The COS/UVES Absorption Survey of the Magellanic Stream: I. One-Tenth Solar Abundances along the Body of the Stream
Andrew J. Fox, Philipp Richter, Bart P. Wakker, Nicolas Lehner, J. Christopher Howk, Nadya Ben Bekhti, Joss Bland-Hawthorn, Stephen Lucas
aa r X i v : . [ a s t r o - ph . GA ] A p r Draft version July 16, 2018
Preprint typeset using L A TEX style emulateapj v. 5/2/11
THE COS/UVES ABSORPTION SURVEY OF THE MAGELLANIC STREAM: I. ONE-TENTH SOLARABUNDANCES ALONG THE BODY OF THE STREAM
Andrew J. Fox , Philipp Richter , , Bart P. Wakker , Nicolas Lehner , J. Christopher Howk , Nadya BenBekhti , Joss Bland-Hawthorn , & Stephen Lucas Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218 Institut f¨ur Physik und Astronomie, Universit¨at Potsdam, Haus 28, Karl-Liebknecht-Str. 24/25, 14476, Potsdam, Germany Leibniz-Institut f¨ur Astrophysik Potsdam (AIP), An der Sternwarte 16, 14482 Potsdam, Germany Department of Astronomy, University of Wisconsin–Madison, 475 North Charter St., Madison, WI 53706 Department of Physics, University of Notre Dame, 225 Nieuwland Science Hall, Notre Dame, IN 46556 Argelander-Institut f¨ur Astronomie, Universit¨at Bonn, Auf dem H¨ugel 71, 53121 Bonn, Germany Institute of Astronomy, School of Physics, University of Sydney, NSW 2006, Australia Department of Physics & Astronomy, University College London, Gower Street, London, WC1E 6BT, UK
Draft version July 16, 2018
ABSTRACTThe Magellanic Stream (MS) is a massive and extended tail of multi-phase gas stripped out of theMagellanic Clouds and interacting with the Galactic halo. In this first paper of an ongoing programto study the Stream in absorption, we present a chemical abundance analysis based on
HST /COSand VLT/UVES spectra of four AGN (RBS 144, NGC 7714, PHL 2525, and HE 0056–3622) lyingbehind the MS. Two of these sightlines yield good MS metallicity measurements: toward RBS 144 wemeasure a low MS metallicity of [S/H]=[S II /H I ]= − ± I /H I ]= − ± ≈ ≈ ≈ I /H I ]= − ± v LSR =150 km s − cloud towardHE 0056–3622, which belongs to a population of anomalous velocity clouds near the South GalacticPole. This suggests these clouds are associated with the Stream or more distant structures (possiblythe Sculptor Group, which lies in this direction at the same velocity), rather than tracing foregroundGalactic material. Subject headings:
ISM: abundances – Magellanic Clouds – Galaxy: halo – Galaxy: evolution – Mag-ellanic Clouds INTRODUCTION
Satellite interactions are one of the important mecha-nisms by which massive galaxies acquire gas and sustaintheir ongoing star formation. These interactions gener-ate extended tidal gas streams which pass through thegalaxies’ multi-phase halos toward their disks. The gasbrought in by tidal streams, and other processes includ-ing cold and hot accretion, clearly plays an importantrole in galaxy evolution (see Putman et al. 2012), but therelative importance of these processes, and the detailedmanner in which gas enters galaxies, is poorly known.The extended halo of the Milky Way, which contains awell-mapped population of high-velocity clouds (HVCs)and is pierced by hundreds of quasar sightlines, repre- [email protected] Based on observations taken under program 12604 of theNASA/ESA Hubble Space Telescope, obtained at the SpaceTelescope Science Institute, which is operated by the Associationof Universities for Research in Astronomy, Inc., under NASAcontract NAS 5-26555, and under proposal ID 085.C-0172(A)with the Ultraviolet and Visual Echelle Spectrograph (UVES)on the Very Large Telescope (VLT) Unit 2 (Kueyen) operatedby the European Southern Observatory (ESO) at Paranal, Chile. sents an ideal locale to investigate these galaxy feedingprocesses.The Magellanic Stream (MS) stands as the moststriking example of a satellite interaction in theGalactic neighborhood. Massive (1.5–5 × M ⊙ in H I , depending on its distance; Br¨uns et al.2005; Nidever et al. 2010; Besla et al. 2012), extended( ≈ ◦ long, or ≈ ◦ when including the LeadingArm Hulsbosch & Wakker 1988; Braun & Thilker 2004;Br¨uns et al. 2005; Nidever et al. 2010), and of low-metallicity ( ≈ HST measurements of the MS metallicity using the S II /H I and Si II /H I ratios found values Z MS =0.28 ± Z MS =0.25 ± Z MS =0.10 ± I /H I ratio, whichis robust against ionization and dust corrections. Thesemeasurements support the view that the Stream origi-nates in the SMC, which has a current-day mean oxy-gen abundance of 0.22 solar, and not the LMC, whichhas a current-day mean oxygen abundance of 0.46 solar(Russell & Dopita 1992), but until now, good statisticson the Stream’s metal abundance have been lacking. Inaddition, caution must be used in comparing gas-phaseabundances in the Stream with current-day abundancesin the Clouds, since the Stream was stripped some timeago ( ≈ LMCSMCFAIRALL 9RBS 144 HE 0056-3622PHL 2525NGC 7714NGC 7469
Figure 1. H I
21 cm map of the MS generated from the LABsurvey and color-coded by H I column density. The color scaleranges from 5 × to 3 × cm − . The map is shown in Galac-tic coordinates centered on the South Galactic Pole. The inte-gration range in deviation velocity is v dev = −
500 to −
80 and 50to 500 km s − , chosen to minimize contamination by foregroundemission. The four directions from this paper plus Fairall 9 (Pa-per II) and NGC 7469 (F10) are marked. Note how the RBS 144and Fairall 9 directions sample the two principal filaments of theStream. Stream, as can be seen in Figure 1, which shows a21 cm emission map of the entire Magellanic region gen-erated from the Leiden-Argentine-Bonn (LAB) survey(Kalberla et al. 2005). The sightlines sample a widerange of H I column density, from log N (H I ) MS =20.17toward RBS 144 down to 18.24 toward PHL 2525. Basicproperties of these sightlines are given in Table 1. In acompanion paper, Richter et al. (2013, hereafter PaperII), we present detailed results from the Fairall 9 sight-line, lying only 8.3 ◦ away on the sky from RBS 144.We use the term “main body” of the Stream to re-fer to its principal H I -emitting filaments passing fromagellanic Stream Abundances 3 Table 1
