty in 𝑣 is 15 km s − , this deviation from pure rotation isprobably real. The same seems to be true for the 𝜎 field, shown inFig. 9. The uncertainties in the Component 1 𝜎 are of ∼ − inthe central region, and up to 15 km s − at the borders, as shown inFig. B1. Therefore, the lack of structure in the 𝜎 map is, in fact, realand not due to observational artefacts .These types of disturbances, in the both 𝑣 and 𝜎 , are predicted bysimulations (Kronberger et al. 2007; Bois et al. 2011) and observed(Torres-Flores et al. 2014; Hung et al. 2016; Bloom et al. 2018) ingalaxies which are, or recently were, undergoing merging processes,as is the case of NGC 2992, and are attributed to perturbationsin the gravitational potential due to tidal forces. In Paper II we willinvestigate whether these disturbances are also present in the velocityfield of NGC 2993.The fainter component of the NLR (Component 2) is detected ina smaller region, originating near the continuum peak and extendingtowards the NW. For the brightest line, [ O iii ] 𝜆 ∼ (cid:48)(cid:48) , as shown by both the flux map, top left panel ofFig. 9 and the P-V plot in Fig. 10. In Fig. 9, we show the maps ofequivalent width (EW) of [O iii], 𝜎 and 𝑣 for the Component 2. The 𝑣 map shows only blueshifted velocities, with values ranging from-250 to -200 km s − . The velocity dispersion is higher than that ofComp. 1, ranging from 200 to 215 km s − . We also show in all thepanels, the contours of the 8-shaped region detected in the 6 cm radioemission (in grey) in which the smaller loop seems to be co-spatialwith Component 2.The position-velocity plot along the major kinematical axis (PA= 32 ◦ ) and the major axis of the Radio emission (PA = -26 ◦ ) ofNGC 2992 are shown in Fig. 10. It is clear that while Comp. 1 has a There is no clear evidence of the "Beam Smearing" effect too, given thatthere is no increase in velocity dispersion values of Comp.1 in the centralregions, where the gradient of the velocity field is maximum. The high valuesof velocity dispersion in Comp. 2, also cannot be purely due to this effect,because it is not co-spatial with the galactic centre and shows very blueshifitedvelocities from -250 to -200 km s − , as can be seen Fig. 9. Furthermore when 𝜎 of both Comp. 2 and Comp. 1 are taken into account their values arein agreement with the ones measured from Br 𝛾 IR lines by Friedrich et al.(2010). MNRAS , 1–20 (2021) M. Guolo-Pereira et al. -3-2-10123 Y ( a r c s e c ) log Flux [OIII] (Comp. 1) E N (Comp. 1) -3-2-10123 Y ( a r c s e c ) NEARFAR150pc V (Comp. 1) NEARFAR E N V Model -3 -2 -1 0 1 2 3 X (arcsec)-3-2-10123 Y ( a r c s e c ) NEARFAR150pc V V Model -3 -2 -1 0 1 2 3 X (arcsec) NEARFAR E N Error V (Comp. 1) Figure 8. Component 1 properties. Top left: Log Flux [O iii] 𝜆 − cm − ). Top right: Velocity dispersion 𝜎 ( km s − ). Middle left:Radial velocity 𝑣 ( km s − ). Middle right: Modelled radial velocity ( km s − ),from the Bertola et al. (1991) rotation model fitting. Bottom left: Residualbetween the measured and the modelled velocity field. Bottom right: Uncer-tainties in the radial velocity 𝑣 ( km s − ). Black cross is the continuum peakand the dashed black line is the fitted major axis PA. rotational pattern, Comp. 2 deviates considerably from this pattern.Considering: • Component 2 is not following the galaxy’s rotation; • It is co-spatial with the radio emission; • It has higher velocity dispersion values than those of Comp. 1, 𝜎 > 200 km s − ; • It has blueshifted velocities in the near side of the galaxy;we are led to the conclusion that it is tracing a radio-jet driven gasoutflow since all these are well-known characteristics of this type ofphenomena. In fact, this ionised gas outflow was already reportedby other IFU studies, e.g. García-Lorenzo et al. (2001), Friedrichet al. (2010) and Mingozzi et al. (2019). Particularly, García-Lorenzoet al. (2001) interpreted this blueshifted structure as the interactionof the minor loop of the radio-jet with the interstellar medium (ISM).García-Lorenzo et al. (2001) also found a similar, however redshifted,structure in the far side of the galaxy, co-spatial with the farthestpart of the radio emission major loop, which is mostly not visiblein our FoV. We have extracted the contours of this structure fromFigure 9 of their paper and show it as the red dashed line in Fig. 9.This finding supports the scenario in which a bipolar over-pressurerelativistic plasma is expanding, accelerating and compressing theISM gas in its path (Chapman et al. 2000). Interestingly, the only -5-4-3-2-10123 Y ( a r c s e c ) NEARFAR log Flux [OIII] (Comp. 2) NEARFAR E N EW [OIII] (Comp. 2) -4 -3 -2 -1 0 1 2 3 X (arcsec) -5-4-3-2-10123 Y ( a r c s e c ) NEARFAR (Comp. 2) -4 -3 -2 -1 0 1 2 3 X (arcsec) NEARFAR E N V (Comp. 2) Figure 9. Component 2 properties. Top left: Log Flux [O iii] 𝜆 − cm − ). Top right: Equivalent Width of [O iii] 𝜆 𝜎 ( km s − ). Bottom right: Radial velocity, 𝑣 (km s − ). Solid black lines represent our FoV. Black cross is the continuumpeak. Grey contours are the 6cm radio emission (Ulvestad & Wilson 1984).Red dashed line is the contour of the redshifted structure found by García-Lorenzo et al. (2001), at the farthest part of the radio emission major loop,taken from Fig. 9 of their paper. V [ k m / s ] Major Kinematical Axis (PA = 32 o ) Component 1Component 24 3 2 1 0 1 2 3 4 Distance [arcsec] V [ k m / s ] Radio Axis (PA = 26 o ) Component 1Component 2 Figure 10. Position-Velocity plot in virtual slits with 0 . (cid:48)(cid:48)000 Position-Velocity plot in virtual slits with 0 . (cid:48)(cid:48)000 , 1–20 (2021) apping the Inner kpc of NGC 2992 -3 -2 -1 0 1 2 3 X (arcsec)-3-2-10123 Y ( a r c s e c ) log L(H ) (Comp. 2) -3 -2 -1 0 1 2 3 X (arcsec) E N log 10 g Figure 11. Left: Logarithm of Component 2 dereddened H 𝛼 luminosity inerg s − . Right: Logarithm of Component 2 surface mass density in units ofM (cid:12) (0 . (cid:48)(cid:48) − . Black cross is the peak of the continuum emission. portion of this structure which falls within our FoV is co-spatialwith the largest velocity residuals seen in figure Fig. 8. Althoughour statistical analysis does not indicate the necessity of a separatecomponent, the large residuals clearly indicate a perturbation in thevelocity field.A very similar geometry was found by Riffel & Storchi-Bergmann(2011) in the galaxy Mrk 1157, in which non-rotational blueshiftedand redshifted structures are present and co-spatial to the two loopsof a very similar nuclear 8-shaped radio emission. The authors alsointerpreted these structures as the effect of the expanding radio-emitting gas into the ionised ISM. If we consider Component 2 as ionised outflowing