Mapping the Supernovae Driven Winds of the Large Magellanic Cloud in Hα Emission I
Drew A. Ciampa, Kathleen A. Barger, Nicolas Lehner, Madeline Horn, Michael Hernandez, L. Matthew Haffner, Brianna Smart, Chad Bustard, Sam Barber, Henry Boot
DDraft version January 29, 2021
Typeset using L A TEX twocolumn style in AASTeX63
Mapping the Supernovae Driven Winds of the Large Magellanic Cloud in H α Emission I
Drew A. Ciampa, Kathleen A. Barger, Nicolas Lehner, Madeline Horn,
1, 3
Michael Hernandez, L. Matthew Haffner,
4, 5, 6
Brianna Smart,
7, 6
Chad Bustard, Sam Barber,
1, 9 and Henry Boot
1, 10 Department of Physics & Astronomy, Texas Christian University, Fort Worth, TX 76129, USA Department of Physics, University of Notre Dame, Notre Dame, IN 46556, USA Department of Astronomy, Smith College, Northampton, MA 01063, USA Embry-Riddle Aeronautical University, Daytona Beach, FL 32114, USA Space Science Institute, Boulder, CO 80301, USA Department of Astronomy, University of Wisconsin-Madison, Madison, WI 53706, USA Department of Physics, Astronomy, and Mathematics, University of Hertfordshire, Hatfield AL10 9AB, UK Department of Physics, University of Wisconsin-Madison, Madison, WI 53706, USA Trinity Valley High School, Fort Worth, TX 76132, USA Burleson High School, Burleson, TX 76028, USA
Submitted to ApJABSTRACTWe present the first spectroscopically resolved H α emission map of the Large Magellanic Cloud’s(LMC) galactic wind. By combining new Wisconsin H-alpha Mapper (WHAM) observations ( I H α (cid:38)
10 mR) with existing H i ≤ v LSR ≤ +250 km s − , (2) determinedits morphology and extent, and (3) estimated its mass, outflow rate, and mass-loading factor. Weobserve H α emission from this wind to typically 1-degree off the LMC’s H i disk. Kinematically, we findthat the diffuse gas in the warm-ionized phase of this wind persists at both low ( (cid:46)
100 km s − ) and high( (cid:38)
100 km s − ) velocities, relative to the LMC’s H i disk. Furthermore, we find that the high-velocitycomponent spatially aligns with the most intense star-forming region, 30 Doradus. We, therefore,conclude that this high-velocity material traces an active outflow. We estimate the mass of the warm( T e ≈ K) ionized phase of the near-side LMC outflow to be log ( M ionized /M (cid:12) ) = 7 . ± .
15 forthe combined low and high velocity components. Assuming an ionization fraction of 75% and thatthe wind is symmetrical about the LMC disk, we estimate that its total (neutral and ionized) mass islog ( M total /M (cid:12) ) = 7 .
93, its mass-flow rate is ˙ M outflow ≈ . M (cid:12) yr − , and its mass-loading factor is η ≈ .
54. Our average mass-loading factor results are roughly a factor of 2.5 larger than previous H α imaging and UV absorption line studies, suggesting that those studies are missing nearly half the gasin the outflows. Keywords:
ISM: kinematics and dynamics — ISM: outflows — galaxies: evolution — galaxies: indi-vidual: Large Magellanic Cloud INTRODUCTIONGalactic feedback, such as stellar winds, supernovae,and active galactic nuclei, expel both energy and mo-mentum into the interstellar medium (ISM) of the hostgalaxy. These processes can further drive gas out of thegalaxies in galactic winds and fountains. As these pro- [email protected] cesses cycle gaseous material through the galaxy andinto its surroundings, they transport enriched gas tothe outskirts of the galaxy and into the circumgalac-tic medium (CGM; Heckman 2003, Veilleux et al. 2005,Tumlinson et al. 2011). Furthermore, if the gas thatis ejected into the CGM is lost from the galaxy or ifit stagnates into the galaxy’s halo (e.g., Ford et al.2014; Peeples et al. 2014), the star-formation rate of thegalaxy will likely decline unless it is able to procure ad- a r X i v : . [ a s t r o - ph . GA ] J a n Ciampa et al. ditional gaseous material from different sources. In mostcases, this baryon cycle is difficult to resolve because thegaseous material in the ISM and CGM is faint. However,the nearby Large Magellanic Cloud (LMC) provides anunparalleled view of gaseous material both within andsurrounding its disk. By investigating the gaseous mate-rial in the CGM of the LMC, we can better understandits behavior, enabling us to decipher how the baryoncycle is connected to galaxy evolution.At a distance of about d (cid:12) ≈
50 kpc (Pietrzy´nski et al.2013; de Grijs et al. 2014; Walker 1999), the LMC isclose enough to resolve spatial features within its ISM.The stellar and total mass, M (cid:63) = 3 × M (cid:12) (vander Marel et al. 2009) and M total = 1 . × M (cid:12) (r enclosed = 8 . ◦ (cid:46) i (cid:46) ◦ (Kim et al.1998; Staveley-Smith et al. 2003; Choi et al. 2018), pro-viding an unobstructed view of its interstellar mediumand activity within the disk. Observations have shownnumerous neutral hydrogen super shells and holes thatexist throughout the LMC’s (Meaburn 1980; Kim et al.1999), which could be a result of interactions with theSmall Magellanic Cloud (SMC) and possibly the MilkyWay (MW; e.g., Besla et al. 2010, 2012), as well as recentperiods of intense star formation due to its interactionwith the SMC (Harris & Zaritsky 2009). Each of thesestudies suggests an active history and stellar lifecyclethat could lead to outflows with large amounts of energyand material blown into their surroundings during timesof increased stellar activity (Erb 2015). The additionof energy and momentum from numerous supernovaethroughout the galaxy could be a way for a large-scaleoutflow to originate in a galaxy like the LMC. Observa-tions capturing a complete picture of any galactic-wideoutflowing material proved difficult due to the LMC’ssize on the sky. Observations of the H α emission fromthe ISM of the LMC in prior studies focused primarilyinside the LMC’s disk (Rosado et al. 1990, Laval et al.1992, and Reid & Parker 2012). While these studies re-vealed activity within the ISM of the LMC (Pellegriniet al. 2012 and Winkler et al. 2015), they only observedthe brighter inner region of the LMC rather than thefaint emission from its extended diffuse disk and thegalaxy’s circumgalactic medium.Although the LMC is our nearest gas-rich neighbor-ing galaxy, it was not until recently that many stud-ies began to directly detect signatures of a large-scalegalactic outflow. Recent ultraviolet (UV) absorption- line spectroscopy studies, using the Hubble Space Tele-scope ( HST ) and
Far Ultraviolet Spectroscopic Explorer ( FUSE ), investigated gas flows of the LMC (Howk et al.2002; Lehner & Howk 2007; Lehner et al. 2009; Pathaket al. 2011). Howk et al. (2002) used 12 stars embed-ded within the LMC as background targets to explorethe gas on the near-side of the LMC; they found thatthe absorption along every sightline had kinematic sig-natures consistent with gaseous material flowing out ofthe LMC. A study performed by Staveley-Smith et al.(2003) using the H i ii regionsand super shells, possibly signaling they are a result ofsupernovae in the disk. The blueshifted material, rela-tive to the LMC, was also detected along sightlines thatprojected onto relatively quiescent regions. Lying in thedirection of the LMC, Lehner et al. (2009) found furtherevidence that the high-velocity cloud (HVC) may haveoriginated from an earlier LMC outflow event. They ob-served that this cloud has a velocity gradient with R.A.in H i emission and low-ionization species (see their Fig-ure 5), a similar oxygen abundance as the LMC, and de-pletion patterns that indicate the presence of dust —astrong indicator that this material is of LMC origin.However, Werner & Rauch (2015) were able to de-termine an upper distance limit of an HVC at a sim-ilar velocity that is positioned only a few degrees off-set from the LMC’s disk using absorption-line spec-troscopy toward a halo star at a known distance. Theyfound the cloud along ( l, b ) = (279 . ◦ , − . ◦
1) onlylies d (cid:12) (cid:46) . LSR = 150 km s − is likely of MW origin (Richter et al.2015), this does not eliminate the possibility of two sep-arate HVC complexes. It remains that toward the LMC,an HVC is observed where the H i position-velocity mapshows a physical association with the LMC (Staveley-Smith et al. 2003), not the MW. It is one of the soleHVCs where dust depletion is observed (Lehner et al.2009). The occurrence of both a MW and LMC HVCaround the same projected area is plausible given theprior work and the large angular region being explored.Each of the previously mentioned absorption-linestudies that detected blueshifted material in the di-rection of the LMC use background targets embedded MC Galactic Outflow ii bubbles and super shells), Lehneret al. (2009) (gradient found in the HVC toward theLMC), and Pathak et al. (2011) (strong absorption inOVI across the entire face of the LMC), provide convinc-ing evidence that the LMC drives a global, large-scaleoutflow across its entire disk.Our study supplements the previous UV work by sup-plying the first spectroscopically resolved H α map ofthe LMC and its surrounding environment. This ap-proach removes limitations of previous works that hadsmall numbers of pointings that were dependent on thelocation of background targets, which severely reducedthe spatial coverage of the material they observed. Thework we present is unique in that (1) our observationsare roughly an order of magnitude more sensitive thanprevious studies—enabling us to map the diffuse opti-cal emission from these clouds—and that (2) we wereable to spatially resolve the entirety of the LMC and itssurroundings at an angular resolution of θ resolution = 1 ◦ .Both of these results cannot be performed with the vastmajority of other more distant gas-rich galaxies. Withemission-line observations of the ionized component ofthe LMC’s IVCs and HVCs, we explore their global mor-phology and kinematic distribution in Section 5. Inthis section, we further assess whether the IVC mate-rial could be associated with an LMC wind origin. InSection 6, we discuss how a portion of the emission is from gas currently being driven from the galaxy as anIVC. This is followed up with an estimate of this mate-rial’s mass as well as it’s role in the LMC’s neighborhood(Sections 6.1 and 6.2). Section 7 discusses a HVC that ismoving at speeds upward of ∆v LMCSR ≈ −
150 km s − .The significance and possible explanation for this mate-rial’s origin is considered in Section 7.1. DATAThis study utilizes archival radio and newly acquiredoptical emission-line observations to trace the neutraland ionized hydrogen gas in and around the LMC.2.1.