Sightline Properties and Details of ObservationsTarget Type l b α aSMC v (H I ) bMS log N (H I ) cMS Exposure Time (s) S/N f ( ◦ ) ( ◦ ) ( ◦ ) (km s − ) GASS LAB UVES d G130M e G160M e − g − · · · h − < σ ) 18.24 5605 2146 2772 22 19HE 0056–3622 Sey1 293.719 − i i i a Angular separation of sightline from center of SMC. b Central LSR velocity of H I emission from the MS (or, in the case of HE 0056–3622, from the AVC). c Logarithmic H I column density in the Stream measured from GASS (14.4 ′ beam) and LAB (30 ′ beam) surveys. d VLT/UVES exposure time with Dichroic 1 390+580 setting. e HST /COS exposure time with G130M/1289 and G160/1589 settings. f Signal-to-noise ratio per resolution element of COS data at 1300 ˚A and 1550 ˚A. g Extended galaxy nucleus, not point-source. h No UVES data taken for this target. i The absorption centered at 150 km s − toward HE 0056–3622 traces the AVCs near the South Galactic Pole, not the MS. the LMC and SMC through the South Galactic Pole andover into the Western Galactic hemisphere (see Figure 1).The term does not refer to the Leading Arm or InterfaceRegion of the Magellanic System, which are discussedin Br¨uns et al. (2005). 21 cm studies have shown themain body to be bifurcated both spatially (Putman et al.2003) and kinematically (Nidever et al. 2008, hereafterN08) into two principal filaments. N08 reported thatone of the two filaments (hereafter the LMC filament)can be traced kinematically back to a star-forming areaof the LMC known as the South-Eastern H I Overden-sity (SEHO) region, which contains the giant H II region30 Doradus, but that the other filament (hereafter thesecond filament) cannot be traced to either MagellanicCloud, so its origin until now has been unknown.We note that RBS 144 lies behind the second fila-ment reported by N08, whereas PHL 2525 and NGC 7714lie near the tip of the Stream, where the filamentarystructure is difficult to discern. We also note that to-ward HE 0056–3622, the Stream is centered near v LSR =–10 km s − , where MS absorption overlaps with fore-ground Galactic absorption. For this reason HE 0056–3622 is not a good probe of the MS, even though thesightline passes through it. However, the HE 0056–3622spectrum shows high-velocity absorption at LSR veloc-ities of 80–200 km s − (see § I emission in directions (like HE 0056–3622) near theSouth Galactic Pole, where the Sculptor Group of galax-ies is located (Mathewson et al. 1975; Haynes & Roberts1979; Putman et al. 2003). HE 0056–3622 is therefore auseful probe of the AVCs. We use metallicity measure-ments to investigate the origin of these clouds, includingthe possibility that they are associated with the MS.This paper is structured as follows. In § § §
4, we derive and discuss thechemical abundances in the MS, analyze the dust con-tent of the Stream, and present
Cloudy models to in-vestigate the effects of ionization. A discussion on theimplications of these results on the origin of the Streamis presented in §
5. We summarize our findings in § ≡ log (X/H)–log (X/H) ⊙ . We adopt solar (photospheric) abundancesfrom Asplund et al. (2009). OBSERVATIONS
HST /COS observations of the four target AGN weretaken under
HST program ID 12604 (PI: A. Fox), us-ing the G130M/1291 and G160M/1589 settings, togethergiving wavelength coverage from 1132–1760 ˚A. Details ofthese observations are given in Table 1. The COS instru-ment is described in Green et al. (2012). In all cases, thedifferent exposures were taken at different FP-POS posi-tions, to move the spectra on the detector and mitigatethe effects of fixed-pattern noise. The CALCOS pipeline(v2.17.3) was used to process and combine the raw data,yielding a set of co-added x1dsum.fits files that wereused for the analysis. No velocity offsets larger than a fewkm s − were found between interstellar absorption fea-tures in these spectra, i.e. there was no evidence for sig-nificant velocity-scale calibration errors. The COS spec-tra have a velocity resolution R ≈
16 000 (instrumentalFWHM ≈
19 km s − ), and were rebinned by three pixelsto ≈ − pixels for display and measurement. TheS/N ratios of each spectra measured at 1300 ˚A (G130Mgrating) and at 1550 ˚A (G160M grating) are given in Ta-ble 1. A second, orbital night-only reduction of the datawas completed, in which the data were extracted onlyover those time intervals when the Sun altitude (FITSkeyword SUN ALT ) was less than 20 ◦ . This reduction isnecessary to reduce the strong geocoronal emission (air-glow) in O I II α at veloc-ities of | v LSR | .
200 km s − . Finally, the spectra werenormalized around each line of interest using low-orderpolynomial fits to the local continuum.Three of the four target AGN (all except NGC 7714)were observed with VLT/UVES under ESO program ID085.C-0172(A) (PI: A. Fox). The UVES instrument is de-scribed in Dekker et al. (2000). These observations weretaken in Service Mode with Dichroic 1 in the 390+580setting, no binning, a 0.6 ′′ slit, and under good seeingconditions (FWHM < ′′ ), giving wavelength coveragefrom 3260-4450 ˚A and 4760–6840 ˚A. The spectral resolu-tion in this setting ( R ≈
70 000) corresponds to a ve- Fox et al.locity resolution FWHM=4.3 km s − . The data werereduced using the standard UVES pipeline (based onBallester et al. 2000) running in the Common PipelineLibrary (CPL) environment, using calibration framestaken close in time to the corresponding science frames.The S/N per resolution element of the UVES data at3930 ˚A (near Ca II ) is ≈
68 toward RBS 144, ≈
40 towardHE 0056–3622, and ≈
101 toward PHL 2525.For our 21 cm analysis, we use spectra from tworadio surveys, the Leiden-Argentine-Bonn (LAB) sur-vey (Kalberla et al. 2005) and the GASS survey(McClure-Griffiths et al. 2009; Kalberla et al. 2010) ,using the closest pointings to our target directions. Us-ing measurements from two radio telescopes allows usto investigate beamsize effects, which limit the precisionwith which one can derive the H I column density (and,in turn, the metallicity) in a pencil-beam direction. TheLAB survey observations were taken with the 30 m VillaElisa telescope with a beam size of 30 ′ . The GASS obser-vations were taken with the 64 m Parkes telescope witha beam size of 14.4 ′ .The H I columns in the Stream were calculated fromthe equation N (H I )=1 . × cm − R v max v min T b d v (e.g.Dickey & Lockman 1990), where T b is the brightnesstemperature in Kelvin, v min and v max are the velocityintegration limits in km s − , and the line is assumed tobe optically thin. The differences in the H I columnsderived from the LAB and GASS spectra are visible inTable 1. We adopt these differences as the beamsize un-certainty in the H I column density in the MS along theline-of-sight to each AGN. This uncertainty translates di-rectly to the derived metal abundances, and we includeit as a systematic error in our results, although towardRBS 144 we are able to confirm the MS H I column inthe pencil-beam line-of-sight using a fit to the dampingwings of Lyman- α (see § MAGELLANIC STREAM ABSORPTION
Figure 2 shows the VLT/UVES data covering Ca II ,Na I , and Ti II for the RBS 144, PHL 2525, andHE 0056–3622 directions. Figures 3, 4, 5, and 6 showthe HST /COS data for the same three directions plusNGC 7714, for which no UVES data are available. Ineach case we include the 21 cm emission-line profile forcomparison, from either the GASS or LAB survey. The21 cm data are shown unbinned in Figure 2, but for dis-play are rebinned to five pixels in Figures 3 to 6. Wenow discuss the measurement techniques and present anoverview of the MS (or AVC) absorption seen in eachdirection.