gas, we can mea-sure its properties such as mass, mass outflow rate and kinetic power.We followed the prescriptions described in Lena et al. (2015), inwhich a biconical outflow geometry is assumed, even though in ourcase we can see only one of the cones.The mass outflow rate can be estimated as the ratio between themass of the outflowing gas and the dynamical time (time for the gasto reach the present distance from the nucleus), M 𝑔 /t d . The mass ofthe gas in outflow (Component 2) is estimated as: 𝑀 𝑔 = 𝑛 𝑒 𝑚 𝑝 𝑉 𝑓 (5)where 𝑚 𝑝 is the proton mass, 𝑛 𝑒 and 𝑉 are the electron densityand the volume of the region where the component is detected, 𝑓 is the filling factor. The filling factor and volume are eliminatedby combining Equation 5 with the following expression for the H 𝛼 luminosity: 𝐿 ( 𝐻𝛼 ) ≈ 𝑓 𝑛 𝑒 𝑗 𝐻 𝛼 𝑉 (6)with 𝑗 𝐻 𝛼 = . × − erg cm s − at 𝑇 = 𝑀 𝑔 = 𝑚 𝑝 𝐿 ( 𝐻𝛼 ) 𝑛 𝑒 𝑗 𝐻 𝛼 (7)Our estimates for 𝑛 𝑒 are based on the flux ratio between [S ii] lines, asdescribed in Appendix C. Please note the 𝐿 ( 𝐻𝛼 ) used here is basedon the flux of Comp. 2 H 𝛼 emission only, corrected by extinction asdescribed in Appendix C. The spatially resolved 𝐿 ( 𝐻𝛼 ) and M 𝑔 (i.e.gas mass surface density, Σ g ) are shown in Fig. 11. The integrated gas mass is 𝑀 𝑔 = . ± . × M (cid:12) . The t 𝑑 wasestimated as the ratio between the maximum extension of the outflow ∼ ≈ 450 pc (see Fig. 10) to its mean velocity, (cid:104) 𝑣 𝑜𝑢𝑡 (cid:105) ≈ − , which gives t d ≈ . × yr. Combining these values weestimated an outflow mass rate of (cid:164) 𝑀 𝑜𝑢𝑡 ≈ ± (cid:12) .We also estimated the outflow kinetic power as: (cid:164) 𝐸 𝑜𝑢𝑡 = (cid:164) 𝑀 𝑜𝑢𝑡 𝑣 𝑜𝑢𝑡 𝑣 𝑜𝑢𝑡 is the radial velocity of Component 2 (Fig. 9). Integratingthe entire outflow we obtained a total outflow kinetic power (cid:164) 𝐸 𝑜𝑢𝑡 ≈ . ± . × erg s − . In subsection 6.3, we compare these valueswith the ones found in the literature and those expected by outflow-AGN relations. Given the presence of young stars discussed in section 3, X-ray(section 1) and BLR emission (section 4) which indicate AGN ex-citation, and the presence of radio emission that can drive shocks(Dopita 2002; Allen et al. 2008), we used the [N ii]/H 𝛼 vs. [O iii]/H 𝛽 and [S ii]/H 𝛼 vs. [O iii]/H 𝛽 Baldwin et al. (1981, BPT) diagnosticdiagrams to investigate which of these excitation mechanisms aredominant in the innermost region of the galaxy.In the bottom panels of Fig. 12, we show the maps of [O iii]/H 𝛽 ,[N ii]/H 𝛼 and [S ii]/H 𝛼 line ratios for the entire line profile (by sum-ming the kinematical components; we do that to increase the preci-sion in the flux measurement and to avoid regions where a secondcomponent is detected in high S/N lines, like [O iii], but not detectedin lower S/N ones, like H 𝛽 .) We excluded the regions where uncer-tainties in the total flux, as shown in Appendix B, are higher than 30per cent. The [O iii]/H 𝛽 shows a clear AGN signature, with its peak,up to [O iii]/H 𝛽 ≈ 25, at the centre of the FoV and decreasing radiallydown to [O iii]/H 𝛽 ≈ 𝛼 and[S ii]/H 𝛼 , however, do not show the same pattern: their maximumvalues are off-nuclear, and they do not present a radial decrease as inthe [O iii]/H 𝛽 one. This indicates a lower energy mechanism (lowerthan an AGN, but higher than H ii regions) can be ionising the gastoo. In fact, shock ionised regions are characterised by having smaller[O iii]/H 𝛽 ratios and larger [S ii] ([N ii])/H 𝛼 than those ionised byAGN (Veilleux & Osterbrock 1987; Kewley et al. 2006; Allen et al.2008).We selected limiting regions in the line ratio maps and shown it inthe respectively BPT diagrams in the top panels of Fig. 12. We alsoshow some division lines: the solid black curves on both diagnosticdiagrams trace the Kewley et al. (2001) theoretical upper limit ofregions ionised by pure star formation (K01 line). All spectra lyingabove the K01 line are dominated by ionisation mechanisms moreenergetic than star formation. Several excitation mechanisms can in-crease the collisional excitation rate, enhancing the ratios of the for-bidden to recombination lines, among them energetic photons from apower law-like emission (AGN); shock excitation and evolved stellarpopulations (pos-AGB) can also produce line ratios along this branchof the diagram. The latter case does not apply to the nuclear regionof NGC 2992 given that it requires EW(H 𝛼 ) < 3 Å (Stasinska et al.2008; Cid Fernandes et al. 2011) and our entire FoV has EW(H 𝛼 )> 10 Å. We also show the dashed black curve on the [N ii]/H 𝛼 vs.[O iii]/H 𝛽 diagnostic diagram: the Kauffmann et al. (2003) empiricalclassification line (Ka03 line), which traces the upper boundary ofthe Sloan Digital Sky Survey Strauss et al. (SDSS 2002) star forma-tion sequence. All spectra lying below the Ka03 line are dominated MNRAS , 1–20 (2021) M. Guolo-Pereira et al. -1.5 -1 -0.5 0 0.5log ([NII]/H )1.00.50.00.51.01.5 l o g ([ O III ] / ) ABCDE 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4log ([SII]/H )1.00.50.00.51.01.5 l o g ([ O III ] / H ) ([NII]/H )0.000.250.500.751.001.251.50 l o g ([ O III ] / ) B 300 600 900 ([SII]/H )0.000.250.500.751.001.251.50 l o g ([ O III ] / H ) B -3 -2 -1 0 1 2 3 X (arcsec)-3-2-10123 Y ( a r c s e c ) A B [O III]/H -3 -2 -1 0 1 2 3 X (arcsec)A BC [N II]/H -3 -2 -1 0 1 2 3 X (arcsec)A BCD E E N [S II]/H Figure 12. Top: [N ii]/H 𝛼 vs. [O iii]/H 𝛽 and [S ii]/H 𝛼 vs. [O iii]/H 𝛽 BPT diagrams; grey countours are galaxies from the SDSS main galaxy sample. red pointsare the spaxel of NGC 2992. Colored markers are the selected regions of NGC 2992, as shown in the bottom panels, representing: red (Region A), blue (B),green (C), pink (D) and cyan (E). Middle: Zoomed BPT diagrams with shock+precursor grid models as measured with MAPPINGS V code (Sutherland et al.2018); the (cid:174) 𝐵 indicates increasing magnetic field of the model (from 0.01 to 4 𝜇 G); the black numbers indicates the velocities of the shocks, ranging from 300 to900 km s − . Bottom: Respectively, from left to right, [O iii]/H 𝛽 , [N ii]/H 𝛼 , and [S ii]/H 𝛼 emission lines ratio maps; red circles show the selected regions shownin the upper panels. In BPT diagrams the grey crosses are the typical errors in the emission line ratios; the solid black lines are Kewley et al. (2001) theoreticalupper limits for pure star forming galaxies. In the [N ii]/H 𝛼 vs. [O iii]/H 𝛽 diagram the dashed black line represents Kauffmann et al. (2003) line which tracesthe upper boundary of the SDSS star formation sequence. In the [S ii]/H 𝛼 vs. [O iii]/H 𝛽 diagram the dashed-dotted black line represents the empirical divisionbetween AGN like and low-ionisation nuclear emission-line region (LINER) (Kewley et al. 2006). by star formation. The dashed-dotted black curve on the [S ii]/H 𝛼 vs. [O iii]/H 𝛽 diagnostic diagram traces the Kewley et al. (2006) em-pirical classification line (K06 line), which separates high ionisationspectra associated with Seyfert (AGN) and other intermediary ioni-sation mechanisms including shock excitation. We also show as greypoints in the top BPT diagrams a collection of single-fibre emissionline measurement from SDSS main galaxy sample by Kewley et al.