Wisconsin H α Mapper
We surveyed the faint H α ( λ H α = 6562 . α Mapper (WHAM)telescope over the +50 ≤ v LSR ≤ +250 km s − veloc-ity range. Equipped with a Fabry-P´erot spectrometer,WHAM is roughly an order of magnitude more sensi-tive than other currently available instruments, enablingus to detect the faint emission at a sensitivity limit of I H α ≈
10 mR per 30 second exposure. WHAM’s sen-sitivity is achieved due to its high throughput result-ing from its 1 degree beam size. By adjusting the gaspressures between the instruments etalon, we can tuneour observations to center on the H α emission-line thatis associated with the LMC. More information aboutthe WHAM telescope and its capabilities are outlinedin Haffner et al. (2003).Our WHAM observations have a θ resolution = 1 ◦ an-gular resolution, which corresponds to a ∆ d resolution ≈ α surveycontains more than 6,600 individual observations thatare positioned both on and off the LMC’s H i disk, cov-ering an area that spans from ( l, b ) = (246 . ◦ , − . ◦
8) to(315 . ◦ , − . ◦ Radio Data We use the kinematic definition of the LSR in which theSun moves 20 km s − in the direction of (R . A ., DEC . ) =(18 h m . s , ◦ (cid:48) . (cid:48)(cid:48)
8) for the Julian 2000 epoch (J2000). A Rayleigh (R) is a unit of measure for the sur-face brightness of emission lines that is equal to1 R = 10 / π photons cm − sr − s − , which is5 . × − erg s − cm − arcsec − for H α . Ciampa et al.
We use archival H i ;McClure-Griffiths+2009) to probe the neutral hydro-gen gas phase in the LMC at an angular resolutionof θ resolution = 16 (cid:48) , corresponding to an angular areathat is roughly 14 × smaller than that of the WHAMbeam. This survey spans a velocity range of − ≤ v LSR ≤ +500 km s − and has a spectral resolution of∆v resolution = 0 .
82 km s − . The RMS brightness tem-perature noise is T B , RMS = 57 mK, corresponding to a3 σ sensitivity of log (cid:0) N H i / cm − (cid:1) ≈ . − (McClure-Griffiths et al. 2009). In this paper, the survey’s 3 σ limit is far lower than the practical limit used for ourmaps and mass calculation.In addition to the GASS data, we use a combinedH i ≤ v LSR ≤ +375 km s − and has a∆v resolution = 1 . − spectral resolution. Com-pared to the GASS survey, this combined map has acolumn density sensitivity of log (cid:0) N H i / cm − (cid:1) ≈ . − wide line, but a much higherspatial resolution at θ resolution = 1 (cid:48) . This data resolvessmaller physical structures than GASS, improving ourability to discern smaller-scale morphological features inthe disk.We also use data from the Leiden/Argentine/Bonn(LAB) Survey to measure extinction along our sight-lines due to its similar beam size as WHAM. LAB datahas an effective FWHM beam of θ resolution = 35 (cid:48) fordeclination ≤ − ◦ . The survey covers a velocity rangeof − ≤ v LSR ≤ +400 km s − with a spectral res-olution of ∆v resolution = 1 . − . The rms noiseis T B , rms = 0 .
07 K resulting in a 3 σ sensitivity oflog (cid:0) N H i / cm − (cid:1) ≈ . DATA REDUCTIONThe H α reduction process we performed was carriedout in two stages. In the first stage, we used thestandard WHAM reduction pipeline, which performsthe bias subtraction, flat-fielding, ring-summing, cosmicray contamination removal, and air mass corrections.In the second stage, we velocity calibrated our spec- We convert H i brightness temperatures ( T B ) to col-umn densities using the relationship N H i = 1 . × (cid:82) ( T B / K)(dv / km s − ) cm − , which assumes that theseclouds optically thin 21 cm radiation. tra, removed atmospheric signatures from observations,masked out observations affected by foreground stars,and corrected for dust extinction.3.1. WHAM Pipeline
We utilized the WHAM pipeline that is described indetail in Haffner et al. (2003). During this data pro-cessing, pixels warmed by cosmic rays were first re-moved. The circular interference patterns that resultfrom our Fabry-P´erot spectrometer observations weresummed in annuli to produce a linear spectrum thatis a function of velocity. These linear spectra span a∆v = 200 km s − velocity range and are uniformlybinned to ∆v bin = 2 km s − intervals. The pipelinenormalizes the spectra by exposure time, scales themfor the air mass of observations, and applies an intensitycorrection factor to account for sensitivity degradationof the WHAM instrumentation that occurs over time.3.2. Velocity Calibration
Our observations span +50 ≤ v GEO ≤ +250 km s − inthe geocentric (GEO) velocity frame. Over this range,these observations do not overlap with the bright geo-coronal H α line at v GEO = − . − and onlyoverlaps with the blue wing of a bright OH line atv GEO = +272 .
44 km s − . Therefore, we were unable touse either of these lines to calibrate our velocities usingthe method described in Hausen et al. (2002) and Bargeret al. (2013). Instead, we used the velocity calibrationtechnique that is described by Barger et al. (2017) andAntwi-Danso et al. (2020) for WHAM observations thatdo not overlap with bright atmospheric lines at wellestablished transitions. Using this technique, we cali-brated our velocity by monitoring the pressure of theSF gas in the WHAM Fabry-P´erot etalons and by fur-ther refining the calibration by comparing our observa-tions with an atmospheric template.Using the linear relationship between the pressure ofthe Fabry-P´erot etalons and ∆ λ measured by Tufte(1997), we calculated the velocity offset between the rawand geocentric velocity frames. This is essentially the re-verse of our tuning process, which enabled us to calibratethe velocity frame to an accuracy of ∆v GEO (cid:46) − .Because all of our observations were taken at the sametune (i.e., at the same interference order), the relativevelocities of the calibrated observations agree within0 . − of each other as described in Barger et al.(2017). We improved our calibration further by aligningour observations with the faint atmospheric lines in theatmospheric template presented by Barger et al. (2013)(see their Figure 3). This enabled our velocity solutionto be calibrated to an accuracy of ∆v GEO (cid:46) − . MC Galactic Outflow
Atmospheric Subtraction
Atmospheric Template
There are significant faint atmospheric emission fea-tures which populate the entire velocity range of our ob-servations. While these lines are abundant, they behavepredictably and vary primarily with air mass. Becauseof this, we modelled these lines using a template cre-ated by Barger et al. (2013) (see their Figure 3), whichcharacterizes the atmospheric emission present in ourobservations. The template was scaled to account fordifferences in air mass between observations and night-to-night variations due to humidity and temperature.Brighter lines are more variable and need to be fitindividually. This includes a bright OH molecular lineat v
GEO = 272 .
44 km s − whose line strength dependson the angle between the Sun and Earth’s upper atmo-sphere. In the direction of the LMC, this bright OHline is overwhelmed by emission from the LMC’s disk.To subtract this emission feature in the direction of theLMC, we kept the area and width of the line fit con-stant so that it matched with off-disk observations thatwere taken during the same night and within roughly 15minutes of the on-disk observations. Because this nar-row OH line is kinematically unresolved by the WHAMtelescope, we fit the line assuming that it has a widthof ∆v OH width = 1 km s − before we convolved it withWHAM’s ∆v WHAM IP ≈
12 km s − instrument profile.We fit the background emission with a constantterm added to the atmospheric template. The to-tal fit —which includes the bright OH line at v geo =272 .
