HI observations of the Magellanic Bridge Erik Muller

HI observations of the Magellanic Bridge Erik Muller UOW/ATNF Supervisors: Bill Zealey (UOW) Lister Staveley-Smith (ATNF)

Overview: HI observations of the Magellanic Bridge • • The Magellanic system HI Data collection and reduction Shell formation mechanisms Magellanic Bridge HI expanding Shell census – Selection criteria – Statistical results • OB Stellar associations, HI shells and HI gas • Shell formation mechanisms applied to the Bridge shell population.

The Magellanic System: • Detected in HI (spin-flip transition of Neutral Hydrogen) by Kerr, Hindman & Robinson, (1954), Parkes Telescope (ATNF). • Magellanic clouds are ~60 kpc (SMC) to ~50 kpc (LMC) • Their nearness makes them an excellent laboratory in which to observe physical processes with high spatial resolution • Magellanic system comprises five elements: – – – Large Magellanic Cloud (LMC) (Kim 1998) Small Magellanic Cloud (SMC) (Staveley-Smith 1998, Stanimirovic 1999) The Magellanic Stream (Putman, Gibson, Stanimirovic etc. 1998) The Leading Arm (Putman 2000) The Magellanic Bridge (Mathewson & Cleary 1984) • Bridge spans the ~14 kpc from western edge of LMC to eastern edge of SMC – Formed through tidal interaction of SMC with LMC (Simulations predict 150 -200 Myr old - eg. Gardiner & Noguchi 1996) – Populated by young O-B (>7 Myr), as well as older, stars. (eg. Irwin et al, 1995)

The Magellanic System in HI: Multibeam, Parkes. To the Magellanic Stream LMC The Magellanic Bridge To the Leading Arm SMC Peak Pixel map, Linear trans. func. Tmax=0. 3 MJy/beam

HI Data collection & Reduction: • 144 pointings with ATCA (375 m configuration) – ~16 minutes/pointing • Scanning with Parkes multibeam (inner seven beams) – Scanning rate: 1 o/min • ATCA Data reduced with MIRIAD – conventional procedures for data flagging and calibration – Parkes and ATCA data merged post-convolution using IMMERGE (Stanimirovic, Ph. D, 1999) • Parkes data reduced on line with ‘LIVEDATA’ – Bandpass calibrations, velocity corrections • Resulting cube: – – ~7 ox 7 o region, Vel range~100 -350 km/s (Heliocentric) RMS ~ 15. 2 m. Jy/Beam (eq 1. 7 x 1018 cm 2 for each channel) 98” spatial resolution ~2 x 108 M (SMC ~4 x 108 M )

Right Ascension-Declination RMS=15. 2 m. Jy/beam (1. 7 x 1018 atm cm-2) Velocity-Declination ? Peak pixel maps of ATCA/Parkes HI datacube 38 km/s [VGSR] Total observed HI Mass=200 x 106 M Mass of centre region=72 x 106 M 8 km/s [VGSR] Right Ascension-Velocity (2 x 4. 7)kpc cylinder ρ=0. 2 atm cm-3 (2 x 4. 7 x 5)kpc slab ρ=0. 06 atm cm-3

Formation mechanisms of HI expanding Shells: • Stellar wind and SNe driven shells (Weaver et al, 1977): – Hot, energetic stars ionise local gas, and blow open an expanding sphere of hot gas. – Detailed study by Rhode et al (1999) using HI data of Holmburg II galaxy find that the distribution and brightness of HOII clusters do not support the idea of expansion from SNe. • HVC collisions (Tenorio-Tagle 1987, 1988, Ehlerova & Palous, 1996) – Capable of producing low energy, spherical expanding structures for impacts by low Ek clouds. Rc ~10 pc – Difficult/impossible to differentiate from stellar wind formation mechanism. • Gamma Ray Bursts (Efremov, Elmegreen & Hodge, 1998, Loeb & Perna, 1998) – Release relatively large amounts of energy (10% of progenitor mass) ~1053 erg • Shells formed from GRB are more energetic for lower radii and more quickly expanding shells. • GBR frequency in a our galaxy ~0. 1 Myr – 1 (Portegies Zwart, & Spreeuw, 1996),

Shell selection Criteria • Adapted from Puche et. al. (1992) i. ii. iii. • • A (rough) ring shape must be observed in all three projections (RA-Dec, RA-Vel, Dec-Vel), and must be present across the velocity range occupied by the shell Expansion must be present across at least three velocity channels (~5 km/s) The rim of the ring has good contrast with ambient column density of channel (i. e. the shell is rim brightened). Criteria target rim brightened, expanding spherical structures (not cylindrical or blown out volumes) To reduce subjectivity, criteria must be strictly satisfied

HI Peak Pixel map. Size and location of 163 Magellanic Bridge HI expanding shells. Crosses locate OB associations (Bica et al. 1995) Ret

Comparison of Magellanic Bridge shells to SMC population: Dynamic Age Luminosity: (Weaver 1977) Magellanic Bridge SMC Mean Stdev Kinematic Age (Myr) 6. 2 3. 4 5. 7 2. 8 Shell Radius (pc) 58. 6 33. 2 91. 9 65. 5 Expansion Velocity (km/s) 6. 5 3. 8 10. 3 6. 3 Energy (log [ergs]) 48. 1 51. 8 (n=1 cm 3) • Bridge shells, compared to the SMC population are (on average): • Marginally older (!) • 60% smaller + expand 60% more slowly = Much less energetic.

