A study of the interplay between ionized gas and star clusters in the central region of NGC 5253 with 2D spectroscopy
A. Monreal-Ibero, J. M. Vilchez, J. R. Walsh, C. Munoz-Tunon
aa r X i v : . [ a s t r o - ph . C O ] M a r Astronomy&Astrophysicsmanuscript no. ms c (cid:13)
ESO 2018October 16, 2018
A study of the interplay between ionized gas and star clusters inthe central region of NGC 5253 with 2D spectroscopy ⋆ A. Monreal-Ibero , J. M. V´ılchez , J. R. Walsh , and C. Mu˜noz-Tu˜n´on European Organisation for Astronomical Research in the Southern Hemisphere, Karl-Schwarzschild-Strasse 2, D-85748 Garchingbei M¨unchen, Germanye-mail: [amonreal,jwalsh]@eso.org Instituto de Astrof´ısica de Andaluc´ıa (CSIC), C / Camino Bajo de Hu´etor, 50, 18008 Granada, Spaine-mail: jvm@@iaa.es Instituto de Astrof´ısica de Canarias, C / V´ıa L´actea, s / n, 38205 La Laguna, Spaine-mail: [email protected] manuscript ABSTRACT
Context.
Starbursts are one of the main contributors to the chemical enrichment of the interstellar medium. However, mechanismsgoverning the interaction between the recent star formation and the surrounding gas are not fully understood. Because of their a priori simplicity, the subgroup of H ii galaxies constitute an ideal sample to study these mechanisms. Aims.
A detailed 2D study of the central region of NGC 5253 has been performed to characterize the stellar and ionized gas structureas well as the extinction distribution, physical properties and kinematics of the ionized gas in the central ∼
210 pc ×
130 pc.
Methods.
We utilized optical integral field spectroscopy (IFS) data obtained with FLAMES.
Results.
A detailed extinction map for the ionized gas in NGC 5253 shows that the largest extinction is associated with the prominentGiant H ii region. There is an o ff set of ∼ . ′′ ff er less extinction than gas by a factor of ∼ ii ] λ / [S ii ] λ N e ) gradient declining from the peak of emission in H α (790 cm − ) outwards, while the argon line ratio tracesareas with N e ∼ − − . The area polluted with extra nitrogen, as deduced from the excess [N ii ] λ / H α , extends up todistances of 3 . ′′ ∼
60 pc) from the maximum pollution, which is o ff set by ∼ . ′′ ∼
100 pc ×
100 pc) and associated with young stellar clusters. Wemeasured He + abundances over most of the field of view and values of He ++ / H + ∼ < . ii region, presents supersonic widths and [N ii ] λ ii ] λλ − with respect to H α . Similarly, one of the narrow components presents o ff setsin the [N ii ] λ ∼ <
20 km s − . This is the first time that maps with such velocity o ff sets for a starburst galaxy have beenpresented. The observables in the giant H ii region fit with a scenario where the two super stellar clusters (SSCs) produce an outflowthat encounters the previously quiescent gas. The south-west part of the FLAMES IFU field is consistent with a more evolved stagewhere the star clusters have already cleared out their local environment. Key words.
Galaxies: starburst — Galaxies: dwarf — Galaxies: individual, NGC 5253 — Galaxies: ISM — Galaxies: abundances— Galaxies: kinematics and dynamics
1. Introduction
Starbursts are events characterized by star-formation rates muchhigher than those found in gas-rich normal galaxies. They areconsidered one of the main contributors to the chemical en-richment of the interstellar medium (ISM) and can be found ingalaxies covering a wide range of masses, luminosities, metal-licities and interaction stages such as blue compact dwarfs, nu-clei of spiral galaxies, or (Ultra)luminous Infrared Galaxies (seeConti et al. 2008, and references therein).A particularly interesting subset are the H ii galaxies, iden-tified for the first time by Haro (1956): gas-rich, metal poor(1 / Z ⊙ ∼ < Z ∼ < / Z ⊙ ) dwarf systems characterized by the pres-ence of large ionized H ii regions that dominate their optical Send o ff print requests to : A. Monreal-Ibero ⋆ Based on observations collected at the European Organisationfor Astronomical Research in the Southern Hemisphere, Chile (ESOProgramme 078.B-0043). spectra (see Kunth & ¨Ostlin 2000, for a review of these galax-ies). These systems are a priori simple, which makes them theideal laboratories to test the interplay between massive star for-mation and the ISM.NGC 5253, an irregular galaxy located in the Centarus A / M 83 galaxy complex (Karachentsev et al. 2007), is a lo-cal example of an H ii galaxy. This galaxy is su ff ering a burstof star formation which is believed to have been triggeredby an encounter with M 83 (van den Bergh 1980). This issupported by the existence of the H i plume extending alongthe optical minor axis which is best explained as tidal debris(Kobulnicky & Skillman 2008).NGC 5253 constitutes an optimal target for the study of thestarburst phenomenon. On the one hand, its proximity allows alinear spatial resolution to be achieved that is good enough tostudy the details of the interplay between the di ff erent compo-nents (i.e. gas, dust and star clusters) in the central region. Onthe other hand, this system has been observed in practically all A. Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES
Table 1.
Basic data for NGC 5253
Parameter Value Ref.Name NGC 5253 (a)Other designations ESO 445 − G004, Haro 10 (a)RA (J2000.0) 13h39m55.9s (a)Dec(J2000.0) − z D ( M pc ) 3.8 (b)scale (pc / ′′ ) 18.4 m B M B − .
13 (c) U − B − .
30 (c) B − V .
50 (c) V − R .
32 (c) M HI ( M ⊙ ) 1 . × M ⊙ (d) Z / Z ⊙ ∼ ( ∗ ) (e)log( L fir / L ⊙ ) 8.95 (f)log( L ir / L ⊙ ) 9.21 (f) ( ∗ ) We assumed 12 + log(O / H) ⊙ = .
66 (Asplund et al. 2004). (a)
NASA / IPAC Extragalactic Database (NED). (b)
Sakai et al. (2004). (c)
Taylor et al. (2005). (d)
Kobulnicky & Skillman (2008). (e)
Kobulnicky et al. (1999). (f)
Sanders et al. (2003). Re-scaled to the distance adopted here. spectral ranges from the X-ray to the radio, and therefore a largeamount of ancillary information is available.The basic characteristics of this galaxy are compiled inTable 1. Its stellar content has been widely studied and morethan 300 stellar clusters have been detected (Cresci et al. 2005).Multi-band photometry with the WFPC2 has revealed that thosein its central region present typical masses of ∼ − × M ⊙ and are very young, with ages of ∼ −
12 Myr (e.g. Harris et al.2004). In particular, HST-NICMOS images have revealed thatthe nucleus of the galaxy is made out of two very massive ( ∼ − × M ⊙ ) super stellar clusters (SSCs), with ages of about ∼ . ∼ . ′′ ii ] λ µ memission (Labrie & Pritchet 2006).NGC 5253 presents a filamentary structure in H α (e.g.Martin 1998) associated with extended di ff use emission in X-ray which can be explained as multiple superbubbles around itsOBs associations and SSCs that are the results of the combinedaction of stellar winds and supernovae (Strickland & Stevens1999; Summers et al. 2004).The measured metallicity of this galaxy is relatively low (seeTable 1) and presents a generally uniform distribution. However,an increase in the abundance of nitrogen in the central region of ∼ − Table 2.
Observation log
Grating Spectral range Resolution t exp
Airmass(Å) (s)L682.2 6 438–7 184 13 700 5 × × On account of their irregular structure, a proper character-ization of the physical properties of H ii galaxies, necessary toexplore the interplay of mechanisms acting between gas andstars, requires high quality two-dimensional spectral informa-tion able to produce a continuous mapping of the relevant quan-tities. Such observations have traditionally been done in the op-tical and near-infrared by mapping the galaxy under study witha long-slit (e.g. V´ılchez & Iglesias-P´aramo 1998; Walsh & Roy1989). This is, however, expensive in terms of telescope timeand might be a ff ected by some technical problems such as mis-alignment of the slit or changes in the observing conditionswith time. The advent and popularization of integral field spec-troscopy (IFS) facilities, able to record simultaneously the spec-tra of an extended continuous field, overcomes these di ffi cul-ties. Nevertheless, work based on this technique devoted to thestudy of H ii galaxies is still relatively rare (e.g. Lagos et al.2009; Bordalo et al. 2009; James et al. 2009; Kehrig et al. 2008;Garc´ıa-Lorenzo et al. 2008; Izotov et al. 2006).Here, we present IFS observations of the central area ofNGC 5253 in order to study the mechanisms that govern theinteraction between the young stars and the surrounding ionizedgas. The paper is organized as follows: section 2 contains theobservational and technical details regarding the data reductionand derivation of the required observables; section 3 describesthe stellar and ionized gas structure as well as the extinction dis-tribution and the physical and kinematic properties of the ion-ized gas; section 4 discusses the evolutionary stage of the gassurrounding the stellar clusters, focusing on the two most rel-evant areas of the field of view (f.o.v.). Section 5 itemises ourresults and conclusions.
2. Observations, data reduction and line fitting
Data were obtained with the
Fibre Large Array Multi ElementSpectrograph , FLAMES (Pasquini et al. 2002) at Kueyen,Telescope Unit 2 of the 8 m VLT at ESO’s observatory onParanal, on February 10, 2007. The central region of the galaxywas observed with the ARGUS Integral Field Unit (IFU) whichhas a field of view of 11 . ′′ × . ′′ ′′ / lens.In addition, ARGUS has 15 fibers that can simultaneously ob-serve the sky and which were arranged forming a circle aroundthe IFU. The precise covered area is shown in Figure 1 whichcontains the FLAMES field of view over-plotted on an HST B,H α , I colour image.We utilized two di ff erent gratings in order to obtain informa-tion for the most important emission lines in the optical spectralrange. Data were taken under photometric conditions and see-ing ranged typically between 0 . ′′ . ′′
0. The covered spectralrange, resolving power, exposure time and airmass for each con-figuration are shown in Table 2. In addition to the science frames,continuum and ThAr arc lamps exposures as well as frames forthe spectrophotometric standard star CD-32 9927 were obtained. . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 3
Fig. 1.
