Star Cluster Complexes and the Host Galaxy in Three HII Galaxies: Mrk 36, UM 408, and UM 461
Patricio Lagos, Eduardo Telles, A. Nigoche-Netro, Eleazar Rodrigo Carrasco
aa r X i v : . [ a s t r o - ph . C O ] S e p Star Cluster Complexes and the Host Galaxy in Three H ii Galaxies: Mrk 36, UM 408, and UM 461
P. Lagos , [email protected] E. Telles [email protected] A. Nigoche-Netro [email protected] andE. R. Carrasco [email protected] ABSTRACT
We present a stellar population study of three H ii galaxies (Mrk 36, UM408, and UM 461) based on the analysis of new ground-based high resolutionnear-infrared J, H and K p broad-band and Br γ narrow-band images obtainedwith Gemini/NIRI. We identify and determine relative ages and masses of theelementary star clusters and/or star cluster complexes of the starburst regionsin each of these galaxies by comparing the colors with evolutionary synthesismodels that include the contribution of stellar continuum, nebular continuumand emission lines. We found that the current star cluster formation efficiencyin our sample of low luminosity H ii galaxies is ∼ Centro de Astrof´ısica da Universidade do Porto, Rua das Estrelas, 4150-762 Porto, Portugal Observat´orio Nacional, Rua Jos´e Cristino, 77, Rio de Janeiro, 20921-400, Brazil Instituto de Astrof´ısica de Andaluc´ıa (IAA), Glorieta de la Astronom´ıa s/n, 18008, Granada, Spain Gemini Observatory/AURA, Southern Operations Center, Casilla 603, La Serena, Chile & M ⊙ . Theage distribution of these star cluster complexes shows that the current burststarted recently and likely simultaneously over short time scales in their hostgalaxies, triggered by some internal mechanism. Finally, the fraction of the totalcluster mass with respect to the low surface brightness (or host galaxy) mass,considering our complete range in ages, is less than 1%. Subject headings: galaxies: individual (Mrk 36, UM 408, UM 461) – galaxies:stellar content – galaxies: dwarf – galaxies: star clusters – infrared: galaxies
1. Introduction H ii galaxies are dwarf galaxies undergoing an intense episode of star formation (SF)that dominates their total optical luminosity. Most H ii galaxies overlap in their observedproperties, with Blue Compact Dwarf (BCD) galaxies; indeed, many are found in bothsamples. Typically, they have relatively small physical sizes (with a few kpc in size), and thestarburst regions spatially cover the visible extent of the galaxy, making it difficult to accessthe presence of an underlying (or host galaxy) stellar population. Originally, their low heavyelement abundance and the non-detection of an old population have given rise to the questionof whether they may be presently forming their first generation of stars (Sargent & Searle1970). Recent works, however, have shown that most H ii galaxies seem to present anunderlying population from previous episodes of SF (e.g., Telles 1995; Papaderos et al. 1996;Telles & Terlevich 1997; Cair´os et al. 2003). So, the integrated light observed in H ii galaxiesis formed by the contribution of two main components, the young stellar population withages of a few Myr to tens Myr and an underlying population of intermediate (with ages ofhundreds Myr) to old stars (with ages ≥ ii galaxies (Lagos et al. 2007,and references therein) have profound implications for the origin of the present starburstand the dominant large scale mode of SF. While many high luminosity H ii galaxies showevidence of morphological disturbances that may be associated with interactions or ongoingminor mergers (e.g., Telles et al. 1997; Bergvall & ¨Ostlin 2002), less luminous and compactH ii galaxies defy our attempt to find morphological signatures of an external triggering agent(Telles et al. 1997). In fact, H ii galaxies seem to be typically isolated and not associatedwith giant galaxies (Telles & Terlevich 1995), and their clustering properties seem to besimilar to those of normal galaxies (Telles & Maddox 2000). There have been, however,many attempts to search for HI companions (Taylor 1997) and intrinsically faint opticalcompanions (Noeske et al. 2001; Pustilnik et al. 2001). In these studies, it was concluded 3 –that galaxy interactions with both massive and dwarfs are probably the main mechanism thattriggers SF bursts in BCD progenitors. Although interactions are often invoked to explainburst of SF in H ii /BCD galaxies, it is possible that internal processes (e.g., gravitationalcloud collapse and/or infall in conjunction with small perturbations) have been responsiblefor triggering the present episode of SF in at least a significant fraction of dwarf galaxies(see Hunter & Elmegreen 2004; Pelupessy et al. 2004; Telles 2009; Simpson et al. 2011, andreferences therein).It has been apparent for over a decade now, with the advent of the HST, that the star-burst regions in these galaxies are composed of a myriad of star clusters (e.g., Billett et al.2002), with masses & − M ⊙ and sizes of a few pc, that are typically more massive thannormal clusters in our Galaxy (Meurer et al. 1995; Conti & Vacca 1994). These Super StarCluster (SSCs) or Young Massive Clusters (YMCs) were found originally in classical starburstgalaxies and in galaxies with evident signs of interaction or merger such as in the AntennaeNGC 4038/4039 (Whitmore et al. 1999), dwarf irregular galaxies such as NGC 1569 andNGC 1705 (O’Connell et al. 1994), and other star-forming dwarf galaxies or BCD galaxiessuch as M82 (Melo et al. 2005) and SBS 0335-052 (Thuan et al. 1997; Papaderos et al. 1998).The formation of these massive clusters and the cluster mass function (CMF) is directly con-nected to the SF processes in the galaxies, in the sense that different physical factors, in thesegalaxies, put constrains on the mass of the star clusters. The CMF appears to be a singlepower law (with index ∼ -2), which implies the same SF mechanism for the massive clus-ters and their lower mass analogues. This type of young and massive clusters were possiblyformed in high pressure conditions (e.g., Elmegreen & Efremov 1997; Billett et al. 2002),hence the extremely high pressure regions give rise to more massive and compact clusters.We know that SF in clusters is a common phenomenon in starburst galaxies, and that themassive clusters play an important role on the evolution of the ISM of their host galaxies,producing large scale structures such as supershells (or bubbles) and creating galactic windsthat cause, in some cases, the blowout of freshly produced metals from the galaxy into theintergalactic medium (IGM). The similarity in mass and size between SSCs and Milky Wayglobular clusters (GCs; M ∼ × M ⊙ , Harris 1991) suggests the possibility of an evolu-tionary connection in the sense that SSCs are proto-globular cluster systems. The classicexample of a YMC is R136 at the center of the 30 Doradus nebula in the Large MagellanicCloud. This cluster has a diameter of ∼ ∼ × M ⊙ (Walborn et al. 2002, and references therein).Apart from evidence for massive star clusters and its effect on the surrounding medium,H ii galaxies show an underlying low surface brightness (LSB) component likely to representtheir host galaxies, whose stellar population is a product of previous episodes of SF. Thisunderlying stellar host generally extends a few kpc from the central star-forming regions, 4 –showing regular and elliptical isophotes with red colors indicative of an evolved stellar pop-ulation. The host galaxy in H ii galaxies has been studied in the literature based on theanalysis of both optical (e.g., Telles & Terlevich 1997) and near infrared (near-IR) images(e.g., Cair´os et al. 2003; Noeske et al. 2003, 2005) and spectroscopy (e.g., Raimann et al.2000; Westera et al. 2004). Thuan (1983) observed, in the near-IR, a sample of BCD galax-ies and showed that the light in these galaxies is due to the presence of K- and M-giant stars.The study of the structural parameters of the host galaxy (exponential scale length ( α ) andcentral surface brightness ( µ )) has been commonly used in the literature to evaluate theevolutionary relations between the different types of dwarf galaxies (e.g., Papaderos et al.1996; Telles et al. 1997; Gil de Paz & Madore 2005, and references therein). The derivationof the ages and spatial distribution of these stellar populations is the first step towardsestablishing the evolutionary state and the SF history of these galaxies.We present near-IR broad-band J, H and K p and Br γ narrow-band images of three lowluminosity H ii galaxies: Mrk 36, UM 408, and UM 461. Our aim is to describe the propertiesof the star clusters or complexes which are distinguishable with our superb ground-based highspatial resolution images with NIRI on the Gemini North telescope as well to determine thestructural properties of the underlying galaxy using surface photometry. These observationsin combination with a proper assessment of recent stellar population synthesis models allowus to put some constraints on the recent and past history, and the dominant large scale modeof SF in these galaxies.The paper is arranged as follows: In § § § §
2. Sample, observations and data reduction2.1. Our sample
We targeted two galaxies Mrk 36 and UM 461, that are particularly rich in giant star-forming knots and one compact H ii galaxy, UM 408, with less evidence of multiple knots ofSF. The light collecting power of the Gemini North telescope and the high spatial resolutionpermitted by NIRI are an excellent combination for resolving the star cluster or/and starcluster complex populations in these compact galaxies. In the following paragraphs wedescribe the main properties of our sample of galaxies. Mrk 36 (Haro 4, UGCA 225) is a compact H ii galaxy showing at least two large star- 5 –forming knots (Lagos et al. 2007) in the south-eastern part of the galaxy. Thuan (1983)reports integrated near-IR colors (de-reddened) of J − H = 0 .