(2006).Region A is the peak of the [O iii]/H 𝛽 map, it shows clear AGN-like ratios, with an upper [O iii]/H 𝛽 value higher than any SDSSAGN, which corresponds to a 3 (cid:48)(cid:48) diameter aperture, and thereforehas a diluted AGN/host-galaxy emission line ratio. Region B has thelowest value in all the maps; it lies below K01, showing the radialdecrease in the AGN contribution to the emission lines and significantcontributions from both star formation and AGN excitation. Regions C, D, and E display the highest values in the [S ii]/H 𝛼 and [N ii]/H 𝛼 maps; however, they have lower [O iii]/H 𝛽 values than region A.Region E, in fact, crosses the K06 line in the [S ii]/H 𝛼 diagram,which again indicates the presence of another ionisation mechanism.In order to search for an explanation to the increase in this in-termediary ionisation energy lines ([N ii] and [S ii]) we used theMAPPINGS V shock models by Sutherland et al. (2018). We ex-plored the parameter space in the pure shock and shock+precursorgrids of the models. None of the pure shock models could ex-plain these line ratios, given the high [O iii]/H 𝛽 values. Howevershock+precursor (Shock ionisation in a medium already ionisedby power-law like emission) with shock velocities between 300km s − and 900 km s − and magnetic field ( (cid:174) B) between 0.01 𝜇 G and4 𝜇 G, with a two-solar metallicity and a pre-shock density of 1 cm − ,successfully reproduce the line ratios of regions C, D and E. In the MNRAS000 Top: [N ii]/H 𝛼 vs. [O iii]/H 𝛽 and [S ii]/H 𝛼 vs. [O iii]/H 𝛽 BPT diagrams; grey countours are galaxies from the SDSS main galaxy sample. red pointsare the spaxel of NGC 2992. Colored markers are the selected regions of NGC 2992, as shown in the bottom panels, representing: red (Region A), blue (B),green (C), pink (D) and cyan (E). Middle: Zoomed BPT diagrams with shock+precursor grid models as measured with MAPPINGS V code (Sutherland et al.2018); the (cid:174) 𝐵 indicates increasing magnetic field of the model (from 0.01 to 4 𝜇 G); the black numbers indicates the velocities of the shocks, ranging from 300 to900 km s − . Bottom: Respectively, from left to right, [O iii]/H 𝛽 , [N ii]/H 𝛼 , and [S ii]/H 𝛼 emission lines ratio maps; red circles show the selected regions shownin the upper panels. In BPT diagrams the grey crosses are the typical errors in the emission line ratios; the solid black lines are Kewley et al. (2001) theoreticalupper limits for pure star forming galaxies. In the [N ii]/H 𝛼 vs. [O iii]/H 𝛽 diagram the dashed black line represents Kauffmann et al. (2003) line which tracesthe upper boundary of the SDSS star formation sequence. In the [S ii]/H 𝛼 vs. [O iii]/H 𝛽 diagram the dashed-dotted black line represents the empirical divisionbetween AGN like and low-ionisation nuclear emission-line region (LINER) (Kewley et al. 2006). by star formation. The dashed-dotted black curve on the [S ii]/H 𝛼 vs. [O iii]/H 𝛽 diagnostic diagram traces the Kewley et al. (2006) em-pirical classification line (K06 line), which separates high ionisationspectra associated with Seyfert (AGN) and other intermediary ioni-sation mechanisms including shock excitation. We also show as greypoints in the top BPT diagrams a collection of single-fibre emissionline measurement from SDSS main galaxy sample by Kewley et al.(2006).Region A is the peak of the [O iii]/H 𝛽 map, it shows clear AGN-like ratios, with an upper [O iii]/H 𝛽 value higher than any SDSSAGN, which corresponds to a 3 (cid:48)(cid:48) diameter aperture, and thereforehas a diluted AGN/host-galaxy emission line ratio. Region B has thelowest value in all the maps; it lies below K01, showing the radialdecrease in the AGN contribution to the emission lines and significantcontributions from both star formation and AGN excitation. Regions C, D, and E display the highest values in the [S ii]/H 𝛼 and [N ii]/H 𝛼 maps; however, they have lower [O iii]/H 𝛽 values than region A.Region E, in fact, crosses the K06 line in the [S ii]/H 𝛼 diagram,which again indicates the presence of another ionisation mechanism.In order to search for an explanation to the increase in this in-termediary ionisation energy lines ([N ii] and [S ii]) we used theMAPPINGS V shock models by Sutherland et al. (2018). We ex-plored the parameter space in the pure shock and shock+precursorgrids of the models. None of the pure shock models could ex-plain these line ratios, given the high [O iii]/H 𝛽 values. Howevershock+precursor (Shock ionisation in a medium already ionisedby power-law like emission) with shock velocities between 300km s − and 900 km s − and magnetic field ( (cid:174) B) between 0.01 𝜇 G and4 𝜇 G, with a two-solar metallicity and a pre-shock density of 1 cm − ,successfully reproduce the line ratios of regions C, D and E. In the MNRAS000 , 1–20 (2021) apping the Inner kpc of NGC 2992 middle panels of Fig. 12 we show the grid models in these shockvelocity and magnetic field ranges superimposed with the emissionline ratios of the regions. The shocks in NGC 2992 have been exten-sively studied by Allen et al. (1999), who conclude that shocks arethe predominant ionisation mechanism in the extended NLR conesat a distance of several kpc from the nucleus.We can conclude that a more simple Starburst-AGN mixture (e.g.Davies et al. 2014a,b; D’Agostino et al. 2018) in which the maxi-mum values of all [O iii]/H 𝛽 , [N ii]/H 𝛼 and [S ii]/H 𝛼 are located inthe innermost spaxel (location of the AGN), and all decrease radiallytowards the H ii ionisation region of the BPT in a straight line (seefor example figure 1 in Davies et al. 2014b), cannot describe theionisation at the circumnuclear region of NGC 2992. Instead, a morecomplex Starburst-AGN-Shocks mixture is responsible for its ioni-sation. However, the very high values of [O iii]/H 𝛽 indicate the mostimportant among these three mechanisms is AGN ionisation, at leastin the innermost portion (see further discussion in subsection 6.4). In general terms, the stellar population synthesis discussed in sec-tion 3 shows that stellar content in the inner 1.1 kpc of NGC 2992 ismainly composed by an old metal-rich population, with a smaller butconsiderable contribution from young metal-poor stars. The presenceof the latter is supported by other studies, for instance, using the Bica& Alloin (1986) base of star clusters for stellar population synthesis,Storchi Bergmann et al. (1990) found a predominance of old stellarpopulations (10 yr) and a small contribution of at least 5 per centby recent star formation in the inner 5 (cid:48)(cid:48) . Moreover, Friedrich et al.