MB and SMC HI shell population Expansion Velocity Dynamic age Radius Slight excess at Discontinuity at Decreasing shell radius Higher RA with increasing RA MB/SMC transition Discontinuity at (effect of selection MB/SMC criteria)(effect of transition selection criteria) Decreasing expansion velocity with increasing RA Generally continuous age distribution

Comparison of power law parameters of expanding HI structures from other surveys αx = 1 -γx Holmberg II SMC Magellanic Bridge (Puche et al. 1992) (Staveley. Smith et al. 1997) 51 509 163 Expansion Velocity αv 2. 9± 0. 6 2. 8± 0. 4 2. 6± 0. 6 Shell Radius αr 2. 2± 0. 3 3. 6± 0. 4 Number of Shells 2. 0± 0. 2 • αv is in agreement with other systems • αr is much steeper for the Magellanic Bridge population – Due to a strict selection criteria that manifests as an overall deficiency of small radii shells, and ultimately as an older shell population.

Distribution of OB associations and HI shells • Visually, OB associations, HI and shell centres appear to correlate reasonably well. Map • A more quantitative study shows that: – ~50% of shells have one or more OB association within 8’ (140 pc) – ~18% of shells have one or more OB associations within 3. 5’ (60 pc) (mean shell radius) – ~40% of shells have at least one or more associations within one radius • Poor spatial correlation statistic – Are these associations really responsible for HI shell expansion? – Alternatives include Gamma ray bursts (Efremov, Elmegreen and Hodge, 1998), HVC collisions (Tenorio-Tagle 1981, 1987), ram pressure drag (Bureau et al, 2001).

HI around OB associations • HI ramps almost linearly to centre of OB positions • Excess of HI <80 pc of association centre, in disagreement with Grondin & Demers, 1993. (No discernable depletion of the local H I) Diamonds: Mean HI averaged in concentric annuli around OB catalogued positions. Triangles: Mean HI averaged in concentric annuli, offset 90 pc (10 pixels) south of OB centres Error bars mark one standard error of the mean, vertical line marks resolution of Parkes observations by Matthewson, Cleary & Murray (1974)

Formation mechanisms of Bridge Shells • Stellar Wind and SNe – The most recent burst of starformation 10 -25 Myr ago (Demers & Battinelli 1998) , C/W mean shell kinematic age ~6 Myr – ‘Constant energy input rate is generally invalid’ (Shull, & Saken 1995) • Input from WR and stellar wind at 3~10 Myr for coeval and noncoeval associations, increased expansion velocity, and mis-estimation of ‘true’ age by up to 40% - lower limit of starburst date by Demers & Battinelli • Bridge Associations & Clusters are very poorly populated, typically N ~ 8 (N increases towards SMC) • Some Assocations & Clusters ‘may be of type later than O-B’ (priv comm. Bica 2002) • Poor spatial correlation of OB associations, clusters and expanding shells in the Magellanic Bridge – Given a ‘normal’ IMF, we would expect a significant energy input from SNe after ~5 Myr. Shells of this age not found around, or even near, most observed Magellanic Bridge OB associations – why not? .

Formation mechanisms of Bridge Shells • HVC impacts – The distribution of holes shows preference for high HI column density (not withstanding selection effects) • There is no reason for HVCs to preferentially impact in a specific region. – Many shells are deeply embedded in the HI, rather than being found near the surface. • GRB • Under this model, mean shell energy is ~1. 3 x 1051 erg, Mean Kinematic age is 1. 2 x 105 yr (c/w 1. 3 x 1048 erg and 6. 1 Myr for stellar wind model), expansion velocities are ~10 -2 of predicted velocity.

Summary: • General appearance: – ATCA and Parkes have uncovered chaotic and intricate structure of HI comprising the Magellanic Bridge. – Loops, filaments and clumps observable to smallest scales of 98” (~29 pc) – Much of the Bridge is bifurcated into two velocity sheets, converging at ~2 hr 30 min – Large loop R~1 kpc off the northeastern edge of SMC. • Shell survey: – 163 shells found within the Magellanic Bridge – Kinematic age is consistent with that of , shells of the SMC although Magellanic Bridge shells are considerably smaller and less energetic. – Power law distribution of expansion velocity is consistent with Ho. II and SMC. – Strict selection criteria is insensitive to incomplete and fragmented shells

Summary: • Shells, stars and HI : – Good correlation of HI with OB assocations and Clusters, and also with HI shells (NB. Selection criteria), Poor correlation of OB associations and clusters with expanding shell centres – HI distribution about OB associations and Clusters shows a mean excess at short radii (<80 pc), and a decreasing slope with increasing radii • Shell formation: – Shell Energies and spatial distribution do not agree with theories of formation by stellar wind by OB associations and Clusters or by SNe – Theoretical frequency of GBRs is too low to be generally applied to Magellanic Bridge shells. – HVCs are capable of producing the observed structures, however, the surface distribution shows preferential distribution (selection effects!), and many shells are found too deeply embedded throughout the HI Bridge. – Alternatives ? ?