Microarcsecond Astrometry in the Local Group and Beyond

Microarcsecond Astrometry in the Local Group and Beyond Andreas Brunthaler Max-Planck-Institut für Radioastronomie with M. J. Reid, K. M. Menten, H. Falcke, C. Henkel, G. C. Bower, …

The Local Group of Galaxies • about 35 -40 member galaxies • Milky Way and Andromeda subgroups

Dynamical models of the Local Group • dynamics needed to estimate mass distribution • only radial velocities of galaxies known • Local Group Timing (Kahn & Woltjer 1959) - Milky Way and Andromeda (M 31) have probably formed nearby and completed the larger part of at least one orbit. - Keplers law gives a mass of > 2 1012 Msol for the Local Group 12 assuming zero proper motion of M 31 - dark matter needed

Dynamical models of the Local Group • proper motions needed, but not easy: Photographic plates (50 years): 0. 56± 0. 25 mas/yr for Sculptor dwarf galaxy at 88 kpc (Schweitzer et al. 1995) HST: 0. 61 ± 0. 23 mas/yr for Fornax dwarf at 140 kpc (Dinescu et al. 2004) 0. 55 ± 0. 23 mas/yr for Ursa Minor at 63 kpc (Piatek et al. 2005) LMC (Kallivayalil et al. 2006) SMC (Kallivayalil et al. 2006) Canis Major (Dinescu et al. 2005) Sagittarius d. Sph (Dinescu et al. 2005) Carina d. Sph (Piatek et al. 2003) • limited to closest neighbours of Milky Way (< 150 kpc) impossible for M 31 subgroup Expected proper motions of few 10 as/yr detectable with VLBI

VLBI Astrometry • strong and compact radio sources needed 1. Radio cores • no strong cores in Local Group • S 8. 4 GHz(M 31*) = 28± 5 μJy (Crane et al. 1992) 2. strong maser emission • H 2 O masers known in M 33 and IC 10 • M 33 and IC 10 in Andromeda subgroup (~ 800 kpc)

VLBI Astrometry Δτ0 τ0 Z } τ0 sec(Z 1) τ0 sec(Z 2) Dominant source of error: - inaccurate zenith delay τ0 at each antenna in correlator-model - different phase-errors for different sources Δφ1= 2πν·Δτ0 sec(Z 1) degrade image quality Δφ2= 2πν·Δτ0 sec(Z 2)

VLBI Astrometry • Include geodetic-like observations - many quasars at different elevations - 16 IFs spanning 450 MHz • Allows determination of atmospheric zenith delay error • CLCOR to correct for this error • Improved image quality

VLBI Astrometry • Include geodetic-like observations - many quasars at different elevations - 16 IFs spanning 450 MHz • Allows determination of atmospheric zenith delay error • CLCOR to correct for this error • Improved image quality • GPS gives similar results

Observations of M 33 • observations with VLBA • M 33 - two maser regions - reference QSO 1º away - check source 14’ away 6 epochs between 03/2001 and 06/2005 IC 133 M 33/19

Rotation of M 33 Relative motion between IC 133 and M 33/19 depends only on rotation Assume vpec < 20 km/s (radial velocities of maser and H I agree) Rotation model of HI: Relative proper motions (Corbelli & Schneider 1997) In RA: 106. 4 20 km/s 29. 9 2 as/yr In DEC: 35. 0 20 km/s 9. 7 5 as/yr v = D D = 750 140 50 kpc

Distance of M 33 Reference Method Distance [kpc] Lee et al. 2002 Cepheids (revised) 802 51 Kim et al. 2002 TRGB + RC 916 55 Mc. Connachie et al. 2005 TRGB 809 24 This work 750 140 50 rotational parallax • Accuracy of < 10 % possible - better rotation model (higher resolution H I data) - new observations and third maser in M 33

Space motion of M 33 • Observed motion of IC 133: vobs = vrot + vprop • Internal rotation of M 33: - assume vpec < 20 km/s (radial velocities agree) • Reflex of Solar motion: - measured proper motion of Sgr A* 6. 379 0. 024 mas/yr (Reid & Brunthaler 2004) - peculiar motion of Sun from Hipparcos: (10. 0, 5. 25, 7. 17) km/s (Dehnen & Binney 1998)

Space motion of M 33 • Proper motion: vprop= vobs-vrot-v RA: vprop = (4. 7 3) – (-18. 5 6) – (53 3) as/yr = -29. 3 7. 6 as/yr = -101 35 km/s DEC: vprop = (-14 6) – (-22 6) – (-38 2. 4) as/yr = 45. 2 9. 1 as/yr = 159 47 km/s vrad = (-179 2) + (140 9) km/s = -39 9 km/s

Space motion of M 33 (Brunthaler et al. 2006)

Observations of IC 10 • observations with VLBA • IC 10 - one maser region - reference QSO 1º away - check source 8’ away 6 epochs between 02/2001 and 06/2005 1 degree