(2010) find that 10-20 per cent of the IR flux is attributable to a star-burst, based on the emission from polycyclic aromatic hydrocarbon(PAH) molecules. Additionally, these authors estimate the starburstto have occurred between 40-50 Myr, based on the equivalent widthof the Br 𝛾 line.Major merger processes are known to be responsible for gas inflowtowards the central regions of galaxies enhancing circumnuclear starformation and possibly triggering nuclear activity. Merger galaxieshave an increase in the SFR, mainly at the nuclear regions (Elli-son et al. 2013; Pan et al. 2019). Thus, these sources have youngernuclear populations than those found in isolated galaxies, in whichmost of their recent formed stars are concentrated at the spiral arms.Moreover, major mergers are also able to modify the metallicitygradients of galaxies (Barrera-Ballesteros et al. 2015). Galaxy simu-lations show that the gas, originally located at external portions of theisolated galaxy (metal-poor regions), moves towards internal regions(metal-rich regions) of the companion galaxy during the encounterand is capable of cooling and forming stars.. Such processes are ableto explain both the increase in the SFR and the modifications inthe metallicity gradient (Dalcanton 2007; Torrey et al. 2012; Silleroet al. 2017). In fact, the pericentre passage between NGC 2992 andits companion NGC 2993 is estimated to have occurred ∼ 100 Myrago (Duc et al. 2000). Thus, a possible scenario is that metal-poorgas inflow has led to interaction-driven circumnuclear star forma-tion which can explain the presence of such young metal-poor stellarpopulation in the nucleus. Such inflows could also be responsiblefor triggering NGC 2992’s nuclear activity, a scenario which will befurther explored using numerical simulations in Paper III. We can compare the mass outflow rate, estimated in subsection 4.3,with the accretion rate necessary to power the AGN at the nucleus ofNGC 2992, calculated as follows: (cid:164) 𝑀 BH = 𝐿 𝑏𝑜𝑙 𝑐 𝜂 (9)where 𝜂 is the mass-energy conversion efficiency, which for Seyfertgalaxies is usually assumed to be 𝜂 = . 𝐿 𝑏𝑜𝑙 is the bolometric luminosity of the AGN, and 𝑐 is the speed of light.To measure 𝐿 𝑏𝑜𝑙 we applied the bolometric correction by Marconiet al. (2004) on the most recent X-ray luminosity measurement ofNGC 2992 made by Marinucci et al. (2018). The author used a2015 NuSTAR observation and obtained an absorption corrected 2-10 Kev Luminosity of 7 . ± . × ergs − that translates to 𝐿 𝑏𝑜𝑙 = . ± . × ergs − . We use these values to derive anaccretion rate of (cid:164) 𝑀 BH ≈ . 02 M (cid:12) 𝑦𝑟 − .The nuclear accretion rate is two orders of magnitude smallerthan the mass outflow rate ( (cid:164) 𝑀 𝑜𝑢𝑡 ≈ ± (cid:12) yr − , see sub-section 4.3). This implies that most of the outflowing gas does notoriginate in the AGN, but in the surrounding ISM, and this resultsupports the scenario in which the plasma bubbles (radio loops) areexpanding, pushing gas away from the nuclear region. However, theratio between the kinetic power of this blueshifted outflow loop andthe bolometric luminosity is ∼ . (cid:164) 𝐸 𝑜𝑢𝑡 / 𝐿 𝑏𝑜𝑙 should be closer to 5 per cent or above in order to effec-tively suppress growth in the most massive galaxies. However, somemodels show that coupling efficiencies of (cid:164) 𝐸 𝑜𝑢𝑡 / 𝐿 𝑏𝑜𝑙 ≈ Comparing our outflow measurements with those from the litera-ture we noted ours results disagree with those presented in Müller-Sánchez et al. (2011), who reported (cid:164) 𝑀 𝑜𝑢𝑡 = 120 M (cid:12) yr − and (cid:164) 𝐸 𝑜𝑢𝑡 = 2.5 × erg s − , both almost one hundred times greater than ours.We attribute this difference to two factors: i) they assumed the fillingfactor 𝑓 to be constant equal to 0.001, while we have measured thefilling factor from L(H 𝛼 ) and 𝑛 𝑒 (see Equation 6). This assumptionmakes their M 𝑔 proportional to 𝑛 𝑒 , and not inversely proportional(as in Equation 7); ii) they assume an electronic density of 𝑛 𝑒 = 5000cm − , a value more than 5 times larger than the maximum valuewe measured in the nuclear region of the galaxy. These two factorscombined explain their overestimation of both (cid:164) 𝑀 𝑜𝑢𝑡 and (cid:164) 𝐸 𝑜𝑢𝑡 . Wealso argue these differences cannot be due to a larger coverage ofthe outflow by a larger field of view, in the sense of seeing the far-thest part of the radio emission major loop and its redshifted outflowstructure, given the authors have used the OSIRIS/VLT instrument,which has an even smaller FoV than GMOS.We can compare our values to the properties of outflows found inother precious papers across the literature. In Fig. 11, we show theFiore et al. (2017) compilation of a set of outflow energetics fromvarious studies over the years spanning a large range of AGN lumi-nosity ( 𝐿 𝑏𝑜𝑙 ≥ ), which they use to establish AGN wind scalingrelations, mass outflow rate and outflow power as a function of bolo-metric luminosity, together with several measurements of nearby MNRAS , 1–20 (2021) M. Guolo-Pereira et al. galaxies from individual studies compiled by us from the literatureor obtained by our collaborators: NGC 1068, NGC 3783, NGC 6814,NGC 7469 (Müller-Sánchez et al. 2011), NGC 4151 (Crenshawet al. 2015), Mrk 573 (Revalski et al. 2018), Mrk 34 (Revalskiet al. 2018), NGC 7582 (Riffel et al. 2009), Mrk 1066 (Riffel et al.2010b), Mrk 1157 (Riffel & Storchi-Bergmann 2011), Mrk 79 (Riffelet al. 2013a), NGC 5929 (Riffel et al. 2013b), NGC 5728 (Shimizuet al. 2019), and NGC 3081 (Schnorr-Müller et al. 2014), ESO 362-G18 (Humire et al. 2018), NGC 1386 (Lena et al. 2015), ESO 153-G20 (Soto-Pinto et al. 2019), 3C 33 (Couto et al. 2017), includingNGC 2992 measurement by Müller-Sánchez et al. (2011). Althoughthese estimates are made using different recipes, leading to differ-ences between methods, they are generally similar to the one adoptedhere. Therefore, this comparison with the literature may be useful atleast as an order of magnitude approximation.As we can see, the overestimated previous measurement ofNGC 2992’s outflow made this source to be an outlier in the Fioreet al. (2017) relations, while our values make the galaxy more likelyto be explained by these relations. However, the other measurementsin nearby Seyfert galaxies support the existence of a larger scatterof these relations, both in (cid:164) 𝐸 𝑜𝑢𝑡 and (cid:164) 𝑀 𝑜𝑢𝑡 , at lower luminosities. Toconfirm whether this extension is valid or not, more detailed studieslike this, mainly using IFU, are needed so that outflow properties canbe properly and systematically measured, without assuming fixedvalues of 𝑛 𝑒 and 𝑓 . In section 5 we showed that despite being dominated by AGN ex-citation, star formation and shocks cannot be excluded as ionisationmechanisms in the nuclear region of the galaxy. The existence of starformation ionisation in the nucleus is confirmed by Friedrich et al.