arXiv:1101.3092v2 [astro-ph.GA] 18 Jan 2011 rsneo a eg,Eas&Clet1994). Collett & induce Evans origin (e.g., bar coun internal a of an of have cases presence could special discs some stellar Nevertheless, Co rotating & review). Bertola a acquisit (see for retrograde formation 1999 a star c subsequent of is and result and gas end external disc the several as othe in interpreted detected to monly been respect has with gas counter-rotating and/or stars of presence The INTRODUCTION 1 o.Nt .Ato.Soc. Astron. R. Not. Mon. esta 0 ftesuidsia aaishs counter- a host galaxies ⋆ spiral f studied In the detectable spirals. of in L a al 10% phenomenon et host 1996). (Bertola rare than them al. origin a et less (Kuijken of external is stars of counter-rotation ex- 10% scale counter-rotating is we than S0’s that of less in 50% fraction contrast, gas the In coun the with 1992). a all assem consistent with if their is galaxies understand pect disc lenticular to of gaseous key fraction a rotating The is process. spirals and bly S0’s in ponents 1 Bertola F. † 383.B-0632. program the for .Coccato L. populations stellar its disentangling i by disc 5719 stellar NGC counter-rotating the of formation the Dating 2 3 cetd. Received... Accepted... 4 c uoenSuhr bevtr,Karl-Schwarzschild-Stra Observatory, Southern European iatmnod srnma nvri` iPdv,vicolo Padova, Universit`a di Astronomia, di Dipartimento NF sevtroAtooiod aoa ioodell’Oss vicolo Padova, di Astronomico Osservatorio INAF, NF sevtroAtooiod oon,vaRnai1 I 1, Ranzani via Bologna, di Astronomico Osservatorio INAF, -al [email protected] E-mail: ae nosrain olce tteErpa otenOb Southern European the at collected observations on Based 01RAS 2011 h eorpyo aeu n tla one-oaigcom counter-rotating stellar and gaseous of demography The 2 1 † .Morelli L. , 000 1L 21)Pitd1 etme 08(NL (MN 2018 September 18 Printed (2011) L1–L5 , one,ls ihi eas more ionised- metals, the in of rich kinematics less younger, and distribution the derive also eevi vial nisnihorod stersl ft of the result fuelled ret the subsequently a and as onto 5713, neighbourhoods 5719 NGC its NGC by in first available of accreted reservoir body was gas stellar where main scenario the the respect with counter-rotating both bnacs–glxe:sias–glxe:selrcontent stellar galaxies: – spirals galaxies: – abundances epeetterslso h L/IO nerlfil spect integral-field VLT/VIMOS the of inner results the present We ABSTRACT opnn ihsnl tla ouainmdl htaccou that models population stellar a component single counter-rotating kinematics with stellar component the two the measure of We indices Lick components. contributi ionised-gas the into one decomposed e and is At spectrum discs. stellar galaxy counter-rotating observed co-spatial two host to e words: Key 2 .M Corsini M. E. , 28 ′′ × aais niiul(G 79 aais ieaisand kinematics galaxies: – 5719) (NGC individual galaxies: 28 β elOsraoi ,I312Pdv,Iay. Italy Padova, I-35122 3, dell’Osservatorio ′′ raoi ,I312Pdv,Italy. Padova, I-35122 5, ervatorio ,D878Grhn e M¨unchen, Germany. bei Garching D-85748 2, e 417 oon,Italy. Bologna, -40127, ( 3 ythe by d servatory . kpc 1 stars r o of ion arge- rsini om- act, ter- ter- 2 - - . .Buson L. , × 3 . kpc 1 etdt emd yyugrsasta hs fterhs gal host their of those than stars disc younger by stellar made form counter-rotating be star to scenario pected subsequent this by an in produced 1998) Therefore, are al. discs Garc´ıa-Burillo Co et 2003). stellar al. 1995; Garc´ıa-Burillorotating et 1994; al. al. et et (Braun (Ciri 4826 det NGC 3626 components counter-rotating NGC gaseous in Thi purely content. the gas of pre-existing acquisi the case retrograde than the larger gas gas from Counter-rotating of only amounts formed gas. acquired are pre-existing spirals newly the in the discs by spirals away gas-rich swept in is while disc only, gaseous amoun counter-rotating S0’s small to poor rise of give acquisition can gas retrograde inte external Pizze the To counter-rotations, that 2007). of argue al. et (2004) frequencies (Vergani observed 5719 the NGC NGC and pret 2000), 1996), al. 2004) al. Garc´ıa-Burillo et et al. al. 1998; (Jore et et al. Emsellem (Bertola et 3593 Corsini 1992; NGC 1994), 