Neutral gas in Lyman-alpha emitting galaxies Haro 11 and ESO 338-IG04 measured through sodium absorption
A. Sandberg, G. Östlin, M. Hayes, K. Fathi, D. Schaerer, J.M. Mas-Hesse, T. Rivera-Thorsen
aa r X i v : . [ a s t r o - ph . GA ] M a r Astronomy&Astrophysicsmanuscript no. sandbergetal2013 c (cid:13)
ESO 2018October 16, 2018
Neutral gas in Lyman-alpha emitting galaxies Haro 11 and ESO338-IG04 measured through sodium absorption ⋆ A. Sandberg , G. ¨Ostlin , M. Hayes , , K. Fathi , D. Schaerer , , J.M. Mas-Hesse , T. Rivera-Thorsen The Oskar Klein Centre, Department of Astronomy, Stockholm University, AlbaNova, 106 91 Stockholm, Sweden.e-mail: [email protected] Universit´e de Toulouse; UPS-OMP; IRAP; Toulouse, France CNRS; IRAP; 14, avenue Edouard Belin, F-31400 Toulouse, France Observatoire de Gen`eve, Universit´e de Gen`eve, 51 Ch. des Maillettes, 1290 Versoix, Switzerland Centro de Astrobiolog´ıa (CSIC–INTA), Madrid, SpainPreprint online version: October 16, 2018
ABSTRACT
Context.
The Lyman alpha emission line of galaxies is an important tool for finding galaxies at high redshift, and thus probe thestructure of the early universe. However, the resonance nature of the line and its sensitivity to dust and neutral gas is still not fullyunderstood.
Aims.
We present measurements of the velocity, covering fraction and optical depth of neutral gas in front of two well known localblue compact galaxies that show Lyman alpha in emission: ESO 338-IG 04 and Haro 11. We thus test observationally the hypothesisthat Lyman alpha can escape through neutral gas by being Doppler shifted out of resonance.
Methods.
We present integral field spectroscopy from the GIRAFFE / Argus spectrograph at VLT / FLAMES in Paranal, Chile. The ex-cellent wavelength resolution allows us to accurately measure the velocity of the ionized and neutral gas through the H α emission andNa D absorption, which traces the ionized medium and cold interstellar gas, respectively. We also present independent measurementswith the VLT / X-shooter spectrograph which confirm our results.
Results.
For ESO 338-IG04, we measure no significant shift of neutral gas. The best fit velocity is -15 ±
16 km / s. For Haro 11, we seean outflow from knot B at 44 ±
13 km / s and infalling gas towards knot C with 32 ±
12 km / s. Based on the relative strength of the Na Dabsorption lines, we estimate low covering fractions of neutral gas (down to 10%) in all three cases. Conclusions.
The Na D absorption likely occurs in dense clumps with higher column densities than where the bulk of the Ly α scattering takes place. Still, we find no strong correlation between outflowing neutral gas and a high Lyman alpha escape fraction.The Lyman alpha photons from these two galaxies are therefore likely escaping due to a low column density and / or covering fraction. Key words. galaxies: kinematics and dynamics – galaxies: ISM – galaxies: starburst – galaxies: individual: ESO338-IG04 and Haro11
1. Introduction
The Lyman alpha (Ly α ) emission line was suggested asa probe for discovering high redshift galaxies already byPartridge & Peebles (1967). When absorbed by neutral hydro-gen gas, approximately two thirds of the ionizing photons fromhot, massive stars are reprocessed into Ly α photons follow-ing case B recombination. The fraction of the bolometric fluxcontained in the Ly α line should be as high as 6-7 per centin a young, star-forming stellar population. However, early sur-veys designed for finding Lyman alpha emitting galaxies (LAEs)came up blank. (see e.g. Pritchet 1994, and references therein).It was only with deeper and larger surveys (e.g. Cowie & Hu1998) that targeting Ly α became the successful methodthat it is today for finding galaxies at redshifts z &
2. It has since then been widely used for mapping outthe large-scale structure of the high redshift universe (e.g.Rhoads et al. 2000; Kudritzki et al. 2000; Malhotra & Rhoads2002; Ouchi et al. 2003, 2005; Gawiser et al. 2006; Ajiki et al.2006; Gronwall et al. 2007; Pirzkal et al. 2007; Finkelstein et al.2008; Nilsson et al. 2009; Yuma et al. 2010; Ouchi et al. ⋆ Based on observations made with ESO Telescopes at the ParanalObservatory under programme IDs 083.B-0470 and 60.A-9433 α tobe detected by standard criteria. The first candidate to be blamedfor this apparent discrepancy between theory and observationwas absorption by dust, which is prominent in the ultraviolet.Early surveys in the local Universe hinted at an anti-correlationbetween the metallicity (which generally correlates with dustcontent) and Ly α luminosity (e.g. Meier & Terlevich 1981), butit soon became clear that dust alone could not explain the de-viation from recombination theory (Giavalisco et al. 1996). Inparticular one galaxy, I Zw 18, showed a very low metallicitycombined with strong Ly α absorption (Kunth et al. 1994).The attention then turned to the resonant scattering of Ly α in H i , which had been explored theoretically for a time (e.g.Osterbrock 1962; Adams 1972). If the path length of the Ly α photons is greatly increased in multiple scatterings on the waythrough the galaxy, even small amounts of dust can cause a largeabsorption (Neufeld 1990). However, if the neutral gas is shiftedin velocity, the Ly α photons are shifted out of resonance and can escape from the galaxy more easily. Kunth et al. (1998) showedfrom a sample of eight local gas-rich dwarf galaxies that theemission of Ly α in each case exhibited a P Cygni profile ac-companied by blueshifted low ionization state (LIS) absorptionlines, suggesting that an outflow of neutral gas would allow theLy α emission to escape towards us. Modern simulations (e.g.Verhamme et al. 2006; Garel et al. 2012, Duval et al. 2012, sub-mitted) thus take into account a combination of the relative ve-locities of ionized and neutral hydrogen gas, as well as dust, andtheir respective morphologies, that all govern