1. Accretion of brown dwarfs: Clues from spectroscopic variability Alexander Scholz (University of Toronto) Ray Jayawardhana (UoT) Jochen Eislöffel (Tautenburg) Alexis Brandeker (UoT)

2. Brown dwarfs in T Tauri phase Accretion flow  variability

3. T Tauri stars are variable „T Tauri has a magnitude range of from 9.4 to 13.5 or 14, but no regular period has as yet been detected.“ (Knott, 1891)

4. Example: RW Aur 1945-2005 Alencar et al. 2005 Joy 1945

5. Monitoring brown dwarfs Long-term project: five years 1999-2004 of deep photometric monitoring in young open clusters with ages from 3 to 700 Myr Using data from ESO, TLS, Calar Alto Main goal: evolution of rotation and activity for M<0.3Ms

6. Scholz & Eislöffel, A&A, 2005 T Tauri lightcurves 11 objects with large amplitudes, partly irregular variability  typical `T Tauri lightcurves`

7. Natta et al. 2002 Brown dwarfs have accretion disks Jayawardhana et al. 2003 Muzerolle et al. 2003 Scholz & Eislöffel 2005

8. Origin of variability shockfront (hot gas) + rotation + instabilities = strong, irregular variability (in photometry and emission lines)

9. Spectroscopic monitoring three observing runs with MIKE/Magellan, Jan-March 2005 six targets: accreting brown dwarfs in star forming regions Hα line: σ(EW) = 22-90% σ(10%width) = 4-30%

10. 2M1207 Brown dwarf at 8 Myr with wide, planetary-mass companion No NIR (but MIR) colour excess, clear signature of accretion and wind Final stage of accretion?

11. Profile Variability broad emission plus redshifted absorption feature absorption disappears and re-appears on timescales of ~1 day 4 hours 4 hours Scholz, Jayawardhana, Brandeker, ApJL, 2005

12. Interpretation: Accretion flow cool, infalling, co-rotating material:   accretion column   close to edge-on geometry   asymmetric flow geometry Scholz, Jayawardhana, Brandeker, ApJL, 2005

13. ISO217 profile asymmetry AND profile variability   nonspherical accretion   indirect evidence for magnetically funneled flow Scholz & Jayawardhana, ApJ, 2006

14. Linewidth variations for 2M1207 variations in the linewidth by ~30% on a timescale of 6 weeks Scholz, Jayawardhana, Brandeker, ApJL, 2005

15. 2M1101-7718 strong variations of the accretion-related emission lines 10% width: 122 232 194 km/s 10% width: 122 232 194 km/s EW: 12 92 126 Å EW: 12 92 126 Å other lines: +HeI,CaII,Hβ +HeI,CaII,Hβ,Hγ other lines: +HeI,CaII,Hβ +HeI,CaII,Hβ,Hγ 8 hours 24 hours Scholz & Jayawardhana, ApJ, 2006

16. Accretion rate variations Accretion rate changes by 0.5-1 order of magnitude in 2M1207 and 2M1101   clumpy, unsteady accretion flow Natta et al. (2004)

17. Accretion rate vs. mass accretion rate vs. object mass: ~M 2 large scatter: influenced by variability?   variability studies essential to study accretion fundamentals Mohanty et al. (2005) Natta et al. (2004)

18. Most important conclusion: Keep an eye on them...

19. ... because you never know

20. Conclusions 1. Brown dwarfs show T Tauri like variability 1. Brown dwarfs show T Tauri like variability 2. monitoring allows close-up view on accretion behaviour 2. monitoring allows close-up view on accretion behaviour 3. strong accretion rate variations up to one order of 3. strong accretion rate variations up to one order of magnitudes on timescales of days to weeks magnitudes on timescales of days to weeks 4. profile asymmetry and variability: evidence for 4. profile asymmetry and variability: evidence for asymmetric flow (large-scale magnetic fields?) asymmetric flow (large-scale magnetic fields?)

21. Brown dwarfs in T Tauri phase Accretion flow 1st part: Variability Dusty disk 2nd part: mm/MIR SEDs

22. VLM rotation periods Scholz & Eislöffel: A&A, 2004, 419, 249 A&A, 2004, 421, 259 A&A, 2005, 429, 1007 PhD thesis A. Scholz 2003: 6 periods (squares) 2004: 80 periods (large dots)

23. Dusty disks constraints from MIR-SEDs: flared disk, flat disks, grain growth but only 1 object with SED from NIR to mm Mohanty et al. (2004) Pascucci et al. (2003)

24. Scholz & Eislöffel, A&A, 2004 Connection to disk accretion near-infrared colour excess, emission line spectrum  variability related to accretion from circumsubstellar disk

25. A 1.3mm survey in Taurus 20 sources with SpT>M6, noise level <1mJy for all objects IRAM 30m telescope with MAMBOII, Pico Veleta (Spain)

26. Fluxes and disk masses 20 sources, 6 detections, flux levels: <0.7... 7 mJy transformation to disk masses: <0.4... 2.4 Jupiter masses Scholz, Jayawardhana, Wood, ApJ, 2006

27. Disk mass vs. object mass Relative disk masses comparable from 0.02 to 3 Ms No trend to lower disk masses in the brown dwarf regime

28. Enter Spitzer IRAC+MIPS available for all Taurus sources: NIR (2MASS) + MIR (Spitzer) +mm (IRAM) IRAC photometry: 3-8 μm IRS spectroscopy: 8-13 μm MIPS photometry: 24 μm

29. SED modeling Minimum outer disk radius for objects with mm detection: 10 AU >25% of the disks have radii >10AU NO evidence for truncated disks Ejection excluded as dominant formation mechanism Scholz, Jayawardhana, Wood 2006

30. Evolution of brown dwarf disks Spitzer GO program to study 35 brown dwarf disks in Up Sco IRS spectroscopy + MIPS photometry Inner disks and chemistry after 5 Myr All observations finished

31. Young and old Taurus UpSco Taurus UpSco 2Myr - strongly flared disk 5Myr - dust settling more to come! more to come!

32. Origins of brown dwarfs ‘in situ’ formation - ultra-low-mass stars - ejection as embryos - failed stars - signature of formation: binarity, kinematics, accretion disks

33. Young stars and variability H  linewidths for stars in young associations (age 6-30 Myr) `errorbars` show scatter over multi-epoch observations   variability common phenomenon in young stars Jayawardhana et al., ApJ, in prep.