44 km s − , faint atmospheric lines, a flat back-ground, and Gaussian modelled astronomical emission—utilized a chi-square minimization with conservative cri-teria to avoid over-fitting emission features that are notphysically realistic. These criteria account for the linewidth and lower limits for the strength of the astronom-ical emission, which are dependent on the gas tempera-ture and instrument sensitivity, respectively. An exam-ple of the pre and post atmospheric corrected spectrumis shown in Figure 1 with a sightline piercing throughthe LMC. A detailed description of these bright linesand how they were handled, including their origin andthe nature of their variability, is outlined in Barger et al.(2013). 3.3.2. Removal of Systematic Signatures
Following the removal of the atmospheric contami-nation with the atmospheric template described above, we discovered systematic spectral signatures in the re-duced spectra. The cause of these structured residu-als is likely due to very faint, unresolved atmosphericlines that are not described in the atmospheric templateor from a slight velocity misalignment between the ob-served spectra and the atmospheric template causing thesignatures during subtraction. These signatures appearat the same geocentric rest frame (GEO) velocity overnarrow, 5 (cid:46) ∆v (cid:46)
10 km s − , velocity widths. Thesesignatures are visible when the spectra in our survey arestacked across the same velocity range in the geocentricframe, as shown in the top panel of Figure 2. At severalvelocities in this frame, there are multiple relatively co-herent vertical signatures across the spectra. However,the two bright astronomical horizontal structures are as-sociated with the Magellanic Bridge (+150 (cid:46) v GEO (cid:46) +210 km s − and 0 ≤ spectrum channel ≤
35) and theLMC and its wind (+150 (cid:46) v GEO (cid:46) +300 km s − and100 ≤ spectrum channel ≤ α emission above our sensitivity limit. Wethen subtract this average off target spectrum from ourobserved spectra to effectively remove these signatures.To ensure that the off target spectra accurately char-acterized the sky for that region of the map, this pro-cedure was repeated for four Galactic latitudinal sub-regions. We display stacked spectral image with before(top panel) and after (bottom panel) samples of this re-duction process for a subset of our observations in Fig-ure 2.Overall, the vertical atmospheric features at v GEO ≈ +150, +180, and +230 km s − have been greatly re-duced. However, some of this residual atmospheric emis-sion remains, especially at v GEO ≈ +180 km s − forspectrum channels >
350 that correspond to a region onthe sky between ( l, b ) = (265 ◦ , − ◦ ) and (285 ◦ , − ◦ ).These atmospheric residuals persist in these spectra be-cause the Galactic latitude region had few good H α faintoff locations. Although the presence of this lingeringatmospheric emission in our final reduced H α emissionmap is not ideal, it does not impact the final results ofthis paper as we do not use the data in that region ofthe sky in our mass calculation.3.4. Observational Cutoffs and Degradation Correction
We removed sightlines that are within a 0 . ◦
55 angu-lar radius of stars that are m V ≤ . α spectra. This cutoff removes 689 observations from our Ciampa et al.
Geocentric Velocity [km/s] m illi - R a y l e i gh [ m R ( k m / s ) - ] Figure 1. (Top panel) A pre-atmospheric subtracted WHAM spectrum toward ( l, b ) = (279 . ◦ , − . ◦
0) drawn as theblack line. The atmospheric template we used to reduce our H α observations is indicated with a red overlaid line. Theemission contributions associated with the Milky Way at v GEO = +66 .
21 km s − and a bright OH line at a velocity ofv GEO = +272 .
44 km s − are traced with dashed dot gray lines. The solid, thick gray line traces all of the atmospheric emission(atmospheric template and OH line) that we subtracted from the WHAM spectrum during our reduction procedure. (Bottompanel) The final reduced spectrum, where emission from the LMC (v GEO (cid:38) +260 km s − ) is shaded light gray, LMC IVCmaterial (+210 (cid:46) v GEO (cid:46) +260 km s − ) is shaded light blue, LMC HVC material (+150 (cid:46) v GEO (cid:46) +210 km s − ) is shadeddark blue, and the Milky Way (v GEO (cid:46) +90 km s − ) is marked. sample, corresponding to roughly 12% of our observa-tions. Because WHAM is a remote observatory config-ured for queue observing and does not require constantmonitoring during good weather conditions, we doublechecked all of the observations that were taken at theoptimal CCD temperature for the WHAM camera of T CCD = − . ◦ C and not during an automated liq-uid nitrogen fill to reduce noise. We confirmed thatthe etalon pressures were stable and that the monitoredvalues matched the input values during setup for the ob-servations at night. We also ensured that all of the usedobservations were taken when the outside humidity wasless than 85%, at a Zenith Distance less than 75 ◦ , andduring dark time observations. Throughout this pro-cess a total of 291 additional observations (or around4%) of the remaining sample were removed. In total,there were 980 observations (or 14%) removed from oursample leaving 5,931 observations. For sightlines withrepeat observations, we averaged the spectra together.Our survey samples a total of 1,712 unique sightlines, where each sightline was observed an average of 3.5 timeswith the locations toward the LMC’s H i disk sampledthe most.As a result of WHAM instrument degradation overtime, observations suffered up to a 20% decrease in ob-served H α intensity (Smart et al. 2019). The procedureused for determining the WHAM instrument degrada-tion trend with time is outlined in Haffner et al. (2003).However, our intensity correction does not include anight-to-night intensity calibration associated with air-mass variations that are due to variations in atmosphericconditions. This is because there were insufficient cali-bration observations taken each night during this survey,in part because there are few WHAM calibration targetsin the southern sky.3.5. Extinction Correction
Previous absorption-line spectroscopic studies towardLMC stars (e.g., Howk et al. 2002; Lehner & Howk2007; Lehner et al. 2009; Barger et al. 2017) indicatethat the gas clouds surveyed in this study exist be-
MC Galactic Outflow M il k y W a y C o n t a m i n a t i o n M il k y W a y C o n t a m i n a t i o n S p ec t r u m C h a nn e l S p ec t r u m C h a nn e l L M C M B Figure 2. (Top) A latitude sub-region between − ≤ b ≤−
36 containing 466 WHAM spectra before the removal ofresidual signatures. At various locations of v
GEO = +155,+195, and +230 km s − , vertical structures are visible withinthe blue dashed rectangles, labeled 1, 2, and 3. (Bottom) Af-ter the correction there is a large improvement in the removalof the vertical signatures. At those same velocities as the toppanel, the signatures are reduced. In both panels, emissionfrom the LMC is visible as horizontal stripes over channels100–200 while the Magellanic Bridge (MB) is visible fromchannels 0–35 at lower velocities. tween us and the LMC. Additionally, Barger et al.(2017) found compelling evidence that the gas within∆v LOS ≈
100 km s − from the H i disk of the LMC alongthe line of sight is associated with large-scale galacticoutflow. Similarly, the depletion patterns observed byLehner et al. (2009) for the HVC material that is in thesame projected region on the sky also indicates that thiscloud contains dust. We, therefore, applied two extinc-tion corrections: one for attenuation due to the MilkyWay and another for self-extinction of the gas cloudsassuming that they have a chemical composition similarto that of the LMC.To correct for reddening, we used the following rela-tionship from Diplas & Savage (1994) that relates thecolor excess with the average N H i foreground emission: E ( B − V ) = (cid:104) N H i (cid:105) . × atoms / (cm mag) (1)We used the Leiden/Argentine/Bonn Galactic H i sur-vey (LAB; Kalberla et al. 2005; Hartmann & Burton LSR Velocity [km/s] I n t en s i t y [ A DU ( k m / s ) - ] Figure 3.
Four example H α spectra that probe differ-ent positions in our survey. This includes emission toward:(Panel A) 30 Doradus at ( l, b ) = (279 . ◦ , − . ◦ . ◦ , − . ◦ . ◦ , − . ◦ . ◦ , − . ◦ i diskof the LMC in light gray (only Panel C over the displayedLSR velocities) and LMC IVC material out to a line-of-sightvelocity of | v LOS | = 100 km s − of its disk in light blue. i emission.For the MW, we integrated the H i emission over the − ≤ v LSR ≤ +100 km s − velocity range and as-sumed an extinction parameter of R v = 3 .
1. Similar tothe Barger et al. (2013) study, we also adopt the Cardelliet al. (1989) optical extinction curve. Combined, the to-tal extinction for the MW dust is then given by: A (H α ) = 5 . × − (cid:104) N H i (cid:105) cm − atoms − mag , (2)where the extinction corrected intensity is then I H α, corr = I H α, obs e A (H α ) . After correcting for extinctionassociated with foreground MW material, our observedH α intensities increase by roughly 10%.The self extinction by the circumgalactic medium issmall as this gas is diffuse. Following a similar pro-cess to the MW extinction, we can account for the selfextinction of the winds by adopting the extinction pa-rameter for the LMC measured by Gordon et al. (2003)of R v = 3 .