(2010) IR emission line analyses, as mentioned in subsection 6.1.The presence of shocks is also supported by Allen et al. (1999) us-ing long-slit spectroscopy. The authors also argue the dominance ofshocks increases, compared to ionisation by the central source, atlarger radius.Galaxies undergoing rapid phases of evolution (through processessuch as galaxy-galaxy interactions, gas outflows and radio jets) oftenhave multiple ionisation mechanisms contributing to their opticalline emission (e.g. Rich et al. 2011; Lanz et al. 2015; Rich et al.2015; Davies et al. 2017). These multiple ionisation mechanisms arealso in agreement with evolutionary scenarios in which gas inflows(driven either by secular mechanisms, like nuclear bars or nuclearspiral arms, or environmental ones, like minor and major mergers)triggers nuclear star formation and nuclear activity (feeding process)and then the active nucleus can drive both ionising photons from theaccretion disk and radio jets (feedback process) that impacts the ISMphysical conditions (Hopkins et al. 2008; Hopkins & Quataert 2010). We have analysed the stellar population, the ionisation mechanismand kinematics of the ionised gas in the inner 1 . ≈ 120 pc and spectral resolution ≈ 40 km s − . The main results are: • The stellar population in the nuclear region of the galaxy ismainly composed (60 per cent ≤ 𝑥 𝑂 ≤ 80 per cent) by an old (t > M o u t [ M y r ] Nearby SeyfertsFiore+17 Ionised WindsNGC2992 (Mueller-Sanchez+11)NGC2992 (This Work) L Bol [ erg s ] E o u t [ e r g s ] Figure 13. Correlations between the mass outflow rate ( (cid:164) 𝑀 𝑜𝑢𝑡 , Top Panel)and kinetic outflow power ( (cid:164) 𝐸 𝑜𝑢𝑡 , Bottom Panel). Blue points are the datacompiled by Fiore et al. (2017) and the black dashed lines show the best fitcorrelations for these points. Red Squares are the previous measurement ofNGC 2992 by Müller-Sánchez et al. (2011), and green ones are the results ofthis work. Violet diamonds are compiled from various literature sources formore modest AGN luminosities. (cid:104) Z (cid:105) 𝑂 ≥ . (cid:12) ) population with a smaller, butconsiderable contribution (10 per cent ≤ 𝑥 𝑌 ≤ 30 per cent) of anyoung (t < 100 Myr) metal-poor ( (cid:104) Z (cid:105) 𝑌 ≤ . (cid:12) ) population. Apossible scenario is that metal-poor gas inflow during the pericentrepassage of NGC 2993 has led to interaction-driven circumnuclear starformation, which can explain the presence of such young metal-poorstellar population in the nuclear region; • The emission line analyses show the presence of both BLRand NLR emission, confirming its classification as an intermediate-type Seyfert galaxy. The BLR emission profile has a velocity dis-persion of FWHM BLR = − and a Balmer decrement of(H 𝛼 /H 𝛽 ) BLR = . 2. The NLR presents two distinctly kinemati-cal components: one that although disturbed is well fitted by a diskmodel, and therefore can be identified as gas in orbit in the galaxydisk, with radial velocities ranging from -130 to 130 km s − ; anotherthat we interpreted as an outflow associated with the radio emission,with blueshifted velocities of ∼ 200 km s − , a mass outflow rate of (cid:164) 𝑀 𝑜𝑢𝑡 ≈ ± (cid:12) and a kinematic power of (cid:164) 𝐸 𝑜𝑢𝑡 ≈ ± × erg s − ; • The BPT diagnostic diagrams show the galaxy posses multi-ple ionisation mechanisms: a mixture of AGN, star-formation andshocks, the former being the dominant one. The [O iii]/H 𝛽 ratiopeaks at the innermost spaxels, and is located in the upper region MNRAS000 200 km s − , a mass outflow rate of (cid:164) 𝑀 𝑜𝑢𝑡 ≈ ± (cid:12) and a kinematic power of (cid:164) 𝐸 𝑜𝑢𝑡 ≈ ± × erg s − ; • The BPT diagnostic diagrams show the galaxy posses multi-ple ionisation mechanisms: a mixture of AGN, star-formation andshocks, the former being the dominant one. The [O iii]/H 𝛽 ratiopeaks at the innermost spaxels, and is located in the upper region MNRAS000 , 1–20 (2021) apping the Inner kpc of NGC 2992 of the BPT diagram. The lower values, located at the border of theFoV, decreases down to below the Kewley et al. (2001) line, showingan increase in the star-formation ionisation compared to the AGNionisation at larger radius. Off-nuclear peaks in the [N ii]/H 𝛼 and[S ii]/H 𝛼 maps indicate the presence of another mechanism and aresuccessfully explained by Sutherland et al. (2018) shock+precursormodels. ACKNOWLEDGMENTS We thank the anonymous referee whose comments helped us im-prove the methodology and the clarity of this work. G-P, M. thanksthe Brazilian National Council for Scientific and Technological De-velopment (CNPq) for his Master Scholarship and Taro Shimizu forproviding a collection of data points for Fig. 13. G. C. acknowledgesthe support by the Comité Mixto ESO-Chile and the DGI at Uni-versity of Antofagasta. J. A. H. J. thanks to Chilean institution CON-ICYT, Programa de Astronomía, Fondo ALMA-CONICYT 2017,Código de proyecto 31170038. N.Z.D. acknowledges partial supportfrom FONDECYT through project 3190769. This work is based onobservations obtained at the Gemini Observatory, which is operatedby the Association of Universities for Research in Astronomy, Inc.,under a cooperative agreement with the NSF on behalf of the Gem-ini partnership: the National Science Foundation (United States),the Science and Technology Facilities Council (United Kingdom),the National Research Council (Canada), CONICYT (Chile), theAustralian Research Council (Australia), Ministério da Ciência eTécnologia (Brazil) and south-eastCYT (Argentina). DATA AVAILABILITY The data underlying this article will be shared on reasonable requestto the corresponding author. REFERENCES Allen M. G., Dopita M. A., Tsvetanov Z. I., Sutherland R. S., 1999, ApJ, 511,686Allen M. G., Groves B. A., Dopita M. A., Sutherland R. S., Kewley L. J.,2008, ApJS, 178, 20Alloin D., Collin-Souffrin S., Joly M., Vigroux L., 1979, A&A, 78, 200Bacon R., et al., 2010, in McLean I. S., Ramsay S. K., Takami H., eds, Societyof Photo-Optical Instrumentation Engineers (SPIE) Conference SeriesVol. 