the escape of Ly α emission from galaxies.Although neutral gas is best traced through LIS absorp-tion lines in the UV, such a study at low redshift requires aspace telescope due to ultraviolet absorption in the atmosphere.The choice is then either HST / STIS which has a low sensitiv-ity or HST / COS with limited spatial information. However, theemission and absorption of Ly α may vary over small scales(Mas-Hesse et al. 2003; ¨Ostlin et al. 2009) and a detailed studyrequires some degree of spatial resolution. It is therefore use-ful to identify an alternative set of longer wavelength absorp-tion features that still form in the neutral ISM, in order to facil-itate the use of contiguous integral field spectrographs attachedto large-aperture telescopes on the ground. Neutral gas motionsin star-forming galaxies have been studied in the past using,e.g. Mg ii i α .In this paper, we present the first spatially resolved study ofabsorption from Na i in the cold interstellar gas in two nearbyLy α emitting blue compact galaxies. Specifically, we targetthe sodium resonance absorption doublet ( λλ α emis-sion and the possible similarities to LAEs and Lyman BreakGalaxies (LBGs) at high redshift (see Section 2).Na D is generally stronger than any other resonance line inthe optical (such as K i or Ca ii , which are strongly depleted indi ff use, low-velocity clouds, see e.g. Spitzer (1968)). With anionizing potential of 5.14 eV it is a good tracer of neutral hydro-gen gas. By simultaneously studying the H α emission line, wecan accurately determine the location, extent and velocity dis-tribution of the ionized hydrogen in these galaxies. We are thusable to measure the relative velocities between the ionized andneutral hydrogen gas, which is precisely the Doppler shift rele-vant for Ly α transmission. With integral field unit spectra fromArgus at VLT / FLAMES we achieve a high spectral resolutioncombined with spatial information. We include an independentmeasurement from VLT / X-Shooter to confirm these results, andto estimate the stellar contamination of the spectra.The paper proceeds as follows; in Section 2 we describe pre-vious studies of the two galaxies in our sample in more detail.In Section 3 we describe our data and the reduction steps. InSection 4 we present our results. In section 5 we discuss pos-sible complications from stellar contamination. In Section 6 wediscuss our results and in Section 7 we leave our concluding re-marks.
2. Sample
The two galaxies in our study have already been studied exten-sively for their interesting properties, but the neutral gas veloc- ity as estimated from the sodium doublet has not been measuredwith high accuracy and spatial resolution until now. AnalyzingVLT / UVES spectra of ESO 338-IG04, ¨Ostlin et al. (2007) esti-mated an outflow velocity ∼
20 km / s from Na D, but emphasizedthat with their low signal to noise they could not estimate the rel-ative ISM versus stellar contribution.An HST UV imaging program including the two galaxies(Kunth et al. 2003; Hayes et al. 2005, 2007; ¨Ostlin et al. 2009)gives us information on the Ly α and UV continuum morpholo-gies.Haro 11 is a blue compact galaxy at z = .
02 with three largecondensations or ”knots”. The knots are traditionally called A,B and C (Vader et al. 1993) as shown in the left panel of Figure1. Knot C appears to be the brightest knot in the ultraviolet andit also shows Ly α in emission, while knot B is brightest in H α ,where Ly α is instead absorbed. These two knots are thereforeof particular interest as we try to explain why Ly α is seen fromone region but is absorbed in the other. Some evidence suggeststhat Haro 11 might be the result of a merger of dwarf galax-ies ( ¨Ostlin et al. 2001), particularly due to the irregular appear-ance, the high relative velocities of several hundred km / s, thebroad emission lines, and the presence of a tidal arm structurewith a high redshifted velocity relative to the rest of the system( ¨Ostlin et al. 2013, in preparation). The H α line width within astar forming knot is as high as ∼
270 km / s (FWHM), and showsstrong multi-component features.Kunth et al. (1998) present HST / GHRS spectra of Haro 11around Ly α and the interstellar O i ( λ ii ( λ / s. Both absorption lines are very broad, indicatingmultiple components along the line of sight spanning roughly200 km / s in velocity. The Ly α line profile has a strong underly-ing absorption, extending more than 1500 km / s on the blue sideof the emission. The line does not show a clear P Cygni pro-file, and the underlying absorption seems to extend also to thered side. Evidence for several blueshifted Ly α components wasfound, indicating multiple absorbing gas clouds. Unfortunately,it is not clear exactly where in the galaxy the 1 . ′′ × . ′′ ∼
180 km / s) and the Ly α emission is concentrated to abright central region. ESO 338 is riddled with ”super star clus-ters”; small knots of intense star formation ( ¨Ostlin et al. 1998,2003, 2007). The ongoing starburst is about 40 Myr old andwas likely triggered by a merger with a small galaxy or froman interaction with remaining debris from a previous encounterwith the companion galaxy ESO 338-IG04b ( ¨Ostlin et al. 2001;Cannon et al. 2004). When we discuss ESO 338 below, wemainly focus on the bright central region, which we refer to asknot A (Hayes et al. 2005, cf. Figure 1) (also known as cluster / STIS long slit spectrum across ESO 338. The spectrumshows di ff use Ly α emission in several regions along the slit, butover knot A the Ly α emission is weak and shows only a hint ofblueshifted absorption. However, the low resolution of the spec-trum makes a detailed kinematical study di ffi cult, and the emis-sion line lies very close to the strong geocoronal Ly α line. Theslit is also only 0 . ′′ α profile in this region. An outflow ve-locity from this STIS spectrum was estimated by Schwartz et al. Fig. 1.