Measurements
We used the apparent optical depth (AOD) techniqueof Savage & Sembach (1991) to measure the absorptionin the LSR velocity range of the MS for each metalline of interest in the COS and UVES data sets. Inthis technique, the AOD in each pixel as calculated as τ a ( v )=ln [ F c ( v ) /F ( v )], where F c is the continuum leveland F is the flux. The apparent column density in eachpixel is then given by N a ( v ) = ( m e c/πe )[ τ a ( v ) /f λ ] = Table 2
MS Column Densities and Ion Abundances toward RBS144Ion Line log (X/H) ⊙ a IP b log N a (MS) [X i /H] MSc (˚A) (eV) ( N a in cm − )H I
21 cm 0.0 13.6 20.17 d · · · O I − > > − I − < < − II − > > − II − ± e − ± e Ni II − < < − II − > > − II − > > − II − > > − II − ± − ± II − ± − ± III − > > − II − ± f − ± f S II − < < − II − < < − II − ± − ± II − ± − ± I − < < − IV − < < − IV − < < − IV − < < − IV − < < − V − < < − V − < < − II − ± − ± II − ± − ± I − < < − . I − < < − . II − < < − . Note . — LSR velocity integration range for MS in this directionis 65 to 210 km s − . For O I II a Solar photospheric abundance (Asplund et al. 2009). b Ionization potential X i → X i +1 . c [X i /H] MS =log [X i /H I ] MS –log (X/H) ⊙ . Upper/lower limits are3 σ /1 σ . These abundances are not corrected for ionization. Only thestatistical error is given. d Value given is GASS survey measurement (14.4 ′ beam). LAB surveygives log N (H I )=20.27 (30 ′ beam). e Line potentially saturated. f Velocity range adjusted to 65–135 km s − to avoid blend. . × [ τ a ( v ) /f λ ] cm − (km s − ) − , where f isthe oscillator strength of the transition and λ is thewavelength in Angstroms. This can be integrated overthe profile to give the apparent column density, N a = R v max v min N a ( v )d v . The apparent column densities in theMS derived from this technique are given in Tables 2, 3,and 4 for the RBS 144, NGC 7714, and PHL 2525 direc-tions. The apparent column densities in the AVC towardHE 0056–3622 are given in Table 5.All our data have sufficient S/N (see Table 1) to lie inthe regime where the AOD method is reliable (Fox et al.2005b), so long as no unresolved saturation is present.This could occur if narrow unresolved lines are presentin the profiles, which is more likely for the COS data thanthe UVES data. In cases where doublets or multipletsof the same ion are present in the data (e.g. Si II ), weused these to check for saturation. When lines are notdetected at 3 σ significance, we give upper limits on N a based on the noise measured in the continuum.For the UVES data, Voigt-component fits to the op-tical (Ca II , Ti II , and Na I , where present) lines wereused to determine the low-ion component structure in theagellanic Stream Abundances 5 Figure 2.
VLT/UVES spectra of RBS 144, PHL 2525, and HE 0556–3622. The top panels show the 21 cm H I spectra from the LAB(black) and GASS (blue) surveys. All other panels show the UVES spectra in black, and our Voigt-profile fits convolved with the linespread function in red. The total column densities in the MW, MS, and/or AVC components are annotated above the spectra, with thenumber of velocity components given in parentheses. Red tick marks show the centroids of each velocity component. We fit the Ti II MSabsorption toward RBS 144 with a single component (bottom-left panel), but it is not a 3 σ detection, so we treat is as an upper limit inthe analysis. The vertical dotted lines show the velocity centroids of the MW and MS 21 cm emission components. Note the compressedy-scale on each panel. Stream, using the
VPFIT software package . These fitsaccount for the UVES instrumental line spread function,assumed to be a Gaussian with a FWHM of 4.3 km s − .Each ion was fit independently, with the number ofcomponents chosen to be the minimum necessary tomatch the data (see Figure 2): three MS componentsfor RBS 144 and two AVC components for HE 0056–3622.Note that we use the UVES data to derive the componentstructure because of their superior velocity resolutioncompared to the COS data (4.3 km s − vs ≈
19 km s − FWHM), though we note that Ca II does not necessar-ily trace the same low-ion phase as the UV metal linesbecause of its low ionization potential. MS Absorption toward RBS 144
The central LSR velocity of the MS in the RBS 144 di-rection is 92 km s − , as defined by the peak of the GASS21 cm emission profile. The H I profile extends over therange 75–140 km s − . The UV absorption lines show ab-sorption centered near the same velocity, but covering amore extended interval of 65–210 km s − (Figure 3).The following UV metal lines are detected in the MS:O I II II III II II II ∼ rfc/vpfit.html. eral of these lines are saturated; the lines which appearunsaturated, from which we derive ionic column densi-ties, are Si II II II II II IV IV V II II I I − . A secondary component at 180 km s − isseen in C II , Si II , and Si III , but not in O I or S II , orH I
21 cm (Figure 3).The VLT/UVES data give MS detections of Ca II II I II profile, spreadover 25 km s − , a much narrower interval than the UVmetal lines, with narrow b -values of 6.1 ± − ,3.0 ± − , and 6.2 ± − . These lines tracethe cool-gas component structure in the Stream (see Pa-per II). An offset of 6 km s − is seen between the velocity Fox et al. Figure 3.
HST /COS metal-line profiles in the UV spectrum of RBS 144, plus the H I
21 cm profile from the GASS survey. Normalizedflux is plotted against LSR velocity for each absorption line. Gray shading indicates the MS absorption velocity interval (65–210 km s − ),and the vertical dotted red line shows the central velocity of the MS 21 cm emission. MW absorption in each panel is visible near 0 km s − .Blends are indicated at the appropriate velocity with a tick mark and accompanying label. The apparent column density of MS absorptionis indicated in the lower corner of each panel, with upper limits given for non-detections. For O I II centroids of the H I and Ca II MS components (see Fig-ure 2): the H I MS emission is centered at 92 km s − (in both the LAB and the GASS data), whereas theCa II absorption is centered at 98 ± − . Further-more, the two higher-velocity components seen in Ca II at 107 ± − and 120 ± − show no analogues inthe 21 cm data. This indicates that small-scale structureexists in the pencil-beam sightline that is not resolved inthe 21 cm beam.However, since the H I column in the MS in thisdirection is high enough to contribute to the damp-ing wings on the Galactic Lyman- α line, we can de-rive N (H I ) MS in the pencil-beam line-of-sight by fit-ting the observed Lyman- α profile with a two-component(MW+MS) model, following Lehner et al. (2008). Thisis important since it eliminates the systematic error onmetallicities that derives from comparing UV lines mea-sured with an infinitesimal beam with 21 cm H I mea-surements made with a finite beam. We fix the ve- locity components at –5 ± − (MW component)and 95 ± − (MS component) to follow the com-ponent structure seen in the UV metal lines, and al-low only the H I column density in each component tovary. Using a fourth-order polynomial fit to the con-tinuum, we find the Lyman- α profile is well reproduced( χ ν =0.88) by a model with log N (H I ) MW =20.00 ± N (H I ) MS =20.09 ± N (H I ) MS =20.17within 2 σ , illustrating that in this direction the system-atic uncertainty on log N (H I ) MS due to beamsize effectsis small ( . N (H I ) MS along theother three sightlines in this paper since in those direc-tions N (H I ) MW ≫ N (H I ) MS .agellanic Stream Abundances 7 Figure 4.