False colour image in filters F435W ( B , blue), F658N( H α , magenta), and F814W ( I , green) for NGC 5253 using HST-ACS images (programme 10608, P.I.:Vacca) with the area cov-ered by our FLAMES data over-plotted as a white rectangle. Theorientation and scale for a distance of 3.8 Mpc are indicated. The basic reduction steps for the FLAMES data were performedwith a combination of the pipeline provided by ESO (version1.0) via esorex , version 2.0.2 and some IRAF routines. Firstof all we masked a bad column in the raw data using the task fixpix within IRAF. Then, each individual frame was pro-cessed using the ESO pipeline in order to perform bias subtrac-tion, spectral tracing and extraction, wavelength calibration andcorrection of fibre transmission.Uncertainties in the relative wavelength calibration were es-timated by fitting a Gaussian to three isolated lines in every spec-trum of the arc exposure. The standard deviation of the centralwavelength for a certain line gives an idea of the associated errorin that spectral range. We were able to determine the centroid ofthe lines with an uncertainty of ∼ ∼ − . The spectral resolution was very uni-form over the whole field-of-view with values of 0 . ± .
004 Åand 0 . ± .
009 Å, FWHM for the blue and red configurationrespectively, which translates into σ instru ∼ − .For the sky subtraction, we created a good signal-to-noise(S / N) spectrum by averaging the spectra of the sky fibres ineach individual frame. This sky spectrum was subsequently sub-tracted from every spectrum. In several of the sky fibres, thestrongest emission lines, namely [O iii ] λ α in the red frames, could clearly be detected. We at-tributed this e ff ect to some cross-talk from the adjacent fibers.A direct comparison of the flux in the sky and adjacent fibersshowed that this contribution was always ∼ < http: // / projects / dfs / dfs-shared / web / vlt / vlt-instrument-pipelines.html. The Image Reduction and Analysis Facility
IRAF is distributed bythe National Optical Astronomy Observatories which is operated by theassociation of Universities for Research in Astronomy, Inc. under coop-erative agreement with the National Science Foundation. ligible in terms of sky subtraction. However, in order to reducethis contamination to a minimum, we decided not to use thesefibres in the creation of the high S / N sky spectra.Regarding the flux calibration, a spectrum for the calibra-tion star was created by co-adding all the fibers of the standardstar frames. Then, a sensitivity function was determined with theIRAF tasks standard and sensfunc and science frames werecalibrated with calibrate . Afterwards, frames correspondingto each configuration were combined and cosmic rays rejectedwith the task imcombine . As a last step, the data were refor-matted into two easier-to-use data cubes, with two spatial andone spectral dimension, using the known position of the lenseswithin the array.
In order to obtain the relevant emission line information, lineprofiles were fitted using Gaussian functions. This procedurewas done in a semi-automatic way using the IDL based routineMPFITEXPR (Markwardt 2009) which o ff ers ample flexibilityin case constraints on the parameters of the fit are included, suchas lines in fixed ratio. The procedure was as follows.As a first step, we fit all the lines by a single Gaussian. TheH α + [N ii ] complex was fitted simultaneously by one Gaussianper emission line plus a flat continuum first-degree polynomialusing a common width for the three lines and fixing the sepa-ration in wavelength between the lines according to the redshiftprovided at NED and the nitrogen line ratio ( λ / λ ii ] λλ iv ] λ iii ] λ i λ iv ] λ iii ] λ β , [O iii ] λ i λ / or multiple com-ponents in their profiles for a large number of spaxels. In thosecases, multi-component fits were performed. Over the wholefield of view we compared the measured flux from performingthe fit with a single Gaussian to the fit by several componentsin the brightest emission lines (namely: H β , [O iii ] λ α ,[N ii ] λ ii ] λλ ff erences between the twosets of line fluxes ranged typically from 0% to 15%, dependingon the spaxel and the emission line, and translated into di ff er-ences in the line ratios ∼ < map and image , when referring to these. See http: // purl.com / net / mpfit. See http: // / . A. Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES Table 3.
Main reference clusters.
Name FLAMES Other namescoord.( ′′ , ′′ ) C97 (a) H04 (b)
K97 (c)
AH04 (d) . , .
5) N5253-5 1 UV3 C1 + C2 . , − .
0) N5253-4 4,8,24,25 UV1 − − . , .
0) N5253-3 3,5 − C4 + C5 (a) Calzetti et al. (1997). (b)
Harris et al. (2004). (c)
Kobulnicky et al. (1997). (d)
Alonso-Herrero et al. (2004).
3. Results
Figure 2 displays the stellar structure, as traced by a continuumclose to H α , as well as the one for the ionized gas (traced by theH α emission line). The over-plotted contours, which representthe HST-ACS images in the F659N and F814W bands convolvedwith a Gaussian to match the seeing at Paranal, show good cor-respondence between the images created from the IFS data andthe HST images (although obviously with poorer resolution forthe ground-based FLAMES data). A direct comparison of thesemaps shows how the stellar and ionized gas structure di ff ers.The continuum image displays three main peaks of emissionwhich will be used through the paper as reference. We have asso-ciated each of these peaks with one or more star clusters by directcomparison with ACS images. Table 3 compiles their positionswithin the FLAMES field of view together with the names ofthe corresponding clusters according to several reference works.These clusters trace a sequence in age as we move towards theright (south-west) in the FLAMES field of view. The clusters as-sociated with peak ∼ − ∼ −
170 Myr, Harris et al. 2004), while the stars in the pair ofclusters associated with peak α emission line reproduces the structure described byCalzetti et al. (1997) using an HST WFPC2 image. Briefly, thecentral region of NGC5253 is divided into two parts by a dustlane that crosses the galaxy along the east-west direction (from ∼ [2 . ′′ . ′′ to the Complex α emission islocated towards the north of this lane where there is giant H ii region associated with the Complex ∼ ◦ and ∼− ◦ , respectively) as well as a extension at P.A. ∼ ◦ whichcontains the Complex ∼ [-3 . ′′ . ′′
5] which could beassociated with cluster 17 in Harris et al. (2004).
Extinction was derived assuming an intrinsic Balmer emissionline ratio of H α / H β = T e =
10 000 K) and using the extinction curve of Fluks et al.(1994). Since the H α and H β emission lines are separated by ∼ Hereafter, the di ff erent quoted positions will be refereed as [ ′′ , ′′ ]and using the FLAMES f.o.v. as reference. −4−2024−202 Continuum ∆ x (arcsec) ∆ y ( a r cs e c ) N 25 pc ++ + H α∆ x (arcsec) ∆ y ( a r cs e c ) ++ + Fig. 2.
Top:
Stellar component distribution as traced by a con-tinuum map made from the average flux in the spectral range6525 − − Bottom:
Ionized gas distribu-tion as traced by the H α emission line. We have over-plotted con-tours corresponding to the HST-ACS images in the F814W ( top )and F658N ( bottom ) filters (programme 10609, P.I.: Vacca) con-volved with a Gaussian of 0 . ′′ α and 0.9 dex for the con-tinuum map. Flux units are arbitrary.a single exposure, we decided to obtain the extinction maps fromthe LR3 and LR6 exposures observed at the smallest airmass(1.009 and 1.160, respectively), thus minimizing any e ff ect dueto di ff erential atmospheric refraction.We have not included any correction for an underlying stellarpopulation. We inspected carefully each individual spectrum tolook for the presence of the stellar absorption feature in H β . Onlyin those spaxels associated with the area around the Complex ∼ ∼ . ′′ × . ′′ ∼ ∼ ∼ −
20 timeswider and with about half - one third of the flux of the emissionline. This implies an underestimation of H β emission line fluxof about 10%. For the particular area around Complex A V ∼ . A V ∼ . . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 5 −4−2024−202 +0.0+0.7 ∆ x (arcsec) ∆ y ( a r cs e c ) E(B−V) ++ +
Fig. 3.
Reddening map obtained from the hydrogen recombi-nation lines assuming H α / H β = E ( B − V ) = A V / . E ( B − V ) = A V / . E ( B − V ) = .
16 to 0 . A V ∼ . − .
13 mag.However, the larger measured extinction values are associatedwith the giant H ii region, in agreement with the H i distribu-tion (Kobulnicky & Skillman 2008). Dust in this area forms anS-shaped distribution with A V ∼ . − .
15 mag in the arms.In order to explore the relation between the extinction suf-fered by the gas and by the stellar populations, our E ( B − V )measurements were compared with colours defined ad hoc . Forthe covered spectral range, it is not possible to exactly simulateany of the existing standard filters. It is possible, however, tocreate filters relatively similar to the g ′ and R c ones. We havesimulated two set of filters.In the first case, the flux was integrated over two large wave-length ranges (465 – 495 nm and 643 – 673 nm) in order to sim-ulate broad filters. The relation between the reddening derivedfor the ionized gas and the derived colour, hereafter ( g ′ − R c ),is shown in the upper panel of Figure 4, as would be observedwith photometry. The first order polynomial fit to the data andthe Pearson correlation coe ffi cient are included on the plot. Alsoshown is the expected relation for an Im galaxy with foregroundreddening. This latter was derived for the average of two Im tem-plates (NGC 4449 and NGC 4485) from Kennicutt (1992) ap-plying a foreground screen of dust with a standard Galactic red-dening law (Cardelli et al. 1989) with R = ff erence between the expected relation for a foreground screenof dust and the measured values. On the one hand, colours aremuch redder. On the other hand, the slope of the 1-degree poly-nomial fit is much less steep than the expected one. Also, there isa very good correlation between the E ( B − V ) and our synthetic Fig. 4.