Slit positions, field of view (FOV) and approximate aperture positions for Haro 11 (left) and ESO 338-IG04 (right), overlayedon HST / ACS F550M continuum images (i.e. close in wavelength to the Na D feature). The X-Shooter slits are 0.9 (or 1.0 for theUVB-arm, not marked in the figure) arcsec wide and 11 arcsec long and marked with solid lines. The FOV for the LR6 wavelengthrange (which includes H α ) is marked with dash-dotted lines while the shorter LR5 range (including Na D) is marked with dottedlines. The Argus binned ”apertures” are 1.56 arcsec wide squares marked with dashed lines. The VLT / FLAMES FOV shown hereshows the area for which we have full exposure after taking into account the dithering pattern.(2006) from low ionization state absorption lines in the UV to be47 ±
70 km / s.In this paper we look mainly at the strong knots in the respec-tive galaxies; see Figure 1. Even though our field of view withVLT / FLAMES is larger, the Na D signature is very weak and theareas marked are the only regions where we can successfully fitthe Na D lines, even after spatially binning our spectra.
3. Observations and Data reduction
The integral field spectroscopy observations (ESO ID 083.B-0470(A)) were performed between June 9 and July 6 2009 withthe Argus integral field unit of the GIRAFFE spectrograph atthe FLAMES instrument at VLT / UT2. Argus consists of 14x22lenslets arranged in a rectangular grid, corresponding to a fieldof view of approximately 7 x 11 arcseconds. Each lenslet is thenconnected to a fiber which leads the light to a spectrograph. Onespectrum for each lenslet is thus obtained.We observed the two galaxies Haro 11 and ESO 338-IG04,both with two ”low resolution” ( R = − λ α emission line ( λ λ λλ × = × + × = . ′′ − . ′′ ∼ . ′′ . ′′ − . ′′ esorex .The bias frames were median combined with the gimasterbias recipe, producing master bias frames foreach observing night. The gimasterflat recipe was then usedfor calibrating the fiber positioning onto the CCD. The flatframes used for each observing block were always acquiredduring the same night, using the Nasmyth screen (which givesbetter illumination in Argus mode). The giwavecalibration recipe subsequently performs the wavelength calibration bymapping known wavelengths from a calibration lamp. Thisrecipe requires an existing dispersion solution as an initialguess, which is then refined. We checked the convergence ofthis procedure by giving the refined dispersion solution as a newinitial guess and ensuring that the two refined solutions wereidentical for each calibration set. Finally, the giscience recipeapplies all calibrations to the data and creates a data cube withtwo spatial and one spectral dimension.Each science spectrum was reduced individually. As a sanitycheck, we verify the sky subtraction by measuring the continuumflux in regions far from the bright condensations, which we findto be consistent with zero.The subsequent reduction steps were made with the PyFITSmodule version 2.3.1 for Python . Since the science data weredithered (the telescope made small movements of roughly 0.5 to1 arcsec between each science exposure), we created an emptyx-y- λ -cube with spatial dimensions matching those of the regioncorresponding to the region on the sky for which we had con-tributions from each science exposure. The spectral data werethen combined onto this grid using a weighted average intensityfor each integral field unit pixel (or ”spaxel”), calculated fromthe uncertainty values obtained from the pipeline. The ditheringshifts were rounded o ff to whole spaxels before combining. Thismay thus introduce a small astrometric shift of up to 0 . ′′
25 in anindividual exposure. Our final spatial resolution is therefore ex-pected to be similar to our worst seeing of about one arcsec. Thefinal data product has the original 0 . ′′ / lenslet spatial samplingresolution and 0.2 Å spectral sampling resolution. However, wealways consider binned spaxels in a 3x3 configuration in this pa-per, corresponding to a bin size of 1 . ′′ version 0.7.0 Python package, which uses aLevenberg-Marquardt least-squares minimization technique. PyFITS is a product of the Space Telescope Science Institute, whichis operated by AURA for NASA http: // Table 1.
X-Shooter observing specifications
Parameter Haro 11B Haro 11C ESO 338-IG04Airmass 1.016 1.012 1.430Seeing 0 . ′′ . ′′ − . ′′ . ′′ The X-Shooter data were obtained as part of the first scienceverification for the instrument between August 10 and 11 2009(ESO ID 60.A-9433(A)). For the VIS arm ( λ ∼ . ′′ × ′′ was used. For the UVB arm ( λ ∼ ′′ × ′′ . The resolving power R is 8800 and 5100for the VIS and UVB arms, respectively.The spectra were reduced with the X-Shooter pipeline v.1.3.7 using esorex v. 3.9.0. Standard settings were used in thephysical model mode, using the xsh scired slit nod recipeto perform reduction, sky subtraction and extraction of the sci-ence frames.These data are also analyzed by Guseva et al. (2012), but wehave included di ff erent reduction steps and we do not discuss thedata from the NIR arm. With the large wavelength range of thesedata, we can examine the e ff ect of the underlying photosphericBalmer absorption on the H α emission feature. It is strongest inESO 338, but negligible for our discussion in all cases.
