Story of Red Supergiant Winds written in water

Transcription

1 Story of Red Supergiant Winds written in water A.M.S.Richards, I.Bains, A.Bartkiewicz, R.J.Cohen, P.J.Diamond, S.Etoka, M.D.Gray, E.E.Lekht, E.Mendoza-Torres, M.R.W.Masheder, K.Murakawa, M.Szymczak, H.J.van Langevelde, W.Vlemmings, J.A.Yates High-mass AGB analogues: M M Water masers from wet, dusty clouds Mainlines overlap water masers Irregular polarization near star What determines maser properties? Over-dense, over-magnetised, accelerated Birthplace, birth rate, survival Hydroxyl masers from diffuse gas Return >>50% enriched mass to ISM Testing predictions Massive stellar evolution

2 The Cast ordered by increasing post-ms evolutionary stage IRAS Silicate OH 1612 MHz colour feature, shape, peak log(jy) region¹ dust² Red Super Giant Star M (M ) Distance (kpc) T (K) S Per IIIa Emission, compact VX Sgr VII Emission, ~Spherical, 1 compact VY CMa VII Declining, complex NML Cyg VIb Absorption, Elongated, multi-shell detatched, 2 Biconical, 0 ~, 2 ¹van der Veen & Habing (1988) ²Monnier et al 1998, 1999, 2000, 2004

3 S Per VX Sgr Water maser shells Richards AU adjusted to common distance MERLIN proper motion/acceleration Murakawa+03 VY CMa Richards+98 NML Cyg Richards+96

4 Measuring 'true' maser cloud size Cloud properties mas as 0 20 m 0 20 mas 20 mas km/s total line width Component measurements 18 Cloud D=18 AU at 1 kpc Channel maps every 0.2 km s mas Largest angular separation across all channels is actual cloud size (18±5 mas) (to limits of sensitivity)

5 Measuring maser cloud size S Per, VX Sgr Colour velocity OH mainlines global VLBI (over) resolves H2O - MERLIN detects ~all flux For many clouds: 8 mas 18 AU Diameter L > l (gain length) Vcloud > Vth Clouds must be defined by density &/or composition (not spooks!) spooks S Per H2O

6 Some RSG clouds last 5 yr+ but half or more vanish Yet total shell crossing time is yr S Per Identifiable patterns in S Per Sub-/mildly super sonic timescale Proper motions in VX Sgr Consistent with radial velocities VX Sgr Average cloud positions shown Apparent shell extent increased >50% since 1983!

7 Clumps persist, masers flicker S Per cloud at VLSR km s-1 imaged 1994, 1999 Vanishes in Puschino data (Lekht+05) Accident of excitation or beaming? Images show gas clump survives Pattern speed sub/ mildly super-sonic

8 Angular separation from centre of emission Water maser cloud density S Per H2O VX Sgr VLSR Quenching radius Collision rate > masing rate Quenching density nq ~ m-3 dm/dtq = 4 ri2vi q NML Cyg H2O VY CMa OH 1667 OH 1665

9 Water (and OH) maser cloud properties RSG dm/dt Number H O cloud H2O cloud H O OH cloud dm/dt H2O 2 2 CO/IR shell ri H2O nq of H2O size <L> mass yr-1 shell ro size <L> (AU) (AU) (M yr-1) (AU) (M yr-1) clouds (M yr-1) (AU) S Per ( 930 ) (44) >13 (66) VX Sgr ( 190 ) >13 (42) VY CMa ( 315 ) NML Cyg ( 515 ) H2O maser cloud filling factor <few % Solves mass loss rate discrepancy Few clouds formed per stellar period Clouds much denser than surroundings: nh2ocloud (15 70) n OHgas Some clouds have multiple peaks up to larger size estimate OH mainlines <L> lower limit as sensitivity-limited

10 OH mainline masers interleave H2O S Per MERLIN H2O (blue) EVN/global mainline OH (contours) OH mainlines interleave H2O 10 mas Clump radii roh~9 au rh2o ~9 au OH masers so close to star! TOH~500 K max? TH2O ~1000 K