Relation between the derived reddening and a colour sim-ilar to g ′ − R c . The vertical line represent the g ′ − R c colourexpected for a Im type galaxy without any extinction while thegreen dashed line corresponds to the expected relation if gas andstars were su ff ering the same amount of extinction (see text fordetails). The 1-degree polynomial fit to all the data appear asa continuous line. The corresponding fit is shown in the upperleft corner while the Pearson’s coe ffi cient of correlation is indi-cated in the lower right corner. The simulated g ′ − R c colourshave been done by integrating the flux in the spectral ranges of4 653 − g ′ ) and 6 431 − R c ) for the upper paneland 4 600 − g ′ ) and 6 600 − R c ) for the lowerone. −4−2024−202 −0.50+0.40 ∆ x (arcsec) ∆ y ( a r cs e c ) (g−Rc) * ++ + Fig. 5.
Line free colour map. The simulated filters have been de-fined as explained in Figure 4. The position of the three mainpeaks of continuum emission are shown for reference.( g ′ − R c ) colour. All this can be attributed to the contaminationof the gas emission lines, mainly H α and H β , in our filters.In the second set, we restricted the spectral ranges for thesimulated filters to a narrower wavelength range which was freefrom the contamination of the main emission lines. The mapfor this line-free colour is displayed in Figure 5. The structure A. Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES resembles the one presented in Figure 3 (i.e. dust lane, reddercolours associated with the giant H ii region), although there aredi ff erences, that can be attributed to di ff erences in the propertiesof the stellar populations in the di ff erent clusters. The relationbetween the reddening and the corresponding ( g ′ − R c ) is shownin the lower panel of Figure 4. This time colours are more simi-lar to what is expected for a given stellar population su ff ering acertain amount of extinction. However, for a given colour, starsdo not reach the expected reddening if gas and stars were suf-fering the same extinction (i.e. data points are below the greenline). The ratio between the slopes indicates that extinction inthe stars is a factor 0.33 lower than the one for the ionized gas.This is similar to what Calzetti et al. (1997) found using HSTimages who estimated that the extinction su ff ered by the stars isa factor 0.5 lower than for the ionized gas and can be explainedif the dust has a larger covering factor for the ionized gas thanfor the stars (Calzetti et al. 1994).In general, our E ( B − V ) measurements agree with pre-vious ones using the same emission lines in specific areas(e.g. Gonz´alez-Riestra et al. 1987; L´opez-S´anchez et al. 2007)or with poorer spatial resolution (Walsh & Roy 1989). Howeverthere are discrepancies when comparing with the estimation ofthe extinction at other wavelengths. In particular, the peak ofextinction ( A V = . ff set by ∼ . ′′ F W image in Figure 3 show thegood correspondence between our maximum of extinction andC2. Measurements in the near and mid-infrared suggest ex-tinctions of A V ∼
17 mag for this cluster (Turner et al. 2003;Alonso-Herrero et al. 2004; Mart´ın-Hern´andez et al. 2005). Thediscrepancy between these two values indicates that a fore-ground screen model is not the appropriate one to explain thedistribution of the dust in the giant H ii region. Electron density ( N e ) can be determined from the[S ii ] λ / [S ii ] λ iv ] λ / [Ar iv ] λ integrated spectrum .Electron densities were determined assuming an electrontemperature of 11 650 K, the average of the values given inL´opez-S´anchez et al. (2007), and using the task temden , basedon the fivel program (Shaw & Dufour 1995) included in theIRAF package nebular . Derived values of N e for the two lineratios were 180 cm − and 4520 cm − , respectively. Di ff erencesbetween the electron densities derived from the argon andsulphur lines are usually found in ionized gaseous nebulae (seeWang et al. 2004) and are understood in terms of the ionizationstructure of the nebulae under study: [Ar iv ] lines normallycome from inner regions of higher ionization degree than [S ii ]lines. Typically, for giant Galactic and extragalactic H ii regions,derived N e from these two line ratios di ff er in a factor of ∼ integrated spectrum of NGC 5253( ∼ ii region is taken intoaccount (i.e. the area of ∼
90 spaxels where the argon lines aredetected) the di ff erence between the densities derived from the −4−2024−202 +0.95+1.45 ∆ x (arcsec) ∆ y ( a r cs e c ) [SII] λ λ ++ + −4−2024−202 +0.80+1.40 ∆ x (arcsec) ∆ y ( a r cs e c ) [ArIV] λ λ ++ + Fig. 6.
Maps for the line ratios sensitive to the electron density.The position of the three main peaks of continuum emissionare shown as crosses for reference. The displayed ranges in theline ratios imply electron densities of < − and 190-8750 cm − for the sulphur and argon line ratio, respectively.argon and the sulphur lines ( ∼
10, see typical values for thedensities below) is more similar to those found for other H ii regions.Maps for both ratios are shown in Figure 6. According tothe sulphur line ratio - detected over the whole field - densitiesrange from very low values, of the order of the low density limit,in a region of about 5 ′′ × ′′ in the upper right corner of thefield just above the Complex − at the peak ofthe emission in the cluster associated with the H ii region, witha mean (median) over the field of view of ∼
130 (90) cm − .The rest of the H ii regions still present high densities (of about400 cm − as a whole, 480 cm − in the H ii -2, (i.e. the up-per area of the giant H ii region, Kobulnicky et al. 1997). Thisagrees well with the value estimated from long-slit measure-ments (L´opez-S´anchez et al. 2007). The tail and the region as-sociated with the cluster UV-1 present intermediate values (ofabout 200 cm − ).The argon line ratio is used to sample the densest regions.The map for this ratio was somewhat noisier and allowed an es-timation of the electron density only in the giant H ii region. Thedensities derived from this line ratio are comparatively higher,with a mean (median) of 3400 (3150) cm − . As it happened inthe case of the extinction, the peak of electron density accordingto this line ratio is o ff set by ∼ . ′′ − . ′′ / N ratio spectra by co-addition of 3 × + C2, and the re- . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 7 gions H ii -2, H ii -1 and UV-1 of Kobulnicky et al. (1997)). Thelargest values are measured around the core (i.e. C1 + C2) wherethe [Ar IV] electron density can be as high as 6200 cm − . Aswe move further away from this region, the measured electrondensity becomes lower. Thus, H ii -2 presents similar, althoughslightly lower, densities ( ∼ − ), followed by H ii -1 with ∼ − and UV1 with ∼ − . These values agree,within the errors, with those reported in L´opez-S´anchez et al.(2007) for similar apertures.An interesting point arises when the di ff erent density val-ues derived for the integrated spectrum and for each individualspaxel / aperture are compared (180 cm − , and up to 790 cm − , re-spectively when using the sulphur line ratio). The covered f.o.v.( ∼
210 pc ×
135 pc) is comparable to the linear scales that one canresolve from the ground at distances of ∼
40 Mpc (or z ∼ . ff ects can cause im-portant underestimation of the electron density in the H ii regionsin starbursts at such distances, or further away. The ionization structure of the interstellar medium can be stud-ied by means of diagnostic diagrams. Di ff erent areas of a givendiagram are explained by di ff erent ionization mechanisms. Inthe optical spectral range, the most widely used are probablythose proposed by Baldwin et al. (1981) and later reviewed byVeilleux & Osterbrock (1987), the so-called BPT diagrams. InFigure 7 the maps for the three available line ratios involved inthese diagrams - namely [N ii ] λ / H α , [S ii ] λλ / H α ,[O iii ] λ / H β - are shown on a logarithmic scale. This figureshows that the ionization structure in the central region of thisgalaxy is complex. Not only do the line ratios not show a uni-form distribution, but the structure changes depending on theparticular line ratio.Both the [O iii ] λ / H β and the [S ii ] λλ / H α lineratios display a gradient away from the peak of emission atComplex ii ] λλ / H α ([O iii ] λ / H β ) ra-tio is smallest (largest) at Complex ∼ [ − . ′′ , − . ′′ α emission. This is coinci-dent with the structure presented in (Calzetti et al. 2004).The [N ii ] λ / H α line ratio, however, display a di ff er-ent structure. While in the right half of the FLAMES fieldof view, the behavior is quite similar to the one observed forthe [S ii ] λλ / H α ratio (i.e. values relatively high, lo-cal minimum at ∼ [ − . ′′ , − . ′′ ii region, displays a completely di ff erent pattern. Thelowest values are associated with Complex ii ] λ / H α lineratios are highest at ∼ [4 . ′′ , . ′′ ii region-like ioniza-tion from ionization by other mechanisms according to sev-eral authors (Veilleux & Osterbrock 1987; Kewley et al. 2001;Kau ff mann et al. 2003; Stasi´nska et al. 2006). We also showthe predictions for models of photo-ionization caused by stars(Dopita et al. 2006) that take into account the e ff ect of the stellarwinds on the dynamical evolution of the region. In these mod- −4−2024−202 −1.50−0.75 ∆ x (arcsec) ∆ y ( a r cs e c ) log([NII] λ α ) ++ + −4−2024−202 −1.40−0.50 ∆ x (arcsec) ∆ y ( a r cs e c ) log([SII] λλ α ) ++ + −4−2024−202 +0.44+0.90 ∆ x (arcsec) ∆ y ( a r cs e c ) log([OIII] λ β ) ++ + Fig. 7.
Emission line ratio maps.
Up: log ([N ii ] λ / H α . Middle: log ([S ii ] λλ / H α ). Bottom: log([O iii ] λ / H β ). The position of the three main peaks ofcontinuum emission are shown for reference.els, the ionization parameter is replaced by a new variable R thatdepends on the mass of the ionizing cluster and the pressure ofthe interstellar medium ( R = (M Cl / M ⊙ ) / (P o / k), with P o / k mea-sured in cm K). Also, the predictions for shocks models for aLMC metallicity are included. Given the relatively low metal-licity of NGC 5253, these are the most appropriate ones. Theywere calculated assuming a N e = − and cover an amplerange of magnetic parameters, B , and shock velocities, v s , (seeAllen et al. 2008, for details).As demonstrated for the electron density, these diagrams il-lustrate very clearly how resolution e ff ects can influence themeasured line ratios. Values derived for individual spaxelscover a range of ∼ ii ] λ / H α ,[S ii ] λλ / H α , and [O iii ] λ / H β line ratios respec-tively, with mean values similar to the integrated values ( − . .