4. Results
The Na D feature is very weak in these galaxies and we weretherefore forced to bin our FLAMES spectra in the spatial di-mension. Even in this case, only the very strongest continuumfeatures show Na D absorption. This reduces our discussion ofthe larger fields of view of FLAMES to the selected binned aper-tures shown in Figure 1. Except for knot A in Haro 11, theseare the only regions where we can see a clear Na D absorptionprofile. Nevertheless, these apertures still allow us to perform asu ffi ciently detailed analysis in some of the most interesting Ly α emitting and absorbing regions.The most important spectral features from the FLAMESbinned apertures are shown in Figure 2. The spectra are not fluxcalibrated and are instead shown normalized to the continuum.The continuum was in all cases fitted with a low-order poly-nomial across the entire observed range, which yielded a goodfit. The H α and He i (5875.64 Å) lines both show strong multi-component features, as is evident from the deviation from a sim-ple gaussian. Note that these features are very similar in bothemission lines. We verify that the asymmetry is not an instru-mental feature by examining the shape of skylines and of cal-ibration lamp frames. The measured velocities of the di ff erentspectral components are listed in Table 3.We have also attempted multi-component fits to the H α emis-sion lines (not shown). In all cases, we identify two strong com-ponents in the H α lines: One narrow ”main” component andone broader but weaker component. In the case of Haro 11 Band ESO 338, the velocity of the broad component is consistentwith the velocity of the main component, within the uncertain-ties. For Haro 11 A the narrow (broad) component has a velocityof 6242 ± ± α flux from this component is there- Table 2.
Na D properties and dust extinction
Parameter Haro 11B Haro 11C ESO 338-IG04Na D doublet EW (Å) − . ± . − . ± . − . ± . / . ± .
22 1 . ± .
18 0 . ± . > > & . / s) − ±
13 32 ± − ± a a < . b Notes.
References: a) Hayes et al. (2007), b) Bergvall & ¨Ostlin (2002) fore not only expected to be weaker intrinsically but should alsosu ff er a larger resonance e ff ect. In each case, these secondarycomponents do not change the velocity of the main componentby more than a few kilometers per second and for simplicity wehave chosen not to include them in the further discussion of thispaper. Our goal is not to perfectly reproduce the line shapes butto analyze the velocities of the main constituents of the galaxies.We note that the components we measure in Haro 11 agree wellwith the recent results from James et al. (2013).We also attempted a simple wavelength centroid fit, numeri-cally defined as Σ λ f λ / Σ f λ across the continuum subtracted emis-sion lines. The result is perfectly consistent with the numbers weuse in this paper, with a maximum shift in velocity of 5 km / s.The X-Shooter spectra are shown for comparison in Figure 3(see also Section 5). Note that single gaussian fits appear to re-produce the line shapes better for these data, but this is mainlydue to a lower spectral resolution and signal to noise.We see the velocity of emission lines from ionized gas are thesame within each region we measure, independent of the speciesused. In the same region, the velocities as measured by the emis-sion lines of H α ( λ i ( λ i ] ( λλ ii ] ( λλ ii ] ( λλ EW NaD = − . ± .
06Å for knot C, − . ± .
06Å for knot B,and − . ± .
04Å for ESO 338. For Haro 11 A we estimate anupper limit of − . C f = − I where I is the residual intensity in the blue Na D line. Notehowever, that the residual intensity may be low in narrow, unre-solved components in our spectrum and that we therefore mea-sure an artificially stronger intensity. We will come back to thispoint later in the discussion. With this simple assumption how-ever, we estimate as a lower limit the covering fraction of Na Dto be ∼
10 % for Haro 11 knots B and C, and roughly 5 % forESO 338.For Haro 11, we detect no measurable Na D absorption to-ward knot A. This is probably due to the low continuum bright-ness in this region which is necessary for observing absorptionlines, combined with a low covering fraction of neutral gas. Forknot B, we are able to measure the Na D velocity and find a
Fig. 2. H α , He i (5875.6) and Na D line profiles from the VLT / FLAMES data. The blue gaussians are a simple fit to the emissionor absorption lines. In the case of Haro 11 knot A, the best fit is shown in this figure but is never used. The vertical dashed linesshow the velocity given by the H α gaussian fit, and the vertical dotted lines show the Na D velocity. The y-axis is normalized to thecontinuum. Also shown inset in the bottom row are errorbars, representing the standard deviation in the continuum from 5900 to5950 Å which shows no emission or absorption features. Since the spectra are normalized to the continuum, these errorbars reflectthe strength of the continuum in each region.moderate blueshifted velocity (compared to the ionized gas) of44 ±
13 km / s. In knot C the absorption is stronger and we find aredshifted velocity of 32 ±
12 km / s.In ESO 338, we are only able to measure the Na D velocityin the brightest knot (knot A), were both the continuum and theLy α emission is strongest. We find no evidence for a shift inthe velocity, with the best fit giving -15 ±
16 km / s. ¨Ostlin et al.(2007) estimated a shift of circa 20 km / s from UVES spectrawith a lower signal to noise. Our estimate of the stellar versusnebular velocity agrees well with the analysis of Cumming et al.(2008), where the stellar component (as measured by the Ca ii triplet ( λλ ± / s.