38 and H − K = 0 .
42, whileHunter & Elmegreen (2006) reports integrated colors of J − H = 0 . ± .
100 and H − K = 0 . ± . β line emission (Lagos et al. 2007) and HImaps (Bravo-Alfaro et al. 2004) coincide, indicating that the current SF is restricted to thedense region of the parental cloud. Bravo-Alfaro et al. argue that a transient encounterbetween Mrk 36 and the neighboring spiral Haro 26 could explain both the SF in the formerand the pronounced warp in the latter. Finally, recent radio observations with the VLAin 1.4, 4.9 and 8.4 GHz by Rosa-Gonz´alez et al. (2007) have shown that Mrk 36 has anearly flat radio spectrum dominated by thermal emission, similar to the regions detectedby Johnson & Kobulnicky (2003) in Henize 2-10. This may be an indication that the firstcontribution of the synchrotron emission to the low frequency emission, due to the firstsupernova (SN) explosions, has not yet appeared. These integrated properties show definiteevidence of a very young starburst. UM 408 is a compact galaxy with a projected size of ∼ ii region in previous studies (e.g., Gil de Paz et al. 2003), two giantregions in the central part of the galaxy were detected by Lagos et al. (2009), with ages of ∼ ∼ M ⊙ . Using GMOS-IFU observations Lagos et al. (2009)showed that the metal content in the ISM of this galaxy is well mixed and homogeneouslydistributed throughout the galaxy, in the same way as in other dwarf galaxies. UM 461 (PGC 037102) is a well studied H ii galaxy. This object has a very compact andbright off-center nucleus, some small regions spread along the galaxy (Noeske et al. 2003),and an external envelope that is strongly distorted towards the south-west. Doublier et al.(2001) and Noeske et al. (2003) have obtained integrated near-IR colors that differ signif-icantly from each other, with J-H=0.99 and H-K=-0.68 and J-H=0.47 mag and H-K=0.2mag, respectively. Taylor et al. (1995) proposed that UM 461 was formed together with UM462 by tidal interaction. But the HI maps of UM 461 and UM 462 do not show that thesegalaxies are tidally interacting, therefore it is unlikely that these objects induced the starformation on each other (Van Zee et al. 1998).Finally, the presence of He ii λ Broad-band J(1.25 µm ), H(1.65 µm ), K p (2.12 µm ), and narrow-band Br γ (2.17 µm ) imageswere obtained using the NIRI instrument on the Gemini North telescope on 2005 August 02(UM 408), November 24 (Mrk 36), and December 29 (UM 461). We used the f/6 camerawhich provides a field of view of ∼ ′′ × ′′ using the 1024 × ′′ .116 on side. The observations were performed under pho-tometric conditions. Table 1 lists all observational parameters and values adopted in thiswork. In this Table, Column (1) gives the object name. Columns (2) and (3) give the α and δ coordinates (J2000), respectively. Column (4) gives the observed heliocentric velocity(vel.) and the 3K CMB (Cosmic Microwave Background radiation) corrected distance fromNED. Columns (5) and (6) give the oxygen abundance and extinction c(H β ) for these galax-ies obtained from the literature. Column (7) gives the galactic extinction E(B-V) from theextinction maps of Schlegel et al. (1998). In Column (8) we present the date of observation.Column (9) shows the filter used in each observation and Columns (10) and (11) give theexposure time in seconds, considering the different coadd exposures and the mean air massof each observation, respectively. Finally, Column (12) shows the instrumental constants C λ of the transformation equation for flux calibration as described below in this section.The data were reduced following the standard procedures for near-IR imaging usingthe Gemini/NIRI package version 1.8 inside IRAF . For each filter, a normalized flat wasconstructed from the flat images observed with the calibration unit with the shutter closed(lamps off) and with the shutter open (lamps on). The bad pixel mask was constructedby identifying the bad pixels in the flat images with shutter off. The sky images wereconstructed from the raw science images by identifying all objects in each frame, maskedout, and averaging the remaining good pixels (images were observed with a dither offsetsof 10 ′′ and 20 ′′ for all galaxies). The raw science images were processed by subtractionthe sky on a frame-by-frame basis and dividing by the normalized flat field images. Thefinal flat-fielded, sky subtracted images were then registered to a common pixel position andmedian combined.In order to obtain a photometric calibration, we observed two standard stars: FS130(GSPC P264-F) and FS21 (GD140) for the observing run of Mrk 36, FS4 (SA93-317) andFS6 (Feige 22) for UM 408, and FS19 (G162-66, LTT 3870) and FS20 (G163-50, LTT 4099)for the observations of the galaxy UM 461. For each filter, the photometric calibration was IRAF: the Image Reduction and Analysis Facility is distributed by the National Optical AstronomyObservatories, which is operated by the Association of Universities for Research in Astronomy, In. (AURA)under cooperative agreement with the National Science Foundation (NSF). m λ = m λ i + χK λ + C λ , (1) m λ i = − . × log ( c/t ) , (2)where m λ is the magnitude in the standard system, m λ i is the instrumental magnitudewith c the number of counts and t the exposure time, χ the air mass, K λ the extinctioncoefficient and C λ the instrumental constant ( λ =J, H, K p ). The instrumental constant C λ was calculated as the average value from the standard stars. Given the limited numberof standard stars observed, we have used the average value of extinction to the “MaunaKea” observatory (MKO) K λ =0.1, 0.06 and 0.09 (Krisciunas et al. 1987) to the filters J, Hand K, respectively. The instrumental C λ constants used for each galaxy are listed in thelast column of Table 1. To compare the observed K p measures with evolutionary synthesismodels, we transformed the K p into K magnitudes using the relation K p = K + 0.22 × (H-K)(Wainscoat & Cowie 1992). We selected star clusters/complexes by visual inspection of theimages and using daofind in DAOPHOT. We fitted the detection threshold properly, with avalue of 2.5 σ above the local background, in order to detect sources in the starburst regionthat in the case of H ii /BCD galaxies spatially cover the visible extent of the galaxies. Allthe clusters or complexes in this study have been detected in the J, H and K p band. Thefinal catalog of objects is shown in § γ images were calibrated using a procedure similar to that described in Lagos et al.(2007). We convolved the stellar Spectral Energy Distribution (SED) with the response ofthe narrow-band filter. To do this we used the SED, appropriately scaled, of stars with thesame spectral type as our standard stars obtained from the literature. Finally, we subtractedthe continuum, estimating their contribution from the modeled SED for a given age of theregions obtained below in § ′′ , 6 ′′ , and 12 ′′ for the galaxies Mrk 36, UM 408, and UM 461, respectively.Foreground objects have been masked out. 8 – We quantified our detection limits and test the reliability of the derived magnitudesusing a series of completeness tests by adding artificial extended objects to our images. First,we created an empirical point spread function (PSF) for isolated point-like sources in theNIRI images. Given that the observed field of view (FOV) is relatively small ( ∼ ′′ × ′′ ),there is a possibility that these point-like sources are non-stellar. So these objects wereselected by eye and we ruled out sources with irregular PSF shapes. Using this informationwe added 20–50 artificial extended objects distributed in a regular square array. Models ofthese extended sources were generated using mksynth in BAOLAB (Larsen 1999) and wereadded to the science images. Magnitudes were randomly assigned to each position from 16.00mag to 23.00 mag with an interval of 0.5 mag. Finally, we compared the number of addedand recovered sources in each galaxy. The estimated completeness fraction for objects as afunction of magnitude for the K p band is shown in Figure 1. The completeness goes downto 20.2 mag in Mrk 36, 21.5 mag in UM 408 and 22.03 mag in UM 461 with ∼
90 % of theobjects recovered in the K p band.Fig. 1.— Completeness profiles (N rec /N add ) in the K p filter for star clusters in each of ourstudied galaxies, from the left to the right, Mrk 36, UM 408 and UM 461, respectively. Thedashed lines show the recovered fraction of 90% and 50%.