1996; al. Kuijken et st & Rix (Merrifield in large-scale 7217 1992; NGC only al. of observed et been (Rubin case has 4550 galaxies particular disc The in counter-rotation 2004). Fabrican al. & et (Kannappan Pizzella disc stellar and/or gaseous rotating ar hc a enrcnl tde yVraie l (2007 al. et Vergani by studied recently been has which pair, α ehne,adls uiosselrcmoet hyare They component. stellar luminous less and -enhanced, hspcuecnb ietytse nteNC51/3galaxy 5719/13 NGC the in tested directly be can picture This fteitrcigsia G 79 hc sknown is which 5719, NGC spiral interacting the of ) A T E nsitu in 3 tl l v2.2) file style X .Pizzella A. , omto ftecutrrttn tla disc. stellar counter-rotating the of formation c oiini h edo iw the view, of field the in position ach a ic hti soitdwt the with associated is that disc, gas .W oe h aao ahstellar each of data the model We s. tfrthe for nt eitrcinwt t companion its with interaction he h aay hs nig prove findings These galaxy. the 2 n ftesetao w stellar two of spectra the of ons .Vergani D. , ocpcosrain fthe of observations roscopic dteln tegh fthe of strengths line the nd ord ri rmtelarge the from orbit rograde α F vrbnac.We overabundance. /Fe ⋆ h spiral the n yais–galaxies: – dynamics 4 , r ex- are s ngas- in s l tal. et lla 2001; t sthe is s inof tion unter- ation. ected NGC 4138 sof ts eous ellar axy. gas r- ). d , 2 L. Coccato et al.

by comparing the intensity of the night-sky emission lines. The three offset observations were used to construct three sky spectra. To compensate for the time variation of the relative intensity of the night-sky emission lines, we compared the fluxes of the sky lines measured in the offset and on-target exposures. The corrected sky spectra were then subtracted from the corresponding on-target ex- posures. Each exposure was organised in a data cube using the tabu- lated correspondence between each fibre and its position in the field of view. The three sky-subtracted data cubes were aligned using the bright galaxy nucleus as reference and co-added into a single data cube. In order to increase the signal to noise ratio (S/N), spec- tra from fibres mapping adjacent regions in the sky were added together using the Voronoi binning method (Cappellari & Copin 10′′ Figure 1. Fit of the galaxy spectrum (black) in the spatial bin at East 2003). Some of the spatial bins were modified to include only from the galaxy center. The best fit model (green) is the sum of the spec- spectra from the regions associated to intense H emission or the tra of the ionised-gas component (magenta) and the two stellar components β (blue and red). The latter are obtained convolving the synthetic templates dust lanes crossing the galactic disc. This ensured us to minimise with the best fitting Gaussian LOSVDs and multiplying them by the best the contamination of the kinematic and stellar-population proper- fitting Legendre polynomial. The differences in the position of absorption ties to be measured from spectra obtained in regions where stellar line features and in the Hβ equivalent widths between the two stellar com- counter-rotation was detected and chemical decoupling is expected ponents (indicating different kinematics and stellar population content) are (Vergani et al. 2007). In total, we have 130 spatial bins few arc- clearly evident. seconds square large. We tested the robustness of our results using different binning schemes. NGC 5719 is an almost edge-on Sab galaxy with a prominent skewed dust lane at a distance of 23.2 Mpc. Vergani et al. (2007) report a spectacular on-going interaction with its face-on Sbc com- 3 KINEMATICS AND LINE STRENGTH INDICES panion NGC 5713. Two H I tidal bridges loop around NGC 5719 and connect it to NGC 5713 at a projected distance of 77 kpc. The In order to measure the kinematics and stellar population proper- neutral and ionised hydrogen in the disc of NGC 5719 are counter- ties of the two counter-rotating stellar components in NGC 5719, rotating with respect to the main stellar disc. The kinematics of we need to separate their contribution to the observed spectrum in the ionized-gas disc and of both the counter-rotating stellar discs each