41. We used an R v that is measured forthe LMC’s disk as the IVC and HVC material likely Ciampa et al. originated from this galaxy via stellar feedback events.When we integrated the H i emission across the veloc-ity range of our wind, +100 ≤ v LSR ≤ +225 km s − ,we found that the associated self extinction correctionis much less than 1% and, therefore, we neglected thiscorrection in our mass calculations below. LMCSR VELOCITY FRAMEWhen exploring the circumgalactic material of theLMC, it is useful to use a reference frame centeredaround the LMC disk rather than the LSR frame. We re-fer to this reference frame as the Large Magellanic CloudStandard of Rest (LMCSR) frame. Because the LMC isactively forming stars across its H i disk, the kinematicwidth of its disk varies rapidly at small scales. Acrossa slice through the LMC’s disk at Galactic latitude of b = − . ◦
67 and centered on the 30 Doradus starburstregion, the width varies from 25 (cid:46) ∆v H i (cid:46)
50 km s − (see Figure 4). This multiple component structure com-plicates the process of determining where the disk kine-matically ends and where a wind begins.Across the roughly 10 degree Galactic longitude sliceshown in Figure 4, the motion of the H i gas hasa velocity gradient that spans from +225 (cid:46) v LSR (cid:46) +275 km s − at higher Galactic longitudes of l ≈ ◦ to +275 (cid:46) v LSR (cid:46) +325 km s − at l ≈ ◦ . Alongthis velocity slice, there are several locations containingholes where little to no H i exists. This is not surprisingas these holes can be created by energetic stellar feed-back process activity occurring inside the disk, whichheats and ionizes the surrounding gas. This feedbackcan further drive circumstellar and interstellar materialoutward, possibly contributing to a galactic outflow asnoted by Staveley-Smith et al. (2003). In the LMCSRvelocity frame, the spatially varying velocity gradient inthe LSR frame is removed, which helped us disentanglethe disk and the wind material.To convert our spectroscopic observations from theLSR to LMCSR velocity reference frame, we initiallyused the relationship provided by Lehner et al. (2009),which described the motion of the LMC’s disk stars.However, galaxy interactions have disrupted the LMC’sdisk so that the gaseous and stellar components do notalign. While the Lehner et al. (2009) LSR to LMCSRrelationship works reasonably well for the motion of thegas toward the disk, our observations extend ∆ θ (cid:38) ◦ off the LMC’s gaseous disk on all sides and is no longercentered in that velocity reference frame. Instead, wemodeled the H i emission the LMC and its surroundingsto convert our H α observations into a LMCSR velocityreference frame. This is especially beneficial becausethe gaseous H i emission extends much further than the Figure 4. (Bottom) An H i intensity-weighted position-position map of the LMC. The circle marks the location of30 Doradus at ( l, b ) = (279 . ◦ , − . ◦
67) and the dashed lineindicates the extent of the Galactic Longitude range consid-ered the top-left panel. (Top-left) A position-velocity mapof H i emission running through the location of 30 Doradus.(Top-right) H i spectra toward the 30 Doradus sightline. Thedashed line represents a single ATCA H i spectrum toward30 Doradus while the solid blue curve depicts an averagespectrum for all emission within circular area of 1 degreediameter centered on 30 Doradus. stellar disk and because H α emission tends to kinemati-cally follow the H i emission in HVCs (e.g., Haffner et al.2001; Putman et al. 2003; Hill et al. 2009; Barger et al.2012, 2013, 2017; Antwi-Danso et al. 2020).We determined the motion of the LMC’s H i disk byperforming a Gaussian decomposition of H i GASS spec-tra across our surveyed region. We enforce that theGaussian fits meet the following criteria: each fit musthave a column density above log (cid:0) N H i / cm − (cid:1) ≈ − , and a ve-locity centroid between +175 ≤ v LSR ≤ +325 km s − tobe considered part of the LMC disk. We modelled theH i disk as a simple 2D plane using a least-squares fit.To improve the accuracy of this plane, we weighted ourfit by the H i column density. Our resultant relation-ship between the line-of-sight Galactic longitude ( l ) andlatitude ( b ) and the central velocity offset is:∆v LMCSR km s − = 262 . − .
25 ( l − .
66 ( b −
33) (3)
MC Galactic Outflow
LMCSR =v LSR +∆v
LMCSR . We used the width of the H i lines out to 3 standard deviations to describe the thick-ness of the LMC’s H i disk and to distinguish betweendisk and wind material. The near-side and far-side diskboundaries are described by the difference between the3 standard deviation fitted plane and the central ve-locity (Equation 3). This results in adopting an LMCdisk width of roughly 80 km s − across the face of theLMC, or 40 km s − from the kinematic center of thedisk to its edge. The newly constructed velocity frameimproves our ability to separate the wind material fromthe LMC’s disk that lies in front of the galaxy and toidentify the wind material that extends past its H i disk. KINEMATIC MORPHOLOGY OF H α EMISSIONWe observe blueshifted material at the intermediate-and high-velocities relative to the LMC in both H i andH α emission. This emission is consistent with the resultsof previous UV absorption-line spectroscopy studies thatsuggest a large-scale galactic wind emanating from theLMC, driven by the stellar activity within its disk (Howket al. 2002; Lehner & Howk 2007; Barger et al. 2016).The H i IVC emission is strong toward the LMC’s diskand rapidly decreases radially. The H α emission simi-larly decreases radially away from the LMC yet extendsoff the boundary of the LMC stellar ( r (cid:63) ∼ .
15 kpc) andH i disk at log (cid:0) H i / cm − (cid:1) ≈
19 (see the left-hand panelof Figure 5). This H α emission is asymmetric, relativeto the H i , such that it extends farther along the edgeof the LMC disk near the 30 Doradus starburst region.Gaseous debris has littered the surrounding area ofthe LMC due to its interactions with the SMC. There-fore, in addition to the wind that is likely associatedwith the LMC, there is Magellanic tidal material andMW HVCs that pollute this region of the sky. Athigher Galactic latitudes than the LMC ( b ≥ − ◦ ),there are a few sparse H i clouds that are likely asso-ciated with the Leading Arm (LA) complex LA I near( l, b ) ≈ (283 ◦ , − ◦ ); for more details on the H α distri-bution of these offset clouds, see Smart et al. (2021), inpreparation. Likewise, because the Magellanic Bridgeconnects the LMC and SMC, its emission is presentat l ≥ ◦ with similar velocities as the high Galac-tic longitude edge of the LMC. Toward the southernrim of the LMC disk, there are fragmented clouds cen-tered on ( l, b ) = (273 ◦ , − ◦ ), (278 ◦ , − ◦ ) and(281 ◦ , − ◦ ) (see the right-hand panel of Figure 5).The background halo star that Richter et al. (2015) usedto establish the distance of an HVC ( d (cid:12) ≤ . l, b ) = (279 . ◦ , − . ◦
1) lies within theregion of these low-latitude fragmented clouds. These clouds overlap in velocity with the HVC absorption(v
LMCSR ≈ −
150 km s − ) along this stellar sightline,suggesting that it could be associated with the MW.For this reason, we conservatively do not consider gasthat is more than a few degrees off of the LMC’s H i disk to be part of the LMC outflow.We find the gaseous material toward the LMC thatis the least blueshifted (i.e., kinematically closest to theLMC’s motion) morphologically follows the H i disk ofthe LMC. In Figure 6, we separate the H α into four sep-arate emission maps, each with small integration rangesthat allow us to study the bulk properties of the gascloud in discrete slices of velocity. The emission atv LMCSR ≈ −
100 km s − maintains an intensity well over I H α ≈ . α emission of theIVC is consistent with a large-scale LMC outflow.At higher velocities, approaching the more blueshiftedmaterial, our data shows the emission has a strong spa-tial alignment with the most active star-forming regionwithin the LMC, 30 Doradus (see right two panels of Fig-ure 6). This emission spans over ∆v LOS (cid:38)
150 km s − and remains stable across each channel map toward thestar-forming region. Detecting strong H α emission outto velocities of nearly ∆v LMCSR ≈ −
150 km s − whilealso observing a connection to the lower velocities in thesame region indicates an association with the LMC. INTERMEDIATE VELOCITY GASWe find numerous clouds that are bright in H α emis-sion that are blue shifted by roughly 50 −
100 km s − relative to the LMC’s H i disk (Figures 5 and 6). Most ofthis intermediate velocity material spatially aligns withthe disk of the LMC, resembling a galactic outflow previ-ously suggested (Howk et al. 2002; Barger et al. 2016).Using emission across 1,712 H α sightlines toward theLMC IVC, we determine its mass, its mass-flow rate, andmass-loading factor. In the following LMC IVC masscalculations, we only include the material that is within∆ θ ≈ ◦ of the LMC H i disk with log (cid:0) H i / cm − (cid:1) (cid:38) α emission that is pro-jected multiple degrees off its disk could be associatedwith Magellanic tidal debris (i.e., Magellanic Bridge,Leading Arm) or MW HVCs (see Figure 5 and Section5). 6.1. Mass Estimate of IVC
Barger et al. (2016) estimated the mass of theintermediate velocity outflowing LMC winds to belog ( M ionized /M (cid:12) ) (cid:38) .