3. Results
Figure 2 shows the three galaxies observed and studied in this work (Mrk 36, UM 408,and UM 461) in the K p and Br γ bands, respectively. All galaxy images reveal the presenceof bright regions and/or star cluster complexes, given the apparent sizes of these regions in 9 –Figure 2, surrounded by LSB envelopes. Br γ emission is detected in all galaxies, but only thebrightest regions show clear evidence of intense emission given by our detection limit. Theunprecedented high spatial resolution images obtained for Mrk 36 allow us to identify, forthe first time using ground-based telescopes, the elementary structures within the starburst.Mrk 36 shows a plethora of small clusters, distributed over the entire extension of the galaxy,but with a previously unresolved concentration in the central knot. The Br γ emission showsa peak in this region, indicating the young nature of these clusters. We also observed thepresence of other groups of bright clusters in the K p image located in the northern region ofthe galaxy with weaker Br γ emission. Here we named these two groups of regions as complexI and complex II, respectively. The galaxy UM 408 appears quite compact and regular, andthe K p -band image shows that the star-forming regions are distributed in a clearly definedring-like geometry, composed of five almost regularly spaced regions where the current SFis occurring. The largest region is located in the eastern part of the galaxy. We observe thepresence of another region outside the central part of the galaxy. The Br γ emission in thisgalaxy is very regular and weak. Finally, the K p -band morphology of UM 461 is similar tothat reported in previous works (e.g., Noeske et al. 2003). The brightest region is off-centerand located in the eastern part of the galaxy. This region is extremely compact and brightin Br γ . We detected a number of regions spread throughout the body of the galaxy. For the color analysis we considered the objects obtained from our catalog in § Kp ∼
21 mag. We detected 33 regions in Mrk 36, 6 in UM 408 and 13 inUM 461, respectively. All regions identified in the K p -band images of the galaxies are shownin Figure 3. Given the seeing of ∼ . ′′ . ′′ ii /BCD galaxies (e.g., M82; Melo et al.2005), so these regions can be considered as individual star clusters, although likely some ofthese sources are blended. In any case, young clusters are not formed in isolation but ratherare found in cluster complexes (Zhang et al. 2001; Larsen 2004). Hence, the propertiesderived are the luminosity-weighted mean value of the complex. We have to bear in mind 10 –this caveat in our analysis but we are still able to derive the light weighted properties ofthese knots of SF and evaluate the clumpiness properties of the large scale mode of SF inthese galaxies. In Figure 3 we also show the size ( ∼
200 pc; Walborn 1991) of the wholenebular region of 30 Doradus to compare with the sizes of the clusters in Mrk 36.We measured the flux of the individual star cluster/complexes in all filters (includingBr γ ) using circular apertures with the program APER in IDL (an adapted version of the taskDAOPHOT in IRAF). For each aperture, we considered only pixels with 3 σ flux above thebackground. For each cluster we calculated the colors J-H and H-K (after the transformationof K p into K magnitudes). Figure 4 shows the color–color diagram (J-H vs. H-K) for all starclusters/complexes. In this figure we illustrate the evolutionary tracks of these colors usingSTARBURST99 (black line; Leitherer et al. 1999) and GALEV (orange line; Kotulla et al.2009) single stellar population (SSP) models, for a Kroupa IMF, and metallicity Z=0.004(more details in § γ in these regions. Finally, we compare our observed apparent magnitudes with the onesfound in the literature. The only one cluster/complex found in the literature is the region J =17.30 mag, that agree within theerrors with our value of m J =17.34 mag.Table 3 lists the photometric values for each star cluster/complex identified in oursample of galaxies (see Figure 3). In this Table, Column (2) shows the identification number.Column (3) shows the aperture considered to obtain the photometry in arcsec. Columns (4),(5), and (6) give the observed photometry in the J, H, and K p bands for each cluster,respectively. The † symbol over the identification number indicates that Br γ emission wasmeasured. Finally, Columns (7), (8), and (9) give the extinction E(B-V), age in units ofMyr, and the stellar mass in units of 10 M ⊙ of each star cluster or complex derived fromthe best fit comparison of the observed colors with chosen evolutionary synthesis models asdescribed below in § Absolute age dating of young star clusters in starburst galaxies has proven to be a verychallenging task, due to the additional emission by dust plus gas, as well as the effects ofextinction and metallicity. In relative terms the determination of ages and masses of the star 11 –cluster population within these galaxies can show some insight on the overall dominant modeof recent SF. In order to show these effects we derived the extinction and ages using ourcolor–color diagrams (see Figure 4) by comparing the near-IR colors for each star cluster orstar cluster complex with two independent SSP models: STARBURTS99 models (model I;this model includes pure stellar and nebular continuum) and GALEV (model II; this modelincludes stellar continuum, nebular continuum, and emission lines from warm ionized gas) foran instantaneous burst of SF. So we derived the ages by calculating the best fit of the models(using the chi-square method) to the observed colors for the model, varying the extinctionE(B-V) from 0 to 1 in steps of 0.05 mag. We used the Galactic extinction curve given byCardelli et al. (1989) assuming R V =3.1. The star cluster/complex masses were derived usingthe absolute magnitudes M k of each cluster as compared with the models. In the followingparagraphs we describe the main properties of the models used in this work. Model I : We considered the STARBURST99 model, which includes pure stellar and neb-ular continuum, for metallicity Z=0.004 ( ∼ ⊙ ), and Geneva evolutionary stellar tracks,assuming a Kroupa IMF ( ∝ M − α ) with α =1.3 for stellar masses between 0.1 to 0.5M ⊙ and α =2.3 for masses between 0.5 and 100M ⊙ for a total mass of 10 M ⊙ . More details about thephysics of the model in Leitherer et al. (1999). Model II : We compared our data with the GALEV evolutionary track for metallicityZ=0.004 and a Kroupa IMF (0.1-100M ⊙ ) including stellar continuum, nebular continuum,and gas emission. The GALEV models used in this work were kindly provided to us by RalfKotulla. These models were run using the Geneva evolutionary tracks with a minimum ageand time resolution of 0.1 Myr, unlike the models available on the GALEV website which usethe Padova isochrones and have a minimum age and time step of 4 Myr. The models also usea fraction of visible mass, which is used for the standard models available on the Web page,and which are thus twice as bright as the Padova models for the same mass. This does notaffect any of the colors, but we must keep this in mind when comparing masses. The flux ofthe hydrogen lines were computed using atomic physics and the production rate of ionizingphotons, whereas non-hydrogen line strengths are computed using metallicity-dependent lineratios relative to H β . More details about the input physics is given in Kotulla et al. (2009). Extinction, masses and cluster/complex ages