spatial bin, taking advantage of their different velocities that ′′ are measured out to about 40 (4.3 kpc) from the galaxy centre. causes a wavelength shift of their spectra. To do that, we modi- In conclusion, Vergani et al. (2007) propose a scenario where H I fied the penalized pixel fitting code (pPXF, Cappellari & Emsellem from the large reservoir available in the galactic surroundings was 2004) in a similar way to that done by McDermid et al. (2006). accreted by NGC 5719 onto a retrograde orbit and subsequently The code builds two synthetic templates (one for each stellar com- fuelled the in situ formation of the counter-rotating stellar disc. ponent) as linear combination of stellar spectra from the MILES ˚ In this work, we will address the crucial piece of information library (at FWHM = 2.54 A spectral resolution, Beifiori et al. which is still missing, i.e. proving that the stellar population of the 2010) and convolves them with two Gaussian line-of-sight veloc- counter-rotating disc of NGC 5719 is younger with respect to that ity distributions (LOSVDs) with different radial velocities and the of the stars in the galaxy main disc. same velocity dispersion σ. This assumption is in agreement with the findings by Vergani et al. (2007). Results are confirmed if two different velocity dispersions are considered, although the velocity fields and maps of stellar population properties appear more noisy 2 OBSERVATIONS AND DATA REDUCTION due to the additional degree of freedom that increases the degener- The integral-field spectroscopic observations were carried out in acy between measured parameters. Gaussian functions are added to service mode with the Very Large Telescope (VLT) at the European the convolved synthetic templates to account for ionised-gas emis- Southern Observatory (ESO) in Paranal during dark time between sion lines (Hβ, [O III]λλ4959, 5007, and [N I]λ5198) and fit si- 28 April and 16 June 2009. The Unit Telescope 3 was equipped multaneously to the observed galaxy spectra. Multiplicative Leg- with the Visible Multi Object Spectrograph (VIMOS) in the Inte- endre polynomials are included to match the shape of the galaxy gral Field Unit (IFU) configuration. The HR blue grism covering continuum, and are set to be the same for the two synthetic tem- the spectral range 4150 – 6200 A˚ and the 0′′. 67 fibre−1 resolution plates. Our technique represents an improvement with respect to were used. The instrumental spectral resolution measured at 5200 Vergani et al. (2007): in fitting the measured LOSVD with a dou- − A˚ was 2.0 A˚ (FWHM), equivalent to 115 km s 1. Observations ble Gaussian, they assumed that the two counter-rotating compo- were organised into three dithered on-target exposures of 2950 sec- nents had the same stellar population. On the contrary, our syn- onds each, alternated to three offset sky exposures of 280 seconds thetic templates account for different stellar populations since they each. The average seeing measured by the ESO Differential Image are independently built from the MILES library. We check the Meteo Monitor was 1′′. robustness of our results using also the INDO-US Coud´elibrary Data reduction (bias subtraction, fibre identification and trac- of stellar spectra (Valdes et al. 2004) and the single stellar popu- ing, flat fielding, wavelength calibration and correction for instru- lation (SSP) synthesis models by Vazdekis et al. (2010) to build ment transmission) was performed using the VIMOS ESO pipeline the synthetic templates. Figure 1 shows an example of the de- version 2.2.1 (http://www.eso.org/sci/software/pipelines/). The dif- composition of the galaxy spectrum measured in one of the spa- ferent relative transmission of the VIMOS quadrants was corrected tial bins where the two stellar counter-rotating components and

c 2011 RAS, MNRAS 000, L1–L5 Counter-rotating discs in NGC 5719 3

Figure 2. Velocity field of the main stellar component (left panel), counter-rotating stellar component (central panel) and counter-rotating ionised gas (right panel) in NGC 5719. Centre is in R.A. = 14h40m56.3s, Dec. = −00◦19′05.4′′ (J2000.0). Black contours: isophotes derived from the Hβ emission map.