16 for the low-ionization species.Because they found that this wind is roughly symmet-rical on either side of the LMC’s disk, this would cor-respond to a mass of log ( M ionized /M (cid:12) ) (cid:38) .
86 for only0
Ciampa et al.
Figure 5. (Top) H α emission maps of the LMC and its surroundings. This map traces the material that is blueshiftedrelative to the LMC. From left to right, the total wind integrated over the − ≤ v LMCSR ≤ −
55 km s − velocity range,the IVC portion integrated over the − ≤ v LMCSR ≤ −
55 km s − velocity range, and the HVC portion integrated over the − ≤ v LMCSR ≤ −
100 km s − velocity range. The overlaid black contours trace the H i emission across the same integrationrange at log( N H i / cm − ) = 19. (Bottom) H i column density map covering the same region and velocity ranges above usingGASS data. Figure 6. H α emission channel maps centered at v LMCSR = − − − −
160 km s − (left to right). Widths of thesevelocity slices are all ∆v = 30 km s − . The bright emission in the right-hand panel at log( I H α / mR) ≈ . l, b ) = (279 . ◦ , − . ◦ i column density contours are drawn at log( N H i / cm − ) = 19 . . the near-side ionized outflow. However, as that study only sampled the wind along two neighboring sightlines MC Galactic Outflow α emission map of the wind of the LMC, we are ableto measure both of these directly.In contrast with the absorption-line work, our H α sur-vey allows us to obtain a mass estimate more naturallyfrom the wind’s density times its volume, M = ρ V.We calculate its mass density using the electron numberdensity as a proxy for the density of protons as they areroughly equal (i.e., n p ≈ n e ) and use a reduced massof µ ≈ . m H to account for the contribution from he-lium and metals. Calculating the volume of the windrequires knowing its solid angle Ω, line-of-sight depth L ,three dimensional geometry (see Figure 8), and distance D . We also include a cos ( i ) factor to account for theinclination of the cross-sectional area of the wind rel-ative to our line of sight. The mass is then given as: M = µn e Ω D L cos ( i ). For gas at the distance of theLMC, the mass enclosed within one WHAM beam withΩ = 1 ◦ is then: M ionized M (cid:12) = 2 . × cos ( i ) (cid:32) D
50 kpc (cid:33) (cid:32) L H + pc (cid:33)(cid:32) n e cm − (cid:33) (4)We estimate the total mass of the wind by sum-ming the single beam mass across the projected out-flow area. For the 1,712 H α spectra that fill our mapof the LMC (Figure 5), we define the morphologicalextent of the LMC’s galactic wind to include regionsthat are within 1 ◦ of its H i disk with neutral columndensities larger than log( N H i / cm − ) ≥ .
0. This re-gion contains 215 WHAM sightlines contained withinroughly 50 deg that are used for our mass estimate.For each of these sightlines, we integrated across the − ≤ v LMCSR ≤ −
55 km s − velocity range to mea-sure the H α intensity of this wind and to explore its spa-tial distribution (see Equation 3). The strength of theH α recombination line is directly proportional to theelectron density squared along the line-of-sight depthand the electron temperature ( T e ) of the gas as: I H α R = 0 . T − . (cid:18) EM pc cm − (cid:19) , (5)where EM is the emission measure ( EM ≡ (cid:82) n e ( s ) ds )and T is given in units of ten thousand Kelvin (i . e ., T =T e / K). Measurements of the average EM and asso-ciated velocity range are given in Figure 1.Since we cannot measure the line-of-sight depth or theelectron density as a function of depth directly, we adopta few necessary assumptions to estimate the mass of thewind that closely follow the procedures used in prior
Table 1.
Observed Velocities and Emission MeasuresOutflow Component v LMCSR (cid:104) EM (cid:105) a (km s − ) (10 − pc cm − )IVC −
100 to −
55 390HVC −
175 to −
100 205 a This is an average emission measure across the corre-sponding velocity range used to calculate the ionizedmass across all sightlines considered to be part of thewind.
WHAM H α studies for evaluating the mass of HVCs(e.g., Hill et al. 2009; Barger et al. 2012, 2017; Smartet al. 2019). 6.1.1. Line-of-Sight Depth
The most difficult of these assumptions pertains tothe depth and line-of-sight distribution of the wind.Past studies of the HVC component of the LMC windobserved similar kinematics for the neutral and low-ionization species (e.g., Lehner et al. 2009). Moreover,observations of both H α and H i in outflows of othergalaxies (e.g. M82; Lehnert et al. 1999 and Schwartz& Martin 2004) support a multi-phase wind that is wellmixed at large scales. In our study, due to our large an-gular resolution at 1 degree, we are spatially resolvingthe wind at the kiloparsec scale and are unable to resolvesmall-scale structure in the wind. We therefore assumethat the neutral gas and ionized gas are well mixed atthe scales we are probing in our survey such that theionized hydrogen depth is roughly equal to the neutraldepth, i.e., L H + ≈ L H i .To estimate this depth, we analyze the fiducial sim-ulations of Bustard et al. (2020) for LMC-specific out-flows (see Figure 7). These magnetohydrodynamic sim-ulations used the observed LMC star-formation historyfrom Harris & Zaritsky (2009) to the seed star clusterparticles that would subsequently deposit the thermal,kinetic, and cosmic ray energy into surrounding gridcells. In these simulations, gas that emerged from theLMC’s disk due to stellar driven outflows experienced anexternal pressure by surrounding coronal gas. Bustardet al. (2020) found that the apparent coronal gas windpushes against the leading edge of the LMC via rampressure and that its effects are strong enough to sup-press the near-side outflows and alter the shape of theLMC’s halo. While this simulation neglects the gravita-tional influence of the SMC and Milky Way, we expectthe depth of the outflow to be primarily influenced by2 Ciampa et al.
Figure 7.
Simulated H α emission maps from Bustard et al.(2020) of the LMC’s galactic wind from an edge-on perspec-tive, using the Trident package (Hummels et al. 2017). Thiswind is assumed to be in photoionization equilibrium withthe UV background (no local ionizing sources from the LMCor Milky Way are included). This model uses the orbital his-tory of the LMC from the models of Besla et al. (2012), thepresent-day infall velocity of the LMC is 258 km s − directededge-on and 194 km s − directed face-on with respect to theLMC; the ambient medium is assumed to be smooth with atotal gas number density of ∼ − cm − at the present-dayLMC distance of d (cid:12) ≈
50 kpc (Salem et al. 2015). (Top)H α emission at a lookback time of 60 Myr. (Bottom) Cur-rent day edge-on view of the LMC and its H α emission. Thearrow in the lower-left corner of each panel represents themotions of the head wind caused by the LMC’s path throughthe MW halo. On each colorbar, the sensitivity of our ob-servations (10 mR) is marked with a horizontal line and anarrow pointed upward. the LMC’s gravitational potential and ram-pressure ef-fects.On a global scale, the galactic winds produced in theBustard et al. (2020) simulations match well kinemat-ically and spatially with the observed LMC outflow.We use the Bustard et al. (2020) results as a guide forconstraining the depth of this wind. They find thatthe 10 K gas in the near-side outflow penetrated to aheight of z wind ≈ z midplane = 0 kpc) at a lookback time of 60 Myrs. Atpresent-day, they find that this wind stalls at a heightof z wind ≈ i disk ( z H i disk ≈ .
75 kpc), this corresponds to present-day wind depth of roughly1 ≤ L wind ≤ (cid:104) n e (cid:105) = (cid:104) EM (cid:105) / L − / + . We then used this den-sity to calculate the outflow’s mass with Equation 4.Although our main uncertainty involves the depth ofthe wind, it is important to note that the mass scalesas M ionized ∝ L − / + , resulting in only a modest vari-ance in mass when we consider a range of reasonabledepths. Moreover, we assume an electron temperaturein the range 0 . (cid:46) T (cid:46) .
2, which is where the H α emission peaks. We further assume that the tempera-ture of the neutral and ionized hydrogen gas are roughlyequal allowing us to relate the neutral and ionized hy-drogen number densities for a given pressure scenario as P ionized /n ionized = P neutral /n neutral under ideal gas con-ditions.Because the LMC is nearly face-on, we cannot con-strain how the morphology of the wind varies with depthas we only see its 2 dimensional projection on the sky.Therefore, we acknowledge three separate volume sce-narios: cylinder, partial outward flaring cone, and par-tial tapered cone (see Figure 8). For the cylindrical windin scenario (a), we simply assume that the radius ofthe wind is constant and matches the extent of the H α emission. We include the outward flaring partial conegeometry in scenario (b) as it has been observed forother galaxies (e.g., M82); in this scenario, we set theinner cone radius to the stellar radius (2.15 kpc) andthe outer radius to match radius of the H α emission.The tapered, inverted cone in scenario (c) is the ge-ometry that resulted for the near-side LMC wind fromBustard et al. (2020) simulation when they accountedfor ram-pressure effects; in this scenario, we match theradii to match the simulation and the H α observations.For these three geometries, the near-side wind masseswould correspond to log ( M ion /M (cid:12) ) = 7 . ± .