Observations of IC 10 A) ‘core-shift’ in calibrator ‘moves’ both sources - motion similar for maser and check source B) residual geometric errors in correlator model - motion similar for maser and check source C) ‘core-shift’ in check source - motion different for maser and check source Look at relative motion between IC 10 and check source

Space motion of IC 10 • Proper motion: vprop= vobs-vrot-v RA: vprop = (6 5) – (8 6) – (37 4) as/yr = -39 9 as/yr = -122 31 km/s DEC: vprop = (23 5) – (3 6) – (-11 1) as/yr = 31 8 as/yr = 97 27 km/s vrad = (-344 3) + (196 10) km/s = -148 10 km/s

Space motion of M 33 & IC 10 (Brunthaler et al. 2006)

Space motion of M 31 Separation between M 33 and M 31 • M 31’s proper motion important - n-body simulations of history of M 33/M 31/MW system - assume a trial proper motion for M 31 1. Calculate orbits 10 Gyr backwards in time 2. Run n-body simulations forward - simulate M 33 stellar disk with 100 tracer stars (Loeb et al. 2005)

Space motion of M 31 • M 31’s proper motion important - n-body simulations of history of M 33/M 31/MW system - assume a trial proper motion for M 31 1. Calculate orbits 10 Gyr backwards in time 2. Run n-body simulations forward - simulate M 33 stellar disk with 100 tracer stars 3. Calculate fraction of tidally stripped stars (from M 33) (Loeb et al. 2005)

Space motion of M 31 • M 31’s proper motion important 50% - n-body simulations of history of M 33/M 31/MW system 20% - assume a trial proper motion for M 31 1. Calculate orbits 10 Gyr backwards in time 2. Run n-body simulations forward - simulate M 33 stellar disk with 100 tracer stars 3. Calculate fraction of tidally stripped stars (from M 33) 4. exclude proper motions of M 31 with a large amount of stripped stars (Loeb et al. 2005)

Space motion of M 31 • M 31’s proper motion important - n-body simulations of history of M 33/M 31/MW system - assume a trial proper motion for M 31 1. Calculate orbits 10 Gyr backwards in time 2. Run n-body simulations forward - simulate M 33 stellar disk with 100 tracer stars 3. Calculate fraction of tidally stripped stars (from M 33) 4. exclude proper motions of M 31 with a large amount of stripped stars => v 80 km/s prop (Loeb et al. 2005)

The mass of M 31 • If M 33 or IC 10 are bound to M 31, then Vrel < Vesc or MM 31 > (V 2 rel R)/2 G IC 10: Vrel = 147 km/s; R= 262 kpc => MM 31 > 6. 6 1011 Msol M 33: Vrel = 230 km/s; R= 202 kpc => MM 31 > 12 1011 Msol (for zero proper motion of M 31) Lower limits: IC 10 M 33 (0. 7, 1, 2. 5, 5, 10, 25) x 1011 Msol (4, 5, 7. 5, 10, 25) x 1011 Msol

The mass of M 31 Combination of lower limits and excluded regions gives a lower limit 7. 5 1011 Msol 2 11 11

Outlook • The proper motion of M 31 2 - HSA observation of M 31* (512 Mbps) last Sunday - possible in beam calibrator - proper motion will be measured within a few years - If not now, then with 4 Gbps! • Water masers in the LMC - observations using LBA started last year - accuracy less than VLBA, but motions much larger 11 11

The Local Group is not enough… • Accuracy of VLBA is good enough for nearby galaxy groups • M 81 group is ideal laboratory to study galaxy interaction (Yun et al. 1994, Nature)

The Local Group is not enough… • Accuracy of VLBA is good enough for nearby galaxy groups • M 81 group is ideal laboratory to study galaxy interaction - M 81* and three water masers in M 82 - three additional background quasars within 2° - HSA observations started January 2007

SN 2008 iz • A new radio supernova in M 82 (SN 2008 iz) • It exploded in February 2008 (or later) • It expands with ~13, 000 km/s • spectral index = -1. 08, probably type II • Radio lightcurve recovered from Urumqi monitoring • So far, no detection at other wavebands • Possibly another RSN this year! • more to come….

The Local Group is not enough… • Accuracy of VLBA is good enough for nearby galaxy groups • M 81 group is ideal laboratory to study galaxy interaction - M 81* and three water masers in M 82 - three additional background quasars within 2° - HSA observations started January 2007 • Also: Water masers in NGC 253 (Brunthaler et al. 2008, A&A)

The Local Group is not enough… • Virgo Cluster (16. 8 Mpc) - relative motions > 1000 km/s - possible in 5 -10 years (van Gorkom)

The Local Group is not enough… • Virgo Cluster (16. 8 Mpc) - relative motions > 1000 km/s - possible in 5 -10 years - evolution of galaxy in clusters - detailed study of interactions - e. g. ram pressure stripping (Oosterloo & van Gorkom 2005, A&A)