Figure 3. Equivalent width of the Lick indices in the main (red diamonds) and counter-rotating (blue circles) stellar components. Predictions from single stellar population models by Thomas et al. (2010) are superimposed. Crosses indicate mean error bars associated to the equivalent widths. ionised gas are observed. We perform the double stellar compo- We assume that the two synthetic templates are a good approx- nent fit only in 73 spatial bins, where the separation of the two imation of the mean stellar populations of the two counter-rotating kinematic components is reliable, and do the single stellar compo- components. This depends on the adopted library of stellar spec- nent fit to the spectra of the other bins. The spectra of outermost tra and on the accuracy of the fitting program in recovering simul- spatial bins have a low S/N, therefore only the ionised-gas emis- taneously the kinematics (i.e, radial velocity and velocity disper- sion lines are fitted. The stellar kinematics are measured by tak- sion) and the population properties (i.e., the continuum and lines ing into account the difference in spectral resolution between the strength) of the two counter-rotating components. To test this, we VIMOS spectra and the MILES library. We derive the properties carry out Monte Carlo simulations on a set of 1200 artificial galaxy of the stellar populations of the two stellar counter-rotating com- spectra to test the reliability and accuracy of the procedure to mea- ponents of NGC 5719 by measuring the line strength of the Lick sure the kinematics and line strength of the Lick indices of the two indices Hβ, Mgb, Fe5270, and Fe5335 as defined by Worthey et al. counter-rotating components and quantify the errors. We generate (1994) in the synthetic templates. To this aim, the spectra are set each galaxy spectrum by summing two stellar spectra with differ- to rest-frame by adopting the measured radial velocity and con- ent velocity separation (∆V ), different fraction of stars in the main volved with a Gaussian function to match the spectral resolu- component with respect to the total (Fmain), different σ, and dif- tion of the Lick system (FWHM = 8.4 A).˚ We also calculate ferent S/N. The two stellar spectra are chosen to have similar line the mean iron index h Fei = (Fe5270 + Fe5335) /2 (Gorgas et al. strengths to those measured for the corresponding stellar compo- ′ 1990) and the combined magnesium-iron index [MgFe] = nent. Gas emission lines are also added to the artificial galaxy spec- ′ pMgb (0.82 · Fe5270 + 0.28 · Fe5335). The [MgFe] index is al- tra. The modified pPXF routine is then applied to analyse the artifi- most independent from α-enhancement and hence serves best as a cial galaxy spectra as if they are real. Errors on the fitted parameters metallicity tracer (Thomas et al. 2003). are estimated by comparing the input and output values. Errors are

c 2011 RAS, MNRAS 000, L1–L5 4 L. Coccato et al.

Figure 4. Two-dimensional maps of age, [Z/H] and [α/Fe] of the main (upper panels) and counter-rotating (lower panels) stellar components, in the region where the two components are disentangled. Black contours: isophotes derived from the Hβ emission map. Centre is as in Fig. 2. assumed to be normally distributed, with mean and standard devia- and constant. This represents only a rigid shift to the points in Fig. tion corresponding to the systematic (∆S) and random (∆R) errors 3, and therefore does not change our result. on the relevant parameters, respectively. We find that for typical The luminosity-weighted metallicity ([Z/H]), α-enhancement − values observed in NGC 5719 (∆V & 150 km s 1, 30 . S/N . ([α/Fe]) and population age are derived for each component by a −1 90, σ ≃ 90 km s , 30% . Fmain . 70%) the random errors linear interpolation between the model predictions by Thomas et al. −1 on the recovered ∆V and Fmain are smaller than 25 km s and (2010) using an iterative procedure as done in Morelli et al. (2008). 