5. Possible stellar contamination of Na D
A challenge with using Na D to measure neutral gas flows is thatthe stars in the galaxy may also exhibit the absorption lines. Inthis section we demonstrate that the Na D we measure is primar-ily of interstellar origin.Na D is prominent in spectra of cool stars: K- and M-typegiants and supergiants. There are several ways of disentanglingthe stellar and interstellar sodium. One of the more common androbust methods make use of the Mg i b-band triplet ( λλ i b must come from stellar atmospheres.Thanks to their similar origins, the equivalent widths ofthe Mg b triplet and the stellar Na D doublet appear tobe strongly correlated. However, any ground-based spectra ofnearby stars are always contaminated by telluric Na D inemission, which makes the relative ratio di ffi cult to measure.Rupke et al. (2002) compile measurements of Galactic globularclusters and mostly nonactive galaxies (Bica & Alloin 1986) andnuclei of nearby galaxies (Heckman et al. 1980). They find theequivalent widths of the purely stellar features to be correlated as EW NaD ∼ . EW Mg b , with a possible intrinsic scatter of & EW NaD ∼ . EW Mg b , which seems to agree withtheir own K giant spectra to within 0.10 dex. Schwartz & Martin(2004) also make a fit to the data by Jacoby et al. (1984) tofind EW NaD = . ± . EW Mg b . Martin (2005) analyzeKeck II / ESI spectra of A, F, G and K dwarfs and giants andconclude EW NaD = / EW Mg b and that the stellar contri-bution was consistently less than 10 % in their study of 18ULIRGs. Sato et al. (2009) investigate Na D absorption in 493spectra from the AEGIS survey, and find EW NaD = . EW Mg b well describes the purely stellar boundary in their sample (seetheir Fig. 1). Finally, Chen et al. (2010) used Sloan Digital SkySurvey (SDSS) spectra from young disc galaxies and estimatethat an average ∼
80% of the Na D absorption arises in stellaratmospheres. Their estimate of the stellar contribution is based
Fig. 3. H α , He i (5875.6), Na D, K i and Mg i line profiles fromthe VLT / X-Shooter data. The blue gaussians are a simple fit tothe emission or absorption lines. The vertical lines show the sys-temic velocity as given by the H α gaussian fit. The y-axis is nor-malized to the continuum.on fitting and subtracting a stellar population synthesis model tothe continuum and absorption features, and they emphasize thattheir spectral model sometimes shows stronger Na D absorptionthan the actual data. They conclude that the most likely expla-nation is that Na D is sometimes seen in emission in their data,predominantly from face-on galaxies with low dust attenuation,and that a stronger stellar contamination is expected in the spec-tra of normal star-forming galaxies as opposed to that in youngstars comparable to the dominant stellar population in our twogalaxies here.We note that Na D is a resonance absorption feature, andtherefore absorbed photons will be re-emitted in a random di-rection. This can create a di ff use, low surface brightness emis-sion component. Prochaska et al. (2011) model the e ff ect of re-emitted light and conclude that in the extreme case that all ofthis light eventually escapes and is caught within the aperture, itcan significantly reduce the observed equivalent width. We ex-pect this e ff ect to be very small in our analysis, however, sincewe are looking only at narrow slits and small e ff ective apertures(1.5 arcsec corresponds to roughly 0.6 kpc in Haro 11 and evenless in ESO 338).An alternative to measuring outflows with Na D is the K i doublet ( λλ absorption bands.Here, we use our X-shooter data to estimate the stellar con-tamination. Due to the comparatively low signal-to-noise and Table 3.
Best fit observed velocities, in km / s. Line Haro11A Haro 11B Haro 11C ESO 338-IG04H α ±
10 6146 ± ± ± i ± ± ± ± i ... 6162 ±
20 6128 ±
15 2853 ± ± ± ± i ... 6123 ±
20 6146 ±
20 2830 ± Notes. Mg i and K i are measured only from VLT / X-Shooter and theuncertainty is therefore larger. The other lines were measured in bothVLT / FLAMES and VLT / X-Shooter and this represents the weightedaverage. lower resolution of the X-Shooter data, the equivalent widths arenot as easily determined as for the Argus data. However, we canstill use our spectra in Figure 3 to see that it is only in Haro 11C that we can clearly distinguish the Mg b feature, although it isvery weak. We thus let the upper limit from Haro 11 C representthe upper limit on the stellar contamination for all our results. Ifthe EW of Mg b would be larger than this limit in the other re-gions this would be readily seen in Figure 3. We estimate the EWof this feature to be roughly -0.2 Å, i.e. about half the equivalentwidth of Na D.Based on the discussion above we thus estimate the stellarcontribution to be on the order of 25-30%, but we note that it maybe as high as 50% in the extreme case. A significant contributionto the Na D profile at the stellar velocity (which is consistentwith the H-alpha velocity in our galaxies) would decrease theaverage o ff set velocities that we measure. The absolute value ofour Na D velocities may therefore be somewhat lower than thepurely nebular Na D velocity.