Same as Figure 3, for the NGC 7714 direction. Gray shading indicates MS absorption (–420 to –190 km s − ), and the verticaldotted red line shows the central velocity of the MS 21 cm emission. The inset in the 21 cm panel shows a zoom-in around the MScomponent. MS Absorption toward NGC 7714
NGC 7714 is an On-Stream direction showing a well-defined 21 cm emission component at –320 km s − withlog N (H I ) MS =19.09 measured from the GASS data(18.93 measured from the LAB data), and UV absorp-tion centered at the same velocity but extending over awider interval of –420 to –190 km s − (Figure 4). TheUV lines detected in the MS are O I II II III IV IV MS Absorption toward PHL 2525
The LAB 21 cm data toward PHL 2525 (30 ′ beam) show a very weak emission componentlog N (H I ) MS =18.24 at –260 km s − (see Figure 5, top-left panel). We identify this component with theMS. This component is not visible in the GASS data(14.4 ′ beam), which give log N (H I ) MS < .
21 (3 σ ),marginally inconsistent with the LAB data. UV absorp-tion covering the interval –300 to –100 km s − is seen inC II II III IV IV II ,Na I , and Ti II (Figure 2). There is a clear offset of ≈
50 km s − between the center of the UV absorptionand the LAB H I emission, further indicating thebeamsize mismatch. Because of this issue, this sightlineis of little use for deriving robust abundances. However,the combination of the non-detection of O I I emission allows us to place an upper limit on theMS metallicity (see § AVC Absorption toward HE 0056–3622
In this direction high-velocity absorption is detectedin many UV lines centered at 150 km s − and cover- Fox et al. Figure 5.
Same as Figure 3, for the PHL 2525 direction. Gray shading indicates MS absorption (–280 to –120 km s − ), and the verticaldotted red line shows the central velocity of the (weak) MS 21 cm emission. No O I MS can be derived. ing the interval 80–200 km s − (Figure 6). As dis-cussed in §
1, this absorption appears to trace the AVCsfound near the South Galactic Pole (Putman et al. 2003).These AVCs may have an origin with the Sculptorgroup of galaxies (Mathewson et al. 1975), whose veloc-ities are are low as 70 km s − , or may represent frag-ments of the MS (Haynes & Roberts 1979) that haveseparated kinematically from the principal filaments.The 21 cm emission component from the AVC is weak,with log N (H I ) AVC =18.70 measured over the interval80–200 km s − in the GASS data. The AVC is de-tected in absorption in O I II II III II II II II IV IV I II I II II II II ab-sorption centered near 150 km s − (Figure 2), but theCa II absorption does not cover the full velocity inter- val of the AVC absorption seen in the UV lines, whichshow one component centered at 100 km s − , and a sec-ond weaker component seen in Si III II − . CHEMICAL ABUNDANCES IN THE MS
The UV data contain a rich variety of diagnostics on el-emental abundances in the Stream. Among the availableabundance indicators, the O I /H I and S II /H I ratiosare considered the most reliable, being the least affectedby dust and ionization effects (Field & Steigman 1971;Savage & Sembach 1996; Meyer et al. 1998; Jensen et al.2005; Jenkins 2009). However, at the high H I columns[log N (H I ) ≈
20] and low metallicities (O/H ≈ I I I column ismeasurable. In this regime, S II /H I is the better metal-licity indicator. Thus the metallicity indicator we use ineach sightline depends on the H I column.Toward RBS 144, we measure log N a (O I ) > I I /H I ] MS > − .
17. Turn-ing to S II /H I , we use the S II column mea-agellanic Stream Abundances 9 Figure 6.
Same as Figure 3, for the HE 0056–3622 direction. Gray shading indicates anomalous-velocity cloud (AVC) absorption (80–200 km s − ). MS absorption in this direction is centered near –10 km s − and overlaps with Galactic foreground absorption. The inset inthe 21 cm panel shows a zoom-in around the AVC component. sured from an AOD integration of S II II /H I ] MS = − ± ± I column densities and accounts for the beamsize mis-match between the radio and UV observations (see F10for a more detailed discussion of the size of this system-atic error). Adding the two errors in quadrature to givean overall uncertainty gives [S II /H I ] MS = − ± II σ significancein the co-added spectrum (Figure 3), but is seen in bothindividual sub-exposures. A non-detection would serveto lower the derived metallicity. For this S II − instead of 65–210 km s − , to avoid blendingwith Galactic Si II I N (N I ) < I /H I ] MS < − .
90 (3 σ ).Nitrogen is therefore underabundant in the gas phase ofthe Stream, with [N/S] MS =[N I /S II ] MS < − .
77 (3 σ ). Toward NGC 7714, we measurelog N a (O I )=14.54 ± I I /H I ] MS = − ± ± − ± N a (O I ) < σ ) in theMS, again from night-only O I I /H I ] MS < − σ ). Toward HE 0056–3622,we measure log N a (O I )=14.36 ± I I /H I ] AVC = − ± ± − ± I II abundancemeasured from the non-detection of S II II /H I ] AVC < +0.55 (3 σ ), consistentwith the O I /H I abundance, but not constraining. Ion Abundances
Table 3
MS Column Densities and Ion Abundances toward NGC7714Ion Line log (X/H) ⊙ a IP b log N a (MS) [X i /H] MSc (˚A) (eV) ( N a in cm − )H I
21 cm 0.0 13.6 19.09 d · · · O I − ± − ± I − < < − II − ± − ± II − < < − II − < < II − ± − ± II − ± − ± II − ± − ± II − < < − III − ± − ± II − < < II − < < − II − < < − IV − ± − ± IV − ± − ± IV − ± − ± IV − < < − V − < < − Note . — LSR velocity integration range for MS in this directionis -420 to -190 km s − . For O I II a Solar photospheric abundance (Asplund et al. 2009). b Ionization potential X i → X i +1 . c [X i /H] MS =log [X i /H I ] MS –log (X/H) ⊙ . Upper/lower limits are3 σ /1 σ . These abundances are not corrected for ionization. Onlythe statistical error is given. d Value given is GASS survey measurement (14.4 ′ beam). Table 4
MS Column Densities and Ion Abundances toward PHL2525Ion Line log (X/H) ⊙ a IP b log N a (MS) [X i /H] MSc (˚A) (eV) ( N a in cm − )H I
21 cm 0.0 13.6 18.24 d · · · O I − < < − I − < < − II − ± − ± II − < < − II − < < II − ± − ± II − ± − ± II − < < − II − < < III − > > − II − < < II − < < II − < < II − < < IV − ± − ± IV − ± − ± IV − ± − ± V − < < − V − < < − Note . — LSR velocity integration range for MS in this directionis -280 to -120 km s − . For O I II a Solar photospheric abundance (Asplund et al. 2009). b Ionization potential X i → X i +1 . c [X i /H] MS =log [X i /H I ] MS –log (X/H) ⊙ . Upper/lower limits are3 σ /1 σ . These abundances are not corrected for ionization. Onlythe statistical error is given. d Value given is GASS survey measurement (14.4 ′ beam). Table 5
AVC Column Densities and Ion Abundances toward HE0056-3622Ion Line log (X/H) ⊙ a IP b log N a (MS) [X i /H] MSc (˚A) (eV) ( N a in cm − )H I
21 cm 0.0 13.6 18.70 d · · · O I − ± − ± I − < < − II − > > − II − ± − ± II − < < II − > > − II − > > − II − ± − ± II − ± − ± II − ± − ± III − > > − II − < < II − < < II − < < − II − < < − I − < < − IV − < < − IV − < < − IV − < < − V − < < − V − < < − II − ± − ± II − ± − ± I − < < − . I − < < − . II − < < Note . — LSR velocity integration range for AVC in this directionis 80 to 200 km s − . For O I II a Solar photospheric abundance (Asplund et al. 2009). b Ionization potential X i → X i +1 . c [X i /H] MS =log [X i /H I ] MS –log (X/H) ⊙ . Upper/lower limits are3 σ /1 σ . These abundances are not corrected for ionization. Onlythe statistical error is given. d Value given is GASS survey measurement (14.4 ′ beam). LAB surveygives log N (H I )=18.87 (30 ′ beam). Figure 7.