7735, Ground-based and Airborne Instrumentation for AstronomyIII. p. 773508, doi:10.1117/12.856027Baldwin J. A., Phillips M. M., Terlevich R., 1981, PASP, 93, 5Baldwin C., McDermid R. M., Kuntschner H., Maraston C., Conroy C., 2018,MNRAS, 473, 4698Barrera-Ballesteros J. K., et al., 2015, A&A, 579, A45Begelman M. C., Blandford R. D., Rees M. J., 1984, Reviews of ModernPhysics, 56, 255Bennert N., Jungwiert B., Komossa S., Haas M., Chini R., 2006, Astronomy& Astrophysics, 456, 953–966Bertola F., Bettoni D., Danziger J., Sadler E., Sparke L., de Zeeuw T., 1991,ApJ, 373, 369Bica E., Alloin D., 1986, A&A, 162, 21Block A., 2011, Caelum-Observatory:NGC2992. Bloom J. V., et al., 2018, Monthly Notices of the Royal Astronomical Society,476, 2339–2351Blumenthal K. A., Barnes J. E., 2018, MNRAS, 479, 3952Bois M., et al., 2011, MNRAS, 416, 1654 Bruzual G., Charlot S., 2003, MNRAS, 344, 1000Bullock J. S., Boylan-Kolchin M., 2017, ARA&A, 55, 343Calzetti D., Armus L., Bohlin R. C., Kinney A. L., Koornneef J., Storchi-Bergmann T., 2000, ApJ, 533, 682Cappellari M., Copin Y., 2003, MNRAS, 342, 345Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345, 245Cardoso L. S. M., Gomes J. M., Papaderos P., 2017, A&A, 604, A99Chabrier G., 2003, PASP, 115, 763Chapman S. C., Morris S. L., Alonso-Herrero A., Falcke H., 2000, MNRAS,314, 263Chen X. Y., Liang Y. C., Hammer F., Prugniel P., Zhong G. H., Rodrigues M.,Zhao Y. H., Flores H., 2010, Astronomy and Astrophysics, 515, A101Cid Fernandes R., 2018, MNRAS, 480, 4480Cid Fernandes R., González Delgado R. M., 2010, Monthly Notices of theRoyal Astronomical Society, 403, 780Cid Fernandes R., Gu Q., Melnick J., Terlevich E., Terlevich R., Kunth D.,Rodrigues Lacerda R., Joguet B., 2004, MNRAS, 355, 273Cid Fernandes R., Mateus A., Sodré L., Stasińska G., Gomes J. M., 2005,MNRAS, 358, 363Cid Fernandes R., Stasińska G., Mateus A., Vale Asari N., 2011, MonthlyNotices of the Royal Astronomical Society, 413, 1687–1699Cid Fernandes R., et al., 2013, A&A, 557, A86Colina L., Fricke K. J., Kollatschny W., Perryman M. A. C., 1987, A&A, 178,51Conroy C., 2013, Annual Review of Astronomy and Astrophysics, 51,393–455Costa T., Rosdahl J., Sijacki D., Haehnelt M. G., 2018, MNRAS, 479, 2079Couto G. S., Storchi-Bergmann T., Schnorr-Müller A., 2017, MNRAS, 469,1573Crenshaw D. M., Fischer T. C., Kraemer S. B., Schmitt H. R., 2015, ApJ,799, 83Croton D. J., et al., 2006, MNRAS, 365, 11D’Agostino J. J., Poetrodjojo H., Ho I.-T., Groves B., Kewley L., Madore B. F.,Rich J., Seibert M., 2018, Monthly Notices of the Royal AstronomicalSociety, 479, 4907–4935Dalcanton J. J., 2007, The Astrophysical Journal, 658, 941Dametto N. Z., et al., 2019, MNRAS, 482, 4437Davies R. L., Rich J. A., Kewley L. J., Dopita M. A., 2014a, MNRAS, 439,3835Davies R. L., Kewley L. J., Ho I. T., Dopita M. A., 2014b, MNRAS, 444,3961Davies R. L., et al., 2017, Monthly Notices of the Royal Astronomical Society,470, 4974–4988Di Matteo T., Springel V., Hernquist L., 2005, Nature, 433, 604Dopita M. A., 2002, in Henney W. J., Steffen W., Binette L., Raga A., eds,Revista Mexicana de Astronomia y Astrofisica Conference Series Vol.13, Revista Mexicana de Astronomia y Astrofisica Conference Series. pp177–182Duc P.-A., Brinks E., Springel V., Pichardo B., Weilbacher P., Mirabel I. F.,2000, The Astronomical Journal, 120, 1238Ellison S. L., Patton D. R., Mendel J. T., Scudder J. M., 2011, MNRAS, 418,2043Ellison S. L., Mendel J. T., Patton D. R., Scudder J. M., 2013, MNRAS, 435,3627Event Horizon Telescope Collaboration et al., 2019, ApJ, 875, L1Fan L., et al., 2016, ApJ, 822, L32Ferrarese L., Merritt D., 2000, ApJ, 539, L9Ferré-Mateu A., Vazdekis A., Trujillo I., Sánchez-Blázquez P., RicciardelliE., de la Rosa I. G., 2012, MNRAS, 423, 632Fiore F., et al., 2017, A&A, 601, A143Fischer T. C., et al., 2018, ApJ, 856, 102Frank J., King A., Raine D. J., 2002, Accretion Power in Astrophysics: ThirdEditionFreund J., 1992, Annalen Der Physik - ANN PHYS-BERLIN, 504, 380Friedrich S., Davies R. I., Hicks E. K. S., Engel H., Müller-Sánchez F., GenzelR., Tacconi L. J., 2010, A&A, 519, A79García-Bernete I., et al., 2015, MNRAS, 449, 1309García-Lorenzo B., Arribas S., Mediavilla E., 2001, A&A, 378, 787MNRAS , 1–20 (2021) M. Guolo-Pereira et al. Ge J., Mao S., Lu Y., Cappellari M., Yan R., 2019, Monthly Notices of theRoyal Astronomical Society, 485, 1675–1693Gilli R., Maiolino R., Marconi A., Risaliti G., Dadina M., Weaver K. A.,Colbert E. J. M., 2000, A&A, 355, 485Glikman E., Simmons B., Mailly M., Schawinski K., Urry C. M., Lacy M.,2015, Major Mergers Host the Most Luminous Red Quasars at z 2: AHubble Space Telescope WFC3/IR Study ( arXiv:1504.02111 )González Delgado R. M., et al., 2015, A&A, 581, A103Greene J. E., Zakamska N. L., Smith P. S., 2012, ApJ, 746, 86Hopkins P. F., Elvis M., 2010, MNRAS, 401, 7Hopkins P. F., Quataert E., 2010, MNRAS, 407, 1529Hopkins P. F., Hernquist L., Cox T. J., Kereš D., 2008, ApJS, 175, 356Humire P. K., et al., 2018, A&A, 614, A94Hung C.-L., Hayward C. C., Smith H. A., Ashby M. L. N., Lanz L., Martínez-Galarza J. R., Sanders D. B., Zezas A., 2016, ApJ, 816, 99Irwin J. A., et al., 2017, MNRAS, 464, 1333Kauffmann G., et al., 2003, MNRAS, 346, 1055Kennicutt Robert C. J., 1998, ApJ, 498, 541Kewley L. J., Dopita M. A., Sutherland R. S., Heisler C. A., Trevena J., 2001,ApJ, 556, 121Kewley L. J., Groves B., Kauffmann G., Heckman T., 2006, MNRAS, 372,961Koleva M., Prugniel P., Ocvirk P., Le Borgne D., Soubiran C., 2008, MNRAS,385, 1998Kormendy J., Ho L. C., 2013, ARA&A, 51, 511Koski A. T., 1978, ApJ, 223, 56Koss M., Mushotzky R., Veilleux S., Winter L., 2010, ApJ, 716, L125Kronberger T., Kapferer W., Schindler S., Ziegler B. L., 2007, Astronomy &Astrophysics, 473, 761–770L. Trippe M., Crenshaw D., R. P. D., Dietrich M., 2008, AJ, 135Lanz L., Ogle P. M., Evans D., Appleton P. N., Guillard P., Emonts B., 2015,The Astrophysical Journal, 801, 17Le Borgne J.-F., et al., 2003, A&A, 402, 433Lena D., et al., 2015, ApJ, 806, 84Madau P., Dickinson M., 2014, ARA&A, 52, 415Maksym W. P., Ulmer M. P., Roth K. C., Irwin J. A., Dupke R., Ho L. C.,Keel W. C., Adami C., 2014, MNRAS, 444, 866Mallmann N. D., et al., 2018, MNRAS, 478, 5491Maraston C., Strömbäck G., 2011, MNRAS, 418, 2785Marconi A., Risaliti G., Gilli R., Hunt L. K., Maiolino R., Salvati M., 2004,MNRAS, 351, 169Marinucci A., Bianchi S., Braito V., Matt G., Nardini E., Reeves J., 2018,MNRAS, 478, 5638Marquez I., Boisson C., Durret F., Petitjean P., 1998, A&A, 333, 459McElroy R., Croom S. M., Pracy M., Sharp R., Ho I. T., Medling A. M.,2015, MNRAS, 446, 2186Menci N., Gatti M., Fiore F., Lamastra A., 2014, A&A, 569, A37Menezes R. B., Steiner J. E., Ricci T. V., 2014, MNRAS, 438, 2597Mentz J. J., et al., 2016, MNRAS, 463, 2819Mingozzi M., et al., 2019, A&A, 622, A146Müller-Sánchez F., Prieto M. A., Hicks E. K. S., Vives-Arias H., Davies R. I.,Malkan M., Tacconi L. J., Genzel R., 2011, ApJ, 739, 69Osterbrock D. E., Ferland G. J., 2006, Astrophysics of gaseous nebulae andactive galactic nucleiPan H.