11%, respectively, while systematic errors are negligible. The mea- We do not include in the interpolation all the spatial bins with surements of the line strengths are affected both by random and equivalent width of the Lick indices located outside the model grid systematic errors: if the true equivalent width of the spectral index more than the mean errorbar. The two-dimensional maps of age, I of the two components are I1 and I2 with I1 < I2, the program [Z/H] and [α/Fe] of the main and counter-rotating stellar compo- measures (I1 +∆S) ± ∆R and (I2 − ∆S) ± ∆R. For the typical nents are shown in Figure 4. values observed in NGC 5719, random errors are 0.15 − 0.30 A,˚ and systemic errors are 0.10 − 0.20 A,˚ mildly depending on I1 and I2. Figure 2 shows the measured two-dimensional velocity fields 4 DISCUSSION AND CONCLUSIONS of the main and counter-rotating stellar components and ionised gas. Figure 3 shows the Lick indices [MgFe]′,Hβ, Mgb, and h Fei We present the results of the VLT/VIMOS integral-field spectro- ′′ ′′ we measure in the bins where the two counter-rotating components scopic observations of the inner 28 × 28 (3.1kpc × 3.1kpc) of are disentangled. The predictions of the single stellar population the NGC 5719. The kinematics of the stars and ionised models by Thomas et al. (2010) are also shown for reference. Cor- gas are very complex. The observed galaxy spectra are decom- rection for the offset to the Lick system is not applied, since no posed into the contributions of three distinct kinematic components standard Lick stars were observed. Nevertheless, the VIMOS spec- characterised by a regular disc-like rotation: one main and one sec- tra are flux calibrated, therefore we expect the offset to be small ondary stellar component and a ionised-gas component. The spec- tral decomposition is done using an implementation of the pPXF

c 2011 RAS, MNRAS 000, L1–L5 Counter-rotating discs in NGC 5719 5 routine, which allows to measure simultaneously the kinematics rotating disc has been recently assembled. The median overabun- and stellar population properties of the two stellar components. dance of the counter-rotating component (0.14 dex) indicates a The rotation of the main stellar component is receding to- history with a time-scale of 2 Gyr (Thomas et al. wards East, like that of the stellar body of the galaxy as observed 2005). More details about the assembly process could be derived ′′ by Vergani et al. (2007) out to ∼ 40 . We measure a maximum by a further analysis of the east-west asymmetries measured in the rotation velocity of ∼ 150 km s−1 at about 10′′ (1.1 kpc) from maps of the stellar-population properties. The scenario proposed by the centre. The secondary stellar and ionised-gas components are Vergani et al. (2007) that NGC 5719 hosts a counter-rotating stellar counter-rotating with respect to the main stellar component, with a disc originated from the gas accreted during the ongoing merging maximum rotation of ∼ 200 km s−1 at about 10′′ from the cen- with its companion NGC 5713, is finally confirmed. tre. We are able to resolve the two stellar components down to the ′′ innermost ∼ 2 (0.2 kpc). At smaller radii, they have a too small velocity separation to be resolved. The median values of light frac- REFERENCES tion contributed by the two components in the spatial bins where Beifiori A., Maraston C., Thomas D., Johansson J., 2010, ArXiv: they are resolved are Fmain = 56% and Fsecondary = 44% (with 20% standard deviation), respectively. 1012.3428 The ionised gas is detected all over the observed field of Bertola F., Buson L. M., Zeilinger W. W., 1992, ApJ, 401, L79 view. It is characterised by a strong Hβ emission, which is con- Bertola F., Cinzano P., Corsini E. M., Pizzella A., Persic M., centrated in a twin-peaked morphology indicating an edge-on ring Salucci P., 1996, ApJ, 458, L67 ′′ with a semi-major axis of ∼ 7 (0.8 kpc), and an outer asymmet- Bertola F., Corsini E. M., 1999, in J. E. Barnes & D. B. Sanders ric m = 1 and fragmented spiral arc extending southwards with a ed., Galaxy Interactions at Low and High Vol. 186 of ′′ semi-major axis of ∼ 13 (1.5 kpc; see contours in Figs. 2 and 4). IAU Symposium, Counterrotation in Galaxies. p. 149 The two-stream fluid instability present in systems like NGC 3593 Braun R., Walterbos R. A. M., Kennicutt Jr. R. C., Tacconi L. J., (Garc´ıa-Burillo et al. 2000) can favor this morphology. An hint of 1994, ApJ, 420, 558 Cappellari M., Copin