Derivation of N (H I ) toward RBS 144 from a two-component (MW+MS) fit to the damping wings of Lyman- α (shown in yellow), using a fourth-order polynomial fit to the contin-uum (shown in blue). The upper and lower panels show the raw andnormalized profiles, respectively. The strong emission feature nearthe center of each panel is geocoronal Lyman- α emission. The fityields log N (H I ) MW =20.00 ± N (H I ) MS =20.09 ± − . agellanic Stream Abundances 11In this sub-section we focus on the low-ion (singlyionized) species and their relative proportion with H I .These ratios form empirical indicators of the gas-phaseabundances. They are listed in Tables 2, 3, 4, and 5 andare plotted in Figure 8 for ten low-ionization species, inorder of increasing atomic number.The top panel of Figure 8 compares the MSion abundances measured toward RBS 144 (this pa-per; log N (H I ) MS =20.17) and Fairall 9 (PaperII; log N (H I ) MS =19.95) with the compilation ofLMC and SMC reference abundances presented byRussell & Dopita (1992), which includes abundancesmeasured in stars and in interstellar emission-line objectssuch as supernova remnants and H II regions. Compar-ing the RBS 144 and Fairall 9 data points, a clear patternis seen, in which the ion abundances (or limits) are lowertoward RBS 144 than toward Fairall 9 for all ions shown.This indicates that as the H I column goes down, the low-ion metal columns do not decrease in linear proportion.The MS ion abundances measured toward RBS 144 areconsistently lower than the current-day elemental abun-dances in the LMC and SMC. The difference is smallestfor sulfur, which shows a 0.6 dex difference between theMS and the current-day SMC value. The MagellanicCloud abundances vary substantially from one elementto another, which complicates the interpretation of theMS abundance pattern.In the lower panel of Figure 8, we compare the dustdepletions of each element measured toward RBS 144and Fairall 9 with those measured in the MagellanicBridge by Lehner et al. (2001) and Lehner (2002), andto the compilation of Milky Way halo depletions pre-sented by Welty et al. (1997), which was based onSavage & Sembach (1996) and Fitzpatrick (1996). TheBridge measurements are made in the direction of O-star DI 1388, whereas the halo cloud measurements rep-resent the averages over a sample of many sightlines. Wedefine the depletion of element X relative to sulfur as δ (X) ≡ [X/S II ]=[X/H]–[S II /H]. The depletions measuredin the two MS sightlines are generally similar, with anotable discrepancy for Ca II : δ (Ca) is 0.44 dex lowertoward Fairall 9 than toward RBS 144. The Bridge andStream show a depletion pattern that is consistent for allions shown, particularly for δ (Fe) and δ (Si), where val-ues rather than limits are available. The MS dust deple-tion pattern toward RBS 144 is also consistent with theGalactic halo depletion pattern for all elements shownexcept Nitrogen, which is underabundant in the MS inthis direction relative to the Galactic halo by & Gas-to-dust ratios
A useful number to derive when discussing the chem-ical properties of the Stream is the gas-to-dust massratio, which is directly related to the depletion of Featoms onto dust grains. We define this ratio follow-ing Wolfe et al. (2003) as G/D ≡ N (H)/ N (Fe) d where N (Fe) d = N (S)(Fe/S) ⊙ (1–10 [Fe / S] ) is the column of ironin dust grains. This assumes that a solar Fe/S ratio ap-plies to the entire (dust+gas) system. The single sight-line in our sample with both Fe II and S II directions(and hence for which we can calculate G/D) is RBS 144;in this direction, we derive (G/D) MS =600000 +320000 − . We can normalize this to the Galactic ISM valueby writing (G/D) norm =(G/D) MS /(G/D) MW . Since(G/D) MW ≈ (Fe/H) − ⊙ as Fe is almost entirely de-pleted in the Galactic ISM, we find (G/D) norm =19 +10 − .Along the nearby Fairall 9 sightline discussed in PaperII, we find a lower ratio (G/D) MS =104000 +17000 − and(G/D) norm =3.3 +0 . − . . This corresponds to a higher dustcontent toward Fairall 9, as expected given the higher MSmetallicity in that direction.For comparison, measurements of (G/D) norm inthe SMC lie in the range ≈ norm ≈ Ionization Corrections
Up to now we have considered ion abundances, not elemental abundances. The abundance of the speciesO I and S II are robust indicators of the elementalabundances of O and Si, as discussed above, but forother species, ionization corrections must be appliedto derive the elemental abundances. These correctionscan be derived under the standard assumption that thelow-ionization species (singly and doubly ionized) arephotoionized by the incident radiation field. We rana set of photoionization models using Cloudy v10.00(Ferland et al. 1998) to investigate this process, assum-ing the gas exists in a plane-parallel slab of uniform den-sity. We follow the two-step method outlined in F10.First, we solve for the value of the ionization parameterlog U ≡ log ( n γ /n H ) necessary to produce the observedSi III /Si II ratio, where n γ is the density of H-ionizingphotons ( λ <
912 ˚A). Second, we solve for the abundancesof all the low-ion elements by comparing their observedcolumns with those predicted by the model.This method does not assume a priori that the heavyelement abundances are in their solar ratios; instead itsolves for the abundance of each element separately, as-suming the low-ions arise in the same gas phase as theH I . See Nigra et al. (2012) for a discussion of effectsoccurring if clumpiness changes this assumption. Themodels combine the UV background radiation calculatedby Haardt & Madau (2012) with a model of the escap-ing ionizing Milky Way radiation from Fox et al. (2005a),which was based on Bland-Hawthorn & Maloney (1999,2002). The MW radiation field is non-isotropic, with theescape fraction highest normal to the Galactic disk, so n γ changes with latitude (Fox et al. 2005a).We ran Cloudy models for two of the four sight-lines in this paper, RBS 144 (chosen to model the MSsince it has the most UV metal lines detected) andHE 0056–3622 (to model the AVCs near the South Galac-tic Pole). For the RBS 144 sightline, we use a modelwith l, b, R =299.5 ◦ , –65.8 ◦ , 50 kpc appropriate if theStream is at the same distance as the Magellanic Clouds,which gives log ( n γ /cm − )=–5.61. For the HE 0056–3622sightline, we adopt l, b, R =293.7 ◦ , –80.9 ◦ , 50 kpc givinglog ( n γ /cm − )=–5.34. We do not include the