92, and 0.74). This is particularly relevant when interpreting
A. Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES
Fig. 8.
Position of the individual spaxels in NGC 5253 in thediagnostic diagrams proposed by Veilleux & Osterbrock (1987).Data points above and below the 3- σ level of the first-degreepolynomial fit in the [N ii ] λ / H α vs. [S ii ] λλ / H α diagram (see text for details) have been represented with blue di-amonds and yellow crosses respectively. The solid curves showthe empirical borders found by Veilleux & Osterbrock betweenionization caused by di ff erent mechanisms, while the dottedlines show the theoretical borders proposed by Kewley et al.(2001) to delimit the area where the line ratios can be explainedby star formation. Black dashed and dot-dashed lines show therevised borders by Kau ff mann et al. (2003) and Stasi´nska et al.(2006), respectively. Solid horizontal lines show the border be-tween classical Seyfert galaxies (above the line) and LINERs(below the line), at [O iii ] λ / H β =
3. Long-dashed lines in or-ange represent the predictions from models of photo-ionizationby stars presented in Dopita et al. (2006) for R = − N e = − .Lines of constant magnetic parameter (blue) cover a range of B = . − µ G, while lines of constant shock velocity (red)cover a range of v s = −
500 km s − . Green asterisks markthe derived values for the integrated spectrum of the whole IFUfield.the ionization mechanisms in galaxies at larger distances wherethe spectrum can sample a region with a range in ionization prop-erties. This loss of spatial resolution thus ’smears’ the determi-nation of the ionization mechanism by a set of line ratios. Evenif this given set of line ratios is typical of photoionization causedby stars, it is not possible to exclude some contribution due toother mechanisms at scales unresolved by the particular obser-vations. Regarding the individual measurements, although all line ra-tios are within the typical values expected for an H ii region-likeionization, two di ff erences between these diagrams arise. Thefirst one is that the diagram involving the [S ii ] λλ / H α line ratio indicates a somewhat higher ionization degree than theone involving the [N ii ] λ / H α line ratio. That is: values forthe diagram involving the [S ii ] λλ / H α line ratio areat the limit of what can be explained by pure photo-ionizationin an H ii region according to the Kewley et al. (2001) theoreti-cal borders. On the contrary, most of the data points in the di-agram involving the [N ii ] λ / H α are clearly in the area as-sociated to photoionization caused by stars. A comparison withthe predictions of the models for metallicities similar to the oneof NGC 5253 shows how the measured line ratios present in-termediate values between those predicted by ionization causedby shocks and those by pure stellar photoionization. That this isexactly what one would expect if shocks caused by the mechan-ical input from stellar winds or supernovae within the starburstwere contributing to the observed spectra. Also, this comparisonsupports previous studies that show how models of photoioniza-tion caused by stars underpredict the [O iii ] λ / H β line ratios,specially in the low-metallicity cases (Brinchmann et al. 2008;Dopita et al. 2006). Fig. 9. [S ii ] λλ / H α vs. [N ii ] λ / H α diagram. Thered box includes the data points utilized to determine thefirst-degree polynomial fit (dark blue continuous line) betweenthe [S ii ] λλ / H α and [N ii ] λ / H α line ratios. Datapoints above and below the 3- σ level (light blue continuousline) have been represented with blue diamonds and yellowcrosses respectively. The expected line ratios for H ii regions for Z = . Z ⊙ and Z = . Z ⊙ with ages in the range of 2-6 Myr and R = ii ] λλ / H α vs. [O iii ] λ / H β diagram form a se-quence, data points in the [N ii ] λ / H α vs. [O iii ] λ / H β diagram are distributed in two groups: a sequence similar tothe one in the [S ii ] λλ / H α vs. [O iii ] λ / H β dia-gram and a cloud of data points above that sequence with larger[N ii ] λ / H α . This result can be interpreted either by localvariations in the relative abundances or by changes in the ion-ization parameter. Here we will explore the first option, which . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 9 −4−2024−202 H α∆ x (arcsec) ∆ y ( a r cs e c ) ++ + Nuc HII−2WR1 WR2 WR3 WR4WR5HeII 1HeII 2 HeII 3 HeII 4
Fig. 10.
Spaxels with [N ii ] λ / H α line ratio above the 3- σ level of the fit presented in Figure 9 on top of our H α emissionline map. The size of the white circles is proportional to the ex-cess in the [N ii ] λ / H α line ratio. Dashed and dotted linesdelimit the apertures utilized to extract the spectra with Wolf-Rayet and nebular He ii features displayed in Figures 11 and 12(see sections 3.5 and 3.6). The position of the three peaks of con-tinuum emission in the map are shown by crosses for reference.is the most accepted explanation (e.g. L´opez-S´anchez et al.2007, and references therein) and is supported by the rela-tively constant ionization parameter found in specific areasvia long-slit (log( U ) ∼ -3, Kobulnicky et al. 1997). Long-slitmeasurements in specific areas of this field have shown outhow this galaxy present some regions with an over-abundanceof nitrogen (e.g. Walsh & Roy 1989; Kobulnicky et al. 1997).For our measured line ratios and using expression (22)in P´erez-Montero & Contini (2009), we measure a range inlog( N / O ) of − .
70 to − .
46. Here we will assume that thisover-abundance is the cause of our excess in the [N ii ] λ / H α line ratio and will use this excess to precisely delimit the areapresenting this over-abundance. To this aim, we placed the in-formation of each of the spaxels in the [S ii ] λλ / H α vs. [N ii ] λ / H α diagram, which better separates the twodi ff erent groups described above. This is presented in Figure9. We have assumed that in the so-called un-polluted ar-eas, the [N ii ] λ / H α and [S ii ] λλ / H α follow alinear relation. This is a reasonable assumption since the[N ii ] λ / [S ii ] λλ standard relation was deter-mined by fitting a first-degree polynomial to the data points with[S ii ] λλ / H α> − . ii ] λ / H α line ratio was in ex-cess of more than 3- σ from the relation determined by this fit,have been identified as having an [N II] / H α excess, and are iden-tified by diamonds in Figure 9. As can be seen from this figure,there are a number of spaxels where this excess is much abovethe standard relation.The data points thus identified with [N ii ] λ / H α excessare shown as a map in Figure 10 where the location and magni-tude of the [N ii ] λ / H α excess is indicated by white circles,whose size is proportional to the size of the [N ii ] λ / H α ex-cess. This figure can be interpreted as a snapshot in the pollutionprocess of the interstellar medium by the SSCs in the central areaof NGC 5253. The pollution is a ff ecting almost the whole giantH ii region. The largest values are found at ∼ . ′′ ii -1 and H ii -2, whilethe N / O ratio in UV-1 (see Table 3) was typical for metal-poorgalaxies.
Wolf-Rayet (W-R) stars are very bright objects with strong broademission lines in their spectra. They are classified as WN (thosewith strong lines of helium and nitrogen) and WC (those withstrong lines of helium, carbon and oxygen) and are understoodas the result of the evolution of massive O stars. As they evolve,they loose a significant amount of their mass via stellar windsshowing the products of the CNO-burning first - identified asWN stars - and the He-burning afterwards - as WC stars (Conti1976). The presence of W-R stars can be recognized via the W-R bumps around λ blue bump , characteristic ofWN stars) and λ red bump , characteristic of WCstars, but not covered by the present data).Schaerer et al. (1997, 1999) carried out a thorough searchand characterization of the W-R population in NGC 5253. Theydetected W-R features (both WN and WC) at the peak of emis-sion in the optical (our Complex / O ratio, incomparison with similar non-W-R galaxies (Brinchmann et al.2008). Other suggested possibilities to cause the enrichment innitrogen include planetary nebulae, O star winds, He-deficientW-R star winds, and luminous blue variables (Kobulnicky et al.1997).Here, we characterize the W-R population in NGC 5253 andexplore the hypothesis of W-R stars as the cause of the nitro-gen enhancement by using the 2D spectral information providedby the present data. In the previous section we have delimitedvery precisely the area that presents nitrogen enhancement. In asame manner, it is possible to look for and localize the areas thatpresent W-R emission. Note that due to the continuous samplingof the present data this can be done in a completely unbiasedway.We visually inspected each spectrum looking for the moreprominent W-R features in the blue bump (i.e. N iii λ ii λ ii line as-sociated with the north-west and south-east extensions (i.e. H ii -2 and W-R 2, respectively). In addition, there are three moreareas which present W-R features, called W-R 1, W-R 4 and W-R 5, relatively far ( ∼ −
83 pc) from the main area of activity.
Fig. 11.
Spectra showing Wolf-Rayet features. The first spectrum has been extracted from a low surface brightness area in the upperright corner of the FLAMES field of view and is presented here as reference. The positions of the nebular emission lines have beenindicated with blue ticks and labels, while those corresponding to Wolf-Rayet features appear in red. The position and extent of theregions are indicated on Figure 10.Interestingly, two of these regions (W-R 4 and W-R 5) present anarrow nebular He ii on top of the broad W-R feature.The short phase of W-R stars during star evolution makestheir detection a very precise method for estimating the age ofa given stellar population. According to Leitherer et al. (1999),typically an instantaneous starburst shows these features at agesof ∼ − Z = . − . < α equivalent width. We esti- mated the ages by means of two indicators: the ratio betweenthe number of W-R and O stars; and the H β equivalent width.The ratio between the number of W-R and O stars was estimatedfrom F(bb) / F(H β ), where F(bb) and F(H β ) are the flux in the bluebump (measured with splot ) and in H β respectively and usingthe relation proposed by Schlegel et al. (1998). Uncertainties arelarge, mainly due to the di ffi culty to define the continuum andto avoid the contamination of the nebular emission lines whenmeasuring F(bb) but indicates a range in log(WR / (WR + O)) of ∼ − . − . ii region. TheH β equivalent widths are extremely high, consistent again withthe expected youth of the stellar population. Predicted ages fromthese two age tracers are reported in Table 4, columns 5 and 7.They were estimated by using STARBURST99 (Leitherer et al.1999), assuming an instantaneous burst of Z = . . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 11 Table 4.