Y., 2003, MNRAS, 342, 345 the inner ring is visible also in the H I position-velocity diagram measured along the major axis of NGC 5719 (Vergani et al. 2007). Cappellari M., Emsellem E., 2004, PASP, 116, 138 The secondary stellar component is associated to these features Ciri R., Bettoni D., Galletta G., 1995, Nature, 375, 661 of the gas distribution, i.e. the counter-rotating stars are detected Corsini E. M., Pizzella A., Funes J. G., Vega Beltran J. C., Bertola in a region enclosed by the ionised-gas structures. A similar phe- F., 1998, A&A, 337, 80 nomenon is observed in NGC 3593, where a concentrated gaseous Emsellem E., Cappellari M., Peletier R. F., et al. 2004, MNRAS, ring is associated to the counter-rotating stellar disc (Corsini et al. 352, 721 1998; Garc´ıa-Burillo et al. 2000). Evans N. W., Collett J. L., 1994, ApJ, 420, L67 The two counter-rotating stellar components are characterised Garc´ıa-Burillo S., Combes F., Hunt et al. 2003, A&A, 407, 485 by different chemical properties as it results from the measured line Garc´ıa-Burillo S., Sempere M. J., Bettoni D., 1998, ApJ, 502, 235 strength of the Lick indices (Fig. 3). On average, the main stel- Garc´ıa-Burillo S., Sempere M. J., Combes F., et al. 2000, A&A, lar component has lower Hβ and higher [MgFe]′, Mgb, and h Fei 363, 869 with respect to the secondary component. This translates imme- Gorgas J., Efstathiou G., Aragon Salamanca A., 1990, MNRAS, diately into different properties of their stellar populations (Fig. 245, 217 4). The main stellar component has ages ranging from 2 to 13.5 Jore K. P., Broeils A. H., Haynes M. P., 1996, AJ, 112, 438 Kannappan S. J., Fabricant D. G., 2001, AJ, 121, 140 Gyr (median age = 4 Gyr with 4 Gyr standard deviation). It Kuijken K., Fisher D., Merrifield M. R., 1996, MNRAS, 283, 543 has nearly-solar metallicity (median [Z/H] = 0.08 dex with 0.20 dex standard deviation). It displays super-solar enhancement (me- McDermid R. M., Emsellem E., Shapiro K. L., et al. 2006, MN- RAS, 373, 906 dian [α/Fe] = 0.10 dex with 0.10 dex standard deviation). The counter-rotating stellar population component is younger, with ages Merrifield M. R., Kuijken K., 1994, ApJ, 432, 575 ranging from 0.7 Gyr to 2.0 Gyr (median age = 1.3 Gyr with 0.6 Morelli L., Pompei E., Pizzella A., et al. 2008, MNRAS, 389, 341 Gyr standard deviation). Its metallicity shows a radial gradient: it Neff S. G., Thilker D. A., Seibert M., et al. 2005, ApJ, 619, L91 Pizzella A., Corsini E. M., Vega Beltr´an J. C., Bertola F., 2004, changes from sub-solar ([Z/H] ∼ −1.0 dex) in the outskirts, to A&A, 424, 447 solar ([Z/H] = 0.0 dex) and super-solar ([Z/H] ∼ 0.3 dex) in the centre. The youngest ages and highest metallicities are found Rix H., Franx M., Fisher D., Illingworth G., 1992, ApJ, 400, L5 Rubin V. C., Graham J. A., Kenney J. D. P., 1992, ApJ, 394, L9 in correspondence of the regions where the Hβ emission is more Thomas D., Maraston C., Bender R., 2003, MNRAS, 339, 897 intense. The α-enhancement is super-solar (median [α/Fe] = 0.14 dex with 0.05 dex standard deviation). Thomas D., Maraston C., Bender R., Mendes de Oliveira C., 2005, ApJ, 621, 673 These findings extend the results by Neff et al. (2005) based Thomas D., Maraston C., Johansson J., 2010, MNRAS, in press on GALEX observations. They reported the presence of a young (arXiv: 1010.4569) stellar component in the disc of NGC 5719 by analysing UV and Valdes F., Gupta R., Rose J. A., et al. 2004, ApJS, 152, 251 optical images. Unfortunately, the poor GALEX angular resolution Vazdekis A., S´anchez-Bl´azquez P., Falc´on-Barroso J., et al. 2010, prevented them to associate it to the Hβ ring as we successfully do MNRAS, 404, 1639 (Fig. 4). Our estimate is in agreement with their lower limit to the Vergani D., Pizzella A., Corsini E. M., et al. 2007, A&A, 463, 883 actual age of such a young population (0.3 ± 0.1 Gyr). Worthey G., Faber S. M., Gonzalez et al. 1994, ApJS, 94, 687 With our new observations and spectral decomposition tech- nique, we prove that the mean age of the counter-rotating disc, which is associated to the neutral and ionised gas disc, is indeed younger than the main stellar disc. This result shows that counter-

c 2011 RAS, MNRAS 000, L1–L5