The two models (model I and II) displayed on the J-H versus H-K color–color diagram(Figure 4) show a different path for ages . ii /BCD galaxies show a clear excess in the near-IR, that is reflected in red H-Kcolors with respect to the models. This observed red excess had been reported earlier byVanzi et al. (2000, 2002); Hunt et al. (2003), Johnson et al. (2004), and more recently byReines et al. (2008) and Adamo et al. (2010, 2011) in the star cluster population of otherH ii /BCD galaxies. In fact Adamo et al. (2010) showed that the I and H bands are signifi-cantly affected by a similar excess. Reines et al. (2010) and Adamo et al. (2010) showed thatthis red excess clearly introduces a systematic offset in all of the derived parameters when allfilters, from UV-optical to near-IR are included in the determination of the observed SEDs.This suggests that the models commonly used in the literature to calculate the propertiesof the clusters, models that include pure stellar continuum, are inadequate for age-datingstudies (Papaderos et al. 1998; Anders & Fritze-v. Alvensleben 2003; Reines et al. 2010).Some of the causes analyzed in the literature that have been proposed to explain thisexcess are: i) there may be an important contribution from nebular continuum and line emis-sion (e.g., Papaderos et al. 1998; Vanzi et al. 2002; Reines et al. 2010), ii) hot dust emission(e.g., Vanzi et al. 2000; Reines et al. 2008), iii) the presence of young Red Super Giant(RSG) stars in older clusters, not properly modeled at low metalicity (Maeder & Meynet2001; V´azquez et al. 2007, and references therein), and iv) other sources as for exampleextended red emission (ERE; Witt & Vijh 2004), produced by photoluminescence processand the presence of a population of young stellar objects (YSOs; Adamo et al. 2011) stillsurrounded by circumstellar disks.Since there is no degeneracy in the color space between age and extinction at youngages (Figure 4), at least qualitatively, model II explains the colors of young star clusters(Reines et al. 2010) given that only this model is able to reproduce the trend seen in thedata, in Figure 4, at young ages.The presence of nebular continuum and emission lines in the near-IR can have a largeimpact on the inferred properties (affecting the mass determination) of the star clusters in 13 –H ii /BCD galaxies, and models that include this effect are the most appropriate in the studyof young stellar population with ages . γ , Br δ and Pa β ), and excited lines. The line emission detected by Vanzi et al. issufficient to produce their observed broad-band excess. Although hot dust and the presenceof RSGs may be important factors that can contribute to the near-IR excess, we assume thatthe excess in the SF regions in our sample of H ii galaxies is mainly produced by nebularcontinuum and emission lines, hence we can use our near-IR photometric bands and modelII, that include this contribution, in order to estimate the properties of the detected starclusters/complexes. In order to estimate the uncertainties in our age calculation procedure, we calculate thebest fit ages while varying the colors by their typical observational errors (as listed in Table2). In Figure 6 we show the ages obtained from model II for three different extinctions E(B-V)=0.0, 0.5 and 1.0 as a function of the mean age ( < age > ) obtained varying the colors bytheir errors. We showed that in the worst case (if all the errors conspire), they can modifythe solutions, but not changing significantly the properties of the sample. In this figure,we can see that for high extinction the clusters/complexes are better fitted than for lowerextinctions, producing that the best fits are obtained at higher values of reddening. This isan indication that the starburst regions or complexes are dominated by highly extinguishedand very young star clusters.A major concern in our method stems from the fact that we are using near-IR bandsto obtain the properties of the star cluster/complexes. The determination of absoluteages/masses of star clusters and star cluster complexes using near-IR colors, alone, couldbe highly uncertain and highly model dependent. It seems that better constraints of theproperties of the star clusters/complexes are obtained with a large broad band coverage plusemission line equivalent widths. However, we depend on the improvement of population syn-thesis models and the contraints of the input parameters and ingredients, as stellar tracks,libraries at different metallicities, and inclusion of the effects of nebular emission, dust, etc.Given these uncertainties, we cannot obtain absolute ages, but our analysis, based on theseobservations give us a qualitative view of a rather homogeneous star cluster population. Rel-ative ages and masses can, however, be useful in order to indicate the mode of SF at galacticscales.We note that future high-resolution ground-based or space telescope observations using 14 –UV-UBVRIJHK broad bands are needed in order to obtain the SED of each star clusteror/and complex in our sample of galaxies and then constrain absolute ages and quantify thenear-IR excess produced by other mechanism, e.g., hot dust, ERE, YSOs, RSGs in metalpoor ambients, etc. In summary, we found that the star cluster population in Mrk 36 is massive with es-timated masses of ∼ − M ⊙ . We detected, in this galaxy, a few clusters with masses of ∼ M ⊙ . Given our detection limits, lower mass clusters are likely not to be detectable.Meanwhile, the star clusters complexes in UM 408 and UM 461 have masses from ∼ M ⊙ to ∼ M ⊙ . The age distribution shows that the detected star clusters/complexes are veryyoung with ages less than ∼
10 Myr in Mrk 36 and ∼ > β ) map are displaced from the peaks of H α emission. The position of thepeaks of extinction found by Lagos et al. are correlated with the position of our detectedstar cluster complexes. This feature suggests that the current starburst episode is sweepingthe gas and dust out of the center into the surrounding regions.Finally, from a comparison of our results with the ages and masses of other star clustersin H ii /BCD galaxies studied in the literature, we see that the ranges are similar. Forexample, the analysis of the cluster population in UM 462 has revealed ages between 4.7and 10 Myr (Vanzi 2003) with masses range from 1.2 to 7.2 × M ⊙ . In the case of Haro 11Adamo et al. (2010) found that 30% of the clusters have masses > M ⊙ , arguing that theseclusters qualify as SSCs. Using HST Imaging Spectrograph (STIS) long-slit far- and near-ultraviolet spectra, Chandar et al. (2004) studied a local sample of SSCs in WR starburstgalaxies, including Mrk 36. They estimated an age < . ii galaxies studied here are characteristic of star clustercomplexes. High spatial resolution observations in space are needed in order to resolve theelementary entities that constitute the starburst population in UM 408 and UM 461. In a self-propagating SF model (Gerola & Seiden 1978) the gas expansion caused bystellar-wind and SNe shock waves will triger the next generation of stars. If this scenario ofSF is plausible, we must derive an age trend of the star clusters, or age sequence of somegroups of star clusters with respect to the spatial position. To illustrate this, we show inFigure 7 the position of the clusters in the galaxy Mrk 36. The star clusters, in this figure,are divided into two groups with masses ∼ M ⊙ and ∼ M ⊙ , considering three ranges inage: 1-5, 5-10 and > cross =R/v prop , where R is the size of the systemand v prop the propagation or expansion velocity. Typically, this velocity varies from ∼ − to ∼
100 kms − in dwarf galaxies (e.g., van Eymeren et al. 2007). Denoting by R thedistance between the central or brightest cluster or complex to the other clusters, we foundthat the propagation of SF is possible within the complexes in Mrk 36, for high expansionvelocities, given that the t cross < ∆age (with a mean t cross and ∆age ∼ cross > ∆age. For low velocities the difference betweenthe ages of the central region (in complex I) with respect to the more distant clusters (incomplex II) are ∆age ≪ t cross (with a mean t cross ∼
30 Myr in Mrk 36), suggesting that thepropagation of SF between the complexes is improbable.Given that the star clusters/complexes are practically coeval, the star formation wasproduced simultaneously within time scales of the order of ∆age on galactic scales. Morelikely a global SF mechanism is responsible for the present SF activity in galactic scales insome H ii galaxies, whereas, self propagating SF on scales of .