Ages of the clusters associated with the W-R regions according to di ff erent indicators. Harris et al. (2004) This workRegion Clusters Age log(WR / (WR + O)) Age WR / O EW(H β ) Age H β Number (Myr) (Myr) (Å) (Myr)Nucleus 1 3 -1.99 2.9 245 2.7H ii -2 20 3 -1.93 2.9 320 2.4W-R 1 13 4 -1.34 3.2 / / / / per mass limit of M up =
100 M ⊙ and a Salpeter-type Initial MassFunction. The two age tracers give consistent age predictionsand in agreement with those reported in Harris et al. (2004).The distribution of the W-R features, in an area of about100 pc ×
100 pc, much larger than the one polluted with nitro-gen, suggests that all the detected W-R stars are not, in general,the cause of this pollution. Since the N-enrichment appears tobe associated with the pair of clusters in the core and, given theposition of maximum, most probably with the obscured SSC C2,the best W-R star candidates to be the cause of this enrichmentare those corresponding to our
Nuc aperture, and perhaps alsothe H ii -2 and W-R 2 regions. ii and heliumabundance The hypothesis that the W-R population is the cause of the ni-trogen enrichment in NGC 5253 requires an enhancement ofthe helium abundance too (e.g. Schaerer 1996). This is nicelyillustrated in Kobulnicky et al. (1997) where di ff erent linear re-lations between the nitrogen and helium abundances (N / H andHe / H) are presented according to di ff erent scenarios of nitrogenenrichment (W-Rs, PNe, etc.). The only scenario able to explainan extra quantity of nitrogen in the ISM without any extra he-lium counterpart would be the one where this nitrogen is causedduring the late O-star wind phase.As in previous sections, we can measure at each spaxelthe total helium abundance and compare it with that for nitro-gen. Since lines like [O ii ] λλ ii -1 and H ii -2 (N / H ∼ . × − ) to estimate how much helium would beneeded in the enriched areas, if the extra nitrogen were causedby W-Rs (i.e. He / H ∼ . / H ∼ . × − ) which re-quires He / H ∼ .
09. Helium abundance can be determined as: He / H = icf × ( He + / H + + He ++ / H + ) (1)where icf is a correction factor due to the presence of neu-tral helium. We assumed ic f ∼ .
0, which is consistent withthe predictions of photoionization models for our measured[O iii ] λ / H β line ratios (Holovatyy & Melekh 2002). Sincethe He i λ / N, for the purpose of this work, we de-termined y + = He + / H + from the He i λ / H α line ratio usingthe expression y + = . t . (2 .
87 He i λ / H α ) (2) −4−2024−202 −2.50−1.80 ∆ x (arcsec) ∆ y ( a r cs e c ) log(HeI λ α ) ++ + Fig. 13. log(He i λ / H α ) line ratio map. The position of thethree peaks in the continuum map are shown for reference.where t is the electron temperature in units of 10 K (Pagel et al.1992). As in section 3.3, we assumed T e =
11 650 K. Figure 13shows the 2D structure of the He i λ / H α line ratio. It is rel-atively uniform with the exception of some spaxels in the upperright corner, close to Complex iii ] λ / H β line ratio would be the only region where one can expect a sub-stantial contribution of neutral helium.Figure 14 presents the derived He + abundances vs. the[O iii ] λ / H β line ratio. With the exception of the data cor-responding to the right upper corner of the FLAMES f.o.v.,He + / H + range between 0.075 and 0.090, being higher in thehigher excitation zones (i.e. the giant H ii region). These val-ues are in agreement with previous measurements of He + / H + in specific areas (Pagel et al. 1992; Kobulnicky et al. 1997;Walsh & Roy 1989). They are consistent with a scenario with-out extra N-enrichment and still far, by a factor ∼ . − .
7, fromthe required ∼ total helium abundance in the W-R scenario,in particular in the areas enriched with nitrogen.What about the He ++ / H + , whose abundance can be deter-mined via the nebular He ii λ ii nebular line in each individual spectrum. Those thatpresented a spatial continuity were taken to define an area andwere co-added before extracting. The selected areas are markedin Figure 10 with dotted lines. The co-added and extracted spec-tra of each individual region appear in Figure 12. In addition tothese regions, as mentioned in section 3.5, nebular He ii in theW-R 4 and W-R 5 has also been detected. Fig. 12.
Spectra showing nebular He ii , but no W-R features. The positions of the nebular emission lines have been indicated withblue ticks and labels, while those corresponding to Wolf-Rayet features appear in red. Fig. 14. He + abundance vs. [O iii ] λ / H β for each individualspaxel. Fig. 15. He ++ abundance vs. [O iii ] λ / H β for each of theextracted spectra with He ii λ ffi cult to reconcilewith a scenario where this enhancement, and the existence ofHe ++ , share a common origin. Moreover, for the purpose of thiswork, we estimated the He ii abundances in this areas using: y ++ = . t . (He ii λ / H β ) (3)from Pagel et al. (1992). Derived values for the individual re-gions are shown Figure 15 as a function of the [O iii ] λ / H β line ratio. They range between 0.0001 and 0.0005. Althoughuncertainties are large, up to 0.0006, due to the weakness ofthe He ii λ ∼ . − .
050 required to bring the he-lium abundance up to ∼ ++ based on optical observations. Thus thepresent data support the scenario suggested by Kobulnicky et al.(1997) where the N-enrichment should arise during the late O-star wind phase. In view of the extra-nitrogen distribution andthe extinction map, the only place where these larger quanti-ties of He ++ could be found (if they existed) is in Complex ii (7-10) at 21 891 Å (Hora et al. 1999).Since the nebular He ii is not associated with the area show-ing N-enhancement, there is still the open question as to its ori-gin. Garnett et al. (1991) explored the di ff erent mechanisms ca-pable of producing this emission in extragalactic H ii regions.The first suggestion is photoionization by fast shocks. However,we have seen in section 3.4, that shocks do not appear to play adominant role in the central parts of NGC 5253. Moreover, themeasured log(He II λ / H β ) are ∼ − . − .
1, much lowerthan those predicted by shocks models with N e = − andLMC abundances ( ∼ − . − .
4, Allen et al. 2008).Another possibility discussed by Garnett et al. (1991) is hot( T ∼ <
70 000 K) stellar ionizing continua. This looks like a plau- . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 13 sible explanation for those cases where we had detected nebularHe ii λ blue bump (WR 4 and WR 5).The last option would be photoionization caused by X-rays.The only point sources detected by Summers et al. (2004) thatfall in our f.o.v. are sources 17, 18, and 19. This last source ap-pears to be associated with the Complex ii was detected for thisregion). Sources 18 and 17 could, however, be associated withHe ii -1 and He ii -4 detections, respectively. In particular, the lat-ter region coincides with the secondary peak of emission in theH α image and is associated with cluster 17 of the sample cata-logued by Harris et al. (2004). No satisfactory explanation wasfound for the cause of the ionization at He ii -2 and He ii -3. Slit observations in specific regions of this galaxy have demon-strated that the kinematics of the ionized gas is rather com-plex, with line profiles revealing asymmetric wings (e.g.Martin & Kennicutt 1995; L´opez-S´anchez et al. 2007). Theseobservations usually include the bright core of the galaxy.However, they might be biased since typically the long slits onlysample specific regions selected by the particular slit placement.The present data permit a 2D spatially resolved analysis of thekinematics of the ionized gas in the central area of the galaxyto be performed, thus overcoming this drawback. We based ouranalysis on the strongest emission lines (i.e. mainly H α , but alsoH β , [O iii ] λ ii ] λ ii ] λλ / N permits the line profiles to be fitted with a high degreeof accuracy.In the following, we will present the results derived fromH α . Similar results were obtained from the H β and [O iii ] λ ff erences of | ∆ v |∼ < − in most casesand always between -5 and 5 km s − . Results for the only emis-sion lines with remarkable di ff erences in the velocity maps (i.e.[N ii ] λ ii ] λλ ff erent profiles for the main emis-sion lines are shown in Figure 16. Lines are ordered by wave-length from bluer (lower) to red (upper) within each panel. Thezero point in the abscissa axis corresponds to the measured sys-temic velocity which is defined as the average of the velocitiesderived from the main emission lines for the peak of the con-tinuum emission. We performed an independent fit for each ofthe brightest emission lines using MPFITEXPR (see section 2.3for details). In general, the line profiles of the individual spec-tra cannot be properly reproduced by a single component. Alarge percentage of them needed two (and even three) indepen-dent components to reproduce the observed profile reasonablywell. We followed the approach of keeping the analysis as sim-ple as possible. Thus in those cases where both fits - the onewith one component and the one with several components - re-produced equally well the line profile, we gave preference tothe fit with one component. Examples of these fits are shown inFigure 17, which contains the H α emission line together withthe total fit and the individual components overplot for the spax-els shown in Figure 16. Note how the fits for the spaxels in thecentral area of our f.o.v. present relatively larger residuals. Thesecan be attributed to a low surface brightness broad extra compo-nent which would be the subject of a future work. Central wave-lengths were translated into heliocentric velocity taking into ac-count the radial velocity induced by the Earth’s motion at thetime of the observation which was evaluated using the IRAFtask rvcorrect . Velocity dispersions were obtained from themeasured FWHM after correcting for the instrumental width and thermal motions. The width of the thermal profile was derivedassuming T e =
11 650 K which translates into a σ ther = √ kT e / m H of ∼
11 km s − for the hydrogen lines. The measured systemicvelocity was 392 km s − . This is slightly lower than the one mea-sured from neutral hydrogen (407 km s − , Koribalski et al. 2004)according to NED.In Figure 18, we present the velocity fields for the three fit-ted components derived from the H α emission line. We also in-cluded the velocity dispersion map for our broadest component.Corrected velocity dispersions for the two narrow componentswere, in general, subsonic and will not be shown here. The onlyexception would be an area at ∼ [4 . ′′ , − . ′′ ∼ [4 . ′′ , − . ′′