11 Angular radius-velocity all masers 1612 MHz 1665 MHz 1667 MHz S Per H2O S Per <1994 VX Sgr 10 mas S Per 1999 VY CMa

12 VY CMa OH mainline masers 1665 MHz S Per 10 mas 1667 MHz

13 Water and OH mainline shells RSG H2O shell H2O shell OH mainline shell OH 1665-MHz shell OH 1667-MHz shell ri (AU) ro (AU) ro (AU) ri (AU) ro (AU) S Per / /2453 VX Sgr / /19504 VY CMa NML Cyg / MHz Etoka & Diamond MHz Diamond 3. Larger outer radius in could be sensitivity or variability 4. Double shell structure

14 OH Asymmetry 1612 MHz OH 1999 MERLIN OH 1612 MHz Axisymmetry for several centuries Box shows H2O/ mainline OH region O/ 1665/7 MHz Similar shape Different velocities along limbs

15 OH Asymmetry 1612 MHz OH 1999 MERLIN OH 1612 MHz Axisymmetry for several centuries Box shows H2O/ mainline OH region O/ 1665/7 MHz Similar shape Different velocities along limbs

16 S Per OH Mainline polarization Circular polarization B Strength in T - mean 0.3 T Consistent with NNE - SSW dipole, S approaching Linear polarization vector at PolA If > 55, PolA B else PolA B PolA typically 16o per clump Linear pol seen from near side only Faraday rotation? Fractional ionisaton ~10-6 Pol. from far side scrambled? But conflicting theories Inferred B strength PolA varies with path length ISM Faraday rotation B B B

17 H2O Zeeman splitting S Per VLBA (Vlemming, Diamond & van Langevelde 02) B 7-20 T ( mg) (model-dependent) Brightest H2O R~90 au (2.3 kpc) - take B ~15 T

18 S Per magnetic field BH2O >> BOH OH mainlines B ~0.3 T at up to 140 au H2O B ~ 15 T at ~90 au (brightest masers) Stretched dipole? B R-2, expect au Selection effect? 6 T would split OH by~18-30 km/s; aligments ambiguous Only strongest/weakest splitting detectable in H2O/OH? H2O clumps ~50x denser than OH gas (Richards+ 99) Frozen-in B n~ (Mouschouvias 87) - n R (no/strong acceleration) so frozen-in B R Ri ~30 T where B(OH) would be ~4 T Extrapolate to 140 au (stretched dipole) B(OH) ~0.6 T This implies clump B >> inter-clump B after dust forms Precursors of water maser clumps already denser?

19 VX Sgr OH/H2O axisymmetry OH 1612 polarization vectors Dipole magnetic field Axis p.a. ~20o S approaching (Szymczak+) H2O ~spherical Lower density bicone ( ) Aligned about mag. axis (Murakawa+)

20 VX Sgr OH/H2O axisymmetry OH 1612 dipole (Szymczak et al.) H2O kinematic axis (Murakawa et al.) H2O magnetic axis (Vlemmings et al.)

21 VX Sgr OH/H2O axisymmetry Vlemmings et al 05

22 VX Sgr from different angles OH 1665/1667 MERLIN Using VO tool TopCat

23 VX Sgr from different angles Inner OH/H 2O Looking down axis OH 1665/1667 MERLIN

24 VX Sgr from different angles All OH 1665/1667/1612 Looking down axis

25 Maser spatial properties Multiplepeaked S Per water maser Brighter clouds resolved Fainter component s have larger FWHM

26 Spatial beaming Spherical Shocked cloud slab Deconvolved observed spot in each individual channel has beamed FWHM s: Spherical clouds have s l 2 [ln(tb /(T0-Tx))]- Gain length l D D Intensity s s Angular distance larger for more saturation Beaming angle = (sd/l)2 Measure s logtb or logs If s inverse function of flux density (typically 10-3 'true' cloud area) indicates ~sphericity But masers propagating perpendicular to a shock are less narrowly beamed (Elitzur+ 92, etc.)