100 pc is still possible withinthe individual complexes. 16 –
In order to obtain the surface brightness profile, for each galaxy, we fitted ellipses tothe isophotes using the IRAF task ellipse . First, we manually masked the bright regionsidentified in the K p -band images, except the central ones, and replacing the values by theaverage of the adjacent regions. Then, we approximated the initial ellipse centers, ellipticities,and position angles, allowing these parameters to vary freely with radius during the fittingprocess. When the routine cannot proceed in the iterations as it reaches the lowest surfacebrightnesses, the ellipse task stops, and we fix the center, ellipticity, and the position anglewith these values to produce our light profiles.Figure 8 shows the surface brightness profiles (upper panels) of the isophotal meanintensity of each galaxy in the sample, considering pixels with 1 σ above the background, forthe J, H, K p and Br γ filters. Br γ is in arbitrary units. We measure the fainter level limits tobe µ J ≃ − , µ H ≃ − , and µ K p ≃ − for Mrk 36, UM 408, and UM 461, respectively.The J-H and H-K p color profiles are shown in the lower panels of Figure 8. We note that thecolor profiles show relatively constant underlying values at intermediate radii with a colorgradient at large radii, indicating that their stellar populations must be fairly homogeneousin the main body of the galaxy with the presence of a old stellar population at large radii.In Figure 9 we show the pixel-to-pixel color map of the galaxies in order to compare thecolor spatial distribution of these galaxies with the profiles previously obtained. Again, wenote that the color spatial distributions are relatively constant through the main body of thegalaxies. The surface photometry of the galaxies were corrected only for Galactic extinctionusing the relation with R V =A V /E(B-V)=3.1 (Cardelli et al. 1989) and adopting the valuesfrom the extinction map of Schlegel et al. (1998) showed in Table 1 Column (7). We did notapply any smoothing procedure to our data in order to obtain the surface brightness at largeradii.We see that the outer part of the light profiles in Figure 8 are well represented by anexponential model (e.g., Papaderos et al. 1996; Telles & Terlevich 1997; Cair´os et al. 2003;Noeske et al. 2003, 2005) thus other functions are not necessary. We can express this profilein terms of the surface brightness, so we can find that µ ( r )= µ o,λ +(1.086/ α λ )r, where µ o,λ corresponds to the central surface brightness and α , ( J,H,K p ) the scale length with λ =J, Hand K p . Column (1) of Table 4 gives the name of the galaxies. Columns (2), (3) and (4) givethe µ , ( J,H,K p ) , α , ( J,H,K p ) parameters and m LSB, ( J,H,K p ) magnitudes of the LSB componentdescribed below in § § The total apparent magnitude of the LSB component can be obtained using that m
LSB = µ - 5 log( α ) - 1.995 - 2.5 log(1- ε ), where ε correspond to the ellipticity derived from thesurface brightness profile (with constant values of ε =0.40, 0.30 and 0.33 for the galaxies Mrk36, UM 408 and UM 461, respectively). So, we calculated the J-H and H-K colors of theLSB component from the results of the exponential fits. The colors of the LSB component inMrk 36 are similar to those obtained previously in the literature by Cair´os et al. (2003) withJ-H=0.37 and H-K s =0.28 from exponential fits. For UM 461, Noeske et al. (2003) foundJ-H and H-K s colors, of the LSB component, that disagree and agree with our values at1 σ level, respectively. The reason for this disagreement could be the uncertainties in thesky estimation and/or the aperture differences in both studies causing the surface brightnessprofile obtained by Noeske et al. (2003) to be deeper with the radius compared with ourprofiles, thus producing higher slopes of the profiles at large radii. In summary, the LSBcomponent of our sample of galaxies show blue J-H colors and red H-K that suggest aphotometrically dominant stellar population of ages & yr. Our interpretation agrees withthe age estimates given by Raimann et al. (2000) and Westera et al. (2004) for a sample ofH ii galaxies (in which they include UM 408 and UM 461 in their analysis) by means ofspectral population synthesis.Finally, we estimated the stellar mass (M ∗ ) of the LSB component for each galaxy usingthe M / L relationship for the H-band (see L´opez-S´anchez 2010). We found that the stellarmass for our sample of galaxies are log(M ∗ )= 7.89, 8.33, and 7.88 M ⊙ for Mrk 36, UM 408,and UM 461, respectively. These stellar masses are typical for BCD galaxies calculated usingdifferent photometric bands, and dominated by the contribution of the intermediate to oldstellar population (e.g., 3 × M ⊙ for Mrk 36; Amor´ın et al. 2009).
4. Discussion
Star formation occurs when the local molecular gas density exceeds a certain threshold(see Leroy et al. 2008, and references therein), given first possibly by collisions of gas clouds,due to turbulence or gravity, and then triggered by the action of massive star evolution, e.g.stellar winds from star clusters and SNe. In many cases, the current star-formation may betriggered by external agents, as in the case of the tidal forces by a neighboring galaxy or 18 –mergers. However, for a fraction of isolated, less luminous and compact H ii galaxies, thisstar-formation activity may occur solely by internal processes.The SF activity may form a fraction of bound star clusters that will survive their infantmortality and evolve. Bastian (2008) has defined the value of the present cluster formationefficiency Γ in a host galaxy as CFR/SFR, where CFR = M tot /∆t is the present clusterformation rate at which the galaxy produces a total cluster mass M tot in a given age interval∆t. On the other hand, Goddard et al. (2010) found a correlation between the value of thepresent cluster formation efficiency and the SF density of the host galaxy in a sample ofstarburst galaxies, Γ(%)=29Σ . SF R M ⊙ yr − kpc − , where Σ SF R is the total SFR per unit ofarea.The first step in the determination of the star cluster formation efficiency in our sampleof galaxies is the determination of the total SF rate (SFR). We calculate the total SFR usingthe Kennicutt (1998) relationship for L(H α ). We use the integrated H β flux obtained usingcontinuum-free emission line broad band images, with L(H α )/L(H β )=2.87, by Lagos et al.(2007) for Mrk 36 and UM 461, and integrated integral field unit H α flux obtained byLagos et al. (2009) for UM 408. So we find that the total SFR(H α ) ≃ ⊙ yr − in Mrk36, 0.017 M ⊙ yr − in UM 408, and 0.085 M ⊙ yr − in UM 461. The SFR in our sampleof galaxies is very low compared with other dwarf galaxies, such as Haro 11 (22 M ⊙ yr − ;Adamo et al. 2010) and NGC 1519 (0.3626 M ⊙ yr − ; Goddard et al. 2010) and in fact withrespect to more irregular and luminous H ii /BCD galaxies, such as Tol 9 with a SFR(H α )=1.82 M ⊙ yr − (L´opez-S´anchez 2010). Assuming a starburst area of ∼ in Mrk 36, ∼ in UM 408 and ∼ in UM 461, we found that the integrated SFR perunit of area in Mrk 36 is Σ SF R =0.039 M ⊙ yr − kpc − , in UM 408 is Σ SF R =0.002 M ⊙ yr − kpc − and in UM 461 is Σ SF R =0.052 M ⊙ yr − kpc − , respectively. We can now calculatethe expected value of Γ using the correlation found by Goddard et al. (2010). Thus, weobtain that the cluster formation efficiency is ∼
13% in Mrk 36, ∼
7% in UM 408 and ∼ ii galaxiesis approximately ∼ ∼ × M ⊙ in Mrk36 (with a total mass in clusters & M ⊙ of 44.69 × M ⊙ ), 13.81 × M ⊙ in UM 408 and We multiplied the SFR by a factor 0.67 obtained from the comparison between the SFR of our galaxiesusing the Kennicutt (1998) and Calzetti et al. (2007) relationships This due to differences in the stellar IMFassumptions given that the Kennicutt (1998) relationship is based on a Salpeter IMF and the models whichwe adopted are based on a Kroupa IMF
19 –43.09 × M ⊙ in UM 461. So, we calculate that the fraction of the total cluster mass withrespect to the LSB host galaxy mass in Mrk 36, considering our complete range in ages, isequal to ∼ ∼ ∼ .