0] as if itwere the result of a strong blending of these two components.However, we were not able to properly deblend these two com-ponents by means of our line fitting technique.From Figure 18, it is clear that the movements of the ion-ized gas are far from simple rotation. For the discussion we willseparate the emitting area into the zone corresponding to the gi-ant H ii region and the rest. The zone of the giant H ii region,occupying roughly the left part of the FLAMES field of view,shows in the upper part, line profiles that can be explained bytwo components while in some spaxels of the lower part a thirdcomponent was required. The area of the giant H ii region itself,which occupies an area of ∼
120 pc ×
60 pc, requires up to threecomponents to properly reproduce the line profiles. They werenamed C1, C2 and C3, according to their relative fluxes. Thefirst component (i.e. C1) accounts for the ∼ −
68% of the fluxin H α , depending on the considered spaxel. It is relatively nar-row and constant, with a ∆ v ∼
10 km s − over a distance of ∼ ′′ ( ∼
110 pc) with slightly bluer velocities in the spaxels associatedwith the edge of the upper and lower extensions.The second component (i.e. C2) accounts for the ∼ − α . It is symmetric with respect of an axis that goesthrough the Complex ∆ v ∼
70 km s − over ∼ . ′′ ∼
86 pc) and is relativelybroad ( σ ∼ −
25 km s − ). Low surface brightness broad com-ponents have been reported in starburst galaxies using a slit sincemore than a decade and have been the subject of several the-oretical (e.g. Tenorio-Tagle et al. 1997) and observational (e.g.Casta˜neda et al. 1990; Gonz´alez-Delgado et al. 1994) studies.They usually represent a small fraction ( ∼ α flux and have widths of σ ∼
700 km s − . Recently, 2D spec-troscopic analysis of very nearby starbursts have shown how lo-cally , the line width is somewhat smaller ( σ ∼ − , seeWestmoquette et al. 2009, and references therein). This is un-derstood in the context of the so-called Turbulent Mixing Layers (e.g. Slavin et al. 1993). However, the high surface brightness ofC2, together with the symmetry in the velocity field and its lowwidths made us to explore an alternative explanation for it (seesection 4.1).The third component (i.e. C3) is present in a small area ofthe field of about 1 . ′′ . ′′ . ′′ ∼
50 km s − towards the blue. This third componentappears in a location about 1 . ′′ We pointed out before that velocity maps derived from the[N ii ] λ ii ] λλ ff erences with respect to the one obtained fromH α . This is the case for the giant H ii region. Figure 19 containsthe measured velocity di ff erences for the two brighter compo-nents (C1 and C2) in the subset of the field corresponding to thegiant H ii region while Figure 20 presents examples of the in-dependent fits for H α , [N ii ] λ ii ] λ ff erencesexist for both, the narrow (i.e. C1) and broad (i.e. C2), compo-nents in [N ii ] λ ff erent directions (P.A. ∼ ◦ and ∼ ◦ for C1 and C2 respectively). Also, the range of velocity di ff er-ences, ( v H α − v [NII] ) is larger for C2 than for C1 ( ∼
70 km s − and ∼
30 km s − , respectively). As illustrated in the right hand mapof Figure 19, the broad component for the [S ii ] λλ ff erences with similar range( v H α − v [SII] ∼
60 km s − ), orientation and sign as in the case ofthe [N ii ] λ ff erences were found. Measured di ff erences in C3, with amean and standard deviation of − ± − and 5 ± − for v H α − v [NII] and v H α − v [SII] respectively, do not appear to besignificant. However, since C3 was only detected in seven spax-els (see Figure 18, bottom left panel) this result has to be treatedwith caution.Similar o ff sets has been detected in galactic H ii regions likeOrion (Garc´ıa-D´ıaz et al. 2008), but to our knowledge, this isthe first time that maps with such o ff sets in velocity for dif-ferent emission lines in starbursts are presented. This can par-tially be caused by the fact that 2D-kinematic analysis of star-bursts, from dwarfs (e.g. Garc´ıa-Lorenzo et al. 2008) to moreextreme events like LIRGs (e.g. Alonso-Herrero et al. 2009), areusually based on fitting techniques that impose restrictions be-tween the H α and [N ii ] λ ff sets. An example of work wherethe main emission lines are fitted independently is presented byWestmoquette et al. (2007). However, since they only analyzedthe kinematic for H α , it is not possible to assess if they founddi ff erent kinematics for the other emission lines. There are how-ever, some works that o ff er examples of o ff sets of this kind us-ing a slit. In particular (L´opez-S´anchez et al. 2007) report alsoan o ff set between the [N ii ] λ α emission line of ∼
10 km s − , similar to what we have measured for C1 in theupper part of our f.o.v.The second region of interest is located in the right (south-west) part of the FLAMES field of view. Emission there showsnarrow lines with velocity dispersion dominated by the thermalwidth. In some areas (the north-east corner in Figure 18), twonarrow lines were needed to better reproduce the line profile.The primary component (i.e. C1) shows a symmetric velocitypattern with respect to the twin clusters associated with the peakof emission ∆ v ∼
40 km − over about 4 . ′′ ∼
75 pc). The secondary component traces a shellblue-shifted ∼
40 km s − in the western corner (see Figure 18,upper right panel). Note that C2, although relatively narrow, isa bit broader than the thermal width (Figure 18, bottom rightpanel). This component accounts for ∼ −
50% of the H α fluxin this area. No significant di ff erences in the velocity fields andthe velocity dispersion maps for the main emission lines havebeen found in this area.
4. Discussion ii region The most interesting area of NGC 5253 in the present data cov-ers the left (north-east) part of the FLAMES field of view. Inprevious sections we have seen that this area is occupied by agiant H ii region which: i) harbors two very massive and youngSSCs at its centre (i.e. Complex ii ] λ / H α line ratiowith respect to [S ii ] λλ / H α which, if interpreted as N-enrichment, indicates an outward gradient of extra nitrogen froma point at ∼ . ′′ − . ′′ Fig. 21.
Sketch showing the di ff erent elements associated withthe area of the giant H ii region. The two massive SSCs (starsin the circles) expel material (red and blue half moons) whichencounter quiescent gas (grey shell). Regions of dust extinctionare represented by the two wavy sheets.In Figure 21, we sketch a plausible scenario compatible withall these results. Here, the broad component (C2) would tracean outflow created by the two SSCs at Complex ff erent grades of grey in Figure 21 in the shellrepresent the di ff erent densities observed in the upper and lowerpart of the FLAMES f.o.v., while two wavy sheets in two gradesof grey have been used to represent the di ff erences in extinctionbetween these two halves. Due to this extinction distribution, inthe upper (i.e. north-western) half of the H ii region, the observercannot see the further part of the shell, while in the lower (i.e.south-eastern) half both parts are visible and are detected as asingle broader component when approaching the vertex of theoval.As in section 3.3, we formed [S ii ] λ / [S ii ] λ × ii region. Although the standard deviations are large( ∼ . . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 15 nent presented similar line ratios in both areas ( ∼ .
07 im-plying densities of ∼
470 cm − ), the narrow component pre-sented somewhat lower line ratios in the upper part than inthe lower one ( ∼ ∼ ∼
390 cm − and ∼
180 cm − , respectively. Also, wecreated (noisier) [N ii ] λ / H α , [S ii ] λλ / H α , and[O iii ] λ / H β line maps and compared the relation between[S ii ] λλ / H α and [N ii ] λ / H α for the individualcomponents. The three fitted components present extra nitrogenin an area coinciding with the one derived from one-gaussian fit-ting. This di ff ers from the findings for Mrk 996, a galaxy withseveral kinematically distinct components where only the broadone presented N-enrichment, with an abundance ∼
20 timeslarger than the one for the narrow component (see James et al.2009). Still, in NGC 5253, the N-enrichment in the broad com-ponent is larger than in the narrow one by a factor of ∼ iii ] λ / H β maps are relatively similar for both compo-nents (not shown), those associated to [N ii ] λ / H α and spe-cially to (the more shock sensitive) [S ii ] λλ / H α lineratio display a relatively di ff erent ionization degree with largervalues for C1 than for C2. This is also consistent with the pre-sented scenario since a larger contribution due to shocks is ex-pected in the area where the outflowing material encounters thepre-existent gas.An interesting result of the previous section was the o ff setsderived for the velocities of the di ff erent species, in particularnitrogen and sulphur. To our knowledge, this is the first time thatmaps showing this kind of o ff sets are reported in an starburstgalaxy. Similar phenomena have already been reported in muchcloser regions of star formation. For example, observations inthe Galactic Orion Nebula, a much less extreme event in termsof star formation, show how H α and [O iii ] λ ii ] λ ii ] λλ ∼ − (Garc´ıa-D´ıaz et al. 2008), an order of magnitudesmaller than the shifts found for NGC 5253. Also, self-consistentdynamic models of steady ionization fronts point towards thedetection of such di ff erences (Henney et al. 2005). In the contextof the scenario sketched in Figure 21, the o ff sets in C2 would fitif [N ii ] λ ii ] λλ ∼
60 pc took place over only ∼ The rightmost (south-west) part of the FLAMES f.o.v. presentsa di ff erent picture. We have seen that this region: i) is associ-ated with two relatively old ( ∼
70 and ∼
110 Myr) and massive(3 and 7 × M ⊙ ) clusters (Harris et al. 2004); ii) presents mod-erate levels of extinction, being higher in the lower part of theFLAMES field of view; iii) has very low N e and lower than100 cm − in the upper corner; iv) the H α surface brightness is C3 is not considered here, since the so-called map would be asso-ciated to ≤ Table 5.