27 Beamed component size as diagnostic of emission region shape S Per: Component FWHM S-1.2 Characteristic beaming effect for a mixture of un/saturated emission from ~spherical cloud NML Cyg: FWHM v. S ~ flat Could indicate brightest emission is from shocked slab

28 Evidence for density-bounded clumps Water maser clouds dusty, rapidly accelerated Cloud diameters can exceed several Vth, maser gain lengths Water maser clouds over-dense: n(h2o cloud) ~50x n(oh gas) Also required for OH maser pumping schemes Direct measurement of spot size and beaming angle Most H2O maser clouds ~spherical, a few shocks Masing lifetimes < cloud lifetimes (shell crossing time) BOH ~ 0.2BH2O at same distance from star Frozen-in B n~ (Mouschouvias 87) 'True' sizes measured across multiple spectral channels R(H2O clouds) & shell size R (RRSG 10x RAGB) Contrast e.g. P, Vexp only vary by factors of a few Cloud size determined by M not micro-physics

29 Not an onion

30 Not an onion H2O clumps dustier, better accelerated Interleaving gas supports OH mainlines near star Mixed, mainly radial beaming OH 1612 further out Tends to tangential beaming Needs ~steady velocity Radial beaming Some H2O & OH mainlines reach high velocities Can overshoot OH 1612 At different latitudes?

31 Not all spherical Zijlstra et al Interaction of lighter wind with equatorial density enhancement Chimney effect? (Icke, Mellema) No disc, no rotation required! Contrast real binary (P)PNe with jets (Vlemmings, Imai)

32 Evolution of RSG shells Some axisymmetry at all ages S Per earliest? OH blobs at various positions in wide-angle bicone H2O ~spherical, dense, rapidly accelerated OH mainlines interleave H2O VX Sgr Far-out OH 1612-MHz shell ~spherical: ex-h2o shell? Bicone almost hollow - OH mainlines on interaction surfaces? VY CMa thickest shell NML Cyg H2O 'jets' in interaction of axial outflow with older shell? OH evidence for increasing bipolar wind speed (Etoka & Diamond) Evidence for Faraday screen - ne few 107 m-3 Multiple shells but still mass-loss (SiO masers)

33 Mass loss from the stellar surface HST Ori lumpy, aspherical RSG clouds 5-10% R at birth Star spots? Chemical inhomogeneity? Convection cells? MERLIN+VLA 5GHz e-merlin, EVLBI Image the star and its masers ALMA Track the dust as it forms Morris et al. VLA 43 GHz

34 Magnetic pressure & energy Thermal/magnetic pressure = n kbt /B2 Typical region densities OH n~5x1013 m-3, H2O n~1015 m-3 (H2O) ~0.1 (T~1000 K) - strong mag. pressure (OH) ~ <4 (T~500 K) - significant but not dominant KE/magnetic energy density =107(B2 R2)/(dM/dt V) BH2O= 15x10-6 T, BOH = 0.3x10-6 T RH2O= 1.35x1013 m, ROH = 2.1x1013 m dm/dth2o(apparent) 2.5x10-3, dm/dtoh 5x10-5 (M /yr 6.5x1022) VH2O= 12x103m/s, VOH = 15x103m/s H2O = 0.2, OH = 0.01 Radiation pressure on dust wins but wind can be deflected

35 Angular radius profiles H2O MERLIN OH VLBI S Per S Per 10 mas OH MERLIN 1612 MHz 1665 MHz 1667 MHz VX Sgr Inner OH mainline shells

36 OH component size s > s S Per 1665 MHz s S -0.9(0.1) 1667 MHz s S -1.2(0.1) VX Sgr 1665 MHz s S -1.9(0.4) 1667 MHz s S -2.6(0.2)

37 Velocity profiles S Per H2O MERLIN OH VLBI S Per 10 mas OH MERLIN 1612 MHz 1665 MHz 1667 MHz VX Sgr

38 Shock in Mira U Ori Bains et al IK Tau, RT Vir, U Her H2O component FWHM narrows as brightness temp increases s TB( ) U Ori behaves differently Many brighter maser spots have larger FWHM s TB+3 Not consistent with beaming from spherical cloud Shocked slab? Strong complex circular polarization (B~0.35 T) (Vlemmings et al 2005) Stellar shock compression?