1% of the underlying galaxy mass, in agreementwith the estimate of Westera et al. (2004) that the past history of SF in the galaxies weremore active than the present one. Additionally, we calculate the SFR per unit of area for theindividual star clusters/complexes using our measured Br γ emission. From these, we obtainthat in the brightest regions (clusters > ⊙ yr − kpc − (Bastian et al. 2005, and references therein). While, the majority ofthe star clusters/complexes have values of the order of ∼ ⊙ yr − kpc − . If we use onlythe total mass in clusters & M ⊙ to calculate SF efficiency, we obtain that in Mrk 36 foran age of 20 Myr our result agree, within the uncertainties, with the one obtained usingthe relationship of Goddard et al. (2010). However, in the case of UM 408 and UM 461 thevalue of Γ is extremely high, indicating that in these cases the star forming complexes arenot resolved into individual clusters, resulting in a overestimation of the Γ parameter.We can also calculate the gas consumption time scale ( τ gas ) or how long it would takebefore all the gas in the galaxies will be consumed at the current SFR. The consumption timescale is defined as the ratio between the available gas and the current SFR, τ gas = M gas /SFR.Assuming the total amount of gas M gas = M HI +M He +M H ≈ × M HI (Leroy et al. 2005),we obtain that τ gas ∼ HI =2.0 × M ⊙ ; Thuan & Martin 1981;Bravo-Alfaro et al. 2004), τ gas ∼
77 Gyr in UM 408 (with M HI =6.53 × M ⊙ ; Salzer et al.2002) and τ gas ∼ HI =0.98 × M ⊙ ; Smoker et al. 2000). Theconsumption timescales in Mrk 36 and UM 461 are significantly less than a Hubble time andcomparable with the times observed in spiral galaxies (Kennicutt et al. 1994). So the SFcannot be sustained for the entire history of the galaxies, which indicates that these objectsundergo a few or several short bursts of SF. While UM 408 has a gas consumption timescalelonger than a Hubble time, indicating that likely the SF is relatively constant through thehistory of the galaxy or the HI halo may not be spatially available for the current SF.Our findings also seem to indicate that the SF mode in our sample of low luminosityH ii galaxies is clumpy, similar to other dwarf galaxies (e.g., NGC 1569 and SBS 0335-052).These complexes or star-forming knots are formed by a few massive star clusters with masses & M ⊙ and high SFR per unit of area. Melena et al. (2009) found that the observed trendsin the number and mass of the SF regions, in a sample of dwarf galaxies, is independent ofthe local environment and even the surrounding galaxy mass and is given by a mass functionwhich stochastically favored SF in clusters (Adamo et al. 2010). Billett et al. (2002) showthat SSCs may require special (or fortunate) circumstances to form in dwarf galaxies, but 20 –when they do, they are very massive ( & − M ⊙ ) and form clumps or groups of similarages. This suggests that these clumps are likely formed in localized regions of high pressuretriggered by large scale ambient gravitational instabilities, given that in dwarf galaxies mostof the ISM is at low pressure (Elmegreen & Hunter 2000). The lack of external perturbers inthe most compact and isolated galaxies indicates that an additional mechanism other thantidal interactions must be considered to explain this current SF activity. This mechanismmay be related to the overall physical conditions of the ISM, particularly the gas surfacedensities, in conjunction with stochastic effects, that allow SF to take place. Alternatively,the low current SFR implies that a burst or a triggering is not necessary, simply that the SFRhas been relatively constant. However, tidal interactions or mergers are likely the primaryagent to trigger the current SF in luminous and more disturbed H ii galaxies as suggested bytheir morphology (e.g., Telles & Terlevich 1995; Lagos et al. 2007; L´opez-S´anchez & Esteban2008).
5. Conclusions
In this paper, a sample of three H ii galaxies (Mrk 36, UM 408, and UM 461) hasbeen analyzed in order to study their stellar populations (star cluster complexes and theunderlying host galaxy or LSB component) using new near-IR high spatial resolution imagesobtained on the Gemini North telescope. In our analysis we used models that include thecontribution of stellar continuum, nebular continuum and emission lines. Our conclusionscan be summarized as follows:1. The presence of nebular continuum and emission lines in the near-IR produces anexcess in the observed SED in young star cluster/complexes. This excess, can have alarge impact in the inferred properties of the star clusters in H ii /BCD galaxies andmodels that include this effect are the most appropriate in the study of young stellarpopulation with ages . ∼ − M ⊙ with afew detected star clusters with masses of ∼ M ⊙ distributed in the main body of thegalaxy. The star cluster complexes in UM 408 and UM 461 have masses from ∼ M ⊙ to ∼ M ⊙ . The age distribution shows that the detected star clusters/complexes arevery young with ages of a few Myr. Two likely old star clusters with colors consistentwith ages > M ⊙ is compared to the currenttotal SFR is about 10% in our sample of galaxies. 21 –3. The spatial distribution and ages of the star cluster/complex population seems to in-dicate that SF is clumpy and simultaneous. We propose that the current SF activityin our sample of low luminosity H ii galaxies is triggered by some internal mechanisminstead of tidal interactions. This mechanism of SF may be related to the overall phys-ical conditions of the ISM that produce the increase of surface densities in conjunctionwith stochastic effects within a time scale comparable to the mean age differences ofthe massive star cluster complexes.4. The LSB component of our sample of galaxies have near-IR colors representative ofevolved stellar population of at least & yr. We found that the stellar mass of thiscomponent for our sample of galaxies are log(M ∗ )= 7.89, 8.33, and 7.88 M ⊙ for Mrk 36,UM 408, and UM 461, respectively. The fraction of the total cluster mass with respectto the LSB hosting galaxy mass in our sample of galaxies, considering our completerange in ages, is less than 1%.P.L. is supported by a Post-Doctoral grant (SFRH/BPD/72308/2010), funded by FCT/MCTES(Portugal) and POPH/FSE (EC). P.L. would like thank to Polychronis Papaderos andDamian Fabbian for their comments and very useful discussions. A.N. acknowledges supportby the projects: AYA2007-67965-C03-02, AYA2010-21887-C04-01 and Consolider-Ingenio2010 CSD2006-00070 First Science with GTC, of the spanish MICINN. We thank RalfKotulla for providing us with the GALEV models used in this work. We also thank theanonymous referee for numerous useful comments and suggestions which led to the overallimprovement of this paper. Based on observations obtained at the Gemini Observatory,which is operated by the Association of Universities for Research in Astronomy, Inc., undera cooperative agreement with the NSF on behalf of the Gemini partnership: the NationalScience Foundation (United States), the Science and Technology Facilities Council (UnitedKingdom), the National Research Council (Canada), CONICYT (Chile), the Australian Re-search Council (Australia), Minist´erio da Ciˆencia e Tecnologia (Brazil) and Ministerio deCiencia, Tecnolog´ıa e Innovaci´on Productiva (Argentina). Program ID: GN-2005B-Q-42.This research has made use of the NASA/IPAC Extragalactic Database (NED) which isoperated by the Jet Propulsion laboratory, California Institute of technology, under contractwith the National Aeronautics and Space Administration. Facilities:
Gemini:North (NIRI) 22 –
REFERENCES
Adamo, A., ¨Ostlin, G., Zackrisson, E., Hayes, M., Cumming, R. J., & Micheva, G. 2010,MNRAS, 407, 870Adamo, A., ¨Ostlin, G., Zackrisson, E., & Hayes, M. 2011, MNRAS, 414, 1793Amor´ın, R., Aguerri, J. A. L., Mu˜noz-Tu˜n´on, C., & Cair´os, L. M. 2009, A&A, 501, 75Anders, P., & Fritze-v. Alvensleben, U. 2003, A&A, 401, 1063Bastian, N., Gieles, M., Efremov, Yu. N., & Lamers, H. J. G. L. M. 2005, A&A, 443, 79Bastian, N. 2008, MNRAS, 390, 759Bergvall, N., & ¨Ostlin, G. 2002, A&A, 390, 891Billett, O. H., Hunter, D. A., & Elmegreen, B. G. 2002, AJ, 123, 1454Bravo-Alfaro, H., Brinks, E., Baker, A. J., Walter, F., & Kunth, D. 2004, AJ, 127, 264Cair´os, L. M., Caon, N., Papaderos, P., Noeske, K., V´ılchez, J, M., Lorenzo, B. G., &Mu˜noz-Tu˜n´on, C. 2003, ApJ, 593, 312Calzetti, D. et al. 2007, ApJ, 666, 870Chandar, R., Leitherer, C., & Tremonti, C. A. 2004, ApJ, 604, 153Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245Conti, P. S. 1991, ApJ, 377, 115Conti, P. S., & Vacca, W. D. 1994, ApJ, 423, 97Doublier, V., Caulet, A., & Comte, G. 2001, A&A, 367, 33Elmegreen, B. G., & Efremov, Y. N. 1997, ApJ, 480, 235Elmegreen, B. G., & Hunter, D. A. 2000, ApJ, 540, 814Garcia-Vargas, M. L., Bressan, A., & Diaz, A. I. 1995, A&AS, 112, 13Gerola, H., & Seiden, P. E. 1978, ApJ, 223, 129Gil de Paz, A., Madore, B. F., & Pevunova, O. 2003, ApJS, 147, 29Gil de Paz, A., & Madore, B. F. 2005, ApJS, 156, 345 23 –Goddard, Q. E., Bastian, N., & Kennicutt, R. C. 2010, MNRAS, 405, 857Harris, W. E. 1991, ARA&A, 29, 543Hunt, L. K., Thuan, T. X., & Izotov, Y. I. 2003, ApJ, 588, 281Hunter, D. A. & Elmegreen, B. G. 2004, AJ, 128, 2170Hunter, D. A., & Elmegreen, B. G. 2006, ApJS, 162, 49Izotov, Y. I., & Thuan, T. X. 1998, ApJ, 500, 188Johnson, K. E., & Kobulnicky, H. A. 2003, ApJ, 597, 923Johnson, K. E., Indebetouw, R., Watson, C., & Kobulnicky, H. A. 2004, AJ, 128, 610Kennicutt, R. C., Tamblyn, P., & Congdon, C. W. 1994, ApJ, 435, 22Kennicutt, R. C. 1998, ARA&A, 36, 189Kotulla, R., Fritze, U., Weilbacher, P., & Anders, P. 2009, MNRAS, 396, 462Krisciunas, K., Sinton, W., Tholen, K., Tokunaga, A., Golisch, W., Griep, D., Kaminski,C., Impey, C., & Christian, C. 1987, PASP, 99, 887Lagos, P., Telles, E., & Melnick, J., 2007, A&A, 476, 89Lagos, P., Telles, E., Mu˜noz-Tu˜n´on, C., Carrasco, E., Cuisinier, F., & Tenorio-Tagle, G.2009, AJ, 137, 5068Larsen, S. S. 1999, A&AS, 139, 393Larsen, S. S. 2004, A&A, 416, 537L´opez-S´anchez, ´A. R., & Esteban, C. 2008, A&A, 491, 131L´opez-S´anchez, ´A. R. 2010, A&A, 521, 63Leitherer, C., Schaerer, D., Goldader, J. D., Delgado, R. M. G., Robert, C., Kune, D. F., deMello, D. F., Devost, D., & Heckman, T. M. 1999, ApJS, 123, 3Leroy, A., Bolatto, A. D., Simon, J. D., & Blitz, L. 2005, ApJ, 625, 763Leroy, A. K., Walter, F., Brinks, E., Bigiel, F., de Blok, W. J. G., Madore, B., & Thornley,M. D. 2008, AJ, 136, 2782 24 –Massey, P., & Hunter, D. A. 1998, ApJ, 493, 180Maeder, A., & Meynet, G. 2001, A&A, 373, 555Melena, N.W., Elmegreen, B. G., Hunter, D. A., & Zernow, L. 2009, AJ, 138, 1203Melo, V. P., Mu˜noz-Tu˜n´on, C., Maiz-Apellaniz, J., & Tenorio-Tagle, G. 2005, ApJ, 619, 270Meurer, G. R., Heckman, T. M., Leitherer, C., Kinney, A., Robert, C., & Garnett, D. R.1995, AJ, 110, 2665Noeske, K. G., Iglesias-P´aramo, J., Vilchez, J. M., Papaderos, P., & Fricke, K. J. 2001, A&A,371, 806Noeske, K. G., Papaderos, P., Cair´os, L. M., & Fricke, K. J. 2003, ApJ, 410, 481Noeske, K. G., Papaderos, P., Cair´os, L. M., & Fricke, K. J. 2005, ApJ, 429, 115O’Connell, R. W., Gallagher, J. S., & Hunter, D. A. 1994, ApJ, 433, 65OPapaderos, P., Loose, H. H., Thuan, T. X., & Fricke, K. J 1996, A&AS, 120, 207Papaderos, P., Izotov, Y. I., Fricke, K. J., Thuan, T. X., & Guseva, N. G. 1998, A&A, 338,43Pelupessy, F. I., van der Werf, P. P., & Icke, V. 2004, A&A, 422, 55Pustilnik, S. A., Kniazev, A. Y., Lipovetsky, V. A., & Ugryumov, A. V. 2001, A&A, 373, 24Raimann, D., Storchi-Bergmann, T., Bica, E., Melnick, J., & Schmitt, H. 2000, MNRAS,316, 559Reines, A. E., Johnson, K. E., & Hunt, L. K. 2008, AJ, 136, 1415Reines A. E., Nidever D. L., Whelan D. G., & Johnson K. E. 2010, ApJ, 708, 26Rosa-Gonz´alez, D., Schmitt, H. R., Terlevich, E., & Terlevich, R. 2007, ApJ, 654, 226Sargent, W. L. W., & Searle, L. 1970, ApJ, 162, 155Salzer, J. J., Rosenberg, J. L., Weisstein, E. W., Mazzarella, J. M., & Bothun, G. D. 2002,ApJ, 124, 191Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525Smoker, J. V., Davies, R. D., Axon, D. J., & Hummel, E. 2000, A&A, 361, 19 25 –Simpson, C. E. et al. 2011, AJ, 142, 82Taylor, C. L. 1997, ApJ, 480, 524Taylor, C. L., Brinks, E., Grashuis, R. M., & Skillman, E. D. 1995, ApJS, 99, 427Telles, E. 1995, Ph.D thesis, Univ. CambridgeTelles, E., & Terlevich, R. 1995, MNRAS, 275, 1Telles, E., & Terlevich R. 1997, MNRAS, 286, 183Telles, E., Melnick, J., & Terlevich, R. 1997, MNRAS, 288, 78Telles, E., & Maddox, S. 2000, MNRAS, 311, 307Telles, E. 2009, arXiv:0908.2966v1Thuan, T. X., Izotov, Y. I., & Lipovetsky, V. A. 1997, ApJ, 477, 661Thuan, T. X., & Martin, G. E. 1981, ApJ, 247, 823Thuan, T. X. 1983, ApJ, 268, 667Van Zee, L., Skillman, E. D., & Salzer, J. J. 1998, AJ, 116, 1186van Eymeren, J., Bomans, D. J., Weis, K., & Dettmar, R.-J. 2007, A&A, 474, 67Vanzi, L., Hunt, L. K., Thuan, T. X., & Izotov, Y. I. 2000, ApJ, 363, 493Vanzi, L., Hunt, L. K., & Thuan, T. X. 2002, ApJ, 390, 481Vanzi, L. 2003, ApJ, 408, 523V´azquez, G. A., Leitherer, C., Schaerer, D., Meynet, G., & Maeder, A. 2007, ApJ, 663, 995Wainscoat, R. J., & Cowie, L. L. 1992, AJ, 103, 332Walborn, N. R. 1991, IAUS, 148, 145Walborn, N. R., Ma´ız-Apell´aniz, J., & Barb´a, R. H. 2002, AJ, 124, 1601Westera, P., Cuisinier, F., Telles, E., & Kehrig, C. 2004, A&A, 423, 133Whitmore, B. C., Zhang, Q., Leitherer, C., Fall, S. M., Schweizer, F., & Miller, B. W. 1999,AJ, 118, 1551 26 –Witt, A. N., & Vijh, U. P. 2004, Astrophysics of Dust, 309, 115Zhang, Q., Fall, S. M., & Whitmore, B. C. 2001, ApJ, 561, 727
This preprint was prepared with the AAS L A TEX macros v5.2.