radiationescaping from the Magellanic Clouds or the radiation2 Fox et al. Figure 8. Upper panel : comparison of MS ion abundances measured toward RBS 144 (this paper) and Fairall 9 (Paper II) with theSMC and LMC abundance patterns presented by Russell & Dopita (1992). The solar, average LMC (0.46 solar), and average SMC (0.22solar) abundances are plotted as horizontal lines. Consistently lower ion abundances are seen toward RBS 144 [log N (H I ) MS =20.17] thantoward Fairall 9 [log N (H I ) MS =19.95]. Lower panel : comparison of the dust depletion levels δ (X)=[X/S II ] measured in the MS towardRBS 144 and Fairall 9 with the dust depletion pattern measured in the Magellanic Bridge by Lehner (2002), and with the Galactic halodepletion compilation of Welty et al. (1997). All values are uncorrected for ionization. The data points have been offset in the x-directionfor clarity. See § produced by the shock cascade or interface regions withinthe Stream. The inclusion of additional radiation fieldswould increase the photon density, so that for a given ion-ization parameter the clouds would become denser andsmaller.The results of our Cloudy models are given in two setsof figures. First, Figure 9 shows the ionization correc-tions for six low ions calculated from the
Cloudy models,as a function of log U , for the RBS 144 and HE 0056–3622 directions. We also include a model appropriateto the Fairall 9 sightline, for use in the discussion andfor comparison to Paper II. The ionization correctionsare defined as the difference between the intrinsic el-emental abundance and the measured ion abundance,i.e. IC(X i )=[X/H]–[X i /H I ]. Second, Figures 10a (forRBS 144) and 10b (for HE 0056–3622) show the columnsof the observed low ions as a function of log U , with thederived abundances of each element and the best-fit log U annotated on the panel. The results drawn from thesetwo sets of figures are discussed in the next two sub-sections. Cloudy Results toward RBS 144
We measure a ratiolog [ N (Si III N (Si II & –0.56 over thefull MS velocity interval 65–210 km s − . Reproducingthis ratio with a Cloudy model yields an ionizationparameter log U & − .
2, corresponding to a gas den-sity log ( n H / cm − ) . − . III N (H II )=19.9. If we repeat the measurement in thevelocity interval 150 to 200 km s − , where Si III II N (Si III N (Si II U & − . n H / cm − ) . − .
8, so there is a ≈ II /H I ] ratio is a good indicator of the overall S abun-dance [S/H] in the MS: in our best-fit model along thissightline at log U & − .
2, the ionization correction isfound to be IC(S II )=0.07 dex, and the correction is < over four orders of magnitude in density (Figure 9,top panel). Physically, this is because the fraction of S inthe form of S +2 or higher is similar to the fraction of Hin the form of H + (see Lu et al. 1998; Howk et al. 2006;Howk & Consiglio 2012, for related discussions).The simulations confirm that Al, Si, and Fe are un-derabundant with respect to S in the MS, as we found inFigure 8 based on the ion abundances alone. As for S, theionization corrections for Al II , Si II , and Fe II derived bythe Cloudy model at log U =–3.2 are small: IC(Al II )=–0.01 dex, IC(Si II )=–0.09 dex, and IC(Fe II )=–0.05 dex,which is a consequence of the high H I column inagellanic Stream Abundances 13 Figure 9.
Ionization corrections for six low ions (O I , S II , Si II ,Al II , Fe II , and Ca II ) in our Cloudy models of the MS to-ward RBS 144 (top) and Fairall 9 (center), and of the AVCs towardHE 0056–3622 (bottom). These corrections show the amount whichmust be added to the observed ion-to-H I ratio to determine theintrinsic abundance. For RBS 144 and Fairall 9, the corrections forall ions shown are . U (derivedby matching the Si III /Si II ratios), indicating that the sub-solarvalues derived for Si/S, Al/S, and Fe/S in the MS can be attributedto dust depletion, not ionization. For HE 0056–3622, the modelssupport the use of O I /H I as a robust abundance indicator evenin the regime log N (H I ) <
19, at least for log U . − .
0. Note thatthe Ca II curves have been offset by –1.0 or –2.0 dex on the y-axisfor comparison to the other ions. the Stream in this direction. Applying these correc-tions we derive gas-phase abundances [Al/H] MS =–1.8,[Si/H] MS =–1.7, and [Fe/H] MS =–1.7, corresponding tomoderate dust depletion levels δ (Al) MS ≈ − . δ (Si) MS ≈− .
6, and δ (Fe) MS ≈ − . Cloudy model is relatively high, [Ca/H] MS =–0.3, even thoughthe Ca II ion abundance is low, [Ca II /H I ]=–2.23 ± +2 (Ca III ). Equivalently, the Ca II ionization cor-rection in the model is large, IC(Ca II )=+1.9. If cor-rect, this would correspond to a negative Ca depletion δ (Ca) MS =–0.7 (i.e., a Ca enhancement relative to S),which would be puzzling for two reasons. First, Ca andS are both α -elements, so no nucleosynthetic differencein their abundances is expected. Second, in the GalacticISM, Ca readily depletes onto dust grains (Welty et al.1996; Savage & Sembach 1996), so sub-solar Ca/S ra-tios are expected in the gas phase, not super-solar ra-tios. In another Stream direction (NGC 7469), F10 found [Ca/O] ≈ II /H I ratio, so the issue is not unique to theRBS 144 direction.The finding of no apparent Ca depletion in the Streamin both these directions probably indicates that thesingle-phase assumption for the low ions, implicit withinthe Cloudy models, breaks down, and that insteadCa II preferentially traces regions of dense, cold neutralmedium (CNM) which are not properly captured by our Cloudy models, whereas S II traces regions of warm neu-tral medium (WNM) and warm ionized medium (WIM).The observation that the Ca II absorption in the MS to-ward RBS 144 occupies a far narrower velocity intervalthan the other low ions supports this idea. For Na I , theproblem is even worse; this ion has an ionization poten-tial of only 5.1 eV, so is destroyed easily even in opticallythick gas, and is therefore only able to survive in denseclumps. We conclude that caution is necessary when us-ing Cloudy to model Na I and Ca II in HVCs (and otherions whose ionization potential is less than 13.6 eV), sincethey likely do not coexist in the same gas phase as the H I and the other low ions (see Paper II and Richter et al.2005, 2011, for more discussion on this point). Cloudy Results toward HE 0056–3622
Integrating over the full AVC velocity interval80–200 km s − , we measure a ratio log [ N (Si III N (Si II U & − .