34. Case study: TWA5A close binary, at least one of the components is accreting Aa + Ab = one solar mass Brandeker et al. 2003

35. Hα variability of TWA5A both components contribute to „flare“ event - delay of broad component? profile decomposition: broad and narrow component dashed: broad dotted: narrow Jayawardhana, et al., ApJL, in prep.

36. Velocity variations comparable periods in both components possible scenario: rotation period of Aa or Ab  two co-rotating spots from accretion and activity broad: P = 19.6 h, FAP = 0.004% narrow: P = 19.2 h, FAP = 0.8% Jayawardhana, et al., ApJ, in prep.

37. Spot properties cool spots, either symmetric distribution or low spot coverage  indication for a change in the magnetic field generation Scholz, Eislöffel & Froebrich, 2005, A&A, 438, 675

38. Amplitudes vs. mass VLM objects: low amplitudes, low rate of active objects  change in spot properties Amplitudes in young open clusters

39. Period vs. Mass ONC: Herbst et al. (2002) VLM objects rotate faster than solar-mass stars average period correlated with mass Scholz & Eislöffel, A&A, 2004, 2005

40. Period vs. Mass I Pleiades (+ literature) IC4665 (+ literature) Pleiades (+ literature) IC4665 (+ literature) VLM objects rotate faster than solar-mass stars

41. Period vs. Mass II VLM regime: period decreases with mass Pleiades (+ Terndrup et al.) IC4665 Pleiades (+ Terndrup et al.) IC4665

42. Period vs. Mass III Median period decreases with mass, even at very young ages Ori + Herbst et al. (2001) Ori + Herbst et al. (2001) σOri + Herbst et al. (2001) εOri + Herbst et al. (2001)

43. The physics of VLM objects 0.35 M S objects are fully convective 0.15 M S degeneracy pressure dominates (radius independent of mass) (radius independent of mass) 0.075 M S no stable hydrogen burning (substellar limit) (substellar limit) 0.060 M S only deuterium burning 0.013 M S no deuterium burning

44. Interior structure fully convective VLM object solar-type star Consequences for magnetic fields, activity, rotation radiative zone

45. Rotation and stellar evolution ´Disk locking´ Stellar winds Bouvier et al. 1997

46. Stellar winds TRACE SOHO

47. 1Myr 10Myr 100Myr 1Myr 10Myr 100Myr 1Gyr σOri, ε Ori 3-10 Myr Scholz & Eislöffel, A&A, 2004, 419, 249 Scholz & Eislöffel, A&A, 2005, 429, 1007 IC4665 36 Myr Eislöffel & Scholz 2002, ESO-Conf. Pleiades 125 Myr Scholz & Eislöffel, A&A, 2004, 421, 259 The clusters Praesepe 700 Myr Time series imaging with TLS Schmidt, ESO/MPG WFI, Calar Alto

48. Lightcurves 90% of all variable objects: regular, periodic variability VLM star in the Pleiades Brown Dwarf in εOri VLM star in the Pleiades Brown Dwarf in εOri

49. Period vs. Mass II VLM regime: period decreases with mass Pleiades (+ Terndrup et al.) IC4665 Pleiades (+ Terndrup et al.) IC4665

50. Models  P(t) = α(t) (R(t)/R i ) 2 P i A) α(t) = const. = 1 only contraction B) α(t) = (t / t i ) ½ Skumanich law (dL/dt ~ω 3 ) C) α(t) = exp((t – t i ) /  )exponential braking (dL/dt ~ ω) Period evolution between 3 and 750 Myr determined by… - hydrostatic contraction - rotational braking by stellar winds - disk-locking (not important)

51. Surface features: Magnetic spots Amplitudes of variability determined by spot properties

52. Spot configuration How do the surfaces of VLM objects look like? Lamm (2003) Barnes & Collier Cameron (2001) b) Only polar spots c) Low spot coverage d) High symmetry e) Low contrast

53. Disks around VLM objects NIR colour excess Strong emission lines NIR colour excess Strong emission lines but: disk frequency only 5-15% in  Ori cluster but: disk frequency only 5-15% in  Ori cluster Colour-colour diagram Optical spectroscopy

54. Accretion vs. rotation Scholz & Eislffel 2004 Scholz & Eislöffel 2004 Basri, Mohanty & Jayawardhana, in prep.

55. Breakup period models not adequate for fastest rotators models not adequate for fastest rotators

56. Rotational evolution

57. Only contraction angular momentum loss necessary to explain slow rotators

58. Contraction + Skumanich Skumanich braking is too strong

59. Contraction + exponential braking best agreement of model and observations

60. Multi-filter monitoring simultaneous monitoring with two telescopes in I, J, H Calar Alto Observatory, 1.2m and 2.2m telescope

61. Magnetic field generation Fully convective objects: no interface layer  solar-type  ω- dynamo,  only small-scale magnetic fields? inefficient wind braking  fast rotation symmetric spot distribution  small amplitudes