6. Discussion
We have measured outflow / inflow velocities and covering frac-tions of the neutral ISM in front of three bright star-formingcondensations, two of which show Ly α emission (ESO 338 A,and Haro 11 C) and one shows Ly α in absorption (Haro 11 B).Further to the information measured in this programme, we alsocompile measurements of the nebular dust attenuation in theseregions from our previous investigations: Hayes et al. (2007) forHaro 11, Bergvall & ¨Ostlin (2002) for ESO 338, see Table 2. For ESO338-IG04, we find thatthe velocity of the neutral gas (as estimated from the Na D ab-sorption) is only very slightly blueshifted (or even static) com-pared to the ionized gas. This would rule out the outflow sce-nario as an explanation for the observed direct Ly α escapefrom this region. However, we note that H α from the centralregion (knot A) of ESO 338 is dominated by a large H ii shell(e.g. Bergvall & ¨Ostlin 2002). The UVES spectra analyzed by¨Ostlin et al. (2007) show a multi-component feature in the [O iii ] ( λ α lines towards knot A, which they interpretas an expanding bubble at ∼
40 km / s. Presumably this shell doesconsists of outflowing ionized gas from stellar winds and super-novae, yet we measure a low outflow velocity in H i . If there isindeed an expanding ionized bubble in ESO 338, it is possiblethat the H α feature is dominated by emission on the side of thebubble facing our way. This would cause us to measure a moreblueshifted velocity for the ionized gas and reduce our inferredneutral gas velocity. However, the bubble seems to be optically thin, which would mean that the e ff ect is very small, and also thestellar component in knot A appears to have the same velocityas the ionized gas (Cumming et al. 2008). Emission – Haro 11 knot C
For Haro 11, we find a small red-shifted velocity of Na D in front of knot C, indicating infall ofcold gas. HST imaging shows Ly α emission from this region,which could partially be explained by this velocity di ff erence.However, it is in the opposite direction to an outflow, and thevelocity di ff erence is not particularly strong.There is evidence for a Ly α / H α ratio higher than the recom-bination case B value in both ESO 338 A ( ¨Ostlin et al. 2009) andHaro 11 C (if extinction is taken into account; Atek et al. 2008).This could be due to Ly α photons actually su ff ering less at-tenuation while scattering on the surface of cool dusty clumpsthan the continuum photons (Neufeld 1991), although this sce-nario seems unlikely based on recent simulations (Laursen et al.2012, Duval et al. 2012, submitted). It is also possible that anextinction correction based on a clumpy dust distribution modelrather than a uniform dust screen would give a result consistentwith recombination values (see e.g. Scarlata et al. 2009). Absorption – Haro 11 knot B
For knot B, we find a moderateblueshifted velocity. Combined with a low covering fraction ∼
10 %, Ly α should escape more easily from knot B. However,our HST UV imaging shows that Ly α is strongly absorbed inthis region.There is some debate as to whether the dust extinction ishigher in knot B than in knot C. Judging from HST images,there do seem to be more dust clouds near and around knot Band based on the X-Shooter data, knot C seems to have a lowerextinction (Guseva et al. 2012). However, the H α and H β imagesfrom which we derive the E(B-V) values presented in Table 2give approximately the same level of extinction in both knots(Hayes et al. 2007; Atek et al. 2008). If the medium is highlyclumpy it is likely that the extinction is varying, even withinthese small regions. Di ff erent results are thus likely due to di ff er-ent aperture sizes and positions, as well as di ff erent techniquesfor measuring the extinction probing varying optical depths. Absorption – Haro 11 knot A
The signal-to-noise and theequivalent width of Na D are too small for us to safely attempta measurement toward knot A. This is partly because the opticalcontinuum emission near Na D is weaker than in the other knots,but it still indicates a low covering fraction of neutral gas. KnotA exhibits both H α and UV continuum emission from ionizedregions, but no significant Ly α radiation appears to escape fromthem. In the three regions that we have identified for study in this arti-cle, we have compiled measurements of H i covering fractions,kinematics, and dust contents.In both knot A and B in Haro 11, absorption of Ly α is seendespite a low covering fraction. However, the covering fractionsthat we estimate from the Na D profiles likely only serves aslower limits, since there may be very narrow, unresolved com-ponents with lower residual intensity in our spectra. We note thatthe O i and Si ii lines in the GHRS spectra of Kunth et al. (1998)are considerably stronger, indicating a high covering fraction of neutral gas, consistent with the strong Ly α absorption on theblue side of the line. The resolution of GHRS ( R ∼ R = ff erence. We note also thatother Na D absorption spectra in the literature often show con-siderably stronger lines (e.g. Heckman et al. 2000; Martin 2005).Our conclusion is that the Na D covering fraction is indeed lowin these regions. The di ff erence from the previous measurementmight in part be explained by the di ff erent apertures pointingat di ff erent sightlines in an inhomogenous ISM. The di ff erencemay also come from the lower photoionization threshold of NaD of 5.14 eV. It is thus possible that the column density of hydro-gen is high enough overall to absorb Ly α but low enough that wedetect Na D only in the densest regions. The interstellar mediumis known to often be very patchy, and it would not be surprisingin these two galaxies with their turbulent pasts. Since Na D hasto be shielded from ionizing radiation by dust (Chen et al. 2010;Murray et al. 2007), it likely only exists in the densest, coolestclumps.For the optical depths that we find from the Na D line ratio,we can put a lower limit on the column density of the gas wherethe absorption arises. We use the relationN(Na i ) = τ b . × − λ f (1)from Spitzer (1978), where τ is the central optical depth, b is the Doppler parameter (in km / s), λ is the vacuum wavelengthin Å and f is the oscillator strength = b = FWHM / (2 √ ln 2) of &
60 km / s for the Na D in ESO 338, a lower limit on the Na i column density is roughly N(Na i ) & × cm − . Convertingthis to a hydrogen column density of course depends heav-ily on the Na / H abundance ratio, but using very conserva-tive estimates based on the conversions given in Rupke et al.(2002); Murray et al. (2007) this corresponds to at least N(H i ) > cm − . Ly α becomes optically thick and resonantly scat-ters on neutral hydrogen already at H i column densities around10 − cm − . It would take an enormous departure from theseassumptions to make the clouds optically thin. In these denseclouds, the Ly α photons would see upwards of a million op-tical depths, and the clouds would be self-shielding. Still, thecovering fraction of these dense clouds where we see Na D isonly ∼ .