6, corresponding to a best-fit densitylog ( n H / cm − ) . − . U , themodel has an ionized gas column log N (H II )=19.5, andan ionization fraction H II /(H I +H II )=86%. If we adoptthe ratio log [ N (Si III )/ N (Si II )]=–0.90 measured overthe smaller interval 150–200 km s − , where both Si III II U =–4.0 and log ( n H / cm − )=–1.4, indicating that the error inthe gas density arising because of possible saturation ofSi III ≈ I /H I ]=[O/H] for all ion-ization parameters log U . − .
5, even at the low neu-tral gas column log N (H I ) AVC =18.70 present in thissightline (see Viegas 1995). The ionization correctionsfor Al II , Si II , and Fe II at the best-fit log U =–3.6 are more significant in this sightline than towardRBS 144 due to the lower H I column (Figure 9, bot-tom panel): IC(Al II )=–0.8 dex, IC(Si II )=–0.6 dex, andIC(Fe II )=–0.4 dex. Using these corrections we derivegas-phase abundances [Al/H] AVC =–1.4, [Si/H]
AVC =–1.1,and [Fe/H]
AVC < − .
9, corresponding to (small) deple-tions relative to oxygen of δ (Al) AVC ≈ − . δ (Si) AVC ≈ .
0, and δ (Fe) AVC < +0 .
1. There is therefore no evidencefor dust depletion in the AVC toward HE 0056–3622. DISCUSSION
The one-tenth-solar metallicity in the Stream mea-sured from [S II /H I ] toward RBS 144 and from [O I /H I ]toward NGC 7714 matches the value measured from[O I /H I ] toward NGC 7469 by F10. These measurementsare shown in Figure 11, where we plot MS metallicityagainst angular distance from the center of the SMC forfour MS directions observed as part of our HST /COSprogram, and one (NGC 7469) observed with
HST /STIS.
There are therefore three independent measurements ofone-tenth-solar metallicity along the main body of the
Figure 10.
Detailed
Cloudy photoionization models of the low-ionization species observed in the MS toward RBS 144 (left), and in theAVCs toward HE 0056–3622 (right). These models assume the low ions exist in a single uniform-density phase exposed to the combinedMilky Way plus extragalactic radiation field. The Si
III /Si II ratio is used to derive the ionization parameter in the plasma (log U & − . U & − . U ; the amount by which each line is scaled determines the gas-phase abundance of that element,as annotated on the plot. In turn, the comparison of the gas-phase abundances with the expected solar ratios indicates the depletion ofeach element onto dust grains. Small offsets have been applied to the observations in the x-direction for clarity. See § MS, and a fourth (PHL 2525) with an upper limit con-sistent with one-tenth solar.
However, these measure-ments differ from the much higher value [S II /H I ] MS =–0.30 ± ◦ away on the skyfrom RBS 144, in a direction with a fairly similar MS H I column density (19.95 for Fairall 9 vs 20.17 for RBS 144).This corresponds to an abundance variation of a factor offive over a small scale. However, the Stream’s LSR veloc-ity centroid changes by ≈
100 km s − between these twodirections, from 92 km s − toward RBS 144 to 194 km s − toward Fairall 9, indicating that the Stream in the twodirections also has very different kinematic properties.Indeed, if we place our targets on the N08 map ofthe two bifurcated filaments of the Stream, we find thatFairall 9 lies behind the LMC filament, whereas RBS 144lies behind the second filament, matching its positionboth spatially and kinematically (see Figure 1). Thus our0.1 solar metallicity measurement in the RBS 144 direc-tion indicates an SMC origin for this second filament (forthe reasons given in the next paragraph). The 0.5 solarmetallicity measurement in the Fairall 9 direction is con-sistent with the LMC origin for that filament reportedby N08, although the low N/ α ratio in that sightline andits proximity to both Magellanic Clouds suggests a com-plex metal enrichment history (Paper II). This is partic-ularly true in light of recent results that both MagellanicClouds experienced a burst of star formation ≈ ≈ ≈ ≈ ≈ I , which is knownto be present in the Stream (Putman et al. 2003;Stanimirovi´c et al. 2002, 2008; Westmeier & Koribalskiagellanic Stream Abundances 15 Figure 11.
Summary of the metal abundance pattern along the body of the Stream. The upper panels show the GASS, LAB, or GreenBank Telescope (GBT) 21 cm profiles, giving the H I columns in the MS. The middle panels show the profiles of O I II II ≈ ′ , variations of up to 20–25% in N (H I ) determina-tions can occur. In F10, we compared N (H I ) MS deter-minations toward NGC 7469 from four radio telescopesof beam sizes ranging from 9.1 ′ to 35 ′ , and found thevalues varied by 0.15 dex (40%). While important, theseuncertainties are far less than the factor of five differ-ence in MS metallicity we observe between the RBS 144and Fairall 9 directions. Furthermore, for RBS 144 thepencil-beam H I column derived from Lyman- α agreeswithin 2 σ with the H I column derived from GASS, andso the metallicities derived in this direction do not havea large systematic beamsize error. We conclude that thehigher-velocity filament of the MS seen toward Fairall 9has a genuinely higher metallicity than the lower-velocityfilament seen toward RBS 144, despite their proximity onthe sky.Finally, we note that our 0.1 solar metallicity mea-surements in most of the MS are consistent with theabundances measured in the Magellanic Bridge (MB) ofgas connecting the LMC and SMC. Lehner et al. (2008)measured [O/H] MB =–0.96 +0 . − . toward O-star DI 1388,and [O/H] MB =–1.36 toward O-star DGIK 975, althoughthese stars appear to lie in front of the principal H I -emitting phase of the Bridge, so only sample the fore-ground part of it. Our MS metallicity measurementsare also consistent with the results of Misawa et al.(2009), who reported –1.0 < [Z/H] MB < –0.5 toward QSOPKS 0312–770, the sightline to which samples the fullradial extent of the Bridge. However, we note that in re- cent simulations (Besla et al. 2012; Diaz & Bekki 2012)the Bridge and the Stream are created at different pointsin time, so their similar present-day metallicities may bepartly coincidental. SUMMARY
In this first paper of a series presenting combined
HST /COS and VLT/UVES spectroscopy of the Mag-ellanic Stream, we have investigated the Stream’s chem-ical abundances using spectra of four AGN: RBS 144,PHL 2525, NGC 7714, and HE 0056–3622. These sight-lines all lie behind the main body of the Stream, andsample a wide range of H I column density, fromlog N (H I ) MS =20.17 toward RBS 144 down to 18.24toward PHL 2525. Three of these targets (RBS 144,PHL 2525, and NGC 7714) have UV spectra that allowMS abundance measurements. The fourth (HE 0056–3622) cannot be used for this purpose since the MS ve-locity centroid in this direction (–10 km s − ) overlapswith foreground Galactic absorption; however, we havemeasured the anomalous velocity cloud (AVC) absorp-tion centered at 150 km s − in that sightline. A widerange of metal-line species is detected in the Stream in-cluding O I , C II , C IV , Si II , Si III , Si IV , S II , Al II ,Fe II , and Ca II . Our main results are as follows.1. Toward RBS 144, we measure a MS sulfur abun-dance [S/H] MS =[S II /H I ] MS =–1.13 ± MS =[O I /H I ] MS =–1.24 ± MS < − .