14 forvolume (a), log ( M ion /M (cid:12) ) = 7 . ± .
14 for volume (b),a mass of log ( M ion /M (cid:12) ) ≤ . ± .
14 for volume (c).As we cannot observationally determine which of thesewind geometries better matches with the LMC’s near-side galactic wind, we will report the values of the cylin-drical scenario in the text as it is the simplest volumethat requires the least assumptions. We report the val-ues and ranges for the other two volume scenarios in Ta-ble 2. To estimate the total mass of the neutral and ion-ized gas of this wind, we assume the outflow is symmet-ric on both the near-side and far-side of the LMC’s diskand that it has an ionization fraction of n H + /n H ≈ . . ≤ log ( M total /M (cid:12) ) ≤ . MC Galactic Outflow a)c)b) 𝑅 = 𝑅 𝑅 %&’() = 𝑅 𝑅 %&’() = 𝑅 𝑅 *++() = 𝑅 ∗ 𝑅 *++() = 𝑅 -*. LMC Disk
Figure 8.
Explored near-side volume geometries of the LMCoutflow, including (a) a cylinder with a uniform outflow thatspans the face of the galaxy out to some height, (b) a partialcone with its narrow end embedded and centered on a regionof intense star-formation, (c) an inverted partial cone with itsnarrow side pointing away from the LMC’s disk to match themorphology of the simulated galactic outflow under influenceof ram-pressure and headwinds.
This is compared to the previous Barger et al. (2016)estimate of log ( M ionized /M (cid:12) ) (cid:38) . t outflow ≈
60 Myr). Thisis calculated using information regarding the last pe-riod of star-formation ( ∼
100 Myrs) as well as thetime necessary for the wind to penetrate through thesurrounding medium and travel approximately 2 kpcoff the H i disk. This results in a total IVC mass-flow rate of 0 . ≤ ˙ M outflow ≤ . M (cid:12) yr − . Themass-loading factor is also calculated to study the ra-tio of the mass-flow rate to the star-formation rate,( η ≡ ˙ M outflow / ˙ M (cid:63) ). We adopt a star-formation rate inthe range 0 . (cid:46) ˙ M (cid:63) (cid:46) . M (cid:12) yr − to agree with thestar-formation history of the LMC (see Figure 11 of Har-ris & Zaritsky 2009). This results in a mass-loading fac- tor between 2 . ≤ η ≤ .
93. Because the mass-loadingfactor, ( η ≡ ˙ M outflow / ˙ M (cid:63) ), is much greater than unity,this indicates that the current star-formation state ofthe LMC is unsustainable such that the galaxy couldbecome quenched if this state is prolonged and if theejected gas is able to escape.6.2. IVC Material - Discussion
Barger et al. (2016) characterized the IVC materialwith respect to the LMC using UV absorption-line spec-troscopy toward a LMC disk star and a backgroundQSO. They found the near-side material to have anestimated mass of log ( M low ions /M (cid:12) ) (cid:38) .
16 for low-ionization species on both the near-side and far-sideof the galaxy. This corresponds to an ionized massof log ( M ionized /M (cid:12) ) (cid:38) . LMCSR ≈ −
100 km s − on the near-side of the LMC.Over the same velocity range, we calculated the ionizedhydrogen mass to be log ( M ionized /M (cid:12) ) ≈ .
36 (see Sec-tion 6). While our estimate is larger than the Bargeret al. (2016) near-side mass, it is important to note thediscrepancies between our estimates can be attributedto how each study determined the masses.In the Barger et al. (2016) study, they assumed (1) thewind has an angular extent similar to the LMC’s H i disk( R H i ≈ . f Ω = 0 . f Ω = 0 . α emission of the wind extends around 1 degree on av-erage off the H i disk, which means that the outflow ra-dius is roughly 1 kpc larger (i.e., R H α ≈ R H i + 1 kpc).Their third assumption may result in a significantly un-derestimated outflow mass as the two sightlines used inthe Barger et al. (2016) study probed a relatively qui-escent region of the LMC. We emphasized the effect oftheir assumptions by taking the average emission mea-sure from the same region as the Barger et al. (2016)sightline—in the opposite quadrant as 30 Doradus—andcalculating a corresponding mass for an area similar totheirs. Using (cid:104) EM (cid:105) ≈ . − , the outflow masswould be log ( M ionized /M (cid:12) ) ≈ .
1, which is nearly halfas large as our IVC cylindrical mass estimate and nearlyin agreement with the Barger et al. (2016) estimate forthe low-ionization species.Ultimately, the fate of this ejected material remainsuncertain. The escape velocity of the LMC is roughly110 km s − (Besla 2015), which corresponds to roughly100 km s − along the line of sight for an inclinationof around 25-degrees. Therefore, the majority of theIVC material is not expected to escape. However, the4 Ciampa et al.
Magellanic System is a crowded environment and tidalinteractions between the SMC, and possibly the MW,can assist in the removal of otherwise bound gas (seeD’Onghia & Fox 2016 for a review). Some of the ma-terial from previous outflow events may have been dis-placed into the trailing Magellanic Stream. In a kine-matic investigation of the H i -21cm emission of the Mag-ellanic Stream, Nidever et al. (2008) found that one ofits two filaments traces back to the 30 Doradus regionof the LMC. Richter et al. (2013) measured the chemi-cal composition of this “30-Doradus” filament and foundthat it has a metallicity that is consistent with an LMCorigin. Fox et al. (2013) explored the chemical composi-tion of the other Magellanic Stream filament and foundthat it has a lower metallicity that is more consistentwith an SMC origin, which kinematically traces back tothe Magellanic Bridge (Nidever et al. 2008).Ram-pressure stripping may also play an importantrole removing gas from the Magellanic Clouds. The im-pact that ram pressure has on the CGM strongly de-pends on the density of the medium that it is travelingthrough and on the motion of the gas within its sur-rounding medium. Based off the work of Salem et al.2015, Bustard et al. 2020 and others (Heckman et al.2000, Mastropietro et al. 2005, and Fujita et al. 2009),the effects of ram pressure on gas is multi-faceted andeither can promote or suppress the removal of gas from agalaxy depending on the circumstance. This is becauseram pressure can work in direct opposition of galac-tic winds positioned on the leading side of a galaxy,where the surrounding coronal gas will act as a headwind that pushes against the outflowing back towardthe galaxy. Therefore, the outflow on the near-side ofthe LMC galaxy, which is the side leading the LMC’s or-bit through the MW’s halo, will be suppressed. Mean-while, the far-side could experience an enhanced out-flow as ram pressure will push the gas away from thegalaxy and it could therefore be more massive. Withram-pressure stripping, the simulated wind in the Bus-tard et al. (2020) study was able to reach a height ofmore than 1 . i disk and had a totalejected mass in the range 6 . ≤ log ( M ejected /M (cid:12) ) ≤ . HVC MATERIALWe also detected a high-velocity component to theLMC’s galactic wind in H α emission over the − ≤ v LMCSR ≤ −
100 km s − velocity range (see top-rightpanel of Figure 5). Much of this material is travelingaway from the LMC at speeds that exceed its escapevelocity and could therefore be permanently lost fromthe galaxy. Furthermore, tidal forces could assist in car-rying this material away. At these high velocities, theH α emission is especially concentrated in the directionof the 30 Doradus starburst region (refer to the rightpanel of Figure 6). We calculate the ionized mass ofthis HVC using the procedures described in Section 6.1.Using the timescale of the IVC material (60 Myrs) aswell as the observed velocities, we estimate the HVCreaches a height up to 9 kpc off the H i disk. It ispossible the material reaches even further as Bustardet al. (2020) shows ejected material to reach heightsin excess of 13 kpc. With these estimates and equa-tion (4), we calculate an ionized hydrogen mass oflog ( M ionized /M (cid:12) ) = 6 . ± .
13 for the HVC. This masscontains 124 WHAM beams that are within roughly20 deg . Following the procedure in Section 6.1, we as-sume the wind to be symmetrical about the LMC diskand to contain both neutral and ionized material. Thiscorresponds to a total mass of log ( M total /M (cid:12) ) ≈ . M outflow ≈ . M (cid:12) yr − and the mass-loadingfactor is η ≈ .
36. The full ranges of these results areprovided in Table 2.When comparing our mass to prior work there is a gen-eral agreement. Lehner et al. (2009) estimated its neu-tral hydrogen gas mass to be 5 . ≤ log ( M H i /M (cid:12) ) ≤ . M total /M (cid:12) ) > . ≤ d (cid:12) ≤
50 kpc away. Us-ing the ionization fraction of 0 . (cid:46) n H II /n H (cid:46) . . ≤ log ( M H /M (cid:12) ) ≤ .
1, which is in agreement with theirtotal mass lower limit. Moreover, simulations from Bus-tard et al. (2020) estimated 6 . ≤ log ( M ejected /M (cid:12) ) ≤ .