1. Our interpretation is that the ISM in these galaxiesis inhomogeneous and likely consists of column densities be-tween these extremes. Ly α photons would then escape throughregions or patches with little neutral gas, and be blocked by gasat slightly higher column densities, while we measure our veloc-ities and covering fractions in the very densest clumps.For Haro 11, a picket-fence model of the ISM agreeswell with the detection of the Lyman continuum (Ly C, < only direct paths are viable (see e.g. Heckman et al. 2001;Zastrow et al. 2011). An indirect method of estimating Ly C es-cape through the residual in the C ii ( λ α is absorbed in knot B, which shows the larger kinematic o ff set.We hypothesize on the existence of di ff use remnant H i gas witha high covering fraction towards knot B. Our measurements of the neutral gas velocities in Haro 11 Bagree rather well with the results in Kunth et al. (1998) where anoutflow of ∼
60 km / s measured by low ionization state absorptionlines in the UV was seen. However, the same spectra show Ly α in emission which would imply that some of the emission fromknot C is included in the aperture as well. It is unfortunatelynot clear exactly where in the galaxy the HST / GHRS aperturewas placed (see discussion in Hayes et al. 2007). Most likely, itwas in between the knots, and it is quite likely that the di ff erentcomponents are mixed in the spectra. Indeed, Kunth et al. (1998)see very broad absorption features (spanning roughly 200 km / s),indicating multiple absorption components.We note that a previous measurement of the neutral gas ve-locity in ESO 338 was made with HST / STIS in Schwartz et al.(2006), which agrees well with our observations. The LIS ab-sorption features are weak and appear close to the systemic ve-locity. Unfortunately, the available HST / STIS spectrum for ESO338 is not optimized for exploring the Ly α emitting region thatwe investigate in this paper. In light of our results, it would bevery interesting to explore both Haro 11 and ESO 338 with mul-tiple HST / COS and HST / STIS pointings to investigate the Ly α emission and absorption line profile in these galaxies in un-precedented spatial and spectral detail, which can be achievedby combining the two instruments.
7. Conclusions
We have presented the first spatially resolved measurements ofthe sodium doublet (Na D) in the two nearby Ly α emittinggalaxies Haro 11 and ESO 338-IG04. Our results can be sum-marized as follows: – We find an outflow of neutral gas from knot B (which showsstrong Ly α absorption) and slow infall towards knot C inHaro 11. In ESO 338-IG04 we find a slow or static interstel-lar medium. In the two latter cases, Ly α is seen in emission.The velocities that we find for the neutral gas are not whatwe would have expected from standard Ly α escape scenar-ios. Typically, a strong outflow is assumed to allow Ly α toescape more easily. – From the Na D line profiles we measure relatively high op-tical depths but small covering fractions of Na D ( ∼ i ) > cm − . Since Ly α is a ff ected by resonant scat-tering already at column densities of N(H i ) ∼ − cm − ,it is likely that the direct Ly α escape seen in e.g. knot C inHaro 11 and from ESO 338 is due to a picket-fence scenariowhere the interstellar medium is highly inhomogeneous andconsists of both dense, neutral clumps as well as ionized gasalong our lines of sight. – We see a larger kinematical o ff set in Haro 11 knot B than C,yet B shows strong Ly α absorption in contrast to the emis-sion from C. Given the relatively similar nebular dust contentof the two knots, we hypothesize on the existence of a dif-fuse remnant H i component with a high covering fractiontowards B, and a possible perpendicular outflow from C. Acknowledgements.
We would like to thank Matthew Lehnert for a fruitful dis-cussion regarding the stellar contamination problem. M.H. received support fromAgence Nationale de la recherche bearing the reference ANR-09-BLAN-0234-01. G ¨O is a Royal Swedish Academy of Sciences Research Fellow, supported bya grant from the Knut and Alice Wallenberg foundation, and also acknowledgessupport from the Swedish Research Council and the Swedish National Space Board. J.M.M.H. is funded by Spanish MINECO grants AYA2010-21887-C04-02 (ESTALLIDOS) and AYA2012-39362-C02-01.
References