Integrated line ratios for the rightmost (south-west) partof the FLAMES f.o.v.
Component log([O iii ] / H β ) log([N ii ] / H α ) log([S ii ] / H α )Upper Blue 0.69 − − − − − − (a) (a) From Osterbrock & Ferland (2006). very low (i.e. one and two orders of magnitude smaller than inthe giant H ii region for the upper and lower portions respec-tively); v) displays two distinct kinematic components in theupper part of the f.o.v.. Noteworthy is that two supernova rem-nant candidates have been detected in the area (Labrie & Pritchet2006). One of them (S001) appears very close in projection tothe massive clusters at ∼ [-5 . ′′ . ′′ Fig. 22.
Sketch showing the di ff erent elements associated withthe right part of the FLAMES field of view. The two massiveSSCs (stars in the circles) have already expelled their mate-rial and are seen as relatively quiescent ionized shells (shownin grey). Di ff erences in extinction are represented by two wavysheets.The first question to consider is if any of the kinematiccomponents is related to the supernova remnant candidates.However, our measured velocity dispersions are much morelower than expansion velocities of typical supernova remnants(i.e. NGC 1952, 1450 km s − Osterbrock & Ferland 2006).Moreover, we created two integrated spectra for the upperand the lower part of the area. Line ratio for both compo-nents of the upper part and for the lower part were relativelysimilar (i.e. within ∼ ff erences between the three com-ponents, this area can be viewed as a snapshot of a more evolved version of what is happening in the left part of the FLAMESf.o.v. The clusters have managed to clear out their environment.Only a broken shell made out of previously quiescent gas re-mains ionized by the remaining hot stars and moving away fromthe clusters with little evidence for high velocity outflow. Figure22 presents a sketch of the di ff erent elements associated with thisarea.
5. Summary
We present a thorough study of the ionized gas and its relationwith the stellar population of NGC 5253 by mapping the central212 pc ×
134 pc in a continuous and unbiased manner using withthe ARGUS IFU unit of FLAMES. The analysis of the data haveyield the following results.1. We obtained a 2D detailed map for the extinction su ff eredby the ionized gas, finding an o ff set of ∼ . ′′ ff ered by gas and stars bydefining ad hoc broad-band colours. We have shown the im-portance of using line-free filters when performing this com-parison and found that stars su ff er less extinction than theionized gas by a factor ∼ .
33, similar to the findings in otherstarburst galaxies.3. We derived N e sensitive line ratio maps. The one involv-ing the sulphur lines shows a gradient from 790 cm − atthe peak of emission in the giant H ii region described byCalzetti et al. (1997) outwards. The argon line ratio is onlydetected in the area associated with the giant H ii region andtraces the highest density ( ∼ − − ) regions.4. We studied the ionization structure by means of the maps ofline ratios involved in the BPT diagrams. The spatial distri-bution of the [S ii ] λλ / H α and [O iii ] λ / H β lineratios follows that for the flux distribution of the ionized gas.On the contrary, the [N ii ] λ / H α map shows a completelydi ff erent structure.5. We evaluated the possible ionization mechanisms throughthe position of these line ratios in the diagnostic diagramsand comparing with the predictions of models. All our lineratios are compatible with photoionization caused by stars.The [S ii ] λλ / H α indicated a somewhat higher ion-ization degree that might be evidence of some contribution ofshocks to the measured line ratios. Part of the data in the di-agram involving the [N ii ] λ / H α line ratio are distributedin a distinct cloud. This can be explained within the localN-enrichment scenario proposed for this galaxy.6. We delimited very precisely the area presenting local N-enrichment. It occupies the whole giant H ii , including thetwo extensions towards the upper and lower part of theFLAMES field of view, peaking at ∼ . ′′ ∼
20 pc) with the peak of extinction.7. We located the areas that could contain Wolf-Rayet stars bylooking for the blue bump . We confirmed the existence of W-R stars associated with the nucleus and the brightest clusterin the ultraviolet. W-R stars are distributed in a wider areathan the one presenting N-enrichment and in a more irregularmanner. We were able to identified one (or more) clusterswith ages compatible with the existence of W-R stars in allbut one (i.e. W-R 5) of our delineated regions with a W-Rsignature. 8. If the scenario of
N-enrichment caused by W-R stars turnsout to be applicable, only the W-R detected at the core(Complex ii region, can be considered the cause of the localN-enrichment, according to the correlation of the spatial dis-tribution of W-R features and N-enrichment.9. We measured the He + and He ++ abundances. He + / H + is ∼ ∼ ′′ × ′′ in the upper right corner, far away from themain ionization source. We detected the nebular He ii λ ii were always ∼ < ++ based on optical observations in the nu-clear region of NGC 5253. This result is di ffi cult to reconcilewith the scenario of N-enrichment caused by W-R stars andfavours a suggestion where the N-enrichment arises duringthe late O-star wind phase.10. We studied the kinematics of the ionized gas by using ve-locity fields and velocity dispersion maps for the main emis-sion lines. We needed up to three components to properlyreproduce the line profiles. In particular, one of the compo-nents associated with the Giant H ii region presents super-sonic widths and [N ii ] λ ii ] λλ − with respect to H α . Also,one of the narrow components shows velocity o ff sets in the[N ii ] λ − . This is the first time thatmaps providing such o ff sets for a starburst galaxy have beenpresented.11. We provide a scenario for the event occurring at the GiantH ii region. The two SSCs are producing an outflow that en-counters previously existing quiescent gas. The scenario isconsistent with the measured extinction structure, electrondensities and kinematics.12. We explain the di ff erent elements in the right (south-west)part of the FLAMES field of view as a more evolved stageof a similar scenario where the clusters have now clearedtheir local environment. This is supported by the low electrondensities and H α surface brightness as well as the kinematicsin this area. Acknowledgements.
We thank Peter Weilbacher for his help in the initial stagesof this project. We also thank the anonymous referee for his / her careful anddetailed review of the manuscript. Based on observations carried out at theEuropean Southern Observatory, Paranal (Chile), programme 078.B-0043(A).This paper uses the plotting package jmaplot , developed by Jes´us Ma´ız-Apell´aniz, http://dae45.iaa.csic.es:8080/ ∼ jmaiz/software . This re-search made use of the NASA / IPAC Extragalactic Database (NED), which isoperated by the Jet Propulsion Laboratory, California Institute of Technology,under contract with the National Aeronautics and Space Administration. AMI issupported by the Spanish Ministry of Science and Innovation (MICINN) underthe program ”Specialization in International Organisations”, ref. ES2006-0003.This work has been partially funded by the Spanish PNAYA, projects AYA2007-67965-C01 and C02 from the Spanish PNAYA and CSD2006 - 00070 ”1stScience with GTC” from the CONSOLIDER 2010 programme of the SpanishMICINN.