78 worth of material reaching over 13 kpc away fromthe LMC disk. If we consider that the bulk of the HVCmaterial distance of the wind to be at d (cid:12) ≈
13 kpc,
MC Galactic Outflow
Origins of the HVC
We observe high-velocity material toward the LMCat velocities greater than 100 km s − off the LMC H i disk (+90 ≤ v LSR ≤ +175 km s − ), detailed in Sec-tion 7. This emission is persistent at intermediate- tolow-velocities relative to the LMC and spatially alignswell with the LMC’s H i disk (see Figure 6). These ob-served properties are consistent with an LMC origin inwhich the gas is expelled from the galaxy by its stellaractivity. This is a conclusion that has also been reachedby Staveley-Smith et al. (2003) using H i emission-lineand by Lehner et al. (2009) using UV absorption lines(also see Lehner & Howk 2007 and Barger et al. 2016).Staveley-Smith et al. (2003) found that the H i columndensities of the HVC peaked at locations that align withH i voids within the LMC disk (such as supergiant shells,e.g., LMC 3); they further identified spatial and kine-matic H i bridges that linked back to the LMC’s disk.Lehner et al. (2009) used 139 stars embedded in theLMC as background targets in an UV absorption-linestudy to explore the properties of this HVC; they foundthat the HVC has (i) a LSR velocity gradient in rightascension that follows the LMC’s velocity gradient, (ii)dust based on depletion patterns—signifying a galacticorigin (see also Smoker et al. 2015), (iii) an oxygen abun-dance similar to the LMC of [O i / H i ] = − . +0 . − . , and(iv) a high covering fraction in the direction of the LMC.However, since the works mentioned above, the originof this HVC has been strongly debated. This is becauseWerner & Rauch (2015) found C ii , Si ii , and Si iii absorption consistent with this HVC in the spectra ofa background star at a distance of d (cid:12) = 9 . +4 . − . kpc.Richter et al. (2015) confirmed the presence of the HVCin the direction of this star, which places the HVC at adistance of d (cid:12) < . ∼
150 km s − .This is because, during the cloud’s journey it wouldsustain numerous collisions with not only the CGM ofthe LMC, but also the halo of the MW. Consequently,Richter et al. (2015) argue that this would strip a largeamount of the cloud’s material and drag forces wouldslow the cloud. Needless to say, the time required tomake this journey (250 −
400 Myr; Barger et al. 2016)would consequently lead to a transverse displacementand the cloud would no longer be toward the LMC. In light of the results presented in this paper, and byprevious studies, we offer a mutual theory on the dis-tribution and association of material observed betweenthe MW and LMC. We postulate that there are twoHVCs with different origins near the LMC on the sky:(1) an HVC is associated with the galactic winds ofthe LMC and (2) an HVC that is associated with theMW. Strong evidence for this latter HVC was presentedby Werner & Rauch (2015) and Richter et al. (2015),who confirmed that there is high-velocity material at( l, b ) = (279 . ◦ , − . ◦ − d (cid:12) < . α emission thatspans across the LMC (see Figure 5). Meanwhile, simi-larities in kinematics, dust depletion, and oxygen abun-dances strongly indicate that most of the high-velocitymaterial in the direction of the LMC has an LMC origin.Unless the previous evidence gathered from Staveley-Smith et al. (2003), Lehner et al. (2009), and our studyis entirely coincidental, there is likely to be more thanone complex toward the LMC. Therefore, we argue thatthe LMC HVC spans a much larger area of the sky inthe direction of the LMC and that there is also a smallerHVC positioned just offset from the LMC on the sky,which is associated with the MW.7.2. Feedback of Low-Mass Galaxies
Galactic winds and the ability of a galaxy to retainejected gas are directly related to the galaxy’s capac-ity to form future stars. These winds tend to intensifyas stellar activity increases. Moreover, because lowermass galaxies have smaller gravitational potential wells,their gas is more easily ejected out of them. For massivegalaxies, the halo and extraplanar gas that surroundstheir disks suppresses outflowing material. This resultsin a general trend in which the mass-loading factor is di-minished for massive galaxies and enhanced in low-massgalaxies. In the case of the LMC, with a stellar mass of M (cid:63) = 3 × M (cid:12) (van der Marel et al. 2009) and totalmass M total = 1 . × M (cid:12) (van der Marel & Kalli-vayalil 2014), as well as an active star-formation history(Harris & Zaritsky 2009), it is expected that this galaxywill have a relatively elevated mass-loading factor.We estimate that the LMC’s mass-loading factorranges from 3 . ≤ η ≤ .
56 when including both theIVC and HVC gas in its winds and when assuming cylin-drical geometry. Our values are relatively large whencompared to other studies for galaxies with similar stel-lar mass (see Figure 9). However, the Barger et al.(2016), Chisholm et al. (2017), and Leethochawalit et al.(2019) studies all used absorption-line spectroscopy,which spatially probes less of the wind; this results in6
Ciampa et al.
Table 2.
Mass Estimates for IVC & HVC Material of the LMC outflowOutflow Geometry M ion M total ˙ M outflow η ( M (cid:12) ) ( M (cid:12) ) ( M (cid:12) yr − ) IVC
Cylinder 7 .
28 – 7 .
43 7 .
70 – 7 .
85 0 .
83 – 1 .
18 2 .
44 – 3 . .
02 – 7 .
17 7 .
44 – 7 .
59 0 .
46 – 0 .
65 1 .
35 - 2 . .
01 – 7 .
16 7 .
43 – 7 .
58 0 .
45 – 0 .
63 1 .
32 – 2 . HVC
Cylinder 6 .
93 – 7 .
05 7 .
35 – 7 .
47 0 .
37 – 0 .
49 1 .
09 – 1 . Note —All values for M total assume a symmetrical wind on the near-sideand far-side of the LMC with an ionization fraction of n H ii /n H = 0 .
75 thatincludes neutral and ionized gas assuming a reduced mass of µ ≈ . m H to account for helium and metals. We used the values listed in Table 1 tocalculate these outflow masses and rates. Figure 9.
Mass-loading factors from various studies. Themass-loading factor we calculated for the total LMC wind(IVC + HVC) is marked with a cyan bar while solely theIVC component is the darker hashed blue bar. Our resultsfor the LMC wind are found in Table 2. The Barger et al.(2016) LMC loading factor (IVC only) is indicated with agray diamond after we adjust their outflow mass to matchthe assumptions used in our study. Green squares mark massloading-factors from McQuinn et al. (2019), which only in-clude the ionized gas mass in outflows. The brown and purpleenvelopes are the limits of uncertainty for mass-loading re-lations from the Chisholm et al. (2017) and Leethochawalitet al. (2019) studies (refer to their equations 16 and 8, re-spectively). a more uncertain mass estimate as the physical extentof winds are poorly constrained. In the case of the Mc-Quinn et al. (2019) H α imaging study, although theywere able to measure the extent of the wind, their ob-servations were more than a magnitude less sensitive than ours and were unable to detect the diffuse gas inthe wind.The impact of geometry assumptions is most tellingwhen comparing our results with those of the Bargeret al. (2016) absorption-line study of the LMC. Althoughour study shares many of the same assumptions as theirstudy, we were able to map the extent and morphol-ogy of the wind, whereas they were forced to makesimplistic assumptions for its solid angle. The mass-flow rates from the Barger et al. (2016) and the star-formation rates from Harris & Zaritsky (2009) corre-spond to a mass-loading factor for LMC that is in therange 0 . ≤ η ≤ .
8. When we increase their solid angleto match what we observe in H α emission, their mass-loading factor becomes 1 . ≤ η ≤ .
5. This revisedrange is in better agreement with our mass-loading fac-tor for conical geometries (Outward and inward cone;see Table 2).In studies that explore a wide range of galaxy masses,the trend that the mass-loading factor decreases withgalaxy mass is clear (see Figure 9). Chisholm et al.(2017) find mass-loading factors ranging from 0 . − . . ≤ log ( M (cid:63) /M (cid:12) ) ≤ .
7) using UV absorption-line spec-troscopy. At the LMC mass, this study predicts amass-loading factor around 1 .
1. Across similar masses,McQuinn et al. (2019) observed their galaxies via H α imaging and found mass-loading factors ranging from0 . ≤ η ionized ≤ . . ≤ log ( M (cid:63) /M (cid:12) ) ≤ . . − . MC Galactic Outflow SUMMARYWe completed the first kinematically resolved surveyof the LMC’s near-side galactic wind in H α λ α mapper. These mapped observationsspan 20 x 20 degrees across the sky and are comprised of1,712 sightlines. By combining these observations withexisting H i observations, we are able to determine theextent and morphology of the neutral and warm ionized( T e ≈ K) phases of this wind. Here we summarizethe main conclusions of this study:1.
Morphology and Extent:
We find that diffusegas in the galactic wind spans across the entireface of the LMC. We additionally find numerousfaint I H α ≈
100 mR clouds offset from the mainwind structure, but we are unable to confidentlydetermine whether or not they are physically asso-ciated with the LMC’s galactic wind as tidally dis-placed Magellanic Cloud material and MW HVCsalso pollute this region of the sky.2.