-A., et al., 2019, ApJ, 881, 119Peterson B. M., 1997, An Introduction to Active Galactic NucleiPeterson B. M., Wanders I., Bertram R., Hunley J. F., Pogge R. W., WagnerR. M., 1998, ApJ, 501, 82Pettini M., Pagel B. E. J., 2004, MNRAS, 348, L59Proxauf B., Öttl S., Kimeswenger S., 2014, A&A, 561, A10Raimann D., Storchi-Bergmann T., González Delgado R. M., Cid FernandesR., Heckman T., Leitherer C., Schmitt H., 2003, MNRAS, 339, 772Revalski M., Crenshaw D. M., Kraemer S. B., Fischer T. C., Schmitt H. R.,Machuca C., 2018, ApJ, 856, 46Rich J. A., Kewley L. J., Dopita M. A., 2011, ApJ, 734, 87Rich J. A., Kewley L. J., Dopita M. A., 2015, The Astrophysical JournalSupplement Series, 221, 28 Riffel R. A., Storchi-Bergmann T., 2011, Monthly Notices of the Royal As-tronomical Society, 417, 2752–2769Riffel R. A., Storchi-Bergmann T., Dors O. L., Winge C., 2009, MNRAS,393, 783Riffel R., Pastoriza M. G., Rodríguez-Ardila A., Bonatto C., 2010a, in StellarPopulations - Planning for the Next Decade.Riffel R. A., Storchi-Bergmann T., Nagar N. M., 2010b, MNRAS, 404, 166Riffel R. A., Storchi-Bergmann T., Winge C., 2013a, MNRAS, 430, 2249Riffel R. A., Storchi-Bergmann T., Riffel R., 2013b, The Astrophysical Jour-nal, 780, L24Rupke D. S. N., Veilleux S., 2011, The Astrophysical Journal, 729, L27Ruschel-Dutra D., 2020, danielrd6/ifscube v1.0,doi:10.5281/zenodo.3945237, https://doi.org/10.5281/zenodo.3945237 Sánchez-Blázquez P., et al., 2006, MNRAS, 371, 703Satyapal S., Ellison S. L., McAlpine W., Hickox R. C., Patton D. R., MendelJ. T., 2014, MNRAS, 441, 1297Schawinski K., Dowlin N., Thomas D., Urry C. M., Edmondson E., 2010,The Astrophysical Journal, 714, L108Schimoia J. S., Storchi-Bergmann T., Grupe D., Eracleous M., Peterson B. M.,Baldwin J. A., Nemmen R. S., Winge C., 2015, ApJ, 800, 63Schlafly E. F., Finkbeiner D. P., 2011, ApJ, 737, 103Schnorr-Müller A., Storchi-Bergmann T., Nagar N. M., Robinson A., LenaD., Riffel R. A., Couto G. S., 2014, MNRAS, 437, 1708Schnorr-Müller A., et al., 2016, MNRAS, 462, 3570Schnorr-Müller A., Storchi-Bergmann T., Nagar N. M., Robinson A., LenaD., 2017, MNRAS, 471, 3888Shimizu T. T., et al., 2019, Monthly Notices of the Royal Astronomical SocietyShuder J. M., 1980, ApJ, 240, 32Silk J., Mamon G. A., 2012, Research in Astronomy and Astrophysics, 12,917Sillero E., Tissera P. B., Lambas D. G., Michel-Dansac L., 2017, MonthlyNotices of the Royal Astronomical Society, 472, 4404–4413Silverman J. D., et al., 2008, ApJ, 679, 118Soto-Pinto P., et al., 2019, MNRAS, 489, 4111Springel V., et al., 2005, Nature, 435, 629Stasinska G., Asari N. V., Fernandes R. C., Gomes J. M., Schlickmann M.,Mateus A., Schoenell W., Sodré Jr L., 2008, Monthly Notices of the RoyalAstronomical Society: LettersStorchi Bergmann T., Bica E., Pastoriza M. G., 1990, MNRAS, 245, 749Storchi-Bergmann T., Raimann D. I., González Delgado R. M., Schmitt H. R.,Cid Fernandes R., Heckman T., Leitherer C., 2003, in Perez E., GonzalezDelgado R. M., Tenorio-Tagle G., eds, Astronomical Society of the Pa-cific Conference Series Vol. 297, Star Formation Through Time. p. 363( arXiv:astro-ph/0211474 )Storchi-Bergmann T., et al., 2018, ApJ, 868, 14Strauss M. A., et al., 2002, AJ, 124, 1810Sun A.-L., Greene J. E., Zakamska N. L., 2017, The Astrophysical Journal,835, 222Sutherland R., Dopita M., Binette L., Groves B., 2018, MAPPINGS V: As-trophysical plasma modeling code (ascl:1807.005)Theureau G., Hanski M. O., Coudreau N., Hallet N., Martin J. M., 2007,A&A, 465, 71Torres-Flores S., Amram P., Mendes de Oliveira C., Plana H., Balkowski C.,Marcelin M., Olave-Rojas D., 2014, MNRAS, 442, 2188Torrey P., Cox T. J., Kewley L., Hernquist L., 2012, The Astrophysical Journal,746, 108Treister E., Schawinski K., Urry C. M., Simmons B. D., 2012, The Astro-physical Journal, 758, L39Trippe M. L., Crenshaw D. M., Deo R., Dietrich M., 2008, AJ, 135, 2048Ulvestad J. S., Wilson A. S., 1984, ApJ, 285, 439Veilleux S., Osterbrock D. E., 1987, ApJS, 63, 295Veilleux S., Shopbell P. L., Miller S. T., 2001, AJ, 121, 198Veilleux S., et al., 2013, ApJ, 776, 27Veron P., Lindblad P. O., Zuiderwijk E. J., Veron M. P., Adam G., 1980,A&A, 87, 245Villar-Martín M., Humphrey A., Delgado R. G., Colina L., Arribas S., 2011,MNRAS, 418, 2032MNRAS000 Sánchez-Blázquez P., et al., 2006, MNRAS, 371, 703Satyapal S., Ellison S. L., McAlpine W., Hickox R. C., Patton D. R., MendelJ. T., 2014, MNRAS, 441, 1297Schawinski K., Dowlin N., Thomas D., Urry C. M., Edmondson E., 2010,The Astrophysical Journal, 714, L108Schimoia J. S., Storchi-Bergmann T., Grupe D., Eracleous M., Peterson B. M.,Baldwin J. A., Nemmen R. S., Winge C., 2015, ApJ, 800, 63Schlafly E. F., Finkbeiner D. P., 2011, ApJ, 737, 103Schnorr-Müller A., Storchi-Bergmann T., Nagar N. M., Robinson A., LenaD., Riffel R. A., Couto G. S., 2014, MNRAS, 437, 1708Schnorr-Müller A., et al., 2016, MNRAS, 462, 3570Schnorr-Müller A., Storchi-Bergmann T., Nagar N. M., Robinson A., LenaD., 2017, MNRAS, 471, 3888Shimizu T. T., et al., 2019, Monthly Notices of the Royal Astronomical SocietyShuder J. M., 1980, ApJ, 240, 32Silk J., Mamon G. A., 2012, Research in Astronomy and Astrophysics, 12,917Sillero E., Tissera P. B., Lambas D. G., Michel-Dansac L., 2017, MonthlyNotices of the Royal Astronomical Society, 472, 4404–4413Silverman J. D., et al., 2008, ApJ, 679, 118Soto-Pinto P., et al., 2019, MNRAS, 489, 4111Springel V., et al., 2005, Nature, 435, 629Stasinska G., Asari N. V., Fernandes R. C., Gomes J. M., Schlickmann M.,Mateus A., Schoenell W., Sodré Jr L., 2008, Monthly Notices of the RoyalAstronomical Society: LettersStorchi Bergmann T., Bica E., Pastoriza M. G., 1990, MNRAS, 245, 749Storchi-Bergmann T., Raimann D. I., González Delgado R. M., Schmitt H. R.,Cid Fernandes R., Heckman T., Leitherer C., 2003, in Perez E., GonzalezDelgado R. M., Tenorio-Tagle G., eds, Astronomical Society of the Pa-cific Conference Series Vol. 297, Star Formation Through Time. p. 363( arXiv:astro-ph/0211474 )Storchi-Bergmann T., et al., 2018, ApJ, 868, 14Strauss M. A., et al., 2002, AJ, 124, 1810Sun A.