References
Allen, M. G., Groves, B. A., Dopita, M. A., Sutherland, R. S., & Kewley, L. J.2008, ApJS, 178, 20Alonso-Herrero, A., Garc´ıa-Mar´ın, M., Monreal-Ibero, A., et al. 2009, A&A,506, 1541Alonso-Herrero, A., Rieke, G. H., Rieke, M. J., & Scoville, N. Z. 2002, AJ, 124,166Alonso-Herrero, A., Takagi, T., Baker, A. J., et al. 2004, ApJ, 612, 222 . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 17
Asplund, M., Grevesse, N., Sauval, A. J., Allende Prieto, C., & Kiselman, D.2004, A&A, 417, 751Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5Bordalo, V., Plana, H., & Telles, E. 2009, ApJ, 696, 1668Brinchmann, J., Kunth, D., & Durret, F. 2008, A&A, 485, 657Calzetti, D., Harris, J., Gallagher, III, J. S., et al. 2004, AJ, 127, 1405Calzetti, D., Kinney, A. L., & Storchi-Bergmann, T. 1994, ApJ, 429, 582Calzetti, D., Meurer, G. R., Bohlin, R. C., et al. 1997, AJ, 114, 1834Campbell, A., Terlevich, R., & Melnick, J. 1986, MNRAS, 223, 811Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245Casta˜neda, H. O., V´ılchez, J. M., & Copetti, M. V. F. 1990, ApJ, 365, 164Conti, P. S. 1976, Memoires of the Societe Royale des Sciences de Li`ege, 9, 193Conti, P. S., Crowther, P. A., & Leitherer, C. 2008, From Luminous Hot Starsto Starburst Galaxies, ed. Conti, P. S., Crowther, P. A., & Leitherer, C.(Cambridge University Press)Cresci, G., Vanzi, L., & Sauvage, M. 2005, A&A, 433, 447Dopita, M. A., Fischera, J., Sutherland, R. S., et al. 2006, ApJS, 167, 177Esteban, C., Peimbert, M., Torres-Peimbert, S., & Rodr´ıguez, M. 2002, ApJ, 581,241Fluks, M. A., Plez, B., The, P. S., et al. 1994, A&AS, 105, 311Garc´ıa-D´ıaz, M. T., Henney, W. J., L´opez, J. A., & Doi, T. 2008, RevistaMexicana de Astronom´ıa y Astrof´ısica, 44, 181Garc´ıa-Lorenzo, B., Cair´os, L. M., Caon, N., Monreal-Ibero, A., & Kehrig, C.2008, ApJ, 677, 201Garnett, D. R., Kennicutt, Jr., R. C., Chu, Y., & Skillman, E. D. 1991, ApJ, 373,458Gonz´alez-Delgado, R. M., P´erez, E., Tenorio-Tagle, G., et al. 1994, ApJ, 437,239Gonz´alez-Riestra, R., Rego, M., & Zamorano, J. 1987, A&A, 186, 64Haro, G. 1956, AJ, 61, 178Harris, J., Calzetti, D., Gallagher, III, J. S., Smith, D. A., & Conselice, C. J. 2004,ApJ, 603, 503Henney, W. J., Arthur, S. J., Williams, R. J. R., & Ferland, G. J. 2005, ApJ, 621,328Holovatyy, V. V. & Melekh, B. Y. 2002, Astronomy Reports, 46, 779Hora, J. L., Latter, W. B., & Deutsch, L. K. 1999, ApJS, 124, 195Izotov, Y. I., Schaerer, D., Blecha, A., et al. 2006, A&A, 459, 71James, B. L., Tsamis, Y. G., Barlow, M. J., et al. 2009, MNRAS, 398, 2Karachentsev, I. D., Tully, R. B., Dolphin, A., et al. 2007, AJ, 133, 504Kau ff mann, G., Heckman, T. M., Tremonti, C., et al. 2003, MNRAS, 346, 1055Kehrig, C., V´ılchez, J. M., S´anchez, S. F., et al. 2008, A&A, 477, 813Kennicutt, Jr., R. C. 1992, ApJS, 79, 255Kewley, L. J. & Dopita, M. A. 2002, ApJS, 142, 35Kewley, L. J., Dopita, M. A., Sutherland, R. S., Heisler, C. A., & Trevena, J.2001, ApJ, 556, 121Kobulnicky, H. A., Kennicutt, Jr., R. C., & Pizagno, J. L. 1999, ApJ, 514, 544Kobulnicky, H. A. & Skillman, E. D. 2008, AJ, 135, 527Kobulnicky, H. A., Skillman, E. D., Roy, J.-R., Walsh, J. R., & Rosa, M. R. 1997,ApJ, 477, 679Koribalski, B. S., Staveley-Smith, L., Kilborn, V. A., et al. 2004, AJ, 128, 16Kunth, D. & ¨Ostlin, G. 2000, A&A Rev., 10, 1Labrie, K. & Pritchet, C. J. 2006, ApJS, 166, 188Lagos, P., Telles, E., Mu˜noz-Tu˜n´on, C., et al. 2009, AJ, 137, 5068Leitherer, C., Schaerer, D., Goldader, J. D., et al. 1999, ApJS, 123, 3L´opez-S´anchez, ´A. R., Esteban, C., Garc´ıa-Rojas, J., Peimbert, M., & Rodr´ıguez,M. 2007, ApJ, 656, 168Markwardt, C. B. 2009, in Astronomical Society of the Pacific ConferenceSeries, Vol. 411, Astronomical Society of the Pacific Conference Series, ed.D. A. Bohlender, D. Durand, & P. Dowler, 251– + Martin, C. L. 1998, ApJ, 506, 222Martin, C. L. & Kennicutt, Jr., R. C. 1995, ApJ, 447, 171Mart´ın-Hern´andez, N. L., Schaerer, D., & Sauvage, M. 2005, A&A, 429, 449Osterbrock, D. E. & Ferland, G. J. 2006, Astrophysics of gaseous nebulae andactive galactic nuclei, ed. D. E. Osterbrock & G. J. FerlandPagel, B. E. J., Simonson, E. A., Terlevich, R. J., & Edmunds, M. G. 1992,MNRAS, 255, 325Pasquini, L., ´Avila, G., Blecha, A., et al. 2002, The Messenger, 110, 1P´erez-Montero, E. & Contini, T. 2009, MNRAS, 398, 949Rieke, G. H. & Lebofsky, M. J. 1985, ApJ, 288, 618Sakai, S., Ferrarese, L., Kennicutt, Jr., R. C., & Saha, A. 2004, ApJ, 608, 42Sanders, D. B., Mazzarella, J. M., Kim, D.-C., Surace, J. A., & Soifer, B. T.2003, AJ, 126, 1607Schaerer, D. 1996, ApJ, 467, L17 + Schaerer, D., Contini, T., Kunth, D., & Meynet, G. 1997, ApJ, 481, L75 + Schaerer, D., Contini, T., & Pindao, M. 1999, A&AS, 136, 35Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525Shaw, R. A. & Dufour, R. J. 1995, PASP, 107, 896Slavin, J. D., Shull, J. M., & Begelman, M. C. 1993, ApJ, 407, 83 Stasi´nska, G., Cid Fernandes, R., Mateus, A., Sodr´e, L., & Asari, N. V. 2006,MNRAS, 371, 972Strickland, D. K. & Stevens, I. R. 1999, MNRAS, 306, 43Summers, L. K., Stevens, I. R., Strickland, D. K., & Heckman, T. M. 2004,MNRAS, 351, 1Taylor, V. A., Jansen, R. A., Windhorst, R. A., Odewahn, S. C., & Hibbard, J. E.2005, ApJ, 630, 784Tenorio-Tagle, G., Munoz-Tunon, C., Perez, E., & Melnick, J. 1997, ApJ, 490,L179 + Tsamis, Y. G., Barlow, M. J., Liu, X., Danziger, I. J., & Storey, P. J. 2003,MNRAS, 345, 186Turner, J. L., Beck, S. C., Crosthwaite, L. P., et al. 2003, Nature, 423, 621Turner, J. L., Beck, S. C., & Ho, P. T. P. 2000, ApJ, 532, L109van den Bergh, S. 1980, PASP, 92, 122Veilleux, S. & Osterbrock, D. E. 1987, ApJS, 63, 295V´ılchez, J. M. & Iglesias-P´aramo, J. 1998, ApJ, 508, 248Walsh, J. R. & Roy, J.-R. 1989, MNRAS, 239, 297Wang, W., Liu, X., Zhang, Y., & Barlow, M. J. 2004, A&A, 427, 873Westmoquette, M. S., Gallagher, J. S., Smith, L. J., et al. 2009, ApJ, 706, 1571Westmoquette, M. S., Smith, L. J., Gallagher, J. S., & Exter, K. M. 2007,MNRAS, 381, 913
Fig. 16.
Profiles of the main emission lines for representative spectra. The distribution within the FLAMES f.o.v. has been roughlyretained. That is: spaxel (1,1) is located at the upper right corner of the FLAMES IFU and numbering increases towards the leftand bottom. The middle row contains the profiles corresponding to the positions of the peaks of continuum emission (left: Complex ii region(upper: extension towards the north-west; lower: extension towards the south-east), those in the central column to areas in the centreof the FLAMES f.o.v. and those in the right column to the area associated with Complex . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 19 λ (nm)−10010203040 F ( a . u . ) Fit (17,3) 656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)0510 F ( a . u . ) Fit (12,4) 656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)−0.50.00.51.01.52.0 F ( a . u . ) Fit (4,3)656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)−20020406080100 F ( a . u . ) Fit (18,6) 656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)02040 F ( a . u . ) Fit (13,9) 656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)012 F ( a . u . ) Fit (3,5)656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)051015 F ( a . u . ) Fit (20,9) 656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)0510 F ( a . u . ) Fit (9,10) 656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)051015 F ( a . u . ) Fit (5,11)
Fig. 17.
Fits in H α for the representative spectra shown in Figure 16. Flux is in arbitrary units. The observed spectra are presentedin black. The first, second and third components have been displayed in blue, red, and yellow, respectively, while the underlyingcontinuum and total fit are in violet and green. Residuals are shown as a dotted line below the profile. Figura 4: Diferences between the H α and [O III ] (left) and H α and H β (right) velocity fields fittedwith one component. −4−2024−202 +370+415 ∆ x (arcsec) ∆ y ( a r cs e c ) C1: v(km s −1 ) ++ + −4−2024−202 +330+405 ∆ x (arcsec) ∆ y ( a r cs e c ) C2: v(km s −1 ) ++ + −4−2024−202 +335+375 ∆ x (arcsec) ∆ y ( a r cs e c ) C3: v(km s −1 ) ++ + −4−2024−202 +0+40 ∆ x (arcsec) ∆ y ( a r cs e c ) C2: σ (km s −1 ) ++ + Figura 5: Campo de velocidades (izquierda) y dispersi ´on de velocidad (derecha) para las diferentescomponentes de H α . 4 Fig. 18.
Kinematic information derived from the H α emission line. Left:
Velocity fields for the two narrow components.
Right:
Velocity field (upper) and velocity dispersion map (lower) for the broad component. −30+30 ∆ x (arcsec) ∆ y ( a r cs e c ) C1: v H α − v [NII] (km s −1 ) ++ −40+40 ∆ x (arcsec) ∆ y ( a r cs e c ) C2: v H α − v [NII] (km s −1 ) ++ XO −55+55 ∆ x (arcsec) ∆ y ( a r cs e c ) C2: v H α − v [SII] (km s −1 ) ++ Fig. 19. Di ff erence between the velocity fields derived from the H α and the [N ii ] λ α and [S ii ] λλ ii region is shown. The “X” and the “O” symbolsin the central panel show the position of the spaxels (17,4) and (17,9), repectively (see line fitting in Figure 20). . Monreal-Ibero et al.: The central region of NGC 5253 with FLAMES 21 λ (nm)−100102030 F ( a . u . ) Fit (17,4) 658.6 658.8 659.0 659.2 659.4 659.6 659.8 λ (nm)-101234 F ( a . u . ) Fit (17,4) 672.0 672.2 672.4 672.6 672.8 673.0 673.2 λ (nm)-0.50.00.51.01.52.0 F ( a . u . ) Fit (17,4)656.6 656.8 657.0 657.2 657.4 657.6 657.8 λ (nm)−10010203040 F ( a . u . ) Fit (17,9) 658.6 658.8 659.0 659.2 659.4 659.6 659.8 λ (nm)-10123 F ( a . u . ) Fit (17,9) 672.0 672.2 672.4 672.6 672.8 673.0 673.2 λ (nm)-10123 F ( a . u . ) Fit (17,9)
Fig. 20.
Fits in H α (left), [N ii ] λ ii ] λ ff sets in velocity forthe di ffff