Adams, J. J., Blanc, G. A., Hill, G. J., et al. 2011, ApJS, 192, 5Adams, T. F. 1972, ApJ, 174, 439Ajiki, M., Mobasher, B., Taniguchi, Y., et al. 2006, ApJ, 638, 596Atek, H., Kunth, D., Hayes, M., ¨Ostlin, G., & Mas-Hesse, J. M. 2008, A&A,488, 491Bergvall, N. & ¨Ostlin, G. 2002, A&A, 390, 891Bergvall, N., Zackrisson, E., Andersson, B.-G., et al. 2006, A&A, 448, 513Bica, E. & Alloin, D. 1986, A&A, 166, 83Blanc, G. A., Adams, J. J., Gebhardt, K., et al. 2011, ApJ, 736, 31Cannon, J. M., Skillman, E. D., Kunth, D., et al. 2004, ApJ, 608, 768Chen, Y.-M., Tremonti, C. A., Heckman, T. M., et al. 2010, AJ, 140, 445Churchill, C. W. & Vogt, S. S. 2001, AJ, 122, 679Cowie, L. L. & Hu, E. M. 1998, AJ, 115, 1319Cumming, R. J., Fathi, K., ¨Ostlin, G., et al. 2008, A&A, 479, 725Dijkstra, M., Wyithe, J. S. B., & Haiman, Z. 2007, MNRAS, 379, 253Finkelstein, S. L., Rhoads, J. E., Malhotra, S., Grogin, N., & Wang, J. 2008, ApJ,678, 655Garel, T., Blaizot, J., Guiderdoni, B., et al. 2012, MNRAS, 422, 310Gawiser, E., van Dokkum, P. G., Gronwall, C., et al. 2006, ApJ, 642, L13Giavalisco, M., Koratkar, A., & Calzetti, D. 1996, ApJ, 466, 831Gronwall, C., Ciardullo, R., Hickey, T., et al. 2007, ApJ, 667, 79Guaita, L., Acquaviva, V., Padilla, N., et al. 2011, ApJ, 733, 114Guseva, N. G., Izotov, Y. I., Fricke, K. J., & Henkel, C. 2012, A&A, 541, A115Hayes, M., ¨Ostlin, G., Atek, H., et al. 2007, MNRAS, 382, 1465Hayes, M., ¨Ostlin, G., Mas-Hesse, J. M., et al. 2005, A&A, 438, 71Hayes, M., ¨Ostlin, G., Schaerer, D., et al. 2010, Nature, 464, 562Heckman, T. M., Crane, P. C., & Balick, B. 1980, A&AS, 40, 295Heckman, T. M., Lehnert, M. D., Strickland, D. K., & Armus, L. 2000, ApJS,129, 493Heckman, T. M., Sembach, K. R., Meurer, G. R., et al. 2001, ApJ, 558, 56Jacoby, G. H., Hunter, D. A., & Christian, C. A. 1984, ApJS, 56, 257James, B. L., Tsamis, Y. G., Walsh, J. R., Barlow, M. J., & Westmoquette, M. S.2013, ArXiv e-printsJensen, H., Laursen, P., Mellema, G., et al. 2012, ArXiv e-printsKemp, S. N., Bates, B., Beckman, J. E., et al. 2002, MNRAS, 333, 561Kudritzki, R.-P., M´endez, R. H., Feldmeier, J. J., et al. 2000, ApJ, 536, 19Kunth, D., Leitherer, C., Mas-Hesse, J. M., ¨Ostlin, G., & Petrosian, A. 2003,ApJ, 597, 263Kunth, D., Lequeux, J., Sargent, W. L. W., & Viallefond, F. 1994, A&A, 282,709Kunth, D., Mas-Hesse, J. M., Terlevich, E., et al. 1998, A&A, 334, 11Laursen, P., Duval, F., & ¨Ostlin, G. 2012, ArXiv e-printsLeitet, E., Bergvall, N., Piskunov, N., & Andersson, B.-G. 2011, A&A, 532,A107Malhotra, S. & Rhoads, J. E. 2002, ApJ, 565, L71Malhotra, S. & Rhoads, J. E. 2004, ApJ, 617, L5Martin, C. L. 2005, ApJ, 621, 227Martin, C. L. & Bouch´e, N. 2009, ApJ, 703, 1394Mas-Hesse, J. M., Kunth, D., Tenorio-Tagle, G., et al. 2003, ApJ, 598, 858Meier, D. L. & Terlevich, R. 1981, ApJ, 246, L109Mshar, A. C., Charlton, J. C., Lynch, R. S., Churchill, C., & Kim, T.-S. 2007,ApJ, 669, 135Murray, N., Martin, C. L., Quataert, E., & Thompson, T. A. 2007, ApJ, 660, 211Nestor, D. B., Johnson, B. D., Wild, V., et al. 2011, MNRAS, 412, 1559Neufeld, D. A. 1990, ApJ, 350, 216Neufeld, D. A. 1991, ApJ, 370, L85Nilsson, K. K., Tapken, C., Møller, P., et al. 2009, A&A, 498, 13Ono, Y., Ouchi, M., Shimasaku, K., et al. 2010, ApJ, 724, 1524Osterbrock, D. E. 1962, ApJ, 135, 195¨Ostlin, G., Amram, P., Bergvall, N., et al. 2001, A&A, 374, 800¨Ostlin, G., Bergvall, N., & R¨onnback, J. 1998, A&A, 335, 85¨Ostlin, G., Cumming, R. J., & Bergvall, N. 2007, A&A, 461, 471¨Ostlin, G., Hayes, M., Kunth, D., et al. 2009, AJ, 138, 923¨Ostlin, G., Zackrisson, E., Bergvall, N., & R¨onnback, J. 2003, A&A, 408, 887Ouchi, M., Shimasaku, K., Akiyama, M., et al. 2005, ApJ, 620, L1Ouchi, M., Shimasaku, K., Furusawa, H., et al. 2003, ApJ, 582, 60Ouchi, M., Shimasaku, K., Furusawa, H., et al. 2010, ApJ, 723, 869Partridge, R. B. & Peebles, P. J. E. 1967, ApJ, 147, 868Pirzkal, N., Malhotra, S., Rhoads, J. E., & Xu, C. 2007, ApJ, 667, 49Pritchet, C. J. 1994, PASP, 106, 1052Prochaska, J. X., Kasen, D., & Rubin, K. 2011, ApJ, 734, 24Rhoads, J. E., Malhotra, S., Dey, A., et al. 2000, ApJ, 545, L85
Rupke, D. S., Veilleux, S., & Sanders, D. B. 2002, ApJ, 570, 588Sato, T., Martin, C. L., Noeske, K. G., Koo, D. C., & Lotz, J. M. 2009, ApJ, 696,214Scarlata, C., Colbert, J., Teplitz, H. I., et al. 2009, ApJ, 704, L98Schwartz, C. M. & Martin, C. L. 2004, ApJ, 610, 201Schwartz, C. M., Martin, C. L., Chandar, R., et al. 2006, ApJ, 646, 858Shibuya, T., Kashikawa, N., Ota, K., et al. 2012, ApJ, 752, 114Spitzer, L. 1968, Di ff use matter in spaceSpitzer, L. 1978, Physical processes in the interstellar mediumVader, J. P., Frogel, J. A., Terndrup, D. M., & Heisler, C. A. 1993, AJ, 106, 1743Verhamme, A., Schaerer, D., & Maselli, A. 2006, A&A, 460, 397Yuma, S., Ohta, K., Yabe, K., et al. 2010, ApJ, 720, 1016Zastrow, J., Oey, M. S., Veilleux, S., McDonald, M., & Martin, C. L. 2011, ApJ,741, L17use matter in spaceSpitzer, L. 1978, Physical processes in the interstellar mediumVader, J. P., Frogel, J. A., Terndrup, D. M., & Heisler, C. A. 1993, AJ, 106, 1743Verhamme, A., Schaerer, D., & Maselli, A. 2006, A&A, 460, 397Yuma, S., Ohta, K., Yabe, K., et al. 2010, ApJ, 720, 1016Zastrow, J., Oey, M. S., Veilleux, S., McDonald, M., & Martin, C. L. 2011, ApJ,741, L17