-L., Greene J. E., Zakamska N. L., 2017, The Astrophysical Journal,835, 222Sutherland R., Dopita M., Binette L., Groves B., 2018, MAPPINGS V: As-trophysical plasma modeling code (ascl:1807.005)Theureau G., Hanski M. O., Coudreau N., Hallet N., Martin J. M., 2007,A&A, 465, 71Torres-Flores S., Amram P., Mendes de Oliveira C., Plana H., Balkowski C.,Marcelin M., Olave-Rojas D., 2014, MNRAS, 442, 2188Torrey P., Cox T. J., Kewley L., Hernquist L., 2012, The Astrophysical Journal,746, 108Treister E., Schawinski K., Urry C. M., Simmons B. D., 2012, The Astro-physical Journal, 758, L39Trippe M. L., Crenshaw D. M., Deo R., Dietrich M., 2008, AJ, 135, 2048Ulvestad J. S., Wilson A. S., 1984, ApJ, 285, 439Veilleux S., Osterbrock D. E., 1987, ApJS, 63, 295Veilleux S., Shopbell P. L., Miller S. T., 2001, AJ, 121, 198Veilleux S., et al., 2013, ApJ, 776, 27Veron P., Lindblad P. O., Zuiderwijk E. J., Veron M. P., Adam G., 1980,A&A, 87, 245Villar-Martín M., Humphrey A., Delgado R. G., Colina L., Arribas S., 2011,MNRAS, 418, 2032MNRAS000 , 1–20 (2021) apping the Inner kpc of NGC 2992 Figure A1. Maps of the recovered properties from the BC03 synthesis. Top:from left to right, percentage contribution to the total light from young,intermediate and old SSPs ( 𝑥 𝑌 , 𝑥 𝐼 and 𝑥 𝑂 ). Middle: from left to right, light-weighted mean stellar metallicity for the three SSP age groups, (cid:104) Z (cid:105) 𝑌 , (cid:104) Z (cid:105) 𝐼 and (cid:104) Z (cid:105) 𝑂 . Bottom left: 𝐴 𝑉 extinction parameter. Bottom middle: Featurelesscontinuum (FC) component percentage contribution to the total light ( 𝑥 𝑎𝑔𝑛 ).Bottom Right: Light-weighted mean stellar age, (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) . The grey contourin all the maps is the continuum emission, left panel of Fig. 1. The 𝑥 𝑎𝑔𝑛 isnull in all points by definition, because there is no FC component in BC03 synthesis spectra base.Ward M., Penston M. V., Blades J. C., Turtle A. J., 1980, MNRAS, 193, 563Wilkinson D. M., Maraston C., Goddard D., Thomas D., Parikh T., 2017,MNRAS, 472, 4297Zakamska N. L., et al., 2016, MNRAS, 459, 3144 APPENDIX A: STELLAR POPULATIONS SYNTHESISMAPS We show the maps of the recovered properties from three spatiallyresolved starlight synthesis we have performed, namely BC03 , BC03+FC and M11+FC . The top panels show, from left to right, theper cent contribution to the total light for young, intermediate andold SSPs ( 𝑥 𝑌 , 𝑥 𝐼 and 𝑥 𝑂 ). In the middle panels the light-weightedmean stellar metallicity for the three age groups, from left to right, (cid:104) Z (cid:105) 𝑌 , (cid:104) Z (cid:105) 𝐼 and (cid:104) Z (cid:105) 𝑂 . In the bottom panel the V -Band extinction( 𝐴 𝑉 ), the per cent contribution to the total light from the FC com-ponent ( 𝑥 𝑎𝑔𝑛 ) and the light-weighted mean stellar age, (cid:104) 𝑙𝑜𝑔 𝑡 ∗ (cid:105) , areshown, respectively, from left to right. APPENDIX B: EMISSION LINE FITTINGUNCERTAINTIES We present here the uncertainties in the parameters of the emissionlines fitting described in subsection 4.1. The uncertainties are thestandard deviation over the one hundred Monte Carlo iterations. Theuncertainties in the radial velocity (V) of Component 1 was alreadypresented in Fig. 8. In Fig. B1 we show the uncertainties in thevelocity dispersion ( 𝜎 ) of Component 1, the values are mostly smallerthan 5 km s − , reaching a maximum value of 15 km s − . Figure A2. Same as Fig. A1for BC03+FC . Figure A3. Same as Fig. A1 for M11+FC . In the top panel we also show the uncertainties of 𝑣 and 𝜎 ofComponent 2, values are less than 10 km s − in most the region,reaching up to 30 km s − in a very small region, the 𝜎 values areless then 10 km s − in the peak of component emission and up to 30km s − in the borders of the detection region. In the middle and bottompanels the percentile uncertainties in the integrated flux (Component1 + Component 2) of H 𝛽 , [O iii] 𝜆 𝛼 , [N ii] 𝜆 𝜆 𝛽 line uncertainties range from 10 per cent at the centre ofthe FoV, up to 45 per cent in the dust lane region. [O iii] and H 𝛼 values range from less than 5 per cent up to 25 per cent at the dustlane region. [S ii] and [N ii] values range from less than 10 per centup to 35 per cent at the dust lane region. MNRAS , 1–20 (2021) M. Guolo-Pereira et al. -3-2-10123 Y ( a r c s e c ) Error (Comp. 1) Error V (Comp. 2) E N Error (Comp. 2)-3-2-10123 Y ( a r c s e c ) Error Flux [H ] Error Flux [OIII] E N Error Flux [H ] -3 -2 -1 0 1 2 3 X (arcsec) -3-2-10123 Y ( a r c s e c ) Error Flux [NII] -3 -2 -1 0 1 2 3 X (arcsec) Error Flux [SII] 6716 -3 -2 -1 0 1 2 3 X (arcsec) E N Error Flux [SII] 6731 Figure B1. Emission line fitting uncertainties. Top: from right to left, velocitydispersion of Component 1 ( km s − ), radial velocity of Component 2 (km s − ), velocity dispersion of Component 2 ( km s − ). Middle and Bottom:percentage uncertainties in the integrated flux (Component 1 + Component)of H 𝛽 , [O iii] 𝜆 𝛼 , [N ii] 𝜆 𝜆 APPENDIX C: NEBULAR EXTINCTION AND ELECTRONDENSITY In order to measure the extinction in the NLR, we adopt the Calzettiet al. (2000) extinction law, and assume the case B recombination andthe intrinsic H 𝛼 /H 𝛽 value for the NLR (Osterbrock & Ferland 2006),which combined leads to the following expression to the V -bandextinction: 𝐴 v = . × log (cid:18) H 𝛼 / H 𝛽 . (cid:19) (C1)In order to improve the S/N and given the fact that H 𝛽 Component2 is detected in a very small region, we used the integrated flux(Component 1 + Component 2) of the NLR H 𝛼 and H 𝛽 lines. The 𝐴 v map is presented in the top left panel of Fig. C1, and its uncertaintyin the bottom left panel. The 𝐴 v peak at ∼ > 𝛽 flux measurement, as seen in Fig. B1. Although the 𝐴 v are not the same in the continuum Appendix A and in the NLR, thespatial patterns are similar: the higher values are found the centralspaxels and towards the dust lane, and the lower towards the SEdirection.The electron density ( 𝑛 𝑒 ) of the ionised gas in the NLR can beobtained from the ratio between the [S ii] lines, [ S ii ] 𝜆 𝜆 𝑛 𝑒 values, as follows: -3-2-10123 Y ( a r c s e c ) A v (NLR) E N n e -3 -2 -1 0 1 2 3 X (arcsec)-3-2-10123 Y ( a r c s e c ) Error A v (NLR) -3 -2 -1 0 1 2 3 X (arcsec) E N Error n e Figure C1. Top left and bottom left panels, shows, respectively the V -bandextinction 𝐴 v map and its uncertainties, both in units of magnitudes. Topright and bottom right panel show, respectively, the electron density ( 𝑛 𝑒 ) andits uncertainties both in cm − units. The black cross shows the peak of thecontinuum emission. log ( 𝑛 𝑒 [ 𝑐𝑚 ]) = . (− . 𝑅 + . )+ . − . 𝑅 + . 𝑅 − . 𝑅 (C2)with R = 𝐹 / 𝐹 , the ratio between the [S ii] lines. The 𝑛 𝑒 mapin its uncertainties are presented, respectively, in the top right andbottom right panel of Fig. C1. The maximum value of 𝑛 𝑒 is 950 𝑐𝑚 − reached at the peak of the emission lines, decreasing radially downto 100 𝑐𝑚 − , another high ( ∼ 𝑐𝑚 − ) density knot is presentat the south of the nucleus. Both the values as well as the spatialprofiles of both 𝐴 v and 𝑛 𝑒 measured here are in agreement with therecent published analyses by Mingozzi et al. (2019) using MUSEinstrument, see their Figures A1 and B1. This paper has been typeset from a TEX/L A TEX file prepared by the author.MNRAS000