63 (3 σ ), based on a non-detection in O I ≈ ≈ independently age it at ≈ Thus we conclude based on chemical ev-idence that most of the Stream originated in theSMC about 2 Gyr ago.
We note that the RBS 144sightline passes through the second (non-LMC) fil-ament identified in H I
21 cm emission by N08, andthus our metallicity measurements resolve the na-ture of this filament’s origin (in the SMC).2. The Stream’s [S/H] abundance we measure to-ward RBS 144 differs by ≈ MS =–0.30 ± ◦ away) in Paper II ofthis series. Fairall 9 lies behind the LMC fila-ment of the Stream (N08). Our Cloudy sim-ulations indicate that ionization corrections arenot the reason for the different abundances: wefind that the ionization correction IC(S II )=[S/H]–[S II /H I ] ≈ | IC(S II ) | < .
01 dex toward Fairall 9,indicating the difference in the abundances is real.Furthermore, the Stream exhibits very differentgas-to-dust ratios and velocity centroids in the twodirections.
This shows that the bifurcation of theStream is seen not only in its spatial extent and itskinematics, but also in its chemical enrichment.
3. We find evidence for dust depletion in the Streamin the form of sub-solar Si/S, Al/S, and Fe/S ratiostoward RBS 144, where S is used as an undepletedreference element. Toward RBS 144, we measuredepletion levels of δ (Al) MS ≈ − . δ (Si) MS ≈ − . δ (Fe) MS ≈ − .
6. These depletions are similarto those found in the Leading Arm by (Lu et al.1998), and indicate that dust grains survive theprocess that ejected the Stream from the Magel-lanic Clouds.4. Toward HE 0056–3622, we measure an oxygenabundance [O/H]
AVC =[O I /H I ] AVC =–1.03 ± v LSR =150 km s − , whichbelongs to a population of anomalous velocityclouds (AVCs) found near the South Galactic Pole.The similarity between this metallicity and the MSmetallicity suggests the AVCs are associated withthe Stream rather than tracing foreground Galacticmaterial. However, we cannot rule out the possibil-ity that they are associated with more distant ma-terial, e.g. in the Sculptor Group, which lies in thisdirection at similar velocities (Mathewson et al.1975; Putman et al. 2003). A detailed study of the Stream’s chemical abundancesin the Fairall 9 direction is presented in Paper II. InPaper III we will address the Stream’s ionization leveland fate, and the more general question of the role oftidal streams in fueling L ∗ galaxies. Acknowledgments
We gratefully thank Justin Ely for providing a night-onlyreduction of the COS data, and we acknowledge valuableconversations with Julia Roman-Duval, David Nidever,Gurtina Besla, Karl Gordon, Mary Putman, and KatBarger. We are grateful to the referee for a useful andinsightful report. Support for program
REFERENCESAsplund, M., Greenness, N., Jacques Sauval, A., & Scott, P. 2009,ARA&A, 47, 481Ballester, P., Modigliani, A., Boitquin, O., et al. 2000, TheMessenger, 101, 31Bekki, K., & Tsujimoto, T. 2012, ApJ, 761, 180Besla, G., Kallivayalil, N., Hernquist, L., et al. 2010, ApJ, 721,L97Besla, G., Kallivayalil, N., Hernquist, L., et al. 2012, MNRAS,421, 2109Bland-Hawthorn, J., & Maloney, P. R. 1999, ApJ, 510, L33Bland-Hawthorn, J., & Maloney, P. R. 2002, ASPC, 254, 267Bland-Hawthorn, J., Sutherland, R., Agertz, O., & Moore, B.2007, ApJ, 670, L109Braun, R., & Thilker, D. A. 2004, A&A, 417, 421Br¨uns, C, Kerp, J., Staveley Smith, L., et al. 2005, A&A, 432, 45Connors, T. W., Kawata, D., & Gibson, B. K. 2006, MNRAS,371, 108Dekker, H., D’Odorico, S., Kaufer, A., Delabre, B., & Kotzlowski,H. 2000, SPIE, 4008, 534Diaz, J. D., & Bekki, K. 2011a, ApJ, 413, 2015Diaz, J. D., & Bekki, K. 2011b, PASP, 28, 117Diaz, J. D., & Bekki, K. 2012, ApJ, 750, 36Dickey, J. M., & Lockman, F. J. 1990, ARA&A, 28, 215Draine, B. T. 2003, ARA&A, 41, 241Ferland, G. J., Korista, K. T., Verner, D. A., et al. PASP, 110,761Field, G. B., & Steigman, G. 1971, ApJ, 166, 59Fitzpatrick, E. L. 1996, ApJ, 473, L55Fox, A. J., Wakker, B. P., Savage, B. D., et al. 2005a, ApJ, 630,332Fox, A. J., Savage, B. D., & Wakker, B. P. 2005b, AJ, 130, 2418Fox, A. J., Wakker, B. P., Smoker, J. V., et al. 2010, ApJ, 718,1046 (F10)Fujimoto, M., & Sofue, Y. 1976, A&A, 47, 263Gordon, K. D, Bot, C., Muller, E., et al. 2009, ApJ, 690, L76Green, J. C., Froning, C. S., Osterman, S., et al. 2012, ApJ, 744,60Gardiner, L. T., Sawa, T., & Fujimoto, M. 1994, MNRAS, 266,567Gardiner, L. T., & Noguchi, M. 1996, MNRAS, 278, 191Gibson, B. K., Giroux, M. L., Penton, S. V., et al. 2000, AJ, 120,1803Haardt, F. & Madau, P. 2012, ApJ, 746, 125Harris, J., & Zaritsky, D. 2004, AJ, 127, 1532Harris, J., & Zaritsky, D. 2009, AJ, 138, 1243Haynes, M. P., & Roberts, M. S. 1979, ApJ, 227, 767Heitsch, F., & Putman, M. E. 2009, ApJ, 698, 1485Heller, P., & Rohlfs, K. 1994, A&A, 291, 743Howk, J. C., Sembach, K. R., & Savage, B. D. 2006, ApJ, 637, 333Howk, J. C., & Consiglio, S. M. 2012, ApJ, 759, 97Hulsbosch, A. N. M., & Wakker, B. P. 1988, A&AS, 75, 191Jenkins, E. B. 2009, ApJ, 700, 1299Jensen, A. G., Rachford, B. L., & Snow, T. P. 2005, ApJ, 619, 891 agellanic Stream Abundances 17agellanic Stream Abundances 17