Kinematic Distribution:
We find the bulk ofthe LMC’s galactic wind is moving with veloci-ties of v
LMCSR (cid:46) −
110 km s − relative to the H i disk, which is less than the escape velocity. How-ever, roughly log ( M ion , hvc /M (cid:12) ) ≈ .
0, or roughly44%, of this wind is moving away from the LMCat speeds greater than the escape velocity. Specifi-cally, we find the gas that is spatially aligned with30 Doradus is moving at the greatest speeds rela-tive to the LMC at v
LMCSR (cid:46) −
175 km s − .3. Two HVC Complexes toward the LMC:
Wefind H α emission at high velocities relative to theLMC that is strongly spatially correlated with the 30 Doradus. This emission similarly persistsat lower velocities. Our results—in addition tothe results from the Staveley-Smith et al. (2003),Lehner & Howk (2007) Lehner et al. (2009), andBarger et al. (2016) studies—lead us to concludethat this starburst region is responsible for gen-erating this HVC (see Figures 3 and 6). TheHVC discussed in Richter et al. (2015), which liesa few degrees off the log (cid:0) N H i / cm − (cid:1) = 19 con-tour of the LMC’s H i disk and at a distance of d (cid:12) (cid:46) . Outflow Mass, Flow Rate, & Loading Fac-tor:
We measure an ionized gas mass in the range7 . ≤ log ( M ionized /M (cid:12) ) ≤ .
43 for the outflow-ing material on the near-side of the LMC thatis moving at intermediate-velocities, i.e., speedsthat are within ∼
100 km s − of the LMC’s H i disk. The high-velocity component of this windhas an ionized mass of 6 . ≤ log ( M ionized /M (cid:12) ) ≤ .
05. Combined, we estimate that the total ion-ized gas mass in this near-side wind is in therange 7 . ≤ log ( M ionized /M (cid:12) ) ≤ .
58. Thiscorresponds to a total neutral and ionized massof the entire wind that ranges between 7 . ≤ log ( M total /M (cid:12) ) ≤ .
0, assuming that it is sym-metrical on the near-side and far-side of the LMCand that it is 75% ionized (see Lehner & Howk2007 and Barger et al. 2016). We further cal-culate a total mass-flow rate and mass-loadingfactor of 1 . ≤ ˙ M outflow ≤ . M (cid:12) yr − and3 . ≤ η ≤ .
56. Table 2 summarizes these re-sults.5.
Undetected Diffuse Material:
We comparedour results with existing mass-loading factortrends that vary with stellar mass. We find thatour average mass-loading factors are on averageroughly 2.5 times larger than both optical H α imaging and UV absorption-line studies at thestellar mass of the LMC. This indicates that ei-ther the observational sensitivity (optical imaging:McQuinn et al. 2019) may be insufficient to detectdiffuse gas in these outflows or that the geometricassumptions are too conservative (UV absorption-line spectroscopy: Barger et al. 2016, Chisholmet al. 2017, and Leethochawalit et al. 2019).ACKNOWLEDGMENTSWe thank Lister Staveley-Smith and Sungeun Kimfor providing us with the ATCA and Parkes tele-scopes LMC H i survey datacube. This paper includes8 Ciampa et al. archived LAB and GASS H i data obtained through theAIfA H i Antwi-Danso, J., Barger, K. A., & Haffner, L. M. 2020,ApJ, 891, 176, doi: 10.3847/1538-4357/ab6ef9Barger, K. A., Haffner, L. M., & Bland-Hawthorn, J. 2013,The Astrophysical Journal, 771, 132,doi: 10.1088/0004-637x/771/2/132Barger, K. A., Haffner, L. M., & Bland-Hawthorn, J. 2013,ApJ, 771, 132, doi: 10.1088/0004-637X/771/2/132Barger, K. A., Haffner, L. M., Wakker, B. P., et al. 2012,ApJ, 761, 145, doi: 10.1088/0004-637X/761/2/145Barger, K. A., Lehner, N., & Howk, J. C. 2016, ApJ, 817,91, doi: 10.3847/0004-637X/817/2/91Barger, K. A., Madsen, G. J., Fox, A. J., et al. 2017, ApJ,851, 110, doi: 10.3847/1538-4357/aa992aBesla, G. 2015, The Orbits of the Magellanic Clouds, 311,doi: 10.1007/978-3-319-10614-4 26Besla, G., Kallivayalil, N., Hernquist, L., et al. 2010, ApJL,721, L97, doi: 10.1088/2041-8205/721/2/L97—. 2012, MNRAS, 421, 2109,doi: 10.1111/j.1365-2966.2012.20466.xBustard, C., Zweibel, E. G., D’Onghia, E., Gallagher, J. S.,I., & Farber, R. 2020, ApJ, 893, 29,doi: 10.3847/1538-4357/ab7fa3Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ,345, 245, doi: 10.1086/167900Chisholm, J., Tremonti, C. A., Leitherer, C., & Chen, Y.2017, MNRAS, 469, 4831, doi: 10.1093/mnras/stx1164Choi, Y., Nidever, D. L., Olsen, K., et al. 2018, ApJ, 866,90, doi: 10.3847/1538-4357/aae083de Boer, K. S., & Savage, B. D. 1980, ApJ, 238, 86,doi: 10.1086/157960de Grijs, R., Wicker, J. E., & Bono, G. 2014, AJ, 147, 122,doi: 10.1088/0004-6256/147/5/122Diplas, A., & Savage, B. D. 1994, ApJ, 427, 274,doi: 10.1086/174139D’Onghia, E., & Fox, A. J. 2016, ARA&A, 54, 363,doi: 10.1146/annurev-astro-081915-023251Erb, D. K. 2015, Nature, 523, 169,doi: 10.1038/nature14454Ford, A. B., Dav´e, R., Oppenheimer, B. D., et al. 2014,MNRAS, 444, 1260, doi: 10.1093/mnras/stu1418Fox, A. J., Richter, P., Wakker, B. P., et al. 2013, ApJ, 772,110, doi: 10.1088/0004-637X/772/2/110 Fujita, A., Martin, C. L., Mac Low, M.-M., New, K. C. B.,& Weaver, R. 2009, ApJ, 698, 693,doi: 10.1088/0004-637X/698/1/693Gordon, K. D., Clayton, G. C., Misselt, K. A., Landolt,A. U., & Wolff, M. J. 2003, ApJ, 594, 279,doi: 10.1086/376774Haffner, L. M., Reynolds, R. J., & Tufte, S. L. 2001, ApJL,556, L33, doi: 10.1086/322867Haffner, L. M., Reynolds, R. J., Tufte, S. L., et al. 2003,ApJS, 149, 405, doi: 10.1086/378850Harris, J., & Zaritsky, D. 2009, The Astronomical Journal,138, 1243, doi: 10.1088/0004-6256/138/5/1243Hartmann, D., & Burton, W. B. 1997, Atlas of GalacticNeutral Hydrogen (”Cambridge University Press”)Hausen, N. R., Reynolds, R. J., Haffner, L. M., & Tufte,S. L. 2002, ApJ, 565, 1060, doi: 10.1086/324692Heckman, T. M. 2003, in Revista Mexicana de Astronomiay Astrofisica Conference Series, Vol. 17, RevistaMexicana de Astronomia y Astrofisica Conference Series,ed. V. Avila-Reese, C. Firmani, C. S. Frenk, & C. Allen,47–55Heckman, T. M., Lehnert, M. D., Strickland , D. K., &Armus, L. 2000, ApJS, 129, 493, doi: 10.1086/313421Hill, A. S., Haffner, L. M., & Reynolds, R. J. 2009, ApJ,703, 1832, doi: 10.1088/0004-637X/703/2/1832Howk, J. C., Sembach, K. R., Savage, B. D., et al. 2002,ApJ, 569, 214, doi: 10.1086/339322Hummels, C. B., Smith, B. D., & Silvia, D. W. 2017, ApJ,847, 59, doi: 10.3847/1538-4357/aa7e2dKalberla, P. M. W., Burton, W. B., Hartmann, D., et al.2005, A&A, 440, 775, doi: 10.1051/0004-6361:20041864Kim, S., Dopita, M. A., Staveley-Smith, L., & Bessell,M. S. 1999, AJ, 118, 2797, doi: 10.1086/301116Kim, S., Staveley-Smith, L., Dopita, M. A., et al. 1998,ApJ, 503, 674, doi: 10.1086/306030—. 2003, ApJS, 148, 473, doi: 10.1086/376980Laval, A., Rosado, M., Boulesteix, J., et al. 1992, A&A,253, 213Leethochawalit, N., Kirby, E. N., Ellis, R. S., Moran, S. M.,& Treu, T. 2019, ApJ, 885, 100,doi: 10.3847/1538-4357/ab4809Lehner, N., & Howk, J. C. 2007, MNRAS, 377, 687,doi: 10.1111/j.1365-2966.2007.11631.x
MC Galactic Outflow19