27 –Fig. 2.— K p and Br γ images (not continuum subtracted) of the galaxies Mrk 36 (top), UM408 (middle) and UM 461 (bottom) with NIRI at Gemini North. The field of view of eachimage is 48 ′′ × ′′ . The images are displayed on a logarithmic scale. The arrows in the K p image of Mrk 36 point to three background galaxies close to the galaxy. North is at the topand east to the left. 28 –Fig. 3.— Images of Mrk 36, UM 408, and UM 461 in the K p filter. Circles mark theposition (and the apertures used for photometry) of each star cluster/complex identified inthe galaxies. The clusters have been labeled with the designation used throughout the paper(see Table 3). In the upper panel of Mrk 36 we show the position of each cluster identifiedin the northern region of the galaxy. The circle at the corner of this panel represents the sizeof the nebular region of 30 Doradus (Walborn 1991). The lower panel shows the clusters inthe central region of Mrk 36 (dashed square in the upper panel). 29 –Fig. 4.— Observed near-IR color–color diagram J-H vs. H-K for the star clusters/complexesin our sample not corrected for extinction. The lines show the evolutionary tracks of thesecolors from STARBURST99 (black line; this model includes stellar and nebular continuum)and GALEV models (orange line; this model includes stellar continuum, nebular continuumand the contribution of nebular emission lines) for a metallicity Z=0.004. Red trianglesrepresent the observed values for clusters in Mrk 36, the blue stars represent the valuesfor UM 408 and green circles represent the values for clusters/complexes in UM461. Filledsymbols indicate the detection of Br γ in the regions. The error bars show the average errorvalue for the colors. Open circles along the tracks indicate ages of 1, 3, 6, 10, and 100 Myr.An additional circle indicating an age of 1 Gyr is included in the model of STARBURST99. 30 –Fig. 5.— Extinction, age and mass distribution of our sample of star clusters/complexes,obtained using the STARBURST99 (model I; black distribution) and GALEV (model II;orange distribution) models. For more details see § σ uncertainties. 32 –Fig. 7.— Spatial distribution of the star clusters in Mrk 36. The different circle sizescorrespond to masses ∼ M ⊙ and ∼ M ⊙ considering three ranges in age: 1-5, 5-10 and > p (corrected for galactic extinction) and Br γ in arbitrary units. We considered pixelswith 1 σ above the background. For a better visualization, the H and K p profiles are shiftedby -1 and -2 mag, respectively. Bottom panels: J-H and H-K p color profiles. 34 –Fig. 9.— J-H and H-K color maps of the galaxies Mrk 36 (top), UM 408 (middle), and UM461 (bottom). Black contours in the J-H maps corresponds with J-band contour maps andin the H-K maps corresponds with K p -band contour maps. Orange contour are from Br γ images. Table 1. Sample data and NIRI observations.
Object α δ V el. β ) E(B-V) Gal
Date of Filter Exposure Air mass C λ (J2000) (J2000) (km/s) Observation Time (s) (mag)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)Mrk 36 11:04:44.0 +29:07:48 646 7.81 a a × (3 ×
60) 1.37 23.93 ± × (4 ×
65) 1.27 24.01 ± p × (3 ×
75) 1.15 23.57 ± γ × (3 ×
40) 1.08UM 408 02:11:23.4 +02:20:30 3598 7.87 b b × (3 ×
60) 1.26 23.91 ± × (4 ×
65) 1.19 23.97 ± p × (3 ×
75) 1.10 23.49 ± γ × (3 ×
40) 1.06UM 461 11:51:33.1 -02:22:22 1039 7.78 a a × (3 ×
60) 1.62 23.95 ± × (4 ×
65) 1.46 24.05 ± p × (3 ×
75) 1.28 23.54 ± γ × (3 ×
40) 1.18 a Izotov & Thuan (1998) b Lagos et al. (2009)Note. — Column (1) galaxy name. Columns (2) and (3) α and δ coordinates (J2000), respectively. Column (4) heliocentric velocity (vel.) and the 3KCMB corrected distance from NED. Columns (5) and (6) oxygen abundance and extinction adopted in this work. Column (7) date of observation. Column(8) Galactic extinction obtained from the extinction maps of Schlegel et al. (1998). Column (9) filter used in each observation and finally columns (10) and(11) exposure time in s and the mean air mass of each observation, respectively. Column (12) shows the instrumental zero points C λ with λ = J, H and K p obtained for each run of observation.
36 –Table 2: Integrated magnitudes of our sample of galaxies.Object J H K p (mag) (mag) (mag)(1) (2) (3) (4)Mrk 36 14.46 ± ± ± ± ± ± ± ± ± Table 3. Observed aperture photometry in the near-IR bands J, H and K p and measured properties of the stellarclusters. Name Cluster Aperture radii J H K p E(B-V) Age Mass Log L(Br γ )(arcsec) (mag) (mag) (mag) (mag) (Myr) ( × M ⊙ ) (erg s − )(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)Mrk 36 1 † ± ± ± ± ± ± · · · † ± ± ± † ± ± ± † ± ± ± † ± ± ± † ± ± ± † ± ± ± ± ± ± · · ·
10 0.28 19.99 ± ± ± ∼ ± ± ± · · · † ± ± ± ± ± ± · · ·
14 0.32 20.62 ± ± ± · · ·
15 0.43 20.97 ± ± ± · · ·
16 0.38 21.10 ± ± ± · · · † ± ± ± ± ± ± · · · † ± ± ± † ± ± ± ± ± ± · · ·
22 0.27 21.19 ± ± ± · · ·
23 0.41 20.28 ± ± ± · · · † ± ± ± † ± ± ± † ± ± ± ± ± ± ∼ · · · † ± ± ± † ± ± ± ± ± ± · · · † ± ± ± ± ± ± · · · Table 3—Continued
Name Cluster Aperture radii J H K p E(B-V) Age Mass Log L(Br γ )(arcsec) (mag) (mag) (mag) (mag) (Myr) ( × M ⊙ ) (erg s − )(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)33 † ± ± ± ± ± ± · · · † ± ± ± † ± ± ± † ± ± ± † ± ± ± ± ± ± · · · UM 461 1 0.66 19.87 ± ± ± · · · † ± ± ± ± ± ± · · · ± ± ± · · · ± ± ± · · · ± ± ± · · · ± ± ± · · · ± ± ± · · · ± ± ± · · ·
10 0.66 20.14 ± ± ± · · ·
11 0.61 20.34 ± ± ± · · ·
12 0.53 20.48 ± ± ± · · ·
13 0.60 20.60 ± ± ± · · · Note. — Column (1) galaxy name. Column (2) identification number of the star clusters/complexes. The † symbol indicates that Br γ emission were measured in the cluster. Column (3) aperture considered to obtain the photometry. Columns (4), (5) and (6) observedphotometry in magnitude of each star cluster/complex in the filters J, H and K p , respectively. Columns (7), (8) and (9) extinction E(B-V),age in units of Myr and the mass in units of M ⊙ for models I and II, respectively. Finally Colum (10) give the Br γ luminosity.
39 –Table 4: Structural parameters from the exponential fits to the host galaxies of Mrk 36, UM408 and UM 461. Column (1) galaxy name. Columns (2), (3) and (4) µ , ( J,H,K p ) , α , ( J,H,K p ) parameters and m LSB, ( J,H,K p ) magnitudes.Object µ ,J µ ,H µ ,K p (mag arcsec − ) (mag arcsec − ) (mag arcsec − ) α ,J α ,H α ,K p (arcsec) (arcsec) (arcsec)m LSB,J m LSB,H m LSB,K pp